Simultaneous Ascending Auctions (SAA): Strategies

In this appendix we describe bidding strategies we have designed to study simultaneous ascending auction games [Reeves et al., 2003, MacKie-Mason et al., 2004, Reeves et al., 2005, Osepayshvili et al., 2005, Wellman et al., 2007]. Throughout this appendix we use the concept of an SAA environment, which comprises an SAA mechanism over M goods, a set of N agents, and a probability distribution D over M -good value functions for each agent. We refer to this distribution as the distribution of the agents’ preferences, and label the environment D(M,N). In the first five sections we describe the strategy families and the details of their implementation. Refer to Table 1 for the content of these sections. In sections 6 through 8 we report the strategy pools in the restricted games we studied in [Osepayshvili et al., 2005] and [Wellman et al., 2007].