Toward an Integration of Social Learning and Individual Learning in Agent-Based Computational Stock Markets: The Approach Based on Population Genetic Programming

Abstract

Artificial stock markets has becomea fast-growing field in the past few years. The essence of this framework is the interaction between many heterogeneous agents. In order to model this complex adaptive system, the techniques of evolutionary computation have been employed. Chen and Yeh (2000) proposed a new architecture to construct the artificial stock market. This framework is composed of a single-population genetic programming (SGP) based adaptive agents and a business school. \r\n\r\nHowever, one of the drawbacks of a SGP-based framework is that the traders can`t work out new ideas by themselves. The only way is to consult researchers in the business school. In other works, traders only follow a kind of social learning, while the individual learning is totally missing. In order to model our traders more realistically, we employ a multi-population GP (MGP) based framework with the mechanism of a school. This extension is not only reasonable, but also has economic implications. How do the agents with different learning behavior influence the economy? Are the econometric properties of the simulation results based on MGP more like the phenomena found in the real stock market? In this paper, the comparison between SGP and MGP is studied from two sides. One is related to the micro-structure, traders` behavior and belief. The other to macro-properties, the econometric properties of time series.