arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1210.7230q-fin.TR
收藏

基于随机PDKEEPAAA的限价盘模型、参数估计及投资优化

A Model of Market Limit Orders By Stochastic PDE's, Parameter Estimation, and Investment Optimization

Zhi Zheng, Richard B. Sowers

摘要:

本文基于Zheng和Sowers(2012)给出的一类随机偏微分方程解的存在性、唯一性和正则性,建立了一个完全连续的时变限价盘演化模型.与文献中提出和研究的几个模型相反,该模型提供了从随机PDE继承的时间和价格的完全连续性,因此特别适合于交易以极快的速度发生的情况,例如由高频交易者(HFT)交付的交易。首先给出了模型及其参数的精确定义,并在给定参数的情况下,从相关的数学结果中证明了模型的存在唯一性。然后,我们统计得出的模型参数估计方案,使用最大似然和最小均方误差估计方法在一定的标准,如AIC,以适应不同数量的参数。最后,作为模型的典型经济金融应用案例,通过分析将模型及其参数作为外生变量的投资者或交易者的效用函数的随机演化,解决了静态和动态意义下的投资优化问题。证明了两个定理,为确定最佳(极限)价格和交易时间点提供了标准。

英文摘要:

In this paper we introduce a completely continuous and time-variate model of the evolution of market limit orders based on the existence, uniqueness, and regularity of the solutions to a type of stochastic partial differential equations obtained in Zheng and Sowers (2012). In contrary to several models proposed and researched in literature, this model provides complete continuity in both time and price inherited from the stochastic PDE, and thus is particularly suitable for the cases where transactions happen in an extremely fast pace, such as those delivered by high frequency traders (HFT's). We first elaborate the precise definition of the model with its associated parameters, and show its existence and uniqueness from the related mathematical results given a fixed set of parameters. Then we statistically derive parameter estimation schemes of the model using maximum likelihood and least mean-square-errors estimation methods under certain criteria such as AIC to accommodate to variant number of parameters . Finally as a typical economics and finance use case of the model we settle the investment optimization problem in both static and dynamic sense by analysing the stochastic (Itô) evolution of the utility function of an investor or trader who takes the model and its parameters as exogenous. Two theorems are proved which provide criteria for determining the best (limit) price and time point to make the transaction.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