| The bullwhip effect may lead to tremendous inefficiency in supply chain. An important source of the bullwhip effect is due to the inventory control policy, which is originally aimed to smooth production in response to demand variation along the supply chain arisen from the customers. In the first part of the dissertation, we propose a control variate technique for reducing the bullwhip effect, or more generally the demand variability in supply chains. By this approach, a stabilizing inventory control policy is derived by imposing a correlated control variate in suitable format on the original order policy if it leads to detrimentally large variation in the demand to upstream supplier. We apply the control variate technique to some inventory models to study its effectiveness as well.; Generally, the derivation of the inventory policies requires the complete knowledge of the specific demand distribution. In the second part, we formulate an applicable and challenging inventory model in which only partial demand information is observable. Then we derive the asymptotic estimators of the first two moments of the (daily) demand by applying the renewal theory in different scenarios and propose a regression-based approximation as well to improve the estimation. A series of numerical studies are carried out to evaluate the accuracy and precision of the estimation and to investigate the impact of the estimation to the optimality of the inventory policies.; The last part of this dissertation studies a periodic review inventory system with two supply options. Regular orders are periodically placed at the review epoch. Emergency orders, characterized by shorter lead time, higher ordering and setup cost, are placed in cases of inventory level being dangerously low. Based on our assumptions on the formulated model, we formulate a dynamic programming model and prove the optimality of (s, S) type policies for both emergency and regular orders. Some analytic properties of the optimal policies are derived and presented. We gain some managerial insights into the optimal policies and cost performance from the numerical studies by applying the approach of backward dynamic programming. |
