Cognitive Radio Networks In Dynamic Spectrum Allocation And Sharing

The dramatic increase of service quality and channel capacity in wireless networks is severely limited by the scarcity of energy and bandwidth, which are the two fundamental resources for communications. New communications and networking paradigms such as cognitive radio networks emerged in the recent that can intelligently and efficiently utilize these scarce resources. The cognitive radio built on a software-defined radio, which is defined as an intelligent wireless communication system. With the development of this new technique, how to design efficient spectrum allocation and sharing schemes become very important, due to the challenges brought by the new technique.In this dissertation, we address the problem of spectrum sharing in cognitive radio networks where multiple primary service providers compete with each other to offer spectrum access opportunities to the secondary systems. We formulate this problem as an oligopoly market competition and present spectrum trading model of the oligopoly market, each of the primary service providers aim to maximize its profit under quality of service (QoS) constraint for primary service. For the secondary systems, we adopt a utility function to obtain the inverse demand function. With a Cournot game model, Nash Equilibrium is considered as the solution of this game. However, this assumption may not be realistic in some cognitive radio systems. Therefore, we present distributed algorithms to obtain the solution for the dynamic game, which the primary service providers gradually and iteratively adjust their strategies based on the observations on their previous strategies. The speed of adjustment of the strategies is controlled by the learning rate. The stability condition of the dynamic behavior for this spectrum sharing is investigated. The numerical results reveal the solution convergence to the Nash equilibrium. Stackelberg dynamic game algorithm was present. With a Stackelberg game model, the sub-game perfect Nash equilibrium is considered as the solution of this game. The Cournot game algorithm and the Stackelberg game algorithm are compared. However, the Nash equilibrium and the sub-game perfect Nash equilibrium are not efficient in the sense that the total profit of the primary service providers is not maximized. An optimal solution to gain the highest total profit can be obtained.The collusion can be established among the primary service providers so that they gain higher profit than that for the sub-game perfect Nash equilibrium. However, since one or more of the primary service providers may deviate from the optimal solution, a punishment mechanism may be applied to the deviating primary service provider. A repeated game among primary service providers is formulated to show that the collusion can be maintained if all of the primary service providers are aware of this punishment mechanism, therefore, properly discount factor (or weight) of their profits to be obtained in the future. The result of numerical analysis indicates that these algorithms improve the profit of the primary service.