Complex network characteristics and evolution research of software architecture

Quantitative description and complexity understanding of software architecture are difficult questions of software development. Taking The complexity evolution of VTK software version as an example, this paper proposes a model for predicting the complexity of software evolution based on logistic equation and quantitative validation of Lehman's law. Its basic idea is that the software architecture was mapped to the network, and the topology was extracted as subjects to establish the new complexity characteristic quantity CM to describe the complexity of the software structure. Finally, The relationship between the complexity of software and the cost of software development is analyzed, the result indicate that large software has higher software cost, the cost and the degree of the graph is linear, and limiting the maximum value of the degree is an availability way to reduce the cost.