面向用户体验的动态无线网络资源管理技术研究

With the explosive increase of video service in wireless networks, the investigation of quality of experience (QoE) oriented resource management and resource optimization from the perspective of end users has become a new research trend in wireless networks. In this project, we will investigate dynamic resource management and resource optimization for adaptive bitrate (ABR) streaming over wireless networks. Firstly, we will design a multiple time-scale resource optimization scheme for ABR streaming over wireless networks under time-varying channels, establish the model of QoE influencing factors and reveal the effects of influencing factors on QoE. Secondly, we will formulate the optimization problem as a stochastic optimization problem with consideration of several QoE influencing factors. In order to solve the stochastic optimization problem, we propose an optimal resource optimization algorithm and a near-optimal resource optimization algorithm based on deep reinforcement learning and Lyapunov optimization technique, respectively. Finally, we will design and develop a system-level simulation platform to evaluate the theoretical analysis results and the designed resource optimization algorithms.

针对无线网络中爆炸式增长的视频业务需求，从用户角度出发探究面向用户体验的资源管理和优化已成为无线网络新兴的研究趋势。本项目以无线自适应视频传输业务为主题，拟从动态自适应的角度，探索在动态时变的无线信道状态下面向用户体验的资源管理机制与资源优化算法。首先，本项目将设计多维时间尺度的面向用户体验资源管理方案，对用户体验的各个影响因素进行数学建模，揭示用户体验的各个影响因素对资源优化的影响机理；其次，通过动态优化理论对面向用户体验的资源优化问题进行数学建模，提出基于深度增强学习的最优资源优化算法和基于李亚普诺夫优化的近似最优算法；最后，开发系统级仿真平台，对理论分析结果以及设计的资源优化算法进行验证。

针对无线网络中爆炸式增长的视频业务需求，从用户角度出发探究面向用户体验的资源管理和优化已成为无线网络新兴的研究趋势。本项目以无线自适应视频传输业务为主题，拟从动态自适应的角度，探索在动态时变的无线信道状态下面向用户体验的资源管理机制与资源优化算法。首先，本项目设计面向用户体验资源管理方案，对用户体验的各个影响因素进行数学建模，揭示用户体验的各个影响因素对资源优化的影响机理；其次，通过动态优化理论对面向用户体验的资源优化问题进行数学建模，提出基于深度增强学习的最优资源优化算法；最后，开发系统级仿真平台，对理论分析结果以及设计的资源优化算法进行验证。