Study On Transshipments Policy And Inventory Policy For Logistics System

In the logistic system, inventory is the guarantee of satisfying services, but it's also one of the principle causes of logistic costs. Information pooling in modern logistic system makes the lateral inventory pooling and fast transshipments possible, as a result to improve service level and to reduce the total operating cost of entire logistics system as well.The study of this paper is based on the model of one upper supplier with capacitated productivity, furnishing two non identical retailers with products. SC managers apply periodic review on retailer's inventory to make replenishments decisions. The retailers in the same echelon can use preventive or emergency lateral transshipments to adjust the inventory-demand imbalance between them.The programming objective is to minimize the expectation of total system cost in limit horizon. The logistic cost is composed of unit unmet demand penalty cost, unit inventory cost, fixed replenishment cost, unit replenishment cost and unit transshipment cost.The paper will compare the three transshipments policies: preventive transshipments, emergency complete inventory pooling transshipments and emergency controlled inventory transshipment, and their influences on the total expected cost. Meanwhile, for the model with productivity capacitated and transshipment, we will study the structure of the periodic replenishments policy, observing the bands structures; and also that of transshipment policy.We use the standard backward dynamic programming method to obtain the optimal replenishments and transshipments solutions, and related expected total cost. However, the backward DP algorithm would lead to the curses of dimensionality in real problem with high stats and decisions dimensions. We therefore try to use approximate dynamic programming method to solve transshipments problem, experimenting linear approximation and square function approximation.