The Research Of Software Reliability Optimized-allocation Technique Application

Computer has been widely used in each field as an equipment of acquiring and handling information in present. The demand for the software is very great and the software system is more complicated. The quality of the software product has become the problem more concerned. The software reliability is an important property of software and affected the reliability of whole computer system very greatly. The software reliability allocation is one of important design technique in software reliability engineering, so it is significant to research the application of software reliability optimized-allocation technique.In this paper, we concentrate on the research and realization of the software reliability optimized-allocation technique. Such as Analytic Hierarchy Process of software reliability, software reliability allocation model based on the frequency ratio, software reliability allocation model based on the maximum system utility under the budget constraint, software reliability allocation model based on the Fuzzy Linear Programming and software reliability allocation model based on the software Fault Tree Analysis are discussed and analyzed in detail. Also it will be introduced that Genetic Algorithm has many evident advantages on solving complicated Nonlinear Programming problem.In this paper, considering the functional characteristics of launching system controlled-software for a ground-launched cruise missile (GLCM), by utilizing Analytic Hierarchy Process to quantitatively determine the relative weight value for each software component first, and then by applying the obtained datum to the above-mentioned models respectively, we can get the reliability index of each component by solving the model. Finally, the necessary parameters for the algorithm can be obtained by using Software Fault Tree Analysis, and then we can get the reliability target of the modules by performing the algorithm. Through comparing and analyzing the simulated experiment output data, we can find that the result is reasonable, realistic and reliable, and that the deviation is trifling and acceptable.