基于复杂网络的软件复杂性度量研究

软件开发者对于日趋复杂的软件系统的理解和控制越来越困难，传统软件工程正接近其复杂性和可扩展性的极限．复杂性使软件开发困难，质量难以保证．复杂网络理论的最新研究成果，为软件复杂性度量提供了新的数学基础．讨论了软件复杂性的形成原因和度量方法，介绍了目前复杂网络与软件复杂性结合的研究工作．探讨了基于复杂网络的软件结构复杂性度量方法，提出一种结合复杂网络和演化算法的软件演化复杂性度量模型．

Software Complexity Metrics Based on Complex Networks

Human understanding and control of complex systems development often seem unattainable goals. The problem is substantial for future construction of network-centric, large-scale software systems. Evidence suggests that software engineering is reaching complexity and scalability limits of technologies. Software development and software quality is troubled with software com- plexity. New and emerging research efforts on complex networks introduce a new mathematical foundation for software complexity metrics. This paper discusses the forming mechanism and metrics methods of software complexity, and introduces recent progresses about associating software complexity with complex networks. Based on software complex networks, some methods of software structure complexity metrics are proposed. Focusing on combination of complex networks model and evolutionary algorithm, a model for software evolutionary complexity metrics is presented.