Cycle time enhancement by simulated annealing for a practical assembly line balancing problem

In the garment industry, assembly line balancing is one of the most significant tasks. To make a product, a manufacturing technique called assembly line is utilized, where components are assembled and transferred from workstation to workstation until the final assembly is finished. Assembly line should always be as balanced as possible in order to maximize efficiency. Different types of assembly line balancing problems were introduced along with many proposed solutions. In this paper, we focus on an assembly line balancing problem where the upper bound of the number of workers is given, tasks and workers have to be grouped into workstations so that the cycle time is minimized, the total number of workers is minimized and balance efficiency is maximized. With unfixed number of workstations and other various constraints, our problem is claimed to be novel. We propose three different approaches: exhaustive search, simulated annealing and simulated annealing with greedy. Computational results affirmed that our SA algorithm performed extremely good in terms of both accuracy and running time. From these positive outcomes, our algorithms clearly show their applicability potential in practice.