针对多个源点需要同时组播数据至所有宿点的多源多宿组播网络,给出了采用线性网络编码实现数据传输并达到最大吞吐率的编码构造方法。通过添加一个虚拟源点以及该虚拟源点至所有源点的虚拟链路,形成了一个含有约束条件的单源组播问题,借助于这个单源组播问题,建立了求各源点组播率的数学模型,该模型是一个组合优化问题——背包问题,并给出了基于遗传算法的实施方法。然后运用实现单源组播连接的线性网络编码构造技术确定各信道的编码矢量。理论分析和仿真测试结果表明,本文提出的方法是可行的。
Aiming at the multi-source multi-sink multicast network in which each source node needs to multicast data to all sink nodes simultaneously,this paper proposes a linear network coding construction approach to implementing data transmission with maximum throughput.By appending an imaginary source node and the imaginary links from the imaginary source node to source nodes,a single-source multicast problem with constraint condition is generated from the original network.To determine the multicast rates of source nodes,a mathematical model is established by virtue of this single-source multicast problem.The model is a combinatorial optimization problem,or knapsack problem,and is solved by the genetic algorithm.Then,the coding vectors of all channels are constructed by linear network coding construction technology for single-source multicast connection.Theoretical analysis and simulation results show that the proposed approach is feasible.