Date of Award

Fall 12-21-2011

Degree Type


Degree Name

Doctor of Philosophy (PhD)

Graduate Group

Electrical & Systems Engineering

First Advisor

Saswati Sarkar


Cognitive radio networks (CRNs) are emerging as a promising technology for the efficient use of radio spectrum. In these networks, there are two levels of networks on each channel, primary and secondary, and secondary users can use the channel whenever the primary is not using it. Spectrum allocation in CRNs poses several challenges not present in traditional wireless networks; the goal of this dissertation is to address some of the economic aspects thereof. Broadly, spectrum allocation in CRNs can be done in two ways- (i) one-step allocation in which the spectrum regulator simultaneously allocates spectrum to primary and secondary users in a single allocation and (ii) two-step allocation in which the spectrum regulator first allocates spectrum to primary users, who in turn, allocate unused portions on their channels to secondary users. For the two-step allocation scheme, we consider a spectrum market in which trading of bandwidth among primaries and secondaries is done. When the number of primaries and secondaries is small, we analyze price competition among the primaries using the framework of game theory and seek to find Nash equilibria. We analyze the cases both when all the players are located in a single small location and when they are spread over a large region and spatial reuse of spectrum is done. When the number of primaries and secondaries is large, we consider different types of spectrum contracts derived from raw spectrum and analyze the problem of optimal dynamic selection of a portfolio of long-term and short-term contracts to sell or buy from the points of view of primary and secondary users. For the one-step allocation scheme, we design an auction framework using which the spectrum regulator can simultaneously allocate spectrum to primary and secondary users with the objective of either maximizing its own revenue or maximizing the social welfare. We design different bidding languages, which the users can use to compactly express their bids in the auction, and polynomial-time algorithms for choosing the allocation of channels to the bidders.