基于股票价格动态特征的多阶段指数优化复制问题研究

项目名称： 基于股票价格动态特征的多阶段指数优化复制问题研究

项目编号： No.71201121

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 管理科学与工程

项目作者： 李倩

作者单位： 西安交通大学

项目金额： 19万元

中文摘要： 指数化投资是一种以获取基准指数的收益为目标的投资管理模式，在投资实践中被广泛使用。指数优化复制是该投资管理中的重要决策。以往文献侧重于单阶段的静态指数复制过程，忽视了指数化投资行为的多阶段性及影响投资组合的因素在动态环境下的变化。本项目以分析股票指数及其成份股的价格动态特征和动态关系为切入点，以离散时间动态情形为背景，建立多阶段纯指数复制模型，并设计解决该问题的随机快速算法；在此基础上，进一步建立多阶段的增强型指数复制模型并设计有效算法；最后应用世界和我国主要股票指数及其成份股的历史数据进行实证研究，测试模型的合理性和算法的有效性，以获得多阶段指数优化复制问题的进一步理解，并应用到投资管理实践中。本项目充分考虑了指数化投资的多阶段性和各种投资模型的融合性，有助于丰富指数化投资管理的理论和方法，促进指数化投资产品的创新发展；有助于提高指数化投资的风险管理水平，为投资者提供理论指导和方法支持。

中文关键词： 指数化投资；股票价格动态特征；多尺度分析；随机占优；投资者情绪

英文摘要： Index investment is a type of passive management that describes the process of investing in a portfolio that attempts to match the performance of some specified benchmark indexes. As index investments have received a lot of attention in the past decade, the problem of optimal index tracking has grasped the interest of more than just portfolio managers. Previous literatures of index tracking have simplified it as a one-period static problem. However, index tracking is, by its nature, a multi-period dynamic problem that influenced by the dynamics of price factors. In this project, we concentrate on the study of the relationships among the time dynamics of the price series for the securities contained in the original index, the index itself, and the replica. We then formulate the pure index tracking portfolio models that includes a comprehensive set of real-world portfolio elements in the context of discrete time series. We also propose the algorithms to search for the solution of the stochastic problems. Based on the pure index tracking models and algorithms for them, we further formulate the multi-period enhanced index tracking portfolio models and propose the algorithms for such stochastic multi-objectives problems. Our proposed approaches are implemented to the main data sets drawn from major world markets incl

英文关键词： index tracking；stock price dynamics；multi-scale analysis；stochastic dominance；investor sentiment