The Research Of Software Reliability Models Based On Random Walks

Since the development of the computer science,the cost of the hardware reduced inch by inch.All kind of computer systems with different complex structures and powerful functions had been used in every pivotal domains.Although the computer science,hardware and software had become more and more progressive,the technology of the software reliability detection is short of improvement.As for researcher,software reliability is the most important thing for estimating the software can be successful operation or not.As for consumer,software reliability is the most important thing for preferring to choose this product rather than the others.Nowadays,there are thousands of software reliability models for the cause of the question from detections and predictions.But there is no model can deal with all the problems from all kinds of environments and all kinds of objects.In other words,the software reliability models appeared in the world can only deal one reliability question or one kind of question.The work we should analyze and think deeply is that we would like to find out the master key to deal with all question we meet.The main works we did appeared in these sides:(1)Introduces the conceptions and background of software reliability and software reliability models.(2)Describe four kinds of typical software reliability models:J-M Model,G-O Model,Nelson Model and Seeding Model,it shows the examples using these models and analyzes the disadvantages.(3)Put up the conceptions of Random Walks,theoretical foundations arithmetic using mathematics ways and put up the one-dimensional and two-dimensional arithmetic.(4)Put up a new software reliability model based on Random Walks arithmetic and conduct tests to insure the reliability.We can make the software structures as a map so that it can adapt all kinds of environments.The model can also reduce the period and payment.The software reliability model put from this paper is easier than other models,not only for theoretical foundations but also for executions.When the model become mature and used by plenty of situations,we can take part in the knowledges like simulation and prediction to enhance applied range and precision.