Worker allocation in U-shaped assembly lines with multiple objectives

Abstract

The multi-objective worker allocation problems of single and mixed-model assembly lines having manually operated machines in several fixed U-shaped layouts are developed. Three objective functions are simultaneously minimized, i.e. number of workers, deviation of operation times of workers, and walking time. NSGA-II, an evolutionary optimization algorithm, searches only good solutions from the population, but a selection method should also discard the below-average solutions in order to avoid producing bad solutions. The COINcidence algorithm makes use of both good and not-good solutions with reward and punishment schemes to update a better generator. After getting the first objective of minimum number of workers, the rest are compared between NSGA-II and COIN in four measures for all testing problems. Each of the computational results gives a worker good task assignment and walking path. COIN is preferable to NSGA-II in most cases.