DefinePK

Abstract

In this study, the performance of software reliability is evaluated by applying the Log-Type (Log-Poisson, Log-Power, Log-Logistic) distributions to the infinite fault NHPP model. For this study, a research solution was developed by an analysis algorithm using software failure time data. Also, maximum likelihood estimation (MLE) was applied to solve the parameter values, and the nonlinear equation was calculated using the binary method. As a result, first, in the strength function analysis, the Log-Logistic model was evaluated as an efficient model because the failure rate showed the greatest decrease as the failure time passed. Second, in the mean value function analysis, the proposed models showed either underestimation or overestimation in estimating the true value, but the Log-Logistic model was effective as it showed the smallest error estimation along with the Log-Poisson model. Third, in the future mission reliability analysis, the Log-Logistic model was evaluated to show a higher reliability trend than other models whose reliability gradually decreased as the mission time passed. In conclusion, by analyzing these evaluation results, it can be seen that the Log-Logistic model is an efficient model with the best performance among the proposed models. Through this study, the reliability performance of the Log-Type distribution model was newly compared and evaluated, and it was able to help software developers to find the optimal software reliability model.