A numerical algorithm is proposed for solving a new two-stage portfolio optimization model with a stochastic dominance constraint on the recourse function based on a sample average approximation method. After applicat...
详细信息
A numerical algorithm is proposed for solving a new two-stage portfolio optimization model with a stochastic dominance constraint on the recourse function based on a sample average approximation method. After application of sample average approximation with some assumptions, the two-stage model becomes an equivalent one-stage non-linear program with nonsmooth plus function which can be solved by smoothing technique. Then, a smoothing algorithm is proposed for solving the two-stage portfolio model. The convergence of the proposed smoothing algorithm is studied. Numerical tests illustrated that the smoothing algorithm has better performance than the linear program algorithm introduced by Dentcheva and Ruszczy & nacute;ski in the following three aspects: (i) The number of constraints in our proposed algorithm is independent on the number of samples;(ii) The smoothing algorithm can deal with the nonlinear portfolio models with nonlinear transaction cost function;(iii) The smoothing algorithm has high computational efficiency.
暂无评论