A New Mean-Risk Quality Portfolio Optimization Model

This paper introduces a new mean-risk quality portfolio optimization model, grounded in uncertainty theory. It assumes that the return of assets is an uncertain variable and utilizes data estimated by experts to test the model. The approach begins by using the expected value of return and risk quality to represent the return and risk of the portfolio respectively. A new risk measure, termed “risk quality,” is introduced, leading to the establishment of the mean-risk quality portfolio optimization model. Next, to more closely align with real financial markets, the paper integrates constraints reflecting realistic financial market characteristics. These include transaction costs, financing constraints, and threshold constraints. By doing so, it establishes a mean-risk quality model with minimized risk, providing a more robust portfolio optimization framework. Finally, both the basic mean-risk-quality model and the version incorporating realistic constraints are empirically analyzed using genetic algorithms. Sensitivity tests are conducted on the robust portfolio model with minimized risk to further validate its effectiveness.