14280

Tuning the Unified Robust Model of Uncertain Linear Programs: An Application to Portfolio Selection

49

54

6

This paper reformulates the unified robus t model and reveals the relationship between uncertainty sets using norm bodie s and their corresponding robust counterparts defined by dual norms. It show s how to app ly a unified robust modeling approach to portfolio selection . It is shown how the robust model of an uncertain portfolio se lection can be adjusted according to the deci sion makerי s utility function and the uncertainty of the return parameters. The model can be tuned by choosing an appropriate norm body and the radius of the uncertainty region . Simulation experiments have been carried out using 10000 samples of the return parameters for vario us lp -norm solutions. The computational results provide some general guidelines as how to choose a suitable lp -norm in the unified robust model according to the degree of risk aversion of an investor. In addition, when the uncertainty is large , the inves tor personality does not play an important role and loc norm is the choice.

1197971