Re: [jgap-users] Promoting Genotype Diversity
Brought to you by:
klausikm
From: Klaus M. <jg...@kl...> - 2013-02-07 13:43:45
|
Simon, thanx for your interest in JGAP. Regarding the fitness function you are free to define your own (to be honest: You have to define your owu fitness function for your own problems to solve). Regarding promoting diversity, it would be a good idea to implement specific genetic operators, such as crossover and mutation operator in order to generate diverse chromosomes. Please see the JGAP source code. There you find a lot of examples with different fitness functions. In package org.jgap.impl., you find several crossover and mutation operators that may guide you in implementing your own. Best Klaus www.klaus-meffert.com <http://www.klaus-meffert.com/> _____ From: Simon Greaves [mailto:si...@ya...] Sent: Wednesday, February 06, 2013 7:21 PM To: jga...@li... Subject: [jgap-users] Promoting Genotype Diversity Hello My name is Simon and I'm trying to implement JGAP to solve a problem that has potentially multiple acceptable solutions, so I want to promote diversity during evolution, so that I won't in the end have the top n chromosomes simply be almost identical to the optimal chromosome. In the theory one of the ways to do this is by using the Hamming Distance to devalue the Fitness Function of the chromosomes. Is there a way to do this in JGAP? Or some other way to promote population diversity? Otherwise I would need to change the current Fitness Function in order to compare one chromosome with the others in the Genotype, which I'm not sure I can do... Thanks in advance for any help you can give me! Simon |