Scheduling of No-wait Stochastic Distributed Assembly Flowshop by Hybrid PSO

Distributed assembly permutation flowshop problem (DAPFSP) is a typical NP-hard combinatorial optimization problem where N jobs are to be processed in the one of the F identical factories (processing stage) and then assembled into S final products by a single machine (assembly stage). In this study, we propose a new type of DAPFSP, i.e., the stochastic DAPFSP with no-wait constraint, in which the factories in the processing stage are characterized by flowshop with no-wait constraint between consecutive processing machines, meanwhile the time for processing and assembly is stochastic. To address the difficulties of the proposed scheduling model with respect to minimizing the makespan, we adapt the PSOSAHT algorithm framework proposed previously, where particle swarm optimization based exploration and simulated annealing based exploitation are well integrated, and the hypothesis test is to compare the stochastic makespan in statistical sense to avoid repeated search. Computational results on typical benchmark problems demonstrate the effectiveness and robustness of the PSOSAHT algorithm. It is worth noting that it is the first attempt to address the no-wait stochastic distributed assembly flowshop by hybrid PSO strategy.