OPTIMAL POLICIES UNDER THE SHORTAGE PROBABILITY CRITERION FOR AN INVENTORY MODEL WITH UNKNOWN DEPENDENT DEMAND.
DECISION STUDIES GROUP PALO ALTO CA
Pagination or Media Count:
This report considers inventory models with a new objective function and a model with incompletely known demands which depend on the state of the inventory. The shortage probability criterion is defined, and the form of optimal policies under this new criterion is proved. Optimal policies are obtained for a particular model with unknown binomial demands when Bayesian estimation methods are used. The object of the shortage probability criterion is to minimize the order costs subject to the restriction that the probability that inventory falls below a given level not to exceed a given value. Inventory models with state dependent and independent demands and with various delivery lags are considered, and the optimal ordering policy in all cases is shown to be a myopic single critical level policy. A Bayesian estimation procedure is developed for a model with binomially distributed demands where the parameter p is unknown. The Bayesian estimate is incorporated into a myopic policy which is shown to be optimal. Author
- Logistics, Military Facilities and Supplies
- Operations Research