修正Lagrangian算法在证券组合模型中的应用

The Application of the Modified Lagrangian Algorithm in Mini-max Portfolio Investment Model

鉴于极大熵法因控制参数太大而会遇到数值计算困难的不足(文献[2])以及修正Lagrangian算法(文献[3])在计算中不需要控制参数太大的优点,采用了修正Lagrangian算法求解极小极大证券组合投资模型,并给出了数值算例。数值结果表明只需控制参数足够大,该算法即可求得有效的证券组合投资方案。

The paper presents the modified Lagrangian algorithm（Ref [3]） to solve the min-max portfolio model,to improve the limitation of the maximum entropy method（Ref [2]） that the controlling parameter needs to be very large to induce the numerical difficulty and the advantage of the modified Lagrangian algorithm.The given numerical results show that when the controlling parameter is chosen to be large enough,the efficient investment schemes can be obtained by using the algorithm.