9414680 Sen This research addresses the problem of multi-stage decision making with uncertainty in a long-range planning process, especially those that are typically undertaken by industry in planning for the entire life cycle of a product. The work is aimed at developing some algorithms to extend earlier work in stochastic decomposition models for two stage problems. This research extends the two-stage problem to a problem involving more than two stages. Stochastic decomposition SD uses randomly generated observations within Benders, decomposition of a two stage stochastic linear program to identify an optimal solution (asymptotically). Many key decisions aimed at improving a company's global competitiveness involve the integration of several aspects planning and design. The techniques developed as part of this research will provide the tools necessary to integrate several decision elements of a company in designing high quality and long lasting products while at the same time keeping cost down. In this respect, the developments from this research can find applications in such areas as concurrent engineering, total quality management, and production planning and control. Furthermore, the results from the research could provide some insight to help bridge the gap between stochastic and deterministic Operations Research.
