everdyx / Deep-reinforcement-learning-for-dynamic-scheduling-of-a-flexible-job-shop

IJPR paper: Deep Reinforcement Learning for dynamic scheduling of a flexible job shop

Deep reinforcement learning for dynamic scheduling of a flexible job shop

This repository includes the code of algorithms used in the following paper:

Liu, R., Piplani, R., & Toro, C. (2022). Deep reinforcement learning for dynamic scheduling of a flexible job shop. International Journal of Production Research 2022 Vol. 60 Issue 13 Pages 4049-4069. https://doi.org/10.1080/00207543.2022.2058432

Free Eprint link: https://www.tandfonline.com/eprint/ITWCDGRSWFXS7G37PGBQ/full?target=10.1080/00207543.2022.2058432

Feel free to revise the code, and good luck with your research!

If you make use of the code in your work, please consider citing our paper (Bibtex below):

@article{Liu_Piplani_Toro_2022,
author = {Liu, Renke and Piplani, Rajesh and Toro, Carlos},
title = {Deep reinforcement learning for dynamic scheduling of a flexible job shop},
journal = {International Journal of Production Research},
volume = {60},
number = {13},
pages = {4049-4069},
year = {2022},
publisher = {Taylor & Francis},
doi = {10.1080/00207543.2022.2058432},
URL = {https://doi.org/10.1080/00207543.2022.2058432}
}

Repository Overview

This repo includes the code of simulation model, learning algorithm, and experimentation. Those files can be identified by their prefix or suffix:

1. "agent": assets on shop floor: machines and work centers;
2. "brain": learning algorithm for job sequencing and machine routing;
3. "creation": dynamic events on shop floor, such as job arrival and machine breakdown;
4. "main": actual simulation processes, to train, or validate deep MARL-based algorithms;
5. "routing" and "sequencing: priority rules
6. "validation": modules to import trained parameters for experiment.

Data and trained parameters can be found in folders:

1. "experiment_result": results of experiments in dynamic environments;
2. "routing models": trained routing agent parameters;
3. "sequencing_models": trained sequencing agent parameters.

User Guide

Wanna use our code as the benchmark? Kindly refer to "routing_models" and "sequencing_models" folders for the trained parameters for routing agent (RA) and sequencing agent (SA), respectively.

To use trained models:

Use the class "network_validated" within "brain_machine_S.py" module to build the sequencing neural network

Use "build_network_small", "build_network_medium", and "build_network_large" classes in "brain_workcenter-R" module to build routing networks of different size.

The state and action functions can also be found in "brain_machine_S.py" or "brain_workcenter_R.py".

An alternative way is to test your approach in our simulation model and context, you may create your algorithm and run the simulation in "main_experiment_R.py", "main_experiment_S.py", "main_experiment_integrated.py" files for comparison. Please refer to the comments in each module to see how to use the simulation model.