| In recent years, organizations have recognized the large potential for savings through internal and external coordination. An example of such coordination is the proactive use of transshipments, transfer of a product among locations at the same echelon level. In this dissertation study, we consider coordination among stocking locations through replenishment strategies that explicitly take into consideration transshipments. The basic contribution of this dissertation is to incorporate capacity concerns to traditional emergency transshipment model.; Different cases of transshipment capacity are formulated as a capacitated network flow problem embedded in a stochastic optimization problem. A solution procedure based on infinitesimal perturbation analysis is developed to solve the stochastic optimization problem numerically.; In the first part, we model two different cases of transshipment capacity as a capacitated network flow problem: The focus is on the cases where the capacity is on either transshipment quantities between any two stocking locations or total outgoing transshipment from any location. After validating the solution approach from literature, the solution algorithm is experimented in different problem settings. The impact of capacity on transshipment arcs---in interaction with network structure---on the system behavior is shown. We find that transportation capacity constraints not only increase the total cost, they also modify the inventory distribution throughout the retailers. Depending on the transportation capacity, the distribution of safety stocks can change in a centralized retailer system.; In the second part of the dissertation, the transshipment model and related network flow problem is adopted to examine the case where supplier has limited production. In addition, the validity of the sample-path-based optimization procedure (based on IPA) for the setting with limited production is proved. The impact on system behavior and performance measures of the stocking locations is analyzed when the supplier may fail to satisfy all the replenishment order of locations. To deal with the demand shortfalls due to limited production capacity, an allocation rule is incorporated to the model, as well. However, inventory distribution in a production-inventory system is robust with respect to the choice of the allocation rule. |
