alpsayin / genetic-algorithm-matlab Goto Github PK

I replaced the target function in EvaluateIndividual.m file：
x1=x(1); x2=x(2); fitnessValue =- (-40*x1*cos(x2)-35*x2*sin(x1)+15*x1^2-sin(x1)*20+12*x2^2-25*x1*x2*sin(x1*x2^2*cos(x1+x2)));

at the same time， I changed the genetic algorithm Settings parameters in FunctionOptimization.m file：
`
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%
% Alp Sayin - alp_sayin[at]ieee[dot]org
% Matlab Genetic Algorithm
% Spring 2012
%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

%% CLEAN-UP
clear;close all;clc;
tic

%% PARAMETERS
populationSize = 1000 ;
numberOfGenes = 40;
crossoverProbability = 0.8 ;
mutationProbability = 0.0625;
tournamentSelectionParameter = 0.5;
variableRange = 50.0;
numberOfGenerations = 5000;
numberOfVariables = 2;
tournamentSize = 10;
numberOfReplications = 2;
verbose = true;
draw_plots = true;
% UNLESS THE FITNESS FUNCTION IS REALLY DIFFICULT TO COMPUTE, IT'S FASTER
% NOT TO USE PARALLEL COMPUTATION
runparallel = false;
`
The maximum value should be 735.484897639184，The corresponding two parameter values should be ：
x1: -20.3956600843891
x2: -22.0150462042061

but this matlab code can't find the best solution：

Maximum Fitness: 7.286375e+02
Best Solution: 17.1875
Best Solution: -18.843
时间已过 120.739511 秒。

Is there a way you can improve the search capability of your genetic algorithm ？

How do we perform multivariable optimization?