Research On Stability Analysis Method Of Software System Based On Multi-granularity Weighted Network Model

Computer software has been deep into all aspects of our lives.With the continuous expansion of software applications,the increasing demand for the functionality and complexity of the software system,the scale and complexity of the software system are increasing,and the stability of the software system is more and more concerned.The software system structure is considered as one of the important factors that affect the stability of the software system.Therefore,to deepen the stability of the software system to explore and research,we must be able to properly describe the structure of the system information,and to quantify the information.The traditional structural metrics(ie,process-oriented/object-oriented metrics)take into account the attribute metrics of software elements(such as the number of packages,classes,and features,etc.),but the complex information contained in the software system itself does not fully excavated.Because of the lack of corresponding theory and methods,the existing research results are rarely combined with the overall and local analysis of the structure of the system,leading to the inability to correctly understand the inherent characteristics of the software system.In recent years,some researchers began to introduce the theory of complex network to software engineering,using networked thinking to describe the structure of software system,that is,software network,the structure of the software system is studied from the perspective of global topology,which provides a powerful theoretical support for correctly understanding the topology of the software system and its dynamic behavior.But there are still some of the following problems:? lack of analysis of the structure of the software from a different granularity,and a measure of the stability of the software system;? there is no unified theoretical basis for the study of the relationship between the structural characteristics of software systems and their stability.In this paper,with the help of complex network theory,firstly,the paper analyzes the dependencies between the elements in the software,such as packets,classes and features,and constructs the corresponding weighted network model.Secondly,based on the importance of nodes in software network and the overall structure of the software system,a stability analysis method based on multi-granularity weighted network SSAM-WN(Software stability analysis method based on multi granularity weighted network)is proposed,and the stability metrics of software system is defined.Finally,The validity of SS and the relationship between SS and software network structure metrics are verified by experiments,and practical conclusions are drawn,which provide guidance for software engineering practice.