Portfolio selection with second order uncertain dominance constraint

This paper proposes an uncertain mean-second order dominance model in the framework of uncertainty theory. By giving mean-expected utility equivalent, we show that the proposed model is suitable for rational and risk-averse investors because the portfolio produced by the model can give the investors the maximum expected return and in the meantime bring the investors expected utility value equal to or higher than the reference return no matter what specific utility functions the investors may take. By offering deterministic equivalents and comparing them with the uncertain mean-variance and uncertain mean-risk index models, we clarify the advantages of the proposed model, i.e., being easier to use and safer in investment. Furthermore, we give a numerical example and some experiments to illustrate the application of the model and the advantages of it.