A Machine Learning Approach for Self-Diagnosing Multiprocessors Systems under the Generalized Comparison Model

Support Vector Machines (SVMs) have been successfully applied to pattern recognition, regression, and classification. Because of their good performance and their mathematical foundations, SVMs are gaining popularity in solving various diagnosis problems. In this paper, we introduce a novel approach using a SVMs to solve the system-level fault diagnosis problem under the generalized comparison model (GCM). The GCM assumes that a set of tasks is assigned to pairs of nodes and their outcomes are compared by neighboring nodes. Given that comparisons are performed by the nodes themselves, faulty nodes can incorrectly claim that fault-free nodes are faulty or that faulty ones are fault-free. The collections of all matches and mismatches, i.e., The comparison outcomes, among the nodes are used to identify the set of permanently faulty nodes. First, we show how SVMs can be adapted to the GCM-based diagnosis problem. Then, from the results of an extensive simulation study we show that the new diagnosis approach succeeded in identifying all faulty nodes in the faults situations considered under t-diagnosable systems. The simulations demonstrate that the SVM-based diagnosis approach remarkably identified all faulty nodes, with a diagnosis correctness of 100% and with very low diagnosis latencies, providing hence an effective solution to the system-level self-diagnosis problem.