SBX and Polynomial Mutation: eta parameter

[Guilherme Prado]

"Deb and Agrawal suggested a polynomial mutation operator with a user-defined index parameter (ηm)...
They also found that a value ηm ∈ [20,100] is adequate in most problems that they tried."

I was trying SBX and polynomial mutation operators with ηm ∈ [20,100] to minimizing the sphere function with 20 genes and the GA founds best solutions around 1e-3. When I used small values of  ηm, like ηm ∈ [0,1] and the GA founds best solutions around 1e-13. There's any problem by using values for ηm very smaller than the set recommended by Deb?
I'm using the eaMuPlusLambda algorithm.

Thank in advance!

Deb, K. and ayan Deb, 2014. “Analysing mutation schemes for real-parameter genetic algorithms”. International Journal
of Artificial Intelligence and Soft Computing, Vol. 4, No. 1, p. 1. doi:10.1504/ijaisc.2014.059280.

[François-Michel De Rainville]

I don’t see any problem with small values. It will inject a lot more diversification in your population. In some case it might be good. However, I encourage you to look at the values of your individuals’ gene if they don’t get extreme values or always bounded.

--
You received this message because you are subscribed to the Google Groups "deap-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to deap-users+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/deap-users/e3826325-2c59-4bdb-85ce-5f96650035f9o%40googlegroups.com.