Multiple objectives in portfolio construction

In this article we attempt to expand the limited framework of single-objective optimisation and broaden Markowitz's market standard, within which the problem of portfolio selection is conventionally addressed. The decision maker's Investment Policy Statement (IPS) is formulated through a multiobjective mixed-integer linear programming model. The model is then solved through the augmented e-constrained (AUGMECON) method, a novel a posteriori Multiobjective Mathematical Programming (MMP) technique. Our purpose for transforming the traditional two-dimension portfolio selection problem into a mixed-integer multiobjective optimisation problem in space is twofold: 1) to incorporate three objective functions in the decision process beyond the mean-variance approach, i.e., an additional measure of return, the portfolio's relative dividend yield, 2) to model even the most complex preferences that the investor may express, regarding specific cardinal constraints. The proposed model is tested through an illustrative application in the Athens Stock Exchange (ASE).