Menu
▾
▴
octave-cvsupdate
|
From: <par...@us...> - 2011-07-27 15:51:56
|
Revision: 8417
http://octave.svn.sourceforge.net/octave/?rev=8417&view=rev
Author: paramaniac
Date: 2011-07-27 15:51:50 +0000 (Wed, 27 Jul 2011)
Log Message:
-----------
control: prepare release of control-2.1.52
Modified Paths:
--------------
trunk/octave-forge/main/control/DESCRIPTION
trunk/octave-forge/main/control/devel/RELEASE_PACKAGE
trunk/octave-forge/main/control/doc/NEWS
Modified: trunk/octave-forge/main/control/DESCRIPTION
===================================================================
--- trunk/octave-forge/main/control/DESCRIPTION 2011-07-27 13:12:27 UTC (rev 8416)
+++ trunk/octave-forge/main/control/DESCRIPTION 2011-07-27 15:51:50 UTC (rev 8417)
@@ -1,6 +1,6 @@
Name: Control
-Version: 2.1.51
-Date: 2011-07-21
+Version: 2.1.52
+Date: 2011-07-27
Author: Lukas Reichlin <luk...@gm...>
Maintainer: Lukas Reichlin <luk...@gm...>
Title: Control Systems
Modified: trunk/octave-forge/main/control/devel/RELEASE_PACKAGE
===================================================================
--- trunk/octave-forge/main/control/devel/RELEASE_PACKAGE 2011-07-27 13:12:27 UTC (rev 8416)
+++ trunk/octave-forge/main/control/devel/RELEASE_PACKAGE 2011-07-27 15:51:50 UTC (rev 8417)
@@ -18,12 +18,12 @@
rm -R ~/octave/__TEMP__/control/devel
cd ~/octave/__TEMP__
grep -i version control/DESCRIPTION
-tar czf control-2.1.51.tar.gz control/
-md5 control-2.1.51.tar.gz
-md5 control-2.1.51.tar.gz > md5_control_pkg.txt
-uuencode control-2.1.51.tar.gz < control-2.1.51.tar.gz > control-2.1.51.tar.gz.uue
+tar czf control-2.1.52.tar.gz control/
+md5 control-2.1.52.tar.gz
+md5 control-2.1.52.tar.gz > md5_control_pkg.txt
+uuencode control-2.1.52.tar.gz < control-2.1.52.tar.gz > control-2.1.52.tar.gz.uue
octave -q --eval \
-"pkg install control-2.1.51.tar.gz"
+"pkg install control-2.1.52.tar.gz"
octave -q --eval \
"pkg load generate_html; generate_package_html ('control', 'control-html', 'octave-forge')"
tar czf control-html.tar.gz control-html
@@ -38,7 +38,7 @@
=====================================================================================
rm -R ~/octave/__TEMP__
-rm -R ~/octave/control-2.1.51
+rm -R ~/octave/control-2.1.52
=====================================================================================
Modified: trunk/octave-forge/main/control/doc/NEWS
===================================================================
--- trunk/octave-forge/main/control/doc/NEWS 2011-07-27 13:12:27 UTC (rev 8416)
+++ trunk/octave-forge/main/control/doc/NEWS 2011-07-27 15:51:50 UTC (rev 8417)
@@ -1,7 +1,7 @@
Summary of important user-visible changes for releases of the control package
===============================================================================
-control-2.x.yy Release Date: 2011-xx-yy Release Manager: Lukas Reichlin
+control-2.1.52 Release Date: 2011-07-27 Release Manager: Lukas Reichlin
===============================================================================
** hsvd.m
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-08-04 19:18:47
|
Revision: 8431
http://octave.svn.sourceforge.net/octave/?rev=8431&view=rev
Author: paramaniac
Date: 2011-08-04 19:18:41 +0000 (Thu, 04 Aug 2011)
Log Message:
-----------
control: various changes
Modified Paths:
--------------
trunk/octave-forge/main/control/devel/MDSSystem.m
trunk/octave-forge/main/control/devel/PROJECTS
trunk/octave-forge/main/control/devel/ncfsyn/ncfsyn.m
trunk/octave-forge/main/control/inst/@frd/display.m
Modified: trunk/octave-forge/main/control/devel/MDSSystem.m
===================================================================
--- trunk/octave-forge/main/control/devel/MDSSystem.m 2011-08-03 06:58:20 UTC (rev 8430)
+++ trunk/octave-forge/main/control/devel/MDSSystem.m 2011-08-04 19:18:41 UTC (rev 8431)
@@ -142,6 +142,7 @@
% Compute the suboptimal positive feedback controller
K = ncfsyn (G, W1, W2, factor); % positive feedback controller
+[K, N, gamma, info] = ncfsyn (G, W1, W2, factor)
K = -K; % negative feedback controller
L = G * K; % open loop
Modified: trunk/octave-forge/main/control/devel/PROJECTS
===================================================================
--- trunk/octave-forge/main/control/devel/PROJECTS 2011-08-03 06:58:20 UTC (rev 8430)
+++ trunk/octave-forge/main/control/devel/PROJECTS 2011-08-04 19:18:41 UTC (rev 8431)
@@ -37,6 +37,16 @@
* Check whether RASP-DESCRIPT Fortran routines are free software and include them
for dss system conversions and other stuff if possible.
+ * Add input and output groups to LTI models.
+ -- sys.ingroup.controls = [1, 3] (stored as a struct)
+ -- sys(:, "controls") (instead of sys(:, [1, 3]))
+ -- Handle __sys_group__ case when indices of the second lti need to be increased
+ by the size of the first lti.
+ -- Handle __sys_prune__ case when indices need to be deleted from all struct fields
+ and channels numbers after the deleted ones need to be decreased.
+ -- Handle conflicting names for __sys_group__.
+ -- Beware of indices not in ascending order.
+
-----------------------
Frequency Response Data:
-----------------------
Modified: trunk/octave-forge/main/control/devel/ncfsyn/ncfsyn.m
===================================================================
--- trunk/octave-forge/main/control/devel/ncfsyn/ncfsyn.m 2011-08-03 06:58:20 UTC (rev 8430)
+++ trunk/octave-forge/main/control/devel/ncfsyn/ncfsyn.m 2011-08-04 19:18:41 UTC (rev 8431)
@@ -16,7 +16,7 @@
## along with LTI Syncope. If not, see <http://www.gnu.org/licenses/>.
## -*- texinfo -*-
-## @deftypefn{Function File} {[@var{K}, @var{N}, @var{gamma}] =} ncfsyn (@var{G}, @var{W1}, @var{W2}, @var{factor})
+## @deftypefn{Function File} {[@var{K}, @var{N}, @var{gamma}, @var{info}] =} ncfsyn (@var{G}, @var{W1}, @var{W2}, @var{factor})
## Normalized Coprime Factor (NCF) H-infinity synthesis.
## Compute positive feedback controller using the McFarlane/Glover Loop Shaping Design Procedure.
## Uses SLICOT SB10ID, SB10KD and SB10ZD by courtesy of
@@ -46,9 +46,19 @@
## @item K
## State-space model of the H-infinity loop-shaping controller.
## @item N
-## State-space model of the lower LFT of @var{P} and @var{K}.
+## State-space model of closed loop.
## @item gamma
## L-infinity norm of @var{N}.
+## @item info
+## Structure containing additional information.
+## @item info.emax
+## Nugap robustness. @code{emax = inv (gamma)}.
+## @item info.Gs
+## Shaped plant. @code{Gs = W2 * G * W1}.
+## @item info.Ks
+## Controller for shaped plant. @code{Ks = ncfsyn (Gs)}.
+## @item info.rcond
+## Estimates of the reciprocal condition numbers of the Riccati equations.
## @end table
## @end deftypefn
@@ -56,8 +66,7 @@
## Created: July 2011
## Version: 0.1
-% function [K, N, gamma, info] = ncfsyn (G, W1 = [], W2 = [], factor = 1.0)
-function K = ncfsyn (G, W1 = [], W2 = [], factor = 1.0)
+function [K, varargout] = ncfsyn (G, W1 = [], W2 = [], factor = 1.0)
if (nargin == 0 || nargin > 4)
print_usage ();
@@ -93,9 +102,25 @@
Ks = ss (ak, bk, ck, dk, tsam);
K = W1 * Ks * W2;
-
- %struct ("Gs", Gs, "Ks", Ks, "rcond", rcond);
+ if (nargout > 1)
+ N = append (eye (p), K, G);
+ M = [zeros(p,p), zeros(p,m), eye(p);
+ eye(p), zeros(p,m), zeros(p,p);
+ zeros(m,p), eye(m), zeros(m,p)];
+ in_idx = [1:p, 2*p+(1:m)];
+ out_idx = 1:p+m;
+ N = mconnect (N, M, in_idx, out_idx);
+ varargout{1} = N;
+ if (nargout > 2)
+ gamma = norm (N, inf);
+ varargout{2} = gamma;
+ if (nargout > 3)
+ varargout{3} = struct ("emax", inv (gamma), "Gs", Gs, "Ks", Ks, "rcond", rcond);
+ endif
+ endif
+ endif
+
endfunction
Modified: trunk/octave-forge/main/control/inst/@frd/display.m
===================================================================
--- trunk/octave-forge/main/control/inst/@frd/display.m 2011-08-03 06:58:20 UTC (rev 8430)
+++ trunk/octave-forge/main/control/inst/@frd/display.m 2011-08-04 19:18:41 UTC (rev 8431)
@@ -118,22 +118,22 @@
endfunction
-function col = __vec2str__ (vec, post)
+function str = __vec2str__ (vec, post)
vec = vec(:);
tmp = isfinite (vec);
tmp = abs (vec(tmp & vec != 0));
vec = num2cell (vec);
if (isempty (tmp) || min (tmp) < 1e-3 || max (tmp) > 1e4)
- col = cellfun (@(x) sprintf ("%.3e", x), vec, "uniformoutput", false);
+ str = cellfun (@(x) sprintf ("%.3e", x), vec, "uniformoutput", false);
elseif (all (floor (tmp) == tmp))
- col = cellfun (@(x) sprintf ("%d", x), vec, "uniformoutput", false);
+ str = cellfun (@(x) sprintf ("%d", x), vec, "uniformoutput", false);
else
- col = cellfun (@(x) sprintf ("%.4f", x), vec, "uniformoutput", false);
+ str = cellfun (@(x) sprintf ("%.4f", x), vec, "uniformoutput", false);
endif
- col = strjust (char (col), "right");
+ str = strjust (char (str), "right");
if (nargin > 1)
- col = [col, repmat(post, length (vec), 1)];
+ str = [str, repmat(post, length (vec), 1)];
endif
endfunction
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-08-05 12:34:21
|
Revision: 8435
http://octave.svn.sourceforge.net/octave/?rev=8435&view=rev
Author: paramaniac
Date: 2011-08-05 12:34:08 +0000 (Fri, 05 Aug 2011)
Log Message:
-----------
control: clean up ncfsyn
Modified Paths:
--------------
trunk/octave-forge/main/control/INDEX
trunk/octave-forge/main/control/devel/makefile_all.m
trunk/octave-forge/main/control/devel/makefile_ncfsyn.m
trunk/octave-forge/main/control/doc/NEWS
Added Paths:
-----------
trunk/octave-forge/main/control/inst/MDSSystem.m
Removed Paths:
-------------
trunk/octave-forge/main/control/devel/MDSSystem.m
trunk/octave-forge/main/control/devel/ncfsyn/
Modified: trunk/octave-forge/main/control/INDEX
===================================================================
--- trunk/octave-forge/main/control/INDEX 2011-08-05 12:14:15 UTC (rev 8434)
+++ trunk/octave-forge/main/control/INDEX 2011-08-05 12:34:08 UTC (rev 8435)
@@ -80,6 +80,7 @@
h2syn
hinfsyn
mixsyn
+ ncfsyn
Matrix Equation Solvers
care
dare
Deleted: trunk/octave-forge/main/control/devel/MDSSystem.m
===================================================================
--- trunk/octave-forge/main/control/devel/MDSSystem.m 2011-08-05 12:14:15 UTC (rev 8434)
+++ trunk/octave-forge/main/control/devel/MDSSystem.m 2011-08-05 12:34:08 UTC (rev 8435)
@@ -1,156 +0,0 @@
-% ===============================================================================
-% Robust Control of a Mass-Damper-Spring System Lukas Reichlin August 2011
-% ===============================================================================
-% Reference: Gu, D.W., Petkov, P.Hr. and Konstantinov, M.M.
-% Robust Control Design with Matlab, Springer 2005
-% ===============================================================================
-
-% Tabula Rasa
-clear all, close all, clc
-
-% ===============================================================================
-% System Model
-% ===============================================================================
-% +---------------+
-% | d_m 0 0 |
-% +-----| 0 d_c 0 |<----+
-% u_m | | 0 0 d_k | | y_m
-% u_c | +---------------+ | y_c
-% u_k | | y_k
-% | +---------------+ |
-% +---->| |-----+
-% | G_nom |
-% u ----->| |-----> y
-% +---------------+
-
-% Nominal Values
-m_nom = 3; % mass
-c_nom = 1; % damping coefficient
-k_nom = 2; % spring stiffness
-
-% Perturbations
-p_m = 0.4; % 40% uncertainty in the mass
-p_c = 0.2; % 20% uncertainty in the damping coefficient
-p_k = 0.3; % 30% uncertainty in the spring stiffness
-
-% State-Space Representation
-A = [ 0, 1
- -k_nom/m_nom, -c_nom/m_nom ];
-
-B1 = [ 0, 0, 0
- -p_m, -p_c/m_nom, -p_k/m_nom ];
-
-B2 = [ 0
- 1/m_nom ];
-
-C1 = [ -k_nom/m_nom, -c_nom/m_nom
- 0, c_nom
- k_nom, 0 ];
-
-C2 = [ 1, 0 ];
-
-D11 = [ -p_m, -p_c/m_nom, -p_k/m_nom
- 0, 0, 0
- 0, 0, 0 ];
-
-D12 = [ 1/m_nom
- 0
- 0 ];
-
-D21 = [ 0, 0, 0 ];
-
-D22 = [ 0 ];
-
-inname = {'u_m', 'u_c', 'u_k', 'u'}; % input names
-outname = {'y_m', 'y_c', 'y_k', 'y'}; % output names
-
-G_nom = ss (A, [B1, B2], [C1; C2], [D11, D12; D21, D22], ...
- 'inputname', inname, 'outputname', outname);
-
-G = G_nom(4, 4); % extract output y and input u
-
-
-% ===============================================================================
-% Frequency Analysis of Uncertain System
-% ===============================================================================
-
-% Uncertainties: -1 <= delta_m, delta_c, delta_k <= 1
-[delta_m, delta_c, delta_k] = ndgrid ([-1, 0, 1], [-1, 0, 1], [-1, 0, 1]);
-
-% Bode Plots of Perturbed Plants
-w = logspace (-1, 1, 100); % frequency vector
-figure (1)
-
-for k = 1 : numel (delta_m)
- Delta = diag ([delta_m(k), delta_c(k), delta_k(k)]);
- G_per = lft (Delta, G_nom);
- bode (G_per, w)
- subplot (2, 1, 1)
- hold on
- subplot (2, 1, 2)
- hold on
-endfor
-
-
-% ===============================================================================
-% Mixed Sensitivity H-infinity Controller Design (S over KS Method)
-% ===============================================================================
-% +-------+
-% +--------------------->| W_p |----------> e_p
-% | +-------+
-% | +-------+
-% | +---->| W_u |----------> e_u
-% | | +-------+
-% | | +---------+
-% | | ->| |->
-% r + e | +-------+ u | | G_nom |
-% ----->(+)---+-->| K |----+--->| |----+----> y
-% ^ - +-------+ +---------+ |
-% | |
-% +-----------------------------------------+
-
-% Weighting Functions
-s = tf ('s'); % transfer function variable
-W_p = 0.95 * (s^2 + 1.8*s + 10) / (s^2 + 8.0*s + 0.01); % performance weighting
-W_u = 10^-2; % control weighting
-
-% Synthesis
-K = mixsyn (G, W_p, W_u); % mixed-sensitivity H-infinity synthesis
-
-% Interconnections
-L = G * K; % open loop
-T = feedback (L); % closed loop
-
-% Plotting
-figure (2)
-sigma (T) % singular values
-
-figure (3)
-step (T) % step response
-
-
-% ===============================================================================
-% H-infinity Loop-Shaping Design
-% ===============================================================================
-
-% Settings
-W1 = 8 * (2*s + 1) / (0.9*s); % precompensator
-W2 = 1; % postcompensator
-factor = 1.1; % suboptimal controller
-
-% Synthesis
-K = ncfsyn (G, W1, W2, factor); % positive feedback controller
-
-% Interconnections
-K = -K; % negative feedback controller
-L = G * K; % open loop
-T = feedback (L); % closed loop
-
-% Plotting
-figure (4)
-sigma (T) % singular values
-
-figure (5)
-step (T) % step response
-
-% ===============================================================================
Modified: trunk/octave-forge/main/control/devel/makefile_all.m
===================================================================
--- trunk/octave-forge/main/control/devel/makefile_all.m 2011-08-05 12:14:15 UTC (rev 8434)
+++ trunk/octave-forge/main/control/devel/makefile_all.m 2011-08-05 12:34:08 UTC (rev 8435)
@@ -19,6 +19,7 @@
makefile_lqr
makefile_lyap
makefile_minreal
+makefile_ncfsyn
makefile_norm
makefile_place
makefile_scale
Modified: trunk/octave-forge/main/control/devel/makefile_ncfsyn.m
===================================================================
--- trunk/octave-forge/main/control/devel/makefile_ncfsyn.m 2011-08-05 12:14:15 UTC (rev 8434)
+++ trunk/octave-forge/main/control/devel/makefile_ncfsyn.m 2011-08-05 12:34:08 UTC (rev 8435)
@@ -1,3 +1,19 @@
+## ==============================================================================
+## Developer Makefile for OCT-files
+## ==============================================================================
+## USAGE: * fetch control from Octave-Forge by svn
+## * add control/inst, control/src and control/devel to your Octave path
+## * run makefile_*
+## NOTES: * The option "-Wl,-framework" "-Wl,vecLib" is needed for MacPorts'
+## octave-devel @3.3.52_1+gcc44 on MacOS X 10.6.4. However, this option
+## breaks other platforms. See MacPorts Ticket #26640.
+## ==============================================================================
+
+homedir = pwd ();
+develdir = fileparts (which ("makefile_ncfsyn"));
+srcdir = [develdir, "/../src"];
+cd (srcdir);
+
## H-infinity loop shaping - continuous-time
mkoctfile "-Wl,-framework" "-Wl,vecLib" \
slsb10id.cc \
@@ -19,4 +35,6 @@
slsb10zd.cc \
SB10ZD.f MA02AD.f SB02OD.f select.f MB01RX.f \
MB02VD.f SB02OY.f SB02OW.f SB02OV.f SB02OU.f \
- SB02MR.f MA02GD.f SB02MV.f
\ No newline at end of file
+ SB02MR.f MA02GD.f SB02MV.f
+
+cd (homedir);
Modified: trunk/octave-forge/main/control/doc/NEWS
===================================================================
--- trunk/octave-forge/main/control/doc/NEWS 2011-08-05 12:14:15 UTC (rev 8434)
+++ trunk/octave-forge/main/control/doc/NEWS 2011-08-05 12:34:08 UTC (rev 8435)
@@ -1,17 +1,29 @@
Summary of important user-visible changes for releases of the control package
===============================================================================
+control-2.x.yy Release Date: 2011-xx-yy Release Manager: Lukas Reichlin
+===============================================================================
+
+** ncfsyn
+ -- Added support for McFarlane/Glover loop shaping design procedure. ncfsyn
+ stands for Normalized Coprime Factor Synthesis.
+
+** MDSSystem
+ -- Added example script which demonstrates the usage of the robust control
+ commands "mixsyn" and "ncfsyn".
+
+===============================================================================
control-2.1.52 Release Date: 2011-07-27 Release Manager: Lukas Reichlin
===============================================================================
-** hsvd.m
+** hsvd
-- Use scaling unless state-space model property "scaled" is set to true.
-** @lti/norm.m
+** @lti/norm
-- Use scaling for computation of L-infinity norm unless state-space model
property "scaled" is set to true.
-** @lti/minreal.m
+** @lti/minreal
-- Use scaling for state-space and descriptor state-space models unless
property "scaled" is set to true.
-- More accurate results are to be expected for descriptor state-space
@@ -19,11 +31,11 @@
reduction occurs. This is achieved by saving the system matrices before
each phase and restoring them if no order reduction took place.
-** @lti/zero.m
+** @lti/zero
-- Use scaling for state-space and descriptor state-space models unless
property "scaled" is set to true.
-** @lti/frdata.m
+** @lti/frdata
-- The frequency response is now returned correctly as an array and not as a
vector, unless the "vector" option is set and the system is single-input
single-output.
@@ -34,7 +46,7 @@
control-2.1.51 Release Date: 2011-07-21 Release Manager: Lukas Reichlin
===============================================================================
-** frd.m
+** frd
-- Support for Frequency Response Data (frd) measurement "models".
@@ -42,14 +54,14 @@
control-2.1.50 Release Date: 2011-07-06 Release Manager: Lukas Reichlin
===============================================================================
-** ss.m
+** ss
-- Support for property "scaled". By default, it is set to "false".
-** @lti/prescale.m
+** @lti/prescale
-- Scaling for state-space models (SLICOT TB01ID) and descriptor models
(SLICOT TG01AD).
-** @lti/freqresp.m
+** @lti/freqresp
-- Scale state-space models using @lti/prescale.m if property "scaled" is
set to "false". Frequency response commands now perform automatic
scaling unless model property "scaled" is set to "true".
@@ -59,7 +71,7 @@
control-2.0.2 Release Date: 2011-03-18 Release Manager: Lukas Reichlin
===============================================================================
-** lsim.m
+** lsim
-- Fixed a logical error that refused valid initial state vectors. It was
due to a thinko introduced with the changes in control-2.0.1.
(Thanks to Rob Frohne)
@@ -69,11 +81,11 @@
control-2.0.1 Release Date: 2011-03-06 Release Manager: Lukas Reichlin
===============================================================================
-** lsim.m
+** lsim
-- Support time vectors not starting at zero. (Thanks to Rob Frohne)
-- Improved help text.
-** @lti/zero.m
+** @lti/zero
-- The gain of descriptor state-space models is now computed correctly.
(fingers crossed)
Copied: trunk/octave-forge/main/control/inst/MDSSystem.m (from rev 8434, trunk/octave-forge/main/control/devel/MDSSystem.m)
===================================================================
--- trunk/octave-forge/main/control/inst/MDSSystem.m (rev 0)
+++ trunk/octave-forge/main/control/inst/MDSSystem.m 2011-08-05 12:34:08 UTC (rev 8435)
@@ -0,0 +1,156 @@
+% ===============================================================================
+% Robust Control of a Mass-Damper-Spring System Lukas Reichlin August 2011
+% ===============================================================================
+% Reference: Gu, D.W., Petkov, P.Hr. and Konstantinov, M.M.
+% Robust Control Design with Matlab, Springer 2005
+% ===============================================================================
+
+% Tabula Rasa
+clear all, close all, clc
+
+% ===============================================================================
+% System Model
+% ===============================================================================
+% +---------------+
+% | d_m 0 0 |
+% +-----| 0 d_c 0 |<----+
+% u_m | | 0 0 d_k | | y_m
+% u_c | +---------------+ | y_c
+% u_k | | y_k
+% | +---------------+ |
+% +---->| |-----+
+% | G_nom |
+% u ----->| |-----> y
+% +---------------+
+
+% Nominal Values
+m_nom = 3; % mass
+c_nom = 1; % damping coefficient
+k_nom = 2; % spring stiffness
+
+% Perturbations
+p_m = 0.4; % 40% uncertainty in the mass
+p_c = 0.2; % 20% uncertainty in the damping coefficient
+p_k = 0.3; % 30% uncertainty in the spring stiffness
+
+% State-Space Representation
+A = [ 0, 1
+ -k_nom/m_nom, -c_nom/m_nom ];
+
+B1 = [ 0, 0, 0
+ -p_m, -p_c/m_nom, -p_k/m_nom ];
+
+B2 = [ 0
+ 1/m_nom ];
+
+C1 = [ -k_nom/m_nom, -c_nom/m_nom
+ 0, c_nom
+ k_nom, 0 ];
+
+C2 = [ 1, 0 ];
+
+D11 = [ -p_m, -p_c/m_nom, -p_k/m_nom
+ 0, 0, 0
+ 0, 0, 0 ];
+
+D12 = [ 1/m_nom
+ 0
+ 0 ];
+
+D21 = [ 0, 0, 0 ];
+
+D22 = [ 0 ];
+
+inname = {'u_m', 'u_c', 'u_k', 'u'}; % input names
+outname = {'y_m', 'y_c', 'y_k', 'y'}; % output names
+
+G_nom = ss (A, [B1, B2], [C1; C2], [D11, D12; D21, D22], ...
+ 'inputname', inname, 'outputname', outname);
+
+G = G_nom(4, 4); % extract output y and input u
+
+
+% ===============================================================================
+% Frequency Analysis of Uncertain System
+% ===============================================================================
+
+% Uncertainties: -1 <= delta_m, delta_c, delta_k <= 1
+[delta_m, delta_c, delta_k] = ndgrid ([-1, 0, 1], [-1, 0, 1], [-1, 0, 1]);
+
+% Bode Plots of Perturbed Plants
+w = logspace (-1, 1, 100); % frequency vector
+figure (1)
+
+for k = 1 : numel (delta_m)
+ Delta = diag ([delta_m(k), delta_c(k), delta_k(k)]);
+ G_per = lft (Delta, G_nom);
+ bode (G_per, w)
+ subplot (2, 1, 1)
+ hold on
+ subplot (2, 1, 2)
+ hold on
+endfor
+
+
+% ===============================================================================
+% Mixed Sensitivity H-infinity Controller Design (S over KS Method)
+% ===============================================================================
+% +-------+
+% +--------------------->| W_p |----------> e_p
+% | +-------+
+% | +-------+
+% | +---->| W_u |----------> e_u
+% | | +-------+
+% | | +---------+
+% | | ->| |->
+% r + e | +-------+ u | | G_nom |
+% ----->(+)---+-->| K |----+--->| |----+----> y
+% ^ - +-------+ +---------+ |
+% | |
+% +-----------------------------------------+
+
+% Weighting Functions
+s = tf ('s'); % transfer function variable
+W_p = 0.95 * (s^2 + 1.8*s + 10) / (s^2 + 8.0*s + 0.01); % performance weighting
+W_u = 10^-2; % control weighting
+
+% Synthesis
+K = mixsyn (G, W_p, W_u); % mixed-sensitivity H-infinity synthesis
+
+% Interconnections
+L = G * K; % open loop
+T = feedback (L); % closed loop
+
+% Plotting
+figure (2)
+sigma (T) % singular values
+
+figure (3)
+step (T) % step response
+
+
+% ===============================================================================
+% H-infinity Loop-Shaping Design
+% ===============================================================================
+
+% Settings
+W1 = 8 * (2*s + 1) / (0.9*s); % precompensator
+W2 = 1; % postcompensator
+factor = 1.1; % suboptimal controller
+
+% Synthesis
+K = ncfsyn (G, W1, W2, factor); % positive feedback controller
+
+% Interconnections
+K = -K; % negative feedback controller
+L = G * K; % open loop
+T = feedback (L); % closed loop
+
+% Plotting
+figure (4)
+sigma (T) % singular values
+
+figure (5)
+step (T) % step response
+
+% ===============================================================================
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-08-05 12:14:26
|
Revision: 8434
http://octave.svn.sourceforge.net/octave/?rev=8434&view=rev
Author: paramaniac
Date: 2011-08-05 12:14:15 +0000 (Fri, 05 Aug 2011)
Log Message:
-----------
control: move ncfsyn into place
Modified Paths:
--------------
trunk/octave-forge/main/control/inst/test_control.m
trunk/octave-forge/main/control/src/Makefile
Added Paths:
-----------
trunk/octave-forge/main/control/devel/makefile_ncfsyn.m
trunk/octave-forge/main/control/inst/ncfsyn.m
trunk/octave-forge/main/control/src/SB10ID.f
trunk/octave-forge/main/control/src/SB10JD.f
trunk/octave-forge/main/control/src/SB10KD.f
trunk/octave-forge/main/control/src/SB10ZD.f
trunk/octave-forge/main/control/src/slsb10id.cc
trunk/octave-forge/main/control/src/slsb10kd.cc
trunk/octave-forge/main/control/src/slsb10zd.cc
Removed Paths:
-------------
trunk/octave-forge/main/control/devel/ncfsyn/SB10ID.dat
trunk/octave-forge/main/control/devel/ncfsyn/SB10ID.f
trunk/octave-forge/main/control/devel/ncfsyn/SB10ID.res
trunk/octave-forge/main/control/devel/ncfsyn/SB10JD.f
trunk/octave-forge/main/control/devel/ncfsyn/SB10KD.dat
trunk/octave-forge/main/control/devel/ncfsyn/SB10KD.f
trunk/octave-forge/main/control/devel/ncfsyn/SB10KD.res
trunk/octave-forge/main/control/devel/ncfsyn/SB10ZD.dat
trunk/octave-forge/main/control/devel/ncfsyn/SB10ZD.f
trunk/octave-forge/main/control/devel/ncfsyn/SB10ZD.res
trunk/octave-forge/main/control/devel/ncfsyn/common.cc
trunk/octave-forge/main/control/devel/ncfsyn/makefile_ncfsyn.m
trunk/octave-forge/main/control/devel/ncfsyn/ncfsyn.m
trunk/octave-forge/main/control/devel/ncfsyn/slsb10id.cc
trunk/octave-forge/main/control/devel/ncfsyn/slsb10kd.cc
trunk/octave-forge/main/control/devel/ncfsyn/slsb10zd.cc
Copied: trunk/octave-forge/main/control/devel/makefile_ncfsyn.m (from rev 8433, trunk/octave-forge/main/control/devel/ncfsyn/makefile_ncfsyn.m)
===================================================================
--- trunk/octave-forge/main/control/devel/makefile_ncfsyn.m (rev 0)
+++ trunk/octave-forge/main/control/devel/makefile_ncfsyn.m 2011-08-05 12:14:15 UTC (rev 8434)
@@ -0,0 +1,22 @@
+## H-infinity loop shaping - continuous-time
+mkoctfile "-Wl,-framework" "-Wl,vecLib" \
+ slsb10id.cc \
+ SB10ID.f SB02RD.f select.f SB10JD.f MB02VD.f \
+ MA02GD.f SB02MS.f MA02ED.f SB02RU.f SB02SD.f \
+ MB01RU.f SB02QD.f SB02MV.f SB02MW.f SB02MR.f \
+ MA02AD.f MB02PD.f MB01SD.f MB01UD.f SB03SY.f \
+ MB01RX.f SB03MX.f SB03SX.f MB01RY.f SB03QY.f \
+ SB03QX.f SB03MY.f SB04PX.f SB03MV.f SB03MW.f
+
+## H-infinity loop shaping - discrete-time - strictly proper case
+mkoctfile "-Wl,-framework" "-Wl,vecLib" \
+ slsb10kd.cc \
+ SB10KD.f SB02OD.f select.f SB02OY.f SB02OW.f \
+ SB02OV.f SB02MV.f SB02OU.f SB02MR.f
+
+## H-infinity loop shaping - discrete-time - proper case
+mkoctfile "-Wl,-framework" "-Wl,vecLib" \
+ slsb10zd.cc \
+ SB10ZD.f MA02AD.f SB02OD.f select.f MB01RX.f \
+ MB02VD.f SB02OY.f SB02OW.f SB02OV.f SB02OU.f \
+ SB02MR.f MA02GD.f SB02MV.f
\ No newline at end of file
Deleted: trunk/octave-forge/main/control/devel/ncfsyn/SB10ID.dat
===================================================================
--- trunk/octave-forge/main/control/devel/ncfsyn/SB10ID.dat 2011-08-05 11:57:01 UTC (rev 8433)
+++ trunk/octave-forge/main/control/devel/ncfsyn/SB10ID.dat 2011-08-05 12:14:15 UTC (rev 8434)
@@ -1,21 +0,0 @@
- SB10ID EXAMPLE PROGRAM DATA
- 6 2 3
- -1.0 0.0 4.0 5.0 -3.0 -2.0
- -2.0 4.0 -7.0 -2.0 0.0 3.0
- -6.0 9.0 -5.0 0.0 2.0 -1.0
- -8.0 4.0 7.0 -1.0 -3.0 0.0
- 2.0 5.0 8.0 -9.0 1.0 -4.0
- 3.0 -5.0 8.0 0.0 2.0 -6.0
- -3.0 -4.0
- 2.0 0.0
- -5.0 -7.0
- 4.0 -6.0
- -3.0 9.0
- 1.0 -2.0
- 1.0 -1.0 2.0 -4.0 0.0 -3.0
- -3.0 0.0 5.0 -1.0 1.0 1.0
- -7.0 5.0 0.0 -8.0 2.0 -2.0
- 1.0 -2.0
- 0.0 4.0
- 5.0 -3.0
- 1.0
Deleted: trunk/octave-forge/main/control/devel/ncfsyn/SB10ID.f
===================================================================
--- trunk/octave-forge/main/control/devel/ncfsyn/SB10ID.f 2011-08-05 11:57:01 UTC (rev 8433)
+++ trunk/octave-forge/main/control/devel/ncfsyn/SB10ID.f 2011-08-05 12:14:15 UTC (rev 8434)
@@ -1,584 +0,0 @@
- SUBROUTINE SB10ID( N, M, NP, A, LDA, B, LDB, C, LDC, D, LDD,
- $ FACTOR, NK, AK, LDAK, BK, LDBK, CK, LDCK,
- $ DK, LDDK, RCOND, IWORK, DWORK, LDWORK, BWORK,
- $ INFO )
-C
-C SLICOT RELEASE 5.0.
-C
-C Copyright (c) 2002-2010 NICONET e.V.
-C
-C This program is free software: you can redistribute it and/or
-C modify it under the terms of the GNU General Public License as
-C published by the Free Software Foundation, either version 2 of
-C the License, or (at your option) any later version.
-C
-C This program is distributed in the hope that it will be useful,
-C but WITHOUT ANY WARRANTY; without even the implied warranty of
-C MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-C GNU General Public License for more details.
-C
-C You should have received a copy of the GNU General Public License
-C along with this program. If not, see
-C <http://www.gnu.org/licenses/>.
-C
-C PURPOSE
-C
-C To compute the matrices of the positive feedback controller
-C
-C | Ak | Bk |
-C K = |----|----|
-C | Ck | Dk |
-C
-C for the shaped plant
-C
-C | A | B |
-C G = |---|---|
-C | C | D |
-C
-C in the McFarlane/Glover Loop Shaping Design Procedure.
-C
-C ARGUMENTS
-C
-C Input/Output Parameters
-C
-C N (input) INTEGER
-C The order of the plant. N >= 0.
-C
-C M (input) INTEGER
-C The column size of the matrix B. M >= 0.
-C
-C NP (input) INTEGER
-C The row size of the matrix C. NP >= 0.
-C
-C A (input) DOUBLE PRECISION array, dimension (LDA,N)
-C The leading N-by-N part of this array must contain the
-C system state matrix A of the shaped plant.
-C
-C LDA INTEGER
-C The leading dimension of the array A. LDA >= max(1,N).
-C
-C B (input) DOUBLE PRECISION array, dimension (LDB,M)
-C The leading N-by-M part of this array must contain the
-C system input matrix B of the shaped plant.
-C
-C LDB INTEGER
-C The leading dimension of the array B. LDB >= max(1,N).
-C
-C C (input) DOUBLE PRECISION array, dimension (LDC,N)
-C The leading NP-by-N part of this array must contain the
-C system output matrix C of the shaped plant.
-C
-C LDC INTEGER
-C The leading dimension of the array C. LDC >= max(1,NP).
-C
-C D (input) DOUBLE PRECISION array, dimension (LDD,M)
-C The leading NP-by-M part of this array must contain the
-C system matrix D of the shaped plant.
-C
-C LDD INTEGER
-C The leading dimension of the array D. LDD >= max(1,NP).
-C
-C FACTOR (input) DOUBLE PRECISION
-C = 1 implies that an optimal controller is required;
-C > 1 implies that a suboptimal controller is required,
-C achieving a performance FACTOR less than optimal.
-C FACTOR >= 1.
-C
-C NK (output) INTEGER
-C The order of the positive feedback controller. NK <= N.
-C
-C AK (output) DOUBLE PRECISION array, dimension (LDAK,N)
-C The leading NK-by-NK part of this array contains the
-C controller state matrix Ak.
-C
-C LDAK INTEGER
-C The leading dimension of the array AK. LDAK >= max(1,N).
-C
-C BK (output) DOUBLE PRECISION array, dimension (LDBK,NP)
-C The leading NK-by-NP part of this array contains the
-C controller input matrix Bk.
-C
-C LDBK INTEGER
-C The leading dimension of the array BK. LDBK >= max(1,N).
-C
-C CK (output) DOUBLE PRECISION array, dimension (LDCK,N)
-C The leading M-by-NK part of this array contains the
-C controller output matrix Ck.
-C
-C LDCK INTEGER
-C The leading dimension of the array CK. LDCK >= max(1,M).
-C
-C DK (output) DOUBLE PRECISION array, dimension (LDDK,NP)
-C The leading M-by-NP part of this array contains the
-C controller matrix Dk.
-C
-C LDDK INTEGER
-C The leading dimension of the array DK. LDDK >= max(1,M).
-C
-C RCOND (output) DOUBLE PRECISION array, dimension (2)
-C RCOND(1) contains an estimate of the reciprocal condition
-C number of the X-Riccati equation;
-C RCOND(2) contains an estimate of the reciprocal condition
-C number of the Z-Riccati equation.
-C
-C Workspace
-C
-C IWORK INTEGER array, dimension max(2*N,N*N,M,NP)
-C
-C DWORK DOUBLE PRECISION array, dimension (LDWORK)
-C On exit, if INFO = 0, DWORK(1) contains the optimal value
-C of LDWORK.
-C
-C LDWORK INTEGER
-C The dimension of the array DWORK.
-C LDWORK >= 4*N*N + M*M + NP*NP + 2*M*N + N*NP + 4*N +
-C max( 6*N*N + 5 + max(1,4*N*N+8*N), N*NP + 2*N ).
-C For good performance, LDWORK must generally be larger.
-C An upper bound of LDWORK in the above formula is
-C LDWORK >= 10*N*N + M*M + NP*NP + 2*M*N + 2*N*NP + 4*N +
-C 5 + max(1,4*N*N+8*N).
-C
-C BWORK LOGICAL array, dimension (2*N)
-C
-C Error Indicator
-C
-C INFO INTEGER
-C = 0: successful exit;
-C < 0: if INFO = -i, the i-th argument had an illegal
-C value;
-C = 1: the X-Riccati equation is not solved successfully;
-C = 2: the Z-Riccati equation is not solved successfully;
-C = 3: the iteration to compute eigenvalues or singular
-C values failed to converge;
-C = 4: the matrix Ip - D*Dk is singular;
-C = 5: the matrix Im - Dk*D is singular;
-C = 6: the closed-loop system is unstable.
-C
-C METHOD
-C
-C The routine implements the formulas given in [1].
-C
-C REFERENCES
-C
-C [1] McFarlane, D. and Glover, K.
-C A loop shaping design procedure using H_infinity synthesis.
-C IEEE Trans. Automat. Control, vol. AC-37, no. 6, pp. 759-769,
-C 1992.
-C
-C NUMERICAL ASPECTS
-C
-C The accuracy of the results depends on the conditioning of the
-C two Riccati equations solved in the controller design (see the
-C output parameter RCOND).
-C
-C CONTRIBUTORS
-C
-C P.Hr. Petkov, D.W. Gu and M.M. Konstantinov, October 2000.
-C
-C REVISIONS
-C
-C V. Sima, Research Institute for Informatics, Bucharest, Oct. 2000,
-C Feb. 2001.
-C
-C KEYWORDS
-C
-C H_infinity control, Loop-shaping design, Robust control.
-C
-C ******************************************************************
-C
-C .. Parameters ..
- DOUBLE PRECISION ZERO, ONE
- PARAMETER ( ZERO = 0.0D+0, ONE = 1.0D+0 )
-C ..
-C .. Scalar Arguments ..
- INTEGER INFO, LDA, LDAK, LDB, LDBK, LDC, LDCK, LDD,
- $ LDDK, LDWORK, M, N, NK, NP
- DOUBLE PRECISION FACTOR
-C ..
-C .. Array Arguments ..
- INTEGER IWORK( * )
- LOGICAL BWORK( * )
- DOUBLE PRECISION A( LDA, * ), AK( LDAK, * ), B( LDB, * ),
- $ BK( LDBK, * ), C( LDC, * ), CK( LDCK, * ),
- $ D( LDD, * ), DK( LDDK, * ), DWORK( * ),
- $ RCOND( 2 )
-C ..
-C .. Local Scalars ..
- CHARACTER*1 HINV
- INTEGER I, I1, I2, I3, I4, I5, I6, I7, I8, I9, I10,
- $ I11, I12, I13, INFO2, IWRK, J, LWA, LWAMAX,
- $ MINWRK, N2, NS, SDIM
- DOUBLE PRECISION SEP, FERR, GAMMA
-C ..
-C .. External Functions ..
- LOGICAL SELECT
- EXTERNAL SELECT
-C ..
-C .. External Subroutines ..
- EXTERNAL DGEES, DGEMM, DLACPY, DLASET, DPOTRF, DPOTRS,
- $ DSYRK, DTRSM, MB02VD, SB02RD, SB10JD, XERBLA
-C ..
-C .. Intrinsic Functions ..
- INTRINSIC DBLE, INT, MAX, SQRT
-C ..
-C .. Executable Statements ..
-C
-C Decode and Test input parameters.
-C
- INFO = 0
- IF( N.LT.0 ) THEN
- INFO = -1
- ELSE IF( M.LT.0 ) THEN
- INFO = -2
- ELSE IF( NP.LT.0 ) THEN
- INFO = -3
- ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
- INFO = -5
- ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
- INFO = -7
- ELSE IF( LDC.LT.MAX( 1, NP ) ) THEN
- INFO = -9
- ELSE IF( LDD.LT.MAX( 1, NP ) ) THEN
- INFO = -11
- ELSE IF( FACTOR.LT.ONE ) THEN
- INFO = -12
- ELSE IF( LDAK.LT.MAX( 1, N ) ) THEN
- INFO = -15
- ELSE IF( LDBK.LT.MAX( 1, N ) ) THEN
- INFO = -17
- ELSE IF( LDCK.LT.MAX( 1, M ) ) THEN
- INFO = -19
- ELSE IF( LDDK.LT.MAX( 1, M ) ) THEN
- INFO = -21
- END IF
-C
-C Compute workspace.
-C
- MINWRK = 4*N*N + M*M + NP*NP + 2*M*N + N*NP + 4*N +
- $ MAX( 6*N*N + 5 + MAX( 1, 4*N*N + 8*N ), N*NP + 2*N )
- IF( LDWORK.LT.MINWRK ) THEN
- INFO = -25
- END IF
- IF( INFO.NE.0 ) THEN
- CALL XERBLA( 'SB10ID', -INFO )
- RETURN
- END IF
-C
-C Quick return if possible.
-C
- IF( N.EQ.0 .OR. M.EQ.0 .OR. NP.EQ.0 ) THEN
- RCOND( 1 ) = ONE
- RCOND( 2 ) = ONE
- DWORK( 1 ) = ONE
- RETURN
- END IF
-C
-C Workspace usage.
-C
- I1 = N*N
- I2 = I1 + N*N
- I3 = I2 + M*N
- I4 = I3 + M*N
- I5 = I4 + M*M
- I6 = I5 + NP*NP
- I7 = I6 + NP*N
- I8 = I7 + N*N
- I9 = I8 + N*N
- I10 = I9 + N*N
- I11 = I10 + N*N
- I12 = I11 + 2*N
- I13 = I12 + 2*N
-C
- IWRK = I13 + 4*N*N
-C
-C Compute D'*C .
-C
- CALL DGEMM( 'T', 'N', M, N, NP, ONE, D, LDD, C, LDC, ZERO,
- $ DWORK( I2+1 ), M )
-C
-C Compute S = Im + D'*D .
-C
- CALL DLASET( 'U', M, M, ZERO, ONE, DWORK( I4+1 ), M )
- CALL DSYRK( 'U', 'T', M, NP, ONE, D, LDD, ONE, DWORK( I4+1 ), M )
-C
-C Factorize S, S = T'*T, with T upper triangular.
-C
- CALL DPOTRF( 'U', M, DWORK( I4+1 ), M, INFO2 )
-C
-C -1
-C Compute S D'*C .
-C
- CALL DPOTRS( 'U', M, N, DWORK( I4+1 ), M, DWORK( I2+1 ), M,
- $ INFO2 )
-C
-C -1
-C Compute B*T .
-C
- CALL DLACPY( 'F', N, M, B, LDB, DWORK( I3+1 ), N )
- CALL DTRSM( 'R', 'U', 'N', 'N', N, M, ONE, DWORK( I4+1 ), M,
- $ DWORK( I3+1 ), N )
-C
-C Compute R = Ip + D*D' .
-C
- CALL DLASET( 'U', NP, NP, ZERO, ONE, DWORK( I5+1 ), NP )
- CALL DSYRK( 'U', 'N', NP, M, ONE, D, LDD, ONE, DWORK( I5+1 ), NP )
-C
-C Factorize R, R = U'*U, with U upper triangular.
-C
- CALL DPOTRF( 'U', NP, DWORK( I5+1 ), NP, INFO2 )
-C
-C -T
-C Compute U C .
-C
- CALL DLACPY( 'F', NP, N, C, LDC, DWORK( I6+1 ), NP )
- CALL DTRSM( 'L', 'U', 'T', 'N', NP, N, ONE, DWORK( I5+1 ), NP,
- $ DWORK( I6+1 ), NP )
-C
-C -1
-C Compute Ar = A - B*S D'*C .
-C
- CALL DLACPY( 'F', N, N, A, LDA, DWORK( I7+1 ), N )
- CALL DGEMM( 'N', 'N', N, N, M, -ONE, B, LDB, DWORK( I2+1 ), M,
- $ ONE, DWORK( I7+1 ), N )
-C
-C -1
-C Compute the upper triangle of Cr = C'*R *C .
-C
- CALL DSYRK( 'U', 'T', N, NP, ONE, DWORK( I6+1 ), NP, ZERO,
- $ DWORK( I8+1 ), N )
-C
-C -1
-C Compute the upper triangle of Dr = B*S B' .
