A cooperative population-based iterated greedy algorithm for distributed permutation flowshop group scheduling problem

Abstract

This paper studies the distributed permutation flowshop group scheduling problem (DPFGSP) with the consideration of minimizing total flowtime (TF), which has important applications in the modern manufacturing process. Based on the characteristics of the problem, a cooperative population-based iterated greedy (CPIG) algorithm is proposed by combining the advantages of the divide-and-rule policy, population-based evolution and iterated greedy algorithm. The CPIG divides the DPFGSP into two coupled sub-problems of group scheduling sub-problem and job scheduling sub-problem, and starts with a single population for simplicity. Unlike in the traditional cooperative co-evolutionary algorithms, the two-coupled sub-problems are addressed with a certain probability that can be determined in favor of solving the whole scheduling problem. Some advanced technologies are used, including the constructive heuristics based initialization, the critical factories based destruction and construction, the new best solution based population updating mechanism. The comprehensive experimental evaluation of 810 instances shows that the CPIG algorithm performs much better than the five state-of-the-art metaheuristics in the literature which are closely related to the considered scheduling problem.