Portfolio Selection Under Social Network

In this paper,we use Agent-based Computational Finance to establish a portfolio selection model which based on complex networks.Different from traditional portfolio selection research,this paper introduce the concept of complex network to describe the link of investors in the actual world.Through simulation,we observe the influence of social network and characteristics of risk assets and then we focus on multiple-assets model to study the efficient frontier.We found that when there is social network,due to the connection between the investors,they can receive more information,so as to make more rational judgment,the market is also becoming more smoothly.The returns and risk of risk assets will greatly affect the choice of the portfolio,and both have different effect on different types of investors.Finally,we found that due to investors’ irrationality and heterogeneity,efficient frontier of the model is quite different from traditional mean-variance model.The significance of this paper lies in:first,put forward the new method to study portfolio selection theory.Agent-based Computational Finance makes it easier for us to observe investor behavior.The use of complex network enables us to describe actual contact between the investors.Second,by simulating the effect of various factors on portfolio selection,we summarize the different portfolio selection under different circumstances,and explain the reasons behind the action,this allows us to better understand investor behavior,on this basis,make better investment behavior and the market norms.