提出了一种新的基于非合作博弈的动态频谱分配方案,考虑多个蜂窝用户服务中心和具有认知能力的设备到设备(D2D)通信用户组,利用伯川德(Bertrand)博弈理论来解决用户频谱分配问题,分别对D2D对用户组和蜂窝用户服务中心的效用函数进行了改进,并给出了蜂窝用户服务中心的最优定价和D2D对用户组的动态价格调整策略,进一步证明了纳什均衡解的存在性和算法的收敛性.通过仿真实验,分析了不同蜂窝用户数和学习因子对所提出方案性能的影响.与现有结果进行比较显示,新方案在频谱利用率和系统公平性方面均有改进.
The author develops a new non-cooperative game dynamic spectrum allocation scheme.This scheme considers multiple cellular service providers and D2D service group which have the cognitive ability, and uses the Bertrand game theory to solve the problem of user spectrum allocation.The utility function of D2D service group and cellular service providers are improved respectively.This article also puts forward optimal pricing of cellular service providers and dynamic price adjustment strategy of D2D service group.Furthermore, the existence of a Nash equilibrium state and the convergence of the algorithm are proved.By simulation and experiment, the impact of different numbers of cellular users and different learning factors on the performance of the proposed scheme is analyzed.Compared with existing ones, this scheme improves both the spectrum utilization and system fairness.