An inventory optimization model with Markov-modulated commodity prices
Classic inventory models embody optimal cost constructs under the constraints of lead times, ordering and carrying costs, stock-out penalties, price-to-customer-fluctuations, and stochastic demand. While appropriate to finished goods or work-in-process inventory with deterministic costs, classic models do not speak to the main source of uncertainty for a raw material procurement manager; that of uncertainty in the purchase price of the commodity. This approach to the procurement manager's inventory problem is anchored by two elements: a simplification of the stochastic nature of prices and a natural order-up-to policy. Even under such considerations, there is complexity in formulating the model and devising a method for computing the optimal solution. Given a set of order-up-to thresholds, this study yields a method for finding the invariant distribution of the new process tempered by the underlying price Markov Chain and the procurement approach. Then, once the stationary distribution is known, the set of thresholds is found for optimizing the long-run performance of the strategy
Thesis, Dissertation, English, ©2004