Social Optimization Research On The Spectrum Alloction Strategy In Cognitive Radio Networks

Dynamic spectrum management is one of the important technologies to achievestrong and rich features in cognitive radio network. The spectrum allocation strategy basedon this technology can greatly improve the spectrum efficiency of cognitive radio network.By combining discrete time queue theory and game theory, the research on socialoptimization of dynamic spectrum allocation strategy in cognitive radio networks isperformed.First of all, setting an infinite buffer and a finite buffer for the secondary user (SU)packets respectively, and considering primary users (PUs) may interrupt the transmissionof SU packets, discrete time queueing models with a possible interrupted transmission anda retransmission policy are established.Secondly, based on the discrete time queuing model with an infinite capacity, in theview of an observable scenario, a two-dimensional embedded Markov process isestablished, and the queueing model is analyzed in steady state. The number of SU packetsin the buffer with an individually optimal strategy and that with a socially optimal strategyare given respectively. Based on an unobservable scenario, by using the method of amatrix geometry solution, the average delay of SU packets is given accordingly. Moreover,it is investigated that the join probability of a SU packet with an individually optimalstrategy is greater than that with a socially optimal strategy.Then, based on the discrete time queuing model with a finite capacity, in the view ofan unobservable scenario, by analyzing the queueing model in steady state, the blockingprobability of a new arriving SU packet and the blocking probability of the interrupted SUpackets are given. On this basis, the arrival rates and the join probability of an SU packetwith an individually optimal strategy and that with a socially optimal strategy are given.Finally, for different arrival rates and service rates of packets, different transmissionreward and sojourn cost, numerical experiments are provided for the spectrum allocationstrategy with individual optimization and with social optimization. Spectrum admission fee of SU packets is proposed, then the spectrum allocation strategy is optimized socially.