求两点间最短路的Floyd算法及其matlab程序详解

#################本文为学习《图论算法及其MATLAB实现》的学习笔记#################

Floyd算法利用动态规划算法的基本思想，核心思想是在找到最短路径矩阵后，将所有顶点设置为中转点重新遍历更新最短路径矩阵。

Floyd算法的基本思想

Floyd算法的基本步骤

 程序参数说明

W: 图的权值矩阵 
P: 最短路
u: 最短路的权和

算法的matlab程序详解

%Floyd算法
function [ P, u ] = f_path(W)
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%% 输入：     W: 图的权值矩阵 
%%%%%%%%% 输出：     P: 最短路
%%%%%%%%%            u: 最短路的权和
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%


%%%%%%%%% 初始化 %%%%%%%%%%
n = length(W); % 计算顶点数
U = W;         % 保存任意两点之间最短路径的长度
m = 1;

%%%%%%%%%%%%%%% 步骤2 %%%%%%%%%%%%%%%
while m <= n   % 判断是否满足停止条件
    for i = 1:n
        for j = 1:n
            if U(i,j) > U(i,m) + U(m,j)
                U(i,j) = U(i,m) + U(m,j);       %更新dij
            end
        end
    end
    m = m+1;
end
u = U(1,n);     % 始点到终点的最短距离

%%%%%%%%%%%%% 输出最短路的顶点 %%%%%%%%%%%%%
P1 = zeros(1,n);        % P1 为最短路径的反序
k = 1;
P1(k) = n;
V = ones(1,n) * inf;
kk = n;
while kk ~= 1           % 判断此时终点是否为始点，即是否找完所有中转点
    for i = 1:n         % 寻找所有中转点
        V(1,i) = U(1,kk) - W(i,kk);  % 计算始点到终点的最短距离与中转点到终点最短距离的差值
        if V(1,i) == U(1,i)
            P1(k+1) = i;        % 将此时的中转点增加到最短路径队列
            kk = i;             % 将此时的中转点设置为新的终点
            k = k+1;
        end
    end
end

%%%%%%%%% 将 P1 的反序路径重新正序排列 %%%%%%%%%
k = 1;
wrow = find(P1 ~= 0);           % 找到最短路径中的所有顶点
for j = length(wrow):(-1):1     % 由最短路径的总顶点数以 -1 为步长反向遍历至始点
    P(k) = P1(wrow(j));         % 重新排列最短路径队列
    k = k+1;
end
P;