Sequencing Mixed Model Assembly Lines Based on a Modified Particle Swarm Optimization Multi-objective Algorithm

Abstract

Mixed model assembly lines are attractive means of mass and large-scale series production. Determination of the production sequence for different models is a key issue in the mixed model assembly line. Particle swarm optimization (PSO) is a novel metaheuristic inspired by the flocking behaviour of birds which has be used in consecutive problems successfully. However, it's applications in the mixed model assembly line sequencing are extremely few. This paper attempts to use a modified particle swarm optimization algorithm to solve the mixed model assembly line sequencing problem in discrete space with two objectives: the total setup cost and total idle-overload cost. Compared with the original PSO, we modified the particle position representation and adapted it to the discrete code, and introduced a self-adaptive escape scheme to enhance the diversity of particles. A comparison between the basic PSO and our modified PSO show that our modified PSO algorithm is an effective sequencing method for mixed model assembly lines which possesses rich diversity.