Inventory management with multiple procurement modes

Inventory systems with multiple procurement modes are widely used in industry and have become more popular with the growth of the globalization and e-commerce. In this dissertation we explore the stocking strategies for periodically reviewed production/inventory systems with multiple procurement modes, where each procurement mode is associated with a different lead time and different unit ordering costs.; In the first part, we investigate periodic-review discounted and average cost procurement system with deterministic lead times and no economies of scale. A streamlined analysis of the two procurement mode problem with independent and identically distributed demands for the discounted cost case is presented, and an efficient algorithm to compute the optimal policy parameters for the average cost case is developed. Also sufficient conditions under which a myopic base stock policy is optimal are developed when demands are not stationary. We also extend the analysis to the case of three or more procurement modes and develop a heuristic algorithm for the average cost case. These results not only offer guidelines for the operation of the inventory systems, but also provide valuable insight into the complex cost tradeoffs when more than two procurement modes are available.; The second part of the dissertation extends the results to multiple-procurement-mode inventory systems with stochastic lead times. Optimal stationary ordering policies for the cases of two procurement modes, one of which has stochastic lead times, are developed. Sensitivity analysis is carried out to reveal how the optimal order quantities and total average cost change with the parameters of the problem. We also extend these results to an inventory system with three procurement modes.; The third part considers the finite-horizon inventory problem with two procurement modes in which there are economies of scale in ordering. The optimality of (s, S) policies for both the express mode and the regular mode is established. We derive explicit formula for optimal order-up-to levels when the set-up cost for the regular mode is zero. We also show that when only the set-up cost of the express mode is zero, the optimal ordering policies are more complicated.; The final part of the dissertation explores inventory systems with two demand classes. At the end of every period the inventory is issued to satisfy high priority demand, followed by low priority demand. We consider the quantities of sales to each demand class as decision variables. Also we show that an (s, S) policy is optimal under this discretionary sales policy scheme.