Resumen de The portfolio selection problem: Hybrid approach to solve the mean-variance rational model
Kaoutar Senhaji, Mohamed Ettaouil
In this paper, we propose to complete the Markowitz portfolio problem by a new rational constraint. Besides, the mean goal of this work is to use The Continuous Hopfield Networks (CHN) and the genetic algorithm to solve the proposed model. As know, (CHN) is a robust neural tool, which was widely used to solve combinatorial problems. For that, we propose an adequate energy function. The equilibrium points of our neural network represent the local’s minimums of this energy function. These solutions are not enough good in the general case. As the CHN is fast, we start from several initial points to construct a set of local solutions, this litter is given as the initial population to the genetic algorithm to construct better solutions. As it is showing in the experimental results, the proposed hybrid approach allows improving largely the CHN portfolio profile for the rational mean-variance model. © 2020 SEIO.
