An improved genetic algorithm with variable neighborhood search to solve the assembly line balancing problem

Abstract

Purpose

This study aims to propose an efficient optimization algorithm to solve the assembly line balancing problem (ALBP). The ALBP arises in high-volume, lean production systems when decision-makers aim to design an efficient assembly line while satisfying a set of constraints.

Design/methodology/approach

An improved genetic algorithm (IGA) is proposed in this study to deal with ALBP to optimize the number of stations and the workload smoothness.

Findings

To evaluate the performance of the IGA, it is used to solve a set of well-known benchmark problems and a real-life problem faced by an automobile manufacturer. The solutions obtained are compared against two existing algorithms in the literature and the basic genetic algorithm. The comparisons show the high efficiency and effectiveness of the IGA in dealing with ALBPs.

Originality/value

The proposed IGA benefits from a novel generation transfer mechanism that improves the diversification capability of the algorithm by allowing population transfer between different generations. In addition, an effective variable neighborhood search is used in the IGA to enhance its local search capability.

Citation

Fathi, M., Nourmohammadi, A., H.C. Ng, A., Syberfeldt, A. and Eskandari, H. (2019), "An improved genetic algorithm with variable neighborhood search to solve the assembly line balancing problem", Engineering Computations, Vol. 37 No. 2, pp. 501-521. https://doi.org/10.1108/EC-02-2019-0053