中文摘要：

本文在均值-方差模型的框架下，建立了以Worst-Case VaR（WCVaR）代替方差作为风险测量指标的均值-WCVaR模型。同时，将对数型隶属函数引入到模型中，以证券组合期望收益率极大化和WCVaR极小化为目标，建立了对数型满意程度的模糊决策投资组合选择模型，更好地反映了投资者对目标值的取值意图。依据上海证券市场的实际数据，采用遗传算法进行了模拟计算，验证了模型的有效性。

英文摘要：

Based on the mean - variance model framework the essay has established the Mean - WCVaR model with a worstcase VaR（WCVaR）to replace variance as a risk measurement indicator. Meanwhile, the logistic membership function is introduced into the model. The essay regards the maximization of portfolio return and the minimization of worst - case VaR as its goal and sets up a fuzzy selection model of investment portfolio which is satisfied with the logistic membership. It better reflects the value of intent given by the investor to the goal value. According to the actual data of Shanghai Security, we use the genetic algorithm to simulate and test the validity of the model.