ga minlp algorithm default parameters

7 次查看（过去 30 天）

Wouter Dillen 2017-6-6

1
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/343528-ga-minlp-algorithm-default-parameters

评论： Doug Rank 2019-11-18

Hello there,

I am solving an optimization problem with integer variables using the GA function. The documentation states that, for integer-valued problems, the GA overrides any custom functions for fitness scaling, selection, crossover and mutation, and replaces them with functions based on the algorithm by Deep et al (reference [2] of the documentation on Mixed Integer Optimization).

My question is this: where can I find the default values for the MATLAB implementation of this GA MINLP algorithm (e.g. the index of mutation p for the "Power mutation" function)? I am unable to extract them from the source files because gaminlp.p and its derived files are protected.

Thanks in advance.

2 个评论
显示无隐藏无

Abdullah Kokcam 2017-7-31

Hi,

Similar problem here, I also wonder how ga performs behind the scenes. I'm tackling with mutation ratio for a custom mutation function.

I appreciate all suggestions.

Doug Rank 2019-11-18

I similarly am curious about the details behind the scenes with respect to crossover.

The first generation of an integer (binary) problem with linear constraints produces a huge fraction of infeasible population members. This fraction decreases gradually with subsequent generations until the entire population is feasible. I'm wondering if I can reformulate the solution to optimize these garbage crossover children.

请先登录，再进行评论。

请先登录，再回答此问题。