Research On Software Reliability Modeling Method Based On State Transition Matrix

With the promotion of hardware technology and the accumulation of software technology,software complexity increases with the software size.How to ensure that software reliability is getting more and more attention.It is an effective method to model the software and correct the validity of the model by using the model-based formal method.This topic aims at the different stages of software development,using formal modeling method to ensure the correctness of the demand.After the software is completed,we use the adaptive random test method to ensure the correctness of the program.State transition matrix(STM)formal modeling method can be applied at all stages of software development.Aiming at the modeling problems that multivariate nesting judgment in STM cell,this paper proposes an STM formal modeling method with decision table.Through confirming the multiple variables and variables of the division interval defined in demand during modeling,decision table should be filled.The decision tree is generated by algorithm C4.5 and the decision tree has a good structure that average judgment path is the shortest.Introducing the weight in the decision tree,can make a decision in the order of different attribute judgment of generated more accord with the actual situation of branch,more logical.Then,through the program code conversion algorithm,which can generate code with coding specification.So the code can be understand easily and the cost of future revision is reduced.Finally,it shows the operability and correctness of the modeling method through the hierarchical modeling examples and the generated program code.In order to solve the problem of correctness of STM,an adaptive test method of STM for black box testing is proposed,which is low in test cost and to make up for the deficiency of some of the other methods.Method using the FSCS algorithm,improving the variable distance calculation,through each selected has been generated with the measured case farthest from the test case as the next test case.The distance measurement of random event sequence generation and test cases in STM is the focus of the study.By combining the event sequence with the input variable to generate a test path,the test path distance is converted to an edit distance between strings.The effect of event sequence distance on the distance of the whole test case is determined by experiment and its weight is determined.Finally,the experiment shows the effectiveness of our method by comparing the random test method and our method...