-C
- CALL DSYRK( 'U', 'N', N, M, ONE, DWORK( I3+1 ), N, ZERO,
- $ DWORK( I9+1 ), N )
-C
-C Solution of the Riccati equation Ar'*X + X*Ar + Cr - X*Dr*X = 0 .
-C Workspace: need 10*N*N + M*M + NP*NP + 2*M*N + N*NP + 4*N +
-C 5 + max(1,4*N*N+8*N).
-C prefer larger.
-C AK is used as workspace.
-C
- N2 = 2*N
- CALL SB02RD( 'A', 'C', HINV, 'N', 'U', 'G', 'S', 'N', 'O', N,
- $ DWORK( I7+1 ), N, DWORK( I10+1 ), N, AK, LDAK,
- $ DWORK( I9+1 ), N, DWORK( I8+1 ), N, DWORK, N, SEP,
- $ RCOND( 1 ), FERR, DWORK( I11+1 ), DWORK( I12+1 ),
- $ DWORK( I13+1 ), N2, IWORK, DWORK( IWRK+1 ),
- $ LDWORK-IWRK, BWORK, INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 1
- RETURN
- END IF
- LWA = INT( DWORK( IWRK+1 ) ) + IWRK
- LWAMAX = MAX( MINWRK, LWA )
-C
-C Solution of the Riccati equation Ar*Z + Z*Ar' + Dr - Z*Cr*Z = 0 .
-C
- CALL SB02RD( 'A', 'C', HINV, 'T', 'U', 'G', 'S', 'N', 'O', N,
- $ DWORK( I7+1 ), N, DWORK( I10+1 ), N, AK, LDAK,
- $ DWORK( I8+1 ), N, DWORK( I9+1 ), N, DWORK( I1+1 ),
- $ N, SEP, RCOND( 2 ), FERR, DWORK( I11+1 ),
- $ DWORK( I12+1 ), DWORK( I13+1 ), N2, IWORK,
- $ DWORK( IWRK+1 ), LDWORK-IWRK, BWORK, INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 2
- RETURN
- END IF
- LWA = INT( DWORK( IWRK+1 ) ) + IWRK
- LWAMAX = MAX( LWA, LWAMAX )
-C
-C -1 -1
-C Compute F1 = -( S D'*C + S B'*X ) .
-C
- CALL DTRSM( 'R', 'U', 'T', 'N', N, M, ONE, DWORK( I4+1 ), M,
- $ DWORK( I3+1 ), N )
- CALL DGEMM( 'T', 'N', M, N, N, -ONE, DWORK( I3+1 ), N, DWORK, N,
- $ -ONE, DWORK( I2+1 ), M )
-C
-C Compute gamma .
-C
- CALL DGEMM( 'N', 'N', N, N, N, ONE, DWORK, N, DWORK( I1+1 ), N,
- $ ZERO, DWORK( I7+1 ), N )
- CALL DGEES( 'N', 'N', SELECT, N, DWORK( I7+1 ), N, SDIM,
- $ DWORK( I11+1 ), DWORK( I12+1 ), DWORK( IWRK+1 ), N,
- $ DWORK( IWRK+1 ), LDWORK-IWRK, BWORK, INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 3
- RETURN
- END IF
- LWA = INT( DWORK( IWRK+1 ) ) + IWRK
- LWAMAX = MAX( LWA, LWAMAX )
- GAMMA = ZERO
- DO 10 I = 1, N
- GAMMA = MAX( GAMMA, DWORK( I11+I ) )
- 10 CONTINUE
- GAMMA = FACTOR*SQRT( ONE + GAMMA )
-C
-C Workspace usage.
-C Workspace: need 4*N*N + M*N + N*NP.
-C
- I4 = I3 + N*N
- I5 = I4 + N*N
-C
-C Compute Ac = A + B*F1 .
-C
- CALL DLACPY( 'F', N, N, A, LDA, DWORK( I4+1 ), N )
- CALL DGEMM( 'N', 'N', N, N, M, ONE, B, LDB, DWORK( I2+1 ), M,
- $ ONE, DWORK( I4+1 ), N )
-C
-C Compute W1' = (1-gamma^2)*In + Z*X .
-C
- CALL DLASET( 'F', N, N, ZERO, ONE-GAMMA*GAMMA, DWORK( I3+1 ), N )
- CALL DGEMM( 'N', 'N', N, N, N, ONE, DWORK( I1+1 ), N, DWORK, N,
- $ ONE, DWORK( I3+1 ), N )
-C
-C Compute Bcp = gamma^2*Z*C' .
-C
- CALL DGEMM( 'N', 'T', N, NP, N, GAMMA*GAMMA, DWORK( I1+1 ), N, C,
- $ LDC, ZERO, BK, LDBK )
-C
-C Compute C + D*F1 .
-C
- CALL DLACPY( 'F', NP, N, C, LDC, DWORK( I5+1 ), NP )
- CALL DGEMM( 'N', 'N', NP, N, M, ONE, D, LDD, DWORK( I2+1 ), M,
- $ ONE, DWORK( I5+1 ), NP )
-C
-C Compute Acp = W1'*Ac + gamma^2*Z*C'*(C+D*F1) .
-C
- CALL DGEMM( 'N', 'N', N, N, N, ONE, DWORK( I3+1 ), N,
- $ DWORK( I4+1 ), N, ZERO, AK, LDAK )
- CALL DGEMM( 'N', 'N', N, N, NP, ONE, BK, LDBK,
- $ DWORK( I5+1 ), NP, ONE, AK, LDAK )
-C
-C Compute Ccp = B'*X .
-C
- CALL DGEMM( 'T', 'N', M, N, N, ONE, B, LDB, DWORK, N, ZERO,
- $ CK, LDCK )
-C
-C Set Dcp = -D' .
-C
- DO 30 I = 1, M
- DO 20 J = 1, NP
- DK( I, J ) = -D( J, I )
- 20 CONTINUE
- 30 CONTINUE
-C
- IWRK = I4
-C
-C Reduce the generalized state-space description to a regular one.
-C Workspace: need 3*N*N + M*N.
-C Additional workspace: need 2*N*N + 2*N + N*MAX(5,N+M+NP).
-C prefer larger.
-C
- CALL SB10JD( N, NP, M, AK, LDAK, BK, LDBK, CK, LDCK, DK, LDDK,
- $ DWORK( I3+1 ), N, NK, DWORK( IWRK+1 ), LDWORK-IWRK,
- $ INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 3
- RETURN
- END IF
- LWA = INT( DWORK( IWRK+1 ) ) + IWRK
- LWAMAX = MAX( LWA, LWAMAX )
-C
-C Workspace usage.
-C Workspace: need 4*N*N + M*M + NP*NP + 2*M*N + 2*N*NP.
-C (NK <= N.)
-C
- I2 = NP*NP
- I3 = I2 + NK*NP
- I4 = I3 + M*M
- I5 = I4 + N*M
- I6 = I5 + NP*NK
- I7 = I6 + M*N
-C
- IWRK = I7 + ( N + NK )*( N + NK )
-C
-C Compute Ip - D*Dk .
-C
- CALL DLASET( 'Full', NP, NP, ZERO, ONE, DWORK, NP )
- CALL DGEMM( 'N', 'N', NP, NP, M, -ONE, D, LDD, DK, LDDK, ONE,
- $ DWORK, NP )
-C
-C -1
-C Compute Bk*(Ip-D*Dk) .
-C
- CALL DLACPY( 'F', NK, NP, BK, LDBK, DWORK( I2+1 ), NK )
- CALL MB02VD( 'N', NK, NP, DWORK, NP, IWORK, DWORK( I2+1 ), NK,
- $ INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 4
- RETURN
- END IF
-C
-C Compute Im - Dk*D .
-C
- CALL DLASET( 'Full', M, M, ZERO, ONE, DWORK( I3+1 ), M )
- CALL DGEMM( 'N', 'N', M, M, NP, -ONE, DK, LDDK, D, LDD, ONE,
- $ DWORK( I3+1 ), M )
-C
-C -1
-C Compute B*(Im-Dk*D) .
-C
- CALL DLACPY( 'F', N, M, B, LDB, DWORK( I4+1 ), N )
- CALL MB02VD( 'N', N, M, DWORK( I3+1 ), M, IWORK, DWORK( I4+1 ), N,
- $ INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 5
- RETURN
- END IF
-C
-C Compute D*Ck .
-C
- CALL DGEMM( 'N', 'N', NP, NK, M, ONE, D, LDD, CK, LDCK, ZERO,
- $ DWORK( I5+1 ), NP )
-C
-C Compute Dk*C .
-C
- CALL DGEMM( 'N', 'N', M, N, NP, ONE, DK, LDDK, C, LDC, ZERO,
- $ DWORK( I6+1 ), M )
-C
-C Compute the closed-loop state matrix.
-C
- CALL DLACPY( 'F', N, N, A, LDA, DWORK( I7+1 ), N+NK )
- CALL DGEMM( 'N', 'N', N, N, M, ONE, DWORK( I4+1 ), N,
- $ DWORK( I6+1 ), M, ONE, DWORK( I7+1 ), N+NK )
- CALL DGEMM( 'N', 'N', NK, N, NP, ONE, DWORK( I2+1 ), NK, C, LDC,
- $ ZERO, DWORK( I7+N+1 ), N+NK )
- CALL DGEMM( 'N', 'N', N, NK, M, ONE, DWORK( I4+1 ), N, CK, LDCK,
- $ ZERO, DWORK( I7+(N+NK)*N+1 ), N+NK )
- CALL DLACPY( 'F', NK, NK, AK, LDAK, DWORK( I7+(N+NK)*N+N+1 ),
- $ N+NK )
- CALL DGEMM( 'N', 'N', NK, NK, NP, ONE, DWORK( I2+1 ), NK,
- $ DWORK( I5+1 ), NP, ONE, DWORK( I7+(N+NK)*N+N+1 ),
- $ N+NK )
-C
-C Compute the closed-loop poles.
-C Additional workspace: need 3*(N+NK); prefer larger.
-C The fact that M > 0, NP > 0, and NK <= N is used here.
-C
- CALL DGEES( 'N', 'N', SELECT, N+NK, DWORK( I7+1 ), N+NK, SDIM,
- $ DWORK, DWORK( N+NK+1 ), DWORK( IWRK+1 ), N,
- $ DWORK( IWRK+1 ), LDWORK-IWRK, BWORK, INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 3
- RETURN
- END IF
- LWA = INT( DWORK( IWRK+1 ) ) + IWRK
- LWAMAX = MAX( LWA, LWAMAX )
-C
-C Check the stability of the closed-loop system.
-C
- NS = 0
- DO 40 I = 1, N+NK
- IF( DWORK( I ).GE.ZERO ) NS = NS + 1
- 40 CONTINUE
- IF( NS.GT.0 ) THEN
- INFO = 6
- RETURN
- END IF
-C
- DWORK( 1 ) = DBLE( LWAMAX )
- RETURN
-C *** Last line of SB10ID ***
- END
Deleted: trunk/octave-forge/main/control/devel/ncfsyn/SB10ID.res
===================================================================
--- trunk/octave-forge/main/control/devel/ncfsyn/SB10ID.res 2011-08-05 11:57:01 UTC (rev 8433)
+++ trunk/octave-forge/main/control/devel/ncfsyn/SB10ID.res 2011-08-05 12:14:15 UTC (rev 8434)
@@ -1,32 +0,0 @@
- SB10ID EXAMPLE PROGRAM RESULTS
-
-
- The controller state matrix AK is
-
- -39.0671 9.9293 22.2322 -27.4113 43.8655
- -6.6117 3.0006 11.0878 -11.4130 15.4269
- 33.6805 -6.6934 -23.9953 14.1438 -33.4358
- -32.3191 9.7316 25.4033 -24.0473 42.0517
- -44.1655 18.7767 34.8873 -42.4369 50.8437
-
- The controller input matrix BK is
-
- -10.2905 -16.5382 -10.9782
- -4.3598 -8.7525 -5.1447
- 6.5962 1.8975 6.2316
- -9.8770 -14.7041 -11.8778
- -9.6726 -22.7309 -18.2692
-
- The controller output matrix CK is
-
- -0.6647 -0.0599 -1.0376 0.5619 1.7297
- -8.4202 3.9573 7.3094 -7.6283 10.6768
-
- The controller matrix DK is
-
- 0.8466 0.4979 -0.6993
- -1.2226 -4.8689 -4.5056
-
- The estimated condition numbers are
-
- 0.13861D-01 0.90541D-02
Deleted: trunk/octave-forge/main/control/devel/ncfsyn/SB10JD.f
===================================================================
--- trunk/octave-forge/main/control/devel/ncfsyn/SB10JD.f 2011-08-05 11:57:01 UTC (rev 8433)
+++ trunk/octave-forge/main/control/devel/ncfsyn/SB10JD.f 2011-08-05 12:14:15 UTC (rev 8434)
@@ -1,355 +0,0 @@
- SUBROUTINE SB10JD( N, M, NP, A, LDA, B, LDB, C, LDC, D, LDD, E,
- $ LDE, NSYS, DWORK, LDWORK, INFO )
-C
-C SLICOT RELEASE 5.0.
-C
-C Copyright (c) 2002-2010 NICONET e.V.
-C
-C This program is free software: you can redistribute it and/or
-C modify it under the terms of the GNU General Public License as
-C published by the Free Software Foundation, either version 2 of
-C the License, or (at your option) any later version.
-C
-C This program is distributed in the hope that it will be useful,
-C but WITHOUT ANY WARRANTY; without even the implied warranty of
-C MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-C GNU General Public License for more details.
-C
-C You should have received a copy of the GNU General Public License
-C along with this program. If not, see
-C <http://www.gnu.org/licenses/>.
-C
-C PURPOSE
-C
-C To convert the descriptor state-space system
-C
-C E*dx/dt = A*x + B*u
-C y = C*x + D*u
-C
-C into regular state-space form
-C
-C dx/dt = Ad*x + Bd*u
-C y = Cd*x + Dd*u .
-C
-C ARGUMENTS
-C
-C Input/Output Parameters
-C
-C N (input) INTEGER
-C The order of the descriptor system. N >= 0.
-C
-C M (input) INTEGER
-C The column size of the matrix B. M >= 0.
-C
-C NP (input) INTEGER
-C The row size of the matrix C. NP >= 0.
-C
-C A (input/output) DOUBLE PRECISION array, dimension (LDA,N)
-C On entry, the leading N-by-N part of this array must
-C contain the state matrix A of the descriptor system.
-C On exit, the leading NSYS-by-NSYS part of this array
-C contains the state matrix Ad of the converted system.
-C
-C LDA INTEGER
-C The leading dimension of the array A. LDA >= max(1,N).
-C
-C B (input/output) DOUBLE PRECISION array, dimension (LDB,M)
-C On entry, the leading N-by-M part of this array must
-C contain the input matrix B of the descriptor system.
-C On exit, the leading NSYS-by-M part of this array
-C contains the input matrix Bd of the converted system.
-C
-C LDB INTEGER
-C The leading dimension of the array B. LDB >= max(1,N).
-C
-C C (input/output) DOUBLE PRECISION array, dimension (LDC,N)
-C On entry, the leading NP-by-N part of this array must
-C contain the output matrix C of the descriptor system.
-C On exit, the leading NP-by-NSYS part of this array
-C contains the output matrix Cd of the converted system.
-C
-C LDC INTEGER
-C The leading dimension of the array C. LDC >= max(1,NP).
-C
-C D (input/output) DOUBLE PRECISION array, dimension (LDD,M)
-C On entry, the leading NP-by-M part of this array must
-C contain the matrix D of the descriptor system.
-C On exit, the leading NP-by-M part of this array contains
-C the matrix Dd of the converted system.
-C
-C LDD INTEGER
-C The leading dimension of the array D. LDD >= max(1,NP).
-C
-C E (input/output) DOUBLE PRECISION array, dimension (LDE,N)
-C On entry, the leading N-by-N part of this array must
-C contain the matrix E of the descriptor system.
-C On exit, this array contains no useful information.
-C
-C LDE INTEGER
-C The leading dimension of the array E. LDE >= max(1,N).
-C
-C NSYS (output) INTEGER
-C The order of the converted state-space system.
-C
-C Workspace
-C
-C DWORK DOUBLE PRECISION array, dimension (LDWORK)
-C On exit, if INFO = 0, DWORK(1) contains the optimal value
-C of LDWORK.
-C
-C LDWORK INTEGER
-C The dimension of the array DWORK.
-C LDWORK >= max( 1, 2*N*N + 2*N + N*MAX( 5, N + M + NP ) ).
-C For good performance, LDWORK must generally be larger.
-C
-C Error Indicator
-C
-C INFO INTEGER
-C = 0: successful exit;
-C < 0: if INFO = -i, the i-th argument had an illegal
-C value;
-C = 1: the iteration for computing singular value
-C decomposition did not converge.
-C
-C METHOD
-C
-C The routine performs the transformations described in [1].
-C
-C REFERENCES
-C
-C [1] Chiang, R.Y. and Safonov, M.G.
-C Robust Control Toolbox User's Guide.
-C The MathWorks Inc., Natick, Mass., 1992.
-C
-C CONTRIBUTORS
-C
-C P.Hr. Petkov, D.W. Gu and M.M. Konstantinov, October 1999.
-C
-C REVISIONS
-C
-C V. Sima, Research Institute for Informatics, Bucharest, Oct. 2000,
-C Feb. 2001.
-C
-C KEYWORDS
-C
-C Descriptor systems, state-space models.
-C
-C ******************************************************************
-C
-C .. Parameters ..
- DOUBLE PRECISION ZERO, ONE
- PARAMETER ( ZERO = 0.0D+0, ONE = 1.0D+0 )
-C ..
-C .. Scalar Arguments ..
- INTEGER INFO, LDA, LDB, LDC, LDD, LDE, LDWORK, M, N,
- $ NP, NSYS
-C ..
-C .. Array Arguments ..
- DOUBLE PRECISION A( LDA, * ), B( LDB, * ), C( LDC, * ),
- $ D( LDD, * ), DWORK( * ), E( LDE, * )
-C ..
-C .. Local Scalars ..
- INTEGER I, IA12, IA21, IB2, IC2, INFO2, IS, ISA, IU,
- $ IV, IWRK, J, K, LWA, LWAMAX, MINWRK, NS1
- DOUBLE PRECISION EPS, SCALE, TOL
-C ..
-C .. External Functions ..
- DOUBLE PRECISION DLAMCH
- EXTERNAL DLAMCH
-C ..
-C .. External Subroutines ..
- EXTERNAL DGEMM, DGESVD, DLACPY, DLASET, DSCAL, XERBLA
-C ..
-C .. Intrinsic Functions ..
- INTRINSIC DBLE, INT, MAX, SQRT
-C ..
-C .. Executable Statements ..
-C
-C Decode and Test input parameters.
-C
- INFO = 0
- IF( N.LT.0 ) THEN
- INFO = -1
- ELSE IF( M.LT.0 ) THEN
- INFO = -2
- ELSE IF( NP.LT.0 ) THEN
- INFO = -3
- ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
- INFO = -5
- ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
- INFO = -7
- ELSE IF( LDC.LT.MAX( 1, NP ) ) THEN
- INFO = -9
- ELSE IF( LDD.LT.MAX( 1, NP ) ) THEN
- INFO = -11
- ELSE IF( LDE.LT.MAX( 1, N ) ) THEN
- INFO = -13
- END IF
-C
-C Compute workspace.
-C
- MINWRK = MAX( 1, 2*N*( N + 1 ) + N*MAX( 5, N + M + NP ) )
- IF( LDWORK.LT.MINWRK ) THEN
- INFO = -16
- END IF
- IF( INFO.NE.0 ) THEN
- CALL XERBLA( 'SB10JD', -INFO )
- RETURN
- END IF
-C
-C Quick return if possible.
-C
- IF( N.EQ.0 ) THEN
- NSYS = 0
- DWORK( 1 ) = ONE
- RETURN
- END IF
-C
-C Set tol.
-C
- EPS = DLAMCH( 'Epsilon' )
- TOL = SQRT( EPS )
-C
-C Workspace usage.
-C
- IS = 0
- IU = IS + N
- IV = IU + N*N
-C
- IWRK = IV + N*N
-C
-C Compute the SVD of E.
-C Additional workspace: need 5*N; prefer larger.
-C
- CALL DGESVD( 'S', 'S', N, N, E, LDE, DWORK( IS+1 ), DWORK( IU+1 ),
- $ N, DWORK( IV+1 ), N, DWORK( IWRK+1 ), LDWORK-IWRK,
- $ INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 1
- RETURN
- END IF
- LWAMAX = MAX( MINWRK, INT( DWORK( IWRK+1 ) + IWRK ) )
-C
-C Determine the rank of E.
-C
- NS1 = 0
- DO 10 I = 1, N
- IF( DWORK( IS+I ).GT.TOL ) NS1 = NS1 + 1
- 10 CONTINUE
- IF( NS1.GT.0 ) THEN
-C
-C Transform A.
-C Additional workspace: need N*max(N,M,NP).
-C
- CALL DGEMM( 'T', 'N', N, N, N, ONE, DWORK( IU+1 ), N, A, LDA,
- $ ZERO, DWORK( IWRK+1 ), N )
- CALL DGEMM( 'N', 'T', N, N, N, ONE, DWORK( IWRK+1 ), N,
- $ DWORK( IV+1 ), N, ZERO, A, LDA )
-C
-C Transform B.
-C
- CALL DLACPY( 'Full', N, M, B, LDB, DWORK( IWRK+1 ), N )
- CALL DGEMM( 'T', 'N', N, M, N, ONE, DWORK( IU+1 ), N,
- $ DWORK( IWRK+1 ), N, ZERO, B, LDB )
-C
-C Transform C.
-C
- CALL DLACPY( 'Full', NP, N, C, LDC, DWORK( IWRK+1 ), NP )
- CALL DGEMM( 'N', 'T', NP, N, N, ONE, DWORK( IWRK+1 ), NP,
- $ DWORK( IV+1 ), N, ZERO, C, LDC )
-C
- K = N - NS1
- IF( K.GT.0 ) THEN
- ISA = IU + K*K
- IV = ISA + K
- IWRK = IV + K*MAX( K, NS1 )
-C
-C Compute the SVD of A22.
-C Additional workspace: need 5*K; prefer larger.
-C
- CALL DGESVD( 'S', 'S', K, K, A( NS1+1, NS1+1 ), LDA,
- $ DWORK( ISA+1 ), DWORK( IU+1 ), K,
- $ DWORK( IV+1 ), K, DWORK( IWRK+1 ), LDWORK-IWRK,
- $ INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 1
- RETURN
- END IF
- IA12 = IWRK
- IB2 = IA12 + NS1*K
- IC2 = IB2 + K*M
-C
- LWA = INT( DWORK( IWRK+1 ) ) + IWRK
- LWAMAX = MAX( LWA, LWAMAX, IC2 + K*NP )
-C
-C Compute the transformed A12.
-C
- CALL DGEMM( 'N', 'T', NS1, K, K, ONE, A( 1, NS1+1 ), LDA,
- $ DWORK( IV+1 ), K, ZERO, DWORK( IA12+1 ), NS1 )
-C
-C Compute CC2.
-C
- CALL DGEMM( 'N', 'T', NP, K, K, ONE, C( 1, NS1+1 ), LDC,
- $ DWORK( IV+1 ), K, ZERO, DWORK( IC2+1 ), NP )
-C
-C Compute the transformed A21.
-C
- IA21 = IV
- CALL DGEMM( 'T', 'N', K, NS1, K, ONE, DWORK( IU+1 ), K,
- $ A( NS1+1, 1 ), LDA, ZERO, DWORK( IA21+1 ), K )
-C
-C Compute BB2.
-C
- CALL DGEMM( 'T', 'N', K, M, K, ONE, DWORK( IU+1 ), K,
- $ B( NS1+1, 1 ), LDB, ZERO, DWORK( IB2+1 ), K )
-C
-C Compute A12*pinv(A22) and CC2*pinv(A22).
-C
- DO 20 J = 1, K
- SCALE = ZERO
- IF( DWORK( ISA+J ).GT.TOL ) SCALE = ONE/DWORK( ISA+J )
- CALL DSCAL( NS1, SCALE, DWORK( IA12+(J-1)*NS1+1 ), 1 )
- CALL DSCAL( NP, SCALE, DWORK( IC2+(J-1)*NP+1 ), 1 )
- 20 CONTINUE
-C
-C Compute Ad.
-C
- CALL DGEMM( 'N', 'N', NS1, NS1, K, -ONE, DWORK( IA12+1 ),
- $ NS1, DWORK( IA21+1 ), K, ONE, A, LDA )
-C
-C Compute Bd.
-C
- CALL DGEMM( 'N', 'N', NS1, M, K, -ONE, DWORK( IA12+1 ), NS1,
- $ DWORK( IB2+1 ), K, ONE, B, LDB )
-C
-C Compute Cd.
-C
- CALL DGEMM( 'N', 'N', NP, NS1, K, -ONE, DWORK( IC2+1 ), NP,
- $ DWORK( IA21+1 ), K, ONE, C, LDC )
-C
-C Compute Dd.
-C
- CALL DGEMM( 'N', 'N', NP, M, K, -ONE, DWORK( IC2+1 ), NP,
- $ DWORK( IB2+1 ), K, ONE, D, LDD )
- END IF
- DO 30 I = 1, NS1
- SCALE = ONE/SQRT( DWORK( IS+I ) )
- CALL DSCAL( NS1, SCALE, A( I, 1 ), LDA )
- CALL DSCAL( M, SCALE, B( I, 1 ), LDB )
- 30 CONTINUE
- DO 40 J = 1, NS1
- SCALE = ONE/SQRT( DWORK( IS+J ) )
- CALL DSCAL( NS1, SCALE, A( 1, J ), 1 )
- CALL DSCAL( NP, SCALE, C( 1, J ), 1 )
- 40 CONTINUE
- NSYS = NS1
- ELSE
- CALL DLASET( 'F', N, N, ZERO, -ONE/EPS, A, LDA )
- CALL DLASET( 'F', N, M, ZERO, ZERO, B, LDB )
- CALL DLASET( 'F', NP, N, ZERO, ZERO, C, LDC )
- NSYS = N
- END IF
- DWORK( 1 ) = DBLE( LWAMAX )
- RETURN
-C *** Last line of SB10JD ***
- END
Deleted: trunk/octave-forge/main/control/devel/ncfsyn/SB10KD.dat
===================================================================
--- trunk/octave-forge/main/control/devel/ncfsyn/SB10KD.dat 2011-08-05 11:57:01 UTC (rev 8433)
+++ trunk/octave-forge/main/control/devel/ncfsyn/SB10KD.dat 2011-08-05 12:14:15 UTC (rev 8434)
@@ -1,17 +0,0 @@
- SB10KD EXAMPLE PROGRAM DATA
- 6 2 2
- 0.2 0.0 0.3 0.0 -0.3 -0.1
- -0.3 0.2 -0.4 -0.3 0.0 0.0
- -0.1 0.1 -0.1 0.0 0.0 -0.3
- 0.1 0.0 0.0 -0.1 -0.1 0.0
- 0.0 0.3 0.6 0.2 0.1 -0.4
- 0.2 -0.4 0.0 0.0 0.2 -0.2
- -1.0 -2.0
- 1.0 3.0
- -3.0 -4.0
- 1.0 -2.0
- 0.0 1.0
- 1.0 5.0
- 1.0 -1.0 2.0 -2.0 0.0 -3.0
- -3.0 0.0 1.0 -1.0 1.0 -1.0
- 1.1
Deleted: trunk/octave-forge/main/control/devel/ncfsyn/SB10KD.f
===================================================================
--- trunk/octave-forge/main/control/devel/ncfsyn/SB10KD.f 2011-08-05 11:57:01 UTC (rev 8433)
+++ trunk/octave-forge/main/control/devel/ncfsyn/SB10KD.f 2011-08-05 12:14:15 UTC (rev 8434)
@@ -1,650 +0,0 @@
- SUBROUTINE SB10KD( N, M, NP, A, LDA, B, LDB, C, LDC, FACTOR,
- $ AK, LDAK, BK, LDBK, CK, LDCK, DK, LDDK, RCOND,
- $ IWORK, DWORK, LDWORK, BWORK, INFO )
-C
-C SLICOT RELEASE 5.0.
-C
-C Copyright (c) 2002-2010 NICONET e.V.
-C
-C This program is free software: you can redistribute it and/or
-C modify it under the terms of the GNU General Public License as
-C published by the Free Software Foundation, either version 2 of
-C the License, or (at your option) any later version.
-C
-C This program is distributed in the hope that it will be useful,
-C but WITHOUT ANY WARRANTY; without even the implied warranty of
-C MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-C GNU General Public License for more details.
-C
-C You should have received a copy of the GNU General Public License
-C along with this program. If not, see
-C <http://www.gnu.org/licenses/>.
-C
-C PURPOSE
-C
-C To compute the matrices of the positive feedback controller
-C
-C | Ak | Bk |
-C K = |----|----|
-C | Ck | Dk |
-C
-C for the shaped plant
-C
-C | A | B |
-C G = |---|---|
-C | C | 0 |
-C
-C in the Discrete-Time Loop Shaping Design Procedure.
-C
-C ARGUMENTS
-C
-C Input/Output Parameters
-C
-C N (input) INTEGER
-C The order of the plant. N >= 0.
-C
-C M (input) INTEGER
-C The column size of the matrix B. M >= 0.
-C
-C NP (input) INTEGER
-C The row size of the matrix C. NP >= 0.
-C
-C A (input) DOUBLE PRECISION array, dimension (LDA,N)
-C The leading N-by-N part of this array must contain the
-C system state matrix A of the shaped plant.
-C
-C LDA INTEGER
-C The leading dimension of the array A. LDA >= max(1,N).
-C
-C B (input) DOUBLE PRECISION array, dimension (LDB,M)
-C The leading N-by-M part of this array must contain the
-C system input matrix B of the shaped plant.
-C
-C LDB INTEGER
-C The leading dimension of the array B. LDB >= max(1,N).
-C
-C C (input) DOUBLE PRECISION array, dimension (LDC,N)
-C The leading NP-by-N part of this array must contain the
-C system output matrix C of the shaped plant.
-C
-C LDC INTEGER
-C The leading dimension of the array C. LDC >= max(1,NP).
-C
-C FACTOR (input) DOUBLE PRECISION
-C = 1 implies that an optimal controller is required;
-C > 1 implies that a suboptimal controller is required
-C achieving a performance FACTOR less than optimal.
-C FACTOR >= 1.
-C
-C AK (output) DOUBLE PRECISION array, dimension (LDAK,N)
-C The leading N-by-N part of this array contains the
-C controller state matrix Ak.
-C
-C LDAK INTEGER
-C The leading dimension of the array AK. LDAK >= max(1,N).
-C
-C BK (output) DOUBLE PRECISION array, dimension (LDBK,NP)
-C The leading N-by-NP part of this array contains the
-C controller input matrix Bk.
-C
-C LDBK INTEGER
-C The leading dimension of the array BK. LDBK >= max(1,N).
-C
-C CK (output) DOUBLE PRECISION array, dimension (LDCK,N)
-C The leading M-by-N part of this array contains the
-C controller output matrix Ck.
-C
-C LDCK INTEGER
-C The leading dimension of the array CK. LDCK >= max(1,M).
-C
-C DK (output) DOUBLE PRECISION array, dimension (LDDK,NP)
-C The leading M-by-NP part of this array contains the
-C controller matrix Dk.
-C
-C LDDK INTEGER
-C The leading dimension of the array DK. LDDK >= max(1,M).
-C
-C RCOND (output) DOUBLE PRECISION array, dimension (4)
-C RCOND(1) contains an estimate of the reciprocal condition
-C number of the linear system of equations from
-C which the solution of the P-Riccati equation is
-C obtained;
-C RCOND(2) contains an estimate of the reciprocal condition
-C number of the linear system of equations from
-C which the solution of the Q-Riccati equation is
-C obtained;
-C RCOND(3) contains an estimate of the reciprocal condition
-C number of the linear system of equations from
-C which the solution of the X-Riccati equation is
-C obtained;
-C RCOND(4) contains an estimate of the reciprocal condition
-C number of the matrix Rx + Bx'*X*Bx (see the
-C comments in the code).
-C
-C Workspace
-C
-C IWORK INTEGER array, dimension 2*max(N,NP+M)
-C
-C DWORK DOUBLE PRECISION array, dimension (LDWORK)
-C On exit, if INFO = 0, DWORK(1) contains the optimal value
-C of LDWORK.
-C
-C LDWORK INTEGER
-C The dimension of the array DWORK.
-C LDWORK >= 15*N*N + 6*N +
-C max( 14*N+23, 16*N, 2*N+NP+M, 3*(NP+M) ) +
-C max( N*N, 11*N*NP + 2*M*M + 8*NP*NP + 8*M*N +
-C 4*M*NP + NP ).
-C For good performance, LDWORK must generally be larger.
-C
-C BWORK LOGICAL array, dimension (2*N)
-C
-C Error Indicator
-C
-C INFO INTEGER
-C = 0: successful exit;
-C < 0: if INFO = -i, the i-th argument had an illegal
-C value;
-C = 1: the P-Riccati equation is not solved successfully;
-C = 2: the Q-Riccati equation is not solved successfully;
-C = 3: the X-Riccati equation is not solved successfully;
-C = 4: the iteration to compute eigenvalues failed to
-C converge;
-C = 5: the matrix Rx + Bx'*X*Bx is singular;
-C = 6: the closed-loop system is unstable.
-C
-C METHOD
-C
-C The routine implements the method presented in [1].
-C
-C REFERENCES
-C
-C [1] McFarlane, D. and Glover, K.
-C A loop shaping design procedure using H_infinity synthesis.
-C IEEE Trans. Automat. Control, vol. AC-37, no. 6, pp. 759-769,
-C 1992.
-C
-C NUMERICAL ASPECTS
-C
-C The accuracy of the results depends on the conditioning of the
-C two Riccati equations solved in the controller design. For
-C better conditioning it is advised to take FACTOR > 1.
-C
-C CONTRIBUTORS
-C
-C P.Hr. Petkov, D.W. Gu and M.M. Konstantinov, October 2000.
-C
-C REVISIONS
-C
-C V. Sima, Katholieke University Leuven, January 2001,
-C February 2001.
-C
-C KEYWORDS
-C
-C H_infinity control, Loop-shaping design, Robust control.
-C
-C ******************************************************************
-C
-C .. Parameters ..
- DOUBLE PRECISION ZERO, ONE
- PARAMETER ( ZERO = 0.0D+0, ONE = 1.0D+0 )
-C ..
-C .. Scalar Arguments ..
- INTEGER INFO, LDA, LDAK, LDB, LDBK, LDC, LDCK, LDDK,
- $ LDWORK, M, N, NP
- DOUBLE PRECISION FACTOR
-C ..
-C .. Array Arguments ..
- INTEGER IWORK( * )
- LOGICAL BWORK( * )
- DOUBLE PRECISION A( LDA, * ), AK( LDAK, * ), B( LDB, * ),
- $ BK( LDBK, * ), C( LDC, * ), CK( LDCK, * ),
- $ DK( LDDK, * ), DWORK( * ), RCOND( 4 )
-C ..
-C .. Local Scalars ..
- INTEGER I, I1, I2, I3, I4, I5, I6, I7, I8, I9, I10,
- $ I11, I12, I13, I14, I15, I16, I17, I18, I19,
- $ I20, I21, I22, I23, I24, I25, I26, INFO2,
- $ IWRK, J, LWA, LWAMAX, MINWRK, N2, NS, SDIM
- DOUBLE PRECISION GAMMA, RNORM
-C ..
-C .. External Functions ..
- LOGICAL SELECT
- DOUBLE PRECISION DLANSY, DLAPY2
- EXTERNAL DLANSY, DLAPY2, SELECT
-C ..
-C .. External Subroutines ..
- EXTERNAL DGEMM, DGEES, DLACPY, DLASET, DPOTRF, DPOTRS,
- $ DSYCON, DSYEV, DSYRK, DSYTRF, DSYTRS, SB02OD,
- $ XERBLA
-C ..
-C .. Intrinsic Functions ..
- INTRINSIC DBLE, INT, MAX, SQRT
-C ..
-C .. Executable Statements ..
-C
-C Decode and Test input parameters.
-C
- INFO = 0
- IF( N.LT.0 ) THEN
- INFO = -1
- ELSE IF( M.LT.0 ) THEN
- INFO = -2
- ELSE IF( NP.LT.0 ) THEN
- INFO = -3
- ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
- INFO = -5
- ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
- INFO = -7
- ELSE IF( LDC.LT.MAX( 1, NP ) ) THEN
- INFO = -9
- ELSE IF( FACTOR.LT.ONE ) THEN
- INFO = -10
- ELSE IF( LDAK.LT.MAX( 1, N ) ) THEN
- INFO = -12
- ELSE IF( LDBK.LT.MAX( 1, N ) ) THEN
- INFO = -14
- ELSE IF( LDCK.LT.MAX( 1, M ) ) THEN
- INFO = -16
- ELSE IF( LDDK.LT.MAX( 1, M ) ) THEN
- INFO = -18
- END IF
-C
-C Compute workspace.
-C
- MINWRK = 15*N*N + 6*N + MAX( 14*N+23, 16*N, 2*N+NP+M, 3*(NP+M) ) +
- $ MAX( N*N, 11*N*NP + 2*M*M + 8*NP*NP + 8*M*N +
- $ 4*M*NP + NP )
- IF( LDWORK.LT.MINWRK ) THEN
- INFO = -22
- END IF
- IF( INFO.NE.0 ) THEN
- CALL XERBLA( 'SB10KD', -INFO )
- RETURN
- END IF
-C
-C Quick return if possible.
-C
- IF( N.EQ.0 .OR. M.EQ.0 .OR. NP.EQ.0 ) THEN
- RCOND( 1 ) = ONE
- RCOND( 2 ) = ONE
- RCOND( 3 ) = ONE
- RCOND( 4 ) = ONE
- DWORK( 1 ) = ONE
- RETURN
- END IF
-C
-C Workspace usage.
-C
- N2 = 2*N
- I1 = N*N
- I2 = I1 + N*N
- I3 = I2 + N*N
- I4 = I3 + N*N
- I5 = I4 + N2
- I6 = I5 + N2
- I7 = I6 + N2
- I8 = I7 + N2*N2
- I9 = I8 + N2*N2
-C
- IWRK = I9 + N2*N2
- LWAMAX = 0
-C
-C Compute Cr = C'*C .
-C
- CALL DSYRK( 'U', 'T', N, NP, ONE, C, LDC, ZERO, DWORK( I2+1 ), N )
-C
-C Compute Dr = B*B' .
-C
- CALL DSYRK( 'U', 'N', N, M, ONE, B, LDB, ZERO, DWORK( I3+1 ), N )
-C -1
-C Solution of the Riccati equation A'*P*(In + Dr*P) *A - P + Cr = 0.
-C
- CALL SB02OD( 'D', 'G', 'N', 'U', 'Z', 'S', N, M, NP, A, LDA,
- $ DWORK( I3+1 ), N, DWORK( I2+1 ), N, DWORK, M, DWORK,
- $ N, RCOND( 1 ), DWORK, N, DWORK( I4+1 ),
- $ DWORK( I5+1 ), DWORK( I6+1 ), DWORK( I7+1 ), N2,
- $ DWORK( I8+1 ), N2, DWORK( I9+1 ), N2, -ONE, IWORK,
- $ DWORK( IWRK+1 ), LDWORK-IWRK, BWORK, INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 1
- RETURN
- END IF
- LWA = INT( DWORK( IWRK+1 ) ) + IWRK
- LWAMAX = MAX( LWA, LWAMAX )
-C
-C Transpose A in AK (used as workspace).
-C
- DO 40 J = 1, N
- DO 30 I = 1, N
- AK( I,J ) = A( J,I )
- 30 CONTINUE
- 40 CONTINUE
-C -1
-C Solution of the Riccati equation A*Q*(In + Cr*Q) *A' - Q + Dr = 0.
-C
- CALL SB02OD( 'D', 'G', 'N', 'U', 'Z', 'S', N, M, NP, AK, LDAK,
- $ DWORK( I2+1 ), N, DWORK( I3+1 ), N, DWORK, M, DWORK,
- $ N, RCOND( 2 ), DWORK( I1+1 ), N, DWORK( I4+1 ),
- $ DWORK( I5+1 ), DWORK( I6+1 ), DWORK( I7+1 ), N2,
- $ DWORK( I8+1 ), N2, DWORK( I9+1 ), N2, -ONE, IWORK,
- $ DWORK( IWRK+1 ), LDWORK-IWRK, BWORK, INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 2
- RETURN
- END IF
- LWA = INT( DWORK( IWRK+1 ) ) + IWRK
- LWAMAX = MAX( LWA, LWAMAX )
-C
-C Compute gamma.
-C
- CALL DGEMM( 'N', 'N', N, N, N, ONE, DWORK( I1+1 ), N, DWORK, N,
- $ ZERO, AK, LDAK )
- CALL DGEES( 'N', 'N', SELECT, N, AK, LDAK, SDIM, DWORK( I6+1 ),
- $ DWORK( I7+1 ), DWORK( IWRK+1 ), N, DWORK( IWRK+1 ),
- $ LDWORK-IWRK, BWORK, INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 4
- RETURN
- END IF
- LWA = INT( DWORK( IWRK+1 ) ) + IWRK
- LWAMAX = MAX( LWA, LWAMAX )
- GAMMA = ZERO
- DO 50 I = 1, N
- GAMMA = MAX( GAMMA, DWORK( I6+I ) )
- 50 CONTINUE
- GAMMA = FACTOR*SQRT( ONE + GAMMA )
-C
-C Workspace usage.
-C
- I3 = I2 + N*NP
- I4 = I3 + NP*NP
- I5 = I4 + NP*NP
- I6 = I5 + NP*NP
- I7 = I6 + NP
- I8 = I7 + NP*NP
- I9 = I8 + NP*NP
- I10 = I9 + NP*NP
- I11 = I10 + N*NP
- I12 = I11 + N*NP
- I13 = I12 + ( NP+M )*( NP+M )
- I14 = I13 + N*( NP+M )
- I15 = I14 + N*( NP+M )
- I16 = I15 + N*N
- I17 = I16 + N2
- I18 = I17 + N2
- I19 = I18 + N2
- I20 = I19 + ( N2+NP+M )*( N2+NP+M )
- I21 = I20 + ( N2+NP+M )*N2
-C
- IWRK = I21 + N2*N2
-C
-C Compute Q*C' .
-C
- CALL DGEMM( 'N', 'T', N, NP, N, ONE, DWORK( I1+1 ), N, C, LDC,
- $ ZERO, DWORK( I2+1 ), N )
-C
-C Compute Ip + C*Q*C' .
-C
- CALL DLASET( 'Full', NP, NP, ZERO, ONE, DWORK( I3+1 ), NP )
- CALL DGEMM( 'N', 'N', NP, NP, N, ONE, C, LDC, DWORK( I2+1 ), N,
- $ ONE, DWORK( I3+1 ), NP )
-C
-C Compute the eigenvalues and eigenvectors of Ip + C'*Q*C
-C
- CALL DLACPY( 'U', NP, NP, DWORK( I3+1 ), NP, DWORK( I5+1 ), NP )
- CALL DSYEV( 'V', 'U', NP, DWORK( I5+1 ), NP, DWORK( I6+1 ),
- $ DWORK( IWRK+1 ), LDWORK-IWRK, INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 4
- RETURN
- END IF
- LWA = INT( DWORK( IWRK+1 ) ) + IWRK
- LWAMAX = MAX( LWA, LWAMAX )
-C -1
-C Compute ( Ip + C'*Q*C ) .
-C
- DO 70 J = 1, NP
- DO 60 I = 1, NP
- DWORK( I9+I+(J-1)*NP ) = DWORK( I5+J+(I-1)*NP ) /
- $ DWORK( I6+I )
- 60 CONTINUE
- 70 CONTINUE
- CALL DGEMM( 'N', 'N', NP, NP, NP, ONE, DWORK( I5+1 ), NP,
- $ DWORK( I9+1 ), NP, ZERO, DWORK( I4+1 ), NP )
-C
-C Compute Z2 .
-C
- DO 90 J = 1, NP
- DO 80 I = 1, NP
- DWORK( I9+I+(J-1)*NP ) = DWORK( I5+J+(I-1)*NP ) /
- $ SQRT( DWORK( I6+I ) )
- 80 CONTINUE
- 90 CONTINUE
- CALL DGEMM( 'N', 'N', NP, NP, NP, ONE, DWORK( I5+1 ), NP,
- $ DWORK( I9+1 ), NP, ZERO, DWORK( I7+1 ), NP )
-C -1
-C Compute Z2 .
-C
- DO 110 J = 1, NP
- DO 100 I = 1, NP
- DWORK( I9+I+(J-1)*NP ) = DWORK( I5+J+(I-1)*NP )*
- $ SQRT( DWORK( I6+I ) )
- 100 CONTINUE
- 110 CONTINUE
- CALL DGEMM( 'N', 'N', NP, NP, NP, ONE, DWORK( I5+1 ), NP,
- $ DWORK( I9+1 ), NP, ZERO, DWORK( I8+1 ), NP )
-C
-C Compute A*Q*C' .
-C
- CALL DGEMM( 'N', 'N', N, NP, N, ONE, A, LDA, DWORK( I2+1 ), N,
- $ ZERO, DWORK( I10+1 ), N )
-C -1
-C Compute H = -A*Q*C'*( Ip + C*Q*C' ) .
-C
- CALL DGEMM( 'N', 'N', N, NP, NP, -ONE, DWORK( I10+1 ), N,
- $ DWORK( I4+1 ), NP, ZERO, DWORK( I11+1 ), N )
-C
-C Compute Rx .
-C
- CALL DLASET( 'F', NP+M, NP+M, ZERO, ONE, DWORK( I12+1 ), NP+M )
- DO 130 J = 1, NP
- DO 120 I = 1, NP
- DWORK( I12+I+(J-1)*(NP+M) ) = DWORK( I3+I+(J-1)*NP )
- 120 CONTINUE
- DWORK( I12+J+(J-1)*(NP+M) ) = DWORK( I3+J+(J-1)*NP ) -
- $ GAMMA*GAMMA
- 130 CONTINUE
-C
-C Compute Bx .
-C
- CALL DGEMM( 'N', 'N', N, NP, NP, -ONE, DWORK( I11+1 ), N,
- $ DWORK( I8+1 ), NP, ZERO, DWORK( I13+1 ), N )
- DO 150 J = 1, M
- DO 140 I = 1, N
- DWORK( I13+N*NP+I+(J-1)*N ) = B( I, J )
- 140 CONTINUE
- 150 CONTINUE
-C
-C Compute Sx .
