基于Ticket探测的非精确单播路由算法仿真  

Simulation of Ticket-based Unicast Routing Algorithm for Network with Imprecise State Information

作　　者：王文凯[1] 黄胜华[1] 

机构地区：[1]中国科学技术大学电子工程与信息科学系,安徽合肥230027

出　　处：《计算机仿真》2007年第2期110-113,共4页Computer Simulation

摘　　要：TBP(Ticket-based Probing)算法是一种有效的非精确网络状态下单播路由算法。它具有较好的可行性,并且优化了平均路由代价,但仿真时该算法仅针对均匀分布的网络时延变化,因此在应用时有一定的局限性。本文仿真分析了其在均匀分布和正态分布两种时延变化模型下算法的性能,以及不同探测包生成方案对算法性能的影响。仿真结果表明,TBP算法在正态分布下的寻路成功率通常比均匀分布下的要低,而不同探测包生成方案对算法性能各指标的影响各有不同。TBP（Ticket - based Probing） algorithm is an effective unicast routing algorithm under imprecise network state information. It is easy to be implemented and the average routing cost is optimized, but in simulation the algorithm only considers network delay under uniform distribution, so it has certain limitation in application. Via extensive simulation, the performances of the algorithm under both the uniform distribution and normal distribution are analyzed and the impacts of different probe generation methods on performance are compared. Results indicate that TBP algorithm has relatively lower success ratio under normal distribution than under uniform distribution, and different probe generation methods have different influences on the performance metrics.

关 键 词：单播路由 服务质量 时延约束 非精确 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]