IMA Journal of Mathematical Control and Information Advance Access originally published online on August 21, 2009
IMA Journal of Mathematical Control and Information 2009 26(3):325-355; doi:10.1093/imamci/dnp016
| ||||||||||||||||||||||||||||||||||||||||||||||||
On the rapid convergence of a class of decentralized decision processes: quantized progressive second-price auctions

Department of Electrical and Computer Engineering, McGill University, 3480 University Street, Montreal, QC, H3A 2A7 Canada

Capital One, Toronto, ON, M2N 6L7 Canada

Department of Electrical and Computer Engineering, McGill University, 3480 University Street, Montreal, QC, H3A 2A7 Canada
Email: pjia{at}cim.mcgill.ca
Email: wqu29{at}hotmail.com
Corresponding author. Email: peterc{at}cim.mcgill.ca
Received on May 2, 2008; Revision received April 3, 2009. Accepted on June 30, 2009
A progressive second price (PSP) auction mechanism was proposed in Semret et al. (2000, IEEE J. Select. Areas Commun., 18, 2499–2513) for network bandwidth allocation. In this paper, a quantized version of this mechanism (Q-PSP) is analysed where the agents have similar demand functions and submit bids synchronously. It is shown that the nonlinear dynamics induced by this mechanism are such that the prices bid by the various agents and the quantities allocated to these agents converge in at most five iterations or oscillate indefinitely; this behaviour is not only independent of the number of agents involved but also independent of the number of quantization levels.
Keywords: nonlinear dynamics; multiagent systems; progressive second price auctions; markets.