Based On Efsm Uio Sequence Generation

Objected-oriented software is composed of Classes.Their behaviors are usually modeled with state diagrams or finite state machines(FSM).Testing classes is regarded as FSMs testing in which Unique Input/Output(UIO) sequences are widely applied.The generation of UIO sequences is shown a NP problem.To these problems,genetic algorithms(GA) may offer much promise.This thesis studies the test case generation technology of object-oriented software. The target is to synchronously generate code-independent test case in early period of software development and improve the productivity and quality of software.The main research contents are as followed:first this paper analyzes the method of test case generation based on FSM,second compares the advantages and drawbacks of FSM methods.Aiming at state transitions uncertain and testing sequence's unable performance in EFSM,this paper presents a method to partition the domain of precondition,transform EFSM to certain test scenes and generate test sequence by FSM-based method.This paper reports primary results of an on-going research on evolutionary testing classes.The author explain how to define UIO sequence generation as a search problem,and then describe the adapting of genetic algorithms to generating UIO sequences.Special issues of using genetic algorithms such as solution representation,valid checking and fitness definition are discussed in detailed.Primary experiments confirm the applicability and feasibility of applying GA to UIO sequence generation.