Research Of Software Test Case Automatic Generation Based On Efsm

Software has been used in many fields of society with the rapid development of the information technology, so the quality of software becomes the common focus of users and developers. Software test is one of the important approaches of software quality assurance. According to statistics, the cost of software test accounts for about 50% of the total cost of software development. That is to say, software test is very important to the dependability of software. Because of the large cost on software test, people try to automate the software test process automatically. The automatic generation of software test case will chang the traditonal methods greatly which employ only intuition and experience, and lighten loads of work on writing the case of software test.This thesis researched automatic generation of software test case based on extanded finite state machine(EFSM). Preconditions have effect on the feasibility of the routes when states had been changed, so we analysed this problem and turned the undeterminate EFSM to determinate EFSM. When choosing the input data for the test, the 'subsection gradient optimal descent algorithm' will be used. Before using the method, turning the expression to the normal expression and shortening the variable value's area have been done in this thesis in order to automate software test input data generation.EFSM is used in many areas like behavior of objects and relations among objects in objected oriented software systems. At the same time, because the state diagram is much similar to EFSM, The automatic generation of software test case makes sense in pracice to research the automatic generation of software test case.