A Multi-Manned Assembly Line Worker Assignment and Balancing Problem With Positional Constraints

One of main operational issues in assembly lines is to assign tasks and workers to workstations so that the maximum productivity can be achieved. Many previous studies have addressed this problem, but most of them have assumed one worker in each workstation or multiple workers with the same skill level in a workstation. In this study, we consider the worker assignment and balancing problem for a real automotive parts assembly line where multiple workers simultaneously perform different tasks in a workstation and the processing time of each task varies by the worker. We also introduce two new positional constraints to secure each worker's working space, required from a company. The objective is to minimize the cycle time. We propose a filtered beam search (FBS) algorithm that adjusts beam and filter widths depending on the gap between an upper bound and a lower bound to quickly obtain the efficient assignment for large-sized instances. Several dominance rules for eliminating redundant nodes and lower bounds for selecting promising nodes are developed for the FBS algorithm. The FBS algorithm is compared to a branch and bound algorithm and a simulated annealing algorithm.