Knight不确定环境下稳健决策模型研究

Study on the Robust Decision Models under Knightian Uncertainty

Knight不确定性是决策者经常面临并且不容忽视的一个因素。在Knight不确定环境下,决策者对未来自然状态的信念用多主观概率测度来表示。本文对基于多主观概率测度的常用稳健决策模型进行了比较分析研究,其中总结了多主观概率测度下稳健效用函数的构造形式,探讨了基于相对熵约束的最大最小期望效用模型、最大乘数效用模型和α-最大最小期望效用模型,并分析了这三个稳健决策模型的区别与联系,最后对稳健决策模型的发展进行了展望。

Knightian uncertainty is a factor that decision-makers often faced but can not be ignored. Under Knightian uncertainty, decision-makers predict the future economy states with a set of subjective probability measures. The paper studied the robust decision-making methods based on multiple priori probability measures under Knightian uncertainty. First, it summarized the Multiple-Priors Utility functions. Then, it discussed three classical robust models of the max-min expected utility model, maximizing multiplier utility model and α- maxmin expected utility model, and analyzed their difference and relation. Finally, it prospected the evolution of the robust decisionmaking methods.