Concurrent optimization in designing for logistics support.

摘要

The military community has considerable experience in the areas of procuring and managing large systems. These systems are often expected to perform their intended function over a period of several years and require an extensive support structure. Logistics Management has always been an important field within the military and is gaining recognition within private industry as well.; Studies have shown that the decisions which are made initially, during the design of the product, will determine 80% of the total system costs. In order to satisfy the overall logistics objectives, there must be a full appreciation of the impact that design decisions have on operations, maintenance, transportation and supply. This dissertation specifically addresses the links between product configuration, component choice, design of the manugistics (manufacturing and logistics) system and operational control policies in terms of production costs, support costs and system availability.; The goal of Concurrent Engineering is to consider the links mentioned above and incorporate them into the decision-making process during product design. Optimal product design, process selection and operating policy solutions can not be discovered unless the links among these decisions are quantified at the design phase of the life cycle. Concurrent optimization achieves this by addressing all of the decision variables associated with each stage of a product's life cycle and determining their optimal, or near-optimal, values at the design stage. This research presents a model which carries out such an optimization.; The contributions of this modeling approach are outlined as follows: (1) A model is constructed which links together the product design stage, the manugistics system design stage and the operating policy of a product's life cycle. (2) An optimization scheme is applied to the overall model structure to concurrently optimize the objective function criteria. (3) The final solution computed by the model is based on a bi-criteria value function formed from the two summary performance measure, Life Cycle Cost and System Availability.