Hybrid Particle Swarm Optimization with parameter selection approaches to solve Flow Shop Scheduling Problem

Abstract

A Hybrid Particle Swarm Optimization (HPSO) with parameter selection approaches is proposed to solve Flow Shop Scheduling Problem (FSSP) with the objective of minimizing makespan. The HPSO integrates the basic structure of a Particle Swarm Optimization (PSO) together with features borrowed from the fields of Tabu Search (TS), Simulated Annealing (SA). The algorithm works from a population of candidate schedules and generates new populations of neighbor and cooling schedules by applying suitable small perturbation schemes. Furthermore, PSO is very sensitive to efficient parameter setting such that modifying a single parameter may cause a considerable change in the result. Another two classes of new adaptive selection of value for inertia weight and acceleration coefficients are introduced into it. Extensive experiments on different scale benchmarks validate the effectiveness of our approaches, compared with other well-established methods. The experimental results show that new upper bounds of some unsolved problems and better solutions in a relatively reasonable time. In addition, proposed algorithms converge to stopping criteria significantly faster.