ABSTRACT
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 390
T HE DECISION ABOUT OPTIMAL COMPOSITION of a portfolio is a complex process, not justa single optimization task. It comprises of the selection of a statistical model, the collection of data, the estimation of the model in a parametric, semi-parametric or
nonparametric way, the choice of an appropriate optimization criterion and finally the
numerical solution of an optimization problem. It is well understood that the precision of
the final decision depends on the quality of the described complete chain of subtasks.