Diagnosability in the decomposition approach for fault location in large analog networks

Diagnosability of analog networks using Salama, Starzyk, and Bandler's fault-location algorithm is investigated. A network is said to be
n
-diagnosable if, under the assumption that the network contains at most
n
faulty subnetworks, all the faulty subnetworks can be located. A theorem for
n
-diagnosability of a network is derived.