A GAME THEORETIC STUDY OF SPECTRUM ALLOCATION AND PRICING FOR SPECTRUM SHARING IN COGNITIVE RADIO NETWORKS

Abstract

With the increase in demand for wireless communication, there is a need for efficient spectrum management and access. According to FCC, a significant amount of spectrum in the licensed bands remains unutilized. Cognitive radio is an enabling technology to improve the efficiency of the spectrum utilization. In a cognitive radio network, the cognitive/secondary users utilize the spectrum from licensed/primary users without causing interference to those licensed users. The cognitive users are intelligent and compete for spectrum resources. Game theoretical dynamic spectrum allocation has been extensively used for more flexible, efficient and fair spectrum usage through analysis of the intelligent behaviours of network users equipped with cognitive radio devices. In spectrum sharing approach for dynamic spectrum allocation, the primary and secondary users exchange different resources (e.g., frequency band and time slot) or money. This process can also be referred to as spectrum trading and economic games are used for analysis of spectrum sharing. In this dissertation, spectrum sharing approach/game/problem is examined and the problem is formulated as an oligopoly market competition. Firstly, we consider competitive spectrum sharing among the secondary users and the situation is modeled as a Cournot game with Nash equilibrium as the solution. Initially, a static game is formulated for the centralized scenario; later; a dynamic game is used to analyze the distributed scenario. The stability analysis of the dynamic game is also carried out. Secondly, a competitive pricing among the multiple primary services to offer spectrum to a single secondary service is modeled as a Bertrand game model with Nash equilibrium as the solution. Similar to the Cournot game model, both static and dynamic game analyses are carried out along with the stability analysis of the dynamic game. Finally, a Stackelberg game model is proposed and both competitive spectrum sharing among the secondary users and competitive pricing among the primary users are modeled using Stackelberg game. The solution of this game is the Stackelberg equilibrium. The improvement of Stackelberg equilibrium over the Nash equilibrium is shown through simulation results. iii