Assembly line design with incompatible task assignments

Abstract Assembly lines are widely used for the mass production of consumer goods and components in large volume production systems. Design of these lines warrants taking into consideration not only cycle time and precedence constraints, but also other restrictions. An important recurring restriction is that some pairs of tasks cannot be assigned to the same station due to factors such as safety, physical demands placed on workers, quality, and technological considerations. We investigate this problem and a current industry practice used for solving this problem. Our investigation of this problem yields three findings. First, we identify a new class of heuristic procedures which dynamically updates task priorities. Our investigations show that this class of heuristics yields better results. Second, extremely greedy procedures such as knapsack heuristics (Hoffmann procedure) continue to perform better than competing heuristics for industrial grade problems even when additional restrictions on task assignments are present. However, when task restrictions are severe, this is no longer true. Lastly, we present an enumerative procedure to search for optimal solutions in the presence of restrictions. Our study shows that while most simple assembly line balancing problems can be solved optimally, presence of additional restrictions such as task assignments makes them inherently more difficult. We provide insights into this aspect.