The robotic assembly line balancing problem under task time uncertainty

Abstract Consideration is given to the robotic assembly line balancing problem (RALBP) under uncertain task (operation) times, a critical challenge encountered in automated manufacturing systems.. RALBP is a decision problem which seeks the optimal assignment of the assembly work as well as the most suitable robots to the workstations of the assembly line with respect to objectives related to the capacity of the line or/and its cost of operation. When multiple types of robots with different capabilities are being used, task times may vary depending on robot type and the nature of the task. Task variation is expected to be small for simple tasks but may be quite large for complex and failure sensitive operations. To deal with uncertainty in task variation we used fuzzy logic theory. First, we introduce formally the fuzzy RALBP and then we describe deeply the fuzzy representation of the task times. We address RALBP with respect to two optimization objectives namely, the production rate and workload smoothing . Since the problem is known to be NP-hard, we explore its heuristic solution through a new robust multi-objective genetic algorithm (MOGA) aiming to determine the Pareto optimal set. Simulation experiments assess MOGA’s efficiency in comparison to the famous NSGA-II and MOPSO algorithms, while also exploring the trade-off between the two conflicting objectives.