Capacity, coding and optimization for network information theory

网络信息论是研究多用户环境下通信的基础理论。在过去的十年里，人们对这一领域的研究兴趣激增，部分原因是对高速无线服务和互联网接入的永不满足的需求。人们越来越认识到，网络信息理论的容量定理不仅非常适合为通信系统的物理层设计提供理论框架，而且在为网络和应用层的网络和内容分发协议提供实用的工程见解方面也发挥着重要作用。提出的研究计划旨在通过关注以下领域，获得对网络信息理论中长期存在的容量问题的基本见解。首先，提出的研究将调查无线协作通信的容量限制，并开发实用的编码和调制策略。拟议的研究将利用纠错码和统计物理学方面的最新进展，为广播频道和中继频道等多用户频道制定接近容量的策略。其次，本研究将利用凸优化技术解决网络信息理论中容量限制的计算问题。这在实际的无线多天线通信场景中尤其重要，因为在发射机设计中有许多自由度。最后，将制定受网络问题启发的新型多用户信道模型，并从这些信道的容量和编码研究中获得的见解将用于解决实际网络协议的设计。总体研究计划旨在解决多用户通信中的理论问题，并对网络中信息流的本质获得新的见解。

Network information theory is a fundamental theory of communication in a multiuser environment. This past decade has witnessed a tremendous surge of research interests in this area, driven, in part, by insatiable demands for high-speed wireless service and Internet access. It has been increasingly recognized that not only are the capacity theorems of network information theory ideally suited in providing a theoretical framework for the physical-layer design of communication systems, they also play an instrumental role in providing practical engineering insights for networking and content-distribution protocols at the network and application layers. The proposed research program aims to gain fundamental insights into long-standing capacity problems in network information theory by focusing on the following areas. First, the proposed research will investigate capacity limits and develop practical coding and modulation strategies for wireless cooperative communications. The proposed research will take advantage of recent advances in error-correcting codes and in statistical physics to develop capacity-approaching strategies for multiuser channels, such as the broadcast channel and the relay channel. Second, the proposed research will address the computation of capacity limits in network information theory using convex optimization techniques. This is especially important in practical wireless multi-antenna communication scenarios where many degrees of freedom in transmitter design are available. Finally, novel multiuser channel models inspired by networking problems will be formulated, and insights gained from the study of capacity and coding for these channels will be used to address the design of practical networking protocols. Together, the overall research program aims to tackle theoretical issues in multiuser communication, and to gain fresh insights into the nature of information flow in a network.