HYBRID EXPERT MAINTENANCE ADVISORY SYSTEM FOR NAVAL SYSTEMS APPLICATIONS

This paper presents an approach for diagnosis and maintenance of compl ex naval systems that incorporates both rule-based and probability-bas ed reasoning. This type of diagnosis is useful in situations where exp eriential knowledge and procedures coexist with statistical reliabilit y data about the system. Effectively using the two types of knowledge is typically difficult, but in this study a blackboard-like control ar chitecture is used to successfully interface these two knowledge compo nents into an integrated diagnostic model. A case study of diagnostic maintenance of a shipboard weapon system is presented.