Portfolio Optimization and Risk Assessment Based on Random Weighting, Markowitz Mean-variance Model and Genetic Algorithm

This paper explores the portfolio optimization and risk assessment problem by exploiting the random weighting method, Markowitz mean-variance model, and genetic algorithm. First, the historical K-line data provided by the official documents of Baostock is obtained and cleaned, and the normalized processing results of the trading day and closing price are given. Then, the selected stocks are analyzed in detail through the visual analysis of the return rate data, the calculation of the covariance matrix and correlation coefficient matrix heat map, and the average annual return rate and annual volatility of the stock return rate. Next, three different portfolio optimization methods are used to compare their expected returns, volatility, and weight distribution. Finally, the efficient frontier and capital market line of the portfolio are plotted to evaluate the effects of different methods. In this study, the random weighting method, Markowitz mean-variance model, and genetic algorithm are considered as optimization methods to comprehensively analyze the volatility and risk of the selected stock portfolio. This provides a comprehensive and scientific reference for investors to choose stock investment strategies.