Models and Algorithms for Integrated Multi-Stage Production/Distribution Systems

University Of California-Berkeley, Berkeley CA

Investigators

Abstract

This grant provides funding for a rigorous analysis of several classes of multi-stage supply chains in both deterministic and stochastic settings. Novel models will be developed for the structural analysis of optimal production and distribution policies. These models will be used as a foundation for the development of new, efficient, and robust heuristics for integrated production and transportation planning. These models and heuristics will be extended to a variety of relatively complex environments, including settings featuring multi-stage serial production, multi-component assembly networks, multi-class transportation cost functions, multi-class demand, and make-to-stock and make-to-order production strategies. After exploring these models in a centralized setting in which all stages are working toward the same objective, decentralized settings where each stage strives to achieve its own objective will be considered. Loss of efficiency due to decentralization will be explored, and coordination mechanisms will be developed. These models and algorithms will be validated with industrial data, in order to test the applicability of insights and the effectiveness of algorithms. If successful, the results of this research will lead to a deeper understanding of integrated production and transportation planning in modern supply chains, and in particular, will help to explain the impact of finite capacity and distribution economies of scale in typical manufacturing supply chains. This insight, as well as the algorithms which will be developed, will lead to more efficient and effective decision-making in these supply chains, and will ultimately lead to effective heuristics for even more complex real-world models. Finally, the insights developed in this research will be incorporated into a case study and a teaching module.