A particle swarm optimization and constraint programming-based approach for integrated process planning and scheduling with lot streaming problem

This paper studies the integrated process planning and scheduling with lot streaming (IPPS-LS) problem, which consists of lot splitting, process planning, and shop scheduling. Although the IPPS-LS problem is common in the manufacturing of flexible process products, it has not been extensively studied due to its high complexity. Hence, this study develops an enhanced particle swarm optimization algorithm based on constraint programming (CP) to minimize makespan. The proposed algorithm employs finite condition and relaxation models for particle reconfiguration and re-optimization. To achieve it, two types of relaxation models are constructed by decomposing the multiple constraints of the CP model. The algorithm dynamically updates particle encoding sequences based on model accuracy, effectively reducing invalid searches and accelerating the search process. The proposed algorithm is compared with models and other metaheuristic algorithms on 120 test instances. The impact of the relaxed CP strategy and particle swarm optimization algorithm on the proposed algorithm performance is also analyzed. Finally, a significance of difference validation is performed. Computational experiments demonstrate the efficiency of the proposed algorithm in solving the IPPS-LS problem of varying scales. In addition, the relaxed CP strategy exhibits a more significant improvement effect for medium-scale problems compared to small and large-scale problems.