广域网中的快速组播树生成算法

在组播树生成算法中，MPH(minimum path cost heuristic)的费用性能几乎是最好的，但它的计算时间相对较长，提出了两种新的组播树生成算法：TNS-MPH(tree-mode started minimum-cost path heuristic)和NTDS-MPH(non-tree-destination started minimum-cost path heuristic).同时提出了一种使节点平均度非常精确的随机网络产生模型。新算法的仿真结果表明，新算法用较少的费用性能恶化来换取更快的计算速度。新算法比SCTF（selective closest terminal first）算法有更好的扩展性。

FAST ALGORITHMS FOR MULTICAST TREE CONSTRUCTION IN WIDE AREA NETWORKS

The cost performance of MPH (minimum path cost heuristic) is nearly the best in several algorithms for constructing multicast tree, but the running time is longer. Two new multicast tree constructing algorithms are presented in this paper, which are called TNS MPH (tree node started minimum cost path heuristic) and NTDS MPH (non tree destination started minimum cost path heuristic). And a new random networks generating model is given. The model can construct a network with demanded average node degree. The results of the simulation for the two new algorithms show that the new algorithms can gain faster running speed with less cost performance worsening, and have a better scalability than algorithm SCTF (selective closest terminal first).