基于人工蜂群算法的高频交易策略组合配置研究  

Study on Portfolio Selection of High-Frequency Trading Strategies Based on Artificial Bee Colony Algorithm

作　　者：陈炜[1] 石诚[1] 孙梦荣[1] 

机构地区：[1]首都经济贸易大学信息学院,北京100070

出　　处：《山西大学学报（自然科学版）》2013年第3期363-368,共6页Journal of Shanxi University(Natural Science Edition)

基　　金：北京市教委科技发展计划项目(KM201210038001);教育部人文社会科学研究青年基金(13YJC630012);北京市人才强教深化计划项目(PHR201108333);北京市财政和北京市教委"人才培养项目";首都经济贸易大学研究生科技创新项目

摘　　要：针对高频交易的发展以及目前国内关于高频交易策略组合配置问题的研究不足,文章以投资组合相关理论为基础,构建高频交易策略组合配置模型.该模型定义了衡量策略收益波动的"Jump"变量,同时设计了包含此变量的协方差估计方法,进而极大地减少了传统投资组合理论下协方差的计算量.最后,应用人工蜂群算法对模型进行了实证研究.结果表明,与传统的高频交易策略配置方法相比,文章所提出的配置方法具有较好的效果.With the development of high-frequency trading and insufficiency of domestic research on its portfolio configuration,we propose a new portfolio allocation model for high-frequency trading strategies.By defining variable jump to measure the abrupt fluctuation of a strategy’s yield as well as designing the method containing the jump to estimate the covariance,the model can greatly reduce the calculation of the covariance estimation compared with the traditional methods.Finally,we present an example to illustrate our proposed method,in which the artificial bee colony algorithm is used.The results show that comparing to the fixed proportional allocation method,our proposed approach has a significant effect.

关 键 词：高频交易策略 组合配置 人工蜂群算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]