Research On Optimization Of Enterprise Portfolio Based On Two-stage Stochastic Programming

With the integration of industry and finance,the pursuit of wealth by enterprises has also shown a trend of diversification.Among them,stock investment has become a very popular way for various entities.Choosing an ideal investment portfolio to obtain excess returns is a common concern for enterprises when making investment decisions.This Paper takes Lanzhou Huanghe Company,the "Stock King of Northwest China",as the research object,combines accounting knowledge and capital market knowledge to analyze its investment characteristics and problems in the investment process,and proposes to establish a two-stage stochastic programming model to optimize its investment plan,using CVaR to measure risk and considering transaction costs in the model to help the model has a comprehensive set of realistic trading constraints and is therefore more reliable and practical than traditional models.In terms of experimental data,this paper uses Python to obtain historical data on the underlying assets and uses machine learning methods to forecast future closing prices.In terms of model solution,the two-stage stochastic programming model is difficult to solve directly,so the model is reasonably transformed and solved with the help of CPLEX software.In order to verify the validity of the model,four listed companies’ stocks held by Lanzhou Huanghe were selected for example analysis.Finally the standard deviation,Sharpe ratio,expected return and excess return were selected as performance evaluation indicators for comparison with the original portfolio.The conclusion shows that the two-stage stochastic programming model established in this paper has significant practical investment significance and can provide more scientific suggestions for enterprise investment decision-making.