中文摘要：

研究机会约束下基于整数规划的均值-VaR（Value-at-Risk）证券投资基金投资组合选择问题。在验证了股票收益率服从Scaled-t分布的条件下,基于非参数方法且以历史观测数值为序次统计值,结合均值-VaR方法和混合整数规划理论,以收益绝对离差作为目标函数建立了机会约束下基于混合整数规划的均值-VaR证券投资基金投资组合选择模型。它是以VaR收益率阈值与置信水平为导向的。该模型还考虑了证券投资基金中的投资比例限制,使其更具有一定的实际应用价值。

英文摘要：

We analyzed that the yields of stock obey the Scaled-t distribution with the sample of continuous observation in the time window. A portfolio selection model of mixed integer programming of mean-VaR for mutual fund is developed based on the Markowits＇s portfolio investment model, VaR theory and mixed integer programming, which is determined by expected threshold rate of return and confidence level by value at risk. In addition, we presented how to obtain the expected threshold probability of VaR with the minimum expected ratio given. At last, we give the explicit result with MATLAB 7.0 and Eviews 3.0.