-C
- CALL DGEMM( 'T', 'N', N, NP, NP, ONE, C, LDC, DWORK( I8+1 ), NP,
- $ ZERO, DWORK( I14+1 ), N )
- CALL DLASET( 'F', N, M, ZERO, ZERO, DWORK( I14+N*NP+1 ), N )
-C
-C Solve the Riccati equation
-C -1
-C X = A'*X*A + Cx - (Sx + A'*X*Bx)*(Rx + Bx'*X*B ) *(Sx'+Bx'*X*A).
-C
- CALL SB02OD( 'D', 'B', 'C', 'U', 'N', 'S', N, NP+M, NP, A, LDA,
- $ DWORK( I13+1 ), N, C, LDC, DWORK( I12+1 ), NP+M,
- $ DWORK( I14+1 ), N, RCOND( 3 ), DWORK( I15+1 ), N,
- $ DWORK( I16+1 ), DWORK( I17+1 ), DWORK( I18+1 ),
- $ DWORK( I19+1 ), N2+NP+M, DWORK( I20+1 ), N2+NP+M,
- $ DWORK( I21+1 ), N2, -ONE, IWORK, DWORK( IWRK+1 ),
- $ LDWORK-IWRK, BWORK, INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 3
- RETURN
- END IF
- LWA = INT( DWORK( IWRK+1 ) ) + IWRK
- LWAMAX = MAX( LWA, LWAMAX )
-C
- I22 = I16
- I23 = I22 + ( NP+M )*N
- I24 = I23 + ( NP+M )*( NP+M )
- I25 = I24 + ( NP+M )*N
- I26 = I25 + M*N
-C
- IWRK = I25
-C
-C Compute Bx'*X .
-C
- CALL DGEMM( 'T', 'N', NP+M, N, N, ONE, DWORK( I13+1 ), N,
- $ DWORK( I15+1 ), N, ZERO, DWORK( I22+1 ), NP+M )
-C
-C Compute Rx + Bx'*X*Bx .
-C
- CALL DLACPY( 'F', NP+M, NP+M, DWORK( I12+1 ), NP+M,
- $ DWORK( I23+1 ), NP+M )
- CALL DGEMM( 'N', 'N', NP+M, NP+M, N, ONE, DWORK( I22+1 ), NP+M,
- $ DWORK( I13+1 ), N, ONE, DWORK( I23+1 ), NP+M )
-C
-C Compute -( Sx' + Bx'*X*A ) .
-C
- DO 170 J = 1, N
- DO 160 I = 1, NP+M
- DWORK( I24+I+(J-1)*(NP+M) ) = DWORK( I14+J+(I-1)*N )
- 160 CONTINUE
- 170 CONTINUE
- CALL DGEMM( 'N', 'N', NP+M, N, N, -ONE, DWORK( I22+1 ), NP+M,
- $ A, LDA, -ONE, DWORK( I24+1 ), NP+M )
-C
-C Factorize Rx + Bx'*X*Bx .
-C
- RNORM = DLANSY( '1', 'U', NP+M, DWORK( I23+1 ), NP+M,
- $ DWORK( IWRK+1 ) )
- CALL DSYTRF( 'U', NP+M, DWORK( I23+1 ), NP+M, IWORK,
- $ DWORK( IWRK+1 ), LDWORK-IWRK, INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 5
- RETURN
- END IF
- LWA = INT( DWORK( IWRK+1 ) ) + IWRK
- LWAMAX = MAX( LWA, LWAMAX )
- CALL DSYCON( 'U', NP+M, DWORK( I23+1 ), NP+M, IWORK, RNORM,
- $ RCOND( 4 ), DWORK( IWRK+1 ), IWORK( NP+M+1), INFO2 )
-C -1
-C Compute F = -( Rx + Bx'*X*Bx ) ( Sx' + Bx'*X*A ) .
-C
- CALL DSYTRS( 'U', NP+M, N, DWORK( I23+1 ), NP+M, IWORK,
- $ DWORK( I24+1 ), NP+M, INFO2 )
-C
-C Compute B'*X .
-C
- CALL DGEMM( 'T', 'N', M, N, N, ONE, B, LDB, DWORK( I15+1 ), N,
- $ ZERO, DWORK( I25+1 ), M )
-C
-C Compute Im + B'*X*B .
-C
- CALL DLASET( 'F', M, M, ZERO, ONE, DWORK( I23+1 ), M )
- CALL DGEMM( 'N', 'N', M, M, N, ONE, DWORK( I25+1 ), M, B, LDB,
- $ ONE, DWORK( I23+1 ), M )
-C
-C Factorize Im + B'*X*B .
-C
- CALL DPOTRF( 'U', M, DWORK( I23+1 ), M, INFO2 )
-C -1
-C Compute ( Im + B'*X*B ) B'*X .
-C
- CALL DPOTRS( 'U', M, N, DWORK( I23+1 ), M, DWORK( I25+1 ), M,
- $ INFO2 )
-C -1
-C Compute Dk = ( Im + B'*X*B ) B'*X*H .
-C
- CALL DGEMM( 'N', 'N', M, NP, N, ONE, DWORK( I25+1 ), M,
- $ DWORK( I11+1 ), N, ZERO, DK, LDDK )
-C
-C Compute Bk = -H + B*Dk .
-C
- CALL DLACPY( 'F', N, NP, DWORK( I11+1 ), N, BK, LDBK )
- CALL DGEMM( 'N', 'N', N, NP, M, ONE, B, LDB, DK, LDDK, -ONE,
- $ BK, LDBK )
-C -1
-C Compute Dk*Z2 .
-C
- CALL DGEMM( 'N', 'N', M, NP, NP, ONE, DK, LDDK, DWORK( I8+1 ),
- $ NP, ZERO, DWORK( I26+1 ), M )
-C
-C Compute F1 + Z2*C .
-C
- CALL DLACPY( 'F', NP, N, DWORK( I24+1 ), NP+M, DWORK( I12+1 ),
- $ NP )
- CALL DGEMM( 'N', 'N', NP, N, NP, ONE, DWORK( I7+1 ), NP, C, LDC,
- $ ONE, DWORK( I12+1 ), NP )
-C -1
-C Compute Ck = F2 - Dk*Z2 *( F1 + Z2*C ) .
-C
- CALL DLACPY( 'F', M, N, DWORK( I24+NP+1 ), NP+M, CK, LDCK )
- CALL DGEMM( 'N', 'N', M, N, NP, -ONE, DWORK( I26+1 ), M,
- $ DWORK( I12+1 ), NP, ONE, CK, LDCK )
-C
-C Compute Ak = A + H*C + B*Ck .
-C
- CALL DLACPY( 'F', N, N, A, LDA, AK, LDAK )
- CALL DGEMM( 'N', 'N', N, N, NP, ONE, DWORK( I11+1 ), N, C, LDC,
- $ ONE, AK, LDAK )
- CALL DGEMM( 'N', 'N', N, N, M, ONE, B, LDB, CK, LDCK, ONE, AK,
- $ LDAK )
-C
-C Workspace usage.
-C
- I1 = M*N
- I2 = I1 + N2*N2
- I3 = I2 + N2
-C
- IWRK = I3 + N2
-C
-C Compute Dk*C .
-C
- CALL DGEMM( 'N', 'N', M, N, NP, ONE, DK, LDDK, C, LDC, ZERO,
- $ DWORK, M )
-C
-C Compute the closed-loop state matrix.
-C
- CALL DLACPY( 'F', N, N, A, LDA, DWORK( I1+1 ), N2 )
- CALL DGEMM( 'N', 'N', N, N, M, -ONE, B, LDB, DWORK, M, ONE,
- $ DWORK( I1+1 ), N2 )
- CALL DGEMM( 'N', 'N', N, N, NP, -ONE, BK, LDBK, C, LDC, ZERO,
- $ DWORK( I1+N+1 ), N2 )
- CALL DGEMM( 'N', 'N', N, N, M, ONE, B, LDB, CK, LDCK, ZERO,
- $ DWORK( I1+N2*N+1 ), N2 )
- CALL DLACPY( 'F', N, N, AK, LDAK, DWORK( I1+N2*N+N+1 ), N2 )
-C
-C Compute the closed-loop poles.
-C
- CALL DGEES( 'N', 'N', SELECT, N2, DWORK( I1+1 ), N2, SDIM,
- $ DWORK( I2+1 ), DWORK( I3+1 ), DWORK( IWRK+1 ), N,
- $ DWORK( IWRK+1 ), LDWORK-IWRK, BWORK, INFO2 )
- IF( INFO2.NE.0 ) THEN
- INFO = 4
- RETURN
- END IF
- LWA = INT( DWORK( IWRK+1 ) ) + IWRK
- LWAMAX = MAX( LWA, LWAMAX )
-C
-C Check the stability of the closed-loop system.
-C
- NS = 0
- DO 180 I = 1, N2
- IF( DLAPY2( DWORK( I2+I ), DWORK( I3+I ) ).GT.ONE ) NS = NS + 1
- 180 CONTINUE
- IF( NS.GT.0 ) THEN
- INFO = 6
- RETURN
- END IF
-C
- DWORK( 1 ) = DBLE( LWAMAX )
- ...
[truncated message content] |
|
From: <par...@us...> - 2011-08-06 19:07:50
|
Revision: 8444
http://octave.svn.sourceforge.net/octave/?rev=8444&view=rev
Author: paramaniac
Date: 2011-08-06 19:07:44 +0000 (Sat, 06 Aug 2011)
Log Message:
-----------
control: improve texinfo strings (2)
Modified Paths:
--------------
trunk/octave-forge/main/control/devel/pzmap2.m
trunk/octave-forge/main/control/inst/@lti/pole.m
trunk/octave-forge/main/control/inst/@lti/zero.m
trunk/octave-forge/main/control/inst/pzmap.m
Modified: trunk/octave-forge/main/control/devel/pzmap2.m
===================================================================
--- trunk/octave-forge/main/control/devel/pzmap2.m 2011-08-06 18:56:03 UTC (rev 8443)
+++ trunk/octave-forge/main/control/devel/pzmap2.m 2011-08-06 19:07:44 UTC (rev 8444)
@@ -19,6 +19,22 @@
## @deftypefn {Function File} pzmap (@var{sys})
## @deftypefnx {Function File} {[@var{p}, @var{z}] =} pzmap (@var{sys})
## Plot the poles and zeros of an LTI system in the complex plane.
+## If no output arguments are given, the result is plotted on the screen.
+## Otherwise, the poles and zeros are computed and returned.
+##
+## @strong{Inputs}
+## @table @var
+## @item sys
+## LTI model.
+## @end table
+##
+## @strong{Outputs}
+## @table @var
+## @item p
+## Poles of @var{sys}.
+## @item z
+## Transmission zeros of @var{sys}.
+## @end table
## @end deftypefn
## Author: Lukas Reichlin <luk...@gm...>
@@ -67,4 +83,4 @@
zer_r = zer{1};
endif
-endfunction
\ No newline at end of file
+endfunction
Modified: trunk/octave-forge/main/control/inst/@lti/pole.m
===================================================================
--- trunk/octave-forge/main/control/inst/@lti/pole.m 2011-08-06 18:56:03 UTC (rev 8443)
+++ trunk/octave-forge/main/control/inst/@lti/pole.m 2011-08-06 19:07:44 UTC (rev 8444)
@@ -18,6 +18,18 @@
## -*- texinfo -*-
## @deftypefn {Function File} {@var{p} =} pole (@var{sys})
## Compute poles of LTI system.
+##
+## @strong{Inputs}
+## @table @var
+## @item sys
+## LTI model.
+## @end table
+##
+## @strong{Outputs}
+## @table @var
+## @item p
+## Poles of @var{sys}.
+## @end table
## @end deftypefn
## Author: Lukas Reichlin <luk...@gm...>
@@ -32,4 +44,4 @@
pol = __pole__ (sys);
-endfunction
\ No newline at end of file
+endfunction
Modified: trunk/octave-forge/main/control/inst/@lti/zero.m
===================================================================
--- trunk/octave-forge/main/control/inst/@lti/zero.m 2011-08-06 18:56:03 UTC (rev 8443)
+++ trunk/octave-forge/main/control/inst/@lti/zero.m 2011-08-06 19:07:44 UTC (rev 8444)
@@ -19,6 +19,20 @@
## @deftypefn {Function File} {@var{z} =} zero (@var{sys})
## @deftypefnx {Function File} {[@var{z}, @var{k}] =} zero (@var{sys})
## Compute transmission zeros and gain of LTI model.
+##
+## @strong{Inputs}
+## @table @var
+## @item sys
+## LTI model.
+## @end table
+##
+## @strong{Outputs}
+## @table @var
+## @item z
+## Transmission zeros of @var{sys}.
+## @item k
+## Gain of @var{sys}.
+## @end table
## @end deftypefn
## Author: Lukas Reichlin <luk...@gm...>
@@ -33,4 +47,4 @@
[zer, gain] = __zero__ (sys, nargout);
-endfunction
\ No newline at end of file
+endfunction
Modified: trunk/octave-forge/main/control/inst/pzmap.m
===================================================================
--- trunk/octave-forge/main/control/inst/pzmap.m 2011-08-06 18:56:03 UTC (rev 8443)
+++ trunk/octave-forge/main/control/inst/pzmap.m 2011-08-06 19:07:44 UTC (rev 8444)
@@ -19,6 +19,22 @@
## @deftypefn {Function File} pzmap (@var{sys})
## @deftypefnx {Function File} {[@var{p}, @var{z}] =} pzmap (@var{sys})
## Plot the poles and zeros of an LTI system in the complex plane.
+## If no output arguments are given, the result is plotted on the screen.
+## Otherwise, the poles and zeros are computed and returned.
+##
+## @strong{Inputs}
+## @table @var
+## @item sys
+## LTI model.
+## @end table
+##
+## @strong{Outputs}
+## @table @var
+## @item p
+## Poles of @var{sys}.
+## @item z
+## Transmission zeros of @var{sys}.
+## @end table
## @end deftypefn
## Author: Lukas Reichlin <luk...@gm...>
@@ -56,4 +72,4 @@
zer_r = zer;
endif
-endfunction
\ No newline at end of file
+endfunction
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-08-07 04:58:11
|
Revision: 8446
http://octave.svn.sourceforge.net/octave/?rev=8446&view=rev
Author: paramaniac
Date: 2011-08-07 04:58:05 +0000 (Sun, 07 Aug 2011)
Log Message:
-----------
control: touch up example
Modified Paths:
--------------
trunk/octave-forge/main/control/devel/pdfdoc/functions.texi
trunk/octave-forge/main/control/inst/MDSSystem.m
Modified: trunk/octave-forge/main/control/devel/pdfdoc/functions.texi
===================================================================
--- trunk/octave-forge/main/control/devel/pdfdoc/functions.texi 2011-08-06 19:15:53 UTC (rev 8445)
+++ trunk/octave-forge/main/control/devel/pdfdoc/functions.texi 2011-08-07 04:58:05 UTC (rev 8446)
@@ -126,6 +126,8 @@
@strong{Inputs}
@table @var
+ @item sys
+ LTI model to be converted to state-space.
@item a
State transition matrix (n-by-n).
@item b
@@ -172,7 +174,8 @@
@strong{Inputs}
@table @var
@item sys
- LTI model. If second argument @var{w} is omitted, the interesting
+ LTI model to be converted to frequency response data.
+ If second argument @var{w} is omitted, the interesting
frequency range is calculated by the zeros and poles of @var{sys}.
@item H
Frequency response array (p-by-m-by-lw). In the SISO case,
@@ -207,6 +210,8 @@
@strong{Inputs}
@table @var
+ @item sys
+ LTI model to be converted to state-space.
@item a
State transition matrix (n-by-n).
@item b
@@ -280,6 +285,8 @@
@strong{Inputs}
@table @var
+ @item sys
+ LTI model to be converted to transfer function.
@item num
Numerator or cell of numerators. Each numerator must be a row vector
containing the exponents of the polynomial in descending order.
@@ -704,7 +711,36 @@
@deftypefn {Function File} {@var{co} =} ctrb (@var{sys})
@deftypefnx {Function File} {@var{co} =} ctrb (@var{a}, @var{b})
+ Return controllability matrix.
+
+ @strong{Inputs}
+ @table @var
+ @item sys
+ LTI model.
+ @item a
+ State transition matrix (n-by-n).
+ @item b
+ Input matrix (n-by-m).
+ @end table
+
+ @strong{Outputs}
+ @table @var
+ @item co
Controllability matrix.
+ @end table
+
+ @strong{Equation}
+ @iftex
+ @tex
+ $$ C_o = [ B AB A^2B ldots A^{n-1}B ] $$
+ @end tex
+ @end iftex
+ @ifinfo
+ @example
+ 2 n-1
+ Co = [ B AB A B ... A B ]
+ @end example
+ @end ifinfo
@end deftypefn
@subsection @@lti/dcgain
@@ -752,6 +788,20 @@
@deftypefn {Function File} {@var{bool} =} isct (@var{sys})
Determine whether LTI model is a continuous-time system.
+
+ @strong{Inputs}
+ @table @var
+ @item sys
+ LTI system.
+ @end table
+
+ @strong{Outputs}
+ @table @var
+ @item bool = 0
+ @var{sys} is a discrete-time system.
+ @item bool = 1
+ @var{sys} is a continuous-time system or a static gain.
+ @end table
@end deftypefn
@subsection isctrb
@@ -839,6 +889,20 @@
@deftypefn {Function File} {@var{bool} =} isdt (@var{sys})
Determine whether LTI model is a discrete-time system.
+
+ @strong{Inputs}
+ @table @var
+ @item sys
+ LTI system.
+ @end table
+
+ @strong{Outputs}
+ @table @var
+ @item bool = 0
+ @var{sys} is a continuous-time system.
+ @item bool = 1
+ @var{sys} is a discrete-time system or a static gain.
+ @end table
@end deftypefn
@subsection @@lti/isminimumphase
@@ -963,31 +1027,134 @@
@deftypefn {Function File} {@var{ob} =} obsv (@var{sys})
@deftypefnx {Function File} {@var{ob} =} obsv (@var{a}, @var{c})
+ Return observability matrix.
+
+ @strong{Inputs}
+ @table @var
+ @item sys
+ LTI model.
+ @item a
+ State transition matrix (n-by-n).
+ @item c
+ Measurement matrix (p-by-n).
+ @end table
+
+ @strong{Outputs}
+ @table @var
+ @item ob
Observability matrix.
+ @end table
+
+ @strong{Equation}
+ @iftex
+ @tex
+ $$ O_b = left[ matrix{ C cr
+ CA cr
+ CA^2 cr
+ dots cr
+ CA^{n-1} }
+ight ] $$
+ @end tex
+ @end iftex
+ @ifinfo
+ @example
+ @group
+ | C |
+ | CA |
+ Ob = | CA^2 |
+ | ... |
+ | CA^(n-1) |
+ @end group
+ @end example
+ @end ifinfo
@end deftypefn
@subsection @@lti/pole
@deftypefn {Function File} {@var{p} =} pole (@var{sys})
Compute poles of LTI system.
+
+ @strong{Inputs}
+ @table @var
+ @item sys
+ LTI model.
+ @end table
+
+ @strong{Outputs}
+ @table @var
+ @item p
+ Poles of @var{sys}.
+ @end table
@end deftypefn
@subsection pzmap
@deftypefn {Function File} pzmap (@var{sys})
@deftypefnx {Function File} {[@var{p}, @var{z}] =} pzmap (@var{sys})
Plot the poles and zeros of an LTI system in the complex plane.
+ If no output arguments are given, the result is plotted on the screen.
+ Otherwise, the poles and zeros are computed and returned.
+
+ @strong{Inputs}
+ @table @var
+ @item sys
+ LTI model.
+ @end table
+
+ @strong{Outputs}
+ @table @var
+ @item p
+ Poles of @var{sys}.
+ @item z
+ Transmission zeros of @var{sys}.
+ @end table
@end deftypefn
@subsection @@lti/size
- @deftypefn {Function File} {@var{nvec} =} size (@var{ltisys})
- @deftypefnx {Function File} {@var{n} =} size (@var{ltisys}, @var{idx})
- @deftypefnx {Function File} {[@var{ny}, @var{nu}] =} size (@var{ltisys})
+ @deftypefn {Function File} {@var{nvec} =} size (@var{sys})
+ @deftypefnx {Function File} {@var{n} =} size (@var{sys}, @var{dim})
+ @deftypefnx {Function File} {[@var{p}, @var{m}] =} size (@var{sys})
LTI model size, i.e. number of outputs and inputs.
+
+ @strong{Inputs}
+ @table @var
+ @item sys
+ LTI system.
+ @item dim
+ If given a second argument, @command{size} will return the size of the
+ corresponding dimension.
+ @end table
+
+ @strong{Outputs}
+ @table @var
+ @item nvec
+ Row vector. The first element is the number of outputs (rows) and the second
+ element the number of inputs (columns).
+ @item n
+ Scalar value. The size of the dimension @var{dim}.
+ @item p
+ Number of outputs.
+ @item m
+ Number of inputs.
+ @end table
@end deftypefn
@subsection @@lti/zero
@deftypefn {Function File} {@var{z} =} zero (@var{sys})
@deftypefnx {Function File} {[@var{z}, @var{k}] =} zero (@var{sys})
Compute transmission zeros and gain of LTI model.
+
+ @strong{Inputs}
+ @table @var
+ @item sys
+ LTI model.
+ @end table
+
+ @strong{Outputs}
+ @table @var
+ @item z
+ Transmission zeros of @var{sys}.
+ @item k
+ Gain of @var{sys}.
+ @end table
@end deftypefn
@section Model Simplification
@subsection @@lti/minreal
Modified: trunk/octave-forge/main/control/inst/MDSSystem.m
===================================================================
--- trunk/octave-forge/main/control/inst/MDSSystem.m 2011-08-06 19:15:53 UTC (rev 8445)
+++ trunk/octave-forge/main/control/inst/MDSSystem.m 2011-08-07 04:58:05 UTC (rev 8446)
@@ -126,7 +126,7 @@
sigma (T) % singular values
figure (3)
-step (T) % step response
+step (T, 10) % step response for 10 seconds
% ===============================================================================
@@ -151,6 +151,6 @@
sigma (T) % singular values
figure (5)
-step (T) % step response
+step (T, 10) % step response for 10 seconds
% ===============================================================================
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-08-08 14:35:00
|
Revision: 8448
http://octave.svn.sourceforge.net/octave/?rev=8448&view=rev
Author: paramaniac
Date: 2011-08-08 14:34:54 +0000 (Mon, 08 Aug 2011)
Log Message:
-----------
control: prepare release of control-2.1.53
Modified Paths:
--------------
trunk/octave-forge/main/control/DESCRIPTION
trunk/octave-forge/main/control/devel/RELEASE_PACKAGE
trunk/octave-forge/main/control/doc/NEWS
Modified: trunk/octave-forge/main/control/DESCRIPTION
===================================================================
--- trunk/octave-forge/main/control/DESCRIPTION 2011-08-07 07:03:54 UTC (rev 8447)
+++ trunk/octave-forge/main/control/DESCRIPTION 2011-08-08 14:34:54 UTC (rev 8448)
@@ -1,6 +1,6 @@
Name: Control
-Version: 2.1.52
-Date: 2011-07-27
+Version: 2.1.53
+Date: 2011-08-08
Author: Lukas Reichlin <luk...@gm...>
Maintainer: Lukas Reichlin <luk...@gm...>
Title: Control Systems
Modified: trunk/octave-forge/main/control/devel/RELEASE_PACKAGE
===================================================================
--- trunk/octave-forge/main/control/devel/RELEASE_PACKAGE 2011-08-07 07:03:54 UTC (rev 8447)
+++ trunk/octave-forge/main/control/devel/RELEASE_PACKAGE 2011-08-08 14:34:54 UTC (rev 8448)
@@ -18,12 +18,12 @@
rm -R ~/octave/__TEMP__/control/devel
cd ~/octave/__TEMP__
grep -i version control/DESCRIPTION
-tar czf control-2.1.52.tar.gz control/
-md5 control-2.1.52.tar.gz
-md5 control-2.1.52.tar.gz > md5_control_pkg.txt
-uuencode control-2.1.52.tar.gz < control-2.1.52.tar.gz > control-2.1.52.tar.gz.uue
+tar czf control-2.1.53.tar.gz control/
+md5 control-2.1.53.tar.gz
+md5 control-2.1.53.tar.gz > md5_control_pkg.txt
+uuencode control-2.1.53.tar.gz < control-2.1.53.tar.gz > control-2.1.53.tar.gz.uue
octave -q --eval \
-"pkg install control-2.1.52.tar.gz"
+"pkg install control-2.1.53.tar.gz"
octave -q --eval \
"pkg load generate_html; generate_package_html ('control', 'control-html', 'octave-forge')"
tar czf control-html.tar.gz control-html
@@ -38,7 +38,7 @@
=====================================================================================
rm -R ~/octave/__TEMP__
-rm -R ~/octave/control-2.1.52
+rm -R ~/octave/control-2.1.53
=====================================================================================
Modified: trunk/octave-forge/main/control/doc/NEWS
===================================================================
--- trunk/octave-forge/main/control/doc/NEWS 2011-08-07 07:03:54 UTC (rev 8447)
+++ trunk/octave-forge/main/control/doc/NEWS 2011-08-08 14:34:54 UTC (rev 8448)
@@ -1,17 +1,21 @@
Summary of important user-visible changes for releases of the control package
===============================================================================
-control-2.x.yy Release Date: 2011-xx-yy Release Manager: Lukas Reichlin
+control-2.1.53 Release Date: 2011-08-08 Release Manager: Lukas Reichlin
===============================================================================
** ncfsyn
- -- Added support for McFarlane/Glover loop shaping design procedure. ncfsyn
- stands for Normalized Coprime Factor Synthesis.
+ -- Added support for McFarlane/Glover loop shaping design procedure.
+ "ncfsyn" stands for Normalized Coprime Factor Synthesis.
** MDSSystem
-- Added example script which demonstrates the usage of the robust control
commands "mixsyn" and "ncfsyn".
+** Texinfo help strings of several functions have been extended, although
+ documentation still leaves a lot to be desired.
+
+
===============================================================================
control-2.1.52 Release Date: 2011-07-27 Release Manager: Lukas Reichlin
===============================================================================
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-08-14 06:26:05
|
Revision: 8457
http://octave.svn.sourceforge.net/octave/?rev=8457&view=rev
Author: paramaniac
Date: 2011-08-14 06:25:58 +0000 (Sun, 14 Aug 2011)
Log Message:
-----------
control: partially fix generation of pdf manual
Modified Paths:
--------------
trunk/octave-forge/main/control/devel/pdfdoc/control.tex
trunk/octave-forge/main/control/devel/pdfdoc/functions.texi
trunk/octave-forge/main/control/devel/pdfdoc/usage.txt
trunk/octave-forge/main/control/inst/place.m
Modified: trunk/octave-forge/main/control/devel/pdfdoc/control.tex
===================================================================
--- trunk/octave-forge/main/control/devel/pdfdoc/control.tex 2011-08-10 13:20:17 UTC (rev 8456)
+++ trunk/octave-forge/main/control/devel/pdfdoc/control.tex 2011-08-14 06:25:58 UTC (rev 8457)
@@ -3,7 +3,7 @@
@setfilename control.info
@settitle Octave Control Systems Package
@afourpaper
-@set VERSION 2.1.52
+@set VERSION 2.1.53
@c @afourwide
@c %**end of header
Modified: trunk/octave-forge/main/control/devel/pdfdoc/functions.texi
===================================================================
--- trunk/octave-forge/main/control/devel/pdfdoc/functions.texi 2011-08-10 13:20:17 UTC (rev 8456)
+++ trunk/octave-forge/main/control/devel/pdfdoc/functions.texi 2011-08-14 06:25:58 UTC (rev 8457)
@@ -522,6 +522,33 @@
@deftypefnx {Function File} {@var{sys} =} feedback (@var{sys1}, @var{sys2}, @var{feedin}, @var{feedout}, @var{"+"})
Feedback connection of two LTI models.
+ @strong{Inputs}
+ @table @var
+ @item sys1
+ LTI model of forward transmission. @code{[p1, m1] = size (sys1)}.
+ @item sys2
+ LTI model of backward transmission.
+ If not specified, an identity matrix of appropriate size is taken.
+ @item feedin
+ Vector containing indices of inputs to @var{sys1} which are involved in the feedback loop.
+ The number of @var{feedin} indices and outputs of @var{sys2} must be equal.
+ If not specified, @code{1:m1} is taken.
+ @item feedout
+ Vector containing indices of outputs from @var{sys1} which are to be connected to @var{sys2}.
+ The number of @var{feedout} indices and inputs of @var{sys2} must be equal.
+ If not specified, @code{1:p1} is taken.
+ @item "+"
+ Positive feedback sign. If not specified, @var{"-"} for a negative feedback interconnection
+ is assumed. @var{+1} and @var{-1} are possible as well, but only from the third argument
+ onward due to ambiguity.
+ @end table
+
+ @strong{Outputs}
+ @table @var
+ @item sys
+ Resulting LTI model.
+ @end table
+
@strong{Block Diagram}
@example
@group
@@ -1742,7 +1769,11 @@
Place is also suitable to design estimator gains:
@group
L = place (A.', C.', p).'
- L = place (sys.', p).' @subsection rlocus
+ L = place (sys.', p).' # useful for discrete-time systems
+ @end group
+ @end example
+ @end deftypefn
+@subsection rlocus
@deftypefn {Function File} rlocus (@var{sys})
@deftypefnx {Function File} {[@var{rldata}, @var{k}] =} rlocus (@var{sys}[, @var{increment}, @var{min_k}, @var{max_k}])
@@ -2275,7 +2306,7 @@
@item factor
@code{factor = 1} implies that an optimal controller is required.
@code{factor > 1} implies that a suboptimal controller is required,
- achieving a performance taht is @var{factor} times less than optimal.
+ achieving a performance that is @var{factor} times less than optimal.
Default value is 1.
@end table
Modified: trunk/octave-forge/main/control/devel/pdfdoc/usage.txt
===================================================================
--- trunk/octave-forge/main/control/devel/pdfdoc/usage.txt 2011-08-10 13:20:17 UTC (rev 8456)
+++ trunk/octave-forge/main/control/devel/pdfdoc/usage.txt 2011-08-14 06:25:58 UTC (rev 8457)
@@ -1,3 +1,11 @@
+* Check that "pkg list" lists the packages generate_html and control.
+* Run function_doc within Octave. This script collects the Texinfo strings from all
+ functions listed in the package's INDEX file and writes them to the file functions.texi.
+ Don't edit the file functions.texi since your changes will be lost by the next run.
+* Adapt version number in control.tex
+* Run control.tex
+
+
tex control.tex
q
???
Modified: trunk/octave-forge/main/control/inst/place.m
===================================================================
--- trunk/octave-forge/main/control/inst/place.m 2011-08-10 13:20:17 UTC (rev 8456)
+++ trunk/octave-forge/main/control/inst/place.m 2011-08-14 06:25:58 UTC (rev 8457)
@@ -64,7 +64,7 @@
## Place is also suitable to design estimator gains:
## @group
## L = place (A.', C.', p).'
-## L = place (sys.', p).' % useful for discrete-time systems
+## L = place (sys.', p).' # useful for discrete-time systems
## @end group
## @end example
## @end deftypefn
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-08-15 17:06:34
|
Revision: 8460
http://octave.svn.sourceforge.net/octave/?rev=8460&view=rev
Author: paramaniac
Date: 2011-08-15 17:06:28 +0000 (Mon, 15 Aug 2011)
Log Message:
-----------
control: work on ss2tf conversion
Modified Paths:
--------------
trunk/octave-forge/main/control/devel/ss2tf/sltb04bd.cc
trunk/octave-forge/main/control/inst/@lti/size.m
trunk/octave-forge/main/control/inst/@ss/__sys2tf__.m
Modified: trunk/octave-forge/main/control/devel/ss2tf/sltb04bd.cc
===================================================================
--- trunk/octave-forge/main/control/devel/ss2tf/sltb04bd.cc 2011-08-15 16:44:45 UTC (rev 8459)
+++ trunk/octave-forge/main/control/devel/ss2tf/sltb04bd.cc 2011-08-15 17:06:28 UTC (rev 8460)
@@ -146,6 +146,9 @@
retval(1) = gd;
retval(2) = ignm;
retval(3) = igdm;
+ retval(4) = octave_value (md);
+ retval(5) = octave_value (p);
+ retval(6) = octave_value (m);
}
return retval;
Modified: trunk/octave-forge/main/control/inst/@lti/size.m
===================================================================
--- trunk/octave-forge/main/control/inst/@lti/size.m 2011-08-15 16:44:45 UTC (rev 8459)
+++ trunk/octave-forge/main/control/inst/@lti/size.m 2011-08-15 17:06:28 UTC (rev 8460)
@@ -71,7 +71,7 @@
else
stru = "s";
endif
- disp (sprintf ("LTI model with %d output%s and %d input%s.", ny, stry, nu, stru));
+ disp (sprintf ("LTI model with %d output%s and %d input%s.", p, stry, m, stru));
case 1
n = [p, m];
case 2
Modified: trunk/octave-forge/main/control/inst/@ss/__sys2tf__.m
===================================================================
--- trunk/octave-forge/main/control/inst/@ss/__sys2tf__.m 2011-08-15 16:44:45 UTC (rev 8459)
+++ trunk/octave-forge/main/control/inst/@ss/__sys2tf__.m 2011-08-15 17:06:28 UTC (rev 8460)
@@ -20,28 +20,32 @@
## Author: Lukas Reichlin <luk...@gm...>
## Created: October 2009
-## Version: 0.1.1
+## Version: 0.2
function [retsys, retlti] = __sys2tf__ (sys)
- if (! issiso (sys))
- error ("ss: ss2tf: MIMO case not implemented yet");
- endif
+ [a, b, c, d] = ssdata (sys); # system could be a descriptor model
- if (isempty (sys.a)) # static gain
- num = sys.d;
- den = 1;
- else # default case
- [zer, gain] = zero (sys);
- pol = pole (sys);
-
- num = gain * real (poly (zer));
- den = real (poly (pol));
- endif
+ [num, den, ign, igd, md, p, m] = sltb04bd (a, b, c, d);
+ num = reshape (num, md, p, m);
+ den = reshape (den, md, p, m);
+
+ num = mat2cell (num, md, ones(1,p), ones(1,m));
+ den = mat2cell (den, md, ones(1,p), ones(1,m));
+
+ num = squeeze (num);
+ den = squeeze (den);
+
+ ign = mat2cell (ign, ones(1,p), ones(1,m));
+ igd = mat2cell (igd, ones(1,p), ones(1,m));
+
+ num = cellfun (@(x, y) x(1:y+1), num, ign, "uniformoutput", false);
+ den = cellfun (@(x, y) x(1:y+1), den, igd, "uniformoutput", false);
+
retsys = tf (num, den, get (sys, "tsam")); # tsam needed to set appropriate tfvar
retlti = sys.lti; # preserve lti properties
## FIXME: sys = tf (ss (5))
-endfunction
\ No newline at end of file
+endfunction
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-08-15 21:24:52
|
Revision: 8463
http://octave.svn.sourceforge.net/octave/?rev=8463&view=rev
Author: paramaniac
Date: 2011-08-15 21:24:46 +0000 (Mon, 15 Aug 2011)
Log Message:
-----------
control: minor changes
Modified Paths:
--------------
trunk/octave-forge/main/control/devel/PROJECTS
trunk/octave-forge/main/control/devel/pdfdoc/control.tex
trunk/octave-forge/main/control/doc/NEWS
trunk/octave-forge/main/control/inst/@tf/display.m
Modified: trunk/octave-forge/main/control/devel/PROJECTS
===================================================================
--- trunk/octave-forge/main/control/devel/PROJECTS 2011-08-15 18:23:39 UTC (rev 8462)
+++ trunk/octave-forge/main/control/devel/PROJECTS 2011-08-15 21:24:46 UTC (rev 8463)
@@ -59,8 +59,6 @@
* Balancing: TB01TD ?
- * Use SLICOT TB04BD for @ss/__sys2tf__.m
-
------------------------
Transfer Function Models:
------------------------
@@ -93,4 +91,4 @@
Octave-like language.
</pre>
-</html>
\ No newline at end of file
+</html>
Modified: trunk/octave-forge/main/control/devel/pdfdoc/control.tex
===================================================================
--- trunk/octave-forge/main/control/devel/pdfdoc/control.tex 2011-08-15 18:23:39 UTC (rev 8462)
+++ trunk/octave-forge/main/control/devel/pdfdoc/control.tex 2011-08-15 21:24:46 UTC (rev 8463)
@@ -3,7 +3,7 @@
@setfilename control.info
@settitle Octave Control Systems Package
@afourpaper
-@set VERSION 2.1.53
+@set VERSION 2.1.54
@c @afourwide
@c %**end of header
Modified: trunk/octave-forge/main/control/doc/NEWS
===================================================================
--- trunk/octave-forge/main/control/doc/NEWS 2011-08-15 18:23:39 UTC (rev 8462)
+++ trunk/octave-forge/main/control/doc/NEWS 2011-08-15 21:24:46 UTC (rev 8463)
@@ -6,8 +6,8 @@
** tf
-- State-space to transfer function conversion uses now SLICOT TB04BD.
- -- Conversion of MIMO models supported.
- -- Usage: tf_sys = tf (ss_sys)
+ Conversion of MIMO models is now supported. Usage: tf_sys = tf (ss_sys)
+ -- Display an empty line between title and numerator for better readability.
===============================================================================
Modified: trunk/octave-forge/main/control/inst/@tf/display.m
===================================================================
--- trunk/octave-forge/main/control/inst/@tf/display.m 2011-08-15 18:23:39 UTC (rev 8462)
+++ trunk/octave-forge/main/control/inst/@tf/display.m 2011-08-15 21:24:46 UTC (rev 8463)
@@ -34,6 +34,7 @@
for nu = 1 : m
disp (["Transfer function \"", sysname, "\" from input \"", inname{nu}, "\" to output ..."]);
+ disp ("");
for ny = 1 : p
__disp_frac__ (sys.num{ny, nu}, sys.den{ny, nu}, sys.tfvar, outname{ny});
endfor
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-08-16 09:47:41
|
Revision: 8464
http://octave.svn.sourceforge.net/octave/?rev=8464&view=rev
Author: paramaniac
Date: 2011-08-16 09:47:35 +0000 (Tue, 16 Aug 2011)
Log Message:
-----------
control: display whether tf is static, continuous- or discrete-time
Modified Paths:
--------------
trunk/octave-forge/main/control/doc/NEWS
trunk/octave-forge/main/control/inst/@tf/display.m
Modified: trunk/octave-forge/main/control/doc/NEWS
===================================================================
--- trunk/octave-forge/main/control/doc/NEWS 2011-08-15 21:24:46 UTC (rev 8463)
+++ trunk/octave-forge/main/control/doc/NEWS 2011-08-16 09:47:35 UTC (rev 8464)
@@ -8,6 +8,7 @@
-- State-space to transfer function conversion uses now SLICOT TB04BD.
Conversion of MIMO models is now supported. Usage: tf_sys = tf (ss_sys)
-- Display an empty line between title and numerator for better readability.
+ -- Display whether model is static, continuous- or discrete-time.
===============================================================================
Modified: trunk/octave-forge/main/control/inst/@tf/display.m
===================================================================
--- trunk/octave-forge/main/control/inst/@tf/display.m 2011-08-15 21:24:46 UTC (rev 8463)
+++ trunk/octave-forge/main/control/inst/@tf/display.m 2011-08-16 09:47:35 UTC (rev 8464)
@@ -1,4 +1,4 @@
-## Copyright (C) 2009, 2010 Lukas F. Reichlin
+## Copyright (C) 2009, 2010, 2011 Lukas F. Reichlin
##
## This file is part of LTI Syncope.
##
@@ -20,12 +20,12 @@
## Author: Lukas Reichlin <luk...@gm...>
## Created: September 2009
-## Version: 0.2
+## Version: 0.3
function display (sys)
sysname = inputname (1);
- [inname, outname] = __lti_data__ (sys.lti);
+ [inname, outname, tsam] = __lti_data__ (sys.lti);
[inname, m] = __labels__ (inname, "u");
[outname, p] = __labels__ (outname, "y");
@@ -40,8 +40,16 @@
endfor
endfor
- display (sys.lti);
+ display (sys.lti); # display sampling time
+ if (tsam == -2)
+ disp ("Static gain.");
+ elseif (tsam == 0)
+ disp ("Continuous-time model.");
+ else
+ disp ("Discrete-time model.");
+ endif
+
endfunction
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-08-15 18:06:52
|
Revision: 8461
http://octave.svn.sourceforge.net/octave/?rev=8461&view=rev
Author: paramaniac
Date: 2011-08-15 18:06:43 +0000 (Mon, 15 Aug 2011)
Log Message:
-----------
control: move ss2tf conversion into place
Modified Paths:
--------------
trunk/octave-forge/main/control/devel/makefile_all.m
trunk/octave-forge/main/control/doc/NEWS
trunk/octave-forge/main/control/inst/@ss/__sys2tf__.m
trunk/octave-forge/main/control/inst/ltimodels.m
trunk/octave-forge/main/control/src/Makefile
Added Paths:
-----------
trunk/octave-forge/main/control/devel/makefile_ss2tf.m
trunk/octave-forge/main/control/src/MC01PD.f
trunk/octave-forge/main/control/src/MC01PY.f
trunk/octave-forge/main/control/src/TB01ZD.f
trunk/octave-forge/main/control/src/TB04BD.f
trunk/octave-forge/main/control/src/sltb04bd.cc
Removed Paths:
-------------
trunk/octave-forge/main/control/devel/ss2tf/MC01PD.f
trunk/octave-forge/main/control/devel/ss2tf/MC01PY.f
trunk/octave-forge/main/control/devel/ss2tf/TB01ZD.f
trunk/octave-forge/main/control/devel/ss2tf/TB04BD.dat
trunk/octave-forge/main/control/devel/ss2tf/TB04BD.f
trunk/octave-forge/main/control/devel/ss2tf/TB04BD.html
trunk/octave-forge/main/control/devel/ss2tf/TB04BD.res
trunk/octave-forge/main/control/devel/ss2tf/common.cc
trunk/octave-forge/main/control/devel/ss2tf/makefile_ss2tf.m
trunk/octave-forge/main/control/devel/ss2tf/sltb04bd.cc
Modified: trunk/octave-forge/main/control/devel/makefile_all.m
===================================================================
--- trunk/octave-forge/main/control/devel/makefile_all.m 2011-08-15 17:06:28 UTC (rev 8460)
+++ trunk/octave-forge/main/control/devel/makefile_all.m 2011-08-15 18:06:43 UTC (rev 8461)
@@ -23,7 +23,8 @@
makefile_norm
makefile_place
makefile_scale
+makefile_ss2tf
makefile_staircase
makefile_zero
-cd (homedir);
\ No newline at end of file
+cd (homedir);
Copied: trunk/octave-forge/main/control/devel/makefile_ss2tf.m (from rev 8460, trunk/octave-forge/main/control/devel/ss2tf/makefile_ss2tf.m)
===================================================================
--- trunk/octave-forge/main/control/devel/makefile_ss2tf.m (rev 0)
+++ trunk/octave-forge/main/control/devel/makefile_ss2tf.m 2011-08-15 18:06:43 UTC (rev 8461)
@@ -0,0 +1,19 @@
+## ==============================================================================
+## Developer Makefile for OCT-files
+## ==============================================================================
+## USAGE: * fetch control from Octave-Forge by svn
+## * add control/inst, control/src and control/devel to your Octave path
+## * run makefile_*
+## ==============================================================================
+
+homedir = pwd ();
+develdir = fileparts (which ("makefile_ss2tf"));
+srcdir = [develdir, "/../src"];
+cd (srcdir);
+
+mkoctfile sltb04bd.cc \
+ TB04BD.f MC01PY.f TB01ID.f TB01ZD.f MC01PD.f \
+ TB04BX.f MA02AD.f MB02RD.f MB01PD.f MB02SD.f \
+ MB01QD.f
+
+cd (homedir);
Deleted: trunk/octave-forge/main/control/devel/ss2tf/MC01PD.f
===================================================================
--- trunk/octave-forge/main/control/devel/ss2tf/MC01PD.f 2011-08-15 17:06:28 UTC (rev 8460)
+++ trunk/octave-forge/main/control/devel/ss2tf/MC01PD.f 2011-08-15 18:06:43 UTC (rev 8461)
@@ -1,159 +0,0 @@
- SUBROUTINE MC01PD( K, REZ, IMZ, P, DWORK, INFO )
-C
-C SLICOT RELEASE 5.0.
-C
-C Copyright (c) 2002-2009 NICONET e.V.
-C
-C This program is free software: you can redistribute it and/or
-C modify it under the terms of the GNU General Public License as
-C published by the Free Software Foundation, either version 2 of
-C the License, or (at your option) any later version.
-C
-C This program is distributed in the hope that it will be useful,
-C but WITHOUT ANY WARRANTY; without even the implied warranty of
-C MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-C GNU General Public License for more details.
-C
-C You should have received a copy of the GNU General Public License
-C along with this program. If not, see
-C <http://www.gnu.org/licenses/>.
-C
-C PURPOSE
-C
-C To compute the coefficients of a real polynomial P(x) from its
-C zeros.
-C
-C ARGUMENTS
-C
-C Input/Output Parameters
-C
-C K (input) INTEGER
-C The number of zeros (and hence the degree) of P(x).
-C K >= 0.
-C
-C REZ (input) DOUBLE PRECISION array, dimension (K)
-C IMZ (input) DOUBLE PRECISION array, dimension (K)
-C The real and imaginary parts of the i-th zero of P(x)
-C must be stored in REZ(i) and IMZ(i), respectively, where
-C i = 1, 2, ..., K. The zeros may be supplied in any order,
-C except that complex conjugate zeros must appear
-C consecutively.
-C
-C P (output) DOUBLE PRECISION array, dimension (K+1)
-C This array contains the coefficients of P(x) in increasing
-C powers of x. If K = 0, then P(1) is set to one.
-C
-C Workspace
-C
-C DWORK DOUBLE PRECISION array, dimension (K+1)
-C If K = 0, this array is not referenced.
-C
-C Error Indicator
-C
-C INFO INTEGER
-C = 0: successful exit;
-C < 0: if INFO = -i, the i-th argument had an illegal
-C value;
-C > 0: if INFO = i, (REZ(i),IMZ(i)) is a complex zero but
-C (REZ(i-1),IMZ(i-1)) is not its conjugate.
