The All-Pairs Shortest Path Problem FLOYD

       Let G = (V,E) be a directed graph in which each edge (i, j) has a nonnegative length l[i, j]. If there is no edge from vertex i to vertex j, then l[i, j] = ∞. The problem is to find the distance from each vertex to all other vertices, where the distance from vertex x to vertex y is the length of a shortest path from x to y.

       Assuming there is no negative length circuit in the directed graph.

         For simplicity, we will assume that V = {1, 2, . . ., n}. Let i and j be two different vertices in V . Define d**k(i,j) to be the length of a shortest path from i to j that does not pass through any vertex in {k + 1, k + 2, . . . , n}. Thus, for example, d**0(i,j) = l[i, j], d**1(i,j) is the length of a shortest path from i to j that does not pass through any vertex except possibly vertex 1, d**2(i,j) is the length of a shortest path from i to j that does not pass through any vertex except possibly vertex 1 or vertex 2 or both, and so on. Then, by definition, d**n(i,j) is the length of a shortest path from i to j, i.e., the distance from i to j. So we can use n + 1 matrices D0,D1, . . . , Dn of dimension n × n to compute the lengths of the shortest constrained paths, i.e., Dk= d**k(i,j)

        By the definition of d**k(i,j), we have:

 Vi, A list of vertices with each vertex serial number no greater than k, Vj

        We can divide all such paths into two disjoint subsets: paths in one subset do not pass the k-th vertex, while paths in the other subset pass the k-th vertex. In the first subset, the serial number of vertices contained in the path will not be greater than k-1. According to our definition, the length of the shortest path is  d**k-1(i,j).

For the second subset, we have:

Vi, vertices with serial number ≤ k-1, Vk , vertices with serial number ≤ k-1, Vj   .

        In other words, each path is composed of two paths: a path from Vi to Vk, where the number of each vertex in the path is not greater than k-1; A path from Vk  to Vj, where the number of each vertex in the path is not greater than k-1

任意两个节点之间的最短路径。假设输入是一个 n x n 的矩阵，表示有向图中每条边的长度，其中矩阵的第 i 行第 j 列表示从节点 i 到节点 j 的边的长度。输出是一个矩阵 D，其中 D[i, j] 表示从节点 i 到节点 j 的最短路径长度。

1. 首先，将输入矩阵复制到一个新的矩阵 D 中。

2. 然后，使用三重循环来计算最短路径长度。外层循环 k 从 1 到 n，表示允许经过节点 k 的路径。中间两层循环 i 和 j 分别表示起点和终点节点，计算经过节点 k 的路径和不经过节点 k 的路径哪个更短，更新矩阵 D。

3. 最后，矩阵 D 中的值就是从节点 i 到节点 j 的最短路径长度。

D0中的位置（1，1）的0意思是第1个节点到自己本身的距离 （1，2）中的2指的是横轴节点1到纵轴节点2的距离 （3，2）的∞则是横轴节点3到纵轴节点2没有路径则距离为∞ 

D1则发生了变化，所有路径可以经过节点1  距离可能发生变化的有

（2， 3） （3，2） 因为对角线距离不可能发生变化，而从节点1出发或从以节点1为终点都已经经过节点1也不会发生变化即横轴为1纵轴为1的距离都不会发生变化。

如何计算变化后的值（2，3）中看（2，1） （1，3）即横轴为2以及纵轴为3的节点中的没有发生变化且距离不为0的节点，距离之和9+8=17>6则距离不变；（3，2）中看（1，3）（2，1）的距离之和1+2=3<∞ 则距离之和取代原有距离 即（3，2）中的距离变为3

其中的原理是（2，1）即节点2到节点1距离 （1，3）节点1到节点3距离 其总和与原有节点2直接到节点3距离作比较 求出最短路径

一文足矣——动态规划经典之Floyd（弗洛伊德）算法_floyd算法负权回路-CSDN博客

 求最短路径Floyd算法！_哔哩哔哩_bilibili