Callbacks about alns HOT 3 CLOSED

We have an on_best callback that's called whenever a new global best solution is encountered. That's useful for, say, local search around a new good solution.

Of Santini's list, I would really like to have on_many_iters_without_improvement, which can be used to perhaps reset the weights, acceptance criterion, and/or current solution. I need to think about the required method signature, and what needs to change to determine 'many iterations'.

from alns.

With #64, the need for on_many_iters_without_improvement has lessened. One can implement a stopping criterion that returns in that case, and optionally restart ALNS from there. That's different from Santini's approach, but basically equivalent.

from alns.

I am closing this issue because I cannot think of any things I need or want that I cannot already do without new callbacks.

from alns.