An improved multi-population genetic algorithm with a greedy job insertion inter-factory neighborhood structure for distributed heterogeneous hybrid flow shop scheduling problem

Abstract

Distributed heterogeneous hybrid flow shop scheduling problem (DHHFSP) is an extension of the classical hybrid flow shop scheduling problem (HFSP), which is an NP-hard problem. In the context of economic globalization, DHHFSP considers the collaboration and heterogeneity among multiple factories. The neighborhood structure plays an important role in continuously improving the current individuals for DHHFSP. The existing work has fully studied the inner-factory neighborhood structure, but the research on the inter-factory neighborhood structure is still immature. A greedy job insertion inter-factory neighborhood structure and a new move evaluation method are designed to ensure the efficiency of neighborhood movement. And an improved multi-population genetic algorithm (IMPGA) is proposed to solve the DHHFSP with makespan. To enhance the convergence speed and robustness of the IMPGA, a guided sub-populations information interaction and a re-initialization procedure with an individual resurrection strategy are designed respectively. In computational experiments, there are 480 instances (including the same proportion of small, medium, and large-scale problems) randomly generated. The proposed IMPGA obtains the best solutions for 457 instances. The analysis of experimental results shows that IMPGA significantly outperforms the reported state-of-the-art algorithms for DHHFSP. Finally, the proposed method is used to solve a polyester film manufacturing company case effectively.