基于POMDP的流媒体网络数据调度建模与仿真

摘要

For data scheduling problems in P2P streaming networks, a new scheduling algorithm was proposed. The algorithm considers the node's data request as a random event and builds a mathematical model based on POMDP by using event-based method. By applying the policy gradient algorithm, the model was simulated to get the optimal policy. The simulation results show the effectiveness of the proposed method.%针对P2P流媒体网络的数据调度问题,提出了一种新的调度算法,即将节点的数据请求作为随机事件,运用基于事件的方法,对P2P流媒体网络的数据调度问题,建立基于部分可观Markov决策过程(POMDP)的数学模型,并运用梯度算法仿真求解模型的最优策略.仿真结果表明了所提方法的有效性.