C++和Python实现dijkstra算法

迪杰斯特拉(Dijkstra)算法是典型最短路径算法，用于计算一个节点到其他节点的最短路径。它的主要特点是以起始点为中心向外层层扩展(广度优先搜索思想)，直到扩展到终点为止。

C++语言实现：

#include<bits/stdc++.h>
   
using namespace std;

int tot = 0;                               //记录边数
int head[100];                             //记录每个几点对应的起始边，顺时针方向数，最右为其实边
int dis[100];                              //记录到各点的距离
int N;                                     //总节点数
int visit[100];                            //是否访问过

struct arc{                                //边的表示
    int end;                               //边的终止节点
    int next;                              //同起点邻边
    int weight;                            //权重

    arc(int x = 0, int y = -1, int w = -1){    //结构体默认是public的构造函数
        end = x;
        next = y;
        weight = w;
    }
}e[100];

void addE(int u, int v, int w){       //增加一条从u 到 v的边
    e[tot] = arc(v, head[u], w);
    head[u] = tot;
    tot++;
}


int getMin(){
    int min = 99999;
    int pos = -1;
    for(int i = 0; i < N; i++){
        if(dis[i] >0 && dis[i] < min && !visit[i]){
            min = dis[i];
            pos = i;
        }
    }
    return pos;
}

/*
*   dijkstra算法是单源最短路径算法，所以给定参数u为起点
*   u:单源中的起点
*/
void dijkstra(int u){
    for(int i = head[u]; i != -1; i = e[i].next)
        dis[e[i].end] = e[i].weight;   //初始化权重

    int count = 1;
    visit[u] = 1;
    dis[u] = 0;

    while(count < N){
        int pos = getMin();  //找到最小值
        visit[pos] = 1;
        count++;

        for(int i = head[pos]; i != -1; i = e[i].next)     //以pos节点为中间节点，更新单源最短路径
            if((dis[pos] + e[i].weight < dis[e[i].end]) || dis[e[i].end] == -1)
                dis[e[i].end] = dis[pos] + e[i].weight;


    }

    for(int i = 0; i < N; i++){
        if(dis[i] >= 0)
            cout<<u<<" 到 "<<i<<" 的最短路径为: "<<dis[i]<<endl;
        else
            cout<<u<<", "<<i<<" 不可达!"<<endl;
    }
}


int main()
{
    N = 6;

    memset(head, -1, sizeof(head));
    memset(dis, -1, sizeof(dis));    //初始化dis为-1,意为不可达
    memset(visit, 0, sizeof(visit));  //初始化visit为0，表示都未访问过

    addE(0, 2, 10);
    addE(0, 4, 30);
    addE(0, 5, 100);

    addE(1, 2, 5);

    addE(2, 3, 50);

    addE(3, 5, 10);

    addE(4, 3, 20);
    addE(4, 5, 60);

    dijkstra(0);

    return 0;
}

python语言实现：

import math
import sys

class PriorityQueue:
    # Based on Min Heap
    def __init__(self):
        self.cur_size = 0
        self.array = []
        self.pos = {}   # To store the pos of node in array

    def isEmpty(self):
        return self.cur_size == 0

    def min_heapify(self, idx):
        lc = self.left(idx)
        rc = self.right(idx)
        if lc < self.cur_size and self.array(lc)[0] < self.array(idx)[0]:
            smallest = lc
        else:
            smallest = idx
        if rc < self.cur_size and self.array(rc)[0] < self.array(smallest)[0]:
            smallest = rc
        if smallest != idx:
            self.swap(idx, smallest)
            self.min_heapify(smallest)

    def insert(self, tup):
        # Inserts a node into the Priority Queue
        self.pos[tup[1]] = self.cur_size
        self.cur_size += 1
        self.array.append((sys.maxsize, tup[1]))
        self.decrease_key((sys.maxsize, tup[1]), tup[0])

    def extract_min(self):
        # Removes and returns the min element at top of priority queue
        min_node = self.array[0][1]
        self.array[0] = self.array[self.cur_size - 1]
        self.cur_size -= 1
        self.min_heapify(1)
        del self.pos[min_node]
        return min_node

    def left(self, i):
        # returns the index of left child
        return 2 * i + 1

    def right(self, i):
        # returns the index of right child
        return 2 * i + 2

