Print

Print


Can anyone help me?

Thank you =)

On 19 March 2012 10:22, Rodrigo Barros <[log in to unmask]> wrote:

> Hello,
>
> I'm working in a hyper-heuristic GA to evolve machine learning algorithms.
> I was using a traditional breeding pipeline (tournament selection -->
> crossover --> mutation), but I decided to change it so it would look
> something like this: (tournament selection -->
> reproduction/crossover/mutation), in which there is 5% of chance of
> reproducing (cloning) the selected individuals, 5% of chance of mutating
> the individuals and 90% of chance of them suffering crossover. In other
> words, the probabilities should sum to 1 considering that the selected
> individuals will undergo one of the three operations at a time.
>
> I edited my parameter file to look like this:
>
> pop.subpop.0.species.pipe = ec.breed.MultiBreedingPipeline
> pop.subpop.0.species.pipe.num-sources = 3
> pop.subpop.0.species.pipe.source.0 = ec.breed.ReproductionPipeline
> pop.subpop.0.species.pipe.source.0.prob = 0.05
> pop.subpop.0.species.pipe.source.1 =
> ec.vector.breed.VectorCrossoverPipeline
> pop.subpop.0.species.pipe.source.1.prob = 0.90
> pop.subpop.0.species.pipe.source.2 = ec.vector.breed.VectorMutationPipeline
> pop.subpop.0.species.pipe.source.2.prob = 0.05
>
> pop.subpop.0.species.pipe.source.0.source.0 = ec.select.TournamentSelection
> pop.subpop.0.species.pipe.source.1.source.0 = ec.select.TournamentSelection
> pop.subpop.0.species.pipe.source.1.source.1 = same
> pop.subpop.0.species.pipe.source.2.source.0 = ec.select.TournamentSelection
>
> pop.subpop.0.species.crossover-type = one
> pop.subpop.0.species.mutation-prob = 0.5
> pop.subpop.0.species.mutation-type = reset
>
> select.tournament.size = 2
>
>
> Is this correct? Or am I missing something important? Apparently it works,
> but I'm not sure it is working as it should.
>
> Thank you for your help.
>
> --
> Rodrigo Coelho Barros, MSc.
> PhD Candidate
> Bioinspired Computation Laboratory (BioCom)
> Department of Computer Science
> University of São Paulo at São Carlos (ICMC-USP)
>
> www.icmc.usp.br/~rcbarros
> e-mail (1): [log in to unmask]
> e-mail (2): [log in to unmask]
>
> mobile: +55 16 8138-3627
>



-- 
Rodrigo Coelho Barros, MSc.
PhD Candidate
Bioinspired Computation Laboratory (BioCom)
Department of Computer Science
University of São Paulo at São Carlos (ICMC-USP)

www.icmc.usp.br/~rcbarros
e-mail (1): [log in to unmask]
e-mail (2): [log in to unmask]

mobile: +55 16 8138-3627