SourceForge has been redesigned. Learn more.

Undefined function or variable "v...

  • Kenneth Nielsen

    Kenneth Nielsen - 2010-01-28

    Hi there,

    I am currently trying to get MatCont up and running. I've done the two tutorials that were available, they were a big help in getting started.

    I am working on implementing a 5-dim system, for which I have already done a bifurcation analysis with the aid of some pretty sloppy code. So now I am trying to do the analysis with MatCont (plus I want to do some additional things that I could not code myself).  I am running matlab 7.5.0 and matcont 2.4 on windows xp.

    My problem is this: I do a simulation, that runs till I am certain that the system is in a stable equilibrium (that I have also confirmed from my own code).  Then I choose these values as initial values and set them as
    1) an equilibrium point. When I do the continuation I get the "No convergence at x0" message.
    2) I set them as Limit Cycle (I know that there is an unstable limit cycle surrounding the equilibrium). In this case I get a message that says "Undefined function or variable "v" " .. I have no variable or function called "v" in my system. I have however looked in the limitcycle.m file in the LimitCycle directory, and here I found a function called "v" which matlab complained was never used.

    Any comments will be greatly appreciated, as the work is intended for an upcoming article.



  • Qasim Ali

    Qasim Ali - 2010-02-02

    Dear Kenneth
    For your two problems, i recommend the following:

    (1)     For the problem of "No covergence at x0", try the following:

             (a)     Decrease the initial and minimum step size.
             (b)     Increase the tolerance avalues in 'Continuer' window.
             (C)     Try to perform the continuation wrt to some other parameter, if you have more than one parameters in your system.  
             (d)     Try some other equilibrium point to start the continuation.

    (2)     For the problem of 'undefined function or variable v', try the following:

              (a)    Firts try to find a hopf point on equilibrium curve, the go for limit cycles continuation.
              (b)    If you actually see a limit cycle during orbit calculation, then you can  try the 'Select cycle' button in 'Starter' window, as
                       per the procedure described in Tutorial 2.

    Qasim Ali

  • Kenneth Nielsen

    Kenneth Nielsen - 2010-02-02

    Dear Qasim,

    First of all, thank you for replying.
    In regards to the "No covergence at x0" I have tried to decrease the initial and minimum step sizes to 10^(-8) and 10^(-9) respectively. I have not yet tried to increase the tolerance values - I will try this!

    The continuation works fine with the other parameters (I have 9 in all) save for one, called f2, i.e. I have 2 parameters that act out.
    Also I have tried to do the continuation at another equilibrium. That is to say with another parameter composition. By this I was able to do a continuation in terms of f2 as well, so the only one that yielded the No convergence message, was the one I was in fact interested in!! =(
    The weird thing is that I am able to do all the tutorials and exapmles, but as soon as I try it with my own system I get error messages.

    I hope that increasing the tolerance values will remedy my problem, otherwise I may have to switch to XPP or AUTO, which entails using Linux.

    I will keep you posted, and thanks again!



Log in to post a comment.