基于多目标优化和效用理论的高阶矩动态组合投资  

Dynamic Portfolio Analysis for Higher Moments Based on Multi-objective Programming and Utility Theory

作　　者：蒋翠侠[1] 许启发[2] 张世英[3] 

机构地区：[1]山东工商学院数学与信息科学学院 [2]山东工商学院统计学院 [3]天津大学管理学院

出　　处：《统计研究》2009年第10期73-80,共8页Statistical Research

基　　金：国家自然科学基金项目(70471050;70671074);教育部人文社会科学研究青年基金项目(08JC790062);中国博士后科学基金项目(20060400192);全国统计科研计划重点项目(2006B07)的资助

摘　　要：存在高阶矩风险偏好条件下,组合投资选择必须考虑最大化收益、偏度和最小化方差、峰度四个相互冲突的目标。同时,考虑到高阶矩风险的时变特征,应建立高阶矩动态组合投资模型。基于多目标优化技术和效用理论,讨论了高阶矩动态组合投资模型的构建,并利用MATLAB软件中的非线性优化函数fmincon对模型进行了求解,从理论和实证两个层面对两类模型进行对比。Higher moments risk cannot be neglected in portfolio analysis. The optimal portfolio must address the trade-off between conflicting objectives, such as maximizing the expected rate of return and positive skewness, and simultaneously minimizing the variance and kurtosis. Furthermore, since the higher moments risk is time-variant, it is necessary to consider a dynamic portfolio model. Based on multi-objective programming technique and utility function theory, this paper establishes the dynamic portfolio model under higher moments and solves it using a nonlinear programming function ＂fmincon＂ in MATLAB. We also compare the proposed two models from the view of both theory and empiric.

关 键 词：动态组合投资 高阶矩 多目标优化 效用函数 

分 类 号：O212[理学—概率论与数理统计]