Vectorizing the fitness function of a genetic algorithm

Question

Atamert Arslan 2017-4-1

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/333111-vectorizing-the-fitness-function-of-a-genetic-algorithm

编辑： Stephen23 2017-4-9

Dear MATLAB Community,

I am currently trying to solve a binary nonlinear problem through ga, but it just takes too long. I came across the following page (see vectorize for speed): https://www.mathworks.com/help/gads/vectorizing-the-fitness-function.html

I understand the logic, however; I have 6210 dimensions in the decision vector and I was wondering if there was another way to write the function in detail as expressed on that page.

My fitness function currently looks as follows:

A = abs(X-X_Stern);
y = c*A';

where X, X_Stern and c are vectors (1x6210).

Is there a way to vectorize for speed without having to write in open format?

I appreciate your time and answer.

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Carl 2017-4-4

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/333111-vectorizing-the-fitness-function-of-a-genetic-algorithm#answer_261765

编辑：Carl 2017-4-4

Hi Atamert, I believe the way you have it structured now should work fine. If pop is the population size, the input to your fitness function will be popx6210. The output of that calculation is a vector of length pop, which is what's required from a vectorized fitness function.

http://www.mathworks.com/help/gads/ga.html#inputarg_fitnessfcn

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Atamert Arslan 2017-4-8

Thanks for the answer Carl! I had to convert the vectors into matrix depending on the size of the matrix ga each time passes (sometimes it indeed passes only a row vector and not a matrix for the whole population) so I implemented it dynamically and it works now!

请先登录，再进行评论。