ACCELERATED SOFTWARE RELIABILITY GROWTH TESTING BASED ON IMPORTANCE SAMPLING FOR HIGH RELIABILITY SOFTWARE

For high reliability software system, software reliability growth testing (SRGT) by the traditional method is inefficient and time-consuming. This paper discusses a method of accelerated SRGT and the theory of importance sampling used in SRGT. The method is put forward based on the strengthened operational profile (OP) which is constructed on Musa's operational profile. An experiment is done to demonstrate the method. Reliability prediction results show that fewer test cases are required with this approach to achieve the same growth target. It can be considered that even very high software reliability growth target could be accomplished within an acceptable period of time.