Science. 1987 May 1;236(4801):537-43. doi: 10.1126/science.236.4801.537.
Proposed in the 18th century by Cramer and Bernoulli and formally axiomatized in the 20th century by von Neumann and Morgenstern and others, the expected utility model has long been the dominant framework for the analysis of decision-making under risk. A growing body of experimental evidence, however, indicates that individuals systematically violate the key behavioral assumption of this model, the so-called independence axiom. This has led to the development and analysis of nonexpected utility models of decision-making. Although recent work in this area has shown that the analytical results of expected utility theory are more robust than previously supposed, other important issues remain unresolved.
由克莱默和伯努利在 18 世纪提出，并由冯·诺依曼和摩根斯特恩等人在 20 世纪正式形式化，期望效用模型长期以来一直是风险决策分析的主导框架。然而，越来越多的实验证据表明，个体系统地违反了该模型的关键行为假设，即所谓的独立性公理。这导致了决策的非期望效用模型的发展和分析。尽管该领域最近的工作表明，期望效用理论的分析结果比以前认为的更稳健，但其他重要问题仍未解决。
