The Study On Inventory Control Problems With Censored Information For Multi-period Demand

Demand information censored (DIC) means when the demand exceeds the available inventory, the demand information can’t be fully observed because of stockout, sometimes we call this situation partly oberservation of demand. DIC is an important character of the demand of market and is very common in nowadays inventory management. At the same time, with the fierce market competition and the rapid technology development multi-period ordering policies have been wildly used. If we ignore the DIC character of demand in the multi-period odering policies that will cause such a viciou circle that the estiamtion of demand is low because of DIC which also results in stockout and then the stockout leads to DIC which makes a lower demand estimation. Today more and more scholars begin to study the inventory management under DIC, since the problem is so compelicated that the study is still at the beginning.In this paper, we use the method of dynamic progamming to study the multi-period inventory problems under DIC, and the study mainly includes two parts:the first is the two-stage ordering policies and the second is the multi-stage ordering policies. In the study of two-stage ordering policy, we establish three different models to analyze the effect of information in the Bayesian update progress, and draw a conclusion that the optimal solution of DIC update is much lower than the no update condition and a little upper than the fully information condition. In the study of multi-stage ordering policy with DIC, we analyze the different dynamic equations that whether take inventory transfer into account, and give the form of optimal solution, through a numerical example we draw a conclusion that the optimal solution of storable products considering inventory transfer is lower than the optimal solution of perishable products without inventory transfer.