Using Linear Support Vector Machines to Solve the Asymmetric Comparison-Based Fault Diagnosis Problem

Abstract

This paper presents a new diagnosis approach, using linear support vector machines (SVMs). The objective is to identify the set of permanent faulty nodes when at most t nodes can fail simultaneously. We consider the asymmetric comparison diagnosis model which assumes that nodes are assigned a set of tasks and their outcomes are compared. Based on the agreements and disagreements among the nodes' outputs, the diagnosis algorithm must identify all faulty nodes. The new linear SVM-based diagnosis is first trained using various input syndromes with known fault sets. Then, it is extensively tested using randomly generated diagnosable systems of different sizes and under various fault scenarios. Results from the thorough simulation study demonstrate the effectiveness of the SVM-based fault diagnosis algorithm, in terms of diagnosis correctness, diagnosis latency, and diagnosis scalability. We have also conducted extensive simulations using partial syndromes, i.e., when not all the comparison outcomes are available prior to initiating the diagnosis phase. Simulations showed that the SVM-based diagnosis performed efficiently, i.e. diagnosis correctness was around 99% even when at most half of the comparison outcomes are missing, making it a viable alternative to existing diagnosis algorithms.