近似最优机制设计：动机、实例和经验教训
Approximately Optimal Mechanism Design: Motivation, Examples, and Lessons Learned
摘要：
机制设计有着丰富而成熟的理论，也有许多杀手级的应用。从政府如何出售无线频谱许可证到主要搜索引擎如何拍卖在线广告，该领域产生的经验法则影响着一切。然而，有一些基本的问题，传统的最佳机制设计方法是不适合的-要么因为它使过于强大的假设，或因为它提倡过于复杂的设计。本文的论点是，近似最优机制使我们能够推理传统理论似乎无法触及的基本问题。这项调查有三个主要部分。第一部分描述了近似最优的机制设计范式-它是如何工作的，以及我们通过应用它来学习什么。调查的第二和第三部分包括两个案例研究，在那里我们实例化的一般设计范式来研究两个基本问题。在第一个例子中，我们考虑收入最大化在一个单一的项目拍卖异质投标人。我们的目标是了解复杂性-在这个意义上的详细分布知识-是一个重要的特征，好的拍卖这个问题，或者如果有更简单的拍卖是接近最优的。第二个例子考虑多个项目的福利最大化。我们的目标在精神上是相似的：什么时候复杂性-以高维出价空间的形式-是每一个保证合理福利的拍卖的基本特征？有没有有趣的情况下，低维出价空间就足够了？
英文摘要：
Optimal mechanism design enjoys a beautiful and well-developed theory, and also a number of killer applications. Rules of thumb produced by the field influence everything from how governments sell wireless spectrum licenses to how the major search engines auction off online advertising. There are, however, some basic problems for which the traditional optimal mechanism design approach is ill-suited --- either because it makes overly strong assumptions, or because it advocates overly complex designs. The thesis of this paper is that approximately optimal mechanisms allow us to reason about fundamental questions that seem out of reach of the traditional theory. This survey has three main parts. The first part describes the approximately optimal mechanism design paradigm --- how it works, and what we aim to learn by applying it. The second and third parts of the survey cover two case studies, where we instantiate the general design paradigm to investigate two basic questions. In the first example, we consider revenue maximization in a single-item auction with heterogeneous bidders. Our goal is to understand if complexity --- in the sense of detailed distributional knowledge --- is an essential feature of good auctions for this problem, or alternatively if there are simpler auctions that are near-optimal. The second example considers welfare maximization with multiple items. Our goal here is similar in spirit: when is complexity --- in the form of high-dimensional bid spaces --- an essential feature of every auction that guarantees reasonable welfare? Are there interesting cases where low-dimensional bid spaces suffice?
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