    def par(self, i):
        # returns the index of parent
        return math.floor(i / 2)

    def swap(self, i, j):
        # swaps array elements at indices i and j
        # update the pos{}
        self.pos[self.array[i][1]] = j
        self.pos[self.array[j][1]] = i
        temp = self.array[i]
        self.array[i] = self.array[j]
        self.array[j] = temp

    def decrease_key(self, tup, new_d):
        idx = self.pos[tup[1]]
        # assuming the new_d is atmost old_d
        self.array[idx] = (new_d, tup[1])
        while idx > 0 and self.array[self.par(idx)][0] > self.array[idx][0]:
            self.swap(idx, self.par(idx))
            idx = self.par(idx)


class Graph:
    def __init__(self, num):
        self.adjList = {}   # To store graph: u -> (v,w)
        self.num_nodes = num    # Number of nodes in graph
        # To store the distance from source vertex
        self.dist = [0] * self.num_nodes
        self.par = [-1] * self.num_nodes  # To store the path

    def add_edge(self, u, v, w):
        #  Edge going from node u to v and v to u with weight w
        # u (w)-> v, v (w) -> u
        # Check if u already in graph
        if u in self.adjList.keys():
            self.adjList[u].append((v, w))
        else:
            self.adjList[u] = [(v, w)]

        # Assuming undirected graph
        if v in self.adjList.keys():
            self.adjList[v].append((u, w))
        else:
            self.adjList[v] = [(u, w)]

    def show_graph(self):
        # u -> v(w)
        for u in self.adjList:
            print(u, '->', ' -> '.join(str("{}({})".format(v, w))
                                       for v, w in self.adjList[u]))

    def dijkstra(self, src):
        # Flush old junk values in par[]
        self.par = [-1] * self.num_nodes
        # src is the source node
        self.dist[src] = 0
        Q = PriorityQueue()
        Q.insert((0, src))  # (dist from src, node)
        for u in self.adjList.keys():
            if u != src:
                self.dist[u] = sys.maxsize  # Infinity
                self.par[u] = -1

        while not Q.isEmpty():
            u = Q.extract_min()  # Returns node with the min dist from source
            # Update the distance of all the neighbours of u and
            # if their prev dist was INFINITY then push them in Q
            for v, w in self.adjList[u]:
                new_dist = self.dist[u] + w
                if self.dist[v] > new_dist:
                    if self.dist[v] == sys.maxsize:
                        Q.insert((new_dist, v))
                    else:
                        Q.decrease_key((self.dist[v], v), new_dist)
                    self.dist[v] = new_dist
                    self.par[v] = u

        # Show the shortest distances from src
        self.show_distances(src)

    def show_distances(self, src):
        print("Distance from node: {}".format(src))
        for u in range(self.num_nodes):
            print('Node {} has distance: {}'.format(u, self.dist[u]))

    def show_path(self, src, dest):
        # To show the shortest path from src to dest
        # WARNING: Use it *after* calling dijkstra
        path = []
        cost = 0
        temp = dest
        # Backtracking from dest to src
        while self.par[temp] != -1:
            path.append(temp)
            if temp != src:
                for v, w in self.adjList[temp]:
                    if v == self.par[temp]:
                        cost += w
                        break
            temp = self.par[temp]
        path.append(src)
        path.reverse()

        print('----Path to reach {} from {}----'.format(dest, src))
        for u in path:
            print('{}'.format(u), end=' ')
            if u != dest:
                print('-> ', end='')

        print('\nTotal cost of path: ', cost)


if __name__ == '__main__':
    graph = Graph(9)
    graph.add_edge(0, 1, 4)
    graph.add_edge(0, 7, 8)
    graph.add_edge(1, 2, 8)
    graph.add_edge(1, 7, 11)
    graph.add_edge(2, 3, 7)
    graph.add_edge(2, 8, 2)
    graph.add_edge(2, 5, 4)
    graph.add_edge(3, 4, 9)
    graph.add_edge(3, 5, 14)
    graph.add_edge(4, 5, 10)
    graph.add_edge(5, 6, 2)
    graph.add_edge(6, 7, 1)
    graph.add_edge(6, 8, 6)
    graph.add_edge(7, 8, 7)
    graph.show_graph()
    graph.dijkstra(0)
    graph.show_path(0, 4)