AF:Small: The Efficiency of Clock Auctions

点击翻译按钮获取中文摘要

Pricing is a fundamental tool that is very widely used on a daily basis and plays a crucial role in reaching effective resource allocation outcomes. For goods that are very scarce and in high demand, assigning a high price to them ensures that they will be requested only by buyers who value them highly. As a result, given an appropriate pricing, the goods may be efficiently allocated, and the seller may also maximize her revenue. Assigning appropriate prices, however, requires some knowledge regarding the demand for each good, and this information is usually not publicly known. Auctions provide a solution to this problem by interacting with the interested buyers and using the information gained from these interactions to discover appropriate prices. The literature on mechanism design, a field in economics, has provided a list of celebrated auctions, which interact with the buyers in different ways aiming to maximize objectives such as social welfare or revenue. From the perspective of a computer scientist, auctions are, in essence, algorithms that take information from the buyers as input and return a pricing and a resource allocation as output. As a result, in the last two decades, computer scientists have leveraged their long tradition in algorithm design and analysis to analyze classic auctions or design new ones. The focus of this project is on clock auctions, a class of auctions which was recently shown to possess a list of very appealing privacy and incentive properties, and to provide the buyers with a simple interface and a clearly dominant strategy. However, understanding of the algorithmic potential of these auctions is still very limited, and the goal of this project is to address this issue by studying clock auctions from an algorithmic standpoint.This project will approach clock auctions as a class of algorithms in order to evaluate their ability to reach efficient outcomes for well-studied resource-allocation problems. Rather than focusing on the conditions under which these auctions can reach the optimal outcome, which is often the approach in economics, this project will instead consider more demanding settings, where achieving full efficiency may be impossible, and measure the performance of the auctions against the optimal benchmark using worst-case, or average-case, approximation factors. This work will lay the foundations for the design and analysis of practical clock auctions and develop a deeper understanding of the performance guarantees that an auction designer can achieve using clock auctions. This work will also unify the theoretical tools from both the computer-science and economics literature that can be used for designing such auctions. The results of this research will provide a set of algorithmic tools that an auction designer can use in order to produce high-performance practical auctions with compelling incentive guarantees, providing a timely contribution toward the widely-demanded "simplicity in auctions."This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.