Periodic review inventory control with fluctuating purchasing costs

We consider the periodic review inventory control problem in which the purchasing cost of the product changes, in a Markovian fashion, from one period to the next. After establishing (with and without the non-speculation assumption) that an order upto policy is optimal, we develop an efficient recursive solution procedure to compute the optimal levels. In addition, we propose a measure for the magnitude of the cost fluctuations and show that this measure is an accurate indicator of the effectiveness of myopic heuristics. These results are validated using a computational study.