Conditions for Identifiability of Mixed Graphs using Simple Paths

Network Tomography is an upcoming field to identify link metrics and improve the quality of network monitoring. From linear algebra, we know that, if the number of unknown link metrics is equal to the number of links within the network, then the network is uniquely identifiable. Previous works include identifying link metrics of networks, in which the links were either undirected or directed bidirectional. By keeping in mind that the link metrics are additive and constant, the focus of this paper is on mixed networks with both undirected and directed unidirectional paths. The method employed involves dividing the network into subgraphs and obtaining the conditions of identifiability on the network without the inclusion of monitoring cycles.