AF: Small: Towards More Realistic Models in Algorithmic Mechanism Design

AF：小：算法机制设计中迈向更现实的模型

• 批准号: 1420381
• 负责人: Anna Karlin
• 金额: $ 45万
• 依托单位: University of Washington
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2014
• 资助国家: 美国
• 起止时间: 2014-08-01 至 2018-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1420381&HistoricalAwards=false
• 关键词: AF; Small; Towards; More; Realistic

Algorithmic mechanism design, or  "incentive engineering", is a topic of great importance at the intersection of game theory, economics and computer science. Research in this field focuses on the design and analysis of algorithms or protocols that are explicitly designed so that rational participants, motivated solely by their self-interest, will end up achieving the designer's goals. In the age of the Internet, the set of important mechanism design applications is vast, including online auctions, advertising auctions, file sharing, routing and scheduling, among others. This project is aimed at developing the foundations of algorithmic mechanism design and its applications.More specifically, the primary goal of the project is to develop simple and robust mechanisms for fundamental problems in algorithmic mechanism design under more realistic modeling assumptions than those used in the past. Building on the theoretical foundations of game theory and computer science, the research will explore auction design in the fundamental setting of agents whose values are correlated or common, and in more realistic settings where agents that do not know their values can spend time and money to learn about them. The PI will also explore mechanism design in repeated and online settings, as well as game-theoretic matching problems.This research has a great potential to influence mechanisms employed in online auctions, a huge societal benefit.

激励机制设计，又称“激励工程”，是博弈论、经济学和计算机科学交叉的重要课题。这一领域的研究集中在算法或协议的设计和分析上，这些算法或协议被明确设计，以便理性的参与者，仅仅受到他们自身利益的激励，最终将实现设计者的目标。在互联网时代，一组重要的机制设计应用程序是巨大的，包括在线拍卖，广告拍卖，文件共享，路由和调度，等等。本项目旨在开发算法机制设计及其应用的基础，具体而言，本项目的主要目标是在比以往更现实的建模假设下，针对算法机制设计中的基本问题，开发简单而稳健的机制。建立在博弈论和计算机科学的理论基础上，研究将探讨拍卖设计的基本设置的代理人的价值是相关的或共同的，并在更现实的设置，代理人不知道他们的价值可以花时间和金钱来了解他们。PI还将探索重复和在线设置中的机制设计，以及博弈论匹配问题。这项研究具有很大的潜力，影响在线拍卖中使用的机制，这是一个巨大的社会效益。