Skip to content

inzva AI Projects #3 - Solving Combinatorial Optimization Problems with RL

License

Notifications You must be signed in to change notification settings

inzva/combinatorial-optimization-with-RL

Repository files navigation

inzva AI Projects #3 - Solving Combinatorial Optimization Problems with RL

Combinatorial optimization problems, a famous example of which is the Traveling Salesman Problem, are often solved with hand-crafted algorithms that are often costly to develop and difficult generalize. With this project, we investigate how state-of-the-art models can be trained to learn such heuristics automatically via reinforcement learning, and propose additional tweaks to improve its performance.

Contributors

*Both contributed equally.

**We also thank Sercan Amaç for his contributions.

More information

  1. COwRL.pdf file contains the slides describing the problem and the performed analysis.

  2. To train models using different distributions, you can replace the content of attention-learn-to-route/problems/tsp/problem_tsp.py file with the contents of problem_tsp_normal.py and problem_tsp_exp.py for normal and exponential distributions respectively.

  3. pretrained_novel folder contains models trained with data from different distributions rather than the uniform distribution utilied in the original work.

  4. TSP20_Experiments.ipynb and TSP50_Experiments.ipynb are two standalone notebooks for evaluating performance results of existing and novel pretrained models, and baselines on data from different distributions.

Feel free to open issues or contact us.

About

inzva AI Projects #3 - Solving Combinatorial Optimization Problems with RL

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published