-C
-C METHOD
-C
-C The routine computes the coefficients of the real K-th degree
-C polynomial P(x) as
-C
-C P(x) = (x - r(1)) * (x - r(2)) * ... * (x - r(K))
-C
-C where r(i) = (REZ(i),IMZ(i)).
-C
-C Note that REZ(i) = REZ(j) and IMZ(i) = -IMZ(j) if r(i) and r(j)
-C form a complex conjugate pair (where i <> j), and that IMZ(i) = 0
-C if r(i) is real.
-C
-C NUMERICAL ASPECTS
-C
-C None.
-C
-C CONTRIBUTOR
-C
-C Release 3.0: V. Sima, Katholieke Univ. Leuven, Belgium, Mar. 1997.
-C Supersedes Release 2.0 routine MC01DD by A.J. Geurts.
-C
-C REVISIONS
-C
-C V. Sima, May 2002.
-C
-C KEYWORDS
-C
-C Elementary polynomial operations, polynomial operations.
-C
-C ******************************************************************
-C
-C .. Parameters ..
- DOUBLE PRECISION ZERO, ONE
- PARAMETER ( ZERO = 0.0D0, ONE = 1.0D0 )
-C .. Scalar Arguments ..
- INTEGER INFO, K
-C .. Array Arguments ..
- DOUBLE PRECISION DWORK(*), IMZ(*), P(*), REZ(*)
-C .. Local Scalars ..
- INTEGER I
- DOUBLE PRECISION U, V
-C .. External Subroutines ..
- EXTERNAL DAXPY, DCOPY, XERBLA
-C .. Executable Statements ..
-C
-C Test the input scalar arguments.
-C
- IF( K.LT.0 ) THEN
- INFO = -1
-C
-C Error return.
-C
- CALL XERBLA( 'MC01PD', -INFO )
- RETURN
- END IF
-C
-C Quick return if possible.
-C
- INFO = 0
- P(1) = ONE
- IF ( K.EQ.0 )
- $ RETURN
-C
- I = 1
-C WHILE ( I <= K ) DO
- 20 IF ( I.LE.K ) THEN
- U = REZ(I)
- V = IMZ(I)
- DWORK(1) = ZERO
-C
- IF ( V.EQ.ZERO ) THEN
- CALL DCOPY( I, P, 1, DWORK(2), 1 )
- CALL DAXPY( I, -U, P, 1, DWORK, 1 )
- I = I + 1
-C
- ELSE
- IF ( I.EQ.K ) THEN
- INFO = K
- RETURN
- ELSE IF ( ( U.NE.REZ(I+1) ) .OR. ( V.NE.-IMZ(I+1) ) ) THEN
- INFO = I + 1
- RETURN
- END IF
-C
- DWORK(2) = ZERO
- CALL DCOPY( I, P, 1, DWORK(3), 1 )
- CALL DAXPY( I, -(U + U), P, 1, DWORK(2), 1 )
- CALL DAXPY( I, U**2+V**2, P, 1, DWORK, 1 )
- I = I + 2
- END IF
-C
- CALL DCOPY( I, DWORK, 1, P, 1 )
- GO TO 20
- END IF
-C END WHILE 20
-C
- RETURN
-C *** Last line of MC01PD ***
- END
Deleted: trunk/octave-forge/main/control/devel/ss2tf/MC01PY.f
===================================================================
--- trunk/octave-forge/main/control/devel/ss2tf/MC01PY.f 2011-08-15 17:06:28 UTC (rev 8460)
+++ trunk/octave-forge/main/control/devel/ss2tf/MC01PY.f 2011-08-15 18:06:43 UTC (rev 8461)
@@ -1,157 +0,0 @@
- SUBROUTINE MC01PY( K, REZ, IMZ, P, DWORK, INFO )
-C
-C SLICOT RELEASE 5.0.
-C
-C Copyright (c) 2002-2009 NICONET e.V.
-C
-C This program is free software: you can redistribute it and/or
-C modify it under the terms of the GNU General Public License as
-C published by the Free Software Foundation, either version 2 of
-C the License, or (at your option) any later version.
-C
-C This program is distributed in the hope that it will be useful,
-C but WITHOUT ANY WARRANTY; without even the implied warranty of
-C MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-C GNU General Public License for more details.
-C
-C You should have received a copy of the GNU General Public License
-C along with this program. If not, see
-C <http://www.gnu.org/licenses/>.
-C
-C PURPOSE
-C
-C To compute the coefficients of a real polynomial P(x) from its
-C zeros. The coefficients are stored in decreasing order of the
-C powers of x.
-C
-C ARGUMENTS
-C
-C Input/Output Parameters
-C
-C K (input) INTEGER
-C The number of zeros (and hence the degree) of P(x).
-C K >= 0.
-C
-C REZ (input) DOUBLE PRECISION array, dimension (K)
-C IMZ (input) DOUBLE PRECISION array, dimension (K)
-C The real and imaginary parts of the i-th zero of P(x)
-C must be stored in REZ(i) and IMZ(i), respectively, where
-C i = 1, 2, ..., K. The zeros may be supplied in any order,
-C except that complex conjugate zeros must appear
-C consecutively.
-C
-C P (output) DOUBLE PRECISION array, dimension (K+1)
-C This array contains the coefficients of P(x) in decreasing
-C powers of x.
-C
-C Workspace
-C
-C DWORK DOUBLE PRECISION array, dimension (K)
-C If K = 0, this array is not referenced.
-C
-C Error Indicator
-C
-C INFO INTEGER
-C = 0: successful exit;
-C < 0: if INFO = -i, the i-th argument had an illegal
-C value;
-C > 0: if INFO = i, (REZ(i),IMZ(i)) is a complex zero but
-C (REZ(i-1),IMZ(i-1)) is not its conjugate.
-C
-C METHOD
-C
-C The routine computes the coefficients of the real K-th degree
-C polynomial P(x) as
-C
-C P(x) = (x - r(1)) * (x - r(2)) * ... * (x - r(K))
-C
-C where r(i) = (REZ(i),IMZ(i)).
-C
-C Note that REZ(i) = REZ(j) and IMZ(i) = -IMZ(j) if r(i) and r(j)
-C form a complex conjugate pair (where i <> j), and that IMZ(i) = 0
-C if r(i) is real.
-C
-C NUMERICAL ASPECTS
-C
-C None.
-C
-C CONTRIBUTOR
-C
-C V. Sima, Research Institute for Informatics, Bucharest, May 2002.
-C
-C REVISIONS
-C
-C -
-C
-C KEYWORDS
-C
-C Elementary polynomial operations, polynomial operations.
-C
-C ******************************************************************
-C
-C .. Parameters ..
- DOUBLE PRECISION ZERO, ONE
- PARAMETER ( ZERO = 0.0D0, ONE = 1.0D0 )
-C .. Scalar Arguments ..
- INTEGER INFO, K
-C .. Array Arguments ..
- DOUBLE PRECISION DWORK(*), IMZ(*), P(*), REZ(*)
-C .. Local Scalars ..
- INTEGER I
- DOUBLE PRECISION U, V
-C .. External Subroutines ..
- EXTERNAL DAXPY, DCOPY, XERBLA
-C .. Executable Statements ..
-C
-C Test the input scalar arguments.
-C
- IF( K.LT.0 ) THEN
- INFO = -1
-C
-C Error return.
-C
- CALL XERBLA( 'MC01PY', -INFO )
- RETURN
- END IF
-C
-C Quick return if possible.
-C
- INFO = 0
- P(1) = ONE
- IF ( K.EQ.0 )
- $ RETURN
-C
- I = 1
-C WHILE ( I <= K ) DO
- 20 IF ( I.LE.K ) THEN
- U = REZ(I)
- V = IMZ(I)
- DWORK(I) = ZERO
-C
- IF ( V.EQ.ZERO ) THEN
- CALL DAXPY( I, -U, P, 1, DWORK, 1 )
-C
- ELSE
- IF ( I.EQ.K ) THEN
- INFO = K
- RETURN
- ELSE IF ( ( U.NE.REZ(I+1) ) .OR. ( V.NE.-IMZ(I+1) ) ) THEN
- INFO = I + 1
- RETURN
- END IF
-C
- DWORK(I+1) = ZERO
- CALL DAXPY( I, -(U + U), P, 1, DWORK, 1 )
- CALL DAXPY( I, U**2+V**2, P, 1, DWORK(2), 1 )
- I = I + 1
- END IF
-C
- CALL DCOPY( I, DWORK, 1, P(2), 1 )
- I = I + 1
- GO TO 20
- END IF
-C END WHILE 20
-C
- RETURN
-C *** Last line of MC01PY ***
- END
Deleted: trunk/octave-forge/main/control/devel/ss2tf/TB01ZD.f
===================================================================
--- trunk/octave-forge/main/control/devel/ss2tf/TB01ZD.f 2011-08-15 17:06:28 UTC (rev 8460)
+++ trunk/octave-forge/main/control/devel/ss2tf/TB01ZD.f 2011-08-15 18:06:43 UTC (rev 8461)
@@ -1,440 +0,0 @@
- SUBROUTINE TB01ZD( JOBZ, N, P, A, LDA, B, C, LDC, NCONT, Z, LDZ,
- $ TAU, TOL, DWORK, LDWORK, INFO )
-C
-C SLICOT RELEASE 5.0.
-C
-C Copyright (c) 2002-2009 NICONET e.V.
-C
-C This program is free software: you can redistribute it and/or
-C modify it under the terms of the GNU General Public License as
-C published by the Free Software Foundation, either version 2 of
-C the License, or (at your option) any later version.
-C
-C This program is distributed in the hope that it will be useful,
-C but WITHOUT ANY WARRANTY; without even the implied warranty of
-C MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-C GNU General Public License for more details.
-C
-C You should have received a copy of the GNU General Public License
-C along with this program. If not, see
-C <http://www.gnu.org/licenses/>.
-C
-C PURPOSE
-C
-C To find a controllable realization for the linear time-invariant
-C single-input system
-C
-C dX/dt = A * X + B * U,
-C Y = C * X,
-C
-C where A is an N-by-N matrix, B is an N element vector, C is an
-C P-by-N matrix, and A and B are reduced by this routine to
-C orthogonal canonical form using (and optionally accumulating)
-C orthogonal similarity transformations, which are also applied
-C to C.
-C
-C ARGUMENTS
-C
-C Mode Parameters
-C
-C JOBZ CHARACTER*1
-C Indicates whether the user wishes to accumulate in a
-C matrix Z the orthogonal similarity transformations for
-C reducing the system, as follows:
-C = 'N': Do not form Z and do not store the orthogonal
-C transformations;
-C = 'F': Do not form Z, but store the orthogonal
-C transformations in the factored form;
-C = 'I': Z is initialized to the unit matrix and the
-C orthogonal transformation matrix Z is returned.
-C
-C Input/Output Parameters
-C
-C N (input) INTEGER
-C The order of the original state-space representation,
-C i.e. the order of the matrix A. N >= 0.
-C
-C P (input) INTEGER
-C The number of system outputs, or of rows of C. P >= 0.
-C
-C A (input/output) DOUBLE PRECISION array, dimension (LDA,N)
-C On entry, the leading N-by-N part of this array must
-C contain the original state dynamics matrix A.
-C On exit, the leading NCONT-by-NCONT upper Hessenberg
-C part of this array contains the canonical form of the
-C state dynamics matrix, given by Z' * A * Z, of a
-C controllable realization for the original system. The
-C elements below the first subdiagonal are set to zero.
-C
-C LDA INTEGER
-C The leading dimension of array A. LDA >= MAX(1,N).
-C
-C B (input/output) DOUBLE PRECISION array, dimension (N)
-C On entry, the original input/state vector B.
-C On exit, the leading NCONT elements of this array contain
-C canonical form of the input/state vector, given by Z' * B,
-C with all elements but B(1) set to zero.
-C
-C C (input/output) DOUBLE PRECISION array, dimension (LDC,N)
-C On entry, the leading P-by-N part of this array must
-C contain the output/state matrix C.
-C On exit, the leading P-by-N part of this array contains
-C the transformed output/state matrix, given by C * Z, and
-C the leading P-by-NCONT part contains the output/state
-C matrix of the controllable realization.
-C
-C LDC INTEGER
-C The leading dimension of array C. LDC >= MAX(1,P).
-C
-C NCONT (output) INTEGER
-C The order of the controllable state-space representation.
-C
-C Z (output) DOUBLE PRECISION array, dimension (LDZ,N)
-C If JOBZ = 'I', then the leading N-by-N part of this array
-C contains the matrix of accumulated orthogonal similarity
-C transformations which reduces the given system to
-C orthogonal canonical form.
-C If JOBZ = 'F', the elements below the diagonal, with the
-C array TAU, represent the orthogonal transformation matrix
-C as a product of elementary reflectors. The transformation
-C matrix can then be obtained by calling the LAPACK Library
-C routine DORGQR.
-C If JOBZ = 'N', the array Z is not referenced and can be
-C supplied as a dummy array (i.e. set parameter LDZ = 1 and
-C declare this array to be Z(1,1) in the calling program).
-C
-C LDZ INTEGER
-C The leading dimension of array Z. If JOBZ = 'I' or
-C JOBZ = 'F', LDZ >= MAX(1,N); if JOBZ = 'N', LDZ >= 1.
-C
-C TAU (output) DOUBLE PRECISION array, dimension (N)
-C The elements of TAU contain the scalar factors of the
-C elementary reflectors used in the reduction of B and A.
-C
-C Tolerances
-C
-C TOL DOUBLE PRECISION
-C The tolerance to be used in determining the
-C controllability of (A,B). If the user sets TOL > 0, then
-C the given value of TOL is used as an absolute tolerance;
-C elements with absolute value less than TOL are considered
-C neglijible. If the user sets TOL <= 0, then an implicitly
-C computed, default tolerance, defined by
-C TOLDEF = N*EPS*MAX( NORM(A), NORM(B) ) is used instead,
-C where EPS is the machine precision (see LAPACK Library
-C routine DLAMCH).
-C
-C Workspace
-C
-C DWORK DOUBLE PRECISION array, dimension (LDWORK)
-C On exit, if INFO = 0, DWORK(1) returns the optimal value
-C of LDWORK.
-C
-C LDWORK INTEGER
-C The length of the array DWORK. LDWORK >= MAX(1,N,P).
-C For optimum performance LDWORK should be larger.
-C
-C Error Indicator
-C
-C INFO INTEGER
-C = 0: successful exit;
-C < 0: if INFO = -i, the i-th argument had an illegal
-C value.
-C
-C METHOD
-C
-C The Householder matrix which reduces all but the first element
-C of vector B to zero is found and this orthogonal similarity
-C transformation is applied to the matrix A. The resulting A is then
-C reduced to upper Hessenberg form by a sequence of Householder
-C transformations. Finally, the order of the controllable state-
-C space representation (NCONT) is determined by finding the position
-C of the first sub-diagonal element of A which is below an
-C appropriate zero threshold, either TOL or TOLDEF (see parameter
-C TOL); if NORM(B) is smaller than this threshold, NCONT is set to
-C zero, and no computations for reducing the system to orthogonal
-C canonical form are performed.
-C All orthogonal transformations determined in this process are also
-C applied to the matrix C, from the right.
-C
-C REFERENCES
-C
-C [1] Konstantinov, M.M., Petkov, P.Hr. and Christov, N.D.
-C Orthogonal Invariants and Canonical Forms for Linear
-C Controllable Systems.
-C Proc. 8th IFAC World Congress, Kyoto, 1, pp. 49-54, 1981.
-C
-C [2] Hammarling, S.J.
-C Notes on the use of orthogonal similarity transformations in
-C control.
-C NPL Report DITC 8/82, August 1982.
-C
-C [3] Paige, C.C
-C Properties of numerical algorithms related to computing
-C controllability.
-C IEEE Trans. Auto. Contr., AC-26, pp. 130-138, 1981.
-C
-C NUMERICAL ASPECTS
-C 3
-C The algorithm requires 0(N ) operations and is backward stable.
-C
-C CONTRIBUTOR
-C
-C V. Sima, Katholieke Univ. Leuven, Belgium, Feb. 1998.
-C
-C REVISIONS
-C
-C V. Sima, Research Institute for Informatics, Bucharest, Oct. 2001,
-C Sept. 2003.
-C
-C KEYWORDS
-C
-C Controllability, minimal realization, orthogonal canonical form,
-C orthogonal transformation.
-C
-C ******************************************************************
-C
-C .. Parameters ..
- DOUBLE PRECISION ZERO, ONE
- PARAMETER ( ZERO = 0.0D0, ONE = 1.0D0 )
-C .. Scalar Arguments ..
- CHARACTER JOBZ
- INTEGER INFO, LDA, LDC, LDWORK, LDZ, N, NCONT, P
- DOUBLE PRECISION TOL
-C .. Array Arguments ..
- DOUBLE PRECISION A(LDA,*), B(*), C(LDC,*), DWORK(*), TAU(*),
- $ Z(LDZ,*)
-C .. Local Scalars ..
- LOGICAL LJOBF, LJOBI, LJOBZ
- INTEGER ITAU, J
- DOUBLE PRECISION ANORM, B1, BNORM, FANORM, FBNORM, H, THRESH,
- $ TOLDEF, WRKOPT
-C .. Local Arrays ..
- DOUBLE PRECISION NBLK(1)
-C .. External Functions ..
- LOGICAL LSAME
- DOUBLE PRECISION DLAMCH, DLANGE
- EXTERNAL DLAMCH, DLANGE, LSAME
-C .. External Subroutines ..
- EXTERNAL DGEHRD, DLACPY, DLARF, DLARFG, DLASET, DORGQR,
- $ DORMHR, MB01PD, XERBLA
-C .. Intrinsic Functions ..
- INTRINSIC ABS, DBLE, MAX
-C .. Executable Statements ..
-C
- INFO = 0
- LJOBF = LSAME( JOBZ, 'F' )
- LJOBI = LSAME( JOBZ, 'I' )
- LJOBZ = LJOBF.OR.LJOBI
-C
-C Test the input scalar arguments.
-C
- IF( .NOT.LJOBZ .AND. .NOT.LSAME( JOBZ, 'N' ) ) THEN
- INFO = -1
- ELSE IF( N.LT.0 ) THEN
- INFO = -2
- ELSE IF( P.LT.0 ) THEN
- INFO = -3
- ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
- INFO = -5
- ELSE IF( LDC.LT.MAX( 1, P ) ) THEN
- INFO = -8
- ELSE IF( LDZ.LT.1 .OR. ( LJOBZ .AND. LDZ.LT.N ) ) THEN
- INFO = -11
- ELSE IF( LDWORK.LT.MAX( 1, N, P ) ) THEN
- INFO = -15
- END IF
-C
- IF ( INFO.NE.0 ) THEN
-C
-C Error return.
-C
- CALL XERBLA( 'TB01ZD', -INFO )
- RETURN
- END IF
-C
-C Quick return if possible.
-C
- NCONT = 0
- DWORK(1) = ONE
- IF ( N.EQ.0 )
- $ RETURN
-C
-C (Note: Comments in the code beginning "Workspace:" describe the
-C minimal amount of real workspace needed at that point in the
-C code, as well as the preferred amount for good performance.
-C NB refers to the optimal block size for the immediately
-C following subroutine, as returned by ILAENV.)
-C
- WRKOPT = ONE
-C
-C Calculate the absolute norms of A and B (used for scaling).
-C
- ANORM = DLANGE( 'Max', N, N, A, LDA, DWORK )
- BNORM = DLANGE( 'Max', N, 1, B, N, DWORK )
-C
-C Return if matrix B is zero.
-C
- IF( BNORM.EQ.ZERO ) THEN
- IF( LJOBF ) THEN
- CALL DLASET( 'Full', N, N, ZERO, ZERO, Z, LDZ )
- CALL DLASET( 'Full', N, 1, ZERO, ZERO, TAU, N )
- ELSE IF( LJOBI ) THEN
- CALL DLASET( 'Full', N, N, ZERO, ONE, Z, LDZ )
- END IF
- RETURN
- END IF
-C
-C Scale (if needed) the matrices A and B.
-C
- CALL MB01PD( 'S', 'G', N, N, 0, 0, ANORM, 0, NBLK, A, LDA, INFO )
- CALL MB01PD( 'S', 'G', N, 1, 0, 0, BNORM, 0, NBLK, B, N, INFO )
-C
-C Calculate the Frobenius norm of A and the 1-norm of B (used for
-C controlability test).
-C
- FANORM = DLANGE( 'Frobenius', N, N, A, LDA, DWORK )
- FBNORM = DLANGE( '1-norm', N, 1, B, N, DWORK )
-C
- TOLDEF = TOL
- IF ( TOLDEF.LE.ZERO ) THEN
-C
-C Use the default tolerance in controllability determination.
-C
- THRESH = DBLE(N)*DLAMCH( 'EPSILON' )
- TOLDEF = THRESH*MAX( FANORM, FBNORM )
- END IF
-C
- ITAU = 1
- IF ( FBNORM.GT.TOLDEF ) THEN
-C
-C B is not negligible compared with A.
-C
- IF ( N.GT.1 ) THEN
-C
-C Transform B by a Householder matrix Z1: store vector
-C describing this temporarily in B and in the local scalar H.
-C
- CALL DLARFG( N, B(1), B(2), 1, H )
-C
- B1 = B(1)
- B(1) = ONE
-C
-C Form Z1 * A * Z1.
-C Workspace: need N.
-C
- CALL DLARF( 'Right', N, N, B, 1, H, A, LDA, DWORK )
- CALL DLARF( 'Left', N, N, B, 1, H, A, LDA, DWORK )
-C
-C Form C * Z1.
-C Workspace: need P.
-C
- CALL DLARF( 'Right', P, N, B, 1, H, C, LDC, DWORK )
-C
- B(1) = B1
- TAU(1) = H
- ITAU = ITAU + 1
- ELSE
- B1 = B(1)
- TAU(1) = ZERO
- END IF
-C
-C Reduce modified A to upper Hessenberg form by an orthogonal
-C similarity transformation with matrix Z2.
-C Workspace: need N; prefer N*NB.
-C
- CALL DGEHRD( N, 1, N, A, LDA, TAU(ITAU), DWORK, LDWORK, INFO )
- WRKOPT = DWORK(1)
-C
-C Form C * Z2.
-C Workspace: need P; prefer P*NB.
-C
- CALL DORMHR( 'Right', 'No transpose', P, N, 1, N, A, LDA,
- $ TAU(ITAU), C, LDC, DWORK, LDWORK, INFO )
- WRKOPT = MAX( WRKOPT, DWORK(1) )
-C
- IF ( LJOBZ ) THEN
-C
-C Save the orthogonal transformations used, so that they could
-C be accumulated by calling DORGQR routine.
-C
- IF ( N.GT.1 )
- $ CALL DLACPY( 'Full', N-1, 1, B(2), N-1, Z(2,1), LDZ )
- IF ( N.GT.2 )
- $ CALL DLACPY( 'Lower', N-2, N-2, A(3,1), LDA, Z(3,2),
- $ LDZ )
- IF ( LJOBI ) THEN
-C
-C Form the orthogonal transformation matrix Z = Z1 * Z2.
-C Workspace: need N; prefer N*NB.
-C
- CALL DORGQR( N, N, N, Z, LDZ, TAU, DWORK, LDWORK, INFO )
- WRKOPT = MAX( WRKOPT, DWORK(1) )
- END IF
- END IF
-C
-C Annihilate the lower part of A and B.
-C
- IF ( N.GT.2 )
- $ CALL DLASET( 'Lower', N-2, N-2, ZERO, ZERO, A(3,1), LDA )
- IF ( N.GT.1 )
- $ CALL DLASET( 'Full', N-1, 1, ZERO, ZERO, B(2), N-1 )
-C
-C Find NCONT by checking sizes of the sub-diagonal elements of
-C transformed A.
-C
- IF ( TOL.LE.ZERO )
- $ TOLDEF = THRESH*MAX( FANORM, ABS( B1 ) )
-C
- J = 1
-C
-C WHILE ( J < N and ABS( A(J+1,J) ) > TOLDEF ) DO
-C
- 10 CONTINUE
- IF ( J.LT.N ) THEN
- IF ( ABS( A(J+1,J) ).GT.TOLDEF ) THEN
- J = J + 1
- GO TO 10
- END IF
- END IF
-C
-C END WHILE 10
-C
-C First negligible sub-diagonal element found, if any: set NCONT.
-C
- NCONT = J
- IF ( J.LT.N )
- $ A(J+1,J) = ZERO
-C
-C Undo scaling of A and B.
-C
- CALL MB01PD( 'U', 'H', NCONT, NCONT, 0, 0, ANORM, 0, NBLK, A,
- $ LDA, INFO )
- CALL MB01PD( 'U', 'G', 1, 1, 0, 0, BNORM, 0, NBLK, B, N, INFO )
- IF ( NCONT.LT.N )
- $ CALL MB01PD( 'U', 'G', N, N-NCONT, 0, 0, ANORM, 0, NBLK,
- $ A(1,NCONT+1), LDA, INFO )
- ELSE
-C
-C B is negligible compared with A. No computations for reducing
-C the system to orthogonal canonical form have been performed,
-C except scaling (which is undoed).
-C
- CALL MB01PD( 'U', 'G', N, N, 0, 0, ANORM, 0, NBLK, A, LDA,
- $ INFO )
- CALL MB01PD( 'U', 'G', N, 1, 0, 0, BNORM, 0, NBLK, B, N, INFO )
- IF( LJOBF ) THEN
- CALL DLASET( 'Full', N, N, ZERO, ZERO, Z, LDZ )
- CALL DLASET( 'Full', N, 1, ZERO, ZERO, TAU, N )
- ELSE IF( LJOBI ) THEN
- CALL DLASET( 'Full', N, N, ZERO, ONE, Z, LDZ )
- END IF
- END IF
-C
-C Set optimal workspace dimension.
-C
- DWORK(1) = WRKOPT
-C
- RETURN
-C *** Last line of TB01ZD ***
- END
Deleted: trunk/octave-forge/main/control/devel/ss2tf/TB04BD.dat
===================================================================
--- trunk/octave-forge/main/control/devel/ss2tf/TB04BD.dat 2011-08-15 17:06:28 UTC (rev 8460)
+++ trunk/octave-forge/main/control/devel/ss2tf/TB04BD.dat 2011-08-15 18:06:43 UTC (rev 8461)
@@ -1,11 +0,0 @@
- TB04BD EXAMPLE PROGRAM DATA
- 3 2 2 0.0 D I N
- -1.0 0.0 0.0
- 0.0 -2.0 0.0
- 0.0 0.0 -3.0
- 0.0 1.0 -1.0
- 1.0 1.0 0.0
- 0.0 1.0 1.0
- 1.0 1.0 1.0
- 1.0 0.0
- 0.0 1.0
Deleted: trunk/octave-forge/main/control/devel/ss2tf/TB04BD.f
===================================================================
--- trunk/octave-forge/main/control/devel/ss2tf/TB04BD.f 2011-08-15 17:06:28 UTC (rev 8460)
+++ trunk/octave-forge/main/control/devel/ss2tf/TB04BD.f 2011-08-15 18:06:43 UTC (rev 8461)
@@ -1,600 +0,0 @@
- SUBROUTINE TB04BD( JOBD, ORDER, EQUIL, N, M, P, MD, A, LDA, B,
- $ LDB, C, LDC, D, LDD, IGN, LDIGN, IGD, LDIGD,
- $ GN, GD, TOL, IWORK, DWORK, LDWORK, INFO )
-C
-C SLICOT RELEASE 5.0.
-C
-C Copyright (c) 2002-2009 NICONET e.V.
-C
-C This program is free software: you can redistribute it and/or
-C modify it under the terms of the GNU General Public License as
-C published by the Free Software Foundation, either version 2 of
-C the License, or (at your option) any later version.
-C
-C This program is distributed in the hope that it will be useful,
-C but WITHOUT ANY WARRANTY; without even the implied warranty of
-C MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-C GNU General Public License for more details.
-C
-C You should have received a copy of the GNU General Public License
-C along with this program. If not, see
-C <http://www.gnu.org/licenses/>.
-C
-C PURPOSE
-C
-C To compute the transfer function matrix G of a state-space
-C representation (A,B,C,D) of a linear time-invariant multivariable
-C system, using the pole-zeros method. Each element of the transfer
-C function matrix is returned in a cancelled, minimal form, with
-C numerator and denominator polynomials stored either in increasing
-C or decreasing order of the powers of the indeterminate.
-C
-C ARGUMENTS
-C
-C Mode Parameters
-C
-C JOBD CHARACTER*1
-C Specifies whether or not a non-zero matrix D appears in
-C the given state-space model:
-C = 'D': D is present;
-C = 'Z': D is assumed to be a zero matrix.
-C
-C ORDER CHARACTER*1
-C Specifies the order in which the polynomial coefficients
-C are stored, as follows:
-C = 'I': Increasing order of powers of the indeterminate;
-C = 'D': Decreasing order of powers of the indeterminate.
-C
-C EQUIL CHARACTER*1
-C Specifies whether the user wishes to preliminarily
-C equilibrate the triplet (A,B,C) as follows:
-C = 'S': perform equilibration (scaling);
-C = 'N': do not perform equilibration.
-C
-C Input/Output Parameters
-C
-C N (input) INTEGER
-C The order of the system (A,B,C,D). N >= 0.
-C
-C M (input) INTEGER
-C The number of the system inputs. M >= 0.
-C
-C P (input) INTEGER
-C The number of the system outputs. P >= 0.
-C
-C MD (input) INTEGER
-C The maximum degree of the polynomials in G, plus 1. An
-C upper bound for MD is N+1. MD >= 1.
-C
-C A (input/output) DOUBLE PRECISION array, dimension (LDA,N)
-C On entry, the leading N-by-N part of this array must
-C contain the original state dynamics matrix A.
-C On exit, if EQUIL = 'S', the leading N-by-N part of this
-C array contains the balanced matrix inv(S)*A*S, as returned
-C by SLICOT Library routine TB01ID.
-C If EQUIL = 'N', this array is unchanged on exit.
-C
-C LDA INTEGER
-C The leading dimension of array A. LDA >= MAX(1,N).
-C
-C B (input/output) DOUBLE PRECISION array, dimension (LDB,M)
-C On entry, the leading N-by-M part of this array must
-C contain the input matrix B.
-C On exit, the contents of B are destroyed: all elements but
-C those in the first row are set to zero.
-C
-C LDB INTEGER
-C The leading dimension of array B. LDB >= MAX(1,N).
-C
-C C (input/output) DOUBLE PRECISION array, dimension (LDC,N)
-C On entry, the leading P-by-N part of this array must
-C contain the output matrix C.
-C On exit, if EQUIL = 'S', the leading P-by-N part of this
-C array contains the balanced matrix C*S, as returned by
-C SLICOT Library routine TB01ID.
-C If EQUIL = 'N', this array is unchanged on exit.
-C
-C LDC INTEGER
-C The leading dimension of array C. LDC >= MAX(1,P).
-C
-C D (input) DOUBLE PRECISION array, dimension (LDD,M)
-C If JOBD = 'D', the leading P-by-M part of this array must
-C contain the matrix D.
-C If JOBD = 'Z', the array D is not referenced.
-C
-C LDD INTEGER
-C The leading dimension of array D.
-C LDD >= MAX(1,P), if JOBD = 'D';
-C LDD >= 1, if JOBD = 'Z'.
-C
-C IGN (output) INTEGER array, dimension (LDIGN,M)
-C The leading P-by-M part of this array contains the degrees
-C of the numerator polynomials in the transfer function
-C matrix G. Specifically, the (i,j) element of IGN contains
-C the degree of the numerator polynomial of the transfer
-C function G(i,j) from the j-th input to the i-th output.
-C
-C LDIGN INTEGER
-C The leading dimension of array IGN. LDIGN >= max(1,P).
-C
-C IGD (output) INTEGER array, dimension (LDIGD,M)
-C The leading P-by-M part of this array contains the degrees
-C of the denominator polynomials in the transfer function
-C matrix G. Specifically, the (i,j) element of IGD contains
-C the degree of the denominator polynomial of the transfer
-C function G(i,j).
-C
-C LDIGD INTEGER
-C The leading dimension of array IGD. LDIGD >= max(1,P).
-C
-C GN (output) DOUBLE PRECISION array, dimension (P*M*MD)
-C This array contains the coefficients of the numerator
-C polynomials, Num(i,j), of the transfer function matrix G.
-C The polynomials are stored in a column-wise order, i.e.,
-C Num(1,1), Num(2,1), ..., Num(P,1), Num(1,2), Num(2,2),
-C ..., Num(P,2), ..., Num(1,M), Num(2,M), ..., Num(P,M);
-C MD memory locations are reserved for each polynomial,
-C hence, the (i,j) polynomial is stored starting from the
-C location ((j-1)*P+i-1)*MD+1. The coefficients appear in
-C increasing or decreasing order of the powers of the
-C indeterminate, according to ORDER.
-C
-C GD (output) DOUBLE PRECISION array, dimension (P*M*MD)
-C This array contains the coefficients of the denominator
-C polynomials, Den(i,j), of the transfer function matrix G.
-C The polynomials are stored in the same way as the
-C numerator polynomials.
-C
-C Tolerances
-C
-C TOL DOUBLE PRECISION
-C The tolerance to be used in determining the
-C controllability of a single-input system (A,b) or (A',c'),
-C where b and c' are columns in B and C' (C transposed). If
-C the user sets TOL > 0, then the given value of TOL is used
-C as an absolute tolerance; elements with absolute value
-C less than TOL are considered neglijible. If the user sets
-C TOL <= 0, then an implicitly computed, default tolerance,
-C defined by TOLDEF = N*EPS*MAX( NORM(A), NORM(bc) ) is used
-C instead, where EPS is the machine precision (see LAPACK
-C Library routine DLAMCH), and bc denotes the currently used
-C column in B or C' (see METHOD).
-C
-C Workspace
-C
-C IWORK INTEGER array, dimension (N)
-C
-C DWORK DOUBLE PRECISION array, dimension (LDWORK)
-C On exit, if INFO = 0, DWORK(1) returns the optimal value
-C of LDWORK.
-C
-C LDWORK INTEGER
-C The length of the array DWORK.
-C LDWORK >= MAX(1, N*(N+P) +
-C MAX( N + MAX( N,P ), N*(2*N+5)))
-C If N >= P, N >= 1, the formula above can be written as
-C LDWORK >= N*(3*N + P + 5).
-C For optimum performance LDWORK should be larger.
-C
-C Error Indicator
-C
-C INFO INTEGER
-C = 0: successful exit;
-C < 0: if INFO = -i, the i-th argument had an illegal
-C value;
-C = 1: the QR algorithm failed to converge when trying to
-C compute the zeros of a transfer function;
-C = 2: the QR algorithm failed to converge when trying to
-C compute the poles of a transfer function.
-C The errors INFO = 1 or 2 are unlikely to appear.
-C
-C METHOD
-C
-C The routine implements the pole-zero method proposed in [1].
-C This method is based on an algorithm for computing the transfer
-C function of a single-input single-output (SISO) system.
-C Let (A,b,c,d) be a SISO system. Its transfer function is computed
-C as follows:
-C
-C 1) Find a controllable realization (Ac,bc,cc) of (A,b,c).
-C 2) Find an observable realization (Ao,bo,co) of (Ac,bc,cc).
-C 3) Compute the r eigenvalues of Ao (the poles of (Ao,bo,co)).
-C 4) Compute the zeros of (Ao,bo,co,d).
-C 5) Compute the gain of (Ao,bo,co,d).
-C
-C This algorithm can be implemented using only orthogonal
-C transformations [1]. However, for better efficiency, the
-C implementation in TB04BD uses one elementary transformation
-C in Step 4 and r elementary transformations in Step 5 (to reduce
-C an upper Hessenberg matrix to upper triangular form). These
-C special elementary transformations are numerically stable
-C in practice.
-C
-C In the multi-input multi-output (MIMO) case, the algorithm
-C computes each element (i,j) of the transfer function matrix G,
-C for i = 1 : P, and for j = 1 : M. For efficiency reasons, Step 1
-C is performed once for each value of j (each column of B). The
-C matrices Ac and Ao result in Hessenberg form.
-C
-C REFERENCES
-C
-C [1] Varga, A. and Sima, V.
-C Numerically Stable Algorithm for Transfer Function Matrix
-C Evaluation.
-C Int. J. Control, vol. 33, nr. 6, pp. 1123-1133, 1981.
-C
-C NUMERICAL ASPECTS
-C
-C The algorithm is numerically stable in practice and requires about
-C 20*N**3 floating point operations at most, but usually much less.
-C
-C FURTHER COMMENTS
-C
-C For maximum efficiency of index calculations, GN and GD are
-C implemented as one-dimensional arrays.
-C
-C CONTRIBUTORS
-C
-C V. Sima, Research Institute for Informatics, Bucharest, May 2002.
-C Partly based on the BIMASC Library routine TSMT1 by A. Varga.
-C
-C REVISIONS
-C
-C -
-C
-C KEYWORDS
-C
-C Eigenvalue, state-space representation, transfer function, zeros.
-C
-C ******************************************************************
-C
-C .. Parameters ..
- DOUBLE PRECISION ZERO, ONE, C100
- PARAMETER ( ZERO = 0.0D0, ONE = 1.0D0, C100 = 100.0D0 )
-C .. Scalar Arguments ..
- CHARACTER EQUIL, JOBD, ORDER
- DOUBLE PRECISION TOL
- INTEGER INFO, LDA, LDB, LDC, LDD, LDIGD, LDIGN, LDWORK,
- $ M, MD, N, P
-C .. Array Arguments ..
- DOUBLE PRECISION A(LDA,*), B(LDB,*), C(LDC,*), D(LDD,*),
- $ DWORK(*), GD(*), GN(*)
- INTEGER IGD(LDIGD,*), IGN(LDIGN,*), IWORK(*)
-C .. Local Scalars ..
- DOUBLE PRECISION ANORM, DIJ, EPSN, MAXRED, TOLDEF, X
- INTEGER I, IA, IAC, IAS, IB, IC, ICC, IERR, IIP, IM,
- $ IP, IPM1, IRP, ITAU, ITAU1, IZ, J, JJ, JWORK,
- $ JWORK1, K, L, NCONT, WRKOPT
- LOGICAL ASCEND, DIJNZ, FNDEIG, WITHD
-C .. Local Arrays ..
- DOUBLE PRECISION Z(1)
-C .. External Functions ..
- LOGICAL LSAME
- DOUBLE PRECISION DLAMCH, DLANGE
- EXTERNAL DLAMCH, DLANGE, LSAME
-C .. External Subroutines ..
- EXTERNAL DAXPY, DCOPY, DHSEQR, DLACPY, MA02AD, MC01PD,
- $ MC01PY, TB01ID, TB01ZD, TB04BX, XERBLA
-C .. Intrinsic Functions ..
- INTRINSIC ABS, DBLE, INT, MAX, MIN
-C ..
-C .. Executable Statements ..
-C
-C Test the input scalar parameters.
-C
- INFO = 0
- WITHD = LSAME( JOBD, 'D' )
- ASCEND = LSAME( ORDER, 'I' )
- IF( .NOT.WITHD .AND. .NOT.LSAME( JOBD, 'Z' ) ) THEN
- INFO = -1
- ELSE IF( .NOT.ASCEND .AND. .NOT.LSAME( ORDER, 'D' ) ) THEN
- INFO = -2
- ELSE IF( .NOT. ( LSAME( EQUIL, 'S' ) .OR.
- $ LSAME( EQUIL, 'N' ) ) ) THEN
- INFO = -3
- ELSE IF( N.LT.0 ) THEN
- INFO = -4
- ELSE IF( M.LT.0 ) THEN
- INFO = -5
- ELSE IF( P.LT.0 ) THEN
- INFO = -6
- ELSE IF( MD.LT.1 ) THEN
- INFO = -7
- ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
- INFO = -9
- ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
- INFO = -11
- ELSE IF( LDC.LT.MAX( 1, P ) ) THEN
- INFO = -13
- ELSE IF( LDD.LT.1 .OR. ( WITHD .AND. LDD.LT.P ) ) THEN
- INFO = -15
- ELSE IF( LDIGN.LT.MAX( 1, P ) ) THEN
- INFO = -17
- ELSE IF( LDIGD.LT.MAX( 1, P ) ) THEN
- INFO = -19
- ELSE IF( LDWORK.LT.MAX( 1, N*( N + P ) +
- $ MAX( N + MAX( N, P ), N*( 2*N + 5 ) ) )
- $ ) THEN
- INFO = -25
- END IF
-C
- IF ( INFO.NE.0 ) THEN
-C
-C Error return.
-C
- CALL XERBLA( 'TB04BD', -INFO )
- RETURN
- END IF
-C
-C Initialize GN and GD to zero.
-C
- Z(1) = ZERO
- CALL DCOPY( P*M*MD, Z, 0, GN, 1 )
- CALL DCOPY( P*M*MD, Z, 0, GD, 1 )
-C
-C Quick return if possible.
-C
- IF( MIN( N, P, M ).EQ.0 ) THEN
- IF( MIN( P, M ).GT.0 ) THEN
- K = 1
-C
- DO 20 J = 1, M
-C
- DO 10 I = 1, P
- IGN(I,J) = 0
- IGD(I,J) = 0
- IF ( WITHD )
- $ GN(K) = D(I,J)
- GD(K) = ONE
- K = K + MD
- 10 CONTINUE
-C
- 20 CONTINUE
-C
- END IF
- DWORK(1) = ONE
- RETURN
- END IF
-C
-C Prepare the computation of the default tolerance.
-C
- TOLDEF = TOL
- IF( TOLDEF.LE.ZERO ) THEN
- EPSN = DBLE( N )*DLAMCH( 'Epsilon' )
- ANORM = DLANGE( 'Frobenius', N, N, A, LDA, DWORK )
- END IF
-C
-C Initializations.
-C
- IA = 1
- IC = IA + N*N
- ITAU = IC + P*N
- JWORK = ITAU + N
- IAC = ITAU
-C
- K = 1
- DIJ = ZERO
-C
-C (Note: Comments in the code beginning "Workspace:" describe the
-C minimal amount of real workspace needed at that point in the
-C code, as well as the preferred amount for good performance.)
-C
- IF( LSAME( EQUIL, 'S' ) ) THEN
-C
-C Scale simultaneously the matrices A, B and C:
-C A <- inv(S)*A*S, B <- inv(S)*B and C <- C*S, where S is a
-C diagonal scaling matrix.
-C Workspace: need N.
-C
- MAXRED = C100
- CALL TB01ID( 'All', N, M, P, MAXRED, A, LDA, B, LDB, C, LDC,
- $ DWORK, IERR )
- END IF
-C
-C Compute the transfer function matrix of the system (A,B,C,D).
-C
- DO 80 J = 1, M
-C
-C Save A and C.
-C Workspace: need W1 = N*(N+P).
-C
- CALL DLACPY( 'Full', N, N, A, LDA, DWORK(IA), N )
- CALL DLACPY( 'Full', P, N, C, LDC, DWORK(IC), P )
-C
-C Remove the uncontrollable part of the system (A,B(J),C).
-C Workspace: need W1+N+MAX(N,P);
-C prefer larger.
-C
- CALL TB01ZD( 'No Z', N, P, DWORK(IA), N, B(1,J), DWORK(IC), P,
- $ NCONT, Z, 1, DWORK(ITAU), TOL, DWORK(JWORK),
- $ LDWORK-JWORK+1, IERR )
- IF ( J.EQ.1 )
- $ WRKOPT = INT( DWORK(JWORK) ) + JWORK - 1
-C
- IB = IAC + NCONT*NCONT
- ICC = IB + NCONT
- ITAU1 = ICC + NCONT
- IRP = ITAU1
- IIP = IRP + NCONT
- IAS = IIP + NCONT
- JWORK1 = IAS + NCONT*NCONT
-C
- DO 70 I = 1, P
- IF ( WITHD )
- $ DIJ = D(I,J)
- IF ( NCONT.GT.0 ) THEN
-C
-C Form the matrices of the state-space representation of
-C the dual system for the controllable part.
-C Workspace: need W2 = W1+N*(N+2).
-C
- CALL MA02AD( 'Full', NCONT, NCONT, DWORK(IA), N,
- $ DWORK(IAC), NCONT )
- CALL DCOPY( NCONT, B(1,J), 1, DWORK(IB), 1 )
- CALL DCOPY( NCONT, DWORK(IC+I-1), P, DWORK(ICC), 1 )
-C
-C Remove the unobservable part of the system (A,B(J),C(I)).
-C Workspace: need W2+2*N;
-C prefer larger.
-C
- CALL TB01ZD( 'No Z', NCONT, 1, DWORK(IAC), NCONT,
- $ DWORK(ICC), DWORK(IB), 1, IP, Z, 1,
- $ DWORK(ITAU1), TOL, DWORK(IIP), LDWORK-IIP+1,
- $ IERR )
- IF ( I.EQ.1 )
- $ WRKOPT = MAX( WRKOPT, INT( DWORK(IIP) ) + IIP - 1 )
-C
- IF ( IP.GT.0 ) THEN
-C
-C Save the state matrix of the minimal part.
-C Workspace: need W3 = W2+N*(N+2).
-C
- CALL DLACPY( 'Full', IP, IP, DWORK(IAC), NCONT,
- $ DWORK(IAS), IP )
-C
-C Compute the poles of the transfer function.
-C Workspace: need W3+N;
-C prefer larger.
-C
- CALL DHSEQR( 'Eigenvalues', 'No vectors', IP, 1, IP,
- $ DWORK(IAC), NCONT, DWORK(IRP),
- $ DWORK(IIP), Z, 1, DWORK(JWORK1),
- $ LDWORK-JWORK1+1, IERR )
- IF ( IERR.NE.0 ) THEN
- INFO = 2
- RETURN
- END IF
- WRKOPT = MAX( WRKOPT,
- $ INT( DWORK(JWORK1) ) + JWORK1 - 1 )
-C
-C Compute the zeros of the transfer function.
-C
- IPM1 = IP - 1
- DIJNZ = WITHD .AND. DIJ.NE.ZERO
- FNDEIG = DIJNZ .OR. IPM1.GT.0
- IF ( .NOT.FNDEIG ) THEN
- IZ = 0
- ELSE IF ( DIJNZ ) THEN
-C
-C Add the contribution due to D(i,j).
-C Note that the matrix whose eigenvalues have to
-C be computed remains in an upper Hessenberg form.
-C
- IZ = IP
- CALL DLACPY( 'Full', IZ, IZ, DWORK(IAS), IP,
- $ DWORK(IAC), NCONT )
- CALL DAXPY( IZ, -DWORK(ICC)/DIJ, DWORK(IB), 1,
- $ DWORK(IAC), NCONT )
- ELSE
- IF( TOL.LE.ZERO )
- $ TOLDEF = EPSN*MAX( ANORM,
- $ DLANGE( 'Frobenius', IP, 1,
- $ DWORK(IB), 1, DWORK )
- $ )
-C
- DO 30 IM = 1, IPM1
- IF ( ABS( DWORK(IB+IM-1) ).GT.TOLDEF ) GO TO 40
- 30 CONTINUE
-C
- IZ = 0
- GO TO 50
-C
- 40 CONTINUE
-C
-C Restore (part of) the saved state matrix.
