Research On Priority Strategy For Software Testing Based On Complex Network

With the continuously growing of software industry, the complexity of software is increasing, at the same time, the quality of software has come to be a challenge for software engineers. It is a new field to study the complexity and stability of software from it’s structure. As a new subject of system science, The complex network brought a new perspective for in-depth research of software system. Select the complex network as a tool, construct the network model of software system. The model can give out the whole properties of the software system, that greatly simplify the complexity of the software system. Combined with the software test result of the real software system, and make some discussion. Based on the complex network parameters and testing code coverage, define a new parameter:test-first degree, and give out a priority strategy for test.First of all, introduce the basic knowledge of complex theory. Including the research history、the commonly used complex network statistical characteristic、the basic characteristic and evolutionary algorithm of four complex network models. And then introduce five steps of software test. In further, introduce the first step of software test, include several commonly used methods of unit test.Secondly, select the real software system as the object, build the network model of this software system, and draw the diagram of software system. Prove the scale-free and small world properties of the software system, that provide a reliable basis for the following study of introducing the complex network characteristic parameter into the study of software system.Finally, analyze the code coverage of the first unit test, introducing the complex network characteristic parameter into the study of software system. Based on the complex network parameters and testing code coverage, define a new parameter:test-first degree, and give out a priority strategy for test, which is a new idea to improve the stability of software.Examples show that the priority strategy is effective to improve the test coverage, search out software defect class in time, and also helpful for improving the efficiency of software testing.