Software testing and reliability growth models

In this thesis, we survey fundamental concepts of software testing. We review general testing process and the key testing methods.;We discuss the theory of software reliability, which is an important quantitative measure of software quality. We present the derivation of several well-known software reliability growth models and describe how they can be used in real-life software development.;Parameters in the software reliability models are usually estimated based on failure time data. We review several commonly used methods of parameter estimation. Maximum likelihood methods and curve fitting using least error methods are presented for two fundamental models: the exponential model and the logarithmic model. The descending mean algorithm is introduced. We apply a global minimization method based on the descending mean algorithm to parameter estimation problems. Numerical experiments are performed on a set of well-known failure time data. The computation results show that the method is efficient and accurate.