M-CVaR Portfolio Selection Program Under Nonlinear Transacton Costs and Minimum Trading Volumes

In this paper, we propose M-CVaR portfolio selection model under nonlinear transaction costs and minimum trading volumes. We use a quadratic function to approximate origin transaction costs function, set genetic algorithms, and analyse the M-CVaR model by real financial data. A series of numerical experiments shows that the model is reasonable and the algorithm is efficient. Further, we give that the portfolio varies with transaction costs and confidence.