Cross-shareholding Optimization Of Listed Companies Based On Improved Non-dominated Sorting Genetic Algorithm

Cross-shareholding is a special corporate equity structure,which has become more and more common in China.How to deal with the risk of cross-shareholding is an important problem to be solved in the company's investment decision.Modern portfolio theory can well plan the return and risk of investment,but portfolio theory is rarely reported in the company's cross-shareholding.There are more advantages to applying intelligent optimization algorithm slyly to get the best portfolio strategy.Intelligent algorithms can deal with complex portfolio models that are complex with continuous and hybrid types when constraints are considered.This provides investors with a very effective tool for solving portfolio models.Among them,the genetic algorithm is a new non-linear search optimization technology,widely used in the optimization of portfolio.In this paper,the non-dominant sorting genetic algorithm(NSGA)is used for the strategy optimization of cross-shareholding,and the genetic algorithm suitable for the optimization of cross-shareholding scheme is constructed.This includes the construction of stock pool,gene initialization,parameter initialization,population initialization,gene crossover,gene variation,establishment and termination condition setting of fast non-dominant sorting module.According to the case of Baowu company's cross-shareholding,this paper analyzes the motivation of its cross-shareholding,that is,optimizing the layout of the industrial chain,seeking strategic alliance,pursuing profit and national policy promotion.According to the motivation of Baowu company to optimize the layout of the industrial chain,three portfolio models were constructed for its upstream and downstream listed companies and listed companies in the same industry.The optimal solution of these combined models is solved by Using NSGA.The return and risk of the optimal portfolio under different risk aversion indices(?)was calculated and compared with the actual situation.The results obtained using the NSGA-II algorithm are better than the cross-shareholding configuration actually selected by Baowu company in all cases.On the downstream side,the yield on the optimal portfolio under the different risk aversion indices(?)is calculated.By selecting eight representative investment targets,this article analyzes the reasons for the Baowu company to use the calculated optimal cross-share investment portfolio from the positive and negative effects of cross-shareholding.In summary,the above results show that the combination of the theory of portfolio investment with genetic algorithm can be used for the design and strategy optimization of cross-shareholding schemes of listed companies.