Optimal fault diagnosis in comparison models

In comparison models for system-level fault diagnosis, pairs of units are given the same job and results are compared. The result of such a comparison test can be 0 (match) or 1 (mismatch) and diagnosis is based on the collection of test results. Two such models have been studied, among others: the symmetric model of K.Y. Chwa and S.L. Hakimi (Inform. Control, 49, p.212-38, 1981) and the asymmetric model of M. Malek (Proc. 7th Symp. Comput. Architecture, p.31-35, May 1980). The worst-case optimal testing algorithms for t-fault detection, sequential t-fault diagnosis, and one-step t-fault diagnosis in both models are presented. Nonadaptive and adaptive testing is discussed and it is shown that the latter often enables one to decrease the number of tests