The Application Research Of An Improved Artificial Colony Algorithm In Portfolio Optimization Problem

Since Markowitz proposed the "Mean-Variance" theory in the 1950 s,the study of portfolio returns and risks has entered a quantitative period.Since then,many scholars have begun to apply relevant theories and tools of mathematical statistics to the field of financial investment theory.However,with the increase of variable dimensions,the complexity of solution goals,and the diversity of constraints,solving portfolio optimization problems is becoming increasingly difficult.The accurate and fast solution of complex models is of great significance to the development of portfolio theory.Due to the characteristics of fast calculation speed and low logarithmic derivation requirements,the bionic intelligent algorithm has become an important tool for scholars to solve the optimization problems of complex functions in recent years.As a new bionic intelligent algorithm,Artificial Bee Colony algorithm,with its fast convergence speed,high solution accuracy and hard to fall into local optimumhas been used in many fields.With the continuous development of investment theory,portfolio optimization has become more and more complex,and the requirements for algorithm performance are gradually increasing.Therefore,it is necessary and meaningful to study and improve the Artificial Bee Colony algorithm under the framework of portfolio optimization.The focuses of this paper are as follows:(1)Based on the previous research results,this paper uses the standard deviation of asset returns as a measure of risk under the paradigm of the "mean-variance" theoretical model.Instead of the standardized third-order central moment,this paper introduces Pearson skewness coefficient into the model as a measure of asset return skewness andconstructs a "Mean-Standard Deviation-Pearson Skewness Coefficient " of portfolio optimization model.(2)Through in-depth study and analysis of the basic principles and development of the Artificial Bee Colony algorithm,the paper improves the Artificial Bee Colony algorithmand proposes an Artificial Bee Colony algorithm with accelerated asynchronous gradient factor(AAGF-ABC).As a result,the convergence speed and solution accuracy of the improved Artificial Bee Colony algorithm have been proved to be have been obviously improved.(3)This paper selects 8 constituent stocks of the Shenzhen 100 Index and 2 corporate bonds to construct a set of investment portfolios.The standard Artificial Bee Colony algorithm,H-ABCAS algorithm and the improved Artificial Bee Colony algorithm in this paper(AAGF-ABC)are compared in the empirical research of the portfolio optimization model constructed in this paper.The performances of the three algorithms in the environment of portfolio optimization problem solving have been compared and analyzed in this paper.The experimental results show that,under the conditions of three different risk preference investor roles,the improved Artificial Bee Colony algorithm presented in this paper(AAGF-ABC)is superior to the other two algorithms in investment strategy.