Error when I try to run genetic algorithm on peaks function.

Below is my code for applying genetic algorithm to peaks function. my intent is to see what points are added in each iteration.

peaks
problem.solver = 'ga';
problem.fitnessfcn = @(x)peaks(x(1),x(2));
problem.nvars = 2;
problem.lb = [-3 -3];
problem.ub = [3 3];
problem.options = optimoptions('ga', ...
    'OutputFcn', @peaksPlotIterates, ...
    'Display', 'iter', ...
    'PopInitRange', [-3 -3;3 3]);

x = ga(problem);

I tried to replicate what I saw in a youtube video (https://www.youtube.com/watch?v=1i8muvzZkPw) published by MAtlab. However, I get the following error in the console:

Peaks_GA

z =  3*(1-x).^2.*exp(-(x.^2) - (y+1).^2) ... 
    - 10*(x/5 - x.^3 - y.^5).*exp(-x.^2-y.^2) ... 
    - 1/3*exp(-(x+1).^2 - y.^2) 


Single objective optimization:
2 Variables

Options:
CreationFcn:       @gacreationuniform
CrossoverFcn:      @crossoverscattered
SelectionFcn:      @selectionstochunif
MutationFcn:       @mutationadaptfeasible
Unrecognized property 'iteration' for class 'optim.options.GaOptions'.

Error in peaksPlotIterates (line 24)
optimValues.iteration = -1;

Error in gaoutput (line 56)
[state,optnew,changed] = feval(functions{i},options.OutputPlotFcnOptions,state,flag,args{i}{:});

Error in galincon (line 55)
[state,options] = gaoutput(FitnessFcn,options,state,currentState);

Error in ga (line 420)
[x,fval,exitFlag,output,population,scores] = galincon(FitnessFcn,nvars, ...
    
Error in Peaks_GA (line 13)
x = ga(problem);

Answer 1

[Abhinav Aravindan]

I see that you are attempting to apply the genetic algorithm to the peaks function and monitor the progress of points added in each iteration. I assume that you are using the “peakPlotIterates” function mentioned in the following link:

https://www.mathworks.com/matlabcentral/fileexchange/27178

Based on the error message you've encountered, it seems there is an issue with the “peaksPlotIterates” function trying to modify the “optimValues” structure, by adding an “iteration” field.

There are a few points to note here:

• The video mentioned in the question utilizes MATLAB R2015a for the example and in turn uses the “gaoptimset” function which returns the genetic algorithm options as a “struct”.
• In the given implementation above in MATLAB R2024a, you have used the “optimoptions” with “ga” as the solver, which is the alternative provided in the documentation for “gaoptimset”. It is to be noted that for “ga” solver, “optimoptions” returns an object with set of options for “ga”.

In the “peakPlotIterates” function, a new field named “iterations” is added to the Genetic Algorithm options object “optimValues” causing this error while using “optimoptions”. A possible workaround to resolve this error is to modify “peakPlotIterates” function by creating a new local variable “iteration” and replace all “optimValues.iteration” with “iteration” to avoid modifying “optimValues”.

Hope this helps resolve the issue !

Please refer to the below documentation for your reference.

https://www.mathworks.com/help/gads/gaoptimset.html

https://www.mathworks.com/help/optim/ug/optim.problemdef.optimizationproblem.optimoptions.html

https://www.mathworks.com/help/gads/genetic-algorithm-options.html

https://www.mathworks.com/help/gads/options-in-genetic-algorithm.html

[Preetam Kulkarni]

Hi Abhinav,

Thank you for answering my question. Below is the part of the code that needs to be modified. Is there any way I can get access to the generation number from gaoptimset. I tried to look through the documentation but I could'nt find anything regarding generatio number.

function varargout = peaksPlotIterates(varargin)
% Output function that plots the iterates of the optimization algorithm.
%  Copyright (c) 2010, The MathWorks, Inc.
%  All rights reserved.
% Check if caller is from global or optimization toolbox
optimValues = varargin{2};
state = varargin{3};
if nargout > 1
    if isfield(optimValues,'x') % simulated annealing options,optimvalues,flag
        x = optimValues.x;
        varargout{1} = false;
        varargout{2} = x; % options field
        varargout{3} = false;
    else % pattern search optimvalues,options,flag,interval
        optimValues = varargin{1};
        state = varargin{3};
        if isfield(optimValues,'x')
            x = optimValues.x;
            varargout{1} = false;
            varargout{2} = x; % options field
            varargout{3} = false;
        else % gentic algorithm options,state,flag,interval
            x = varargin{2}.Population;
            optimValues.iteration = -1;
            varargout{1} = varargin{2};
            varargout{2} = varargin{1};
            varargout{3} = false;
        end
    end
else
    x = varargin{1};
    varargout{1} = false;
end
% Check for state
switch state
    case 'init'
        % Plot objective function surface
        PlotSurface(x,peaks(x(:,1),x(:,2)));
    case 'iter'
        if ~(optimValues.iteration == 0)
            % Update surface plot to show current solution
            PlotUpdate(x,peaks(x(:,1),x(:,2)));
        end
    case 'done'
        if ~(optimValues.iteration == 0)
            % After optimization, display solution in plot title
            DisplayTitle(x,peaks(x(:,1),x(:,2)))
        end
end

[Abhinav Aravindan]

The use of "gaoptimset" is no longer recommended, you can utilize "optimoptions" as written in your code. The current generation number can be obtained from "varargin{2}" as follows:

currentGen = varargin{2}.Generation

https://www.mathworks.com/help/gads/genetic-algorithm-options.html#:~:text=Generation%20%E2%80%94%20Current%20generation%20number.

The number of generations computed can be found from the fourth output argument of "ga"

[x, fval, exitflag, out] = ga(problem);
disp(out.generations)

https://www.mathworks.com/help/gads/ga.html#mw_cb8d5c20-6d67-4f7c-b1c2-f61fed7772b1:~:text=generations%20%E2%80%94%20Number%20of%20generations%20computed.

[Preetam Kulkarni]

Thank you Abhinav. This worked!