您的位置:首页  > 论文页面

极大极小距离密度多目标微分进化算法在投资组合优化中的应用

发表时间:2014-01-15  浏览量:1355  下载量:585
全部作者: 韦博洋,曾国巍,焦桂梅
作者单位: 兰州大学数学与统计学院
摘 要: 引入极大极小距离密度(max-min distance destiny, MMDD)多目标微分进化(multi-objective differential evolution, MODE)算法求解多目标投资组合优化模型,此改进的多目标微分算法采用MMDD表示个体间的疏密程度,重新定义微分进化算法的选择操作和Pareto候选解集维护规则,可以很好地保证解的多样性和收敛性。同时在建立均值-方差双目标模型的基础上,鉴于投资者希望偏差越大越好,考虑偏度因素,建立均值-方差-偏度三目标投资组合优化模型。选取创业板的100只股票共69个交易日的收益率进行实证分析,结果表明:MMDD MODE算法能够有效地求解多目标投资组合优化模型,考虑偏度的三目标投资组合模型得出的Pareto最优解为投资者提供了更多可参考信息。
关 键 词: 金融学;多目标投资组合模型;微分进化算法;极大极小距离密度
Title: A multi-objective differential evolution algorithm based on max-min distance density for multi-objective portfolio optimization model
Author: WEI Boyang, ZENG Guowei, JIAO Guimei
Organization: School of Mathematics and Statistics, Lanzhou University
Abstract: This paper introduces the multi-objective differential evolution (MODE) algorithm based on max-min distance density (MMDD) to solve the multi-objective portfolio optimization model, this improved multi-objective differential evolution algorithm employs the MMDD to stand for the density of individual, redefines the selection operation of differential evolution algorithm and the maintenance regulation of the Pareto candidate solution set, which enables the algorithm to ensure diversity of the Pareto solution set and the convergence of the algorithm. After building a mean-variance bi-objective portfolio optimization model, this study considers the skewness as well, builds mean-variance-skewness tri-objective portfolio optimization model. This paper chooses the yield of 100 Growth Enterprise Market stocks in 69 trading days for empirical analysis, the experimental results indicate that the MODE algorithm could be able to solve the multi-objective portfolio optimization effectively, the Pareto optimal solution set of the tri-objective portfolio model considering skewness could provide more references for investors.
Key words: finance; multi-objective portfolio optimization model; differential evolution algorithm; max-min distance destiny
发表期数: 2014年1月第1期
引用格式: 韦博洋,曾国巍,焦桂梅. 极大极小距离密度多目标微分进化算法在投资组合优化中的应用[J]. 中国科技论文在线精品论文,2014,7(1):79-87.
 
0 评论数 0
暂无评论
友情链接