PSO combined with ILS for flowshop-based parallel task assignment problem

Abstract

This research explores a new scheduling problem in electronics mounting assembly line: flowshop-based parallel task assignment (FSBPTA), in which the machines are lined flowingly and the tasks-including both the allocation of PCBs and their components, will be assigned to machines for mounting with the objective of minimizing the makespan. The allocation of PCBs has the characteristic of the flowshop scheduling problem (FSSP) while the assignment of the components resembles the job assignment in parallel-machine scheduling problem. To solve this problem, a hybrid algorithm method of Particle Swarm Optimization (PSO) and Iterated Local Search (ILS) is proposed. In this algorithm, ILS is used to help PSO algorithm escape from being trapped into local optima. To testify this hybrid algorithm (PSO&ILS), comparisons between PSO&ILS and PSO, PSO&ILS and multi-start descent algorithm have been advanced. Results show a 9.80% and 3.70% improvement respectively, which illustrate the effectiveness of the proposed algorithm. © 2008 Springer Berlin Heidelberg.