Research On Robust Optimization Of Portfolio With Different Risk Preference

A portfolio is usually a collection of investments owned by individuals or institutions and made up of multiple marketable securities such as stocks,bonds and derivative financial instruments.Traditionally,the classical paradigm of mathematical programming for portfolio models is to build a model under the assumption that the input parameters are accurately known and equal to some nominal value,and to solve the model using existing mathematical programming methods to arrive at the optimal solution.However,these methods do not consider the impact of data uncertainty on modeling quality and viability.Therefore,this paper uses a robust optimization method to construct a portfolio model to solve the problem that the portfolio model is easily affected by the input parameters.This paper analyzes and summarizes Markowitz's mean-variance portfolio theory model,at the same time,analyzes and compiles the development trend of portfolio models at home and abroad,and then discusses the robust optimization method in detail.This paper also introduces the two main risk measurement methods at present,that is VaR and CVaR methods.From the perspective of behavioral finance theory,this dissertation divides investors into three types: risk aversion,risk neutrality and risk appetite.By using robust optimization methods and incorporating several constraints,several different types of investors are established.The two models are compared under each investor type.The empirical data are used to analyze which models are more suitable for investors and the investors' optimal investment strategies are given through empirical data.