Determination of part-delivery dates in a small lot stochastic assembly system

In this paper the authors develop an approach for prescribing part delivery dates in a small lot stochastic assembly system where parts are assembled at each station to a subassembly. The part delivery and processing times at assembly stations follow known probability distributions. The buffers between assembly stations may be of limited capacity. The objective is to minimise the expected total cost of subassembly and part inventory, facility usage, finished goods inventory and delivery tardiness. A heuristic solution technique based on scheduling the individual parts by optimizing single stations in isolation is developed. Results have verified the effectiveness of this approach over all generalizations and it yields good solutions with very little computational effort. The approach is then used to generate insights regarding the effect of system parameters, randomness in delivery and processing times, the distribution used, and buffer limitations on part delivery dates and overall system behaviour.