Game Theoretical Analysis of Congested Wireless Networks

Wireless technology has become an inseparable part of our daily lives. In addition to the traditional wireless applications, such as TV broadcast or satellite communication, the demand of wireless access from tablets and smartphones has increased tremendously. One big challenge of wireless communications is that spectrum resource is limited. When the resource is scarce, resource allocation in congested networks is of particular importance. In this thesis, we are interested to investigate the user behavior and new spectrum sharing mechanisms in congested networks using the tool of game theory.;In the first part of the thesis, we model the interaction of secondary users (SUs) in cognitive radio networks as congestion games. We evaluate the inefficiency of the Nash equilibria with the social optimal in four families of games. Most importantly, we identify the worst case scenario which leads to a severe performance loss. In order to tackle the problem, we propose a partial cooperation scheme which provides an alternative to improve the system throughput and fairness of the congested network. By specifying the allocation for cooperating users, we show that SUs have incentive to cooperate truthfully. Moreover, we give the performance bound of system throughput compared with full cooperation scheme. We also observe from the numerical results that the partial cooperation of SUs can greatly improve the system throughput comparing with no cooperation of SUs. In addition, the performance of the partial cooperation scheme is close to optimal in congested networks with a reduced overhead and complexity of the algorithm.;In the second part, we consider the congestion of current mobile networks. By modeling the interaction between a network service provider and the mobile subscribers with the Stackelberg game, we study two pricing schemes which suggested how prices should be set in different regions with the consideration of negative effects of congestion. Under given conditions of resource allocation and utility of subscribers, we show that regionspecific pricing is more preferable than region-agnostic pricing which can improve both the revenue of the network service provider and the social welfare of the subscribers.