A Gradual Approach for Model-Based Diagnosis

This paper investigates the decomposition of diagnosis problem, gives a theorem for decomposition and combination of the diagnosis. On the basis of above works, an algorithm using gradual approach to decompose the diagnosis problem is proposed. Besides, the correctness, completeness and complexity of the algorithm are proved in this paper. The experimental results indicate that our algorithm can evidently improve the effectiveness of diagnosing multi-output system. Comparing with the method of decomposition by assuming instantiations of some variables, our algorithm is more effective and applies to more general diagnosis problems.