Adjusting Replenishment Orders to Reflect Learning in a Material Requirements Planning Environment
针对高科技制造业中初始良率低但随经验提升的特点，将学习曲线融入MRP逻辑，通过实验证明在低良率环境下考虑学习效应可大幅降低平均库存水平，且对学习率估计误差具有稳健性。
Some manufacturing firms, particularly in the high-technology sector, have production processes which are characterized by very low initial yields followed by steady “experience” based yield improvement. Material Requirements Planning literature reveals that MRP implementations are seldom adjusted in any systematic way to account for such yield improvement. A single product, single stage MRP model is developed which incorporates learning curve behavior into conventional MRP logic. A series of experiments systematically examine the impact on mean inventory level of various combinations of environmental conditions and managerial policies. The research demonstrates that substantial reductions in mean inventory levels can be realized in low yield environments if learning is properly included in the order release logic. This finding proves to be robust with respect to modest errors in the estimation of learning rate.