A Case-Based Reasoning Approach to Parts Obsolescence Prediction

The ultimate aim of this research is to develop a computer system to aid in producing accurate parts obsolescence predictions. This can be accomplished through the automated use of previous experience, which consists of information on existing parts. The feasibility of an automated system which uses existing components as cases to aid the obsolescence prediction for a current part will be demonstrated by the development of a prototype with a broad corss-section of information on existing parts. The use of cases of past experience to make estimates of the current situation is referred to as Case-Based Reasoning (CBR). CBR often overcomes problems encountered with the use of traditional methods. Using a CBR strategy for obsolescence prediction is both innovative and can lead to commercial applications. Obsolescence prediction by its nature resists accurate calculations. By basing the estimate on similar cases, far greater accuracy can be achieved. The Phase I effort will focus on the design of the knowledge base structure for the information on existing components, the evaluation of several obsolescence prediction strategies based on the case base, as well as initial experimentation with user interfaces.