We axiomatize a new class of multi-prior preferences for decision-making under uncertainty. The unique feature of this class of preferences is that it allows for the role of a reference probability measure. The class of preferences has a tractable representation. It takes the form of minimization, over a set of priors, of an expected utility plus a penalty function that penalizes deviation from the reference probability measure. The preference reduces to the standard expected utility when there is no uncertainty. The paper also discusses some potential applications of the axiomatized preferences.
展开
下载论文 
复制链接
新浪微博
微信扫一扫