基于部分可观察马尔可夫决策过程的受控无线网络系统动态资源分配  

Research of dynamic resource allocation in controlled wireless network systems based on POMDP

作　　者：李萌[1] 司鹏搏[1] 孙恩昌[1] 张延华[1] 

机构地区：[1]北京工业大学北京未来网络科技高精尖创新中心北京工业大学信息学部,北京100124

出　　处：《高技术通讯》2017年第3期220-227,共8页Chinese High Technology Letters

基　　金：国家自然科学基金(61372089;61571021;61671029)资助项目

摘　　要：研究了受控无线网络的动态资源分配。针对传统无线通信传输模型的局限性随着无线通信系统架构的发展日益凸显的问题,提出了一种引入反馈控制策略的受控无线网络模型。该模型结合部分可观察马尔可夫决策过程(POMDP),将用户接收功率与数据传输误码率作为反馈观测对象,对通信小区内基站天线开启数与用户接入数进行动态资源最优匹配。仿真结果表明,这种方法能够有效提升系统传输能效性与可靠性,降低传输误码率,改善系统资源动态匹配控制性能。The dynamic resource allocation in controlled wireless networks was studied. To solve the problem that with the rapid development of wireless communication systems, the limitations of the traditional wireless communication transmission model are becoming increasingly evident, a controlled wireless network model using the feedback control strategy was proposed. The model combines the partially observable Markov decision process （POMDP）, and sets the receiving power of users and the bit error rate （BER） of data transmission as the observation targets in the feedback strategy. The antennas that work on the base station and the users which connect the base station can be dynamic optimal matching each other. The simulation results show that the proposed algorithm can improve the en- ergy efficiency and the reliability of the system, and reduce the BER, so that the dynamic matching performance of system resources can be improved effectively.

关 键 词：受控网络 反馈策略 资源分配 部分可观察马尔可夫决策过程(POMDP) 

分 类 号：O225[理学—运筹学与控制论] TN92[理学—数学]