Exploiting System Hierarchy to Compute Repair Plans in Probabilistic Model-Based Diagnosis

The goal of model based diagnosis is to iso late causes of anomalous system behavior and recommend cost e ective repair actions In general precomputing optimal repair policies is intractable To date investigators address ing this problem have explored approxima tions that either impose restrictions on the system model such as a single fault assump tion or that compute an immediate best ac tion with limited lookahead In this paper we develop a formulation of repair in model based diagnosis and a repair algorithm that computes optimal sequences of actions This optimal approach is costly but can be applied to precompute an optimal repair strategy for compact systems We show how we can ex ploit a hierarchical system speci cation to make this approach tractable for larger sys tems When introducing hierarchy we also consider the tradeo between simply replac ing a component and decomposing it to repair its subcomponents The hierarchical repair algorithm is suitable for o line precomputa tion of an optimal repair strategy A modi cation of the algorithm takes advantage of an iterative deepening scheme to trade o in ference time and the quality of the computed strategy