Compromise programming and an approximation of the optimum portfolio

The paper examines the portfolio selection problem from the viewpoint of multiple-criteria programming techniques. Portfolios generated by such safety first models as Roy's, Kataoka's and Telser's safety first rules are treated as compromise efficient ones. The paper also presents a reference portfolio method and compromise programming as tools for generation of the best compromise portfolio. Compromise programming techniques are also used for approximation of optimal portfolio for an investor with particular preferences.