Greetings,
I was wondering if it would be possible to get output files
for various parameters. This would allow to plot and monitor more closely
the timely relationship between the genotype and its evolution. I assume
this is not possible as we speak, but would be a nice thing to play with.
Also, I'm not quite sure about the mechanism of deletion of
genotypes. Why sometime I get up to 100 genotypes and a million iteration
later, there is only 3 left? This is not especially good because it forced
my unsupervised evolution to dwell in a local minimum of slow crawling bugs.
I used the "inversely proportional to fitness" methods.
Thanks for the great software, I will use it to introduce Alife in my
biocomputing class.
Christian
Re: Output
> I was wondering if it would be possible to get output files
> for various parameters. This would allow to plot and monitor more closely
> the timely relationship between the genotype and its evolution. I assume
> this is not possible as we speak, but would be a nice thing to play with.
If you mean saving gene pool every period of time, it is possible.
See SimulatorParameters->Files. Then you are able to analyze
various time relationships.
> Also, I'm not quite sure about the mechanism of deletion of
> genotypes. Why sometime I get up to 100 genotypes and a million iteration
> later, there is only 3 left? This is not especially good because it forced
> my unsupervised evolution to dwell in a local minimum of slow crawling bugs.
> I used the "inversely proportional to fitness" methods.
That is the effect of using double pressure: you select
genotypes according to the roulette wheel rule, and you
delete them according to the same rule. If you use
directed selection, it's enough to use random deletion.
Best,
Maciej