-C
- IZ = IP - IM
- CALL DLACPY( 'Full', IZ, IZ, DWORK(IAS+IM*(IP+1)),
- $ IP, DWORK(IAC), NCONT )
-C
-C Apply the output injection.
-C
- CALL DAXPY( IZ, -DWORK(IAS+IM*(IP+1)-IP)/
- $ DWORK(IB+IM-1), DWORK(IB+IM), 1,
- $ DWORK(IAC), NCONT )
- END IF
-C
- IF ( FNDEIG ) THEN
-C
-C Find the zeros.
-C Workspace: need W3+N;
-C prefer larger.
-C
- CALL DHSEQR( 'Eigenvalues', 'No vectors', IZ, 1,
- $ IZ, DWORK(IAC), NCONT, GN(K), GD(K),
- $ Z, 1, DWORK(JWORK1), LDWORK-JWORK1+1,
- $ IERR )
- IF ( IERR.NE.0 ) THEN
- INFO = 1
- RETURN
- END IF
- END IF
-C
-C Compute the gain.
-C
- 50 CONTINUE
- IF ( DIJNZ ) THEN
- X = DIJ
- ELSE
- CALL TB04BX( IP, IZ, DWORK(IAS), IP, DWORK(ICC),
- $ DWORK(IB), DIJ, DWORK(IRP),
- $ DWORK(IIP), GN(K), GD(K), X, IWORK )
- END IF
-C
-C Form the numerator coefficients in increasing or
-C decreasing powers of the indeterminate.
-C IAS is used here as pointer to the workspace.
-C
- IF ( ASCEND ) THEN
- CALL MC01PD( IZ, GN(K), GD(K), DWORK(IB),
- $ DWORK(IAS), IERR )
- ELSE
- CALL MC01PY( IZ, GN(K), GD(K), DWORK(IB),
- $ DWORK(IAS), IERR )
- END IF
- JJ = K
-C
- DO 60 L = IB, IB + IZ
- GN(JJ) = DWORK(L)*X
- JJ = JJ + 1
- 60 CONTINUE
-C
-C Form the denominator coefficients.
-C
- IF ( ASCEND ) THEN
- CALL MC01PD( IP, DWORK(IRP), DWORK(IIP), GD(K),
- $ DWORK(IAS), IERR )
- ELSE
- CALL MC01PY( IP, DWORK(IRP), DWORK(IIP), GD(K),
- $ DWORK(IAS), IERR )
- END IF
- IGN(I,J) = IZ
- IGD(I,J) = IP
- ELSE
-C
-C Null element.
-C
- IGN(I,J) = 0
- IGD(I,J) = 0
- GN(K) = DIJ
- GD(K) = ONE
- END IF
-C
- ELSE
-C
-C Null element.
-C
- IGN(I,J) = 0
- IGD(I,J) = 0
- GN(K) = DIJ
- GD(K) = ONE
- END IF
-C
- K = K + MD
- 70 CONTINUE
-C
- 80 CONTINUE
-C
- RETURN
-C *** Last line of TB04BD ***
- END
Deleted: trunk/octave-forge/main/control/devel/ss2tf/TB04BD.html
===================================================================
--- trunk/octave-forge/main/control/devel/ss2tf/TB04BD.html 2011-08-15 17:06:28 UTC (rev 8460)
+++ trunk/octave-forge/main/control/devel/ss2tf/TB04BD.html 2011-08-15 18:06:43 UTC (rev 8461)
@@ -1,418 +0,0 @@
-<HTML>
-<HEAD><TITLE>TB04BD - SLICOT Library Routine Documentation</TITLE>
-</HEAD>
-<BODY>
-
-<H2><A Name="TB04BD">TB04BD</A></H2>
-<H3>
-Transfer matrix of a given state-space representation (A,B,C,D), using the pole-zeros method
-</H3>
-<A HREF ="#Specification"><B>[Specification]</B></A>
-<A HREF ="#Arguments"><B>[Arguments]</B></A>
-<A HREF ="#Method"><B>[Method]</B></A>
-<A HREF ="#References"><B>[References]</B></A>
-<A HREF ="#Comments"><B>[Comments]</B></A>
-<A HREF ="#Example"><B>[Example]</B></A>
-
-<P>
-<B><FONT SIZE="+1">Purpose</FONT></B>
-<PRE>
- To compute the transfer function matrix G of a state-space
- representation (A,B,C,D) of a linear time-invariant multivariable
- system, using the pole-zeros method. Each element of the transfer
- function matrix is returned in a cancelled, minimal form, with
- numerator and denominator polynomials stored either in increasing
- or decreasing order of the powers of the indeterminate.
-
-</PRE>
-<A name="Specification"><B><FONT SIZE="+1">Specification</FONT></B></A>
-<PRE>
- SUBROUTINE TB04BD( JOBD, ORDER, EQUIL, N, M, P, MD, A, LDA, B,
- $ LDB, C, LDC, D, LDD, IGN, LDIGN, IGD, LDIGD,
- $ GN, GD, TOL, IWORK, DWORK, LDWORK, INFO )
-C .. Scalar Arguments ..
- CHARACTER EQUIL, JOBD, ORDER
- DOUBLE PRECISION TOL
- INTEGER INFO, LDA, LDB, LDC, LDD, LDIGD, LDIGN, LDWORK,
- $ M, MD, N, P
-C .. Array Arguments ..
- DOUBLE PRECISION A(LDA,*), B(LDB,*), C(LDC,*), D(LDD,*),
- $ DWORK(*), GD(*), GN(*)
- INTEGER IGD(LDIGD,*), IGN(LDIGN,*), IWORK(*)
-
-</PRE>
-<A name="Arguments"><B><FONT SIZE="+1">Arguments</FONT></B></A>
-<P>
-
-<B>Mode Parameters</B>
-<PRE>
- JOBD CHARACTER*1
- Specifies whether or not a non-zero matrix D appears in
- the given state-space model:
- = 'D': D is present;
- = 'Z': D is assumed to be a zero matrix.
-
- ORDER CHARACTER*1
- Specifies the order in which the polynomial coefficients
- are stored, as follows:
- = 'I': Increasing order of powers of the indeterminate;
- = 'D': Decreasing order of powers of the indeterminate.
-
- EQUIL CHARACTER*1
- Specifies whether the user wishes to preliminarily
- equilibrate the triplet (A,B,C) as follows:
- = 'S': perform equilibration (scaling);
- = 'N': do not perform equilibration.
-
-</PRE>
-<B>Input/Output Parameters</B>
-<PRE>
- N (input) INTEGER
- The order of the system (A,B,C,D). N >= 0.
-
- M (input) INTEGER
- The number of the system inputs. M >= 0.
-
- P (input) INTEGER
- The number of the system outputs. P >= 0.
-
- MD (input) INTEGER
- The maximum degree of the polynomials in G, plus 1. An
- upper bound for MD is N+1. MD >= 1.
-
- A (input/output) DOUBLE PRECISION array, dimension (LDA,N)
- On entry, the leading N-by-N part of this array must
- contain the original state dynamics matrix A.
- On exit, if EQUIL = 'S', the leading N-by-N part of this
- array contains the balanced matrix inv(S)*A*S, as returned
- by SLICOT Library routine TB01ID.
- If EQUIL = 'N', this array is unchanged on exit.
-
- LDA INTEGER
- The leading dimension of array A. LDA >= MAX(1,N).
-
- B (input/output) DOUBLE PRECISION array, dimension (LDB,M)
- On entry, the leading N-by-M part of this array must
- contain the input matrix B.
- On exit, the contents of B are destroyed: all elements but
- those in the first row are set to zero.
-
- LDB INTEGER
- The leading dimension of array B. LDB >= MAX(1,N).
-
- C (input/output) DOUBLE PRECISION array, dimension (LDC,N)
- On entry, the leading P-by-N part of this array must
- contain the output matrix C.
- On exit, if EQUIL = 'S', the leading P-by-N part of this
- array contains the balanced matrix C*S, as returned by
- SLICOT Library routine TB01ID.
- If EQUIL = 'N', this array is unchanged on exit.
-
- LDC INTEGER
- The leading dimension of array C. LDC >= MAX(1,P).
-
- D (input) DOUBLE PRECISION array, dimension (LDD,M)
- If JOBD = 'D', the leading P-by-M part of this array must
- contain the matrix D.
- If JOBD = 'Z', the array D is not referenced.
-
- LDD INTEGER
- The leading dimension of array D.
- LDD >= MAX(1,P), if JOBD = 'D';
- LDD >= 1, if JOBD = 'Z'.
-
- IGN (output) INTEGER array, dimension (LDIGN,M)
- The leading P-by-M part of this array contains the degrees
- of the numerator polynomials in the transfer function
- matrix G. Specifically, the (i,j) element of IGN contains
- the degree of the numerator polynomial of the transfer
- function G(i,j) from the j-th input to the i-th output.
-
- LDIGN INTEGER
- The leading dimension of array IGN. LDIGN >= max(1,P).
-
- IGD (output) INTEGER array, dimension (LDIGD,M)
- The leading P-by-M part of this array contains the degrees
- of the denominator polynomials in the transfer function
- matrix G. Specifically, the (i,j) element of IGD contains
- the degree of the denominator polynomial of the transfer
- function G(i,j).
-
- LDIGD INTEGER
- The leading dimension of array IGD. LDIGD >= max(1,P).
-
- GN (output) DOUBLE PRECISION array, dimension (P*M*MD)
- This array contains the coefficients of the numerator
- polynomials, Num(i,j), of the transfer function matrix G.
- The polynomials are stored in a column-wise order, i.e.,
- Num(1,1), Num(2,1), ..., Num(P,1), Num(1,2), Num(2,2),
- ..., Num(P,2), ..., Num(1,M), Num(2,M), ..., Num(P,M);
- MD memory locations are reserved for each polynomial,
- hence, the (i,j) polynomial is stored starting from the
- location ((j-1)*P+i-1)*MD+1. The coefficients appear in
- increasing or decreasing order of the powers of the
- indeterminate, according to ORDER.
-
- GD (output) DOUBLE PRECISION array, dimension (P*M*MD)
- This array contains the coefficients of the denominator
- polynomials, Den(i,j), of the transfer function matrix G.
- The polynomials are stored in the same way as the
- numerator polynomials.
-
-</PRE>
-<B>Tolerances</B>
-<PRE>
- TOL DOUBLE PRECISION
- The tolerance to be used in determining the
- controllability of a single-input system (A,b) or (A',c'),
- where b and c' are columns in B and C' (C transposed). If
- the user sets TOL > 0, then the given value of TOL is used
- as an absolute tolerance; elements with absolute value
- less than TOL are considered neglijible. If the user sets
- TOL <= 0, then an implicitly computed, default tolerance,
- defined by TOLDEF = N*EPS*MAX( NORM(A), NORM(bc) ) is used
- ...
[truncated message content] |
|
From: <par...@us...> - 2011-08-22 16:09:06
|
Revision: 8473
http://octave.svn.sourceforge.net/octave/?rev=8473&view=rev
Author: paramaniac
Date: 2011-08-22 16:08:57 +0000 (Mon, 22 Aug 2011)
Log Message:
-----------
control: prepare release of control-2.1.54
Modified Paths:
--------------
trunk/octave-forge/main/control/DESCRIPTION
trunk/octave-forge/main/control/devel/RELEASE_PACKAGE
trunk/octave-forge/main/control/doc/NEWS
Added Paths:
-----------
trunk/octave-forge/main/control/doc/control.pdf
Modified: trunk/octave-forge/main/control/DESCRIPTION
===================================================================
--- trunk/octave-forge/main/control/DESCRIPTION 2011-08-19 16:20:21 UTC (rev 8472)
+++ trunk/octave-forge/main/control/DESCRIPTION 2011-08-22 16:08:57 UTC (rev 8473)
@@ -1,6 +1,6 @@
Name: Control
-Version: 2.1.53
-Date: 2011-08-08
+Version: 2.1.54
+Date: 2011-08-22
Author: Lukas Reichlin <luk...@gm...>
Maintainer: Lukas Reichlin <luk...@gm...>
Title: Control Systems
Modified: trunk/octave-forge/main/control/devel/RELEASE_PACKAGE
===================================================================
--- trunk/octave-forge/main/control/devel/RELEASE_PACKAGE 2011-08-19 16:20:21 UTC (rev 8472)
+++ trunk/octave-forge/main/control/devel/RELEASE_PACKAGE 2011-08-22 16:08:57 UTC (rev 8473)
@@ -18,12 +18,12 @@
rm -R ~/octave/__TEMP__/control/devel
cd ~/octave/__TEMP__
grep -i version control/DESCRIPTION
-tar czf control-2.1.53.tar.gz control/
-md5 control-2.1.53.tar.gz
-md5 control-2.1.53.tar.gz > md5_control_pkg.txt
-uuencode control-2.1.53.tar.gz < control-2.1.53.tar.gz > control-2.1.53.tar.gz.uue
+tar czf control-2.1.54.tar.gz control/
+md5 control-2.1.54.tar.gz
+md5 control-2.1.54.tar.gz > md5_control_pkg.txt
+uuencode control-2.1.54.tar.gz < control-2.1.54.tar.gz > control-2.1.54.tar.gz.uue
octave -q --eval \
-"pkg install control-2.1.53.tar.gz"
+"pkg install control-2.1.54.tar.gz"
octave -q --eval \
"pkg load generate_html; generate_package_html ('control', 'control-html', 'octave-forge')"
tar czf control-html.tar.gz control-html
@@ -38,7 +38,7 @@
=====================================================================================
rm -R ~/octave/__TEMP__
-rm -R ~/octave/control-2.1.53
+rm -R ~/octave/control-2.1.54
=====================================================================================
Modified: trunk/octave-forge/main/control/doc/NEWS
===================================================================
--- trunk/octave-forge/main/control/doc/NEWS 2011-08-19 16:20:21 UTC (rev 8472)
+++ trunk/octave-forge/main/control/doc/NEWS 2011-08-22 16:08:57 UTC (rev 8473)
@@ -1,7 +1,7 @@
Summary of important user-visible changes for releases of the control package
===============================================================================
-control-2.x.yy Release Date: 2011-xx-yy Release Manager: Lukas Reichlin
+control-2.1.54 Release Date: 2011-08-22 Release Manager: Lukas Reichlin
===============================================================================
** tf
@@ -10,7 +10,11 @@
-- Display an empty line between title and numerator for better readability.
-- Display whether model is static, continuous- or discrete-time.
+** A PDF manual is included for the first time. It is located inside the "doc"
+ folder. It has been generated automatically from the Texinfo help strings
+ and is not yet completely sorted out.
+
===============================================================================
control-2.1.53 Release Date: 2011-08-08 Release Manager: Lukas Reichlin
===============================================================================
Added: trunk/octave-forge/main/control/doc/control.pdf
===================================================================
--- trunk/octave-forge/main/control/doc/control.pdf (rev 0)
+++ trunk/octave-forge/main/control/doc/control.pdf 2011-08-22 16:08:57 UTC (rev 8473)
@@ -0,0 +1,3313 @@
+%PDF-1.3
+%\xC7쏢
+5 0 obj
+<</Length 6 0 R/Filter /FlateDecode>>
+stream
+x\x9C\x95S\xC9n1 \xEDY_\xA1\xA3tZ\x94(\x89\xBA:MS]\xE2d\x8A \x82 \xD7q\x8B\xC6-\xBC\xA4@\xFF\xBE\xE4Ll\xE7\x90 \x8A\xC1\x8C\xF2qyO\x9C\x8D
+\x806\xE8\xF3t.\xD6frU\xEDjg\x82\xBD\x90we6\x80}:k;\xED\xC46(\x99\xB3\xED\xEF͘\x8C\xB6F[B\x82j\xFB\xB5\xB9u \x80\xA8\xA6\xD0\xDC/ߡ\xD8
+\xAB\xFB\xE9c\x81\xA2\xDB{\x84D\xA9\xBA\xAD\xE00p\x8D\xF9\x88\xA3\xEA |\x97jē\xD4,5Fr\xF8\xA27\x9F\xBC4\x9A\xCC\xFC\xA5g\xCE{37K9\x95aڎ\xB0\xAD\xD9\xE6\xD6\x88\x95\xFFtf&\xB3v\xBF}\\x9AɍE3y\xAB\x9F\xE9\xE5\x99 \xB3\xD7\xF6\x959\x9F\xD9\xF9?uA\xF5@
+\x9F \x93b\x82B63ҨΙ\xEFH\xF8\xE6Q\x90ְ\x89" B@E\xBA<\xC4T\x84\xC4и\xB8k\xDF\xF7G(\xB9\x9D\xF2\x92\x81\xE9Y\xC6ғ[\xFB.\xAA_\xC0;\xCDgw\xE9S\xCB\xDFI\xC1\x85Ofv?|N\xA0\xC1\xBD\x92w)9\xB9I0\xB9\xFBS%\x9Db\xABղ4"w\xE1\x81\x8Fb\x85\x90#\xBA\xCF\xD6\xC4\xEC> ) \x88\x8At:X\xFBcΝvo\\xB3\xFB\xED\xD3\xC0!\xE9\xE4#N/I\xF7
+I֢D\x95\xAF\x8BU\xB7\x8E\xE4zX\x84\xFB*½\xD7\xE1rl\xA5\xB9G5a!\xA5S\xA1\x94\xAAM\xC4d\xE0A\x85rtoĔ\xCE,R\xC1\xB8X,\xBD\xAF\x9A\x89P\xC9G\xA8\xBB\xEF\xEA\xC3Ъj\x85\x90ks\xDF\xC66Q8FG r\x94 <\x84_\xDC3\xF9O\xE8\xB0g\x8C "\xE3\xEC\xD9\xDC\xFC\xAE\xCB\xC1Bendstream
+endobj
+6 0 obj
+492
+endobj
+15 0 obj
+<</Length 16 0 R/Filter /FlateDecode>>
+stream
+x\x9C\xADWKo\xDBF\xBE \xFD<r\x8B\x90\xE5\xBE\xC8\xDDk\x8B\xB6@\x91C\xE8\x96\xF6\xA0X\x8A ď\xC6v\\xF4\xDFw\xBE\x99}Q\xB2k'- Fw\xDE\xF3}3\x9BO\xDD4\xEAn\xC2_\xFA<\xBB\xDA|\xF7f\xE9\xCE\xEF6S\xF73=\xE7\x9BO\xCD]\xFA8\xBB\xEA\xBEߒ\x90ƛ1NQw\xDB\xF7\xD1\xD6\xDDb:\xED\xE7\xD1\xCD\xDD\xF6j\xF3\xB6\xFFA
+\xAE\xBFQ\xBE\xFFS\x99y\x8C\xD1\xF8\xFEo5\xE8\xFEV
+\xBE\xFF\xA0LNg\xCA,㤭\xE9\xEF\x95\xFDc\xFB\xCBƻ\xD1\xD8((\xEB\xE32\x93\xAD=\xD9:S\x87\xE4yi=\xB6ւ\xBF\x93u\x9DD\xD7A\xBA\xD1\xEB)P\x963\xE4{JO\xA4gP\xB6\xBC\xD3\xE9y\xA5\xEC8M\xBA\xAD\xD3V\xC3\xDCT\xE14\xCDK\xBF#\x91;\xB1\xC9\xF5?)C\x9F!\xCC\xFDH/M \xA1\xA5\xA3(>z\xABM\xE8\xCAq\xDAgʢ :R\xF2d\xF0\x92^\xE1\xF55\xFD@\xE0\xC3\xEC\xC7\xC5S=\x8FF\xA2\xFDU\xD9I*#\\xC0+r\xC2zw\x8D\xFF\xC1/ԛ,YJwҾ\xC4\xC8?Qq\xD6ޱ\x98\xB4\x95g\xBB{\xC4\xED\xF9\xE7
+}\x9D(u\xED\xC4\xE4?*k\x90\xD4L\xC2\xE4\xC0s\x82b\x88\xD2\xD8'\x8D)\xE0\xEB\x9C=\xA36\xC1EE!\xA5\xFB\xF8$\xFAw\x90\xE5\xC2\x89\xE2\xA5ZR% \xC8eM\x99eᰈ\xC2\xE2C\x9F\xA3J\xE0R |
|
From: <par...@us...> - 2011-09-03 18:50:42
|
Revision: 8488
http://octave.svn.sourceforge.net/octave/?rev=8488&view=rev
Author: paramaniac
Date: 2011-09-03 18:50:35 +0000 (Sat, 03 Sep 2011)
Log Message:
-----------
control: add bilinear transformation (c2d case)
Modified Paths:
--------------
trunk/octave-forge/main/control/devel/makefile_all.m
trunk/octave-forge/main/control/inst/@ss/__c2d__.m
trunk/octave-forge/main/control/inst/@tf/__c2d__.m
trunk/octave-forge/main/control/inst/ltimodels.m
trunk/octave-forge/main/control/src/Makefile
Added Paths:
-----------
trunk/octave-forge/main/control/devel/makefile_tustin.m
trunk/octave-forge/main/control/src/AB04MD.f
trunk/octave-forge/main/control/src/slab04md.cc
Modified: trunk/octave-forge/main/control/devel/makefile_all.m
===================================================================
--- trunk/octave-forge/main/control/devel/makefile_all.m 2011-09-03 16:37:43 UTC (rev 8487)
+++ trunk/octave-forge/main/control/devel/makefile_all.m 2011-09-03 18:50:35 UTC (rev 8488)
@@ -25,7 +25,7 @@
makefile_scale
makefile_ss2tf
makefile_staircase
+makefile_tustin
makefile_zero
-system ("rm *.o");
cd (homedir);
Added: trunk/octave-forge/main/control/devel/makefile_tustin.m
===================================================================
--- trunk/octave-forge/main/control/devel/makefile_tustin.m (rev 0)
+++ trunk/octave-forge/main/control/devel/makefile_tustin.m 2011-09-03 18:50:35 UTC (rev 8488)
@@ -0,0 +1,19 @@
+## ==============================================================================
+## Developer Makefile for OCT-files
+## ==============================================================================
+## USAGE: * fetch control from Octave-Forge by svn
+## * add control/inst, control/src and control/devel to your Octave path
+## * run makefile_*
+## ==============================================================================
+
+homedir = pwd ();
+develdir = fileparts (which ("makefile_tustin"));
+srcdir = [develdir, "/../src"];
+cd (srcdir);
+
+## bilinear transformation
+mkoctfile slab04md.cc \
+ AB04MD.f
+
+system ("rm *.o");
+cd (homedir);
Modified: trunk/octave-forge/main/control/inst/@ss/__c2d__.m
===================================================================
--- trunk/octave-forge/main/control/inst/@ss/__c2d__.m 2011-09-03 16:37:43 UTC (rev 8487)
+++ trunk/octave-forge/main/control/inst/@ss/__c2d__.m 2011-09-03 18:50:35 UTC (rev 8488)
@@ -44,6 +44,19 @@
sys.a = tmp (1:n, 1:n); # F
sys.b = tmp (1:n, n+(1:m)); # G
+ case {"tustin", "bilin"}
+ if (! isempty (sys.e))
+ if (rcond (sys.e) < eps)
+ error ("ss: c2d: tustin method requires proper system");
+ else
+ sys.a = sys.e \ sys.a;
+ sys.b = sys.e \ sys.b;
+ sys.e = []; # require ordinary state-space model
+ endif
+ endif
+
+ [sys.a, sys.b, sys.c, sys.d] = slab04md (sys.a, sys.b, sys.c, sys.d, 1, 2/tsam, false);
+
otherwise
error ("ss: c2d: %s is an invalid method", method);
Modified: trunk/octave-forge/main/control/inst/@tf/__c2d__.m
===================================================================
--- trunk/octave-forge/main/control/inst/@tf/__c2d__.m 2011-09-03 16:37:43 UTC (rev 8487)
+++ trunk/octave-forge/main/control/inst/@tf/__c2d__.m 2011-09-03 18:50:35 UTC (rev 8488)
@@ -26,13 +26,13 @@
[p, m] = size (sys);
- switch (method)
- case {"zoh", "std"}
+ ##switch (method)
+ ## case {"zoh", "std"}
error ("tf: c2d: not implemented yet");
- otherwise
- error ("tf: c2d: %s is an invalid method", method);
+ ## otherwise
+ ## error ("tf: c2d: %s is an invalid method", method);
- endswitch
+ ##endswitch
endfunction
\ No newline at end of file
Modified: trunk/octave-forge/main/control/inst/ltimodels.m
===================================================================
--- trunk/octave-forge/main/control/inst/ltimodels.m 2011-09-03 16:37:43 UTC (rev 8487)
+++ trunk/octave-forge/main/control/inst/ltimodels.m 2011-09-03 18:50:35 UTC (rev 8488)
@@ -1229,3 +1229,71 @@
%!
%!assert (NUM, NUMe, 1e-4);
%!assert (DEN, DENe, 1e-4);
+
+
+## bilinear transformation
+%!shared Mo, Me
+%! A = [ 1.0 0.5
+%! 0.5 1.0 ].';
+%!
+%! B = [ 0.0 -1.0
+%! 1.0 0.0 ].';
+%!
+%! C = [ -1.0 0.0
+%! 0.0 1.0 ].';
+%!
+%! D = [ 1.0 0.0
+%! 0.0 -1.0 ].';
+%!
+%! [Ao, Bo, Co, Do] = slab04md (A, B, C, D, 1.0, 1.0, false);
+%!
+%! Ae = [ -1.0000 -4.0000
+%! -4.0000 -1.0000 ];
+%!
+%! Be = [ 2.8284 0.0000
+%! 0.0000 -2.8284 ];
+%!
+%! Ce = [ 0.0000 2.8284
+%! -2.8284 0.0000 ];
+%!
+%! De = [ -1.0000 0.0000
+%! 0.0000 -3.0000 ];
+%!
+%! Mo = [Ao, Bo; Co, Do];
+%! Me = [Ae, Be; Ce, De];
+%!
+%!assert (Mo, Me, 1e-4);
+
+
+## bilinear transformation
+%!shared Mo, Me
+%! A = [ 1.0 0.5
+%! 0.5 1.0 ].';
+%!
+%! B = [ 0.0 -1.0
+%! 1.0 0.0 ].';
+%!
+%! C = [ -1.0 0.0
+%! 0.0 1.0 ].';
+%!
+%! D = [ 1.0 0.0
+%! 0.0 -1.0 ].';
+%!
+%! [Ao, Bo, Co, Do] = ssdata (c2d (ss (A, B, C, D), 2, "tustin"));
+%!
+%! Ae = [ -1.0000 -4.0000
+%! -4.0000 -1.0000 ];
+%!
+%! Be = [ 2.8284 0.0000
+%! 0.0000 -2.8284 ];
+%!
+%! Ce = [ 0.0000 2.8284
+%! -2.8284 0.0000 ];
+%!
+%! De = [ -1.0000 0.0000
+%! 0.0000 -3.0000 ];
+%!
+%! Mo = [Ao, Bo; Co, Do];
+%! Me = [Ae, Be; Ce, De];
+%!
+%!assert (Mo, Me, 1e-4);
Added: trunk/octave-forge/main/control/src/AB04MD.f
===================================================================
--- trunk/octave-forge/main/control/src/AB04MD.f (rev 0)
+++ trunk/octave-forge/main/control/src/AB04MD.f 2011-09-03 18:50:35 UTC (rev 8488)
@@ -0,0 +1,345 @@
+ SUBROUTINE AB04MD( TYPE, N, M, P, ALPHA, BETA, A, LDA, B, LDB, C,
+ $ LDC, D, LDD, IWORK, DWORK, LDWORK, INFO )
+C
+C SLICOT RELEASE 5.0.
+C
+C Copyright (c) 2002-2010 NICONET e.V.
+C
+C This program is free software: you can redistribute it and/or
+C modify it under the terms of the GNU General Public License as
+C published by the Free Software Foundation, either version 2 of
+C the License, or (at your option) any later version.
+C
+C This program is distributed in the hope that it will be useful,
+C but WITHOUT ANY WARRANTY; without even the implied warranty of
+C MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+C GNU General Public License for more details.
+C
+C You should have received a copy of the GNU General Public License
+C along with this program. If not, see
+C <http://www.gnu.org/licenses/>.
+C
+C PURPOSE
+C
+C To perform a transformation on the parameters (A,B,C,D) of a
+C system, which is equivalent to a bilinear transformation of the
+C corresponding transfer function matrix.
+C
+C ARGUMENTS
+C
+C Mode Parameters
+C
+C TYPE CHARACTER*1
+C Indicates the type of the original system and the
+C transformation to be performed as follows:
+C = 'D': discrete-time -> continuous-time;
+C = 'C': continuous-time -> discrete-time.
+C
+C Input/Output Parameters
+C
+C N (input) INTEGER
+C The order of the state matrix A. N >= 0.
+C
+C M (input) INTEGER
+C The number of system inputs. M >= 0.
+C
+C P (input) INTEGER
+C The number of system outputs. P >= 0.
+C
+C ALPHA, (input) DOUBLE PRECISION
+C BETA Parameters specifying the bilinear transformation.
+C Recommended values for stable systems: ALPHA = 1,
+C BETA = 1. ALPHA <> 0, BETA <> 0.
+C
+C A (input/output) DOUBLE PRECISION array, dimension (LDA,N)
+C On entry, the leading N-by-N part of this array must
+C contain the state matrix A of the original system.
+C On exit, the leading N-by-N part of this array contains
+C _
+C the state matrix A of the transformed system.
+C
+C LDA INTEGER
+C The leading dimension of array A. LDA >= MAX(1,N).
+C
+C B (input/output) DOUBLE PRECISION array, dimension (LDB,M)
+C On entry, the leading N-by-M part of this array must
+C contain the input matrix B of the original system.
+C On exit, the leading N-by-M part of this array contains
+C _
+C the input matrix B of the transformed system.
+C
+C LDB INTEGER
+C The leading dimension of array B. LDB >= MAX(1,N).
+C
+C C (input/output) DOUBLE PRECISION array, dimension (LDC,N)
+C On entry, the leading P-by-N part of this array must
+C contain the output matrix C of the original system.
+C On exit, the leading P-by-N part of this array contains
+C _
+C the output matrix C of the transformed system.
+C
+C LDC INTEGER
+C The leading dimension of array C. LDC >= MAX(1,P).
+C
+C D (input/output) DOUBLE PRECISION array, dimension (LDD,M)
+C On entry, the leading P-by-M part of this array must
+C contain the input/output matrix D for the original system.
+C On exit, the leading P-by-M part of this array contains
+C _
+C the input/output matrix D of the transformed system.
+C
+C LDD INTEGER
+C The leading dimension of array D. LDD >= MAX(1,P).
+C
+C Workspace
+C
+C IWORK INTEGER array, dimension (N)
+C
+C DWORK DOUBLE PRECISION array, dimension (LDWORK)
+C On exit, if INFO = 0, DWORK(1) returns the optimal value
+C of LDWORK.
+C
+C LDWORK INTEGER
+C The length of the array DWORK. LDWORK >= MAX(1,N).
+C For optimum performance LDWORK >= MAX(1,N*NB), where NB
+C is the optimal blocksize.
+C
+C Error Indicator
+C
+C INFO INTEGER
+C = 0: successful exit;
+C < 0: if INFO = -i, the i-th argument had an illegal
+C value;
+C = 1: if the matrix (ALPHA*I + A) is exactly singular;
+C = 2: if the matrix (BETA*I - A) is exactly singular.
+C
+C METHOD
+C
+C The parameters of the discrete-time system are transformed into
+C the parameters of the continuous-time system (TYPE = 'D'), or
+C vice-versa (TYPE = 'C') by the transformation:
+C
+C 1. Discrete -> continuous
+C _ -1
+C A = beta*(alpha*I + A) * (A - alpha*I)
+C _ -1
+C B = sqrt(2*alpha*beta) * (alpha*I + A) * B
+C _ -1
+C C = sqrt(2*alpha*beta) * C * (alpha*I + A)
+C _ -1
+C D = D - C * (alpha*I + A) * B
+C
+C which is equivalent to the bilinear transformation
+C
+C z - alpha
+C z -> s = beta --------- .
+C z + alpha
+C
+C of one transfer matrix onto the other.
+C
+C 2. Continuous -> discrete
+C _ -1
+C A = alpha*(beta*I - A) * (beta*I + A)
+C _ -1
+C B = sqrt(2*alpha*beta) * (beta*I - A) * B
+C _ -1
+C C = sqrt(2*alpha*beta) * C * (beta*I - A)
+C _ -1
+C D = D + C * (beta*I - A) * B
+C
+C which is equivalent to the bilinear transformation
+C
+C beta + s
+C s -> z = alpha -------- .
+C beta - s
+C
+C of one transfer matrix onto the other.
+C
+C REFERENCES
+C
+C [1] Al-Saggaf, U.M. and Franklin, G.F.
+C Model reduction via balanced realizations: a extension and
+C frequency weighting techniques.
+C IEEE Trans. Autom. Contr., AC-33, pp. 687-692, 1988.
+C
+C NUMERICAL ASPECTS
+C 3
+C The time taken is approximately proportional to N .
+C The accuracy depends mainly on the condition number of the matrix
+C to be inverted.
+C
+C CONTRIBUTORS
+C
+C Release 3.0: V. Sima, Katholieke Univ. Leuven, Belgium, and
+C A. Varga, German Aerospace Research Establishment,
+C Oberpfaffenhofen, Germany, Nov. 1996.
+C Supersedes Release 2.0 routine AB04AD by W. van der Linden, and
+C A.J. Geurts, Technische Hogeschool Eindhoven, Holland.
+C
+C REVISIONS
+C
+C -
+C
+C KEYWORDS
+C
+C Bilinear transformation, continuous-time system, discrete-time
+C system, state-space model.
+C
+C ******************************************************************
+C
+C .. Parameters ..
+ DOUBLE PRECISION ZERO, ONE, TWO
+ PARAMETER ( ZERO=0.0D0, ONE=1.0D0, TWO=2.0D0 )
+C .. Scalar Arguments ..
+ CHARACTER TYPE
+ INTEGER INFO, LDA, LDB, LDC, LDD, LDWORK, M, N, P
+ DOUBLE PRECISION ALPHA, BETA
+C .. Array Arguments ..
+ INTEGER IWORK(*)
+ DOUBLE PRECISION A(LDA,*), B(LDB,*), C(LDC,*), D(LDD,*), DWORK(*)
+C .. Local Scalars ..
+ LOGICAL LTYPE
+ INTEGER I, IP
+ DOUBLE PRECISION AB2, PALPHA, PBETA, SQRAB2
+C .. External Functions ..
+ LOGICAL LSAME
+ EXTERNAL LSAME
+C .. External Subroutines ..
+ EXTERNAL DGEMM, DGETRF, DGETRS, DGETRI, DLASCL, DSCAL,
+ $ DSWAP, XERBLA
+C .. Intrinsic Functions ..
+ INTRINSIC ABS, MAX, SIGN, SQRT
+C .. Executable Statements ..
+C
+ INFO = 0
+ LTYPE = LSAME( TYPE, 'D' )
+C
+C Test the input scalar arguments.
+C
+ IF( .NOT.LTYPE .AND. .NOT.LSAME( TYPE, 'C' ) ) THEN
+ INFO = -1
+ ELSE IF( N.LT.0 ) THEN
+ INFO = -2
+ ELSE IF( M.LT.0 ) THEN
+ INFO = -3
+ ELSE IF( P.LT.0 ) THEN
+ INFO = -4
+ ELSE IF( ALPHA.EQ.ZERO ) THEN
+ INFO = -5
+ ELSE IF( BETA.EQ.ZERO ) THEN
+ INFO = -6
+ ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
+ INFO = -8
+ ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
+ INFO = -10
+ ELSE IF( LDC.LT.MAX( 1, P ) ) THEN
+ INFO = -12
+ ELSE IF( LDD.LT.MAX( 1, P ) ) THEN
+ INFO = -14
+ ELSE IF( LDWORK.LT.MAX( 1, N ) ) THEN
+ INFO = -17
+ END IF
+C
+ IF ( INFO.NE.0 ) THEN
+C
+C Error return.
+C
+ CALL XERBLA( 'AB04MD', -INFO )
+ RETURN
+ END IF
+C
+C Quick return if possible.
+C
+ IF ( MAX( N, M, P ).EQ.0 )
+ $ RETURN
+C
+C (Note: Comments in the code beginning "Workspace:" describe the
+C minimal amount of real workspace needed at that point in the
+C code, as well as the preferred amount for good performance.
+C NB refers to the optimal block size for the immediately
+C following subroutine, as returned by ILAENV.)
+C
+ IF (LTYPE) THEN
+C
+C Discrete-time to continuous-time with (ALPHA, BETA).
+C
+ PALPHA = ALPHA
+ PBETA = BETA
+ ELSE
+C
+C Continuous-time to discrete-time with (ALPHA, BETA) is
+C equivalent with discrete-time to continuous-time with
+C (-BETA, -ALPHA), if B and C change the sign.
+C
+ PALPHA = -BETA
+ PBETA = -ALPHA
+ END IF
+C
+ AB2 = PALPHA*PBETA*TWO
+ SQRAB2 = SIGN( SQRT( ABS( AB2 ) ), PALPHA )
+C -1
+C Compute (alpha*I + A) .
+C
+ DO 10 I = 1, N
+ A(I,I) = A(I,I) + PALPHA
+ 10 CONTINUE
+C
+ CALL DGETRF( N, N, A, LDA, IWORK, INFO )
+C
+ IF (INFO.NE.0) THEN
+C
+C Error return.
+C
+ IF (LTYPE) THEN
+ INFO = 1
+ ELSE
+ INFO = 2
+ END IF
+ RETURN
+ END IF
+C -1
+C Compute (alpha*I+A) *B.
+C
+ CALL DGETRS( 'No transpose', N, M, A, LDA, IWORK, B, LDB, INFO )
+C -1
+C Compute D - C*(alpha*I+A) *B.
+C
+ CALL DGEMM( 'No transpose', 'No transpose', P, M, N, -ONE, C,
+ $ LDC, B, LDB, ONE, D, LDD )
+C
+C Scale B by sqrt(2*alpha*beta).
+C
+ CALL DLASCL( 'General', 0, 0, ONE, SQRAB2, N, M, B, LDB, INFO )
+C -1
+C Compute sqrt(2*alpha*beta)*C*(alpha*I + A) .
+C
+ CALL DTRSM( 'Right', 'Upper', 'No transpose', 'Non-unit', P, N,
+ $ SQRAB2, A, LDA, C, LDC )
+C
+ CALL DTRSM( 'Right', 'Lower', 'No transpose', 'Unit', P, N, ONE,
+ $ A, LDA, C, LDC )
+C
+C Apply column interchanges to the solution matrix.
+C
+ DO 20 I = N-1, 1, -1
+ IP = IWORK(I)
+ IF ( IP.NE.I )
+ $ CALL DSWAP( P, C(1,I), 1, C(1,IP), 1 )
+ 20 CONTINUE
+C -1
+C Compute beta*(alpha*I + A) *(A - alpha*I) as
+C -1
+C beta*I - 2*alpha*beta*(alpha*I + A) .
+C
+C Workspace: need N; prefer N*NB.
+C
+ CALL DGETRI( N, A, LDA, IWORK, DWORK, LDWORK, INFO )
+C
+ DO 30 I = 1, N
+ CALL DSCAL(N, -AB2, A(1,I), 1)
+ A(I,I) = A(I,I) + PBETA
+ 30 CONTINUE
+C
+ RETURN
+C *** Last line of AB04MD ***
+ END
Property changes on: trunk/octave-forge/main/control/src/AB04MD.f
___________________________________________________________________
Added: svn:executable
+ *
Modified: trunk/octave-forge/main/control/src/Makefile
===================================================================
--- trunk/octave-forge/main/control/src/Makefile 2011-09-03 16:37:43 UTC (rev 8487)
+++ trunk/octave-forge/main/control/src/Makefile 2011-09-03 18:50:35 UTC (rev 8488)
@@ -4,6 +4,7 @@
sltb01pd.oct slsb03od.oct slsg03bd.oct slag08bd.oct sltg01jd.oct \
sltg01hd.oct sltg01id.oct slsg02ad.oct sltg04bx.oct sltb01id.oct \
sltg01ad.oct slsb10id.oct slsb10kd.oct slsb10zd.oct sltb04bd.oct \
+ slab04md.oct \
is_real_scalar.oct is_real_vector.oct is_real_matrix.oct \
is_real_square_matrix.oct
@@ -212,6 +213,11 @@
TB04BX.f MA02AD.f MB02RD.f MB01PD.f MB02SD.f \
MB01QD.f
+# bilinear transformation
+slab04md.oct: slab04md.cc
+ mkoctfile slab04md.cc \
+ AB04MD.f
+
# helpers
is_real_scalar.oct: is_real_scalar.cc
mkoctfile is_real_scalar.cc
Added: trunk/octave-forge/main/control/src/slab04md.cc
===================================================================
--- trunk/octave-forge/main/control/src/slab04md.cc (rev 0)
+++ trunk/octave-forge/main/control/src/slab04md.cc 2011-09-03 18:50:35 UTC (rev 8488)
@@ -0,0 +1,128 @@
+/*
+
+Copyright (C) 2011 Lukas F. Reichlin
+
+This file is part of LTI Syncope.
+
+LTI Syncope is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+LTI Syncope is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with LTI Syncope. If not, see <http://www.gnu.org/licenses/>.
+
+Discrete-time <--> continuous-time systems conversion
+by a bilinear transformation.
+Uses SLICOT AB04MD by courtesy of NICONET e.V.
+<http://www.slicot.org>
+
+Author: Lukas Reichlin <luk...@gm...>
+Created: September 2011
+Version: 0.1
+
+*/
+
+#include <octave/oct.h>
+#include <f77-fcn.h>
+#include "common.cc"
+
+extern "C"
+{
+ int F77_FUNC (ab04md, AB04MD)
+ (char& TYPE,
+ int& N, int& M, int& P,
+ double& ALPHA, double& BETA,
+ double* A, int& LDA,
+ double* B, int& LDB,
+ double* C, int& LDC,
+ double* D, int& LDD,
+ int* IWORK,
+ double* DWORK, int& LDWORK,
+ int& INFO);
+}
+
+DEFUN_DLD (slab04md, args, nargout,
+ "-*- texinfo -*-\n\
+Slicot AB04MD Release 5.0\n\
+No argument checking.\n\
+For internal use only.")
+{
+ int nargin = args.length ();
+ octave_value_list retval;
+
+ if (nargin != 7)
+ {
+ print_usage ();
+ }
+ else
+ {
+ // arguments in
+ char type;
+
+ Matrix a = args(0).matrix_value ();
+ Matrix b = args(1).matrix_value ();
+ Matrix c = args(2).matrix_value ();
+ Matrix d = args(3).matrix_value ();
+
+ double alpha = args(4).double_value ();
+ double beta = args(5).double_value ();
+ int discrete = args(6).int_value ();
+
+ if (discrete == 0)
+ type = 'C';
+ else
+ type = 'D';
+
+ int n = a.rows (); // n: number of states
+ int m = b.columns (); // m: number of inputs
+ int p = c.rows (); // p: number of outputs
+
+ int lda = max (1, n);
+ int ldb = max (1, n);
+ int ldc = max (1, p);
+ int ldd = max (1, p);
+
+ // workspace
+ int ldwork = max (1, n);
+
+ OCTAVE_LOCAL_BUFFER (int, iwork, n);
+ OCTAVE_LOCAL_BUFFER (double, dwork, ldwork);
+
+ // error indicator
+ int info;
+
+
+ // SLICOT routine AB04MD
+ F77_XFCN (ab04md, AB04MD,
+ (type,
+ n, m, p,
+ alpha, beta,
+ a.fortran_vec (), lda,
+ b.fortran_vec (), ldb,
+ c.fortran_vec (), ldc,
+ d.fortran_vec (), ldd,
+ iwork,
+ dwork, ldwork,
+ info));
+
+ if (f77_exception_encountered)
+ error ("slab04md: exception in SLICOT subroutine AB04MD");
+
+ if (info != 0)
+ error ("slab04md: AB04MD returned info = %d", info);
+
+ // return values
+ retval(0) = a;
+ retval(1) = b;
+ retval(2) = c;
+ retval(3) = d;
+ }
+
+ return retval;
+}
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-09-03 21:01:24
|
Revision: 8490
http://octave.svn.sourceforge.net/octave/?rev=8490&view=rev
Author: paramaniac
Date: 2011-09-03 21:01:18 +0000 (Sat, 03 Sep 2011)
Log Message:
-----------
control: touch up conversion routines
Modified Paths:
--------------
trunk/octave-forge/main/control/doc/NEWS
trunk/octave-forge/main/control/inst/@lti/c2d.m
trunk/octave-forge/main/control/inst/@lti/d2c.m
trunk/octave-forge/main/control/inst/@ss/__d2c__.m
Modified: trunk/octave-forge/main/control/doc/NEWS
===================================================================
--- trunk/octave-forge/main/control/doc/NEWS 2011-09-03 19:42:05 UTC (rev 8489)
+++ trunk/octave-forge/main/control/doc/NEWS 2011-09-03 21:01:18 UTC (rev 8490)
@@ -1,6 +1,18 @@
Summary of important user-visible changes for releases of the control package
===============================================================================
+control-2.1.55 Release Date: 2011-xx-yy Release Manager: Lukas Reichlin
+===============================================================================
+
+** @lti/c2d
+ -- Support for "tustin" method added.
+ -- Improved Texinfo string.
+
+** @lti/d2c
+ -- INDEX
+
+
+===============================================================================
control-2.1.54 Release Date: 2011-08-22 Release Manager: Lukas Reichlin
===============================================================================
Modified: trunk/octave-forge/main/control/inst/@lti/c2d.m
===================================================================
--- trunk/octave-forge/main/control/inst/@lti/c2d.m 2011-09-03 19:42:05 UTC (rev 8489)
+++ trunk/octave-forge/main/control/inst/@lti/c2d.m 2011-09-03 21:01:18 UTC (rev 8490)
@@ -1,4 +1,4 @@
-## Copyright (C) 2009 Lukas F. Reichlin
+## Copyright (C) 2009, 2011 Lukas F. Reichlin
##
## This file is part of LTI Syncope.
##
@@ -18,12 +18,35 @@
## -*- texinfo -*-
## @deftypefn {Function File} {@var{sys} =} c2d (@var{sys}, @var{tsam})
## @deftypefnx {Function File} {@var{sys} =} c2d (@var{sys}, @var{tsam}, @var{method})
-## Convert the continuous lti model into its discrete time equivalent.
