Recent posts to Discussionhttps://sourceforge.net/p/matcont/discussion/Recent posts to DiscussionenFri, 26 May 2017 13:23:05 -0000Error using feval Argument must contain a string or function_handlehttps://sourceforge.net/p/matcont/discussion/762215/thread/e80fa019/?limit=25#f715<div class="markdown_content"><p>Hello,</p>
<p>I tried to run some bifurcation analysis with matcont and faced this error right in the begining. I have a system of differential equations. I was able to run the orbit to find equilibrium. However, when I tried to select the equilibrium point and see how it behaves depending on a parameter it gave the error: "Error using feval Argument must contain a string or function_handle".</p>
<p>Tried to google but no disscussion on this.</p>
<p>Any help is highly appreciated,</p></div>Nguyen Linh PhuongFri, 26 May 2017 13:23:05 -0000https://sourceforge.neta94190f7ac33b09dde85a42572f7d3688ece2efbStep size decreasing while computing LC continuationhttps://sourceforge.net/p/matcont/discussion/762215/thread/a1956cb7/?limit=25#fadd<div class="markdown_content"><p>Solved. The options: "MaxStepize" is what I need.</p></div>Zefirov ArtemSat, 08 Apr 2017 09:05:23 -0000https://sourceforge.neta0efd2b80e38c2aecc5bfc0321176536966431a4Step size decreasing while computing LC continuationhttps://sourceforge.net/p/matcont/discussion/762215/thread/a1956cb7/?limit=25#756f<div class="markdown_content"><p>Hello,<br/>
Is there a possibility to decrease a step size of a parameter while continuing LC from Hopf bifucation? Step size set by default is too small for me and thus requires too much time to compute a whole continuation. Couldn't I see it in documentation? </p></div>Zefirov ArtemThu, 06 Apr 2017 09:43:04 -0000https://sourceforge.neta8e57fe63777799105c02116aef3eac80c3a4a8aError using odearguments.https://sourceforge.net/p/matcont/discussion/762215/thread/61047829/?limit=25#8c88<div class="markdown_content"><p>Well, it is difficult to even start helping, but most likely the file containing the equations is set up in a different way MatCont expects. A few things you could provide:<br/>
1. Are you using the GUI or CL version?<br/>
2. What command do you give when this appears?<br/>
3. Have you followed the tutorial(s)?</p></div>hilmeijerSun, 02 Apr 2017 10:36:10 -0000https://sourceforge.neta2356acd3b912058154de23c084cc9958c7df0e8Error using odearguments.https://sourceforge.net/p/matcont/discussion/762215/thread/61047829/?limit=25#dbad<div class="markdown_content"><p>Hello there,</p>
<p>My ultimate goal is to do a bifurcation analysis of my system for which I started to use Matcont. It gave me various different error. So I stepped back and tried to simply solve the system to get a temporal plot just to see if it works. Although I am aware that i can solve a system using Matlab as well and I have actually done that. It works fine in Matlab. But when I tried it with Matcont, it gave me this error:<br/>
" Error using odearguments(line 90) FUN_EVAL must return a column vector."<br/>
The system that I am working on is:</p>
<p>R' = a - q<em>R -(u</em>q<em>V/(alpha +V))</em>R <br/>
T' = q<em>R - d</em>T - k<em>T</em>V +(u<em>q</em>V/(alpha +V))<em>R + b</em>T<em>V/(alpha +V)<br/>
I1' = k</em>V<em>T - e</em>I1 -m<em>I1+n</em>J <br/>
V' = e<em>p</em>I1 - c1<em>V -k</em>T<em>V <br/>
J' = m</em>I1 -n*J </p>
<p>where R,T,I1,V,J are coordinates and a,q,u,alpha,b,k,e,m,n,c1,p,d are parameters and their values are:</p>
<p>a = 20;<br/>
q = .0001; <br/>
u = 200;<br/>
alpha = 1000; <br/>
d = 1/87; <br/>
b = 0.01;<br/>
k= 0.0065;<br/>
e = 0.25;<br/>
m = 0.000001; <br/>
n = 1; % <br/>
p = 20000;<br/>
c= 100; </p>
<p>Initial data is<br/>
R=2000, T=400, I1=0, V=0, J=0.</p>
<p>Please suggest where am i going wrong and how can I get away with this? Any help is appreciated.</p>
