Sequential Inventory Control and Optimization through Stochastic Approximation.
Abstract
Optimal inventory policies are typically characterized by a stationary post-order inventory level S, which is the level up to which an order is placed including both inventory on hand and on order. Some Inventory systems have explicit solutions presented in the literature and others have solutions characterized by dynamic programming equations. All of these solutions require a knowledge of the distribution function for demand when it is a random variable. If one has no knowledge of the distribution function one is able to formulate certain stochastic approximation techniques which 'hone-in' on the optimal policy values using observations of demands and costs as they occur. These techniques all assume convergence in the mean square and with probability one to the optimal values. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1972
- Accession Number
- AD0741451
Entities
People
- Thomas R. Tower
Organizations
- Air Force Institute of Technology