计算实验金融作为实验经济学的发展和补充已经有了二十多年的发展,虽然其研究方法已经逐渐成熟,但是关于其思想基础和研究范式目前还存在一定的争议和挑战。本文从金融研究需求的视角讨论了计算实验金融的起源与发展,对常见的计算实验金融模型的建模方法与类型进行了评述与总结,进而对目前比较典型的一些挑战性问题进行了讨论,从中展现出计算实验金融的思想基础和研究范式,并重点讨论了其相对于传统研究方法在金融市场异象、市场微观结构、行为金融学、以及交易机制设计等研究领域的优势。
Agent-based computational finance (ACF), which develops and reinforces the experimental economics, has a repid big development for over the 2 decades. Although its research methods is almost perfect, there still some challenges on its ideological foundation and research paradigm. We discussed the development of ACF from the angle of financial research needs, the categories of models and answered these representative challenges. From these answers we can learn ACF's ideological foundation and research paradigm very well, and show some advantages of ACF to traditional methods in the financial markets' anomalous, market microstructure, behavioral finance and trading mechanism design etc.