Evaluation of Efficiency of Application of Functionally Sustainable Generalized Information System of the Enterprise

Abstract

The paper evaluates the effectiveness of the use of functionally stable generalized information system of the enterprise. Diagnosis of the system in general. The following definition-assumption was used in the diagnosis: the amount of diagnostic information is the number of test results accumulated by the module before the execution of the algorithm, recorded during the mathematical modeling of the diagnostic procedure. It is illustrated that the reliability of the diagnosis does not change due to the expansion of the system. This is due to the property of adaptive self-diagnosis - the ability to diagnose with a given reliability. Values of reliability significantly depend on the number of failures. As the number of failures increases, the reliability of diagnosis decreases, due to the peculiarities of the algorithm of analysis of test results, in which the whole set of modules is divided into two subsets - possible good and possible bad modules. It is the power of many possible faulty modules that influences decision-making, ie the reliability of diagnosis. Dependence of reliability on a priori probability of working condition of modules shows insignificant increase of reliability with increase of value of probability of working capacity that is also caused by features of algorithm of analysis of diagnostic information in which the decision on failures of modules is made directly taking into account value of probability of working capacity. The amount of diagnostic information is the number of test results accumulated by the module before executing the algorithm. The property of adaptive self-diagnosis, in contrast to the existing ones, is considered in the work due to the increase in the time of diagnosis and the acceptable decrease in the reliability of the diagnosis. At the same time, the time of diagnosis is predictable and is within acceptable limits, which allows to detect failures in time, roll back the computational process and ensure the insensitivity of the computational process to failures of system modules.