+## Convert the continuous lti model into its discrete-time equivalent.
+##
+## @strong{Inputs}
+## @table @var
+## @item sys
+## Continuous-time LTI model.
+## @item tsam
+## Sampling time in seconds.
+## @item method
+## Optional conversion method. If not specified, default method @var{"zoh"}
+## is taken.
+## @table @code
+## @item "zoh"
+## Zero-order hold or matrix exponential.
+## @item "tustin", "bilin"
+## Bilinear transformation or Tustin approximation.
+## @end table
+## @end table
+##
+## @strong{Outputs}
+## @table @var
+## @item sys
+## Discrete-time LTI model.
+## @end table
## @end deftypefn
## Author: Lukas Reichlin <luk...@gm...>
## Created: October 2009
-## Version: 0.1
+## Version: 0.2
function sys = c2d (sys, tsam, method = "std")
@@ -35,6 +58,10 @@
error ("c2d: first argument is not an lti model");
endif
+ if (isdt (sys))
+ error ("c2d: system is already discrete-time");
+ endif
+
if (! issample (tsam))
error ("c2d: second argument is not a valid sample time");
endif
Modified: trunk/octave-forge/main/control/inst/@lti/d2c.m
===================================================================
--- trunk/octave-forge/main/control/inst/@lti/d2c.m 2011-09-03 19:42:05 UTC (rev 8489)
+++ trunk/octave-forge/main/control/inst/@lti/d2c.m 2011-09-03 21:01:18 UTC (rev 8490)
@@ -19,6 +19,27 @@
## @deftypefn {Function File} {@var{sys} =} d2c (@var{sys})
## @deftypefnx {Function File} {@var{sys} =} d2c (@var{sys}, @var{method})
## Convert the discrete lti model into its continuous-time equivalent.
+##
+## @strong{Inputs}
+## @table @var
+## @item sys
+## Discrete-time LTI model.
+## @item method
+## Optional conversion method. If not specified, default method @var{"zoh"}
+## is taken.
+## @table @code
+## @item "zoh"
+## Zero-order hold or matrix logarithm.
+## @item "tustin", "bilin"
+## Bilinear transformation or Tustin approximation.
+## @end table
+## @end table
+##
+## @strong{Outputs}
+## @table @var
+## @item sys
+## Continuous-time LTI model.
+## @end table
## @end deftypefn
## Author: Lukas Reichlin <luk...@gm...>
@@ -35,6 +56,10 @@
error ("d2c: first argument is not an lti model");
endif
+ if (isct (sys))
+ error ("d2c: system is already continuous-time");
+ endif
+
if (! ischar (method))
error ("c2d: second argument is not a string");
endif
Modified: trunk/octave-forge/main/control/inst/@ss/__d2c__.m
===================================================================
--- trunk/octave-forge/main/control/inst/@ss/__d2c__.m 2011-09-03 19:42:05 UTC (rev 8489)
+++ trunk/octave-forge/main/control/inst/@ss/__d2c__.m 2011-09-03 21:01:18 UTC (rev 8490)
@@ -36,6 +36,7 @@
endif
endif
+ error ("ss: d2c: zoh method not implemented yet");
[n, m] = size (sys.b); # n: states, m: inputs
## TODO: use SLICOT MB05OD
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-09-04 12:38:50
|
Revision: 8492
http://octave.svn.sourceforge.net/octave/?rev=8492&view=rev
Author: paramaniac
Date: 2011-09-04 12:38:43 +0000 (Sun, 04 Sep 2011)
Log Message:
-----------
control: use Slicot SB10JD for descriptor to regular state-space conversion
Modified Paths:
--------------
trunk/octave-forge/main/control/devel/makefile_all.m
trunk/octave-forge/main/control/doc/NEWS
trunk/octave-forge/main/control/inst/@lti/ssdata.m
trunk/octave-forge/main/control/inst/@ss/__c2d__.m
trunk/octave-forge/main/control/inst/@ss/__d2c__.m
trunk/octave-forge/main/control/src/Makefile
Added Paths:
-----------
trunk/octave-forge/main/control/devel/makefile_conversions.m
trunk/octave-forge/main/control/inst/__dss2ss__.m
trunk/octave-forge/main/control/src/slsb10jd.cc
Removed Paths:
-------------
trunk/octave-forge/main/control/devel/makefile_ss2tf.m
Modified: trunk/octave-forge/main/control/devel/makefile_all.m
===================================================================
--- trunk/octave-forge/main/control/devel/makefile_all.m 2011-09-03 21:04:15 UTC (rev 8491)
+++ trunk/octave-forge/main/control/devel/makefile_all.m 2011-09-04 12:38:43 UTC (rev 8492)
@@ -12,6 +12,7 @@
cd (srcdir);
makefile_chol
+makefile_conversions
makefile_h2syn
makefile_hankel
makefile_helpers
@@ -23,7 +24,6 @@
makefile_norm
makefile_place
makefile_scale
-makefile_ss2tf
makefile_staircase
makefile_tustin
makefile_zero
Copied: trunk/octave-forge/main/control/devel/makefile_conversions.m (from rev 8491, trunk/octave-forge/main/control/devel/makefile_ss2tf.m)
===================================================================
--- trunk/octave-forge/main/control/devel/makefile_conversions.m (rev 0)
+++ trunk/octave-forge/main/control/devel/makefile_conversions.m 2011-09-04 12:38:43 UTC (rev 8492)
@@ -0,0 +1,25 @@
+## ==============================================================================
+## Developer Makefile for OCT-files
+## ==============================================================================
+## USAGE: * fetch control from Octave-Forge by svn
+## * add control/inst, control/src and control/devel to your Octave path
+## * run makefile_*
+## ==============================================================================
+
+homedir = pwd ();
+develdir = fileparts (which ("makefile_conversions"));
+srcdir = [develdir, "/../src"];
+cd (srcdir);
+
+## state-space to transfer function
+mkoctfile sltb04bd.cc \
+ TB04BD.f MC01PY.f TB01ID.f TB01ZD.f MC01PD.f \
+ TB04BX.f MA02AD.f MB02RD.f MB01PD.f MB02SD.f \
+ MB01QD.f
+
+## descriptor to regular state-space
+mkoctfile slsb10jd.cc \
+ SB10JD.f
+
+system ("rm *.o");
+cd (homedir);
Deleted: trunk/octave-forge/main/control/devel/makefile_ss2tf.m
===================================================================
--- trunk/octave-forge/main/control/devel/makefile_ss2tf.m 2011-09-03 21:04:15 UTC (rev 8491)
+++ trunk/octave-forge/main/control/devel/makefile_ss2tf.m 2011-09-04 12:38:43 UTC (rev 8492)
@@ -1,20 +0,0 @@
-## ==============================================================================
-## Developer Makefile for OCT-files
-## ==============================================================================
-## USAGE: * fetch control from Octave-Forge by svn
-## * add control/inst, control/src and control/devel to your Octave path
-## * run makefile_*
-## ==============================================================================
-
-homedir = pwd ();
-develdir = fileparts (which ("makefile_ss2tf"));
-srcdir = [develdir, "/../src"];
-cd (srcdir);
-
-mkoctfile sltb04bd.cc \
- TB04BD.f MC01PY.f TB01ID.f TB01ZD.f MC01PD.f \
- TB04BX.f MA02AD.f MB02RD.f MB01PD.f MB02SD.f \
- MB01QD.f
-
-system ("rm *.o");
-cd (homedir);
Modified: trunk/octave-forge/main/control/doc/NEWS
===================================================================
--- trunk/octave-forge/main/control/doc/NEWS 2011-09-03 21:04:15 UTC (rev 8491)
+++ trunk/octave-forge/main/control/doc/NEWS 2011-09-04 12:38:43 UTC (rev 8492)
@@ -10,6 +10,9 @@
** @lti/d2c
-- INDEX
+
+** Conversion from descriptor to regular state-space is now performed by SLICOT
+ routine SB10JD.
===============================================================================
Modified: trunk/octave-forge/main/control/inst/@lti/ssdata.m
===================================================================
--- trunk/octave-forge/main/control/inst/@lti/ssdata.m 2011-09-03 21:04:15 UTC (rev 8491)
+++ trunk/octave-forge/main/control/inst/@lti/ssdata.m 2011-09-04 12:38:43 UTC (rev 8492)
@@ -45,7 +45,7 @@
## Author: Lukas Reichlin <luk...@gm...>
## Created: September 2009
-## Version: 0.3
+## Version: 0.4
function [a, b, c, d, tsam, scaled] = ssdata (sys)
@@ -55,15 +55,8 @@
[a, b, c, d, e, ~, scaled] = __sys_data__ (sys);
- if (! isempty (e))
- if (rcond (e) < eps) # check for singularity
- error ("ss: ssdata: descriptor matrice ""e"" singular");
- endif
+ [a, b, c, d, e] = __dss2ss__ (a, b, c, d, e);
- a = e \ a;
- b = e \ b;
- endif
-
tsam = sys.tsam;
endfunction
\ No newline at end of file
Modified: trunk/octave-forge/main/control/inst/@ss/__c2d__.m
===================================================================
--- trunk/octave-forge/main/control/inst/@ss/__c2d__.m 2011-09-03 21:04:15 UTC (rev 8491)
+++ trunk/octave-forge/main/control/inst/@ss/__c2d__.m 2011-09-04 12:38:43 UTC (rev 8492)
@@ -26,40 +26,20 @@
switch (method)
case {"zoh", "std"}
- if (! isempty (sys.e))
- if (rcond (sys.e) < eps)
- error ("ss: c2d: zero-order hold method requires proper system");
- else
- sys.a = sys.e \ sys.a;
- sys.b = sys.e \ sys.b;
- sys.e = []; # require ordinary state-space model
- endif
- endif
-
+ [sys.a, sys.b, sys.c, sys.d, sys.e] = __dss2ss__ (sys.a, sys.b, sys.c, sys.d, sys.e);
[n, m] = size (sys.b); # n: states, m: inputs
-
## TODO: use SLICOT MB05OD
tmp = expm ([sys.a*tsam, sys.b*tsam; zeros(m, n+m)]);
-
sys.a = tmp (1:n, 1:n); # F
sys.b = tmp (1:n, n+(1:m)); # G
case {"tustin", "bilin"}
- if (! isempty (sys.e))
- if (rcond (sys.e) < eps)
- error ("ss: c2d: tustin method requires proper system");
- else
- sys.a = sys.e \ sys.a;
- sys.b = sys.e \ sys.b;
- sys.e = []; # require ordinary state-space model
- endif
- endif
-
+ [sys.a, sys.b, sys.c, sys.d, sys.e] = __dss2ss__ (sys.a, sys.b, sys.c, sys.d, sys.e);
[sys.a, sys.b, sys.c, sys.d] = slab04md (sys.a, sys.b, sys.c, sys.d, 1, 2/tsam, false);
+ ## TODO: descriptor case
otherwise
error ("ss: c2d: %s is an invalid or missing method", method);
-
endswitch
endfunction
\ No newline at end of file
Modified: trunk/octave-forge/main/control/inst/@ss/__d2c__.m
===================================================================
--- trunk/octave-forge/main/control/inst/@ss/__d2c__.m 2011-09-03 21:04:15 UTC (rev 8491)
+++ trunk/octave-forge/main/control/inst/@ss/__d2c__.m 2011-09-04 12:38:43 UTC (rev 8492)
@@ -26,37 +26,14 @@
switch (method)
case {"zoh", "std"}
- if (! isempty (sys.e))
- if (rcond (sys.e) < eps)
- error ("ss: d2c: zero-order hold method requires proper system");
- else
- sys.a = sys.e \ sys.a;
- sys.b = sys.e \ sys.b;
- sys.e = []; # require ordinary state-space model
- endif
- endif
+ [sys.a, sys.b, sys.c, sys.d, sys.e] = __dss2ss__ (sys.a, sys.b, sys.c, sys.d, sys.e);
error ("ss: d2c: zoh method not implemented yet");
- [n, m] = size (sys.b); # n: states, m: inputs
- ## TODO: use SLICOT MB05OD
- tmp = expm ([sys.a*tsam, sys.b*tsam; zeros(m, n+m)]);
-
- sys.a = tmp (1:n, 1:n); # F
- sys.b = tmp (1:n, n+(1:m)); # G
-
case {"tustin", "bilin"}
- if (! isempty (sys.e))
- if (rcond (sys.e) < eps)
- error ("ss: d2c: tustin method requires proper system");
- else
- sys.a = sys.e \ sys.a;
- sys.b = sys.e \ sys.b;
- sys.e = []; # require ordinary state-space model
- endif
- endif
-
+ [sys.a, sys.b, sys.c, sys.d, sys.e] = __dss2ss__ (sys.a, sys.b, sys.c, sys.d, sys.e);
[sys.a, sys.b, sys.c, sys.d] = slab04md (sys.a, sys.b, sys.c, sys.d, 1, 2/tsam, true);
+ ## TODO: descriptor case
otherwise
error ("ss: d2c: %s is an invalid or missing method", method);
Added: trunk/octave-forge/main/control/inst/__dss2ss__.m
===================================================================
--- trunk/octave-forge/main/control/inst/__dss2ss__.m (rev 0)
+++ trunk/octave-forge/main/control/inst/__dss2ss__.m 2011-09-04 12:38:43 UTC (rev 8492)
@@ -0,0 +1,36 @@
+## Copyright (C) 2011 Lukas F. Reichlin
+##
+## This file is part of LTI Syncope.
+##
+## LTI Syncope is free software: you can redistribute it and/or modify
+## it under the terms of the GNU General Public License as published by
+## the Free Software Foundation, either version 3 of the License, or
+## (at your option) any later version.
+##
+## LTI Syncope is distributed in the hope that it will be useful,
+## but WITHOUT ANY WARRANTY; without even the implied warranty of
+## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+## GNU General Public License for more details.
+##
+## You should have received a copy of the GNU General Public License
+## along with LTI Syncope. If not, see <http://www.gnu.org/licenses/>.
+
+## -*- texinfo -*-
+## Convert descriptor state-space system into regular state-space form.
+
+## Author: Lukas Reichlin <luk...@gm...>
+## Created: September 2011
+## Version: 0.1
+
+function [a, b, c, d, e] = __dss2ss__ (a, b, c, d, e)
+
+ if (isempty (e))
+ return;
+ elseif (rcond (e) < eps) # check for singularity
+ error ("ss: dss2ss: descriptor matrice ""e"" singular");
+ else
+ [a, b, c, d] = slsb10jd (a, b, c, d, e);
+ e = [];
+ endif
+
+endfunction
\ No newline at end of file
Modified: trunk/octave-forge/main/control/src/Makefile
===================================================================
--- trunk/octave-forge/main/control/src/Makefile 2011-09-03 21:04:15 UTC (rev 8491)
+++ trunk/octave-forge/main/control/src/Makefile 2011-09-04 12:38:43 UTC (rev 8492)
@@ -4,7 +4,7 @@
sltb01pd.oct slsb03od.oct slsg03bd.oct slag08bd.oct sltg01jd.oct \
sltg01hd.oct sltg01id.oct slsg02ad.oct sltg04bx.oct sltb01id.oct \
sltg01ad.oct slsb10id.oct slsb10kd.oct slsb10zd.oct sltb04bd.oct \
- slab04md.oct \
+ slab04md.oct slsb10jd.oct \
is_real_scalar.oct is_real_vector.oct is_real_matrix.oct \
is_real_square_matrix.oct
@@ -218,6 +218,11 @@
mkoctfile slab04md.cc \
AB04MD.f
+# descriptor to regular state-space conversion
+slsb10jd.oct: slsb10jd.cc
+ mkoctfile slsb10jd.cc \
+ SB10JD.f
+
# helpers
is_real_scalar.oct: is_real_scalar.cc
mkoctfile is_real_scalar.cc
Added: trunk/octave-forge/main/control/src/slsb10jd.cc
===================================================================
--- trunk/octave-forge/main/control/src/slsb10jd.cc (rev 0)
+++ trunk/octave-forge/main/control/src/slsb10jd.cc 2011-09-04 12:38:43 UTC (rev 8492)
@@ -0,0 +1,122 @@
+/*
+
+Copyright (C) 2011 Lukas F. Reichlin
+
+This file is part of LTI Syncope.
+
+LTI Syncope is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+LTI Syncope is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with LTI Syncope. If not, see <http://www.gnu.org/licenses/>.
+
+Convert descriptor state-space system into regular state-space form.
+Uses SLICOT SB10JD by courtesy of NICONET e.V.
+<http://www.slicot.org>
+
+Author: Lukas Reichlin <luk...@gm...>
+Created: September 2011
+Version: 0.1
+
+*/
+
+#include <octave/oct.h>
+#include <f77-fcn.h>
+#include "common.cc"
+
+extern "C"
+{
+ int F77_FUNC (sb10jd, SB10JD)
+ (int& N, int& M, int& NP,
+ double* A, int& LDA,
+ double* B, int& LDB,
+ double* C, int& LDC,
+ double* D, int& LDD,
+ double* E, int& LDE,
+ int& NSYS,
+ double* DWORK, int& LDWORK,
+ int& INFO);
+}
+
+DEFUN_DLD (slsb10jd, args, nargout,
+ "-*- texinfo -*-\n\
+Slicot SB10JD Release 5.0\n\
+No argument checking.\n\
+For internal use only.")
+{
+ int nargin = args.length ();
+ octave_value_list retval;
+
+ if (nargin != 5)
+ {
+ print_usage ();
+ }
+ else
+ {
+ // arguments in
+ Matrix a = args(0).matrix_value ();
+ Matrix b = args(1).matrix_value ();
+ Matrix c = args(2).matrix_value ();
+ Matrix d = args(3).matrix_value ();
+ Matrix e = args(4).matrix_value ();
+
+ int n = a.rows (); // n: number of states
+ int m = b.columns (); // m: number of inputs
+ int np = c.rows (); // np: number of outputs
+
+ int lda = max (1, n);
+ int ldb = max (1, n);
+ int ldc = max (1, np);
+ int ldd = max (1, np);
+ int lde = max (1, n);
+
+ // arguments out
+ int nsys;
+
+ // workspace
+ int ldwork = max (1, 2*n*n + 2*n + n*max (5, n + m + np));
+ OCTAVE_LOCAL_BUFFER (double, dwork, ldwork);
+
+ // error indicator
+ int info;
+
+
+ // SLICOT routine SB10JD
+ F77_XFCN (sb10jd, SB10JD,
+ (n, m, np,
+ a.fortran_vec (), lda,
+ b.fortran_vec (), ldb,
+ c.fortran_vec (), ldc,
+ d.fortran_vec (), ldd,
+ e.fortran_vec (), lde,
+ nsys,
+ dwork, ldwork,
+ info));
+
+ if (f77_exception_encountered)
+ error ("slsb10jd: exception in SLICOT subroutine SB10JD");
+
+ if (info != 0)
+ error ("slsb10jd: SB10JD returned info = %d", info);
+
+ // resize
+ a.resize (nsys, nsys);
+ b.resize (nsys, m);
+ c.resize (np, nsys);
+
+ // return values
+ retval(0) = a;
+ retval(1) = b;
+ retval(2) = c;
+ retval(3) = d;
+ }
+
+ return retval;
+}
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-09-05 17:58:42
|
Revision: 8500
http://octave.svn.sourceforge.net/octave/?rev=8500&view=rev
Author: paramaniac
Date: 2011-09-05 17:58:35 +0000 (Mon, 05 Sep 2011)
Log Message:
-----------
control: update index and news files
Modified Paths:
--------------
trunk/octave-forge/main/control/INDEX
trunk/octave-forge/main/control/doc/NEWS
Modified: trunk/octave-forge/main/control/INDEX
===================================================================
--- trunk/octave-forge/main/control/INDEX 2011-09-05 17:45:13 UTC (rev 8499)
+++ trunk/octave-forge/main/control/INDEX 2011-09-05 17:58:35 UTC (rev 8500)
@@ -19,6 +19,7 @@
@lti/tfdata
Model Conversions
@lti/c2d
+ @lti/d2c
@lti/prescale
@lti/xperm
Model Interconnections
Modified: trunk/octave-forge/main/control/doc/NEWS
===================================================================
--- trunk/octave-forge/main/control/doc/NEWS 2011-09-05 17:45:13 UTC (rev 8499)
+++ trunk/octave-forge/main/control/doc/NEWS 2011-09-05 17:58:35 UTC (rev 8500)
@@ -9,10 +9,13 @@
-- Improved Texinfo string.
** @lti/d2c
- -- INDEX
+ -- Discrete to continuous-time conversion added. However, support is
+ limited to state-space models and the zero-order hold and tustin methods.
** Conversion from descriptor to regular state-space is now performed by SLICOT
- routine SB10JD.
+ routine SB10JD. Better numerical results are to be expected over the
+ previous naive inversion formula. This conversion is used internally for
+ ssdata and some other functions.
===============================================================================
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-09-05 19:14:09
|
Revision: 8502
http://octave.svn.sourceforge.net/octave/?rev=8502&view=rev
Author: paramaniac
Date: 2011-09-05 19:14:03 +0000 (Mon, 05 Sep 2011)
Log Message:
-----------
control: add pre-warping option to c2d and d2c
Modified Paths:
--------------
trunk/octave-forge/main/control/doc/NEWS
trunk/octave-forge/main/control/inst/@lti/c2d.m
trunk/octave-forge/main/control/inst/@lti/d2c.m
trunk/octave-forge/main/control/inst/@ss/__c2d__.m
trunk/octave-forge/main/control/inst/@ss/__d2c__.m
Modified: trunk/octave-forge/main/control/doc/NEWS
===================================================================
--- trunk/octave-forge/main/control/doc/NEWS 2011-09-05 18:02:37 UTC (rev 8501)
+++ trunk/octave-forge/main/control/doc/NEWS 2011-09-05 19:14:03 UTC (rev 8502)
@@ -5,12 +5,13 @@
===============================================================================
** @lti/c2d
- -- Support for "tustin" method added.
+ -- Support for "tustin" and "prewarp" method added.
-- Improved Texinfo string.
** @lti/d2c
-- Discrete to continuous-time conversion added. However, support is
- limited to state-space models and the zero-order hold and tustin methods.
+ limited to state-space models and the zero-order hold, tustin and pre-
+ warping methods.
** Conversion from descriptor to regular state-space is now performed by SLICOT
routine SB10JD. Better numerical results are to be expected over the
Modified: trunk/octave-forge/main/control/inst/@lti/c2d.m
===================================================================
--- trunk/octave-forge/main/control/inst/@lti/c2d.m 2011-09-05 18:02:37 UTC (rev 8501)
+++ trunk/octave-forge/main/control/inst/@lti/c2d.m 2011-09-05 19:14:03 UTC (rev 8502)
@@ -18,6 +18,7 @@
## -*- texinfo -*-
## @deftypefn {Function File} {@var{sys} =} c2d (@var{sys}, @var{tsam})
## @deftypefnx {Function File} {@var{sys} =} c2d (@var{sys}, @var{tsam}, @var{method})
+## @deftypefnx {Function File} {@var{sys} =} c2d (@var{sys}, @var{tsam}, @var{"prewarp"}, @var{w0})
## Convert the continuous lti model into its discrete-time equivalent.
##
## @strong{Inputs}
@@ -29,11 +30,13 @@
## @item method
## Optional conversion method. If not specified, default method @var{"zoh"}
## is taken.
-## @table @code
+## @table @var
## @item "zoh"
## Zero-order hold or matrix exponential.
## @item "tustin", "bilin"
## Bilinear transformation or Tustin approximation.
+## @item "prewarp"
+## Bilinear transformation with pre-warping at frequency @var{w0}.
## @end table
## @end table
##
@@ -48,9 +51,9 @@
## Created: October 2009
## Version: 0.2
-function sys = c2d (sys, tsam, method = "std")
+function sys = c2d (sys, tsam, method = "std", w0 = 0)
- if (nargin < 2 || nargin > 3)
+ if (nargin < 2 || nargin > 4)
print_usage ();
endif
@@ -70,7 +73,7 @@
error ("c2d: third argument is not a string");
endif
- sys = __c2d__ (sys, tsam, method);
+ sys = __c2d__ (sys, tsam, lower (method), w0);
sys.tsam = tsam;
-endfunction
\ No newline at end of file
+endfunction
Modified: trunk/octave-forge/main/control/inst/@lti/d2c.m
===================================================================
--- trunk/octave-forge/main/control/inst/@lti/d2c.m 2011-09-05 18:02:37 UTC (rev 8501)
+++ trunk/octave-forge/main/control/inst/@lti/d2c.m 2011-09-05 19:14:03 UTC (rev 8502)
@@ -18,6 +18,7 @@
## -*- texinfo -*-
## @deftypefn {Function File} {@var{sys} =} d2c (@var{sys})
## @deftypefnx {Function File} {@var{sys} =} d2c (@var{sys}, @var{method})
+## @deftypefnx {Function File} {@var{sys} =} d2c (@var{sys}, @var{"prewarp"}, @var{w0})
## Convert the discrete lti model into its continuous-time equivalent.
##
## @strong{Inputs}
@@ -27,11 +28,13 @@
## @item method
## Optional conversion method. If not specified, default method @var{"zoh"}
## is taken.
-## @table @code
+## @table @var
## @item "zoh"
## Zero-order hold or matrix logarithm.
## @item "tustin", "bilin"
## Bilinear transformation or Tustin approximation.
+## @item "prewarp"
+## Bilinear transformation with pre-warping at frequency @var{w0}.
## @end table
## @end table
##
@@ -46,9 +49,9 @@
## Created: September 2011
## Version: 0.1
-function sys = d2c (sys, method = "std")
+function sys = d2c (sys, method = "std", w0 = 0)
- if (nargin == 0 || nargin > 2)
+ if (nargin == 0 || nargin > 3)
print_usage ();
endif
@@ -64,7 +67,7 @@
error ("d2c: second argument is not a string");
endif
- sys = __d2c__ (sys, sys.tsam, method);
+ sys = __d2c__ (sys, sys.tsam, lower (method), w0);
sys.tsam = 0;
endfunction
Modified: trunk/octave-forge/main/control/inst/@ss/__c2d__.m
===================================================================
--- trunk/octave-forge/main/control/inst/@ss/__c2d__.m 2011-09-05 18:02:37 UTC (rev 8501)
+++ trunk/octave-forge/main/control/inst/@ss/__c2d__.m 2011-09-05 19:14:03 UTC (rev 8502)
@@ -22,10 +22,10 @@
## Created: October 2009
## Version: 0.2
-function sys = __c2d__ (sys, tsam, method = "zoh")
+function sys = __c2d__ (sys, tsam, method = "zoh", w0 = 0)
- switch (method)
- case {"zoh", "std"}
+ switch (method(1))
+ case {"z", "s"} # {"zoh", "std"}
[sys.a, sys.b, sys.c, sys.d, sys.e] = __dss2ss__ (sys.a, sys.b, sys.c, sys.d, sys.e);
[n, m] = size (sys.b); # n: states, m: inputs
## TODO: use SLICOT MB05OD
@@ -33,13 +33,16 @@
sys.a = tmp (1:n, 1:n); # F
sys.b = tmp (1:n, n+(1:m)); # G
- case {"tustin", "bilin"}
+ case {"t", "b", "p"} # {"tustin", "bilin", "prewarp"}
+ if (method(1) == "p") # prewarping
+ beta = w0 / tan (w0*tsam/2);
+ else
+ beta = 2/tsam;
+ endif
[sys.a, sys.b, sys.c, sys.d, sys.e] = __dss2ss__ (sys.a, sys.b, sys.c, sys.d, sys.e);
- [sys.a, sys.b, sys.c, sys.d] = slab04md (sys.a, sys.b, sys.c, sys.d, 1, 2/tsam, false);
+ [sys.a, sys.b, sys.c, sys.d] = slab04md (sys.a, sys.b, sys.c, sys.d, 1, beta, false);
## TODO: descriptor case
- ## TODO: case "prewarp"
-
otherwise
error ("ss: c2d: %s is an invalid or missing method", method);
endswitch
Modified: trunk/octave-forge/main/control/inst/@ss/__d2c__.m
===================================================================
--- trunk/octave-forge/main/control/inst/@ss/__d2c__.m 2011-09-05 18:02:37 UTC (rev 8501)
+++ trunk/octave-forge/main/control/inst/@ss/__d2c__.m 2011-09-05 19:14:03 UTC (rev 8502)
@@ -22,10 +22,10 @@
## Created: September 2011
## Version: 0.1
-function sys = __d2c__ (sys, tsam, method = "zoh")
+function sys = __d2c__ (sys, tsam, method = "zoh", w0 = 0)
- switch (method)
- case {"zoh", "std"}
+ switch (method(1))
+ case {"z", "s"} # {"zoh", "std"}
[sys.a, sys.b, sys.c, sys.d, sys.e] = __dss2ss__ (sys.a, sys.b, sys.c, sys.d, sys.e);
[n, m] = size (sys.b); # n: states, m: inputs
tmp = logm ([sys.a, sys.b; zeros(m,n), eye(m)]) / tsam;
@@ -35,13 +35,16 @@
sys.a = real (tmp(1:n, 1:n));
sys.b = real (tmp(1:n, n+1:n+m));
- case {"tustin", "bilin"}
+ case {"t", "b", "p"} # {"tustin", "bilin", "prewarp"}
+ if (method(1) == "p") # prewarping
+ beta = w0 / tan (w0*tsam/2);
+ else
+ beta = 2/tsam;
+ endif
[sys.a, sys.b, sys.c, sys.d, sys.e] = __dss2ss__ (sys.a, sys.b, sys.c, sys.d, sys.e);
- [sys.a, sys.b, sys.c, sys.d] = slab04md (sys.a, sys.b, sys.c, sys.d, 1, 2/tsam, true);
+ [sys.a, sys.b, sys.c, sys.d] = slab04md (sys.a, sys.b, sys.c, sys.d, 1, beta, true);
## TODO: descriptor case
- ## TODO: case "prewarp"
-
otherwise
error ("ss: d2c: %s is an invalid or missing method", method);
endswitch
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-09-06 11:36:17
|
Revision: 8506
http://octave.svn.sourceforge.net/octave/?rev=8506&view=rev
Author: paramaniac
Date: 2011-09-06 11:36:10 +0000 (Tue, 06 Sep 2011)
Log Message:
-----------
control: enable c2d and d2c for tf via ss functions
Modified Paths:
--------------
trunk/octave-forge/main/control/doc/NEWS
trunk/octave-forge/main/control/inst/@lti/c2d.m
trunk/octave-forge/main/control/inst/@lti/d2c.m
trunk/octave-forge/main/control/inst/@tf/__c2d__.m
trunk/octave-forge/main/control/inst/@tf/__d2c__.m
Modified: trunk/octave-forge/main/control/doc/NEWS
===================================================================
--- trunk/octave-forge/main/control/doc/NEWS 2011-09-06 06:03:10 UTC (rev 8505)
+++ trunk/octave-forge/main/control/doc/NEWS 2011-09-06 11:36:10 UTC (rev 8506)
@@ -10,8 +10,7 @@
** @lti/d2c
-- Discrete to continuous-time conversion added. However, support is
- limited to state-space models and the zero-order hold, tustin and pre-
- warping methods.
+ limited to the zero-order hold, tustin and pre-warping methods.
** Conversion from descriptor to regular state-space is now performed by SLICOT
routine SB10JD. Better numerical results are to be expected over the
Modified: trunk/octave-forge/main/control/inst/@lti/c2d.m
===================================================================
--- trunk/octave-forge/main/control/inst/@lti/c2d.m 2011-09-06 06:03:10 UTC (rev 8505)
+++ trunk/octave-forge/main/control/inst/@lti/c2d.m 2011-09-06 11:36:10 UTC (rev 8506)
@@ -24,7 +24,7 @@
## @strong{Inputs}
## @table @var
## @item sys
-## Continuous-time LTI model. Only state-space models have been implemented so far.
+## Continuous-time LTI model.
## @item tsam
## Sampling time in seconds.
## @item method
Modified: trunk/octave-forge/main/control/inst/@lti/d2c.m
===================================================================
--- trunk/octave-forge/main/control/inst/@lti/d2c.m 2011-09-06 06:03:10 UTC (rev 8505)
+++ trunk/octave-forge/main/control/inst/@lti/d2c.m 2011-09-06 11:36:10 UTC (rev 8506)
@@ -24,7 +24,7 @@
## @strong{Inputs}
## @table @var
## @item sys
-## Discrete-time LTI model. Only state-space models have been implemented so far.
+## Discrete-time LTI model.
## @item method
## Optional conversion method. If not specified, default method @var{"zoh"}
## is taken.
Modified: trunk/octave-forge/main/control/inst/@tf/__c2d__.m
===================================================================
--- trunk/octave-forge/main/control/inst/@tf/__c2d__.m 2011-09-06 06:03:10 UTC (rev 8505)
+++ trunk/octave-forge/main/control/inst/@tf/__c2d__.m 2011-09-06 11:36:10 UTC (rev 8506)
@@ -1,4 +1,4 @@
-## Copyright (C) 2009 Lukas F. Reichlin
+## Copyright (C) 2009, 2011 Lukas F. Reichlin
##
## This file is part of LTI Syncope.
##
@@ -20,19 +20,23 @@
## Author: Lukas Reichlin <luk...@gm...>
## Created: October 2009
-## Version: 0.1
+## Version: 0.2
-function sys = __c2d__ (sys, tsam, method = "zoh")
+function sys = __c2d__ (sys, tsam, method = "zoh", w0 = 0)
[p, m] = size (sys);
- ##switch (method)
- ## case {"zoh", "std"}
- error ("tf: c2d: not implemented yet");
+ for i = 1 : p
+ for j = 1 : m
+ idx = substruct ("()", {i, j});
+ tmp = subsref (sys, idx);
+ tmp = c2d (ss (tmp), tsam, method, w0);
+ [num, den] = tfdata (tmp, "tfpoly");
+ sys.num(i, j) = num;
+ sys.den(i, j) = den;
+ endfor
+ endfor
- ## otherwise
- ## error ("tf: c2d: %s is an invalid method", method);
+ sys.tfvar = "z";
- ##endswitch
-
-endfunction
\ No newline at end of file
+endfunction
Modified: trunk/octave-forge/main/control/inst/@tf/__d2c__.m
===================================================================
--- trunk/octave-forge/main/control/inst/@tf/__d2c__.m 2011-09-06 06:03:10 UTC (rev 8505)
+++ trunk/octave-forge/main/control/inst/@tf/__d2c__.m 2011-09-06 11:36:10 UTC (rev 8506)
@@ -22,8 +22,21 @@
## Created: September 2011
## Version: 0.1
-function sys = __d2c__ (sys, tsam, method = "zoh")
+function sys = __d2c__ (sys, tsam, method = "zoh", w0 = 0)
- error ("tf: d2c: not implemented yet");
+ [p, m] = size (sys);
-endfunction
\ No newline at end of file
+ for i = 1 : p
+ for j = 1 : m
+ idx = substruct ("()", {i, j});
+ tmp = subsref (sys, idx);
+ tmp = d2c (ss (tmp), method, w0);
+ [num, den] = tfdata (tmp, "tfpoly");
+ sys.num(i, j) = num;
+ sys.den(i, j) = den;
+ endfor
+ endfor
+
+ sys.tfvar = "s";
+
+endfunction
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-09-07 17:45:19
|
Revision: 8508
http://octave.svn.sourceforge.net/octave/?rev=8508&view=rev
Author: paramaniac
Date: 2011-09-07 17:45:09 +0000 (Wed, 07 Sep 2011)
Log Message:
-----------
control: prepare release of control-2.1.55
Modified Paths:
--------------
trunk/octave-forge/main/control/DESCRIPTION
trunk/octave-forge/main/control/devel/PROJECTS
trunk/octave-forge/main/control/devel/RELEASE_PACKAGE
trunk/octave-forge/main/control/devel/pdfdoc/control.tex
trunk/octave-forge/main/control/devel/pdfdoc/functions.texi
trunk/octave-forge/main/control/doc/NEWS
trunk/octave-forge/main/control/doc/control.pdf
Modified: trunk/octave-forge/main/control/DESCRIPTION
===================================================================
--- trunk/octave-forge/main/control/DESCRIPTION 2011-09-06 11:39:16 UTC (rev 8507)
+++ trunk/octave-forge/main/control/DESCRIPTION 2011-09-07 17:45:09 UTC (rev 8508)
@@ -1,6 +1,6 @@
Name: Control
-Version: 2.1.54
-Date: 2011-08-22
+Version: 2.1.55
+Date: 2011-09-07
Author: Lukas Reichlin <luk...@gm...>
Maintainer: Lukas Reichlin <luk...@gm...>
Title: Control Systems
Modified: trunk/octave-forge/main/control/devel/PROJECTS
===================================================================
--- trunk/octave-forge/main/control/devel/PROJECTS 2011-09-06 11:39:16 UTC (rev 8507)
+++ trunk/octave-forge/main/control/devel/PROJECTS 2011-09-07 17:45:09 UTC (rev 8508)
@@ -67,10 +67,6 @@
* Use SLICOT (TD04AD?) for tf2ss conversion (@tf/__sys2ss__.m)
- * Implement c2d conversion (@tf/__c2d__.m)
-
- * Implement d2c conversion (@tf/__d2c__.m)
-
------
Always:
------
Modified: trunk/octave-forge/main/control/devel/RELEASE_PACKAGE
===================================================================
--- trunk/octave-forge/main/control/devel/RELEASE_PACKAGE 2011-09-06 11:39:16 UTC (rev 8507)
+++ trunk/octave-forge/main/control/devel/RELEASE_PACKAGE 2011-09-07 17:45:09 UTC (rev 8508)
@@ -5,6 +5,7 @@
* update version number in the commands below
* adapt version number and release date in DESCRIPTION
* update doc/NEWS
+ * update doc/control.pdf
* copy-paste entire block at once to the terminal
* follow instructions on <http://octave.sourceforge.net/developers.html>
@@ -18,12 +19,12 @@
rm -R ~/octave/__TEMP__/control/devel
cd ~/octave/__TEMP__
grep -i version control/DESCRIPTION
-tar czf control-2.1.54.tar.gz control/
-md5 control-2.1.54.tar.gz
-md5 control-2.1.54.tar.gz > md5_control_pkg.txt
-uuencode control-2.1.54.tar.gz < control-2.1.54.tar.gz > control-2.1.54.tar.gz.uue
+tar czf control-2.1.55.tar.gz control/
+md5 control-2.1.55.tar.gz
+md5 control-2.1.55.tar.gz > md5_control_pkg.txt
+uuencode control-2.1.55.tar.gz < control-2.1.55.tar.gz > control-2.1.55.tar.gz.uue
octave -q --eval \
-"pkg install control-2.1.54.tar.gz"
+"pkg install control-2.1.55.tar.gz"
octave -q --eval \
"pkg load generate_html; generate_package_html ('control', 'control-html', 'octave-forge')"
tar czf control-html.tar.gz control-html
@@ -38,7 +39,7 @@
=====================================================================================
rm -R ~/octave/__TEMP__
-rm -R ~/octave/control-2.1.54
+rm -R ~/octave/control-2.1.55
=====================================================================================
Modified: trunk/octave-forge/main/control/devel/pdfdoc/control.tex
===================================================================
--- trunk/octave-forge/main/control/devel/pdfdoc/control.tex 2011-09-06 11:39:16 UTC (rev 8507)
+++ trunk/octave-forge/main/control/devel/pdfdoc/control.tex 2011-09-07 17:45:09 UTC (rev 8508)
@@ -3,7 +3,7 @@
@setfilename control.info
@settitle Octave Control Systems Package
@afourpaper
-@set VERSION 2.1.54
+@set VERSION 2.1.55
@c @afourwide
@c %**end of header
Modified: trunk/octave-forge/main/control/devel/pdfdoc/functions.texi
===================================================================
--- trunk/octave-forge/main/control/devel/pdfdoc/functions.texi 2011-09-06 11:39:16 UTC (rev 8507)
+++ trunk/octave-forge/main/control/devel/pdfdoc/functions.texi 2011-09-07 17:45:09 UTC (rev 8508)
@@ -449,8 +449,64 @@
@deftypefn {Function File} {@var{sys} =} c2d (@var{sys}, @var{tsam})
@deftypefnx {Function File} {@var{sys} =} c2d (@var{sys}, @var{tsam}, @var{method})
- Convert the continuous lti model into its discrete time equivalent.
+ @deftypefnx {Function File} {@var{sys} =} c2d (@var{sys}, @var{tsam}, @var{"prewarp"}, @var{w0})
+ Convert the continuous lti model into its discrete-time equivalent.
+
+ @strong{Inputs}
+ @table @var
+ @item sys
+ Continuous-time LTI model.
+ @item tsam
+ Sampling time in seconds.
+ @item method
+ Optional conversion method. If not specified, default method @var{"zoh"}
+ is taken.
+ @table @var
+ @item "zoh"
+ Zero-order hold or matrix exponential.
+ @item "tustin", "bilin"
+ Bilinear transformation or Tustin approximation.
+ @item "prewarp"
+ Bilinear transformation with pre-warping at frequency @var{w0}.
+ @end table
+ @end table
+
+ @strong{Outputs}
+ @table @var
+ @item sys
+ Discrete-time LTI model.
+ @end table
@end deftypefn
+@subsection @@lti/d2c
+
+ @deftypefn {Function File} {@var{sys} =} d2c (@var{sys})
+ @deftypefnx {Function File} {@var{sys} =} d2c (@var{sys}, @var{method})
+ @deftypefnx {Function File} {@var{sys} =} d2c (@var{sys}, @var{"prewarp"}, @var{w0})
+ Convert the discrete lti model into its continuous-time equivalent.
+
+ @strong{Inputs}
+ @table @var
+ @item sys
+ Discrete-time LTI model.
+ @item method
+ Optional conversion method. If not specified, default method @var{"zoh"}
+ is taken.
+ @table @var
+ @item "zoh"
+ Zero-order hold or matrix logarithm.
+ @item "tustin", "bilin"
+ Bilinear transformation or Tustin approximation.
+ @item "prewarp"
+ Bilinear transformation with pre-warping at frequency @var{w0}.
+ @end table
+ @end table
+
+ @strong{Outputs}
+ @table @var
+ @item sys
+ Continuous-time LTI model.
+ @end table
+ @end deftypefn
@subsection @@lti/prescale
@deftypefn {Function File} {[@var{scaledsys}, @var{info}] =} prescale (@var{sys})
Modified: trunk/octave-forge/main/control/doc/NEWS
===================================================================
--- trunk/octave-forge/main/control/doc/NEWS 2011-09-06 11:39:16 UTC (rev 8507)
+++ trunk/octave-forge/main/control/doc/NEWS 2011-09-07 17:45:09 UTC (rev 8508)
@@ -1,7 +1,7 @@
Summary of important user-visible changes for releases of the control package
===============================================================================
-control-2.1.55 Release Date: 2011-xx-yy Release Manager: Lukas Reichlin
+control-2.1.55 Release Date: 2011-09-07 Release Manager: Lukas Reichlin
===============================================================================
** @lti/c2d
Modified: trunk/octave-forge/main/control/doc/control.pdf
===================================================================
--- trunk/octave-forge/main/control/doc/control.pdf 2011-09-06 11:39:16 UTC (rev 8507)
+++ trunk/octave-forge/main/control/doc/control.pdf 2011-09-07 17:45:09 UTC (rev 8508)
@@ -3,14 +3,13 @@
5 0 obj
<</Length 6 0 R/Filter /FlateDecode>>
stream
-x\x9C\x95S\xC9n1 \xEDY_\xA1\xA3tZ\x94(\x89\xBA:MS]\xE2d\x8A \x82 \xD7q\x8B\xC6-\xBC\xA4@\xFF\xBE\xE4Ll\xE7\x90 \x8A\xC1\x8C\xF2qyO\x9C\x8D
-\x806\xE8\xF3t.\xD6frU\xEDjg\x82\xBD\x90we6\x80}:k;\xED\xC46(\x99\xB3\xED\xEF͘\x8C\xB6F[B\x82j\xFB\xB5\xB9u \x80\xA8\xA6\xD0\xDC/ߡ\xD8
-\xAB\xFB\xE9c\x81\xA2\xDB{\x84D\xA9\xBA\xAD\xE00p\x8D\xF9\x88\xA3\xEA |\x97jē\xD4,5Fr\xF8\xA27\x9F\xBC4\x9A\xCC\xFC\xA5g\xCE{37K9\x95aڎ\xB0\xAD\xD9\xE6\xD6\x88\x95\xFFtf&\xB3v\xBF}\\x9AɍE3y\xAB\x9F\xE9\xE5\x99 \xB3\xD7\xF6\x959\x9F\xD9\xF9?uA\xF5@
-\x9F \x93b\x82B63ҨΙ\xEFH\xF8\xE6Q\x90ְ\x89" B@E\xBA<\xC4T\x84\xC4и\xB8k\xDF\xF7G(\xB9\x9D\xF2\x92\x81\xE9Y\xC6ғ[\xFB.\xAA_\xC0;\xCDgw\xE9S\xCB\xDFI\xC1\x85Ofv?|N\xA0\xC1\xBD\x92w)9\xB9I0\xB9\xFBS%\x9Db\xABղ4"w\xE1\x81\x8Fb\x85\x90#\xBA\xCF\xD6\xC4\xEC> ) \x88\x8At:X\xFBcΝvo\\xB3\xFB\xED\xD3\xC0!\xE9\xE4#N/I\xF7
-I֢D\x95\xAF\x8BU\xB7\x8E\xE4zX\x84\xFB*½\xD7\xE1rl\xA5\xB9G5a!\xA5S\xA1\x94\xAAM\xC4d\xE0A\x85rtoĔ\xCE,R\xC1\xB8X,\xBD\xAF\x9A\x89P\xC9G\xA8\xBB\xEF\xEA\xC3Ъj\x85\x90ks\xDF\xC66Q8FG r\x94 <\x84_\xDC3\xF9O\xE8\xB0g\x8C "\xE3\xEC\xD9\xDC\xFC\xAE\xCB\xC1Bendstream
+x\x9C\x95SKo1 \xE6\x9C_\x91cro\x9C8\x89s\xDDR\xCA" \xDDvġZ\xB6 \xA2 \xDAG\x91\xF8\xF7\xD83\xFB\xE8\xA1 \xD0h&\x96\xFD\xF9\xF1}\xF1ll |
|
From: <par...@us...> - 2011-09-18 16:40:24
|
Revision: 8556
http://octave.svn.sourceforge.net/octave/?rev=8556&view=rev
Author: paramaniac
Date: 2011-09-18 16:40:17 +0000 (Sun, 18 Sep 2011)
Log Message:
-----------
control: work on tf2ss conversion
Modified Paths:
--------------
trunk/octave-forge/main/control/doc/NEWS
trunk/octave-forge/main/control/inst/@tf/__pole__.m
trunk/octave-forge/main/control/inst/@tf/__sys2ss__.m
trunk/octave-forge/main/control/inst/@tf/__zero__.m
Modified: trunk/octave-forge/main/control/doc/NEWS
===================================================================
--- trunk/octave-forge/main/control/doc/NEWS 2011-09-18 16:25:50 UTC (rev 8555)
+++ trunk/octave-forge/main/control/doc/NEWS 2011-09-18 16:40:17 UTC (rev 8556)
@@ -1,6 +1,21 @@
Summary of important user-visible changes for releases of the control package
===============================================================================
+control-2.2.0 Release Date: 2011-xx-yy Release Manager: Lukas Reichlin
+===============================================================================
+
+** ss
+ -- Transfer function to state-space conversion uses now SLICOT TD04AD.
