A Reinforcement Learning Environment For Job-Shop Scheduling

Pierre Paul Alain Tassel; Martin Gebser; Konstantin Schekotihin

A Reinforcement Learning Environment For Job-Shop Scheduling

Pierre Paul Alain Tassel^*
, Martin Gebser
, Konstantin Schekotihin

^*Corresponding author for this work

Institute of Software Engineering and Artificial Intelligence (7160)

Research output: Contribution to conference › Paper › peer-review

Abstract

Scheduling is a fundamental task occurring in various automated systems applications, e.g., optimal schedules for machines on a job shop allow for a reduction of production costs and waste. However, finding such schedules is often intractable and cannot be achieved by Combinatorial Optimization Problem (COP) methods within a given time limit. Recent advances of Deep Reinforcement Learning (DRL) in learning complex behavior enable new COP application possibilities. This paper presents an efficient DRL environment for Job-Shop Scheduling – an important problem in the field. Furthermore, we design a meaningful and compact state representation as well as a novel, simple dense reward function, closely related to the sparse make-span minimization criteria used by COP methods.
We demonstrate that our approach significantly outperforms existing DRL methods on classic benchmark instances, coming close to state-of-the-art COP approaches.

Original language	English
Publication status	Published - Aug 2021
Event	2021 PRL Workshop – Bridging the Gap Between AI Planning and Reinforcement Learning - Virtuell, China Duration: 5 Aug 2021 → 6 Aug 2021

Conference

Conference	2021 PRL Workshop – Bridging the Gap Between AI Planning and Reinforcement Learning
Country/Territory	China
City	Virtuell
Period	5/08/21 → 6/08/21

Access to Document

https://prl-theworkshop.github.io/prl2021/papers/PRL2021_paper_9.pdfLicence: CC BY 4.0

Cite this

@conference{a363fabe37494f71ab9cfc6bdb6205c8,

title = "A Reinforcement Learning Environment For Job-Shop Scheduling",

abstract = "Scheduling is a fundamental task occurring in various automated systems applications, e.g., optimal schedules for machines on a job shop allow for a reduction of production costs and waste. However, finding such schedules is often intractable and cannot be achieved by Combinatorial Optimization Problem (COP) methods within a given time limit. Recent advances of Deep Reinforcement Learning (DRL) in learning complex behavior enable new COP application possibilities. This paper presents an efficient DRL environment for Job-Shop Scheduling – an important problem in the field. Furthermore, we design a meaningful and compact state representation as well as a novel, simple dense reward function, closely related to the sparse make-span minimization criteria used by COP methods. We demonstrate that our approach significantly outperforms existing DRL methods on classic benchmark instances, coming close to state-of-the-art COP approaches.",

author = "Tassel, \{Pierre Paul Alain\} and Martin Gebser and Konstantin Schekotihin",

year = "2021",

month = aug,

language = "English",

note = "2021 PRL Workshop – Bridging the Gap Between AI Planning and Reinforcement Learning ; Conference date: 05-08-2021 Through 06-08-2021",

}

TY - CONF

T1 - A Reinforcement Learning Environment For Job-Shop Scheduling

AU - Tassel, Pierre Paul Alain

AU - Gebser, Martin

AU - Schekotihin, Konstantin

PY - 2021/8

Y1 - 2021/8

N2 - Scheduling is a fundamental task occurring in various automated systems applications, e.g., optimal schedules for machines on a job shop allow for a reduction of production costs and waste. However, finding such schedules is often intractable and cannot be achieved by Combinatorial Optimization Problem (COP) methods within a given time limit. Recent advances of Deep Reinforcement Learning (DRL) in learning complex behavior enable new COP application possibilities. This paper presents an efficient DRL environment for Job-Shop Scheduling – an important problem in the field. Furthermore, we design a meaningful and compact state representation as well as a novel, simple dense reward function, closely related to the sparse make-span minimization criteria used by COP methods. We demonstrate that our approach significantly outperforms existing DRL methods on classic benchmark instances, coming close to state-of-the-art COP approaches.

AB - Scheduling is a fundamental task occurring in various automated systems applications, e.g., optimal schedules for machines on a job shop allow for a reduction of production costs and waste. However, finding such schedules is often intractable and cannot be achieved by Combinatorial Optimization Problem (COP) methods within a given time limit. Recent advances of Deep Reinforcement Learning (DRL) in learning complex behavior enable new COP application possibilities. This paper presents an efficient DRL environment for Job-Shop Scheduling – an important problem in the field. Furthermore, we design a meaningful and compact state representation as well as a novel, simple dense reward function, closely related to the sparse make-span minimization criteria used by COP methods. We demonstrate that our approach significantly outperforms existing DRL methods on classic benchmark instances, coming close to state-of-the-art COP approaches.

M3 - Paper

T2 - 2021 PRL Workshop – Bridging the Gap Between AI Planning and Reinforcement Learning

Y2 - 5 August 2021 through 6 August 2021

ER -