Creating an Partial Initial Population of Matrices for a Genetic Algorithm

Hello,

I'm running a genetic algortithm that seachers for the minumum value created by a 7x7 matrix of variables.

I want to create an partial initial population where each "point" is actually a 7x7 matrix of values. I want a total population of 100 but I want the partial initial population to consist of 26 matrices each 7x7. From the documentation it seems like each point needs to be a double value, rather than a matrix. Is there a way to trick the GA into accepting these matrices?

Many thanks in advanced!

Answer 1

[Alan Weiss]

ga accepts only row vectors as population members. However, it is easy for you to call reshape in your fitness function to turn a row vector into a matrix. Your population members should be of size 1-by-49. Your initial population matrix should be of size 26-by-49. OK?

Alan Weiss

MATLAB mathematical toolbox documentation

Answer 2

[Mahesh Taparia]

Hi

Double is the variable datatype. If A is the matrix, you can check the variable class of the matrix by

whos A

If it is not in double, you can convert the matrix A into double as

A=double(A);