Galapagos to find shortest path

Hey all,

I am trying to use galapagos to adjust the integer inputs of items to create lines who's length falls within a certain range. It seems pretty straight forward but for some reason when it finishes I always have like 4 lines that meet the criteria and about twenty that dont. any ideas?

Attachments:

Rule2.ghx, 4.6 MB

Replies to This Discussion

Permalink Reply by Kris Hammerberg on December 2, 2010 at 1:48am

I believe its your fitness function that needs work. My understanding of Galapagos is that it requires one number to optimize and it looks like your definition is feeding it 41 numbers. You'll want to create a function that sums the difference (from your target length) of all the lines and then set Galapagos to minimize. You could work the function so that it adds a large value for any difference that falls outside your tolerance so that genome is particularly unfit.

Permalink Reply by Nels Long on December 2, 2010 at 1:55am

yeah that makes sense. i was thinking that the fitness would be maximized if it fell within that range but i guess thats just picking a number within the range. makes sense now thank you!

Permalink Reply by Kris Hammerberg on December 2, 2010 at 2:03am

Sitting down to do it, I ended up taking a different approach. I used < > operations and logic gates to determine if a value was within your range. Then I found the total number that were true and subtracted from it the number that that were not. In this case you'd set Galapagos to maximize. I didn't have points you were using so I didn't run it, but this should provide an answer with more in the range than not. ;)

Attachments: