Reinforcement learning and stochastic dynamic programming for jointly scheduling jobs and preventive maintenance on a single machine to minimise earliness-tardiness

This paper addresses the problem of stochastic jointly scheduling of resumable jobs and preventive maintenance on a single machine, subject to random breakdowns, to minimise the earliness-tardiness cost. The main objective is to investigate using trending machine learning-based methods compared to s...

Full description

Saved in:
Bibliographic Details
Published in:International journal of production research Vol. 62; no. 3; pp. 705 - 719
Main Authors: Sabri, Abderrazzak, Allaoui, Hamid, Souissi, Omar
Format: Journal Article
Language:English
Published: London Taylor & Francis 01.02.2024
Taylor & Francis LLC
Subjects:
Advanced planning and scheduling systems
Algorithms
Computer Science
Computing time
Deep learning
Dynamic programming
Lateness
Machine learning
Preventive maintenance
Run time (computers)
Scheduling
Stochastic programming
ISSN:0020-7543, 1366-588X
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper addresses the problem of stochastic jointly scheduling of resumable jobs and preventive maintenance on a single machine, subject to random breakdowns, to minimise the earliness-tardiness cost. The main objective is to investigate using trending machine learning-based methods compared to stochastic optimisation approaches. We propose two different methods from both fields as we solve the same problem firstly with a stochastic dynamic programming model in an approximation way, then with an attention-based deep reinforcement learning model. We conduct a detailed experimental study according to solution quality, run time, and robustness to analyse their performances compared to those of an existing approach in the literature as a baseline. Both algorithms outperform the baseline. Moreover, the machine learning-based algorithm outperforms the stochastic dynamic programming-based heuristic as we report up to 30.5% saving in total cost, a reduction of computational time from 67 min to less than $ 1s $ 1 s on big instances, and a better robustness. These facts highlight clearly its potential for solving such problems.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0020-7543
1366-588X
DOI:10.1080/00207543.2023.2172472