search

wouterkool / attention-learn-to-route

Attention based model for learning to solve different routing problems
MIT License
1.08k stars 341 forks source link

About the optimal objective of all problems #31

Open bywbilly opened 3 years ago

bywbilly commented 3 years ago

Hi,

I am wondering can u provide us with the optimal objective of all problems' validation set? Or tell me how can I get them from your code.

Thanks!

wouterkool commented 3 years ago

Hi!

I guess you're talking about TSP? You can install Concorde by running ./install_concorde.sh from the directory problems/tsp. Then from the root directory you can run python -m problems.tsp.tsp_baseline concorde data/tsp/tsp20_validation_seed4321.pkl data/tsp/tsp50_validaton_seed4321.pkl data/tsp/tsp100_validation_seed4321.pkl. These sets can be generated using python generate_data.py --problem all --name validation --seed 4321. See also the instructions. Solving all sets should take a reasonable amount of time (not more than an hour at most).