A parallel genetic algorithm for identifying faults in large diagnosable systems

Abstract

This paper deals with the problem of fault identification in large diagnosable systems under the PMC model. Recently, genetic algorithms have been successfully used to solve this system-level fault diagnosis problem; however, they have one major drawback, i.e. even though they have been shown to perform better than the existing diagnosis algorithms, they are still time-consuming especially for large systems composed of hundreds or thousands of nodes. In this paper, we describe a new parallel version of the existing evolutionary diagnosis algorithm, which exploits competing sub-populations to speed up the diagnosis algorithm. The new approach has been implemented using the parallel virtual machine (PVM) environment and has been evaluated on a workstation network using randomly generated large diagnosable systems. Experimental results showed that the new parallel version considerably improved the response time of the diagnosis algorithm, hence, allowing for fast identification of faulty nodes.