Social networks constitute a major channel for the diffusion of information and the formation of attitudes in a society. Introducing a dynamic model of social learning, the first part of this thesis studies the emergence of socially influential individuals and groups, and identifies the characteristics that make them influential. The second part uses a Bayesian network game to analyse the role of social interaction and conformism in the making of decisions whose returns or costs are ex ante uncertain.

The increasing popularity and widespread deployment of wireless data systems fuel the increasing demand for more spectrum. On the other hand, various studies measuring spectrum utilization show that there is a huge variation in spectrum utilization at different times and locations. In view of this, various dynamic spectrum access (DSA) methods have been proposed in order to achieve more efficient utilization of spectrum resources by virtue of exploiting the variations in spectrum demand over time and space. Implementing DSA systems in a centralized way can lead to complexity and scalability problems due to the extensive control signaling involved. Therefore distributed implementations of DSA systems in which the users can access the system resources at their own discretion have been proposed. These distributed mechanisms typically incorporate cognitive radio systems which act as agents on behalf of users to measure the radio environment and make decisions based on these measurements. On the other hand, the freedom of the users in distributed systems to form their actions can lead each user to try to maximize its benefit from the system without regard the overall performance of the DSA system. Therefore, selfish behavior can prevail in distributed systems, which is likely to degrade the system performance.In this thesis we investigate the implications of selfish decision making in dynamic spectrum access systems. To address this broad problem, we focus our analysis on a particular system which can represent the essential properties of DSA systems and thus can shed light on the performance of the broad class of DSA systems with selfish users. Specifically, we model a DSA system as a multichannel random access system which uses ALOHA for medium access, and we analyze the behavior of the selfish users by modeling the system as a non-cooperative game. In this analysis we incorporate the effect of channel state information on the decision making of the users; we consider both cases when the users act on global (complete) and on local (incomplete) information. We determine the behavior of the selfish users at the Nash equilibria of the non-cooperative game and measure the performance of the system in terms of sum and individual utilities for various user loads and amount of available resources. We try to identify how the performance of the DSA system with selfish users compares with its cooperative counterpart. By performing these analyses we provide insights into the broader question of whether selfish users can utilize spectrum resources in a DSA system as well as cooperative users. / <p>QC 20111208</p> / MultiOperator Dynamic Spectrum access (MODyS)

No description available.