Development and Application of a Multi-Objective Ant Colony Op-timization Method for Portfolio Problem

Abstract

A numerical method of multi-objective optimization is proposed for an approximate solution of the problem based on the generation of feasible solutions using the continuous ant colony method, non-dominated sorting and the epsilon-constraint technique. Solving a problem means finding the Pareto front. Solutions of typical model examples are given. The applied problem of optimizing an investment portfolio has been solved, in which the initial data are the tabulated average returns and covariance of stocks.