Application Of Variable Point Theory In Software Reliability Field

As computer and information process technology is widely used in all aspects of work and life, demand and dependence for computers are gradually increased. Hardware of computer continues to improve, what’s making demand of software quality become higher. As an intrinsic property in software quality indicators, the importance of software reliability in the field of software engineering is increasingly reflected, this aroused great attention. Nowadays, study of software reliability has become an important research direction in software engineering discipline.Over the years, software reliability has maintained rapid development. A variety of software reliability models have been introduced and have been continuously improved and optimized. Validation and forth-putting of the model has become the focus in the study of software reliability. However, it is still a problem to develop software reliable, moreover, to test and validate its reliability.Software reliability growth model (SRGM) is using mathematical statistical method for software reliability evaluation and prediction, which is a very important aspect of software reliability engineering. Non-Homogeneous Poisson Process (NHPP) is one of the most widely used in SRGM, and the Goel-Okumoto(GO)model is the basis of most NHPP model. Due to the unreasonable of the assumption which will induce a big difference between the prediction results and the actual situation. In order to make the model closer to the actual situation, the assumptions of the traditional GO model is modified. In the past, many traditional SRGMs assumed that the workload of software testing is constant during SDLC.That is, the changes of TE consumption were not considered or involved in the modeling of software reliability growth.In order to more accurately depict the distribution of TE considering all possible changes, in this paper we will study how to use the Parr-curve with multiple CPs to depict the TE consumption and how to perform further software reliability prediction and evaluation. The performance evaluation of Parr-curve with multiple CPs TE model will be performed and discussed based on two real data sets. Experimental results show that the model has better prediction and evaluation ability.