Optimization of workers and workstations in multimanned assembly lines using genetic algorithms

Abstract

Assembly lines are a production mechanism that has historically presented economic and quality benefits in organizations. However, different problems arise during their execution, among them, the issue of balancing assembly lines with multi-manned stations. This situation occurs more frequently in industrial organizations that manufacture products of medium and large dimensions, compared to a single worker per workstation model. Despite this, we find a greater tendency to study the second case in the literature. In contrast, few studies refer to the first, in addition, exhaustive search methods such as linear programming have encountered barriers due to computational complexity, so research on this problem has focused on using heuristics looking for more efficient algorithms in order to solve it. Therefore, the present work proposes a genetic algorithm that, to our knowledge, has not been used in the search to minimize the number of workers and the number of workstations for the balancing model of assembly lines with multi-manned stations. In addition, a new cost function is proposed that weights the number of workstations and workers, punishing solutions with high idle times to avoid their selection. The results of the proposed algorithm are evaluated by comparing test instances presented in the literature. The algorithm is available at <https://github.com/juanseck/GAMmALB>.