how to use genetic algorithm (GA) in matlab ?

Question

jinwoo lee 2021-2-15

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/745867-how-to-use-genetic-algorithm-ga-in-matlab

评论： Walter Roberson 2021-2-15

采纳的回答： Walter Roberson

在 MATLAB Online 中打开

Hi, all

I found GA function https://kr.mathworks.com/help/gads/options-in-genetic-algorithm.html.

% input_data shape = [ M x N x T ] ( channel, trials, time sample )
% ga : Find minimum of function using genetic algorithm
ga(fitness_function_my_own, input_data)

I want to find a few channels that minimize the fitness function for M channels, and the question is two.

Q1. I wonder if ga can be applied to the matrix data ?

Q2. I know that parameters that need to be set, such as crossover probability and mutation probability, are theoretically necessary to apply ga.

Is it necessary to modify the ga function directly in the file implemented to apply this option to the corresponding function?

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

Walter Roberson 2021-2-15

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/745867-how-to-use-genetic-algorithm-ga-in-matlab#answer_623907

在 MATLAB Online 中打开

Q1. I wonder if ga can be applied to the matrix data ?

Yes, but fitness_function_my_own must return a scalar.

If you want fitness_function_my_own to return multiple values because you are trying to optimize multiple functions at the same time, then you need gamultiobj() instead of ga()

Is it necessary to modify the ga function directly in the file implemented to apply this option to the corresponding function?

No, see gaoptimset() to construct an options structure. The options structure must be passed as either the 10th or 11th parameter to ga(); you can use [] for parameters you are not using. For example,

ga(fitness_function_my_own, 3, [], [], [], [], [], [], [], options)

Note: you do not pass your data directly in to fitness_function_my_own . Instead see http://www.mathworks.com/help/matlab/math/parameterizing-functions.html