Galapagos Example file + Blog post

Hi All,

just put up a blog post that deals with one of the examples I show in lectures about Galapagos. The gh file is downloadable via the post.

http://ieatbugsforbreakfast.wordpress.com/2011/12/01/multiple-local...

David Rutten

david@mcneel.com

Poprad, Slovakia

Replies to This Discussion

Permalink Reply by Michael Pryor on December 1, 2011 at 4:03pm

been waiting for this, Thanks.

Permalink Reply by Yoann Mescam (Systemiq) on December 1, 2011 at 5:46pm

I was wondering, is there a way to collect all the "decent" local optima ? For example those with a fitness above a fixed value ?

At the moment, unless I missed something obvious, you can see these optima during the Galapagos process but when it stops you only have access to variations of global optimum right ?

Permalink Reply by taz on December 1, 2011 at 6:43pm

There were a couple discussions on this...

Finches and Slingshot seem to be able to let you store genome varieties but you'll have to cull the data for your fitness threshold.

http://www.grasshopper3d.com/forum/topics/keep-last-galapagos-round...

http://www.grasshopper3d.com/forum/topics/grasshopper-solution-counter

Permalink Reply by Yoann Mescam (Systemiq) on December 1, 2011 at 6:54pm

Thanks for the links Taz.

Permalink Reply by taz on December 1, 2011 at 7:04pm

Actually I think Nathan's Slingshot can do the sorting too.

Check out his blog:

http://nmillerarch.blogspot.com/2011/07/genome-database-using-sling...

Permalink Reply by Yoann Mescam (Systemiq) on December 1, 2011 at 7:43pm

However, there is a difference between keeping track of solutions (which are just temporary steps to find the absolute winner and tends to disappear during the evolution without being permitted to reach the local optimum near them) , and converging towards local optima.

Or perhaps playing with Maintain and Inbreeding ?

Although those good candidates can be used as seeds for researching optima I guess.

Permalink Reply by David Rutten on December 2, 2011 at 4:19am

I suppose. The information is available in Galapagos, it's just impossible to save and navigate it at the moment. It's also not really obvious whether a group of nearby individuals are populating a single optimum or a collection of related optima.

David Rutten

david@mcneel.com

Poprad, Slovakia

Permalink Reply by Yoann Mescam (Systemiq) on December 2, 2011 at 7:22am

About Maintain and Inbreeding, would acting on those factors favor the preservation of local optima ? High Maintain and low Inbreeding?

I suppose that the research for the optimum would be compromised while maintaining diversity.

I will experiment.

Permalink Reply by David Rutten on December 2, 2011 at 8:13am

Not really. A lower optimum will almost always go extinct very quickly if different genomes are on a higher peak. No matter the settings.

David Rutten

david@mcneel.com

Poprad, Slovakia