Multiagent仿真与资源稀缺下的适应行为  

Multiagent simulation and individual's adaptive behaviors in resource shortage

作　　者：赵纯均[1] 华立[1] 

机构地区：[1]清华大学经济管理学院,北京100084

出　　处：《管理科学学报》1999年第4期1-7,共7页Journal of Management Sciences in China

摘　　要：以信息经济学和多ａｇｅｎｔ系统理论为分析框架，借助美国Ｓａｎｔａ Ｆｅ Ｉｎｓｔｉｔｕｔｅ 开发的仿真工具Ｓｗ ａｒｍ ，研究了短缺经济下个体的选择、竞争与适应行为，以及制度的进化过程．研究表明，简单的个体适应行为可以产生复杂的整体效果；个体基于历史以及相互间的信息交流导致了学习；满足一定的条件，新的制度安排将突现；个体之间信息的不对称可能使系统进化出类似道德风险和逆向选择的复杂行为．本研究有助于理解现实社会中普遍存在的占座现象；有助于解释财产权利以及制度的变迁．Theories of information economics and multiagent system are adopted as analysis frame for this paper, with Swarm as the simulation tool. Swarm is a software package for multiagent simulation of complex systems being developed at the Santa Fe Institute, which has been a useful tool for researchers in a variety of disciplines, especially artificial life. In this paper we use it to describe agents' selection, competition, and adaptive behaviors in resource shortage, and the evolution of institutions as a whole. Agents interact each other by local information and relatively simple reinforcement algorithms, and new data are collected from the environment and processed, which change the internal model of the agent in the end. Research evinces that simple adaptive behaviors of individuals can lead to complex system characteristics in higher level. Agent learns a lot from analysis of the operation history and by communications among them. New institution emerges in some conditions, which is often quite efficient in the context.Furthermore, we observe that complex system behaviors such as moral hazard and reverse selection may evolve out due to informational asymmetries. This research helps to explain the universal occupation phenomenon in real world, and the evolution of property right and institutions. Above all, multiagent simulation used in this paper offered a quite new perspective in research of complex social economic phenomena.

关 键 词：经济系统 仿真 资源稀缺 适应行为 人工智能 

分 类 号：F014.9[经济管理—政治经济学]