The Research On Network Utility Maximization Of Coded Packet Network And Application Of Network Coding

Network coding is introduced by Ahlswede in 2000, which is based on the conception of network in formation flow. If coding is permited by the network vertices the max-flow bound of network multicast can be achieved, i.e .the theoretical upper bound of network resource utilization, which can not always be achieved by traditional copying and routing. Furthermore, through network coding, we can obtain the benefits of saving in bandwidth, load balancing, optimizing the energy usage of energy-limited networks and so on. Now, the research of network coding has received great interests and has become one of the most attractive fields in network information theory.This dissertation mainly studies network coding. Firstly, the design of network coding is deeply analysed. Based on this, in theory aspect, the network utility maximum of coded packet network is studied. In application aspect, we study the effection of random linear network coding on file download success rate or download time in wireless Mesh network, and the problem of continuous real-time data collection based on network coding in sensor network. The dissertation includes the following:Chapter 1 introduces the background of network coding been raised, the concept of network coding, the research status of network coding, and the work of the dissertation.Chapter 2 summarizes the research results on the design of network coding, introduces the design scheme of linear network coding, random linear network coding, delayed network coding used on acyclic network. In addition, the practical network coding used in real network is discribed. The chapter is the theoretical basis of following chapters.Chapter 3 studies the problem of network utility maximization of single session coded packet network. Develops a distributed projection gradient algorithm for the problem, and proves the sufficient but not necessary condition that can make the proposed algorithm converge to the globally optimal solutions. Numerical examples are provided to validate the correctness of the algorithm. Chapter 4 studies the problem of network utility maximization of coding packet network under delivery contracts constraint when coding subgraph is given. Develops a distributed projection gradient algorithm for the problem model. Proves sufficient but not necessary condition that can make the proposed algorithm converge to the globally optimal solutions. Numerical examples are provided to complement the theoretical analysis.Chapter 5 studies the performance gain brought by random network coding in wireless Mesh nework when file-sharing is conducted. Through simulation, proves the performance gain brought by random network coding compared to the traditional route in following three aspects: nodes be upload -limited but MAC layer be ideal conflict-free; MAC layer be conflict; nodes be mobile and MAC be conflict.Chapter 6 studies continuous real-time data collection problem in wireless sensor networks through network coding. Data collector can collect data in continuous real-time manner through using partial network coding in source nodes and random network coding in relay storage nodes.Finally, the dissertation is summarized, and future research on network coding are put forward.