An Statistical Optimization For Large-scaled Cellular Multicast With Scalable Video Streams

The quick development of wireless cellular networks and the popularity of smart mobile devices expedites the coming of network multimedia era. In recent years, the online and live video services have been growing so fast that the wireless networks seem unable to afford the demanding radio resources. In order to exploit such limited resources sufficiently, we require that the compression rate of the video coding should be high and the video steam can be suitable for different network environments. Therefore, this thesis jointly considers the scalable video coding(SVC) and adaptive modulation and coding(AMC). By doing so, this scheme can provide an effective solution to achieving the video multicast communications over different channels.This thesis firstly provides a comprehensive description of the how the multicast communication of scalable video works in a large-scale wireless cellular network. Then, a new framework used for the video multicast communication is proposed, where the twice coding processes(i.e., scalable video coding and network coding), the corresponding decoding, and the two-phase modulation and coding scheme are included. The two-phase coding and decoding includes the scalable video coding and network coding and the corresponding decoding and the two-phase modulation. Specifically, this thesis considers the Nakagami-m distribution to represent the channel fading. The network coding is utilized to reduce the packet loss rate. Moreover, this thesis considers the optimization of system utility in this framework. Subsequently, the system utility optimization problem can be formulated into a convex optimization problem and a Non-convex optimization problem that is solved by the Qusai-Newton method and Genetic algorithm respectively. To be more prominent, considering the increasing network consumption of wireless cellular networks, this thesis jointly considers the system utility and the energy consumption in the proposed framework. Since the system utility and energy consumption is a trade-off, this thesis proposes a scheme to achieve the balance between the system utility and energy consumption and uses Non-dominatedSorting Genetic Algorithm to solve this problem.Simulation results are presented to show the significant time efficiency improvement compared to the existing dynamic programming method under the case that the utility can be guaranteed. Moreover, this thesis discusses the policies for the system utility and energy consumption optimization problem under different requirements demanded by the large-scale video multicast system.