Research On Adaptive Inventory Control Of Non - Stationary Stochastic Demand Chain

A supply chain is a network system composed of some closely related units such as suppliers, manufacturers, retailers and customers. Supply chain management is practiced by more and more enterprises with expect to achieve horizontal integration among each other, thus to improve the speed of response to market demands by cooperation. In a supply chain, inventory plays an important role and directly affects not only the cost but also the efficiency of the whole supply chain. In a supply chain environment, inventory is not only the thing considered by some one supply chain member but also closely related to every member in the chain. Therefore, it must be coordinated properly in order to reach the whole optimum.Adaptive inventory control is a new idea occurred in the inventory control field during recent years. It is based on the idea of adaptive control theory to realize inventory control under uncertain environment. It’s main idea can be expressed as following: for a multiechelon multi-cycle inventory control problem with uncertain customer demand, the inventory control policy can be regulated according to the real effect after periodical review thus to reach the given object.This paper focuses on the adaptive inventory control problem in a multi-echelon multi-cycle supply chain with non-stationary stochastic demand. It mainly studied two problems. Two adaptive inventory control models, i.e. a centralized model and a decentralized one were established for a two-echelon supply chain with on supplier and one retailer with(without) information sharing respectively. EOQ-based order quantity was used to determine the inventory replenishment time and a reinforcement learning was used to determine the safety stock for the first situation. The model with the order time and order quantity as the control parameters will improve the service level, and mitigate the bullwhip effect. Finally, verified information sharing can effectively mitigate the bullwhip effect. However, the simulation data show that in the process of information sharing the bullwhip effect still exists. Base on the causes, on the condition of information sharing, a fuzzy exponential smoothing method was proposed to forecast the future demand. Simulation results show that the proposed control method can improve the service level and reduce the bullwhip effect to some extent.