Experimental evaluation of auction designs for spectrum allocation under interference constraints

Abstract

The rapid expansion of wireless technology in consumer markets is revealing inefficiencies regarding radio spectrum allocation. To improve spectrum utilization and increase potential revenues, the creation of a secondary market is essential. Opportunistic trading in a secondary market raises the possibility of interference (i.e. a situation in which a given spectrum band is over utilized thus reducing its quality of service). Thus, an auction mechanism for the redistribution of spectrum rights in a secondary market must satisfactorily address the possibility of interference. This work focused on the experimental evaluation of various auction mechanisms for spectrum allocation, primarily the Truthful Multichannel Auction (TMCA) developed by Barrera, Garcia, and Hong. Designs were evaluated based on revenue generation, allocative efficiency and incentive compatibility (including the possibility of collusion). The simulation software suite, Netlogo, was used in modeling versions of the mechanism. The auction scheme utilizes a conflict graph to model interference between auction bidders in which adjacent nodes cannot be allocated the same channel. This interference graph is used to create supply curves for each participating bidder based on conflicting bids. To test performance the results from human experimental auctions were compared with the optimal simulated results. These results show the validity of the Truthful Multichannel Auction's incentive structure and dictate recommendations for further modifications.