Hybrid genetic algorithm based on bin packing strategy for the unrelated parallel workgroup scheduling problem

• Published in: Journal of intelligent manufacturing Vol. 32; no. 4; pp. 957 - 969
• Main Authors: Su, Bentao, Xie, Naiming, Yang, Yingjie
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.04.2021; Springer Nature B.V
• Subjects: Advanced manufacturing technologies; Algorithms; Business and Management; Control; Genetic algorithms; Heuristic methods; Human resources; Integer programming; Machines; Manufacturing; Mechatronics; Processes; Production; Resource scheduling; Robotics; Scheduling; Unrelated parallel machine problem; Heuristic strategy; Workgroup scheduling; Genetic algorithm; Two-dimensional bin packing problem
• ISSN: 0956-5515; 1572-8145; 1572-8145
• DOI: 10.1007/s10845-020-01597-8

In this paper we focus on an unrelated parallel workgroup scheduling problem where each workgroup is composed of a number of personnel with similar work skills which has eligibility and human resource constraints. The most difference from the general unrelated parallel machine scheduling with resource constraints is that one workgroup can process multiple jobs at a time as long as the resources are available, which means that a feasible scheduling scheme is impossible to get if we consider the processing sequence of jobs only in time dimension. We construct this problem as an integer programming model with the objective of minimizing makespan. As it is incapable to get the optimal solution in the acceptable time for the presented model by exact algorithm, meta-heuristic is considered to design. A pure genetic algorithm based on special coding design is proposed firstly. Then a hybrid genetic algorithm based on bin packing strategy is further developed by the consideration of transforming the single workgroup scheduling to a strip-packing problem. Finally, the proposed algorithms, together with exact approach, are tested at different size of instances. Results demonstrate that the proposed hybrid genetic algorithm shows the effective performance.