简体繁体 English

为什么不用全网络视图就可以使用 Bellman-Ford 算法？

[英]Why can the Bellman-Ford algorithm be used without the full network view?

原文 2021-05-31 13:33:43 5 1 algorithm/ graph-algorithm/ bellman-ford

I have been reading up some applications of the Bellman-Ford algorithm and have come across this particular thread .我一直在阅读 Bellman-Ford 算法的一些应用，并且遇到了这个特定的线程。 From what I can understand from it, it suggests that the Bellman-Ford algorithm does not need the full network view in order to operate properly (find the shortest path).据我所知，这表明 Bellman-Ford 算法不需要完整的网络视图即可正常运行（找到最短路径）。 Why is this?为什么是这样？

Secondly, I want to ask what exactly does "distributed" Bellman Ford mean?其次，我想问一下“分布式”贝尔曼福特究竟是什么意思？ Specifically, how would one go about splitting a massive network so that we can use said "distributed" Bellman Ford.具体来说，go 如何拆分一个庞大的网络，以便我们可以使用所说的“分布式”贝尔曼福特。

Please correct me if I happen to be interpreting any of this incorrectly.如果我碰巧解释错误，请纠正我。 Thanks in advance.提前致谢。

1 个解决方案

The thread that you mentioned in the question discussed the usage of the Bellman-Ford algorithm to find the best routing path to send data packets between routers.您在问题中提到的线程讨论了使用 Bellman-Ford 算法来找到在路由器之间发送数据包的最佳路由路径。 In the beginning, each router starts with a vector of distances to all directly attached networks.一开始，每个路由器都从一个到所有直接连接网络的距离向量开始。 Then, each router broadcast its current distance vector to all the neighboring routers.然后，每个路由器将其当前距离向量广播给所有相邻路由器。 Each of the neighboring routers then will update the local copy of their distance vector and broadcast it again to its neighbors.然后，每个相邻路由器将更新其距离矢量的本地副本并再次将其广播给其邻居。 In this way, each of the source routers of the network ultimately updates their own distance vector w.r.t.这样，网络的每个源路由器最终都会更新自己的距离向量 w.r.t。 the other destination routers.其他目的路由器。 Now you can imagine, all the routers ultimately calculate the shortest routing path (sourced from it) without knowing the complete network.现在你可以想象，所有的路由器最终都会在不知道完整网络的情况下计算出最短的路由路径（来源于它）。