A Method of Software Reliability Test Based on Relative Reliability in Object-Oriented Programming

This paper presents an approach about how to use the relative reliability test and operation paths’ reliability prediction to adjust the test allocation of software reliability test in object-oriented programming which is based on the original operational profile. In this new method, software reliability test is not only based on the operational profiles but also guided by the relative reliability prediction results of operation paths. The adjustable range is decided by the original operation running rate and the operation independence factor which can be gotten by neural network learning algorithm and differentiated from software to software.