Conditional Distributionally Robust Functionals

• 成果类型: Article
• 署名作者: Shapiro, Alexander; Pichler, Alois
• 署名单位: University System of Georgia; Georgia Institute of Technology
• 刊物名称: OPERATIONS RESEARCH
• ISSN/ISSBN: 0030-364X
• DOI: 10.1287/opre.2023.2470
• 发表日期: 2024
• 页码: 2745-2757
• 关键词: risk

摘要

Many decisions, in particular decisions in a managerial context, are subject to uncertainty. Risk measures cope with uncertainty by involving more than one candidate probability. The corresponding risk averse decision takes all potential candidate probabilities into account and is robust with respect to all potential probabilities. This paper considers conditional robust decision making, where decisions are subject to additional prior knowledge or information. The literature discusses various definitions to characterize the corresponding conditional risk measure, which determines further the decision. The aim of this paper is to compare two different approaches for the construction of conditional functionals used in multistage distributionally robust optimization. As an application, we discuss conditional counterparts of a distance between probability measures.