Modelling and analysis of software reliability with Burr type X testing-effort and release-time determination

Purpose - The purpose of this research paper is to discuss a software reliability growth model (SRGM) based on the non-homogeneous Poisson process which incorporates the Burr type X testing-effort function (TEF), and to determine the optimal release-time based on cost-reliability criteria. Design/methodology/approach - It is shown that the Burr type X TEF can be expressed as a software development/testing-effort consumption curve. Weighted least squares estimation method is proposed to estimate the TEF parameters. The SRGM parameters are estimated by the maximum likelihood estimation method. The standard errors and confidence intervals of SRGM parameters are also obtained. Furthermore, the optimal release-time determination based on cost-reliability criteria has been discussed within the framework. Findings - The performance of the proposed SRGM is demonstrated by using actual data sets from three software projects. Results are compared with other traditional SRGMs to show that the proposed model has a fairly better prediction capability and that the Burr type X TEF is suitable for incorporating into software reliability modelling. Results also reveal that the SRGM with Burr type X TEF can estimate the number of initial faults better than that of other traditional SRGMs. Research limitations/implications - The paper presents the estimation method with equal weight. Future research may include extending the present study to unequal weight. Practical implications - The new SRGM may be useful in detecting more faults that are difficult to find during regular testing, and in assisting software engineers to improve their software development process. Originality/value - The incorporated TEF is flexible and can be used to describe the actual expenditure patterns more faithfully during software development.