Abstract:Software reliability engineering has been developed to examine software systems over the years. Non-homogeneous Poisson process (NHPP) models are widely employed for analyzing software reliability data, as they have the ability to describe the expected number of failures over time using the mean value function. In addition, these models have an intensity function representing the failure rate over time. This article proposes a new non-homogeneous Poisson process model based on extended log-logistic (NHPP ELL) distribution with three parameters. The model properties were derived and represented graphically. Furthermore, the maximum likelihood estimation method was utilized to estimate the model parameters, which were computed numerically using the Newton–Raphson method. Three real software reliability datasets were utilized to evaluate the performance of the new NHPP ELL model based on three distinct criteria: the mean square error, the coefficient of determination (R2), the Theil statistic, the predictive-ratio risk, and the predictive power. Moreover, a comparative study was conducted between the proposed model and several well-known NHPP models regarding goodness-of-fit and predictive performance. All the evaluation measures indicate that the proposed model outperforms the other models across all the considered datasets. In addition, the graphical plots depicting the actual data and the prediction results for all the fitted models show that the proposed model performs well in all three considered software reliability datasets compared to other models.

Study on Renability Growth for Weibull Distributed Products under Zero Failure Data

Study on Sequential Compliance Method Influenced by Shape Parameter of Weibull Distribution

Evaluating the lifetime distribution parameters and reliability of products using successive approximation method

Reliability evaluation in the zero‐failure Weibull case based on double‐modified hierarchical Bayes

Bayesian Analysis Of Reliability-Growth Test For Exponential Life Distribution Case Based On A New Dirichlet Prior Distribution

Low Carbon-Oriented Optimal Reliability Design With Interval Product Failure Analysis And Grey Correlation Analysis

Grey Duane model for reliability growth prediction under small sample uncertain failure data

Software Reliability Modeling withWeibull-type Testing-Effort and Multiple Change-Points

Reliability Analysis of Dynamic Load-Sharing Systems with Constrained and Changing Component Performances

Statistical inference for a competing failure model based on the Wiener process and Weibull distribution

Lognormal and Gamma Mixed Negative Binomial Model for Defects Prediction in Steel Products

New generalized-X family: Modeling the reliability engineering applications

Reliability analysis incorporating exponentiated inverse Weibull distribution and inverse power law

Predictive modeling of reliability engineering data using a new version of the flexible Weibull model

Integrating Generalized Weibull-type Testing-Effort Function and Multiple Change-Points into Software Reliability Growth Models

Analysis of a software reliability growth model with logistic testing-effort function

A new generalized odd log-logistic flexible Weibull regression model with applications in repairable systems

The bivariate exponentiated additive Weibull distribution and its multivariate extension with applications

The Analysis of the Distribution of the Failure Statistical Data Hydrogen-Adjacent Pipelines

A finite failure software reliability model using extended log-logistic distribution

A modification in Weibull parameters to achieve a more accurate probability distribution function in fatigue applications