Probabilistic syndrome decoding in self-diagnosable digital systems

The self-diagnosis of multi-unit digital systems is reconsidered assuming the hypothesis that: (i) faults are not equally probable and (ii) the outcome of tests performed between units are not deterministic. Given a set of test results, the problem of finding the most likely set of faulty units of a system is considered here and an approach to probabilistic diagnosability with repair is presented. it is shown that a significant class of systems exists for which these problems are easily solved and a decoding procedure is given whose complexitY is 0(n) where n is the number of system units.