How to force maltab to continue with next value of for loop index if it encounters NaN or infinity in the present loop.

I have created an automate script to ease my pains of running my simulink simulation again and again manually. I pass the value to the simulink in each loop and let it run and save the generated graphs in a directory of my own. My problem is that the loop terminates abruptly whenever simulink encounters NaN! So, instead of taking the next loop index, it comes out of the loop prematurely without even completing all the indices over which for loop is running. Kindly help me. This is my very simple automate script:

for i=5:7
    f = figure('visible','off');
    rng(i)
    theta_f_init=.1*rand(2,25);
    rng(i)
    theta_g_init=.1*rand(2,100);
    sim('fuzzy_lafifi')
    plot(error,'linewidth',2);
    title('error');
    set(f, 'Position', get(0, 'Screensize'));
    k=num2str(i);
    string=strcat('~/Desktop/yew/','error_afsmc_',k);
    saveas(f,string,'png');
end

Answer 1

[madhan ravi]

if ~isfinite(yourvariable(i))==1
 continue
else 
  do this
end

[madhan ravi]

the above explains the logic if the variable has nan to inf it skips to the next iteration without abrupting the loop

[Amardeep Mishra]

What does "yourvariable" refer to here? I mean can you modify my code, to explain your point? Thanks in advance!

[madhan ravi]

your variable is the vector which you obtain or have

for example

yourvariable = [1:4 nan inf 4 nan]

[Amardeep Mishra]

Can you kindly modify my code to show your point? I am afraid, I dont quite get it.

[Amardeep Mishra]

The variable which is going NaN is basically a time series, and when I put it as argument to isfinite it throws up this error: Undefined function 'isfinite' for input arguments of type 'timeseries'.

[Amardeep Mishra]

Also the thing is as soon the for loop encounters NaN in sim('fuzzy_lafifi'), it comes out of the for loop. I modified my code like this: But still it isnt running!

for i=11:15
    f = figure('visible','off');
    rng(i)
    theta_f_init=.1*rand(2,25);
    rng(i)
    theta_g_init=.1*rand(2,100);
    sim('fuzzy_lafifi')
    if ~isfinite(theta_f.Data)==1
        continue
    end
    plot(error,'linewidth',2);
    title('error');
    set(f, 'Position', get(0, 'Screensize'));
    k=num2str(i);
    string=strcat('~/Desktop/yew/','error_afsmc_',k);
    saveas(f,string,'png');
end

[madhan ravi]

if ~isfinite(theta_f.Data(i))==1

[madhan ravi]

if it doesn't work upload theta_f.Data

[Amardeep Mishra]

Why do I need to write, theta_f.Data(i), I mean every time, theta_f.Data will be overwritten anyway.

[madhan ravi]

whats the size of it?

[Amardeep Mishra]

the size of theta_f is 2x25x10001, and it is this that is blowing up. My point is, the moment sim() encounters NaN it comes out of the for loop.