The Study Of Dynamic Spectrum Sharing Algorithm Based On Game Theory In Cognitive Radio Networks

Cognitive radio is born which brings a promising future to solve the problem of spectrumshortage and improve the low efficiency of spectrum utilization. It can adapt system operationparameters by sensing the environment’s variations. In addition, spectrum sharing is always the corefocus of the cognitive radio system. It allows the non-authorized users dynamic share the sensedauthorized spectrum without affecting the normal communication of the authorized users. Dynamicspectrum sharing can realize the re-use of the non-renewable frequency resources that therebyeffectively alleviate the problem of tension spectrum resources. In addition, strategy selections ofthe spectrum, power, etc. are included in spectrum sharing. Game theory is an effectivemathematical tool that can be used to help decision-makers analyze and select for decision-making.Accordingly, how to apply game to study and analyze the problem of spectrum sharing is one of thecurrent research focuses.From the economic point of view, the cognitive radio spectrum sharing technology is studiedbased on game theory, economic theory and other theories. The theoretical framework of spectrumsharing system in cognitive radio network is established. Additionally, some spectrum sharingalgorithms base on game theory for the problem of spectrum sharing are proposed. Major contentsand contributions of this dissertation can be summarized as follows:(1) A three stage market competition-based dynamic game of spectrum access algorithm wasproposed, with a two-layer dynamic spectrum access network model. Backward induction andBertrand game theory were adopted to solve the spectrum access problem. Two duopoly mobilevirtual network operators (MVNO) determined the amounts of the spectrum in succession andattracted the secondary end users by the price. The end users made their spectrum access optionsbased on the heterogeneous preference over the rates and prices. The accuracy and stability of thealgorithm had been analyzed theoretically and certificated by the existence of the unique Nashequilibrium. The simulation results revealed that the MVNOs could make investment and pricingdecisions optimally, and maximize their profits based on the proposed algorithm. Such algorithmalso provides a reference model for the practical application of MVNO.(2) Dynamic spectrum sharing algorithm based on evolutionary game theory is proposed. Theoptimal price strategy selection of two bounded rational cognitive mobile virtual network operators (MVNOs) in a duopoly spectrum sharing market is investigated. The bounded rational operatorsdynamically compete to sell the leased spectrum to secondary users in order to maximize theirprofits. Meanwhile, the secondary users’ heterogeneous preferences to rate and price are taken intoconsideration. The evolutionary game theory (EGT) is employed to model the dynamic pricestrategy selection of the MVNOs taking into account the response of the secondary users. Thebehavior dynamics and the evolutionary stable strategy (ESS) of the operators are derived viareplicated dynamics theory. Furthermore, a reward and punishment mechanism is developed tooptimize the performance of the operators. Numerical results show that the proposed evolutionaryalgorithm is convergent to the ESS, and the incentive mechanism increases the profits of theoperators.(3) The dynamic competition between two bounded rational mobile virtual network operators(MVNOs) in a duopoly spectrum market is constructed. A two stage game is employed to model theinteraction of the MVNOs and the quality of service of the secondary users is taken into account. Inthe first stage of the game, the two bounded rational MVNOs select what type of strategies tocompete for the secondary users. Then in the second stage they decide their prices or quantitiesbased on their chosen types of strategies. The evolutionary game theory is introduced to model thedynamic strategy selections of MVNOs. Using replicated dynamics, the proposed evolutionarygame algorithm can converge to a unique evolutionary stable strategy. Simulation results verify thatthe proposed algorithm can make the MVNOs adaptively adjust the strategies to approximateoptimal solution.(4) A power control algorithm based on differential game theory is proposed. According to thedynamic nature of network, differential game theory is applied to study the power control of thecognitive radio network in the proposed algorithm. The time continuity of the network is considered.The cognitive users’ selections of power control strategy are modeled as a differential game model.It not only considers the user’s current income, but also considers the long-term significance of theuser’s strategy. Through the proposed model, the open-loop Nash equilibrium solution is obtained.The simulation results show that the proposed algorithm can effectively control the transmit powerof the cognitive user to reach a steady state, maximization the revenue of the users and improve thesystem performance.(5) A power control algorithm based on evolutionary game theory in the cognitive radio network.The bounded rationality assumption of cognitive users is taken into consideration. The evolutionary game theory is modeled to the dynamic evolution of the cognitive users’ power strategy selection.The replicator dynamics is employed to analysis the dynamic evolution of the secondary user’sstrategy. The unique evolutionary stable strategy is derived. The experimental results validate thatthe users can overcome the constraints of bounded rationality. The proposed algorithm makescognitive users study and adjust strategy constantly through repeated games to achieve evolutionarystable equilibrium, which leads to an effective control of power.