Uniform TitlePerformance analysis and design of batch ordering policies in supply chains
NameKaraman, Abdullah Siddik (author), Altiok, Tayfur (chair), Boucher, Thomas (internal member), Chaovalitwongse, Wanpracha (internal member), Melamed, Benjamin (outside member), Rutgers University, Graduate School-New Brunswick,
SubjectIndustrial and Systems Engineering,
Supply and demand,
Physical distribution of goods
DescriptionDevising manufacturing/distribution strategies for supply chains and determining their parameter values have been challenging problems. Linking production management to stock keeping processes improves the planning of the supply chain activities, including material management, culminating in improved customer service levels. In this thesis, we investigate a multi-echelon supply chain consisting of a supplier, a plant, a distribution center and a retailer. Material flow between stages is driven by reorder point/order quantity inventory control policies. We develop a model to analyze supply chain behavior using some key performance metrics such as the time averages of inventory and backorder levels, as well as customer service levels at each echelon. The model is validated via simulation, yielding good agreement of robust performance metrics.
The metrics are then used within an optimization framework to help design the supply chain by calculating optimal parameter values minimizing the expected total cost. Optimal design of the material flow system is part of the overall planning and operation of a supply chain. The outcome of the optimization framework specifies not only how much and where to hold inventory but also how to move inventory across the supply chain.
The developed model requires limited computational requirements, which in turn helps frequently update the performance measures and optimal system parameters so as to be more responsive to short-term changes in demand or supply. In addition, it can be used as a decision support system for effective decision making as opposed to using simplistic inventory models, which results in significantly higher operating costs.
In a similar vein, we consider a distribution inventory system with one warehouse and several retailers. The challenge in this system is to describe the demand arrival process at the warehouse. We propose a procedure to characterize the demand arrival process at the warehouse as a superposition of several independent Erlang processes. An important characteristic of the superposed process is that although the individual processes are independent from each other, the superposed process may be no longer independent. We present a methodology to characterize such arrival streams as Markovian processes. We, then, extend the methodology to phase-type arrival streams as well.
Note[bibliography] Includes bibliographical references (p. 132-138).
CollectionGraduate School - New Brunswick Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.