椭球不确定集下的投资组合鲁棒优化模型

对于含有不确定参数的采用CVaR风险度量的投资组合模型,基于鲁棒优化理论的最新进展,结合统计或时间序列,构造形式较为简单的椭球不确定集作为对参数不确定性的近似,把原问题转化为易于求解的确定型最优化问题,解决了该模型由于参数具有不确定性所造成的缺陷,得到鲁棒性与最优性都较为满意的解.通过市场数据对模型的可操作性和实用性进行验证.

Robust Portfolio Selection under Ellipsoidal Uncertainty

A study of the portfolio selection using CVaR strategy with data uncertainty was conducted.Ways to formulate and solve robust portfolio selection problems based on recent progress in robust optimization were proposed.By using statistic theory and time series techniques,we constructed an ellipsoidal uncertainty set which contained the most possible realizations of uncertain parameters.Then the original problem was converted to a deterministic problem which could obtain a solution good for the most possible r...