A multi-objective optimization for preemptive identical parallel machines scheduling problem

• Published in: Computational & applied mathematics Vol. 36; no. 3; pp. 1367 - 1387
• Main Authors: Aalaei, Amin, Kayvanfar, Vahid, Davoudpour, Hamid
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.09.2017; Springer Nature B.V
• Subjects: Applications of Mathematics; Applied physics; Classification; Computational mathematics; Computational Mathematics and Numerical Analysis; Genetic algorithms; Heuristic methods; Industrial applications; Mathematical Applications in Computer Science; Mathematical Applications in the Physical Sciences; Mathematics; Mathematics and Statistics; Multiple objective analysis; Philosophy; Preempting; Production scheduling; Scheduling; Sorting algorithms; Workflow; Preemption; Secondary 68M20; Identical parallel machines; Primary 90B35; Just-in-time; Non-dominated sorting genetic algorithm; Earliness and tardiness
• DOI: 10.1007/s40314-015-0298-0
• ISSN: 0101-8205; 2238-3603

This research investigates a new notion in just-in-time philosophy on the identical parallel machines considering allowable job preemption with respect to a bi-objective approach. The work-in-process (WIP) is also allowed, since minimization of WIP is desirable in many industrial applications specifically those including perishable items. In this new notion, a new model is defined in which the earliness costs depend on the start times of the jobs. The goal of this study is to minimize two objectives simultaneously: (1) total weighted earliness and tardiness as well as holding cost of all jobs which are waiting to be processed as WIP costs and (2) number of jobs interruptions. In this context, two multi-objective meta-heuristic algorithms, i.e., the non-dominated sorting genetic algorithm II (NSGAII) and non-dominated ranking genetic algorithm (NRGA) are employed to solve such bi-objective problems. Three measurement factors are then employed to evaluate the algorithms performances. Computational results demonstrate that NRGA outperforms NSGAII in all small- and medium-to-large-sized sample-generated problems; however, intangibly.