Indexed by:期刊论文

Date of Publication:2016-12-01

Journal:JAPAN JOURNAL OF INDUSTRIAL AND APPLIED MATHEMATICS

Included Journals:SCIE、EI、SSCI、Scopus

Volume:33

Issue:3

Page Number:767-780

ISSN No.:0916-7005

Key Words:Robust CVaR; Robust reward; Incomplete distribution; Duality theorem

Abstract:In the paper, based on reward-risk optimization models, robust (worst-case) conditional value-at-risk (CVaR) optimization models are presented under partially known information of random variables. Compared with current robust reward-risk portfolio optimization models, the proposed models consider the same distribution of the uncertain variable in the reward and the risk. When an expression of the incomplete distribution information is the discrete distribution, the robust optimization models can be reformulated as non-convex optimization problems by the duality theorem. Finally, the models are used to solve asset allocation problems. Numerical results show that they can give decisions according to personal preferences so that the investors would receive reasonable rewards.