Research On Portfolio Optimization Based On Multi-strategy Grey Wolf Algorith

Against the background of rapid development of market economy,it has long been the research theme of financial industry to counter risks and stabilize returns.With the advent of the information age,the use of intelligent algorithms to optimize the portfolio has become the focus of academic research.Grey wolf optimizer is a new optimization algorithm proposed in recent years.Its characteristics of simple coding,high search accuracy,high convergence rate and not easy to fall into the local optimal make it applied in many fields,such as machine learning,image processing and financial engineering.At present,portfolio optimization is becoming more and more complex,and the requirements for the good performance of the algorithm are increasing.Therefore,it is necessary and meaningful to study the portfolio optimization problem based on the improved grey wolf optimizer.The main work of this paper is as follows:(1)After studying the portfolio theory and its shortcomings in application from the perspective of risk measurement index,CVaR is selected as the risk measurement tool of the portfolio model in this paper.The investor risk preference index is introduced into the portfolio model to make the model more suitable for solving investment problems in practice.(2)The basic principles and research development of the original grey wolf optimizer are introduced.The adaptive evolutionary population dynamic strategy,differential perturbation strategy and greedy selection strategy are added to the original grey wolf optimizer,and the multi-strategy grey wolf optimizer is proposed from the perspectives of improving the diversity of the population and the efficiency of the algorithm.The performance of the improved algorithm and the three comparison algorithms is tested by reference functions and OR-Library test sets.The results show that the multi-strategy grey wolf optimizer is effective in both function and practical problem solving.(3)The constituent stocks of SSE 50 Index and Kechuang 50 Index are selected to construct two groups of investment portfolios.The original grey wolf optimizer,two variants of grey wolf optimizer with good performance and the improved multi-strategy grey wolf optimizer(ADGGWO)are respectively used to solve the portfolio model constructed in this paper.By comparing the fitness values corresponding to the solution results of the four algorithms,and combining the standard deviation,the maximum value and other indicators,the improvement degree of algorithm performance is judged,and the optimal investment strategy is given.On this basis,experiment 2 is conducted to limit the number of stocks invested according to the actual situation.According to the results of experiment 1,the top ten stocks of the two groups of asset portfolios were selected respectively to form a new portfolio and solve it,and the optimal investment strategy was obtained,which made the research having more practical significance and application value.The empirical results show that under the three investor roles with different risk preferences,the evaluation indexes of investment strategies given by the improved multistrategy grey wolf optimizer(ADGGWO)in this paper are superior to the three comparison algorithms,which can be used as a reference for investment decisions.