A new locally t-diagnosable structure under the PMC model with an application to matching composition networks

Abstract

The PMC model is the test-based diagnosis in which a node performs the diagnosis by testing the neighbor nodes via the links between them. If we concentrate on the status of some nodes then instead of doing the global test, Hsu and Tan proposed the concept of local diagnosis and two structures to diagnose a node under the PMC model. To better evaluate the local diagnosability of a node, we propose a new structure and the related algorithm to diagnose a node under the PMC model in this paper. Applying the two structures proposed by Hsu and Tan, and the new structure we propose in this paper, we determine the accurate value of the local diagnosability of each node in matching composition networks. Simulation results are presented, showing the performance of our algorithm. It shows that even if the failure probability of a node is 0.4, our algorithm can still determine the state of a node with the accuracy above 0.9.