A portfolio model with quadratic subsection concave transaction costs based on PSO

In this paper, the optimal portfolio selection problem with transaction costs is studied. In the previous study, the transaction cost is generally assumed as a V-shaped function of difference between the existing and the new portfolio. But, in this study, a portfolio selection model with quadratic subsection concave transaction costs is presented. Due to proposed model is a complex quadratic programming problem which can't be solved by exact algorithms efficiently, an improved particle swarm optimization (IPSO) is designed to solve it. Finally, a numerical example is given to illustrate our proposed effective approach and the performances of IPSO and standard genetic algorithm (SGA) are compared. Experiment results show that IPSO is clearly superior compared to a SGA.