Centralized Spectrum Allocation Strategy For Cognitive Radio Users And Performance Analysis

With the development of communication technology and the growth of wirelessservices, there has been a dramatic increase in the demands for spectrum bands; on theother hand, large part of the licensed bands with a severe under-utilization leads to a greatwaste of spectrum resources. Therefore, how to design a reasonable spectrum allocationstrategy becomes one of the key issues for improving the spectrum utility in a cognitiveradio network. Based on the centralized controlling method, novel spectrum allocationstrategies are proposed in this paper, mathematical models are built, and systemperformance is evaluated.Firstly, only one buffer is set for all the cognitive radio users, a central schedulerallocates the available spectrum to the cognitive radio users according to a first come, firsttransmitted strategy, consequently, a centralized spectrum allocation strategy with a singlebuffer is proposed. A discrete-time queueing model with a service process that can beinterrupted is built. By using the method of an embedded Markov Chain, the queueingmodel is analyzed in steady state.Secondly, considering the adaptive assignment of the spectrum, independent buffer isassigned for each cognitive radio user. By introducing a gated polling strategy, agated-polling based spectrum allocation strategy with multiple buffers is proposed. Agated vacation queueing system with interrupted service is then built. Considering thehandoff procedure between the different cognitive radio users, the system model isanalyzed for steady state by applying the method of regeneration cycle.Thirdly, in order to guarantee the fairness on using the spectrum bands, byintroducing a limited polling strategy, a limited-polling based spectrum allocation strategywith multiple buffers is proposed. A limited vacation queueing system with interruptedservice is therefore built, and then model analysis is given in steady state.Finally, the performance measurement system is constructed for the centralizedspectrum allocation of the cognitive radio users, the formulas of the performance measuresare presented based on the steady state solution of the system models. The dependencyrelationships of the system traffic load for the cognitive radio users and the arrival rate of the primary users on the system performance are investigated with analysis results andsimulation experiments. Taking into account different system measures, such as theaverage response time and the throughput, the optimal design for the arrival rate of theprimary users is given by constructing utility functions.