Software Reliability Testing Method Based On Markov Chain Model

With computer software technology's wide use in more and more fields, software system requires very high reliability. However, software reliability testing is, actually, not put in practice, even in these critical fields, such as aviation, spaceflight, nuclear energy, weapon, medical device and so on. Discrepancy between requirements and facts roots in inefficient software reliability testing system.Actually, Markov model-based method is the most important and comparatively mature software reliability testing method. This thesis does some researches and experiments in Markov chain model-based method, and main contents are as follows:First, software reliability and software reliability testing is introduced. The characteristic of software reliability is analysed in this part, and software reliability's model view is emphasized. Then software reliability testing definition, research subjects, process, etc. are expatiated.Second, we focus on Markov chain model-based method, especially on three most important and basal subjects: Markov chain model, testing adequacy determination and test cases generation. We will point out the limitation of this method at the end of this part.Third, Markov chain model-based strict method is advanced by improving its expressivity about Markov chain model. We pay more attention to Markov chain model, testing adequacy determination and test cases generation.At last, we design and develop Markov chain model-based method and optimized Markov chain model-based method both and make some experiments to prove our point of views.