经济复杂系统的建模研究--人工经济模型

经济系统是一个由大量自由买卖主体构成的复杂适应系统,可以用基于Agent的计算经济学（Agent-based Computational Economics, ACE）方法进行研究.本文提出了一个自主开发的ACE系统：SIMECO模型.用CRA（Classifiers, Rules, Actions）三层结构对Agent建模可以得到比传统基于简单固定规则的ACE模型更丰富的涌现行为,包括社会分工的自发形成、价格围绕某一固定区间动态波动、流的自发出现并增长.进一步,计算机模拟实验中还形成了由集体大规模Agent自发的交易行为导致的交易网络之上的商品流,它会伴随着系统的进化而不断增强,并使得SIMECO中的大量Agent能更好地适应多变的环境.

Artificial Economy Model: Modeling Complex Economic Systems

Economic system is a complex adaptive system composed by massive autonomous buyers and sellers. Agent based modeling is an effective approach to study economic systems. SIMECO is an agent-based artificial economy model which uses an innovative agent modeling approach, CRA ( Classifiers, Rules, Actions) architecture, to investigate more general phenomena in economic systems than the conventional ones those are constructed by simple fixed rules agents. Through the simulation experiments, a-bundant emergent results such as spontaneously form of social labor division, price fluctuation within a certain interval and emergence of flows can be concluded. Furthermore, the goods flow can emerge from the behaviors of massive agents spontaneously to facilitate underling agents adapting their dynamic environment more efficiently.