Can Galapagos fitness match a tree?

I would like to slice surfaces with a moving curve, in such a way that one of the output portions has an area similar to a predefined number. Say I have a tree with cut surfaces with a certain area (fitness) and sliders to control the movement of such cutting curves. I’d like such surfaces to have an area similar to the numbers on a different, matching tree. However, this is not a matter of maximizing or minimizing fitness, but matching a tree with desired results.
Thanks in advance!

Hello @juliamiranda.aloise,

You could probably subtract the sum of region a from the sum of region b, and feed this absolute difference as fitness value to the Galagos solver.
If the difference between both areas is equal to or very close to 0.0, both regions are also equal or at least very similar to each other in terms of their respective area. This mean that inside the solver, you want to minimise your fitness value towards 0.0.
It would also be a good idea to test this first with a single case.

Hello @diff-arch!

Thanks! It did work! I found it to be weirdly inaccurate though, with certain branches approaching fitness much more than others after a certain runtime. Do you think this has to do with the complexity of geometry (actually, the worst geometries were the most orthogonal ones), or with the fact that I am processing multiple geometries at the same time, or sth else?

@juliamiranda.aloise
From testing, Galagos takes the average of all the results and uses that as how close it is to getting to a solution. No idea if this would help.

It’s hard to say! The inaccuracy could be related to many factors.
Would you be willing to upload an example file? It would make helping you much more straightforward.