异构无线网络中基于马尔可夫决策过程的区分业务接纳控制的研究

Research of service-differentiated admission control based on Markov decision processes in heterogeneous wireless networks

对异构无线网络中区分业务类型的接纳控制机制进行研究。分析了语音和数据2种典型业务在CDMA蜂窝网络和WLAN中的容量区域。基于马尔可夫决策过程理论,提出异构无线网络中区分业务类型的接纳控制理论模型,规定了不同类型业务的接纳控制行为并推导了系统状态转移概率。而且,进一步从用户角度对不同类型业务QoS要求和网络状态之间关系进行分析,提出一种基于模糊逻辑的接纳效用评估机制,在保证各类业务接入和切换成功率的基础上,推导出接纳效用最大的最优接纳控制策略。仿真表明,基于模糊逻辑的接纳效用评估能够有效反映网络状态动态变化对接纳控制的影响,最优接纳控制策略在平均接纳效用方面明显优于不考虑业务区分和用户移动性2种接纳控制机制,并且能严格保证各类业务的接入和切换成功率。

Service-differentiated admission control was studied in heterogeneous wireless networks.Capacity regions of voice and data services in CDMA cellular network and WLAN were derived.A theoretical model of admission control with differentiated services was proposed on basis of Markov decision processes theory,in which the admission actions were specified for each traffic service and system state transition probabilities were formulated.Furthermore,a fuzzy logic admission utility evaluation mechanism was presented based on the analysis of relationship of QoS requirements and network state,then the optimal admission control policy that maximizes overall utility with new and handoff call blocking probability constraints was formulated.The simulation results reveal that network dynamics can be captured by the proposed utility evaluation mechanism.And the average utility earned in the optimal admission control policy was significantly larger than two other schemes in which service differentiation and mobility are not considered,in addition,the new and handoff call blocking probability can be strictly constrained.