认知网络中时延最优的频谱决策方法

针对异构认知网络中频谱管理框架的频谱决策问题展开研究，综合考虑物理层频谱感知错误和无线信道衰落、主用户的多次打断带来的不利影响，提出了基于概率的和基于感知的具有负载均衡功能的频谱决策方法。对于基于概率的频谱决策方法，基于短时业务优先的调度策略采用动态规划方法给出了其平均等待时延的闭式解；对于基于感知的频谱决策方法，提出了允许其中断后可以重新搜索可用信道的决策方法，减少了等待主用户传输的时延，并在分析比较两者时延的基础上，提出了一种总体时延最优的频谱决策方法。此外，将认知网络的异构信道特性考虑分析框架中，理论分析较具一般性。仿真结果证实了方法的有效性。

Delay-optimal spectrum decision method for cognitive network

Spectrum decision was investigated within the confine of spectrum management framework for cognitive radio networks.Sensing errors in PHY layer,outage probability on wireless channel and multiple interruptions of primary user were taken into consideration.Probability-based and sensing-based spectrum decision methods were proposed,which were capable of balancing traffic evenly among licensed.For the probability-based method,short-time first discipline was employed to further shorten average residual time,closed-formed expression were presented based on dynamical programing.For the sensing based method,new channel could be chosen for uncompleted transmission,which alleviated the time spent in waiting primary user.Optimal residual time was obtained by choosing the minimal value of the two methods.Furthermore,spectrum heterogeneity was taken into consideration and theoretical analysis was of general sense.Simulation results validate the approach is efficient.