Find values for best and mean fitness value for each generation in Genetic Algorithm

After running GA, using Optimization toolbox and code below, the following mean and best fitness value graph was obtained.

[x,fval,exitflag,output,population,score] = ...

ga(@revisepartialmodespring,nvars,[],[],[],[],lb,ub,[],[],options);

How to get the value of best fitness at each generation ?.. I wanted to show a comparison of the best fitness v/s generation for different problems on the same graph. Hence I need values to best fitness at each generation.

[Debaditya Gupta]

How to get similar type of graph in Particle Swarm Optimization in matlab?

Matlab PSO script is showing only iteration vs best fitnes value, but mean fitness value is not showing

Answer 1

[Alan Weiss]

Use a Custom Output Function for Genetic Algorithm. Model it on the gaplotbestf plot function. To see that function,

edit gaplotbestf

Alan Weiss

MATLAB mathematical toolbox documentation

[Ankur Shah]

Thanks for reply. Could you help me with code. I have tried but am unable to save Generations v/s best fitness data.

[Alan Weiss]

options = optimoptions("ga","OutputFcn",@gascoreoutfun);
rng default % For reproducibility
x = ga(@ps_example,2,[],[],[],[],[],[],[],[],options)

Optimization terminated: average change in the fitness value less than options.FunctionTolerance.

x = 1×2

-4.6793 -0.0860

size(gapopulationhistory)

ans = 1×2

122 2

function [state,options,optchanged] = gascoreoutfun(options,state,flag)
persistent history
optchanged = false;
switch flag
    case 'init'
        history = [min(state.Score),mean(state.Score)];
        assignin('base','gapopulationhistory',history);
    case 'iter'
        history = [history;min(state.Score),mean(state.Score)];
        assignin('base','gapopulationhistory',history);
    case 'done'
end
end

Alan Weiss

MATLAB mathematical toolbox documentation