Research On Similar Path Of Software Testing Based On Complex Network

In recent years, the complex network technology for the research and analysis of software system has become a hot spot. Be aimed at complexity of the software system, complex software topological structure is abstracted into a complex network model which can be simple and effective to mining important execution path in the dynamic execution software. It is important for the research of software structure stability and reliability.Based on previous research, this paper uses the complex network related theory and software engineering research method to combine the real open source software Cflow and Gzip as the research object, and proposes the similarity and the construction of weighted network mining software. The main content of this paper is as follows:First, On the basis of complex network theory, the idea of complex software network modeling is proposed. The establishment of the software node rank model and the software weighted network model is established. Research of the classical sequence mining algorithm is summarized, and the related algorithm is analyzed. Application of fuzzy clustering technique is used to mine software architecture.Secondly, because of the large amount of data and the complexity of data types, the algorithm has a high degree of time complexity. The algorithm is based on the software node rank model. In the dynamic execution of the software, the algorithm of the function node is preprocessed, which can effectively improve the time efficiency of the algorithm.Again, for complex networks, this paper uses the software execution network as a measurement tool, and puts forward the structure similarity algorithm based on software weighted network structure. The software is used to perform similar sequences, which provides a strong basis for reducing test cases.At last, the paper presents the algorithm and the data collected in the dynamic execution of real software, and analyzes the experimental results.