TY - GEN
T1 - Solving a Multi-objective Job Shop Scheduling Problem with an Automatically Configured Evolutionary Algorithm
AU - Para, Jesús
AU - Del Ser, Javier
AU - Nebro, Antonio J.
N1 - Publisher Copyright:
© 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2023
Y1 - 2023
N2 - In this work we focus on optimizing a multi-objective formulation of the Job Shop Scheduling Problem (JSP) which considers the minimization of energy consumption as one of the objectives. In practice, users experts in the problem domain but with a low knowledge in metaheuristics usually take an existing algorithm with default settings to optimize problem instances but, in this context, the use of automatic parameter configuration techniques can help to find ad-hoc configurations of algorithms that effectively solve optimization problems. Our aim is to study what improvement in results can be obtained by applying an autoconfiguration approach versus using a set of well-known multi-objective evolutionary algorithms (NSGA-II, SPEA2, SMS-EMOA and MOEA/D) for different instances of the JSP, with varying dimensionality. Our experiments showcase the potential of automated algorithmic configuration for energy-efficient production scheduling, producing better balanced solutions than the multi-objective solvers considered in the study.
AB - In this work we focus on optimizing a multi-objective formulation of the Job Shop Scheduling Problem (JSP) which considers the minimization of energy consumption as one of the objectives. In practice, users experts in the problem domain but with a low knowledge in metaheuristics usually take an existing algorithm with default settings to optimize problem instances but, in this context, the use of automatic parameter configuration techniques can help to find ad-hoc configurations of algorithms that effectively solve optimization problems. Our aim is to study what improvement in results can be obtained by applying an autoconfiguration approach versus using a set of well-known multi-objective evolutionary algorithms (NSGA-II, SPEA2, SMS-EMOA and MOEA/D) for different instances of the JSP, with varying dimensionality. Our experiments showcase the potential of automated algorithmic configuration for energy-efficient production scheduling, producing better balanced solutions than the multi-objective solvers considered in the study.
KW - Automatic Algorithm Configuration
KW - Job Shop Scheduling
KW - Multi-Objective Optimization
UR - http://www.scopus.com/inward/record.url?scp=85163300084&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-34020-8_4
DO - 10.1007/978-3-031-34020-8_4
M3 - Conference contribution
AN - SCOPUS:85163300084
SN - 9783031340192
T3 - Communications in Computer and Information Science
SP - 48
EP - 61
BT - Optimization and Learning - 6th International Conference, OLA 2023, Proceedings
A2 - Dorronsoro, Bernabé
A2 - Chicano, Francisco
A2 - Danoy, Gregoire
A2 - Talbi, El-Ghazali
PB - Springer Science and Business Media Deutschland GmbH
T2 - 6th International Conference on Optimization and Learning, OLA 2023
Y2 - 3 May 2023 through 5 May 2023
ER -
