Davydov A. Automating the generation and processing of production rules for diagnosis of computer systems

The object of the research is the functional diagnostic processes of the fault conditions of computer systems and networks. The aim of the research is the develop methods of functional diagnostics of computer systems and networks with using structural modeling level, aimed at the effectiveness increase of diagnosis process for account of the intellectual tools improvement forming of expert knowledge and its realization in the fuzzy expert systems. The methods of the research are the set theory, the graph theory, and finite automata theory for the modeling of computer systems conditions and minimization of production rules; the technical diagnosis of digital systems and fuzzy logic for the develop of tracing methods. The equipment - personal computer. Theoretical and practical results of the research - the actual scientific and technical problem of the develop of methods of computer diagnosis system, aimed at the rise of the diagnosis process efficiency for the account of methods of knowledge presentation in the expert systems, that provides the increase of the computer system availability and decrease the time for the research of probable disrepair in the computer systems. The scientific novelty- for the first time the method of transformation of the free verbal description of the computer system defects by way of intellectual fault dictionaries in the logical production rules with using of the sentential calculus system was proposed, that allowed to formalize the description of computer system fault condition by way of the decision tree; for the first time with the using of decision tree and the theory of switching functions the synthetic method of diagnosis production rules was proposed, that foreseen the stages of analysis of organic number of diagnostic variable and making decision if the incomplete or contradictory information is present, that allow to diagnose the fault condition of computer system with big speed in real time; the synthesis procedure of the production rules for ability to determine the fault free functional condition of computer system was improved, which uses binary relations of means of diagnosis parameters features in comparison with legacy tables of fault condition functions, that allow to reduce the number of diagnosis production rules; that developed the method of computer system diagnosis with using of fuzzy logic by way of describing the condition and classification diagnosis features by way of fuzzy variable, that allowed to reduce the number of production rules and to increase the speed of expert diagnosis system. The results of research has a practical application in the PrAT "Severodonetskoe NPO Impuls" for diagnosis of technical condition of specialized computer cluster. Also the scientific applications and results of the research are used in the department of design automation of computer technology of Kharkiv national university of radio electronics with preparation of academic subjects from the projecting and diagnosis of computer systems and networks. Scientific and practical results of the research can be used in the scientific and technical designing and in designing of numeral system with diagnosis and restoration of software and hardware components functionality of computer system and rates.