项目名称: 基于随机规划的多阶段投资组合选择研究
项目编号: No.71271201
项目类型: 面上项目
立项/批准年度: 2013
项目学科: 管理科学
项目作者: 房勇
作者单位: 中国科学院数学与系统科学研究院
项目金额: 56万元
中文摘要: 多阶段投资组合选择一直是金融工程研究领域的前沿问题之一。本项目综合应用金融经济学、随机规划、计量经济学的理论和方法研究多阶段投资组合选择问题,试图为投资决策分析建立一个新的分析框架。采用计量经济模型、随机模拟与优化方法结合,对多阶段投资中资产收益进行情景分析,提出涵盖专家知识的情景树生成方法;将投资者心理纳入动态风险度量,提出一种考虑认知风险的一致性动态风险度量方法;基于上述情景树生成方法和动态风险度量的研究成果,结合不同种类资产的特点和市场限制条件,针对多阶段投资组合选择问题提出随机规划模型;进一步,分析投资组合选择问题的特殊结构,设计出求解随机规划模型的高效算法;最终尝试构建人机交互的多阶段投资决策支持系统的整体框架。随着我国资本市场趋于理性和成熟,对科学化和数量化的金融管理方法提出了新的要求,在这种形势下,应用随机规划探讨多阶段投资组合选择问题,在理论和实践上都具有重要和深远的意义
中文关键词: 投资组合;投资组合选择;随机规划;情景树生成;动态风险度量
英文摘要: Multi-stage portfolio selection is always one of frontier issues in the field of financial engineering. The theory and methods of financial economics,stochastic programming and econometrics are integrated to study multi-stage portfolio selection in this project. We try to set up a new framework for investment analysis. We will propose a generating method for the scenario tree which includes expert knowledge based on econometrics models, stochastic simulation methods and optimization methods. Considering perceived risk, we will propose a coherent risk measure method, which brings investor psychology into dynamic risk. Based on the above achievements of the generating method and dynamic risk measure,we will propose stochastic programming models for multi-stage portfolio selection, considering the characteristics of different kinds of assets and market conditions. Furthermore, we will design algorithms to solve the stochatic programming models through studing special structure of the portfolio selection problem. At last, we will try to build a framework of the interactive decision support system for multi-stage portfolio selection. Getting rational and mature in our capital market, it needs scientific and quantitative financial management. Therefore, applying stochastic programming appoach to study multi-stage port
英文关键词: portfolio;portfolio selection;stochastic programming;scenario tree generation;dynamic risk measure