+ Conversion of MIMO models is now supported. Usage: ss_sys = ss (tf_sys)
+
+** @lti/pole, @lti/zero
+ -- Computation of poles and zeros of MIMO transfer functions is now possible
+ via conversion to state-space. Please note that the state-space
+ realization of SLICOT TD04AD is a minimal one. Therefore certain poles
+ and zeros might be missing.
+
+
+===============================================================================
control-2.1.55 Release Date: 2011-09-07 Release Manager: Lukas Reichlin
===============================================================================
Modified: trunk/octave-forge/main/control/inst/@tf/__pole__.m
===================================================================
--- trunk/octave-forge/main/control/inst/@tf/__pole__.m 2011-09-18 16:25:50 UTC (rev 8555)
+++ trunk/octave-forge/main/control/inst/@tf/__pole__.m 2011-09-18 16:40:17 UTC (rev 8556)
@@ -1,4 +1,4 @@
-## Copyright (C) 2009 Lukas F. Reichlin
+## Copyright (C) 2009, 2011 Lukas F. Reichlin
##
## This file is part of LTI Syncope.
##
@@ -20,14 +20,15 @@
## Author: Lukas Reichlin <luk...@gm...>
## Created: October 2009
-## Version: 0.1
+## Version: 0.2
function pol = __pole__ (sys)
if (issiso (sys))
pol = roots (sys.den{1});
else
- error ("tf: pole: mimo case not implemented yet");
+ warning ("tf: pole: converting to minimal state-space for poles of mimo tf");
+ pol = pole (ss (sys));
endif
endfunction
\ No newline at end of file
Modified: trunk/octave-forge/main/control/inst/@tf/__sys2ss__.m
===================================================================
--- trunk/octave-forge/main/control/inst/@tf/__sys2ss__.m 2011-09-18 16:25:50 UTC (rev 8555)
+++ trunk/octave-forge/main/control/inst/@tf/__sys2ss__.m 2011-09-18 16:40:17 UTC (rev 8556)
@@ -1,4 +1,4 @@
-## Copyright (C) 2009 Lukas F. Reichlin
+## Copyright (C) 2009, 2011 Lukas F. Reichlin
##
## This file is part of LTI Syncope.
##
@@ -20,27 +20,64 @@
## Author: Lukas Reichlin <luk...@gm...>
## Created: October 2009
-## Version: 0.1
+## Version: 0.2
function [retsys, retlti] = __sys2ss__ (sys)
- if (! issiso (sys))
- error ("tf: tf2ss: MIMO case not implemented yet");
- endif
-
+ [p, m] = size (sys);
[num, den] = tfdata (sys);
+
+ numc = cell (p, m);
+ denc = cell (p, 1);
+
+ ## multiply all denominators in a row and
+ ## update each numerator accordingly
+ for i = 1 : p
+ denc(i) = __conv__ (den{i,:});
+ for j = 1 : m
+ idx = setdiff (1:m, j);
+ numc(i,j) = __conv__ (num{i,j}, den{i,idx});
+ endfor
+ endfor
- num = num{1, 1};
- den = den{1, 1};
-
+ len_numc = cellfun (@length, numc);
+ len_denc = cellfun (@length, denc);
+
## tfpoly ensures that there are no leading zeros
- if (length (num) > length (den))
+ tmp = len_numc > repmat (len_denc, 1, m);
+ if (any (tmp(:)))
error ("tf: tf2ss: system must be proper");
endif
- [a, b, c, d] = __tf2ss__ (num, den);
+ max_len_denc = max (len_denc(:));
+ ucoeff = zeros (p, m, max_len_denc);
+ dcoeff = zeros (p, max_len_denc);
+ index = len_denc-1;
+ for i = 1 : p
+ len = len_denc(i);
+ dcoeff(i, 1:len) = denc{i};
+ for j = 1 : m
+ ucoeff(i, j, len-len_numc(i,j)+1 : len) = numc{i,j};
+ endfor
+ endfor
+
+ [a, b, c, d] = sltd04ad (ucoeff, dcoeff, index, sqrt (eps));
+
retsys = ss (a, b, c, d);
retlti = sys.lti; # preserve lti properties
-endfunction
\ No newline at end of file
+endfunction
+
+
+function vec = __conv__ (vec, varargin)
+
+ if (nargin == 1)
+ return;
+ else
+ for k = 1 : nargin-1
+ vec = conv (vec, varargin{k});
+ endfor
+ endif
+
+endfunction
Modified: trunk/octave-forge/main/control/inst/@tf/__zero__.m
===================================================================
--- trunk/octave-forge/main/control/inst/@tf/__zero__.m 2011-09-18 16:25:50 UTC (rev 8555)
+++ trunk/octave-forge/main/control/inst/@tf/__zero__.m 2011-09-18 16:40:17 UTC (rev 8556)
@@ -1,4 +1,4 @@
-## Copyright (C) 2009 Lukas F. Reichlin
+## Copyright (C) 2009, 2011 Lukas F. Reichlin
##
## This file is part of LTI Syncope.
##
@@ -20,18 +20,18 @@
## Author: Lukas Reichlin <luk...@gm...>
## Created: October 2009
-## Version: 0.1
+## Version: 0.2
function [zer, gain] = __zero__ (sys)
if (issiso (sys))
num = get (sys.num{1});
den = get (sys.den{1});
-
zer = roots (num);
gain = num(1) / den(1);
else
- error ("tf: zero: mimo case not implemented yet");
+ warning ("tf: zero: converting to minimal state-space for zeros of mimo tf");
+ [zer, gain] = zero (ss (sys));
endif
endfunction
\ No newline at end of file
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-09-18 19:42:58
|
Revision: 8560
http://octave.svn.sourceforge.net/octave/?rev=8560&view=rev
Author: paramaniac
Date: 2011-09-18 19:42:52 +0000 (Sun, 18 Sep 2011)
Log Message:
-----------
control: add a test for new tf2ss conversion
Modified Paths:
--------------
trunk/octave-forge/main/control/inst/@tf/__pole__.m
trunk/octave-forge/main/control/inst/@tf/__zero__.m
trunk/octave-forge/main/control/inst/ltimodels.m
Added Paths:
-----------
trunk/octave-forge/main/control/devel/__tf2ss__.m
Removed Paths:
-------------
trunk/octave-forge/main/control/inst/__tf2ss__.m
Copied: trunk/octave-forge/main/control/devel/__tf2ss__.m (from rev 8557, trunk/octave-forge/main/control/inst/__tf2ss__.m)
===================================================================
--- trunk/octave-forge/main/control/devel/__tf2ss__.m (rev 0)
+++ trunk/octave-forge/main/control/devel/__tf2ss__.m 2011-09-18 19:42:52 UTC (rev 8560)
@@ -0,0 +1,134 @@
+## Copyright (C) 1996, 1998, 2000, 2002, 2004, 2005, 2007
+## Auburn University. All rights reserved.
+##
+##
+## This program is free software; you can redistribute it and/or modify it
+## under the terms of the GNU General Public License as published by
+## the Free Software Foundation; either version 3 of the License, or (at
+## your option) any later version.
+##
+## This program is distributed in the hope that it will be useful, but
+## WITHOUT ANY WARRANTY; without even the implied warranty of
+## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+## General Public License for more details.
+##
+## You should have received a copy of the GNU General Public License
+## along with this program; see the file COPYING. If not, see
+## <http://www.gnu.org/licenses/>.
+
+## -*- texinfo -*-
+## @deftypefn {Function File} {[@var{a}, @var{b}, @var{c}, @var{d}] =} tf2ss (@var{num}, @var{den})
+## Conversion from transfer function to state-space.
+## The state space system:
+## @iftex
+## @tex
+## $$ \dot x = Ax + Bu $$
+## $$ y = Cx + Du $$
+## @end tex
+## @end iftex
+## @ifinfo
+## @example
+## .
+## x = Ax + Bu
+## y = Cx + Du
+## @end example
+## @end ifinfo
+## is obtained from a transfer function:
+## @iftex
+## @tex
+## $$ G(s) = { { \rm num }(s) \over { \rm den }(s) } $$
+## @end tex
+## @end iftex
+## @ifinfo
+## @example
+## num(s)
+## G(s)=-------
+## den(s)
+## @end example
+## @end ifinfo
+##
+## The vector @var{den} must contain only one row, whereas the vector
+## @var{num} may contain as many rows as there are outputs @var{y} of
+## the system. The state space system matrices obtained from this function
+## will be in controllable canonical form as described in @cite{Modern Control
+## Theory}, (Brogan, 1991).
+## @end deftypefn
+
+## Author: R. Bruce Tenison <bte...@en...>
+## Created: June 22, 1994
+## mod A S Hodel July, Aug 1995
+
+function [a, b, c, d] = __tf2ss__ (num, den)
+
+ if (nargin != 2)
+ print_usage ();
+ elseif (isempty (num))
+ error ("tf2ss: empty numerator");
+ elseif (isempty (den))
+ error ("tf2ss: empy denominator");
+ elseif (! is_real_vector (num))
+ error ("num(%dx%d) must be a vector", rows (num), columns (num));
+ elseif (! is_real_vector (den))
+ error ("den(%dx%d) must be a vector", rows (den), columns (den));
+ endif
+
+ ## strip leading zeros from num, den
+ nz = find (num != 0);
+ if (isempty (nz))
+ num = 0;
+ else
+ num = num(nz(1):length(num));
+ endif
+ nz = find (den != 0);
+ if (isempty (nz))
+ error ("denominator is 0.");
+ else
+ den = den(nz(1):length(den));
+ endif
+
+ ## force num, den to be row vectors
+ num = reshape (num, 1, []);
+ den = reshape (den, 1, []);
+ nn = length (num);
+ nd = length (den);
+ if (nn > nd)
+ error ("deg(num)=%d > deg(den)= %d", nn, nd);
+ endif
+
+ ## Check sizes
+ if (nd == 1)
+ a = b = c = [];
+ d = num(:,1) / den(1);
+ else
+ ## Pad num so that length(num) = length(den)
+ if (nd-nn > 0)
+ num = [zeros(1,nd-nn), num];
+ endif
+
+ ## Normalize the numerator and denominator vector w.r.t. the leading
+ ## coefficient
+ d1 = den(1);
+ num = num / d1;
+ den = den(2:nd)/d1;
+ sw = nd-1:-1:1;
+
+ ## Form the A matrix
+ if (nd > 2)
+ a = [zeros(nd-2,1), eye(nd-2,nd-2); -den(sw)];
+ else
+ a = -den(sw);
+ endif
+
+ ## Form the B matrix
+ b = zeros (nd-1, 1);
+ b(nd-1,1) = 1;
+
+ ## Form the C matrix
+ c = num(:,2:nd)-num(:,1)*den;
+ c = c(:,sw);
+
+ ## Form the D matrix
+ d = num(:,1);
+ endif
+
+endfunction
Modified: trunk/octave-forge/main/control/inst/@tf/__pole__.m
===================================================================
--- trunk/octave-forge/main/control/inst/@tf/__pole__.m 2011-09-18 19:24:53 UTC (rev 8559)
+++ trunk/octave-forge/main/control/inst/@tf/__pole__.m 2011-09-18 19:42:52 UTC (rev 8560)
@@ -27,8 +27,8 @@
if (issiso (sys))
pol = roots (sys.den{1});
else
- warning ("tf: pole: converting to minimal state-space for poles of mimo tf");
+ warning ("tf: pole: converting to minimal state-space for pole computation of mimo tf");
pol = pole (ss (sys));
endif
-endfunction
\ No newline at end of file
+endfunction
Modified: trunk/octave-forge/main/control/inst/@tf/__zero__.m
===================================================================
--- trunk/octave-forge/main/control/inst/@tf/__zero__.m 2011-09-18 19:24:53 UTC (rev 8559)
+++ trunk/octave-forge/main/control/inst/@tf/__zero__.m 2011-09-18 19:42:52 UTC (rev 8560)
@@ -30,8 +30,8 @@
zer = roots (num);
gain = num(1) / den(1);
else
- warning ("tf: zero: converting to minimal state-space for zeros of mimo tf");
+ warning ("tf: zero: converting to minimal state-space for zero computation of mimo tf");
[zer, gain] = zero (ss (sys));
endif
-endfunction
\ No newline at end of file
+endfunction
Deleted: trunk/octave-forge/main/control/inst/__tf2ss__.m
===================================================================
--- trunk/octave-forge/main/control/inst/__tf2ss__.m 2011-09-18 19:24:53 UTC (rev 8559)
+++ trunk/octave-forge/main/control/inst/__tf2ss__.m 2011-09-18 19:42:52 UTC (rev 8560)
@@ -1,134 +0,0 @@
-## Copyright (C) 1996, 1998, 2000, 2002, 2004, 2005, 2007
-## Auburn University. All rights reserved.
-##
-##
-## This program is free software; you can redistribute it and/or modify it
-## under the terms of the GNU General Public License as published by
-## the Free Software Foundation; either version 3 of the License, or (at
-## your option) any later version.
-##
-## This program is distributed in the hope that it will be useful, but
-## WITHOUT ANY WARRANTY; without even the implied warranty of
-## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
-## General Public License for more details.
-##
-## You should have received a copy of the GNU General Public License
-## along with this program; see the file COPYING. If not, see
-## <http://www.gnu.org/licenses/>.
-
-## -*- texinfo -*-
-## @deftypefn {Function File} {[@var{a}, @var{b}, @var{c}, @var{d}] =} tf2ss (@var{num}, @var{den})
-## Conversion from transfer function to state-space.
-## The state space system:
-## @iftex
-## @tex
-## $$ \dot x = Ax + Bu $$
-## $$ y = Cx + Du $$
-## @end tex
-## @end iftex
-## @ifinfo
-## @example
-## .
-## x = Ax + Bu
-## y = Cx + Du
-## @end example
-## @end ifinfo
-## is obtained from a transfer function:
-## @iftex
-## @tex
-## $$ G(s) = { { \rm num }(s) \over { \rm den }(s) } $$
-## @end tex
-## @end iftex
-## @ifinfo
-## @example
-## num(s)
-## G(s)=-------
-## den(s)
-## @end example
-## @end ifinfo
-##
-## The vector @var{den} must contain only one row, whereas the vector
-## @var{num} may contain as many rows as there are outputs @var{y} of
-## the system. The state space system matrices obtained from this function
-## will be in controllable canonical form as described in @cite{Modern Control
-## Theory}, (Brogan, 1991).
-## @end deftypefn
-
-## Author: R. Bruce Tenison <bte...@en...>
-## Created: June 22, 1994
-## mod A S Hodel July, Aug 1995
-
-function [a, b, c, d] = __tf2ss__ (num, den)
-
- if (nargin != 2)
- print_usage ();
- elseif (isempty (num))
- error ("tf2ss: empty numerator");
- elseif (isempty (den))
- error ("tf2ss: empy denominator");
- elseif (! is_real_vector (num))
- error ("num(%dx%d) must be a vector", rows (num), columns (num));
- elseif (! is_real_vector (den))
- error ("den(%dx%d) must be a vector", rows (den), columns (den));
- endif
-
- ## strip leading zeros from num, den
- nz = find (num != 0);
- if (isempty (nz))
- num = 0;
- else
- num = num(nz(1):length(num));
- endif
- nz = find (den != 0);
- if (isempty (nz))
- error ("denominator is 0.");
- else
- den = den(nz(1):length(den));
- endif
-
- ## force num, den to be row vectors
- num = reshape (num, 1, []);
- den = reshape (den, 1, []);
- nn = length (num);
- nd = length (den);
- if (nn > nd)
- error ("deg(num)=%d > deg(den)= %d", nn, nd);
- endif
-
- ## Check sizes
- if (nd == 1)
- a = b = c = [];
- d = num(:,1) / den(1);
- else
- ## Pad num so that length(num) = length(den)
- if (nd-nn > 0)
- num = [zeros(1,nd-nn), num];
- endif
-
- ## Normalize the numerator and denominator vector w.r.t. the leading
- ## coefficient
- d1 = den(1);
- num = num / d1;
- den = den(2:nd)/d1;
- sw = nd-1:-1:1;
-
- ## Form the A matrix
- if (nd > 2)
- a = [zeros(nd-2,1), eye(nd-2,nd-2); -den(sw)];
- else
- a = -den(sw);
- endif
-
- ## Form the B matrix
- b = zeros (nd-1, 1);
- b(nd-1,1) = 1;
-
- ## Form the C matrix
- c = num(:,2:nd)-num(:,1)*den;
- c = c(:,sw);
-
- ## Form the D matrix
- d = num(:,1);
- endif
-
-endfunction
Modified: trunk/octave-forge/main/control/inst/ltimodels.m
===================================================================
--- trunk/octave-forge/main/control/inst/ltimodels.m 2011-09-18 19:24:53 UTC (rev 8559)
+++ trunk/octave-forge/main/control/inst/ltimodels.m 2011-09-18 19:42:52 UTC (rev 8560)
@@ -1369,3 +1369,44 @@
%!
%!assert (Mo, Me, 1e-4);
+
+## transfer function to state-space conversion
+## test from SLICOT TD04AD
+%!shared Mo, Me
+%! INDEX = [ 3 3 ];
+%!
+%! DCOEFF = [ 1.0 6.0 11.0 6.0
+%! 1.0 6.0 11.0 6.0 ];
+%!
+%! UCOEFF = zeros (2, 2, 4);
+%!
+%! u11 = [ 1.0 6.0 12.0 7.0 ];
+%! u12 = [ 0.0 1.0 4.0 3.0 ];
+%! u21 = [ 0.0 0.0 1.0 1.0 ];
+%! u22 = [ 1.0 8.0 20.0 15.0 ];
+%!
+%! UCOEFF(1,1,:) = u11;
+%! UCOEFF(1,2,:) = u12;
+%! UCOEFF(2,1,:) = u21;
+%! UCOEFF(2,2,:) = u22;
+%!
+%! [Ao, Bo, Co, Do] = sltd04ad (UCOEFF, DCOEFF, INDEX, 0);
+%!
+%! Ae = [ 0.5000 -0.8028 0.9387
+%! 4.4047 -2.3380 2.5076
+%! -5.5541 1.6872 -4.1620 ];
+%!
+%! Be = [ -0.2000 -1.2500
+%! 0.0000 -0.6097
+%! 0.0000 2.2217 ];
+%!
+%! Ce = [ 0.0000 -0.8679 0.2119
+%! 0.0000 0.0000 0.9002 ];
+%!
+%! De = [ 1.0000 0.0000
+%! 0.0000 1.0000 ];
+%!
+%! Mo = [Ao, Bo; Co, Do];
+%! Me = [Ae, Be; Ce, De];
+%!
+%!assert (Mo, Me, 1e-4);
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-09-19 11:40:57
|
Revision: 8565
http://octave.svn.sourceforge.net/octave/?rev=8565&view=rev
Author: paramaniac
Date: 2011-09-19 11:40:51 +0000 (Mon, 19 Sep 2011)
Log Message:
-----------
control: fix the pdf manual
Modified Paths:
--------------
trunk/octave-forge/main/control/devel/pdfdoc/control.tex
trunk/octave-forge/main/control/devel/pdfdoc/functions.texi
trunk/octave-forge/main/control/devel/tf2ss/tf2ss.m
trunk/octave-forge/main/control/inst/@lti/mldivide.m
trunk/octave-forge/main/control/inst/test_control.m
Modified: trunk/octave-forge/main/control/devel/pdfdoc/control.tex
===================================================================
--- trunk/octave-forge/main/control/devel/pdfdoc/control.tex 2011-09-19 10:36:15 UTC (rev 8564)
+++ trunk/octave-forge/main/control/devel/pdfdoc/control.tex 2011-09-19 11:40:51 UTC (rev 8565)
@@ -35,3 +35,4 @@
@include functions.texi
@end
+@bye
Modified: trunk/octave-forge/main/control/devel/pdfdoc/functions.texi
===================================================================
--- trunk/octave-forge/main/control/devel/pdfdoc/functions.texi 2011-09-19 10:36:15 UTC (rev 8564)
+++ trunk/octave-forge/main/control/devel/pdfdoc/functions.texi 2011-09-19 11:40:51 UTC (rev 8565)
@@ -2639,7 +2639,7 @@
@subsection @@lti/mldivide
Matrix left division of LTI objects. If necessary, object conversion
- is done by sys_group in mtimes. Used by Octave for "lti1 lti2".
+ is done by sys_group in mtimes. Used by Octave for "lti1 \ lti2".
@subsection @@lti/mpower
Matrix power of LTI objects. The exponent must be an integer.
@@ -2689,6 +2689,6 @@
@end deftypefn
@subsection test_control
- @deftypefn {Script File} test_control
+ @deftypefn {Script File} {} test_control
Execute all available tests at once.
@end deftypefn
Modified: trunk/octave-forge/main/control/devel/tf2ss/tf2ss.m
===================================================================
--- trunk/octave-forge/main/control/devel/tf2ss/tf2ss.m 2011-09-19 10:36:15 UTC (rev 8564)
+++ trunk/octave-forge/main/control/devel/tf2ss/tf2ss.m 2011-09-19 11:40:51 UTC (rev 8565)
@@ -54,7 +54,7 @@
ucoeff(i, j, len-len_numc(i,j)+1 : len) = numc{i,j};
endfor
endfor
-
+index, prod (index), eps*prod (index)
[a, b, c, d] = sltd04ad (ucoeff, dcoeff, index, tol);
retsys = ss (a, b, c, d);
Modified: trunk/octave-forge/main/control/inst/@lti/mldivide.m
===================================================================
--- trunk/octave-forge/main/control/inst/@lti/mldivide.m 2011-09-19 10:36:15 UTC (rev 8564)
+++ trunk/octave-forge/main/control/inst/@lti/mldivide.m 2011-09-19 11:40:51 UTC (rev 8565)
@@ -17,7 +17,7 @@
## -*- texinfo -*-
## Matrix left division of LTI objects. If necessary, object conversion
-## is done by sys_group in mtimes. Used by Octave for "lti1 \ lti2".
+## is done by sys_group in mtimes. Used by Octave for "lti1 \\ lti2".
## Author: Lukas Reichlin <luk...@gm...>
## Created: October 2009
Modified: trunk/octave-forge/main/control/inst/test_control.m
===================================================================
--- trunk/octave-forge/main/control/inst/test_control.m 2011-09-19 10:36:15 UTC (rev 8564)
+++ trunk/octave-forge/main/control/inst/test_control.m 2011-09-19 11:40:51 UTC (rev 8565)
@@ -16,7 +16,7 @@
## along with LTI Syncope. If not, see <http://www.gnu.org/licenses/>.
## -*- texinfo -*-
-## @deftypefn {Script File} test_control
+## @deftypefn {Script File} {} test_control
## Execute all available tests at once.
## @end deftypefn
@@ -55,4 +55,4 @@
test initial
test ctrb
test obsv
-test issample
\ No newline at end of file
+test issample
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
|
From: <par...@us...> - 2011-09-18 19:18:03
|
Revision: 8557
http://octave.svn.sourceforge.net/octave/?rev=8557&view=rev
Author: paramaniac
Date: 2011-09-18 19:17:54 +0000 (Sun, 18 Sep 2011)
Log Message:
-----------
control: move tf2ss conversion into place
Modified Paths:
--------------
trunk/octave-forge/main/control/devel/makefile_conversions.m
trunk/octave-forge/main/control/devel/tf2ss/makefile_tf2ss.m
trunk/octave-forge/main/control/src/Makefile
Added Paths:
-----------
trunk/octave-forge/main/control/src/TD03AY.f
trunk/octave-forge/main/control/src/TD04AD.f
trunk/octave-forge/main/control/src/sltd04ad.cc
Removed Paths:
-------------
trunk/octave-forge/main/control/devel/tf2ss/AB07MD.f
trunk/octave-forge/main/control/devel/tf2ss/MB01PD.f
trunk/octave-forge/main/control/devel/tf2ss/MB01QD.f
trunk/octave-forge/main/control/devel/tf2ss/MB03OY.f
trunk/octave-forge/main/control/devel/tf2ss/TB01ID.f
trunk/octave-forge/main/control/devel/tf2ss/TB01PD.f
trunk/octave-forge/main/control/devel/tf2ss/TB01UD.f
trunk/octave-forge/main/control/devel/tf2ss/TB01XD.f
trunk/octave-forge/main/control/devel/tf2ss/TD03AY.f
trunk/octave-forge/main/control/devel/tf2ss/TD04AD.f
trunk/octave-forge/main/control/devel/tf2ss/common.cc
trunk/octave-forge/main/control/devel/tf2ss/sltd04ad.cc
Modified: trunk/octave-forge/main/control/devel/makefile_conversions.m
===================================================================
--- trunk/octave-forge/main/control/devel/makefile_conversions.m 2011-09-18 16:40:17 UTC (rev 8556)
+++ trunk/octave-forge/main/control/devel/makefile_conversions.m 2011-09-18 19:17:54 UTC (rev 8557)
@@ -21,5 +21,10 @@
mkoctfile slsb10jd.cc \
SB10JD.f
+## transfer function to state-space
+mkoctfile sltd04ad.cc \
+ TD04AD.f TD03AY.f TB01PD.f TB01XD.f AB07MD.f \
+ TB01UD.f TB01ID.f MB01PD.f MB03OY.f MB01QD.f
+
system ("rm *.o");
cd (homedir);
Deleted: trunk/octave-forge/main/control/devel/tf2ss/AB07MD.f
===================================================================
--- trunk/octave-forge/main/control/devel/tf2ss/AB07MD.f 2011-09-18 16:40:17 UTC (rev 8556)
+++ trunk/octave-forge/main/control/devel/tf2ss/AB07MD.f 2011-09-18 19:17:54 UTC (rev 8557)
@@ -1,224 +0,0 @@
- SUBROUTINE AB07MD( JOBD, N, M, P, A, LDA, B, LDB, C, LDC, D, LDD,
- $ INFO )
-C
-C SLICOT RELEASE 5.0.
-C
-C Copyright (c) 2002-2010 NICONET e.V.
-C
-C This program is free software: you can redistribute it and/or
-C modify it under the terms of the GNU General Public License as
-C published by the Free Software Foundation, either version 2 of
-C the License, or (at your option) any later version.
-C
-C This program is distributed in the hope that it will be useful,
-C but WITHOUT ANY WARRANTY; without even the implied warranty of
-C MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-C GNU General Public License for more details.
-C
-C You should have received a copy of the GNU General Public License
-C along with this program. If not, see
-C <http://www.gnu.org/licenses/>.
-C
-C PURPOSE
-C
-C To find the dual of a given state-space representation.
-C
-C ARGUMENTS
-C
-C Mode Parameters
-C
-C JOBD CHARACTER*1
-C Specifies whether or not a non-zero matrix D appears in
-C the given state space model:
-C = 'D': D is present;
-C = 'Z': D is assumed a zero matrix.
-C
-C Input/Output Parameters
-C
-C N (input) INTEGER
-C The order of the state-space representation. N >= 0.
-C
-C M (input) INTEGER
-C The number of system inputs. M >= 0.
-C
-C P (input) INTEGER
-C The number of system outputs. P >= 0.
-C
-C A (input/output) DOUBLE PRECISION array, dimension (LDA,N)
-C On entry, the leading N-by-N part of this array must
-C contain the original state dynamics matrix A.
-C On exit, the leading N-by-N part of this array contains
-C the dual state dynamics matrix A'.
-C
-C LDA INTEGER
-C The leading dimension of array A. LDA >= MAX(1,N).
-C
-C B (input/output) DOUBLE PRECISION array, dimension
-C (LDB,MAX(M,P))
-C On entry, the leading N-by-M part of this array must
-C contain the original input/state matrix B.
-C On exit, the leading N-by-P part of this array contains
-C the dual input/state matrix C'.
-C
-C LDB INTEGER
-C The leading dimension of array B. LDB >= MAX(1,N).
-C
-C C (input/output) DOUBLE PRECISION array, dimension (LDC,N)
-C On entry, the leading P-by-N part of this array must
-C contain the original state/output matrix C.
-C On exit, the leading M-by-N part of this array contains
-C the dual state/output matrix B'.
-C
-C LDC INTEGER
-C The leading dimension of array C.
-C LDC >= MAX(1,M,P) if N > 0.
-C LDC >= 1 if N = 0.
-C
-C D (input/output) DOUBLE PRECISION array, dimension
-C (LDD,MAX(M,P))
-C On entry, if JOBD = 'D', the leading P-by-M part of this
-C array must contain the original direct transmission
-C matrix D.
-C On exit, if JOBD = 'D', the leading M-by-P part of this
-C array contains the dual direct transmission matrix D'.
-C The array D is not referenced if JOBD = 'Z'.
-C
-C LDD INTEGER
-C The leading dimension of array D.
-C LDD >= MAX(1,M,P) if JOBD = 'D'.
-C LDD >= 1 if JOBD = 'Z'.
-C
-C Error Indicator
-C
-C INFO INTEGER
-C = 0: successful exit;
-C < 0: if INFO = -i, the i-th argument had an illegal
-C value.
-C
-C METHOD
-C
-C If the given state-space representation is the M-input/P-output
-C (A,B,C,D), its dual is simply the P-input/M-output (A',C',B',D').
-C
-C REFERENCES
-C
-C None
-C
-C NUMERICAL ASPECTS
-C
-C None
-C
-C CONTRIBUTOR
-C
-C Release 3.0: V. Sima, Katholieke Univ. Leuven, Belgium, Dec. 1996.
-C Supersedes Release 2.0 routine AB07AD by T.W.C.Williams, Kingston
-C Polytechnic, United Kingdom, March 1982.
-C
-C REVISIONS
-C
-C V. Sima, Research Institute for Informatics, Bucharest, Feb. 2004.
-C
-C KEYWORDS
-C
-C Dual system, state-space model, state-space representation.
-C
-C ******************************************************************
-C
-C .. Scalar Arguments ..
- CHARACTER JOBD
- INTEGER INFO, LDA, LDB, LDC, LDD, M, N, P
-C .. Array Arguments ..
- DOUBLE PRECISION A(LDA,*), B(LDB,*), C(LDC,*), D(LDD,*)
-C .. Local Scalars ..
- LOGICAL LJOBD
- INTEGER J, MINMP, MPLIM
-C .. External functions ..
- LOGICAL LSAME
- EXTERNAL LSAME
-C .. External subroutines ..
- EXTERNAL DCOPY, DSWAP, XERBLA
-C .. Intrinsic Functions ..
- INTRINSIC MAX, MIN
-C .. Executable Statements ..
-C
- INFO = 0
- LJOBD = LSAME( JOBD, 'D' )
- MPLIM = MAX( M, P )
- MINMP = MIN( M, P )
-C
-C Test the input scalar arguments.
-C
- IF( .NOT.LJOBD .AND. .NOT.LSAME( JOBD, 'Z' ) ) THEN
- INFO = -1
- ELSE IF( N.LT.0 ) THEN
- INFO = -2
- ELSE IF( M.LT.0 ) THEN
- INFO = -3
- ELSE IF( P.LT.0 ) THEN
- INFO = -4
- ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
- INFO = -6
- ELSE IF( LDB.LT.MAX( 1, N ) ) THEN
- INFO = -8
- ELSE IF( ( N.GT.0 .AND. LDC.LT.MAX( 1, MPLIM ) ) .OR.
- $ ( N.EQ.0 .AND. LDC.LT.1 ) ) THEN
- INFO = -10
- ELSE IF( ( LJOBD .AND. LDD.LT.MAX( 1, MPLIM ) ) .OR.
- $ ( .NOT.LJOBD .AND. LDD.LT.1 ) ) THEN
- INFO = -12
- END IF
-C
- IF ( INFO.NE.0 ) THEN
-C
-C Error return.
-C
- CALL XERBLA( 'AB07MD', -INFO )
- RETURN
- END IF
-C
-C Quick return if possible.
-C
- IF ( MAX( N, MINMP ).EQ.0 )
- $ RETURN
-C
- IF ( N.GT.0 ) THEN
-C
-C Transpose A, if non-scalar.
-C
- DO 10 J = 1, N - 1
- CALL DSWAP( N-J, A(J+1,J), 1, A(J,J+1), LDA )
- 10 CONTINUE
-C
-C Replace B by C' and C by B'.
-C
- DO 20 J = 1, MPLIM
- IF ( J.LE.MINMP ) THEN
- CALL DSWAP( N, B(1,J), 1, C(J,1), LDC )
- ELSE IF ( J.GT.P ) THEN
- CALL DCOPY( N, B(1,J), 1, C(J,1), LDC )
- ELSE
- CALL DCOPY( N, C(J,1), LDC, B(1,J), 1 )
- END IF
- 20 CONTINUE
-C
- END IF
-C
- IF ( LJOBD .AND. MINMP.GT.0 ) THEN
-C
-C Transpose D, if non-scalar.
-C
- DO 30 J = 1, MPLIM
- IF ( J.LT.MINMP ) THEN
- CALL DSWAP( MINMP-J, D(J+1,J), 1, D(J,J+1), LDD )
- ELSE IF ( J.GT.P ) THEN
- CALL DCOPY( P, D(1,J), 1, D(J,1), LDD )
- ELSE IF ( J.GT.M ) THEN
- CALL DCOPY( M, D(J,1), LDD, D(1,J), 1 )
- END IF
- 30 CONTINUE
-C
- END IF
-C
- RETURN
-C *** Last line of AB07MD ***
- END
Deleted: trunk/octave-forge/main/control/devel/tf2ss/MB01PD.f
===================================================================
--- trunk/octave-forge/main/control/devel/tf2ss/MB01PD.f 2011-09-18 16:40:17 UTC (rev 8556)
+++ trunk/octave-forge/main/control/devel/tf2ss/MB01PD.f 2011-09-18 19:17:54 UTC (rev 8557)
@@ -1,271 +0,0 @@
- SUBROUTINE MB01PD( SCUN, TYPE, M, N, KL, KU, ANRM, NBL, NROWS, A,
- $ LDA, INFO )
-C
-C SLICOT RELEASE 5.0.
-C
-C Copyright (c) 2002-2010 NICONET e.V.
-C
-C This program is free software: you can redistribute it and/or
-C modify it under the terms of the GNU General Public License as
-C published by the Free Software Foundation, either version 2 of
-C the License, or (at your option) any later version.
-C
-C This program is distributed in the hope that it will be useful,
-C but WITHOUT ANY WARRANTY; without even the implied warranty of
-C MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-C GNU General Public License for more details.
-C
-C You should have received a copy of the GNU General Public License
-C along with this program. If not, see
-C <http://www.gnu.org/licenses/>.
-C
-C PURPOSE
-C
-C To scale a matrix or undo scaling. Scaling is performed, if
-C necessary, so that the matrix norm will be in a safe range of
-C representable numbers.
-C
-C ARGUMENTS
-C
-C Mode Parameters
-C
-C SCUN CHARACTER*1
-C SCUN indicates the operation to be performed.
-C = 'S': scale the matrix.
-C = 'U': undo scaling of the matrix.
-C
-C TYPE CHARACTER*1
-C TYPE indicates the storage type of the input matrix.
-C = 'G': A is a full matrix.
-C = 'L': A is a (block) lower triangular matrix.
-C = 'U': A is an (block) upper triangular matrix.
-C = 'H': A is an (block) upper Hessenberg matrix.
-C = 'B': A is a symmetric band matrix with lower bandwidth
-C KL and upper bandwidth KU and with the only the
-C lower half stored.
-C = 'Q': A is a symmetric band matrix with lower bandwidth
-C KL and upper bandwidth KU and with the only the
-C upper half stored.
-C = 'Z': A is a band matrix with lower bandwidth KL and
-C upper bandwidth KU.
-C
-C Input/Output Parameters
-C
-C M (input) INTEGER
-C The number of rows of the matrix A. M >= 0.
-C
-C N (input) INTEGER
-C The number of columns of the matrix A. N >= 0.
-C
-C KL (input) INTEGER
-C The lower bandwidth of A. Referenced only if TYPE = 'B',
-C 'Q' or 'Z'.
-C
-C KU (input) INTEGER
-C The upper bandwidth of A. Referenced only if TYPE = 'B',
-C 'Q' or 'Z'.
-C
-C ANRM (input) DOUBLE PRECISION
-C The norm of the initial matrix A. ANRM >= 0.
-C When ANRM = 0 then an immediate return is effected.
-C ANRM should be preserved between the call of the routine
-C with SCUN = 'S' and the corresponding one with SCUN = 'U'.
-C
-C NBL (input) INTEGER
-C The number of diagonal blocks of the matrix A, if it has a
-C block structure. To specify that matrix A has no block
-C structure, set NBL = 0. NBL >= 0.
-C
-C NROWS (input) INTEGER array, dimension max(1,NBL)
-C NROWS(i) contains the number of rows and columns of the
-C i-th diagonal block of matrix A. The sum of the values
-C NROWS(i), for i = 1: NBL, should be equal to min(M,N).
-C The elements of the array NROWS are not referenced if
-C NBL = 0.
-C
-C A (input/output) DOUBLE PRECISION array, dimension (LDA,N)
-C On entry, the leading M by N part of this array must
-C contain the matrix to be scaled/unscaled.
-C On exit, the leading M by N part of A will contain
-C the modified matrix.
-C The storage mode of A is specified by TYPE.
-C
-C LDA (input) INTEGER
-C The leading dimension of the array A. LDA >= max(1,M).
-C
-C Error Indicator
-C
-C INFO (output) INTEGER
-C = 0: successful exit
-C < 0: if INFO = -i, the i-th argument had an illegal
-C value.
-C
-C METHOD
-C
-C Denote by ANRM the norm of the matrix, and by SMLNUM and BIGNUM,
-C two positive numbers near the smallest and largest safely
-C representable numbers, respectively. The matrix is scaled, if
-C needed, such that the norm of the result is in the range
-C [SMLNUM, BIGNUM]. The scaling factor is represented as a ratio
-C of two numbers, one of them being ANRM, and the other one either
-C SMLNUM or BIGNUM, depending on ANRM being less than SMLNUM or
-C larger than BIGNUM, respectively. For undoing the scaling, the
-C norm is again compared with SMLNUM or BIGNUM, and the reciprocal
-C of the previous scaling factor is used.
-C
-C CONTRIBUTOR
-C
-C V. Sima, Katholieke Univ. Leuven, Belgium, Nov. 1996.
-C
-C REVISIONS
-C
-C Oct. 2001, V. Sima, Research Institute for Informatics, Bucharest.
-C
-C ******************************************************************
-C
-C .. Parameters ..
- DOUBLE PRECISION ZERO, ONE
- PARAMETER ( ZERO = 0.0D0, ONE = 1.0D0 )
-C .. Scalar Arguments ..
- CHARACTER SCUN, TYPE
- INTEGER INFO, KL, KU, LDA, M, MN, N, NBL
- DOUBLE PRECISION ANRM
-C .. Array Arguments ..
- INTEGER NROWS ( * )
- DOUBLE PRECISION A( LDA, * )
-C .. Local Scalars ..
- LOGICAL FIRST, LSCALE
- INTEGER I, ISUM, ITYPE
- DOUBLE PRECISION BIGNUM, SMLNUM
-C .. External Functions ..
- LOGICAL LSAME
- DOUBLE PRECISION DLAMCH
- EXTERNAL DLAMCH, LSAME
-C ..
-C .. External Subroutines ..
- EXTERNAL DLABAD, MB01QD, XERBLA
-C .. Intrinsic Functions ..
- INTRINSIC MAX, MIN
-C .. Save statement ..
- SAVE BIGNUM, FIRST, SMLNUM
-C .. Data statements ..
- DATA FIRST/.TRUE./
-C ..
-C .. Executable Statements ..
-C
-C Test the input scalar arguments.
-C
- INFO = 0
- LSCALE = LSAME( SCUN, 'S' )
- IF( LSAME( TYPE, 'G' ) ) THEN
- ITYPE = 0
- ELSE IF( LSAME( TYPE, 'L' ) ) THEN
- ITYPE = 1
- ELSE IF( LSAME( TYPE, 'U' ) ) THEN
- ITYPE = 2
- ELSE IF( LSAME( TYPE, 'H' ) ) THEN
- ITYPE = 3
- ELSE IF( LSAME( TYPE, 'B' ) ) THEN
- ITYPE = 4
- ELSE IF( LSAME( TYPE, 'Q' ) ) THEN
- ITYPE = 5
- ELSE IF( LSAME( TYPE, 'Z' ) ) THEN
- ITYPE = 6
- ELSE
- ITYPE = -1
- END IF
-C
- MN = MIN( M, N )
-C
- ISUM = 0
- IF( NBL.GT.0 ) THEN
- DO 10 I = 1, NBL
- ISUM = ISUM + NROWS(I)
- 10 CONTINUE
- END IF
-C
- IF( .NOT.LSCALE .AND. .NOT.LSAME( SCUN, 'U' ) ) THEN
- INFO = -1
- ELSE IF( ITYPE.EQ.-1 ) THEN
- INFO = -2
- ELSE IF( M.LT.0 ) THEN
- INFO = -3
- ELSE IF( N.LT.0 .OR.
- $ ( ( ITYPE.EQ.4 .OR. ITYPE.EQ.5 ) .AND. N.NE.M ) ) THEN
- INFO = -4
- ELSE IF( ANRM.LT.ZERO ) THEN
- INFO = -7
- ELSE IF( NBL.LT.0 ) THEN
- INFO = -8
- ELSE IF( NBL.GT.0 .AND. ISUM.NE.MN ) THEN
- INFO = -9
- ELSE IF( ITYPE.LE.3 .AND. LDA.LT.MAX( 1, M ) ) THEN
- INFO = -11
- ELSE IF( ITYPE.GE.4 ) THEN
- IF( KL.LT.0 .OR. KL.GT.MAX( M-1, 0 ) ) THEN
- INFO = -5
- ELSE IF( KU.LT.0 .OR. KU.GT.MAX( N-1, 0 ) .OR.
- $ ( ( ITYPE.EQ.4 .OR. ITYPE.EQ.5 ) .AND. KL.NE.KU ) )
- $ THEN
- INFO = -6
- ELSE IF( ( ITYPE.EQ.4 .AND. LDA.LT.KL+1 ) .OR.
- $ ( ITYPE.EQ.5 .AND. LDA.LT.KU+1 ) .OR.
- $ ( ITYPE.EQ.6 .AND. LDA.LT.2*KL+KU+1 ) ) THEN
- INFO = -11
- END IF
- END IF
-C
- IF( INFO.NE.0 ) THEN
- CALL XERBLA( 'MB01PD', -INFO )
- RETURN
- END IF
-C
-C Quick return if possible.
-C
- IF( MN.EQ.0 .OR. ANRM.EQ.ZERO )
- $ RETURN
-C
- IF ( FIRST ) THEN
-C
-C Get machine parameters.
-C
- SMLNUM = DLAMCH( 'S' ) / DLAMCH( 'P' )
- BIGNUM = ONE / SMLNUM
- CALL DLABAD( SMLNUM, BIGNUM )
- FIRST = .FALSE.
- END IF
-C
- IF ( LSCALE ) THEN
-C
-C Scale A, if its norm is outside range [SMLNUM,BIGNUM].
-C
- IF( ANRM.LT.SMLNUM ) THEN
-C
-C Scale matrix norm up to SMLNUM.
-C
- CALL MB01QD( TYPE, M, N, KL, KU, ANRM, SMLNUM, NBL, NROWS,
- $ A, LDA, INFO )
- ELSE IF( ANRM.GT.BIGNUM ) THEN
-C
-C Scale matrix norm down to BIGNUM.
-C
- CALL MB01QD( TYPE, M, N, KL, KU, ANRM, BIGNUM, NBL, NROWS,
- $ A, LDA, INFO )
- END IF
-C
- ELSE
-C
-C Undo scaling.
-C
- IF( ANRM.LT.SMLNUM ) THEN
- CALL MB01QD( TYPE, M, N, KL, KU, SMLNUM, ANRM, NBL, NROWS,
- $ A, LDA, INFO )
- ELSE IF( ANRM.GT.BIGNUM ) THEN
- CALL MB01QD( TYPE, M, N, KL, KU, BIGNUM, ANRM, NBL, NROWS,
- $ A, LDA, INFO )
- END IF
- END IF
-C
- RETURN
-C *** Last line of MB01PD ***
- END
Deleted: trunk/octave-forge/main/control/devel/tf2ss/MB01QD.f
===================================================================
--- trunk/octave-forge/main/control/devel/tf2ss/MB01QD.f 2011-09-18 16:40:17 UTC (rev 8556)
+++ trunk/octave-forge/main/control/devel/tf2ss/MB01QD.f 2011-09-18 19:17:54 UTC (rev 8557)
@@ -1,334 +0,0 @@
- SUBROUTINE MB01QD( TYPE, M, N, KL, KU, CFROM, CTO, NBL, NROWS, A,
- $ LDA, INFO )
-C
-C SLICOT RELEASE 5.0.
-C
-C Copyright (c) 2002-2010 NICONET e.V.
-C
-C This program is free software: you can redistribute it and/or
-C modify it under the terms of the GNU General Public License as
-C published by the Free Software Foundation, either version 2 of
-C the License, or (at your option) any later version.
-C
-C This program is distributed in the hope that it will be useful,
-C but WITHOUT ANY WARRANTY; without even the implied warranty of
-C MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-C GNU General Public License for more details.
-C
-C You should have received a copy of the GNU General Public License
-C along with this program. If not, see
-C <http://www.gnu.org/licenses/>.
-C
-C PURPOSE
-C
-C To multiply the M by N real matrix A by the real scalar CTO/CFROM.
-C This is done without over/underflow as long as the final result
-C CTO*A(I,J)/CFROM does not over/underflow. TYPE specifies that
-C A may be full, (block) upper triangular, (block) lower triangular,
-C (block) upper Hessenberg, or banded.
