基于Copula-GARCH-EVT的资产组合选择模型及其混合遗传算法

在非正态分布的条件下,M arkow itz的均值-方差资产组合选择模型存在不足。为此,以V aR和CV aR作为风险度量方法,EVT反映收益率的尾部分布,GARCH反映收益率的波动性,Copu la函数反映金融资产收益的相关性,构建了基于Copu la函数的资产组合选择模型。针对非正态分布条件下V aR非凸性和分布函数不连续性导致资产组合选择优化计算复杂、不精确的难题,设计了基于单纯形和传统遗传算法的混合遗传算法。最后,根据中国证券市场数据,采用该混合遗传算法对建立的资产组合选择模型求解。

A Portfolio Selection Model on Copula-GARCH-EVT Based and Its Hybrid Genetic Algorithm

There are some drawbacks in Markowitz's mean-variance portfolio selection model under the condition of no normal distributions.So the article constructed a portfolio selection model based on Copula-EVT-GARCH,which measures the risk by VaR and CVaR,reflects the tail distributions by EVT(extreme value theory),reflects the volatility by GARCH,reflects the dependence of financial assets returns by copula function.Then according to the fact VaR's no convexity and the discontinuity in distribution,which make the computation of the portfolio selection optimization very complex and inaccurate,the article designed a hybrid genetic quantitative algorithm based on Nelder-Mead simplex and traditional genetic algorithms.Finally,according to the data from China securities market,the article does empirical research for the portfolio selection model by the hybrid genetic quantitative algorithm.