智能体建模和资本市场复杂性  

Agent-based modeling and complexity of capital market

作　　者：刘兴华[1] 汤兵勇[2] 

机构地区：[1]山东经济学院信息管理学院,济南250014 [2]东华大学旭日工商管理学院,上海200051

出　　处：《管理科学学报》2005年第4期35-42,共8页Journal of Management Sciences in China

摘　　要：以复杂适应系统的思想和智能体建模的方法研究资本市场复杂性是新兴的有价值的研究领域,阐述了资本市场作为复杂适应系统的动力机制,介绍了这一领域一个重要模型———少数派博弈(minority game,MG)模型.仿真发现,处于拥挤阶段的MG具有和实际市场相近的收益率分布.进一步扩展标准MG,提出了快速适应的少数派博弈模型,仿真结果显示,新的模型有着和真实市场相同的特征:收益率分布的尖峰和肥尾现象,揭示了资本市场复杂性的内部动力学机理.Agent-based modeling which is a new approach to the research of the complexity in financial market is promising. Minority game （MG） that is an important agent-based model is introduced in this paper. The results of our simulation show that the fluctuation feature of MG that is in crowded region are more similar to real market than that in prefect cooperation region. So we extend the basic MG and present a high adaptable MG. The results of numerical simulation on the new model show that agents will be more intelligent, and the types of features of fluctuations are the same in real-world market. The new model reveals the intrinsic dynamic of complexity in financial market.

关 键 词：少数派博弈 复杂适应系统 资本市场 智能体建模 

分 类 号：N941[自然科学总论—系统科学]