所有点对最短路径Floyd-Warshall

动态规划

 

点数V (V1, V2, ... Vn) 二维矩阵dist[n][n] , dist[i][j] 表示vi 与vj点之间的最短距离; 遍历点k, 0<k<=n;

查找(vi, vj) 的最短路径时, 获得dist_k-1[vi][vk] + dist_k-1[vk][vj]的值与dist[vi][vj]的对比,  取最小值;

意识就是遍历每个一个点, 并对比(vi, vj)的当前最短路径与 分割成新两段的路径, 取小的; 

算法时间: O(v^3)

foreach u in V do
  foreach v in V do
     dist[u][v] = inf
     pred[u][v] = -1
  dist[u][v] = 0
  foreach neighbor v of u do
     dist[u][v] = weight of edge(u,v)
     pred[u][v] = u
  // 遍历每个节点 并以该节点将路径分段
foreach t in V do
  foreach u in V do
     foreach v in V do
       nL = dist[u][t] + dist[t][v]
       if nL < dist[u][v] then
          dist[u][v] = nL
          pred[u][v] = pred[t][v]