Recently I am trying to use Galapagos to do some self-evolving form. My question is, is there an effective method that I can link massive input into the calculator? For example, if I want to change the size of 50 or 60 holes on a surface individually to Galapagos, is there are a better way to input data rather than make 50 number sliders to Galapagos?
You can use the Gene Pool object to define a large number of sliders.
Thank you very much!
On the other hand, if I hope the input number all be different, is there are better way than using random or random reduce?
What I am trying to do is to reduce 30 circles from point cloud to obtain better light result. Previously I used random reduce, and linked seed to Galapagos. However, as a result, I noticed this method is too random for Galapagos to obtain a clear evolution patter.
What I think might work is to create a gene pool with 1 and 0 to open and close each circle, but I am not sure how to control total number of “0”, or closed point to 30.
Could you please help?
Thank you very much.
- Uploading your file makes it easier/faster to help…
- you could create a Boolean. If total amount of the sliders is 30, it‘s true.
For Galapagos: create a filter - true your calculated value will be streamed, if not, 0 will be streamed.
For octopus: you can just set the Boolean as goal.
What you’re describing is Binary Integer Programming - something Galapagos is not suited to handling, since it needs a smooth gradient function on the solution landscape.
Think of the solver tweaking sliders to see if the fitness moves ‘uphill’ or ‘downhill’ - it cannot do that when the sliders are basically On/Off switches and the output (
N==30) is also discontinuous. You’ll probably have to ‘rephrase’ your inputs in a more continuous manner or use some other tool besides Galapagos.
Of course, no one can help any further without a file-