基于马尔可夫决策过程的接纳控制模型研究

Research of admission control model based on Markov decision processes

针对异构网络中系统容量有限、资源利用率低的问题,在分析对比传统呼叫接纳控制模型的基础上,提出一种基于马尔可夫决策过程理论的接纳控制模型。理论采用定义五元组的方式来描述建模过程,推导出目标评价函数,并通过求解具有QoS约束条件下的方程进行数值分析。仿真结果表明,该模型能满足网络动态实时性,解决系统容量有限情况下的最优接纳控制问题,从而能够在一定程度上降低各类呼叫业务的阻塞概率,达到提高不同用户服务质量的体验性、网络的系统收益最大化的目的。

According to the existing problems of the limited system capacity and low resource utilization in heterogeneous net- work, compared to the traditional admission control, this paper proposed the model of admission control based on the theory of Markov decision. It defined quintuple way to describe the modeling process, and derived objective evaluation function and by solving with QoS constraints equations numerical analysis. Simulation results show that the model can meet the network dynam- ic realtime, solve the problem about optimal admission control under circumstances of the limited system capacity, which re- duced the probability of call blocking for various types of service, to improve the quality of service of different users experien- tial and maximize the network system revenue.