A Differential Evolution Algorithm for Simple Assembly Line Balancing

Abstract

This paper describes the application of differential evolution algorithm (DEA) to the simple assembly line balancing problem (SALBP). DEA is an evolutionary algorithm similar to a real-coded genetic algorithm for global optimization over continues spaces. The paper is concerned with SALBP type-1 whose objective is to minimize the number of workstations required to manufacture a product in an assembly line within a given fixed cycle time. Extensive experimental work over public benchmarks test problems show the effectiveness of the proposed approach. Copyright © 2005 IFAC

Differential evolution algorithm for simple assembly line balancing type 1 (SALBP-1)

“…Although, DE has been successfully applied to solve multiobjective SALBP-2 as proposed in Nourmohammadi and Zandieh [24], the only DE-based algorithm for SALBP-1 was proposed by Nearchou [18] and tested with 64 benchmark test problem instances found in the literature. Nearchou [18] constructs a solution by forming an order of items that will be assigned to workstations and then assigned them to a station according to that order.…”

“…For instances, Tabu search by Scholl and Voß [34], Chiang [6], and Lapierre et al [16]; genetic algorithm by Kim et al [15], Rubinovitz and Levitin [28], Bautista et al [5] and Sabuncuoglu et al [29]; particle swam optimization by Hamtaa et al [10]; simulated annealing by Khorasanian et al [13]; and differential evolution (DE) algorithm by Nearchou [18,20,21] and Nourmohammadi and Zandieh [24].…”

“…In this article, precedence graph will be considered in parallel, while an order of items is constructed so no infeasible solution will be formed. Detail of the decoding scheme proposed by Nearchou [18] and the proposed decoding scheme will be explained in Section 3.1.…”

“…An infeasible solution of SALBP-1 occurs when a task that is a successor of another task is assigned to an earlier workstation that a predecessor task being assigned. If the infeasible solution is formed, Nearchou [18] uses repair strategy to turn an infeasible solution into a feasible solution. The algorithm can find optimal solution in 61 out of 64 test problem instances.…”

“…Nearchou [18] constructs a solution by forming an order of items that will be assigned to workstations and then assigned them to a station according to that order. However, the ordering of items does not consider the precedence diagram so infeasible solutions may be easily formed when tasks are assigned to stations.…”