Collaborative Research: Data-driven Mechanism Design for Combinatorial Auctions and Exchanges

点击翻译按钮获取中文摘要

This award promises to contribute to the nation's economic prosperity by improving the efficiency of market mechanisms. Market mechanisms increase the welfare of market participants by enabling the transfer of goods and services from those who ascribe a low value to their holdings to those who ascribe a higher value to them. Combinatorial auctions and exchanges are market mechanisms that allow agents to specify preferences (via bids and offers) over bundles of assets (goods, services, rights). The value of an item may depend on whether an agent owns a different item, so the value of a bundle of items may differ from the sum of the values of individual items in the bundle. There are many important domains where the need for such mechanisms arise, such as wireless spectrum markets, which transfer radio spectrum rights from governments to mobile phone carriers and between carriers, truckload transportation, and the leasing market for cloud computing resources. Each of these markets involves the transfer of billions of dollars, and in many cases, the assets are publicly owned. The project will involve undergraduate and graduate student training, and will also support the outreach activities of the PIs in organizing college student volunteers to teach STEM subjects to inner-city K-12 public school students. As there are multiple design objectives in the construction of market mechanisms (efficiency, revenue, incentives, simplicity, etc.), there is no single perfect market design involving combinatorial values. Instead, there is a substantial literature in which a large variety of designs have been suggested. What is missing from this literature, however, is a sound methodology for comparing competing designs under realistic models of participants' values. This project addresses this shortcoming by creating a simulation platform that will support data-driven comparisons of combinatorial market designs in the spectrum auction and cloud computing domains. This platform will be open and accessible to the research community, practitioners, and public agencies to rigorously evaluate existing market designs, variations on those designs, and entirely new designs, and to definitively point to features that render some designs better than others in their chosen application domains.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.