以复杂网络理论研究了面向对象软件系统质量的度量方法。以实际软件系统作为研究对象,分析了软件系统在初始化以及功能执行完全两个不同阶段下系统复杂度的变化。在此基础上,使用网络模型中节点的信息量作为评判类在软件系统中重要性的指标,用来发掘软件系统运行过程中功能上更加重要的类;最后,用软件网络的信息熵衡量软件系统的复杂度。通过研究证明了以信息量作为描述类的重要程度的合理性,进一步验证了信息熵在整个软件系统复杂程度方面的作用。
This paper analyzed the quality of object oriented software system based on complex network theory. It analyzed the complexity of software networks at both initialization and fully execution stages, and drew some curves to reflect the variation of the complexity of software. First, it built dynamic network model by the execution process to analyze the features of the software. In addition, the new factor that determined the significance of the node in software network was found out. Moreover, this paper put forward some formulas to calculate the significance of nodes and found out the important nodes in the network. Finally, it proved the information entropy and weight of software networks.