[理学]基于复杂网络结构的.ppt

中国科技大学复杂系统研究组 陈慧平，谢彦波，卜寿亮 蒋品群，赵明，李季 汪洋，王文旭，付传技，高坤，殷传洋，周涛， 金颖滴，陈雯，季莉， 严刚，王杰，褚建勋 研究方向 理论物理：平衡与非平衡统计力学、自组织临界性和相变理论 非线性科学：非线性动力学系统理论、分岔与混沌理论、混沌的同步及其控制、分形理论、时间序列的非线性分析方法 非线性物理及其应用：神经网络理论、元胞自动机模型、交通流复杂系统的微观高速并行模拟、地震预测研究、森林火灾模型 经济物理学：金融数据统计分析、 金融市场自适应行为的基于经纪人相互作用的物理模型 复杂性科学：复杂系统动力学、复杂适应系统、复杂网络 Outline 复杂适应系统与经济物理学 争当少数者博弈模型 可以进化的争当少数者博弈模型 生长有向网络的组织和动力学 Kauffman Boolean Network 网络结构的争当少数者博弈模型的 布尔动力学研究 Econophysics and Complex Adaptive System 最近二十年物理学研究的新方向: 采用精确科学的语言讨论和研究所有社会和经济问题。 用物理学方法处理社会和经济科学问题 例子： 交通流，网络信息流，城市膨胀，社会网络，疾病的传播，科学研究的经济回报，人口姓氏分布，金融物理模型，高频金融数据的经验统计分析， 金融风险预测及规避对策 Motivation 1. Population in which agents are competing for limited resouce ? common in everyday experience! ? losers are always more than winners 2. Agents can adap their behaviour according to past experience and the environment ? common in social, economics, and biological sciences 3. Agents decide based on their limited capabilities and shared (global) information ? information produced, in turn, by activity of agents themselves! (Interaction!) 4. Agents adapt by adjusting their strategies 5. The general situation is that agents are forced to make decisions based on inductive, rather than deductive, thinking [Beyond Reductionism] ? Stock trading ? Traffic flow ? Internet congestion ? Tragedy of the commons Empirical statistical behavior of prices Stochastic process models for price dynamics Models for price formation and money market evolution based on competition and adaptive behavior among agents Applications: Option pricing, Risk controlling, Profit productiveness How to model competition and adaptive behavior among agents in money market？ W.B.Arthur: “ El Farol Bar ” (Amer.Econ.Review, 84(1994)406) N agents, only available information: the numbers who came in past weeks, each has k predictors Seating Capacity: L (cutoff) e.g. 60 seats “Go” or “Stay” ? Pool of V rules of “Predictors” Agents decide based on attendance in past weeks Ba