Multi-objective evolutionary algorithm for solving energy-aware fuzzy job shop problems
- PDF / 1,327,805 Bytes
- 12 Pages / 595.276 x 790.866 pts Page_size
- 102 Downloads / 157 Views
METHODOLOGIES AND APPLICATION
Multi-objective evolutionary algorithm for solving energy-aware fuzzy job shop problems Inés González-Rodríguez1
· Jorge Puente2
· Juan José Palacios2
· Camino R. Vela2
© Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract A growing concern about the environmental impact of manufacturing processes and in particular the associated energy consumption has recently driven some researchers within the scheduling community to consider energy costs in addition to more traditional performance-related measures, such as satisfaction of due-date commitments. Recent research is also devoted to narrowing the gap between real-world applications and academic problems by handling uncertainty in some input data. In this paper, we address the job shop scheduling problem, a well-known hard problem with many applications, using fuzzy sets to model uncertainty in processing times and with the target of finding solutions that perform well with respect to both due-date fulfilment and energy efficiency. The resulting multi-objective problem is solved using an evolutionary algorithm based on the NSGA-II procedure, where the decoding operator incorporates a new heuristic procedure in order to improve the solutions’ energy consumption. This heuristic is based on a theoretical analysis of the changes in energy consumption when a solution is subject to slight changes, referred to as local right shifts. The experimental results support the theoretical study and show the potential of the proposal. Keywords Job shop scheduling · Fuzzy durations · Multi-objective · Due dates · Energy efficiency · Genetic algorithm
1 Introduction Scheduling problems appear in a growing number of domains, including engineering, management science or distributed and parallel computing. One of the most relevant problems Communicated by V. Loia. This research has been financially supported by the Spanish Government under research Grant TIN2016-79190-R and by the Principality of Asturias Government under Grant IDI/2018/000176.
B
Camino R. Vela [email protected] Inés González-Rodríguez [email protected] Jorge Puente [email protected] Juan José Palacios [email protected]
1
Department of Maths, Stats and Computing, University of Cantabria, Santander, Spain
2
Dep. of Computer Science, University of Oviedo, Gijón, Spain
is the job shop problem in its numerous variants, since it is considered to be a reference for many practical applications (e.g. wafer fabs in the semiconductor industry often function as job shops) (Jain and Meeran 1999; Pinedo 2016). It also poses a challenge to the research community due to its complexity (Garey and Johnson 1979). The most common objective in the literature consists in finding solutions minimising the execution time span of the project, known as makespan. However, due-date satisfaction has also occupied researchers (Koulamas 1994) and its interest is growing in recent years as on-time fulfilment gains importance in modern pull-oriented supply chain systems, and keeping job du
Data Loading...
