| The concept of network coding was first put forward by R. Ahlswede et al. from Chinese University of Hong Kong in 2000. In traditional information processing method, the intermediate nodes in the network just respond for copying and forwarding the information.However, network coding allows the intermediate nodes of the communication networks to code the input information flow before forwarding.Then at the sinks, information is retrieved from what they received. Network coding can achieve the maximum capacity of communication networks, which is determined by the min-cut max-flow theorem.The construction of network coding algorithm is used to obtain the corresponding coding vector for each link as effectively as possible, and then gain the information transmitted over the link by calculating the encoding vectors. The complexity of network coding algorithm might reflect whether the algorithm can be effectively implemented. In this paper, the multicast network coding algorithm has been studied and several results are obtained.The main work of the paper consists of the following three parts:1. Several classical network coding algorithms have been researched, such as exponential-time algorithm, polynomial-time algorithm and random network coding algorithm.2. Based on the generic linear network coding algorithm, and combined with the usage of discrete routing, a novel improved multicast network coding algorithm is proposed in this paper. The rationality of the algorithm has been proved. Complexity analysis indicates that the algorithm in this paper reduces the computational complexity obviously compared with generic algorithm.3. A routing algorithm for network coding multicast based on the shortest path is proposed. The algorithm, together with the existing encoding scheme, can construct a practical multicast system. Simulations show that, for network coding multicast, performance has been improved significantly in comparison with the traditional IP multicast in terms of throughput. |
PDF Full Text Request