A KNOWLEDGE-BASED APPROACH TO THE EVALUATION OF FAULT-TREES

A list of critical components is useful for determining the potential problems of a complex system. However, to find this list through evalu ating the fault trees is expensive and time consuming. This paper inte nds to propose an integrated software program which consists of a faul t tree constructor, a knowledge base, and an efficient algorithm for e valuating minimal cut sets of a large fault tree. The proposed algorit hm uses the approaches of top-down heuristic searching and the probabi lity-based truncation. That makes the evaluation of fault trees obviou sly efficient and provides critical components for solving the potenti al problems in complex systems. Finally, some practical fault trees ar e included to illustrate the results. (C) 1996 Elsevier Science Limite d.