Simbiology plot problem when using global solvers

Question

Christopher Sarkizi 2022-4-21

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1701365-simbiology-plot-problem-when-using-global-solvers

编辑： Arthur Goldsipe 2022-4-22

I am trying to model model glycolysis, PPP and TCA using simbiology. While I get some output plots using local solvers, when I do it with global solvers options it simply does not show any results. (the solving procedure seems to be working properly)

Any ideas?

2 个评论
显示无隐藏无

Rick Paxson 2022-4-21

We may need a bit more information to help. What version of MATLAB/SimBiology are you using?

I assume you are running a "Fit program" from within the Analysis app in SimBiology and see the estimation progress until conclusion. Is that the case? Then you don't see plots appear in the analysis app with the results. Can you confirm that my understanding is correct?

Thanks,

Rick Paxson

Christopher Sarkizi 2022-4-21

Yes, Rick that is indeed what I am doing.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Arthur Goldsipe 2022-4-21

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1701365-simbiology-plot-problem-when-using-global-solvers#answer_947915

Can you share the project or provide some other way for us to reproduce what you're seeing? That would help us definitely say what going on

But if I had to guess, the program might be "stuck" simulation. A common problem when using global solvers or performing global sensitivity analysis is trying to simulate with parameter values that drastically slow down the simulation and make it appear like MATLAB or SimBiology is doing nothing. One thing that helps prevent that problem is to set MaximumWallClock to some reasonable value. So if your simulations normally take 1 second, maybe set this to 10 (seconds). You could also try setting narrower bounds on parameters that you're trying to estimate.

2 个评论
显示无隐藏无

Christopher Sarkizi 2022-4-22

Unfortunately I cannot share the project but here are some points that might give you an idea:

1- Simbiology does have this loading/thinking icon after convergence is achieved and the iterations have stopped.

2- The simulation with a global solver takes around 9-10 minutes. (narrowing my range my be one option as you suggested)

3- I did leave overnight (after convergence) to see if there is something that consumes time but with no success.

4- I remember something with SimBiology.web.modelhandler came up when I tried to kill the process (Ctrl+C) in the workspace (but I do not know if this can help you)

Thanks for your time and concern,

Chris

Arthur Goldsipe 2022-4-22

编辑：Arthur Goldsipe 2022-4-22

Ah, the extra information you provided is quite useful. It sounds like you got a plot that showed that your fits converged but then things appeared to "lock up" after that. There are still several things that happen after fit converges. But it's hard for me to say exactly where the problem is or how to fix it without being able to see the problem itself. If the concern is confidentially, could you contact me or Technical Support directly so that you don't have to share the model publicly? Even if you can't privately share the model file, could you screen share while we debug the problem together?

In the interim, I can offer one "hack" that I sometimes use to figure out where things are stuck. If you open the MATLAB editor before you run the program in SimBiology, you get an Editor section in the Toolstrip. Once you run the program, this Toolstrip should have a "Pause" button. When a program appears to be hung you can often click this button and stop in the debugger to figure out where you're stuck. If you can do that, I suggest typing "dbstack" at the command prompt and sharing that info here. That might give me an idea where the program is hanging. If you decide to do that, please also share the output from running "ver" so that I know exactly which version of MATLAB and SimBiology you're using.

请先登录，再进行评论。