An effective hybrid particle swarm optimization for batch scheduling of polypropylene processes

Short-term scheduling for batch processes which allocates a set of limited resources over time to manufacture one or more products plays a key role in batch processing systems of the enterprise for maintaining competitive position in fast changing market. This paper proposes an effective hybrid particle swarm optimization (HPSO) algorithm for polypropylene (PP) batch industries to minimize the maximum completion time, which is modeled as a complex generalized multi-stage flow shop scheduling problem with parallel units at each stage and different inventory storage policies. In HPSO, a novel encoding scheme based on random key representation, a new assignment scheme STPT (smallest starting processing time) by taking the different intermediate storage strategies into account, an effective local search based on the Nawaz–Enscore–Ham (NEH) heuristic, as well as a local search based on simulated annealing with an adaptive meta-Lamarckian learning strategy are proposed. Simulation results based on a set of random instances and comparisons with several adaptations of constructive methods and meta-heuristics demonstrate the effectiveness of the proposed HPSO.