A multiuser wireless environment is a highly structured system where competition
and cooperation coexist. The aim of this thesis is to illustrate the role
and the applications of game theory to the resource allocation problem in selforganizing
multiuser wireless networks. Various game-theoretical tools will be
considered, including strategic form games with complete information, potential
games, Bayesian games, coupled constraints games, and Nash bargaining games.
There are four main results in this thesis. First, non-cooperative games with
complete information are introduced to study the resource allocation problem
in the context of small-cell wireless networks. In this game-theoretical setup,
the wireless devices are assumed to have complete information about the global
network status. It is shown that this game can be viewed as a potential game.
Thus, the existence and convergence of equilibrium can be readily addressed.
Second, non-cooperative games with incomplete information are introduced to
study a distributed resource allocation problem in the context of fading multiple
access channels. This case is formulated as a Bayesian waterfilling game, in
which the wireless devices are assumed to have only local information about
the fading channel states. This Bayesian game formulation is important from a
practical point of view, as wireless devices can have local information but can
barely access to global information on the network status. Further, coupled
constraints games as well as various fairness concepts are introduced to study
the rate allocation problem in the capacity region of multiple access channels.
The concept of normalized equilibrium is adopted to address the equilibrium
selection problem. Remarkably, when the decision making is based only on
statistical information, all fairness concepts coincide with the unique normalized
equilibrium. Finally, Nash bargaining games are introduced to improve the noncooperative
system performance of the small-cell wireless networks.
