Research On Fuzzy Multi-period Portfolio Selection Model

One of the fundamental principles of financial investment is diversification,and investors diversify their assets into different types of assets.How to make decisions in a complex and volatile market environment,improve returns in investment decisions,reduce risks and meet the demands of return and risk have always been a difficult problem for investors.Portfolio diversification can minimize the investor's investment risks,and maximize the investment returns,which will involve a multi-objective optimization problem.They have got many attention and researches of scholars in recent years.In the real world,investors may have to deal with uncertain and inaccurate data as the incomplete information.Therefore,in many cases,these imprecise data or fuzzy information is expressed through natural language and subjective statements.Fuzzy investment decision has become an extremely promising new research direction and subject frontier.However,the research of portfolio selection in fuzzy environment is still mostly in single-period investment.This paper combines fuzzy theory,portfolio optimization theory and dynamic programming to construct a fuzzy multi-period portfolio selection model to simulate the behavior of investors and provide some theoretical suggestions for investment decisions.First of all,based on the possibility theory,we proposed the mean-semivariance fuzzy multi-period portfolio model with transaction costs which take transaction lots,diversification and transaction cost into account.Then using the Shanghai Stock Exchange from January 2010 to January 2016,we analyze investors' investment behavior under different investment preferences,and verify the performance of the proposed model.Secondly,based on the mean-variance model,we assume that the possibility distribution is unknown and construct the mean-variance fuzzy multi-period portfolio selection model by using interval analysis which take transaction cost and turnover ratio into account.Meanwhile,we use the historical data from January 2016 to January 2017 in Shenzhen Stock Exchange for empirical analysis.