Approximation of Pareto set in multi objective portfolio optimization

Abstract

In this chapter we experimentally investigate several evolutionary multi objective optimization methods and compare their efficiency in problems of portfolio selection with the efficiency of specially tailored method of adjustable weights. Test problems were based on standard portfolio quality criteria, and data on stocks of ten Lithuanian companies. We do not concern here in mach between analytical properties of the criteria functions and such properties favorable for the considered methods; we believe, however that general (global) structure of multi-criteria portfolio selection problem will be invariant with respect to switching from criteria defined by simple analytical formula to criteria defined by complicated numerical methods.