ML-based LS Operator Selection #2566
Replies: 2 comments 1 reply
-
|
Do you mean changing this parameter during the search ? or-tools/ortools/constraint_solver/samples/vrp_with_time_limit.py Lines 99 to 100 in b37d9c7 What about running in parallel the same model with different LS metaheuristic ? |
Beta Was this translation helpful? Give feedback.
1 reply
-
|
As an update, I tried accessing the routing model's underlying solver, adding a decision builder, and calling newSearch but to no response from the code. I'm guessing that function isn't available to call. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
I've been looking into the GoogleOR documentation and trying to find a way to change or keep the local search operator being used at any given step. I looked into search monitors and then the metaheuristic models but couldn't seem to find a way to select the operator to use mid-way. I'm not certain if this route is feasible as the examples don't show a method of switching LS Operators either.
The main concern was to also be able to use the routing model. I looked into the simplest method of restarting the search or setting up a new one every time with a 1-step limit but that caused unneeded overhead and was often buggy, especially when accessing the routing model through the python wrapper.
Any thoughts on other steps I can try to customize the model? I'm obviously quite new to GoogleOR as a whole and hoping here would be a good place to get a proper foothold.
Beta Was this translation helpful? Give feedback.
All reactions