Affiliation:
1. Department of Informatics, Technical University of Munich, 85748 Munich, Germany;
2. Department of Economics, Stanford University, Stanford, California 94305
Abstract
Combinatorial auctions have found widespread application for allocating multiple items in the presence of complex bidder preferences. The enumerative exclusive OR (XOR) bid language is the de facto standard bid language for spectrum auctions and other applications, despite the difficulties, in larger auctions, of enumerating all the relevant packages or solving the resulting NP-hard winner determination problem. We introduce the flexible use and efficient licensing (FUEL) bid language, which was proposed for radio spectrum auctions to ease both communications and computations compared with XOR-based auctions. We model the resulting allocation problem as an integer program, discuss computational complexity, and conduct an extensive set of computational experiments, showing that the winner determination problem of the FUEL bid language can be solved reliably for large realistic-sized problem instances in less than half an hour on average. In contrast, auctions with an XOR bid language quickly become intractable even for much smaller problem sizes. We compare a sealed-bid FUEL auction to a sealed-bid auction with an XOR bid language and to a simultaneous clock auction. The sealed-bid auction with an XOR bid language incurs significant welfare losses because of the missing bids problem and computational hardness, the simultaneous clock auction leads to a substantially lower efficiency than FUEL because of the exposure problem. This paper was accepted by Axel Ockenfels, behavioral economics and decision analysis. Funding: This work was supported by Deutsche Forschungsgemeinschaft [Grant BI 1057-1/8]. P. Milgrom gratefully acknowledges support from the U.S. National Science Foundation [Grant SES-1947514]. M. Bichler and G. Schwarz was supported by the German Research Foundation [Grants BI 1057 I-9 and 277991500]. Supplemental Material: The data files are available at https://doi.org/10.1287/mnsc.2022.4465 .
Publisher
Institute for Operations Research and the Management Sciences (INFORMS)
Subject
Management Science and Operations Research,Strategy and Management
Cited by
1 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献
1. Mediated talk: An experiment;Journal of Economic Theory;2023-03