Variable neighborhood based memetic algorithm for Just-in-Time distributed assembly permutation flowshop scheduling

摘要

Distributed Assembly Permutation Flowshop Scheduling Problem (DAPFSP) represents a widely-applied manufacturing pattern, which is formed by two relatively independent stages, i.e., processing stage and assembly stage. In this paper, we introduce the Just-in-Time (JIT) constraint between the two stages of the conventional DAPFSP to form the Just-in-Time DAPFSP (JIT-DAPFSP), with respect to minimizing the maximum weighted earliness/tardiness cost. A variable neighborhood search based Memetic Algorithm (VNS-MA) is proposed by incorporating several novel neighborhoods manipulating the factory assignment. Computational tests are carried out on 10 small-scale and 5 large-scale benchmark problems to compare the results of different operators and algorithms. To the best of our knowledge, this is the first report to model JIT constraints and objective in DAPFSP.