Research On Test Case Generation Of Computer Interlocking Software Based On UML

As a guarantee of train and shunting safety in station and computer interlocking system’ssafety-critical software, interlocking software affects the safety and reliability of the entiresystem. Testing is an effective way to make sure the quality of software, and the test casegeneration is the emphasis and difficulty of the test. Therefore, the study of test casegeneration of interlocking software is meaningful to ensure the safety and reliability ofcomputer interlocking system.This thesis researches on the test case generation of interlocking software:(1) The conception of model testing and the existing testing models are introduced.Testing models are compared, and then the advantages of using unified modeling language(UML) model as the testing models are presented. The features, syntax, semantics of UMLand composition and characteristics of modeling tool Rhapsody are illustrated, and then UMLmodel testability problems are analyzed.(2) Method of generating test cases based on UML use case diagrams and sequencediagrams model is proposed. By traversing the directed graph SUG based on use casecoverage criteria, which is indicated the use cases’ dependencies, test sequences are gained.To get the effective event sequences of sequence diagram conveniently, sequence diagramswhich are showed the information interaction among all objects in use case are definedformally. And then considering combined fragment of UML sequence diagram, such as alt,loop, opt, break and polymorphism information, sequence diagrams are converted to directedgraph SEG. By traversing the SEG based on the testing coverage criteria, test paths areformed, and then the redundant paths are deleted according to constraint condition. Then byanalyzing and eliminating the contradictions in all test scenarios, the systematic consistentrequired scenarios are generated. Finally, the interlocking software’s test case suites aregotten.(3) The effectiveness of the method of test case generation is verified. Interlockingsoftware’s test cases are generated in terms of interlocking software’s use case diagram andthe sequence diagram of LockRoute use case. And then the validity of disposing combinedfragment and polymorphism information and the feasibility of the way of generating test caseand the validity of test cases generated are compared between this method and other methodsin other references. The results indicate that test cases generated using the test case generatingtechnique presented in this thesis have higher error detecting capability, and they also couldcover all functions of the system and all scenarios during the execution of the function.