ABSTRACT
Fuzzy set theory offers new methods for modeling the inexactness and uncertainty concerning decision making. The fuzzy approach improves the potential for modeling human reasoning and for presenting and utilizing linguistic descriptions (e.g. rather sure) in computerized inference. While we believe that the Fuzzy Clustering Algorithm may be generally applicable in every network scenarios. Thus, we study the performance of Fuzzy Clustering Algorithm in the context of a restricted class of faults for which we can indeed construct practical Fuzzy Clustering Algorithm. In each instance, we collect fault signatures from a fault detection system. The signatures are then input to a localization algorithm that outputs a hypothesis corresponding to a set of likely faults in the network. We use the fuzzy membership function describes the uncertainty between the sink nodes and the fault link based on the network traffic and the network topology. We apply the Fuzzy Clustering Algorithm used in the all-optical transport network, give the steps of Fuzzy Clustering Algorithm for fault localization based on the mathematical model of optical transport network and evaluate the effectiveness of Fuzzy Clustering Algorithm.
