摘要

Social propagation issue that appears in both the real world and the cyberspace recently gains great attentions in several communities. It typically is involved with a great number of autonomous participated individuals and their local interactions, and may arise global emergent behaviors or properties. It is therefore necessary to investigate what emergence may occur and how to influence the emergence in the social propagation. This paper proposes an agent-based artificial society approach to analyzing social propagation issue, which includes AVI (Agent-Virus-Interaction) model for modeling social propagation and the systematic method to investigate the emergence. The AVI model that borrows ideas from multiagent and organization metaphors has three core concepts: agent, virus and interaction, representing the propagation carrier, content, and channel, respectively. Based on the AVI model, a systematic method for analyzing social propagation that consists of five analysis activities including modeling artificial society, programing simulation codes, designing experiments, performing simulations, and analyzing results, is proposed. We study a case of an infectious virus propagation to illustrate the approach and show its usability and effectiveness. Besides, we design and conduct a number of experiments to analyze what/how factors influence the infectious virus propagation. Finally, we discuss some potential extensions to the AVI model for satisfying further requirements of social propagation applications.