A dynamic programming algorithm for scheduling mixed-model, just-in-time production systems

A dynamic programming procedure for determining the optimal just-in-time (JIT) production schedule for a mixed-model facility is presented. Compared to previously available optimization methods it substantially reduces computation and storage requirements, enabling practical-sized problems to be solved. However, there are still difficulties for solving very large problems. Heuristics, which have been proposed for these large problems, can now be evaluated by generating problems and comparing the schedules produced by the heuristics with the optimal schedule. This is done for two heuristics.