Research on tracking error minimization based on dynamic clustering and genetic algorithm

In order to improve the method of stock selection based on correlation coefficient with the order from highness to lowness and the method of optimization based on nonlinear programming, by using the CSI 300 index as the target index, the stock was selected by dynamic clustering. Genetic algorithm was applied to optimize the allocation of funds under certain constraints. The goal of constructing optimal index portfolio with minimized tracking error was achieved. The empirical results show that the combination of dynamic clustering and genetic algorithm in constructing an index portfolio can get smaller tracking error and achieve better simulating effect.