Pareto based artificial bee colony algorithm for multi objective single model assembly line balancing with uncertain task times

Abstract

Assembly line balancing is significant for efficient and cost effective production of the products and is therefore gaining popularity in recent years. However, several uncertain events in assembly lines might causes variation in the task time and due to these variations there always remains a possibility that completion time of tasks might exceed the predefined cycle time. To hedge against this issue, a single model assembly line balancing problem with uncertain task times and multiple objectives is presented. Current research is aimed to minimize cycle time in addition to maximize the probability that completion time of tasks on stations will not exceed the cycle time and minimize smoothness index simultaneously. A Pareto based artificial bee colony algorithm is proposed to get Pareto solution of the multiple objectives. The proposed algorithm called Pareto based artificial bee colony algorithm (PBABC) introduces some extra steps i.e., sorting of food sources, niche technique and preserve some elitists in the standard artificial bee colony algorithm (ABC) to get Pareto solution. Furthermore, the effective parameters of the proposed algorithm are tuned using Taguchi method. Experiments are performed to solve standard assembly line balancing problems taken from operations research (OR) library. The performance of proposed PBABC algorithm is compared with a famous multi objective optimization algorithm NSGA II, in literature. Computational result shows that proposed PBABC algorithm outperforms NSGA II in terms of the quality of Pareto solutions and computational time.