A novel discrete particle swarm algorithm for assembly line balancing problems

Abstract

Purpose

The purpose of this paper is to develop a feasible sequence-oriented new discrete particle swarm optimization (NDPSO) algorithm with novel particles’ updating mechanism for solving simple assembly line balancing problems (SALBPs).

Design/methodology/approach

In the NDPSO, a task-oriented representation is adopted to solve type I and type II SALBPs, and a particle directly represents a feasible task sequence (FTS) as a permutation. Then, the particle (permutation) is updated as a whole using the geometric crossover based on the edit distance with swaps for two permutations. Furthermore, the fragment mutation with adaptive mutation probability is incorporated into the NDPSO to improve exploration ability.

Findings

Case study illustrates the effectiveness of the NDPSO. Comparative results between the NDPSO and existing real-encoded PSO (CPSO) and direct discrete PSO (DDPSO) against benchmark instances of type I SALBP and type II SALBP show promising higher performance of the proposed NDPSO.

Originality/value

A novel particles’ updating mechanism for FTS-encoded particle is proposed to solve the SALBPs. The comparative results indicate that updating of FTS as a whole seems superior to existing updating of FTS by fragment with respect to exploration ability for solving SALBPs. The novel particles’ updating mechanism is also applicable to generalized assembly line balancing problems.

Citation

Dou, J., Li, J. and Zhao, X. (2017), "A novel discrete particle swarm algorithm for assembly line balancing problems", Assembly Automation, Vol. 37 No. 4, pp. 452-463. https://doi.org/10.1108/AA-08-2016-104