Abstract: In this study, the Exponential-type lifetime distribution which is often used in reliability testing as a continuous probability distribution for the elapsed time of software failure occurrence was applied to the NHPP model, and the reliability performance of the proposed model was newly compared and analyzed. For this, a solution for analyzing reliability performance was developed according to the order of the analysis algorithm presented in this study by utilizing the failure time data collected during the operation of the software system. In this process, the parameter estimation was solved by applying the maximum likelihood estimation (MLE), and the nonlinear equation was solved using the bisection method. In the criteria analysis for judging by comparing efficient models, the Exponential-exponential and Exponential-based models showed relatively good results. As a result of analyzing reliability performance, first, in the estimation power analysis on the real value using the mean value function, the Exponential-exponential model showed a slight overestimation error but showed excellent performance with the smallest error value. Second, in the analysis of the failure strength using the intensity function, the Exponential-exponential model was evaluated as an efficient model because the failure rate showed a large decrease with the smallest value. Third, in the reliability analysis, the Exponential-exponential model was evaluated as showing a higher and more stable reliability trend than other models whose reliability decreased as the mission time passed. In conclusion, it can be seen that the Exponential-exponential model is an efficient model with the best performance among the proposed models. Through this study, the Exponential-type distribution model was newly analyzed, and the related data is expected to be utilized as basic design data required for software developers.

Key words: Burr-Hatke-exponential, exponential-basic, exponential-exponential, exponential-type distribution, inverse-exponential, reliability performance