The research in this dissertation involves the study of several multi-echelon inventory systems with stochastic capacity and intermediate product demand. Specifically we analyze the behavior of the system which consists of several intermediate product demands. The analysis is primarily three fold i) developed update (relational) equations for all the multi-echelon inventory systems under several inventory allocation policies, ii) develop two simulation optimization approaches 1) OptQuest framework, and 2) IPA (Infinitesimal Perturbation Analysis) framework, used to minimize the total cost of the inventory systems that satisfy the desired customer service level, iii) obtain numerical results for all the multi-echelon inventory systems under several scenarios and instances, and an extensive analysis and implications of the results.
The research done in this dissertation differ from earlier works, since it considers a complex (combination of serial and assembly systems) multi-period multi-echelon inventory system with several sources of demand (specifically intermediate product demands). We obtain the best found base-stock levels for each node in the system that satisfies the required customer service level. A SIO (Simulation based Inventory Optimization) approach is used to obtain the best found base-stock level for the system under several inventory allocation policies. We consider a system which is closer to the actual world and can be used to solve contemporary issues like, 1) manufacturing firm that produces finished products as well as spare parts, 2) manufacturer-warehouse-distribution center-retail outlets etc. I am not aware of any work that studies the impact of inventory allocation polices for multi-period in a multi-echelon inventory system, and obtains best found base stock level for each node using an IPA framework. Moreover the best found base-stock level for each node is obtained under realistic conditions like stochastic demand, stochastic capacity, and lead time.