An exact method to incorporate ergonomic risks in Assembly Line Balancing Problems

Abstract

Since ergonomic risks related to the tasks performed by each worker are generally neglected when solving the Simple Assembly Line Balancing Problem (SALBP) or partially treated in post-processing, this paper aims to incorporate ergonomic issues natively in the decision process (ErgoSALBP). For such a problem, we present a novel Mixed-Integer Linear Programming (MILP) model that linearizes and incorporates into the SALBP the Occupational Repetitive Action (OCRA). This well-known method gives ergonomic risk indicators for musculoskeletal disorders. A lexicographical approach to address the problem is proposed, where the minimum number of stations is obtained for SALBP and used as a starting point to solve a type-F ErgoSALBP (feasibility problem) focused on satisfying the maximum ergonomic risk workload. Once the exact solution is found, we establish an objective function to smooth the ergonomic risk distribution among stations. We tested a total of 2160 instances with up to 297 tasks. Our method was able to solve 94% of all instances optimally. The share of solutions found with only green stations (the ideal ergonomic condition) was 87%, surpassing the state of the art. For the instances where the model did not converge to a solution within the time limit, we present a procedure to assess the optimality gap. Furthermore, the model can be embedded as a building block for other assembly line configurations. Finally, the results provide trade-off insights for managers: the average OCRA for the line improves as the cycle time increases, which may positively impact workers’ satisfaction and product quality in the long run.