Multi-objective assembly line balancing problem under uncertainty using genetic algorithm

Abstract

This paper addresses the single-model multi-objective assembly line balancing (ALB) problem with fuzzy processing time. A fuzzy optimisation model is formulated for the problem. In the classical mathematical formulation of SALBP, the relevant data are considered deterministic. But, the data of the real world problems are imprecise, vague or uncertain. The input data should be estimated within uncertainty represented by a fuzzy number. The concept of fuzzy numbers is introduced to treat imprecise data, such as the processing time of each task. Due to the NP-hard nature of the ALB problem, heuristics are used to solve real life cases. In this paper, an efficient heuristic is proposed to solve the fuzzy multi-objective ALB problem. The proposed heuristic is based on genetic algorithm (GA) that is efficient to handle fuzzy job time through the evolution process. As the proposed algorithm can find out the optimum solution as a trade-off among several objectives, it outperforms the existing heuristics on several test cases. The proposed algorithm can be effectively used in the design phase of an assembly line.