A problem in determining the mutation rate

Hi, everyone

Hope you are well and thanks for your attention. I am using the GLobal optimmisation Tool Boox and I was looking how can I change the default value of Rate of mutation at so I found this syntax options = gaoptimset('MutationFcn', {@mutationuniform, rate}) my problem is that i found also this : For example, if the Population size is 20, the Elite count is 2, and the Crossover fraction is 0.8, the numbers of each type of children in the next generation is as follows: •There are 2 elite children •There are 18 individuals other than elite children, so the algorithm rounds 0.8*18 = 14.4 to 14 to get the number of crossover children. •The remaining 4 individuals, other than elite children, are mutation children. So my question is : what is the influence of determining the mutation rate through options = gaoptimset('MutationFcn', {@mutationuniform, rate}) If the mutation depend on the crossover rate

Thanks for any help or suggestion

Answer 1

[Alan Weiss]

http://www.mathworks.com/help/gads/genetic-algorithm-options.html states "Uniform (mutationuniform) — Uniform mutation is a two-step process. First, the algorithm selects a fraction of the vector entries of an individual for mutation, where each entry has a probability Rate of being mutated. The default value of Rate is 0.01. In the second step, the algorithm replaces each selected entry by a random number selected uniformly from the range for that entry."

http://www.mathworks.com/help/gads/genetic-algorithm-options.html explains what CrossoverFraction does, and how it affects the mutation rate.

Alan Weiss

MATLAB mathematical toolbox documentation

Answer 2

[Abdullah Kokcam]

So I found as crossover fraction and mutation fraction sums up to 100 percent.

For example: If crossover fraction is set to 0.8, it means that mutation fraction can be considered as 0.2

Population size = 100

Elite count = 20

Crossover population = (100-20)*0.8 = 64

Mutation population = (100-20)-64 = 16

Note that mutation fraction may not exactly equal to 0.2 even if crossover fraction is 0.8

If crossover fraction is set to 0.5

Population size = 20

Elite count = 5

Crossover population = (20-5)*0.8 = 7.5 -> 8

Mutation population = (20-5)-8 = 7

Answer 3

[Hector]

Continuing the theme, if I have a custom MutationFcn. How do I change the rate of mutation?

If you read the manual, you really do not go as the percentage change of mutation when a custom MutationFcn.

http://www.mathworks.com/help/gads/gaoptimset.html

How I can change the rate of mutation in a custom MutationFcn?

thanks

Hector R:.

[Abdullah Kokcam]

This question is asked many years ago, but seems that not answered at all. I also wonder this topic. I appreciate if someone share if s/he found the answer. Thanks.