Full Article: PDF
Scientific Object Identifier: http://s-o-i.org/1.1/TAS-09-53-16
DOI: https://dx.doi.org/10.15863/TAS.2017.09.53.16
Language: Russian
Citation: Dimitriev YK (2017) ON THE USE OF BOOLEAN FUNCTIONS FOR COMPARATIVE ANALYSIS OF MODELS OF UNRELIABLE TESTS, AND THE DEVELOPMENT OF ALGORITHMS FOR SELF-DIAGNOSIS OF MULTIPROCESSOR COMPUTING SYSTEMS. ISJ Theoretical & Applied Science, 09 (53): 101-112. Soi: http://s-o-i.org/1.1/TAS-09-53-16 Doi: https://dx.doi.org/10.15863/TAS.2017.09.53.16 |
Pages: 101-112
Published: 30.09.2017
Abstract: Self-diagnosis of multiprocessor computing systems at system level is examined, when multiple faults occurs and under using of the unreliable tests. The state of the system is determined by the results of the comparative analysis of the outcomes of mutual testing of the system's modules. The generalization of the well-known PMC-model of the system is used, that allows to study diagnostic properties of the system as function of properties of the used unreliable tests. For a formal description of the tests, the Boolean functions (called as consistency functions) are used, the submission of which is adapted to the solution of the self-diagnosis problem. The set of basic (elementary) consistency functions is selected, the combination of which describes different test models from the exhaustive group of unreliable test models. The indicators of the worth of elementary consistency functions and their compositions are introduced and studied. The possibility of using of consistency functions and their worthy characteristics for comparative analysis of test outcomes, in particular, in determining the states of the modules of the system according to the fragments of the syndrome that refers to a limited neighborhood of the considered modules (the local self-diagnosis) is shown.
Key words: Local self-diagnosis on system level, theoretical-graph diagnostic model, models of unreliable test, Boolean consistency functions, worthy of consistency functions.
|
