Dependence Analysis Based Model Repair Technology For EFSMS With Ignored Event

Software repair technique needs to understand the whole software system to repair it well,which becomes much harder with the increasing scale of software.Model is a high-level abstraction of software system,which can simplify the analysis to software and make the design of repair algorithm easier.And EFSM model is an important kind of models,which consists of state,transition and the information on transition.When the environment the model in changes,the state of the art repair technology uses static analysis to delete functions after the ignored event.To prevent from ignored event,the original algorithm deletes anything which relates to ignored event directly or potentially.Under certain circumstances,this algorithm will cause the model unavailable totally.This paper proposes a new novel algorithm to repair EFSM model with ignored event based on model dependence analysis and the deficiency of the original algorithm.This method doesn't simply delete anything about ignored event,but tries to identify the function that can be repaired.Then it repairs the model through merging repairable transitions,which can retain functions of the original model as much as possible,on the base of ensuring the availability of repaireempirical d model.This paper demonstrates theoretical and aspects to verify the correctness of the proposed repair algorithm.In theory,this paper proves the consistency of the model after modification by using the concept of model similarity and formal proof.In empirical research,this paper proves in two aspects:comparing the repaired model with the original one,and analyzing the changing information between them;proposing the test case modification method based on the model repair strategy,and comparing the execution results of new test cases in the repaired model with the original test cases in the original model.And the experiment result shows that the repair algorithm proposed by this paper can ensure the repaired model semantical consistency and structural similarity,in the premise of retaining the majority of functions.