Strategies To Accelerate Automatic Test Case Generation Based On Interval Arithmetic

Software testing is a very important method to ensure the quality of software.Unit testing is used to test the smallest unit or module at the beginning of all of test work,and it can be divided into three phases which are test case generation,test case execution and statistics of test result.It is very common to accomplish unit testing work by developers and testers of the program manually.Because of the large costs of time and repetition,manual testing has to spend a lot of manpower.Automated unit testing,especially the development of automatic test cases generation,makes up the shortage of manual testing and becomes the hot research area in software testing.The constraint satisfaction problem in path-wise testing is very hard.Due to the complication of the constraints in the path and the limit of the method to solve the problem,it is hard to come up with test cases fast and correctly.This paper proposes four methods to accelerate test case generation which are based on branch bound algorithm and interval arithmetic.A constraint solving method for library functions based on semantic analysis and interval arithmetic,which could make the constraints which contain library functions could meet the requirement.Interval consistency decision based on iteration calculation,which can eliminate the influence on the result of interval arithmetic because of the constraints' order.Constraint transformation can transform some non-four-arithmetic into an arithmetic which interval arithmetic can deal with.Skip back-tracking strategy based on variable correlation,which can avoid the unnecessary the processes of back-rolling and back-tracking in the past so that it can reduce the unnecessary interval arithmetic.After the implementation of the strategies in CTS(Code Testing System),the result of experiment shows that these methods have a good influence on reducing the time consume and improving the accuracy of test case generation.