> The idea was to allow the user to "freeze" the positions of a bunch of nodes, and then try and place all the remaining nodes in a way that does not violate elasticity constraints and does not put parts of the figure outside of the hole boundaries.
IMHO it was a brilliant idea! What we (team skobochka) had was a separate manual visual solver and a standalone bruteforce solver, and there was always a problem with search space size.
We managed to achieve the best results with monte-carlo solver (simulated annealing) but it wasn't able to score 0 on large problems :(
![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
asenathwaite
(no subject)
Date: 2021-07-17 12:37 pm (UTC)IMHO it was a brilliant idea! What we (team skobochka) had was a separate manual visual solver and a standalone bruteforce solver, and there was always a problem with search space size.
We managed to achieve the best results with monte-carlo solver (simulated annealing) but it wasn't able to score 0 on large problems :(