Deep Reinforcement Learning Guided Improvement Heuristic for Job Shop Scheduling

Paper: https://openreview.net/forum?id=jsWCmrsHHs

If you make use of the code/experiment or L2S algorithm in your work, please cite our paper (Bibtex below).

@InProceedings{
zhang2024deep,
title={Deep Reinforcement Learning Guided Improvement Heuristic for Job Shop Scheduling},
author={Zhang, Cong and Cao, Zhiguang and Song, Wen and Wu, Yaoxin and Zhang, Jie},
booktitle={The Twelfth International Conference on Learning Representations},
year={2024}
}

We have a better version of L2S in TBGAT, published in UAI-2024

System Requirement

Ubuntu 18.04.5 LTS

Nvidia Docker

python 3.8.x

pytorch 1.7.0+cu110 (torchvision 0.8.0, torchaudio 0.7.0)

pip install torch==1.7.0 torchvision==0.8.0 torchaudio==0.7.0 -f https://download.pytorch.org/whl/cu110/torch_stable.html

Then install dependencies: (if using venv or conda, omit --user)

pip install --user --upgrade pip
pip install --user torch-scatter==2.0.6 -f https://pytorch-geometric.com/whl/torch-1.7.0+cu110.html
pip install --user torch-sparse==0.6.9 -f https://pytorch-geometric.com/whl/torch-1.7.0+cu110.html
pip install --user torch-geometric==1.7.2
pip install --user matplotlib==3.4.3
pip install --user ortools==9.0.9048

Docker Setup (preferred)

Clone this repo and within the repo folder run the following command.

To create image l2s_image:

sudo docker build -t l2s_image .

To create container l2s_container from l2s_image, and activate it:

sudo docker run --gpus all --name l2s_container -it l2s_image

Reproducing

To reproduce the result in the paper, first clone the whole repo:

git clone https://github.com/zcaicaros/L2S.git

Then run:

python3 test_learned.py

To train the model, adjust hyper-parameters in parameters.py, then run:

python3 n-step_reinforce.py