Optimal Allocation of Inventory and Demand for Managing Supply Chain Revenues

This dissertation focuses on three distinct yet related problems that are motivated by practices of electronics manufacturers, who satisfy stochastic demand from multiple markets and multisource parts from several suppliers. The first problem investigates joint replenishment and allocation decisions for a supplier who satisfies stochastic demand from a primary market and a spot market. We formulate the problem as a multi-period stochastic dynamic program and show that the optimal policy is characterized by two quantities: the critical produce-up-to level and the critical retain-up-to level. We establish bounds for these two quantities, discuss their economic interpretation, and use them to construct a new and effective heuristic policy. We identify two practical benchmark policies and establish thresholds on the unit revenue earned from the spot market such that one of the two benchmark policies is optimal. Using a computational study, we quantify the benefits of the optimal policy relative to the benchmark policies and examine the effects of demand correlation.

The second problem investigates an important extension where a supplier faces stochas- tic demand from Class 1 along with price-sensitive stochastic demand from Class 2. We investigate the supplier?s joint replenishment, allocation and pricing problem by formulating it as a multi-period, two-stage stochastic dynamic program. We show that a dynamic pricing policy is optimal at stage 2, and the stage 1 optimal policy is characterized by two quantities: the critical produce-up-to level and the critical amount of inventory to be protected from Class 1. In contrast to the optimal policy, myopic policies are less costly to evaluate, and hence, are more practical. We establish two sufficient conditions under which a myopic joint inventory and pricing policy is optimal. Using a computational study, we show that the benefits of dynamic pricing to Class 2 are higher than the benefits of discretionary sales to Class 1.

While the first two problems consider a supplier?s decision under stochastic demand from multiple markets, the third problem considers decisions of a buyer who satisfies stochastic demand by multi-sourcing parts with percentage supply allocations (PSAs). We define PSA as a pre-negotiated percentage of a multi-sourced part?s total demand that the buyer should allocate to a supplier. During recent industry collaboration, we observed that in such settings the buyer?s demand allocation decisions are challenging due to operational changes needed for (temporarily) switching suppliers, and lead to the bullwhip effect. Demand allocation policies that can meet PSAs and the resulting bullwhip effect have not been investigated in the literature before. We contribute to the existing literature by introducing and analyzing the concept of bullwhip effect under multi-sourcing. In addition, we propose and investigate three demand allocation policies: (i) random allocation policy (RAP), which benchmarks the current practice, (ii) time-based (CCP-T) and (iii) quantity-based cyclic consumption (CCP-Q) policies. We show that while RAP and CCP-T always lead to bullwhip effect, the bullwhip ratio under CCP-Q can be less than 1. We demonstrate that CCP-T and CCP-Q can reduce the supplier?s bullwhip effect without increasing the buyer?s expected long-run average number of supplier switches compared to RAP.