中文摘要：

构建一个基于模糊决策学习的人工股市仿真模型，采用灵敏度分析、历史数据法对模型进行了检验和校准，并将其应用于“大小非解禁”问题的实证研究．仿真结果表明，大量抛售行为与股价波动之间存在着复杂的关系，长短期效果并不一致，证明中国股市的“大小非限售”政策难以收到预期的效果．

英文摘要：

The paper constructs a fuzzy learning classifier system （FLCS） which performs decision-making, learning, and evolving of agents in artificial stock markets. On the basis of FLCS, an artificial stock market model is developed, validated and calibrated, and it is applied in an empirical case. The results of simulation show that there are complicated relations between ＂block-selling＂ and volatility of stock price. And farther a- nalysis of simulation data draws shows that the ＂selling-forbidden＂ policy in China stock market will not accomplish its mission.