This paper considers a two-echelon distribution system with N local distribution centers (LDCs) and a single central distribution center (CDC). The purpose is to determine appropriate stock levels and best ordering policies for the system. Each LDC adopts a continuous review (Q, r) policy. On the other hand, the CDC assumes a periodic review (R, T) policy. When a stockout occurs, customers with a known fraction β are willing to wait for their orders to be filled. The remaining fraction 1-β of customers who are impatient to wait will result in lost sales.
In the proposed model, the two echelons are linked by effective lead time, which includes the nominal lead time and the expected delay due to the CDC being out of stock. This lead time reflects interactions between the two echelons. A heuristic algorithm based on a numerical analysis approach is developed to find near-optimal ordering policies that minimize the average total cost per year of the whole distribution system. Computational process and result are given by illustrating an example. Through the analysis of the experimental results, several important conclusions are drawn.
英文關鍵字
Multi-echelon Distribution System, Backorder, Lost Sales, Effective Lead Time