<p>Thanks!</p></div>vraFri, 31 Mar 2017 15:41:24 -0000https://sourceforge.net28c6b3b31af842f2e5d08c8c9bb17786b4134a63Newbie questionshttps://sourceforge.net/p/matcont/discussion/762214/thread/4450889c/?limit=25#c63f<div class="markdown_content"><p>Hi, </p>
<p>I am new to MatCont and have the following questions: <br/>
-- What is the difference between MatCont and MatContm? Should I be using one versus the other? <br/>
-- I work in finite element analysis, particularly with nonlinear structural analysis. Is is possible for me to directly couple my code with MatCont (or its command line equivalent as a library)? I am interested in identifying limit points and equilibrium branches of my structural model at various load parameters. <br/>
-- What is the largest model that the code can handle? </p>
<p>I would greatly appreciate your insights. </p>
<p>Regards,<br/>
Manav</p></div>Manav BhatiaFri, 17 Mar 2017 05:05:57 -0000https://sourceforge.net04988cec22b35fb581a35121013ed44a6bed2a86hessianp GUI computationhttps://sourceforge.net/p/matcont/discussion/762215/thread/dcbe3724/?limit=25#5d03<div class="markdown_content"><ol>
<li>These are second order partial derivatives. So Hessian seems a very adequate name. Note that internally we split the continuation variable into state x and parameters p, this gives different dimensions and this is how we keep track of it.</li>
<li>This is used in the continuation of bifurcations (limit point and Hopf) as it involves conditions on the Jacobian.</li>
</ol></div>hilmeijerFri, 03 Feb 2017 08:30:53 -0000https://sourceforge.net30026a8c48129255429f9e9b6ac4e59db968574bhessianp GUI computationhttps://sourceforge.net/p/matcont/discussion/762215/thread/dcbe3724/?limit=25#b4ed<div class="markdown_content"><p>Hello,<br/>
I've not found the definition of hessianp (Hessian with respect to parameters) in MATCONT documentation. Following to Hessian definition I expect the hessianp looks like df_i/(dp_j * dp_k), where f is a vector-function and p is vector of parameters and the function f deals with its parameters as its independental variables. But MATCONT computes hessianp as df_i/(dx_j * dp_k), where x is an origin vector of variables. Is it right to call it Hessian? And which algorithms use hessianp? </p></div>Zefirov ArtemThu, 02 Feb 2017 13:39:14 -0000https://sourceforge.net5266c50899767b02df35a229b58007b31168246cTolerance of multipliershttps://sourceforge.net/p/matcont/discussion/762214/thread/f148fc2d/?limit=25#7643<div class="markdown_content"><p>Dear Zefirov,<br/>
The multipliers come from the discretization of the limit cycle. The parameters influencing the accuracy of the approximation are the number of mesh points (ntst) and collocation points (ncol). NCOL=4 is a good choice for various numerical reasons. Increasing NTST from 20 to 40 (or 60 depending on memory) will improve the accuracy.<br/>
Note that strong contraction (e.g. large negative eigenvalues of an equilibrium) will result in very small multipliers. So do not worry about those.<br/>
Best regards, Hil Meijer</p></div>hilmeijerMon, 09 Jan 2017 08:47:17 -0000https://sourceforge.net94f532904863e697026e1dc86672e56df78baa2cTolerance of multipliershttps://sourceforge.net/p/matcont/discussion/762214/thread/f148fc2d/?limit=25#c3a2<div class="markdown_content"><p>Hi all,<br/>
Is there a way to know tolerance of multipliers being computed during limit cycle continuation? I found it close to 1e-5 in my case. Is this value always the same or there is an option defining it? Decreasing such options as FunTolerance,VarTolerance, MaxNewtonIters,MaxCorrSize didn't influence the multipliers' tolerance.</p></div>Zefirov ArtemSat, 07 Jan 2017 21:34:15 -0000https://sourceforge.netca669f395366ba8f11b2523cb817d6b1c024eb9a