Abstract
This paper investigates the numerical procedure for solving the continuous-time portfolio optimization under a no-short selling constraint and a leverage constraint. The optimal investment strategy is designed for achieving the pre-determined target wealth. The performance criterion is defined by a suitable function of the difference between the investor’s wealth and the target wealth. However, the explicit boundary condition is no longer available in this situation. To improve the accuracy decreasing by the lack of the boundary conditions, we use the twice integrated radial basis function in the kernel-based collocation method. The obtained investment strategy is evaluated by the Monte-Carlo simulations based on empirical data.
