ai4co / rl4co

A PyTorch library for all things Reinforcement Learning (RL) for Combinatorial Optimization (CO)
https://rl4.co
MIT License
432 stars 80 forks source link

[Feat] Add DeepACO #142

Closed Furffico closed 7 months ago

Furffico commented 7 months ago

Description

Implementing the non-autoregressive algorithm DeepACO on RL4CO.

Motivation and Context

As a non-autoregressive algorithm, DeepACO leverages Ant Colony Optimization (ACO) algorithms to construct solutions without relying on autoregressive decoding, thereby extending the capabilities of RL4CO.

Types of changes

What types of changes does your code introduce? Remove all that do not apply:

Checklist

Go over all the following points, and put an x in all the boxes that apply. If you are unsure about any of these, don't hesitate to ask. We are here to help!

fedebotu commented 7 months ago

LGTM! Thoughts @cbhua?

We can merge now IMO~

Things to do afterward:

  1. Reproduce results from the paper (at least make sure it works the same as the original implementation in terms of runtime and convergence)
  2. Integration with local search CC: @hyeok9855 (and relative checking results~)
cbhua commented 7 months ago

LGTM!