A NOVEL DECENTRALIZED APPROACH FOR PRODUCTION SCHEDULING

Abstract

This paper proposes a novel decentralized approach for multi-stage job-shop scheduling. The method divides the larger job-shop scheduling problem into several smaller problems, which various agents solve. The collaboration among these agents ensures the exploration of globally superior solutions while allowing enhanced local exploration. Based on an extensive investigation, the current work shows that the proposed approach outperforms the centralized approach, especially for problems with increasing problem size, in a faster manner. Since the proposed approach is based on decentralized information processing, it is easily adaptable to next-generation cyber-physical manufacturing systems. © INTERNATIONAL JOURNAL OF INDUSTRIAL ENGINEERING.