项目名称: 不完全信息下的投资组合选择模型研究:一个时间一致性的视角
项目编号: No.71501176
项目类型: 青年科学基金项目
立项/批准年度: 2016
项目学科: 管理科学
项目作者: 李永武
作者单位: 中国科学院数学与系统科学研究院
项目金额: 17.4万元
中文摘要: 投资组合选择理论是金融学、金融工程研究的基本问题之一,从上世纪50年代开始到现在,投资组合选择理论已被系统、深入地研究和发展。本项目拟从动态时间不一致优化的角度,研究不完全信息下的动态均值-方差投资组合选择问题、动态投资消费问题、考虑违约债券的动态投资组合选择问题以及隐Markov 模型下的动态投资组合选择问题,特别是考虑这些模型的时间一致性策略(均衡策略);利用博弈论的思想、概念与方法、非线性滤波理论及随机最优控制的最新研究成果,通过求解扩展的HJB方程系统,给出这些投资组合选择模型的均衡策略的闭形式解或数值解;通过数值方法或数值模拟技术,分析模型参数对均衡策略的影响并给出适当的经济解释。本项目的研究目标是建立和发展一些新的投资组合选择模型,探索研究金融资产配置的新范式,同时也将为我国的金融实践提供决策支持。
中文关键词: 投资组合选择;不完全信息;时间一致性;均值方差模型;期望效用模型
英文摘要: Portfolio selection theory is an one of the basic problems of research on Finance and Financial Engineering, since the 1950s, the portfolio selection theory has been systematically and deeply studied and developed by many scholars. From the perspective of dynamic time inconsistent optimization, we will study the dynamic mean-variance portfolio selection problem, the dynamic investment and consumption problem and the dynamic portfolio choice problem with default bonds under the incomplete information, and study the dynamic portfolio selection problem under hidden Markov model, especially, we will focus on the time consistent strategies (equilibrium strategies). Using the ideas, concept and method of game theory, nonlinear filtering theory and the latest research results of stochastic optimal control, we attempt to derive the closed-form solution or numerical solution of equilibrium strategies of those portfolio selection models by solving the extended HJB equation system. By means of numerical method or numerical simulation technology, we will present some sensitivity analysises to illustrate the impact of the parameters on the equilibrium strategy and give some reasonable economic explanation. The target of this research is to develop some new portfolio selection models and explore the new paradigm of research on financial asset allocation. This research will also provide some decision supports for financial practice in China.
英文关键词: portfolio selection;incomplete information;time-consistency;mean-variance model;expected utility model