一种改进的多源组播网络的线性网络编码构造方案

Construction Project of Linear Network Coding Improved by Multi-source Cast Network

在传统的线性网络编码算法中,网络中间节点需要执行编码操作,这提高了网络吞吐量,但也增加了编码和解码的计算成本。根据网络中间节点的入度与出度之间的关系,将网络中间节点划分为编码节点和非编码节点,并且最小化网络编码信道的数量,进行随机线性网络编码。采用粒子群优化算法对多源组播网络进行子图划分为多个单源组播网络,各单源组播网络具有相互约束的组播容量,求解包含每个单源组播网络组播能力的Pareto解集。根据权重选择解向量用于改进的线性网络编码构造。仿真测试结果表明,不仅组播容量达到理论最大值,而且降低了时空复杂度,提高了网络的容错性和鲁棒性。

In traditional algorithm of linear network coding,intermediate nodes should execute coding operation so that it improves network throughputs but increases the cost of encoding and decoding.According to the relationship between the input and output of intermediate nodes,encoding nodes and decoding nodes can be classified and the number of coding channels can be minimized to have a random linear network coding.The particle swarm optimization method can divide multi-source cast network into several single source cast networks,which constrain the cast capacity each other and solve the Pareto solution set including cast capacities with several single source cast networks.And the solution vector of weight selection can be used to improve the construction of linear network coding.The result of simulation testing shows that the cast capacities can not only reach the maximum but also decrease the complexity of time and space and improve the fault tolerance and robustness.