A Decomposition-Based Coevolutionary Algorithm to Solve Distributed Heterogeneous Hybrid Flow Shop Scheduling Problem with Job Deadlines and Priorities

Distributed heterogeneous hybrid flow shop scheduling with job deadlines and priorities (DHHFSP-JDP) is a combination of scheduling problem and distributionary environment. Addressing complex work sequences and energy consumption in distributed manufacturing with heterogeneous plants is a major challenge. It is necessary for optimizing total weighted delay (TWD) and total energy consumption (TEC) in distributed heterogeneous green hybrid flowshops. A model using mixed integer linear programming is applied to describe DHGHFSP-JDP and a decomposition-based coevolutionary algorithm (DBCEA) is considered to be the solution in this article. In this approach, (1) a decomposition-based heuristic initialization is proposed, in which an initialization strategy with a randomly sized population is adopted to establish effective initial schedules. (2) elite selection strategy based on the integration of an external archive and an elite archive. (3) four problem-based operator selection strategies embedded in a cooperative local search framework, and an Upper Confidence Bound (UCB) mechanism to design a strategy for selecting local search operators. In the end, The superiority of DBCEA is validated through comparative experiments against several advanced algorithms across 20 benchmarks, with results showing it often provides the best Pareto solution set.