【摘要】 仿真实验分析是研究交易系统规律的一种重要手段,带约束的零智能系统(ZI-C)是业界研究双向交易系统的重要基准.本文针对ZI-C的典型仿真实验进行分析,揭示其中系统内价格出清过程和实验设置的关系,并量化了各交易者在市场匹配中的难易程度.在分析中,首次提出了一种概率仿真模型,使用了迭代计算来估计每一时刻系统中各种价格产生的可能性,这种方法对其他更普遍的交易模型的分析也具有普适性.结果表明在ZI-C实验完全可由本文模型来表达,本模型对实验中市场价格形成的轨迹的预测同经典实验结果基本吻合.另外本文结论还支持了零智能系统并不具有自主市场调节能力的理论.更多还原

【Abstract】 Simulation experiment is an important way for trading system research,and the system filled with zero intelligence with budget constraint(ZI-C) agents is commonly used as a benchmark for double auction experiments.This paper analyzed the classic ZI-C experiments,revealed the relationship between the price clearing process and the experimental settings,and quantified the trading process for each trader.A math model was proposed and the iteration algorithm was used to predict the probabilities of the transaction prices in trading process.The results show that the prediction is consistent with the outputs of the well known experiments,demonstrating that it is the mechanism of the ZI-C experiment that determines the price convergent process,not the market discipline itself.It also suggests that,in the ZI-C system,the auction market itself has no predictive power.更多还原