基于移动云计算的协同数据分流联合信道诜择研究

发布时间：2018-02-27 22:40

本文关键词： 移动云计算移动数据分流潜博弈纳什均衡分布式算法　出处：《南京邮电大学》2017年硕士论文　论文类型：学位论文

【摘要】：移动云计算是一种将移动互联网与云计算相结合的新技术,其典型的应用场景就是移动端数据分流。用户可以通过数据分流将移动端的计算密集型任务迁移到云端来改善其移动端的CPU能量消耗问题,从而使得移动端支持繁重的数据业务。但是,由于上行链路上的长距离传输,传统的数据分流技术仍会使得移动设备产生不必要的能量消耗。为了减少移动端在长距离传输上的不必要能量消耗,本论文在传统的数据分流技术上提出了一种协同数据分流联合信道选择的框架。在此框架中,不同的移动设备能够首先形成同信道上的联合,然后再以合作的方式执行数据分流。本论文采用了博弈论的方法将这个信道选择问题建模为非合作博弈,然后证明了这个信道选择博弈是一个潜博弈,因而存在纳什均衡。接着,本论文应用了马尔可夫链的性质为这个博弈设计了一个分布式的信道选择算法,该算法使得其中的移动设备能够在不用了解全局网络信息的情况下自组织的达到稳态。最后,本论文通过理论分析与数值仿真证明了该分布式算法的可行性与稳定性,并且证明了该分布式算法相对于中心式算法的高效性。
[Abstract]:Mobile cloud computing is a new technology that combines mobile Internet with cloud computing. The typical application scenario is mobile data shunt. Users can transfer computing intensive tasks from mobile to cloud to improve the CPU energy consumption of mobile. So that the mobile side supports heavy data traffic. However, because of the long distance transmission on the uplink, Traditional data streaming technology still causes unnecessary energy consumption on mobile devices. In order to reduce unnecessary energy consumption on long distance transmission of mobile terminal, In this paper, a framework of cooperative joint channel selection for data streaming is proposed in traditional data streaming technology. In this framework, different mobile devices can first form a union on the same channel. In this paper, the channel selection problem is modeled as a non-cooperative game by game theory, and it is proved that the channel selection game is a latent game. Therefore, Nash equilibrium exists. Then, this paper designs a distributed channel selection algorithm for the game by using the properties of Markov chain. The algorithm enables mobile devices to achieve steady-state self-organization without knowing the global network information. Finally, the feasibility and stability of the distributed algorithm are proved by theoretical analysis and numerical simulation. The efficiency of the distributed algorithm compared with the central algorithm is proved.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN92

【相似文献】