An effective mixed-model assembly line sequencing heuristic for Just-In-Time production systems

The commonly accepted objective of sequencing a mixed-model assembly line for a Just-In-Time (JIT) production system is to keep part usage rates at the assembly line close to constant. Sumichrast and Russell evaluated four existing sequencing methods and showed that the best method in achieving this objective is Miltenburg’s algorithm 3 using heuristic 2 (M-A3H2). This paper presents a much simpler heuristic procedure compared to M-A3H2. The experiment on 9 problem sets conducted by Sumichrast and Russell is repeated for the new heuristic procedures and M-A3H2. The results show that the new heuristic procedure is as good as M-A3H2 in solution quality regarding the mean squared and absolute deviations and much more efficient in computation time. It is futther noticed that the computation time required by M-A3H2 increases much faster than the new heuristic procedure as the number of products increases. Nevertheless, the most important benefit of this new heuristic procedure is its simplicity which should make it very desirable for practitioners.