Learning Heuristics for Dynamic Graph Combinatorial Optimization

Spatio-Temporal architecture to solve dynamic combinatorial optimization and this code base build on top of the following code (https://github.com/wouterkool/attention-learn-to-route).

Currently, dynamic versions of Travelling Salesmen Problem and Vehicle Routing Problem are supported.

Requirements

• Python>=3.8
• NumPy
• SciPy
• PyTorch>=1.7
• tqdm
• tensorboard_logger
• Matplotlib (optional, only for plotting)

Acknowledgements

We thank attention learning to route [https://github.com/wouterkool/attention-learn-to-route] for easily extendable code base.