A Knowledge-Driven Hybrid Algorithm for Solving the Integrated Production and Transportation Scheduling Problem in Job Shop

Abstract

Intelligent transportation systems, incorporating multiple AGVs, are extensively utilized in manufacturing workshops in various industries. This widespread use has spurred significant research interest in the integrated production and transportation scheduling problem, particularly in job shop environments. However, current research often fails to adequately leverage domain knowledge, leading to algorithms that struggle to find high-quality solutions for large-scale problems. To address this issue, this paper proposes a knowledge-driven hybrid algorithm (KDHA). The domain knowledge incorporated in the KDHA includes: 1) three critical path-based neighborhood structures for comprehensive neighborhood solution searches, 2) three neighborhood cropping methods to avoid ineffective searches for poor solutions, and 3) a new fast evaluation method to enhance the efficiency of neighborhood solution searching. Additionally, a new encoding method is introduced to achieve a one-to-one mapping between the chromosome and the disjunctive graph, allowing valuable information from neighborhood solutions to contribute to the algorithm’s evolution. Comparative experiments between the proposed algorithm and other state-of-the-art approaches are conducted on the small-scale EX and large-scale SWV benchmarks. The results demonstrate that the proposed KDHA is able to output better solutions efficiently and consistently, and updates the best solutions of all 20 SWV instances.