A self-evolving network-based artificial society model for the experiment analysis of complex social system

Abstract

Powered by advanced information technology, more and more complex social systems are exhibiting characteristics of complex networks. In consideration of the cost, legal and institutional constraints on the study of social systems in real world, the computational experiment method has emerged as a novel and powerful computational theory and tool for the quantitative analysis of complex social systems. However, how to map the dynamic network features of the real world into artificial society models is an urgent issue confronting the computational experiments method. In this paper, based on the theory of complex networks and complex adaptive systems, we propose a dynamic self-evolving network-based artificial society modeling method to achieve the autonomous evolution of the whole social network. The method abstracts the artificial society into an individual node model and a global network model: the individual model evolves autonomously through individual behavior rules, which continuously changes the topology of the global network model, and then presents the emergent phenomenon of the artificial society system. And then, the validity of the model is demonstrated by using the world trade network as the case study. The experiment results show that the self-evolving network-based artificial society modeling method proposed in this paper provides a new perspective and tool for studying the network evolution characteristics of complex social systems.