Using optimization variable with wiener2

I have a problem, listed below, where x is the optimization variable. It essentially corresponds to the filter window in wiener2 (where m=n). The function calculateExtractionSharpness has the wiener2 filter called within.

x0 = 10;
xl = 1;
xu = 100;
nvar = 1;
x = optimvar('x',1,'LowerBound',xl,'UpperBound',xu);
prob = optimproblem;
optimFunc = @(filterWind) calculateExtractionSharpness(chart,filterWind,1);
prob.Objective = optimFunc(x);
options = optimoptions('intlinprog',"Display","iter");
sol = solve(prob,x0,'Options',options)

However, I get an error that states that:

Error using ones
Size inputs must be numeric.

Error in wiener2 (line 48)
localMean = filter2(ones(nhood), g) / prod(nhood);

Error in untitled>extractDisplacementCurve (line 40)
        chart = wiener2(chart,[filterWind filterWind]);

Error in untitled>calculateExtractionSharpness (line 86)
curve = extractDisplacementCurve(chart,filterWind,filterOptions);

Error in untitled (line 32)
optimFunc = @(filterWind) calculateExtractionSharpness(chart,filterWind,1);

Is there a way to solve this issue or is optimization not possible with this type of solver?

Answer 1

[Walter Roberson]

That cannot be done with Optimization Variables; you would need to switch to Solver-Based Optimization.

[Aditya Chandramohan]

Ah makes sense now. Thanks a lot!