The Research Of Test Case Generation Approach Based On Markov Time Usage Model For Real-Time Software

Computer software technology play a role in various fields in-depth,computer software reliability requirements are also increasing,especially for real-time software reliability is the focus of attention of people.Software reliability testing is an important means to improve software reliability,in recent years,with the development of object-oriented technology,model-based software reliability testing has become an important branch of software reliability research,domestic and foreign scholars to carry out in-depth and extensive research has also made certain achievements.But there are still some problems in test cases generation method of real-time software reliability test based on models.Through in-depth analysis of relevant literature we found,existing literature solved the conversion problems from UML model to Markov chain usage model,during the conversion process,through the state with time constraints extending time delay state and the dual time violation state,resolved time constraints impact on different states.But they did not consider the impact on the stimulation time constraints of the software under test.And the stimulation time constraint existing in real-time software is a common phenomenon.Aiming at the shortcomings of existing methods,in this paper,by extending the Markov chain usage model,propose an improved concept of Markov time usage model.The model takes into account both the state time constraint,and the stimulation time constraints of the model.Besides,existing test data selection methods of reliability testing basically are random tactics based on satisfying current constraint.This method is simple and easy to implement,but not taking into account the constraints of relationship between input variables.Using this method may generate lots of test cases invalid,make having a very bias between software's estimated reliability and real reliability.In this paper,we present the interval reduction algorithm based on interval algebra for solving linear constraint expression,making target variables get valid ranges.As well as non-linear constraint expression groups,construct objective function by using branch function thought.Using segmented Newton Method solve the problem,making the validity of test cases generated increasing obviously.On this basis,we present a complete and specific technical route of test case generation based on Markov chain time usage model.And implemented a prototype system.Finally,verify the practicability and effectiveness of the method by cases study.