-C
-C ARGUMENTS
-C
-C Mode Parameters
-C
-C TYPE CHARACTER*1
-C TYPE indices the storage type of the input matrix.
-C = 'G': A is a full matrix.
-C = 'L': A is a (block) lower triangular matrix.
-C = 'U': A is a (block) upper triangular matrix.
-C = 'H': A is a (block) upper Hessenberg matrix.
-C = 'B': A is a symmetric band matrix with lower bandwidth
-C KL and upper bandwidth KU and with the only the
-C lower half stored.
-C = 'Q': A is a symmetric band matrix with lower bandwidth
-C KL and upper bandwidth KU and with the only the
-C upper half stored.
-C = 'Z': A is a band matrix with lower bandwidth KL and
-C upper bandwidth KU.
-C
-C Input/Output Parameters
-C
-C M (input) INTEGER
-C The number of rows of the matrix A. M >= 0.
-C
-C N (input) INTEGER
-C The number of columns of the matrix A. N >= 0.
-C
-C KL (input) INTEGER
-C The lower bandwidth of A. Referenced only if TYPE = 'B',
-C 'Q' or 'Z'.
-C
-C KU (input) INTEGER
-C The upper bandwidth of A. Referenced only if TYPE = 'B',
-C 'Q' or 'Z'.
-C
-C CFROM (input) DOUBLE PRECISION
-C CTO (input) DOUBLE PRECISION
-C The matrix A is multiplied by CTO/CFROM. A(I,J) is
-C computed without over/underflow if the final result
-C CTO*A(I,J)/CFROM can be represented without over/
-C underflow. CFROM must be nonzero.
-C
-C NBL (input) INTEGER
-C The number of diagonal blocks of the matrix A, if it has a
-C block structure. To specify that matrix A has no block
-C structure, set NBL = 0. NBL >= 0.
-C
-C NROWS (input) INTEGER array, dimension max(1,NBL)
-C NROWS(i) contains the number of rows and columns of the
-C i-th diagonal block of matrix A. The sum of the values
-C NROWS(i), for i = 1: NBL, should be equal to min(M,N).
-C The array NROWS is not referenced if NBL = 0.
-C
-C A (input/output) DOUBLE PRECISION array, dimension (LDA,N)
-C The matrix to be multiplied by CTO/CFROM. See TYPE for
-C the storage type.
-C
-C LDA (input) INTEGER
-C The leading dimension of the array A. LDA >= max(1,M).
-C
-C Error Indicator
-C
-C INFO INTEGER
-C Not used in this implementation.
-C
-C METHOD
-C
-C Matrix A is multiplied by the real scalar CTO/CFROM, taking into
-C account the specified storage mode of the matrix.
-C MB01QD is a version of the LAPACK routine DLASCL, modified for
-C dealing with block triangular, or block Hessenberg matrices.
-C For efficiency, no tests of the input scalar parameters are
-C performed.
-C
-C CONTRIBUTOR
-C
-C V. Sima, Katholieke Univ. Leuven, Belgium, Nov. 1996.
-C
-C ******************************************************************
-C
-C .. Parameters ..
- DOUBLE PRECISION ZERO, ONE
- PARAMETER ( ZERO = 0.0D0, ONE = 1.0D0 )
-C ..
-C .. Scalar Arguments ..
- CHARACTER TYPE
- INTEGER INFO, KL, KU, LDA, M, N, NBL
- DOUBLE PRECISION CFROM, CTO
-C ..
-C .. Array Arguments ..
- INTEGER NROWS ( * )
- DOUBLE PRECISION A( LDA, * )
-C ..
-C .. Local Scalars ..
- LOGICAL DONE, NOBLC
- INTEGER I, IFIN, ITYPE, J, JFIN, JINI, K, K1, K2, K3,
- $ K4
- DOUBLE PRECISION BIGNUM, CFROM1, CFROMC, CTO1, CTOC, MUL, SMLNUM
-C ..
-C .. External Functions ..
- LOGICAL LSAME
- DOUBLE PRECISION DLAMCH
- EXTERNAL LSAME, DLAMCH
-C ..
-C .. Intrinsic Functions ..
- INTRINSIC ABS, MAX, MIN
-C ..
-C .. Executable Statements ..
-C
- IF( LSAME( TYPE, 'G' ) ) THEN
- ITYPE = 0
- ELSE IF( LSAME( TYPE, 'L' ) ) THEN
- ITYPE = 1
- ELSE IF( LSAME( TYPE, 'U' ) ) THEN
- ITYPE = 2
- ELSE IF( LSAME( TYPE, 'H' ) ) THEN
- ITYPE = 3
- ELSE IF( LSAME( TYPE, 'B' ) ) THEN
- ITYPE = 4
- ELSE IF( LSAME( TYPE, 'Q' ) ) THEN
- ITYPE = 5
- ELSE
- ITYPE = 6
- END IF
-C
-C Quick return if possible.
-C
- IF( MIN( M, N ).EQ.0 )
- $ RETURN
-C
-C Get machine parameters.
-C
- SMLNUM = DLAMCH( 'S' )
- BIGNUM = ONE / SMLNUM
-C
- CFROMC = CFROM
- CTOC = CTO
-C
- 10 CONTINUE
- CFROM1 = CFROMC*SMLNUM
- CTO1 = CTOC / BIGNUM
- IF( ABS( CFROM1 ).GT.ABS( CTOC ) .AND. CTOC.NE.ZERO ) THEN
- MUL = SMLNUM
- DONE = .FALSE.
- CFROMC = CFROM1
- ELSE IF( ABS( CTO1 ).GT.ABS( CFROMC ) ) THEN
- MUL = BIGNUM
- DONE = .FALSE.
- CTOC = CTO1
- ELSE
- MUL = CTOC / CFROMC
- DONE = .TRUE.
- END IF
-C
- NOBLC = NBL.EQ.0
-C
- IF( ITYPE.EQ.0 ) THEN
-C
-C Full matrix
-C
- DO 30 J = 1, N
- DO 20 I = 1, M
- A( I, J ) = A( I, J )*MUL
- 20 CONTINUE
- 30 CONTINUE
-C
- ELSE IF( ITYPE.EQ.1 ) THEN
-C
- IF ( NOBLC ) THEN
-C
-C Lower triangular matrix
-C
- DO 50 J = 1, N
- DO 40 I = J, M
- A( I, J ) = A( I, J )*MUL
- 40 CONTINUE
- 50 CONTINUE
-C
- ELSE
-C
-C Block lower triangular matrix
-C
- JFIN = 0
- DO 80 K = 1, NBL
- JINI = JFIN + 1
- JFIN = JFIN + NROWS( K )
- DO 70 J = JINI, JFIN
- DO 60 I = JINI, M
- A( I, J ) = A( I, J )*MUL
- 60 CONTINUE
- 70 CONTINUE
- 80 CONTINUE
- END IF
-C
- ELSE IF( ITYPE.EQ.2 ) THEN
-C
- IF ( NOBLC ) THEN
-C
-C Upper triangular matrix
-C
- DO 100 J = 1, N
- DO 90 I = 1, MIN( J, M )
- A( I, J ) = A( I, J )*MUL
- 90 CONTINUE
- 100 CONTINUE
-C
- ELSE
-C
-C Block upper triangular matrix
-C
- JFIN = 0
- DO 130 K = 1, NBL
- JINI = JFIN + 1
- JFIN = JFIN + NROWS( K )
- IF ( K.EQ.NBL ) JFIN = N
- DO 120 J = JINI, JFIN
- DO 110 I = 1, MIN( JFIN, M )
- A( I, J ) = A( I, J )*MUL
- 110 CONTINUE
- 120 CONTINUE
- 130 CONTINUE
- END IF
-C
- ELSE IF( ITYPE.EQ.3 ) THEN
-C
- IF ( NOBLC ) THEN
-C
-C Upper Hessenberg matrix
-C
- DO 150 J = 1, N
- DO 140 I = 1, MIN( J+1, M )
- A( I, J ) = A( I, J )*MUL
- 140 CONTINUE
- 150 CONTINUE
-C
- ELSE
-C
-C Block upper Hessenberg matrix
-C
- JFIN = 0
- DO 180 K = 1, NBL
- JINI = JFIN + 1
- JFIN = JFIN + NROWS( K )
-C
- IF ( K.EQ.NBL ) THEN
- JFIN = N
- IFIN = N
- ELSE
- IFIN = JFIN + NROWS( K+1 )
- END IF
-C
- DO 170 J = JINI, JFIN
- DO 160 I = 1, MIN( IFIN, M )
- A( I, J ) = A( I, J )*MUL
- 160 CONTINUE
- 170 CONTINUE
- 180 CONTINUE
- END IF
-C
- ELSE IF( ITYPE.EQ.4 ) THEN
-C
-C Lower half of a symmetric band matrix
-C
- K3 = KL + 1
- K4 = N + 1
- DO 200 J = 1, N
- DO 190 I = 1, MIN( K3, K4-J )
- A( I, J ) = A( I, J )*MUL
- 190 CONTINUE
- 200 CONTINUE
-C
- ELSE IF( ITYPE.EQ.5 ) THEN
-C
-C Upper half of a symmetric band matrix
-C
- K1 = KU + 2
- K3 = KU + 1
- DO 220 J = 1, N
- DO 210 I = MAX( K1-J, 1 ), K3
- A( I, J ) = A( I, J )*MUL
- 210 CONTINUE
- 220 CONTINUE
-C
- ELSE IF( ITYPE.EQ.6 ) THEN
-C
-C Band matrix
-C
- K1 = KL + KU + 2
- K2 = KL + 1
- K3 = 2*KL + KU + 1
- K4 = KL + KU + 1 + M
- DO 240 J = 1, N
- DO 230 I = MAX( K1-J, K2 ), MIN( K3, K4-J )
- A( I, J ) = A( I, J )*MUL
- 230 CONTINUE
- 240 CONTINUE
-C
- END IF
-C
- IF( .NOT.DONE )
- $ GO TO 10
-C
- RETURN
-C *** Last line of MB01QD ***
- END
Deleted: trunk/octave-forge/main/control/devel/tf2ss/MB03OY.f
===================================================================
--- trunk/octave-forge/main/control/devel/tf2ss/MB03OY.f 2011-09-18 16:40:17 UTC (rev 8556)
+++ trunk/octave-forge/main/control/devel/tf2ss/MB03OY.f 2011-09-18 19:17:54 UTC (rev 8557)
@@ -1,388 +0,0 @@
- SUBROUTINE MB03OY( M, N, A, LDA, RCOND, SVLMAX, RANK, SVAL, JPVT,
- $ TAU, DWORK, INFO )
-C
-C SLICOT RELEASE 5.0.
-C
-C Copyright (c) 2002-2010 NICONET e.V.
-C
-C This program is free software: you can redistribute it and/or
-C modify it under the terms of the GNU General Public License as
-C published by the Free Software Foundation, either version 2 of
-C the License, or (at your option) any later version.
-C
-C This program is distributed in the hope that it will be useful,
-C but WITHOUT ANY WARRANTY; without even the implied warranty of
-C MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-C GNU General Public License for more details.
-C
-C You should have received a copy of the GNU General Public License
-C along with this program. If not, see
-C <http://www.gnu.org/licenses/>.
-C
-C PURPOSE
-C
-C To compute a rank-revealing QR factorization of a real general
-C M-by-N matrix A, which may be rank-deficient, and estimate its
-C effective rank using incremental condition estimation.
-C
-C The routine uses a truncated QR factorization with column pivoting
-C [ R11 R12 ]
-C A * P = Q * R, where R = [ ],
-C [ 0 R22 ]
-C with R11 defined as the largest leading upper triangular submatrix
-C whose estimated condition number is less than 1/RCOND. The order
-C of R11, RANK, is the effective rank of A. Condition estimation is
-C performed during the QR factorization process. Matrix R22 is full
-C (but of small norm), or empty.
-C
-C MB03OY does not perform any scaling of the matrix A.
-C
-C ARGUMENTS
-C
-C Input/Output Parameters
-C
-C M (input) INTEGER
-C The number of rows of the matrix A. M >= 0.
-C
-C N (input) INTEGER
-C The number of columns of the matrix A. N >= 0.
-C
-C A (input/output) DOUBLE PRECISION array, dimension
-C ( LDA, N )
-C On entry, the leading M-by-N part of this array must
-C contain the given matrix A.
-C On exit, the leading RANK-by-RANK upper triangular part
-C of A contains the triangular factor R11, and the elements
-C below the diagonal in the first RANK columns, with the
-C array TAU, represent the orthogonal matrix Q as a product
-C of RANK elementary reflectors.
-C The remaining N-RANK columns contain the result of the
-C QR factorization process used.
-C
-C LDA INTEGER
-C The leading dimension of the array A. LDA >= max(1,M).
-C
-C RCOND (input) DOUBLE PRECISION
-C RCOND is used to determine the effective rank of A, which
-C is defined as the order of the largest leading triangular
-C submatrix R11 in the QR factorization with pivoting of A,
-C whose estimated condition number is less than 1/RCOND.
-C 0 <= RCOND <= 1.
-C NOTE that when SVLMAX > 0, the estimated rank could be
-C less than that defined above (see SVLMAX).
-C
-C SVLMAX (input) DOUBLE PRECISION
-C If A is a submatrix of another matrix B, and the rank
-C decision should be related to that matrix, then SVLMAX
-C should be an estimate of the largest singular value of B
-C (for instance, the Frobenius norm of B). If this is not
-C the case, the input value SVLMAX = 0 should work.
-C SVLMAX >= 0.
-C
-C RANK (output) INTEGER
-C The effective (estimated) rank of A, i.e., the order of
-C the submatrix R11.
-C
-C SVAL (output) DOUBLE PRECISION array, dimension ( 3 )
-C The estimates of some of the singular values of the
-C triangular factor R:
-C SVAL(1): largest singular value of R(1:RANK,1:RANK);
-C SVAL(2): smallest singular value of R(1:RANK,1:RANK);
-C SVAL(3): smallest singular value of R(1:RANK+1,1:RANK+1),
-C if RANK < MIN( M, N ), or of R(1:RANK,1:RANK),
-C otherwise.
-C If the triangular factorization is a rank-revealing one
-C (which will be the case if the leading columns were well-
-C conditioned), then SVAL(1) will also be an estimate for
-C the largest singular value of A, and SVAL(2) and SVAL(3)
-C will be estimates for the RANK-th and (RANK+1)-st singular
-C values of A, respectively.
-C By examining these values, one can confirm that the rank
-C is well defined with respect to the chosen value of RCOND.
-C The ratio SVAL(1)/SVAL(2) is an estimate of the condition
-C number of R(1:RANK,1:RANK).
-C
-C JPVT (output) INTEGER array, dimension ( N )
-C If JPVT(i) = k, then the i-th column of A*P was the k-th
-C column of A.
-C
-C TAU (output) DOUBLE PRECISION array, dimension ( MIN( M, N ) )
-C The leading RANK elements of TAU contain the scalar
-C factors of the elementary reflectors.
-C
-C Workspace
-C
-C DWORK DOUBLE PRECISION array, dimension ( 3*N-1 )
-C
-C Error Indicator
-C
-C INFO INTEGER
-C = 0: successful exit;
-C < 0: if INFO = -i, the i-th argument had an illegal
-C value.
-C
-C METHOD
-C
-C The routine computes a truncated QR factorization with column
-C pivoting of A, A * P = Q * R, with R defined above, and,
-C during this process, finds the largest leading submatrix whose
-C estimated condition number is less than 1/RCOND, taking the
-C possible positive value of SVLMAX into account. This is performed
-C using the LAPACK incremental condition estimation scheme and a
-C slightly modified rank decision test. The factorization process
-C stops when RANK has been determined.
-C
-C The matrix Q is represented as a product of elementary reflectors
-C
-C Q = H(1) H(2) . . . H(k), where k = rank <= min(m,n).
-C
-C Each H(i) has the form
-C
-C H = I - tau * v * v'
-C
-C where tau is a real scalar, and v is a real vector with
-C v(1:i-1) = 0 and v(i) = 1; v(i+1:m) is stored on exit in
-C A(i+1:m,i), and tau in TAU(i).
-C
-C The matrix P is represented in jpvt as follows: If
-C jpvt(j) = i
-C then the jth column of P is the ith canonical unit vector.
-C
-C REFERENCES
-C
-C [1] Bischof, C.H. and P. Tang.
-C Generalizing Incremental Condition Estimation.
-C LAPACK Working Notes 32, Mathematics and Computer Science
-C Division, Argonne National Laboratory, UT, CS-91-132,
-C May 1991.
-C
-C [2] Bischof, C.H. and P. Tang.
-C Robust Incremental Condition Estimation.
-C LAPACK Working Notes 33, Mathematics and Computer Science
-C Division, Argonne National Laboratory, UT, CS-91-133,
-C May 1991.
-C
-C NUMERICAL ASPECTS
-C
-C The algorithm is backward stable.
-C
-C CONTRIBUTOR
-C
-C V. Sima, Katholieke Univ. Leuven, Belgium, Feb. 1998.
-C
-C REVISIONS
-C
-C V. Sima, Research Institute for Informatics, Bucharest, Jan. 2009.
-C V. Sima, Jan. 2010, following Bujanovic and Drmac's suggestion.
-C
-C KEYWORDS
-C
-C Eigenvalue problem, matrix operations, orthogonal transformation,
-C singular values.
-C
-C ******************************************************************
-C
-C .. Parameters ..
- INTEGER IMAX, IMIN
- PARAMETER ( IMAX = 1, IMIN = 2 )
- DOUBLE PRECISION ZERO, ONE
- PARAMETER ( ZERO = 0.0D0, ONE = 1.0D0 )
-C .. Scalar Arguments ..
- INTEGER INFO, LDA, M, N, RANK
- DOUBLE PRECISION RCOND, SVLMAX
-C .. Array Arguments ..
- INTEGER JPVT( * )
- DOUBLE PRECISION A( LDA, * ), DWORK( * ), SVAL( 3 ), TAU( * )
-C ..
-C .. Local Scalars ..
- INTEGER I, ISMAX, ISMIN, ITEMP, J, MN, PVT
- DOUBLE PRECISION AII, C1, C2, S1, S2, SMAX, SMAXPR, SMIN,
- $ SMINPR, TEMP, TEMP2, TOLZ
-C ..
-C .. External Functions ..
- INTEGER IDAMAX
- DOUBLE PRECISION DLAMCH, DNRM2
- EXTERNAL DLAMCH, DNRM2, IDAMAX
-C .. External Subroutines ..
- EXTERNAL DLAIC1, DLARF, DLARFG, DSCAL, DSWAP, XERBLA
-C .. Intrinsic Functions ..
- INTRINSIC ABS, MAX, MIN, SQRT
-C ..
-C .. Executable Statements ..
-C
-C Test the input scalar arguments.
-C
- INFO = 0
- IF( M.LT.0 ) THEN
- INFO = -1
- ELSE IF( N.LT.0 ) THEN
- INFO = -2
- ELSE IF( LDA.LT.MAX( 1, M ) ) THEN
- INFO = -4
- ELSE IF( RCOND.LT.ZERO .OR. RCOND.GT.ONE ) THEN
- INFO = -5
- ELSE IF( SVLMAX.LT.ZERO ) THEN
- INFO = -6
- END IF
-C
- IF( INFO.NE.0 ) THEN
- CALL XERBLA( 'MB03OY', -INFO )
- RETURN
- END IF
-C
-C Quick return if possible.
-C
- MN = MIN( M, N )
- IF( MN.EQ.0 ) THEN
- RANK = 0
- SVAL( 1 ) = ZERO
- SVAL( 2 ) = ZERO
- SVAL( 3 ) = ZERO
- RETURN
- END IF
-C
- TOLZ = SQRT( DLAMCH( 'Epsilon' ) )
- ISMIN = 1
- ISMAX = ISMIN + N
-C
-C Initialize partial column norms and pivoting vector. The first n
-C elements of DWORK store the exact column norms. The already used
-C leading part is then overwritten by the condition estimator.
-C
- DO 10 I = 1, N
- DWORK( I ) = DNRM2( M, A( 1, I ), 1 )
- DWORK( N+I ) = DWORK( I )
- JPVT( I ) = I
- 10 CONTINUE
-C
-C Compute factorization and determine RANK using incremental
-C condition estimation.
-C
- RANK = 0
-C
- 20 CONTINUE
- IF( RANK.LT.MN ) THEN
- I = RANK + 1
-C
-C Determine ith pivot column and swap if necessary.
-C
- PVT = ( I-1 ) + IDAMAX( N-I+1, DWORK( I ), 1 )
-C
- IF( PVT.NE.I ) THEN
- CALL DSWAP( M, A( 1, PVT ), 1, A( 1, I ), 1 )
- ITEMP = JPVT( PVT )
- JPVT( PVT ) = JPVT( I )
- JPVT( I ) = ITEMP
- DWORK( PVT ) = DWORK( I )
- DWORK( N+PVT ) = DWORK( N+I )
- END IF
-C
-C Save A(I,I) and generate elementary reflector H(i).
-C
- IF( I.LT.M ) THEN
- AII = A( I, I )
- CALL DLARFG( M-I+1, A( I, I ), A( I+1, I ), 1, TAU( I ) )
- ELSE
- TAU( M ) = ZERO
- END IF
-C
- IF( RANK.EQ.0 ) THEN
-C
-C Initialize; exit if matrix is zero (RANK = 0).
-C
- SMAX = ABS( A( 1, 1 ) )
- IF ( SMAX.EQ.ZERO ) THEN
- SVAL( 1 ) = ZERO
- SVAL( 2 ) = ZERO
- SVAL( 3 ) = ZERO
- RETURN
- END IF
- SMIN = SMAX
- SMAXPR = SMAX
- SMINPR = SMIN
- C1 = ONE
- C2 = ONE
- ELSE
-C
-C One step of incremental condition estimation.
-C
- CALL DLAIC1( IMIN, RANK, DWORK( ISMIN ), SMIN, A( 1, I ),
- $ A( I, I ), SMINPR, S1, C1 )
- CALL DLAIC1( IMAX, RANK, DWORK( ISMAX ), SMAX, A( 1, I ),
- $ A( I, I ), SMAXPR, S2, C2 )
- END IF
-C
- IF( SVLMAX*RCOND.LE.SMAXPR ) THEN
- IF( SVLMAX*RCOND.LE.SMINPR ) THEN
- IF( SMAXPR*RCOND.LE.SMINPR ) THEN
-C
-C Continue factorization, as rank is at least RANK.
-C
- IF( I.LT.N ) THEN
-C
-C Apply H(i) to A(i:m,i+1:n) from the left.
-C
- AII = A( I, I )
- A( I, I ) = ONE
- CALL DLARF( 'Left', M-I+1, N-I, A( I, I ), 1,
- $ TAU( I ), A( I, I+1 ), LDA,
- $ DWORK( 2*N+1 ) )
- A( I, I ) = AII
- END IF
-C
-C Update partial column norms.
-C
- DO 30 J = I + 1, N
- IF( DWORK( J ).NE.ZERO ) THEN
- TEMP = ABS( A( I, J ) ) / DWORK( J )
- TEMP = MAX( ( ONE + TEMP )*( ONE - TEMP ), ZERO)
- TEMP2 = TEMP*( DWORK( J ) / DWORK( N+J ) )**2
- IF( TEMP2.LE.TOLZ ) THEN
- IF( M-I.GT.0 ) THEN
- DWORK( J ) = DNRM2( M-I, A( I+1, J ), 1 )
- DWORK( N+J ) = DWORK( J )
- ELSE
- DWORK( J ) = ZERO
- DWORK( N+J ) = ZERO
- END IF
- ELSE
- DWORK( J ) = DWORK( J )*SQRT( TEMP )
- END IF
- END IF
- 30 CONTINUE
-C
- DO 40 I = 1, RANK
- DWORK( ISMIN+I-1 ) = S1*DWORK( ISMIN+I-1 )
- DWORK( ISMAX+I-1 ) = S2*DWORK( ISMAX+I-1 )
- 40 CONTINUE
-C
- DWORK( ISMIN+RANK ) = C1
- DWORK( ISMAX+RANK ) = C2
- SMIN = SMINPR
- SMAX = SMAXPR
- RANK = RANK + 1
- GO TO 20
- END IF
- END IF
- END IF
- END IF
-C
-C Restore the changed part of the (RANK+1)-th column and set SVAL.
-C
- IF ( RANK.LT.N ) THEN
- IF ( I.LT.M ) THEN
- CALL DSCAL( M-I, -A( I, I )*TAU( I ), A( I+1, I ), 1 )
- A( I, I ) = AII
- END IF
- END IF
- IF ( RANK.EQ.0 ) THEN
- SMIN = ZERO
- SMINPR = ZERO
- END IF
- SVAL( 1 ) = SMAX
- SVAL( 2 ) = SMIN
- SVAL( 3 ) = SMINPR
-C
- RETURN
-C *** Last line of MB03OY ***
- END
Deleted: trunk/octave-forge/main/control/devel/tf2ss/TB01ID.f
===================================================================
--- trunk/octave-forge/main/control/devel/tf2ss/TB01ID.f 2011-09-18 16:40:17 UTC (rev 8556)
+++ trunk/octave-forge/main/control/devel/tf2ss/TB01ID.f 2011-09-18 19:17:54 UTC (rev 8557)
@@ -1,402 +0,0 @@
- SUBROUTINE TB01ID( JOB, N, M, P, MAXRED, A, LDA, B, LDB, C, LDC,
- $ SCALE, INFO )
-C
-C SLICOT RELEASE 5.0.
-C
-C Copyright (c) 2002-2010 NICONET e.V.
-C
-C This program is free software: you can redistribute it and/or
-C modify it under the terms of the GNU General Public License as
-C published by the Free Software Foundation, either version 2 of
-C the License, or (at your option) any later version.
-C
-C This program is distributed in the hope that it will be useful,
-C but WITHOUT ANY WARRANTY; without even the implied warranty of
-C MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-C GNU General Public License for more details.
-C
-C You should have received a copy of the GNU General Public License
-C along with this program. If not, see
-C <http://www.gnu.org/licenses/>.
-C
-C PURPOSE
-C
-C To reduce the 1-norm of a system matrix
-C
-C S = ( A B )
-C ( C 0 )
-C
-C corresponding to the triple (A,B,C), by balancing. This involves
-C a diagonal similarity transformation inv(D)*A*D applied
-C iteratively to A to make the rows and columns of
-C -1
-C diag(D,I) * S * diag(D,I)
-C
-C as close in norm as possible.
-C
-C The balancing can be performed optionally on the following
-C particular system matrices
-C
-C S = A, S = ( A B ) or S = ( A )
-C ( C )
-C
-C ARGUMENTS
-C
-C Mode Parameters
-C
-C JOB CHARACTER*1
-C Indicates which matrices are involved in balancing, as
-C follows:
-C = 'A': All matrices are involved in balancing;
-C = 'B': B and A matrices are involved in balancing;
-C = 'C': C and A matrices are involved in balancing;
-C = 'N': B and C matrices are not involved in balancing.
-C
-C Input/Output Parameters
-C
-C N (input) INTEGER
-C The order of the matrix A, the number of rows of matrix B
-C and the number of columns of matrix C.
-C N represents the dimension of the state vector. N >= 0.
-C
-C M (input) INTEGER.
-C The number of columns of matrix B.
-C M represents the dimension of input vector. M >= 0.
-C
-C P (input) INTEGER.
-C The number of rows of matrix C.
-C P represents the dimension of output vector. P >= 0.
-C
-C MAXRED (input/output) DOUBLE PRECISION
-C On entry, the maximum allowed reduction in the 1-norm of
-C S (in an iteration) if zero rows or columns are
-C encountered.
-C If MAXRED > 0.0, MAXRED must be larger than one (to enable
-C the norm reduction).
-C If MAXRED <= 0.0, then the value 10.0 for MAXRED is
-C used.
-C On exit, if the 1-norm of the given matrix S is non-zero,
-C the ratio between the 1-norm of the given matrix and the
-C 1-norm of the balanced matrix.
-C
-C A (input/output) DOUBLE PRECISION array, dimension (LDA,N)
-C On entry, the leading N-by-N part of this array must
-C contain the system state matrix A.
-C On exit, the leading N-by-N part of this array contains
-C the balanced matrix inv(D)*A*D.
-C
-C LDA INTEGER
-C The leading dimension of the array A. LDA >= max(1,N).
-C
-C B (input/output) DOUBLE PRECISION array, dimension (LDB,M)
-C On entry, if M > 0, the leading N-by-M part of this array
-C must contain the system input matrix B.
-C On exit, if M > 0, the leading N-by-M part of this array
-C contains the balanced matrix inv(D)*B.
-C The array B is not referenced if M = 0.
-C
-C LDB INTEGER
-C The leading dimension of the array B.
-C LDB >= MAX(1,N) if M > 0.
-C LDB >= 1 if M = 0.
-C
-C C (input/output) DOUBLE PRECISION array, dimension (LDC,N)
-C On entry, if P > 0, the leading P-by-N part of this array
-C must contain the system output matrix C.
-C On exit, if P > 0, the leading P-by-N part of this array
-C contains the balanced matrix C*D.
-C The array C is not referenced if P = 0.
-C
-C LDC INTEGER
-C The leading dimension of the array C. LDC >= MAX(1,P).
-C
-C SCALE (output) DOUBLE PRECISION array, dimension (N)
-C The scaling factors applied to S. If D(j) is the scaling
-C factor applied to row and column j, then SCALE(j) = D(j),
-C for j = 1,...,N.
-C
-C Error Indicator
-C
-C INFO INTEGER
-C = 0: successful exit.
-C < 0: if INFO = -i, the i-th argument had an illegal
-C value.
-C
-C METHOD
-C
-C Balancing consists of applying a diagonal similarity
-C transformation
-C -1
-C diag(D,I) * S * diag(D,I)
-C
-C to make the 1-norms of each row of the first N rows of S and its
-C corresponding column nearly equal.
-C
-C Information about the diagonal matrix D is returned in the vector
-C SCALE.
-C
-C REFERENCES
-C
-C [1] Anderson, E., Bai, Z., Bischof, C., Demmel, J., Dongarra, J.,
-C Du Croz, J., Greenbaum, A., Hammarling, S., McKenney, A.,
-C Ostrouchov, S., and Sorensen, D.
-C LAPACK Users' Guide: Second Edition.
-C SIAM, Philadelphia, 1995.
-C
-C NUMERICAL ASPECTS
-C
-C None.
-C
-C CONTRIBUTOR
-C
-C Release 3.0: V. Sima, Katholieke Univ. Leuven, Belgium, Jan. 1998.
-C This subroutine is based on LAPACK routine DGEBAL, and routine
-C BALABC (A. Varga, German Aerospace Research Establishment, DLR).
-C
-C REVISIONS
-C
-C -
-C
-C KEYWORDS
-C
-C Balancing, eigenvalue, matrix algebra, matrix operations,
-C similarity transformation.
-C
-C *********************************************************************
-C
-C .. Parameters ..
- DOUBLE PRECISION ZERO, ONE
- PARAMETER ( ZERO = 0.0D+0, ONE = 1.0D+0 )
- DOUBLE PRECISION SCLFAC
- PARAMETER ( SCLFAC = 1.0D+1 )
- DOUBLE PRECISION FACTOR, MAXR
- PARAMETER ( FACTOR = 0.95D+0, MAXR = 10.0D+0 )
-C ..
-C .. Scalar Arguments ..
- CHARACTER JOB
- INTEGER INFO, LDA, LDB, LDC, M, N, P
- DOUBLE PRECISION MAXRED
-C ..
-C .. Array Arguments ..
- DOUBLE PRECISION A( LDA, * ), B( LDB, * ), C( LDC, * ),
- $ SCALE( * )
-C ..
-C .. Local Scalars ..
- LOGICAL NOCONV, WITHB, WITHC
- INTEGER I, ICA, IRA, J
- DOUBLE PRECISION CA, CO, F, G, MAXNRM, RA, RO, S, SFMAX1,
- $ SFMAX2, SFMIN1, SFMIN2, SNORM, SRED
-C ..
-C .. External Functions ..
- LOGICAL LSAME
- INTEGER IDAMAX
- DOUBLE PRECISION DASUM, DLAMCH
- EXTERNAL DASUM, DLAMCH, IDAMAX, LSAME
-C ..
-C .. External Subroutines ..
- EXTERNAL DSCAL, XERBLA
-C ..
-C .. Intrinsic Functions ..
- INTRINSIC ABS, MAX, MIN
-C ..
-C .. Executable Statements ..
-C
-C Test the scalar input arguments.
-C
- INFO = 0
- WITHB = LSAME( JOB, 'A' ) .OR. LSAME( JOB, 'B' )
- WITHC = LSAME( JOB, 'A' ) .OR. LSAME( JOB, 'C' )
-C
- IF( .NOT.WITHB .AND. .NOT.WITHC .AND. .NOT.LSAME( JOB, 'N' ) )
- $ THEN
- INFO = -1
- ELSE IF( N.LT.0 ) THEN
- INFO = -2
- ELSE IF( M.LT.0 ) THEN
- INFO = -3
- ELSE IF( P.LT.0 ) THEN
- INFO = -4
- ELSE IF( MAXRED.GT.ZERO .AND. MAXRED.LT.ONE ) THEN
- INFO = -5
- ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
- INFO = -7
- ELSE IF( ( M.GT.0 .AND. LDB.LT.MAX( 1, N ) ) .OR.
- $ ( M.EQ.0 .AND. LDB.LT.1 ) ) THEN
- INFO = -9
- ELSE IF( LDC.LT.MAX( 1, P ) ) THEN
- INFO = -11
- END IF
- IF( INFO.NE.0 ) THEN
- CALL XERBLA( 'TB01ID', -INFO )
- RETURN
- END IF
-C
- IF( N.EQ.0 )
- $ RETURN
-C
-C Compute the 1-norm of the required part of matrix S and exit if
-C it is zero.
-C
- SNORM = ZERO
-C
- DO 10 J = 1, N
- SCALE( J ) = ONE
- CO = DASUM( N, A( 1, J ), 1 )
- IF( WITHC .AND. P.GT.0 )
- $ CO = CO + DASUM( P, C( 1, J ), 1 )
- SNORM = MAX( SNORM, CO )
- 10 CONTINUE
-C
- IF( WITHB ) THEN
-C
- DO 20 J = 1, M
- SNORM = MAX( SNORM, DASUM( N, B( 1, J ), 1 ) )
- 20 CONTINUE
-C
- END IF
-C
- IF( SNORM.EQ.ZERO )
- $ RETURN
-C
-C Set some machine parameters and the maximum reduction in the
-C 1-norm of S if zero rows or columns are encountered.
-C
- SFMIN1 = DLAMCH( 'S' ) / DLAMCH( 'P' )
- SFMAX1 = ONE / SFMIN1
- SFMIN2 = SFMIN1*SCLFAC
- SFMAX2 = ONE / SFMIN2
-C
- SRED = MAXRED
- IF( SRED.LE.ZERO ) SRED = MAXR
-C
- MAXNRM = MAX( SNORM/SRED, SFMIN1 )
-C
-C Balance the matrix.
-C
-C Iterative loop for norm reduction.
-C
- 30 CONTINUE
- NOCONV = .FALSE.
-C
- DO 90 I = 1, N
- CO = ZERO
- RO = ZERO
-C
- DO 40 J = 1, N
- IF( J.EQ.I )
- $ GO TO 40
- CO = CO + ABS( A( J, I ) )
- RO = RO + ABS( A( I, J ) )
- 40 CONTINUE
-C
- ICA = IDAMAX( N, A( 1, I ), 1 )
- CA = ABS( A( ICA, I ) )
- IRA = IDAMAX( N, A( I, 1 ), LDA )
- RA = ABS( A( I, IRA ) )
-C
- IF( WITHC .AND. P.GT.0 ) THEN
- CO = CO + DASUM( P, C( 1, I ), 1 )
- ICA = IDAMAX( P, C( 1, I ), 1 )
- CA = MAX( CA, ABS( C( ICA, I ) ) )
- END IF
-C
- IF( WITHB .AND. M.GT.0 ) THEN
- RO = RO + DASUM( M, B( I, 1 ), LDB )
- IRA = IDAMAX( M, B( I, 1 ), LDB )
- RA = MAX( RA, ABS( B( I, IRA ) ) )
- END IF
-C
-C Special case of zero CO and/or RO.
-C
- IF( CO.EQ.ZERO .AND. RO.EQ.ZERO )
- $ GO TO 90
- IF( CO.EQ.ZERO ) THEN
- IF( RO.LE.MAXNRM )
- $ GO TO 90
- CO = MAXNRM
- END IF
- IF( RO.EQ.ZERO ) THEN
- IF( CO.LE.MAXNRM )
- $ GO TO 90
- RO = MAXNRM
- END IF
-C
-C Guard against zero CO or RO due to underflow.
-C
- G = RO / SCLFAC
- F = ONE
- S = CO + RO
- 50 CONTINUE
- IF( CO.GE.G .OR. MAX( F, CO, CA ).GE.SFMAX2 .OR.
- $ MIN( RO, G, RA ).LE.SFMIN2 )GO TO 60
- F = F*SCLFAC
- CO = CO*SCLFAC
- CA = CA*SCLFAC
- G = G / SCLFAC
- RO = RO / SCLFAC
- RA = RA / SCLFAC
- GO TO 50
-C
- 60 CONTINUE
- G = CO / SCLFAC
- 70 CONTINUE
- IF( G.LT.RO .OR. MAX( RO, RA ).GE.SFMAX2 .OR.
- $ MIN( F, CO, G, CA ).LE.SFMIN2 )GO TO 80
- F = F / SCLFAC
- CO = CO / SCLFAC
- CA = CA / SCLFAC
- G = G / SCLFAC
- RO = RO*SCLFAC
- RA = RA*SCLFAC
- GO TO 70
-C
-C Now balance.
-C
- 80 CONTINUE
- IF( ( CO+RO ).GE.FACTOR*S )
- $ GO TO 90
- IF( F.LT.ONE .AND. SCALE( I ).LT.ONE ) THEN
- IF( F*SCALE( I ).LE.SFMIN1 )
- $ GO TO 90
- END IF
- IF( F.GT.ONE .AND. SCALE( I ).GT.ONE ) THEN
- IF( SCALE( I ).GE.SFMAX1 / F )
- $ GO TO 90
- END IF
- G = ONE / F
- SCALE( I ) = SCALE( I )*F
- NOCONV = .TRUE.
-C
- CALL DSCAL( N, G, A( I, 1 ), LDA )
- CALL DSCAL( N, F, A( 1, I ), 1 )
- IF( M.GT.0 ) CALL DSCAL( M, G, B( I, 1 ), LDB )
- IF( P.GT.0 ) CALL DSCAL( P, F, C( 1, I ), 1 )
-C
- 90 CONTINUE
-C
- IF( NOCONV )
- $ GO TO 30
-C
-C Set the norm reduction parameter.
-C
- MAXRED = SNORM
- SNORM = ZERO
-C
- DO 100 J = 1, N
- CO = DASUM( N, A( 1, J ), 1 )
- IF( WITHC .AND. P.GT.0 )
- $ CO = CO + DASUM( P, C( 1, J ), 1 )
- SNORM = MAX( SNORM, CO )
- 100 CONTINUE
-C
- IF( WITHB ) THEN
-C
- DO 110 J = 1, M
- SNORM = MAX( SNORM, DASUM( N, B( 1, J ), 1 ) )
- 110 CONTINUE
-C
- END IF
- MAXRED = MAXRED/SNORM
- RETURN
-C *** Last line of TB01ID ***
- END
Deleted: trunk/octave-forge/main/control/devel/tf2ss/TB01PD.f
===================================================================
--- trunk/octave-forge/main/control/devel/tf2ss/TB01PD.f 2011-09-18 16:40:17 UTC (rev 8556)
+++ trunk/octave-forge/main/control/devel/tf2ss/TB01PD.f 2011-09-18 19:17:54 UTC (rev 8557)
@@ -1,352 +0,0 @@
- SUBROUTINE TB01PD( JOB, EQUIL, N, M, P, A, LDA, B, LDB, C, LDC,
- $ NR, TOL, IWORK, DWORK, LDWORK, INFO )
-C
-C SLICOT RELEASE 5.0.
-C
-C Copyright (c) 2002-2010 NICONET e.V.
-C
-C This program is free software: you can redistribute it and/or
-C modify it under the terms of the GNU General Public License as
-C published by the Free Software Foundation, either version 2 of
-C the License, or (at your option) any later version.
-C
-C This program is distributed in the hope that it will be useful,
-C but WITHOUT ANY WARRANTY; without even the implied warranty of
-C MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-C GNU General Public License for more details.
-C
-C You should have received a copy of the GNU General Public License
-C along with this program. If not, see
-C <http://www.gnu.org/licenses/>.
-C
-C PURPOSE
-C
-C To find a reduced (controllable, observable, or minimal) state-
-C space representation (Ar,Br,Cr) for any original state-space
-C representation (A,B,C). The matrix Ar is in upper block
-C Hessenberg form.
-C
-C ARGUMENTS
-C
-C Mode Parameters
-C
-C JOB CHARACTER*1
-C Indicates whether the user wishes to remove the
-C uncontrollable and/or unobservable parts as follows:
-C = 'M': Remove both the uncontrollable and unobservable
-C parts to get a minimal state-space representation;
-C = 'C': Remove the uncontrollable part only to get a
-C controllable state-space representation;
-C = 'O': Remove the unobservable part only to get an
-C observable state-space representation.
-C
-C EQUIL CHARACTER*1
-C Specifies whether the user wishes to preliminarily balance
-C the triplet (A,B,C) as follows:
-C = 'S': Perform balancing (scaling);
-C = 'N': Do not perform balancing.
-C
-C Input/Output Parameters
-C
-C N (input) INTEGER
-C The order of the original state-space representation, i.e.
-C the order of the matrix A. N >= 0.
-C
-C M (input) INTEGER
-C The number of system inputs. M >= 0.
-C
-C P (input) INTEGER
-C The number of system outputs. P >= 0.
-C
-C A (input/output) DOUBLE PRECISION array, dimension (LDA,N)
-C On entry, the leading N-by-N part of this array must
-C contain the original state dynamics matrix A.
-C On exit, the leading NR-by-NR part of this array contains
-C the upper block Hessenberg state dynamics matrix Ar of a
-C minimal, controllable, or observable realization for the
-C original system, depending on the value of JOB, JOB = 'M',
-C JOB = 'C', or JOB = 'O', respectively.
-C
-C LDA INTEGER
-C The leading dimension of array A. LDA >= MAX(1,N).
-C
-C B (input/output) DOUBLE PRECISION array, dimension (LDB,M),
-C if JOB = 'C', or (LDB,MAX(M,P)), otherwise.
-C On entry, the leading N-by-M part of this array must
-C contain the original input/state matrix B; if JOB = 'M',
-C or JOB = 'O', the remainder of the leading N-by-MAX(M,P)
-C part is used as internal workspace.
-C On exit, the leading NR-by-M part of this array contains
-C the transformed input/state matrix Br of a minimal,
-C ...
[truncated message content] |
|
From: <par...@us...> - 2011-09-21 19:07:30
|
Revision: 8577
http://octave.svn.sourceforge.net/octave/?rev=8577&view=rev
Author: paramaniac
Date: 2011-09-21 19:07:24 +0000 (Wed, 21 Sep 2011)
Log Message:
-----------
control: add zpkdata
Modified Paths:
--------------
trunk/octave-forge/main/control/INDEX
Added Paths:
-----------
trunk/octave-forge/main/control/inst/@lti/zpkdata.m
Modified: trunk/octave-forge/main/control/INDEX
===================================================================
--- trunk/octave-forge/main/control/INDEX 2011-09-21 18:06:41 UTC (rev 8576)
+++ trunk/octave-forge/main/control/INDEX 2011-09-21 19:07:24 UTC (rev 8577)
@@ -17,6 +17,7 @@
@lti/set
@lti/ssdata
@lti/tfdata
+ @lti/zpkdata
Model Conversions
@lti/c2d
@lti/d2c
Added: trunk/octave-forge/main/control/inst/@lti/zpkdata.m
===================================================================
--- trunk/octave-forge/main/control/inst/@lti/zpkdata.m (rev 0)
+++ trunk/octave-forge/main/control/inst/@lti/zpkdata.m 2011-09-21 19:07:24 UTC (rev 8577)
@@ -0,0 +1,66 @@
+## Copyright (C) 2011 Lukas F. Reichlin
+##
+## This file is part of LTI Syncope.
+##
+## LTI Syncope is free software: you can redistribute it and/or modify
+## it under the terms of the GNU General Public License as published by
+## the Free Software Foundation, either version 3 of the License, or
+## (at your option) any later version.
+##
+## LTI Syncope is distributed in the hope that it will be useful,
+## but WITHOUT ANY WARRANTY; without even the implied warranty of
+## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+## GNU General Public License for more details.
+##
+## You should have received a copy of the GNU General Public License
+## along with LTI Syncope. If not, see <http://www.gnu.org/licenses/>.
+
+## -*- texinfo -*-
+## @deftypefn {Function File} {[@var{z}, @var{p}, @var{k}, @var{tsam}] =} zpkdata (@var{sys})
+## @deftypefnx {Function File} {[@var{z}, @var{p}, @var{k}, @var{tsam}] =} zpkdata (@var{sys}, @var{"v"})
+## Access zero-pole-gain data.
+##
+## @strong{Inputs}
+## @table @var
+## @item sys
+## Any type of LTI model.
+## @item "v", "vector"
+## For SISO models, return @var{z} and @var{p} directly as column vectors
+## instead of cells containing a single column vector.
+## @end table
+##
+## @strong{Outputs}
+## @table @var
+## @item z
+## Cell of column vectors containing the zeros for each channel.
+## z@{i,j@} contains the zeros from input j to output i.
+## @item p
+## Cell of column vectors containing the poles for each channel.
+## p@{i,j@} contains the poles from input j to output i.
+## @item k
+## Matrix containing the gains for each channel.
+## k(i,j) contains the gain from input j to output i.
+## @item tsam
+## Sampling time in seconds. If @var{sys} is a continuous-time model,
+## a zero is returned.
+## @end table
+## @end deftypefn
+
+## Author: Lukas Reichlin <luk...@gm...>
+## Created: September 2011
+## Version: 0.1
+
+function [z, p, k, tsam] = zpkdata (sys, rtype = "cell")
+
+ [num, den, tsam] = tfdata (sys);
+
+ z = cellfun (@roots, num, "uniformoutput", false);
+ p = cellfun (@roots, den, "uniformoutput", false);
+ k = cellfun (@(n,d) n(1)/d(1), num, den);
+
+ if (lower (rtype(1)) == "v" && issiso (sys))
+ z = z{1};
+ p = p{1};
+ endif
+
+endfunction
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
×
Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.
Thanks for helping keep SourceForge clean.
X