Clear Filters
Clear Filters

musyn hangs on DG-step

7 views (last 30 days)
Lucas Koerner
Lucas Koerner on 18 Jan 2024
Answered: UDAYA PEDDIRAJU on 25 Jan 2024
% A single parameter is uncertain and used to construct an uncertation system
RI = ureal('RI', RI);
RI.Percentage = [-2, 2];
% H-infinity Design
ncont = 2; % 2 control signals
nmeas = 4; % 3 measurement signals
[K, CL, gamma, info] = hinfsyn(total_sys, nmeas, ncont);
opt = musynOptions('TargetPerf', 2e19, 'TolPerf', 0, 'Display', 'full', 'MixedMU','on', 'maxiter', 4, 'FitOrder', [3 2], 'MaxFrequency', 5e6);
[K,CLPERF,INFO] = musyn(total_sys, nmeas, ncont, opt);
The hinfsyn command returns a reasonable controller that works well. The gamma value is high which I can't explain but the control signals and plant response are as desired.
However, when I run musyn the K-step completes but MATLAB always hangs at the DG-Step. I've allowed at least 20 minutes with no progress past what is displayed below:
======================= ITERATION 1 ====================
Test bounds: 1.265e+19 <= gamma <= 1.278e+19
gamma X>=0 Y>=0 rho(XY)<1 p/f
1.271e+19 7.5e+04 6.2e-04 2.382e-06 p
Best performance (actual): 1.265e+19
How can I debug this? I've tried various musyn options that have not helped. Should some progress of the DG-STEP be indicated?

Answers (1)

UDAYA PEDDIRAJU on 25 Jan 2024
Hello Lucas,
It seems you're encountering an issue where “musyn” hangs during the DG-step. To address this, you can try the following steps:
  1. Disable “MixedMU”: Since the “MixedMU” option could be related to the problem, try turning it off in your “musynOptions”:
opt = musynOptions('MixedMU', 'off', ...);
2. Increase “maxiter”: The default number of iterations may not suffice for convergence, especially in complex scenarios. Increase the “maxiter” value:
opt = musynOptions('maxiter', 10, ...); % Adjust the number as necessary
3. If you have access to a more powerful computing setup, consider running your code there to see if additional resources prevent the hang-up.
4. Ensure that your system model is as lean as necessary for accurate control design without unnecessary complexity.
5. Keep the “Display” option set to “full” to closely monitor the algorithm's progress and potentially identify the step where it stalls.
6. Ensure you're using the latest version of MATLAB and the Robust Control Toolbox, as updates can contain important fixes and improvements.
You can refer to the following documentation to know more about the “musynOptions”




Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!