Multi-Period Portfolio Based on the Forecast of the Fuzzy AR Model
Using the fuzzy AR model where the coefficient is fuzzy triangular number to forecast the investment returns,we expand the single-period mean-variance model of Markowitz into multi-period fuzzy investment programming.Using the fuzzy triangular number to describe the investment returns,and the variance and covariance of fuzzy triangular number to measure the investment risk,we establish the multi-period fuzzy portfolio model,and solve this nonlinear programming with the particle swarm algorithms which includes the factor of genetic cross and mutation.By the analysis of the empirical of the 19 stocks in the SSE 100,and to compare with the discrete multi-period mean-variance model,we discover that the optimal investment returns based on the multi-period fuzzy investment programming is higher than the market returns,that is to show the validity of the model,at the same time,the multi-period fuzzy portfolio programming can effectively avoid the predictive error of the discrete multi-period mean-variance model so that it makes the investment more flexible.
fuzzy triangular number AR model multi-period mean - variance particle swarm algorithms portfolio
Yang LIU Xinlu ZHUANG
School of Business Administration,Northeastern University,Liaoning Shenyang 110004,China;School of C School of Communication and Information Management,Northeast Normal University,Changchun 13024,China
国际会议
2008 International Conference on System Management(2008年系统管理学术研讨会)(2008 CSM)
上海
英文
406-412
2008-05-30（万方平台首次上网日期，不代表论文的发表时间）