[Maxima-bugs] [ maxima-Bugs-3081820 ] lbfgs causes error

[SourceForge.net <no...@so...>]

Bugs item #3081820, was opened at 2010-10-05 21:45
Message generated for change (Tracker Item Submitted) made by rtoy
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=104933&aid=3081820&group_id=4933

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: Lisp Core - Floating point
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Raymond Toy (rtoy)
Assigned to: Nobody/Anonymous (nobody)
Summary: lbfgs causes error

Initial Comment:
>From the maxima mailing list:

(%i2) V: 0.75$
a: 24$
b: 68$
C(r) := 2*%pi*b*r2 + 4*a*%pi*r + 2*b*V/r + a*V/(%pi*r2)$
lbfgs(C(r), [r], [1.0], 1e-4, [1,0]);
(%i3) 
(%i4) 
(%i5) 
(%i6) *************************************************
  N=    1   NUMBER OF CORRECTIONS=25
       INITIAL VALUES
 F=  8.365790735841403D+02   GNORM=  1.042646940618427D+03
*************************************************

   I  NFN     FUNC                    GNORM                   STEPLENGTH

Maxima encountered a Lisp error:

 Error in NTH [or a callee]: T is not of type SYSTEM::PROPER-LIST.

This is because the computation of the gradient returns T to indicate a failure, and indeed this is true because lbfgs tries to find the gradient at 0.

The proposed solution is to catch this case and print an error message (instead of returning a noun form).



----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=104933&aid=3081820&group_id=4933