Research On Modelling Theory And Optimization Method For Spectrum Allocation Strategy In Cognitive Radio Networks

With the rapid development of wireless communication, in recent years, networkenvironment is more complicated, and wireless spectrum shortage becomes more serious.Cognitive radio network has thus emerged as an effective technology to resolve theproblem of spectrum shortage. By distributing the spectrum to the Primary Users (PUs)and the Secondary Users (SUs) reasonably, spectrum allocation strategy in cognitive radionetworks devotes to reduce the spectrum holes in time and space to the greatest extent.This thesis focuses on the research of spectrum allocation strategy in cognitive radionetworks. Based on the admission and transmission control, a series of dynamic spectrumallocation strategies are proposed for different network requirements and applications.Then the mathematical modelling and system optimization for the proposed spectrumallocation strategies are presented by using stochastic theory.Firstly, in order to alleviate network congestion and improve the responseperformance of the SUs, by introducing an admission control scheme to the newly arrivingSU packets, a kind of spectrum allocation strategy with access threshold and accessprobability is proposed. A priority queueing model with adjustable arrival rate isconstructed, and then the influence of the access threshold on the system performance isinvestigated. By building a benefit function, an optimum design scheme for the accessthreshold is suggested.Secondly, in order to balance the throughput and response performance of the SUs,by introducing an admission control scheme to the interrupted SU packets, a kind ofspectrum allocation strategy with feedback mechanism is proposed. By constructing apriority queueing model with retrial scheme, the impact of the retrial probability on thesystem performance is analyzed. By building a benefit function, an optimum designscheme for the retrial probability is provided.Thirdly, in order to meet the concurrent spectrum access demands from SUs, a kindof spectrum allocation strategy with concurrent access control is proposed for the scenariothat SUs generate multiple transmission requests simultaneously. By constructing a priority queueing model with multiple input streams, the influence of the SU buffercapacity on the system performance is evaluated. By building a benefit function, anoptimum design scheme for the SU buffer capacity is given.Fourthly, in order to conserve the network resource properly, a kind of spectrumallocation strategy with dynamic channel closing scheme is proposed, in which parts ofthe channels could be closed periodically to realize data transmission control once there isno any transmission request in the system. By constructing a priority queueing model withworking vacation, the impact for the proportion of closed channels on the systemperformance is studied. By building a benefit function, an optimum design scheme for theproportion of closed channels is provided.Lastly, in order to obtain more stable performance for the response time, a kind ofspectrum allocation strategy with dynamic channel bonding scheme is proposed, in whichsome of the channels could be activated and aggregated based on the number of packets inthe system to control the data transmission. By constructing a priority queueing modelwith adjustable service rate, the influence of the SU buffer capacity on the systemperformance is revealed. The game analysis for the access actions of the SU packets isconducted, and a pricing scheme is presented to achieve social optimization.