The worst-case discounted regret portfolio optimization problem

摘要: Recently a regret portfolio optimization approach is proposed by minimizing the difference between maximum return and sum of each which can efficiently overcome drawback that classical model cannot catch core risk diversification. In this paper, we generalize considering to minimize weighted return. We suppose decision-maker ambiguous about choice weights he choose robust cope with ambiguity. Then aim at maximization return, where be taken in all possible distributions weights. call generalization as worst-case discounted optimization. general solution for problem NP hard some approximation method often proposed. decision maker gets across parts information uncertain weights, example first-order, support set affine first-order information. By applying duality semi-infinite programming, equivalently reformulated linear problem, then established approaches applied our setting. An given show efficiency methods results demonstrate satisfy diversified property under