不允许卖空情况下多阶段均值-方差投资组合优化

Optimization of multiperiod mean-variance portfolio selection without short sales

提出了不允许卖空情况下终期财富最大化的多阶段均值-方差投资组合模型,其目标函数不具有可分离性。将该模型嵌入到一个辅助模型中,从而转化为目标函数可分离的动态规划问题,并用离散近似迭代法进行求解。最后采用源自上海证券交易所的实证数据验证了该模型和算法的有效性。

This paper proposes a multiperiod mean-variance portfolio selection model aiming at terminal wealth maximization without short sales,and its objective function is inseparable.The original model is turned to a dynamic programming problem with separable objective function by embedding it in an auxiliary model,which can be solved by the discrete approximate iteration method.Finally,an example using the real data from Shanghai Stock Exchange is given to illustrate the effectiveness of the model and algorithm.