The Portfolio Selection Based On GARCH Model And Mixed Integer Programming

In the stock market, investors invest in some different securities to reduce risk, how to select securities and how to allocate fund in these securities are core problem in portfolio theory. The Markowitz mean-variance model has received considerable attention in the last few years since its creation. However, these models always get the proportion of fund invested in single asset to whole fund, which need further conversion in actual investment.A mixed integer programming model for portfolio selection with transaction cost has been discussed in this paper, lower and upper bound of amount on each asset are given as constraints. GARCH model method in Eviews 6.0 is used to forecast the returns and risks of assets. The mixed integer programming model is solved by Lingo 9.0. Numerical results show that the model takes some advantages in return rate and utilization ratio of the funds.