图的深度优先遍历思想是:
从图中某结点出发,访问其某一相邻结点,再访问该结点的相邻结点,直至访问完所有的结点。
形象的比喻就是:一条路走到头,回头再走没走过的路。
可见,深度优先遍历是一种递归思想;
需要注意的是:
对于图的邻接矩阵存储和邻接表存储,深度优先遍历输出的次序有有一定去别的。
对于邻接矩阵而言,DFS和BFS得到的序列是唯一的;
对于邻接表而言,DFS和BFS输入的序列不同,得到的输出序列也不相同。
深度优先遍历的核心算法 :
void DFS(GraAdList G, int v) {
EdgeNode* p;
int j;
cout << G.AdList[v].data << " ";
visited[v] = 1;
p = G.AdList[v].first;
while (p)
{
j = p->adjvex;
if (visited[j] == 0)
{
DFS(G, j);
}
p = p->next;
}
}
void DFS(GraAdList G)
{
for (int i = 0; i < G.vexnum; i++)
{
if (visited[i] == 0)
{
DFS(G, i);
}
}
}
完整代码实现:
#include<iostream>
using namespace std;
#define MAX 6
int visited[MAX];
int D[MAX] = { 9999 };
typedef struct EdgeNode {
int adjvex;
EdgeNode* next;
};
typedef struct VexNode {
char data;
EdgeNode* first;
};
typedef struct GraAdList {
VexNode AdList[MAX];
int vexnum;
int edgenum;
};
//创建邻接矩
void Creat(GraAdList& G) {
int i, j, k;
EdgeNode* e = NULL;
EdgeNode* q = NULL;
cout << "请输入顶点数和边数: " << endl;
cin >> G.vexnum >> G.edgenum;
cout << "请输入顶点信息" << endl;
for (k = 0; k < G.vexnum; k++)
{
cin >> G.AdList[k].data;
G.AdList[k].first = NULL;
}
for (k = 0; k < G.edgenum; k++)
{
cout << "请输入边(vi,vj)的下标i,j: " << endl;
cin >> i >> j;
e = new EdgeNode;
e->adjvex = j;
e->next = G.AdList[i].first;
G.AdList[i].first = e;
}
}
void myprint(GraAdList G) {
cout << endl << "邻接表: " << endl;
EdgeNode* p;
for (int i = 0; i < G.vexnum; i++)
{
cout << G.AdList[i].data << ": ";
for (p = G.AdList[i].first; p; p = p->next)
{
cout << p->adjvex << " ";
}
cout << endl;
}
}
void DFS(GraAdList G, int v) {
EdgeNode* p;
int j;
cout << G.AdList[v].data << " ";
visited[v] = 1;
p = G.AdList[v].first;
while (p)
{
j = p->adjvex;
if (visited[j] == 0)
{
DFS(G, j);
}
p = p->next;
}
}
void DFS(GraAdList G)
{
for (int i = 0; i < G.vexnum; i++)
{
if (visited[i] == 0)
{
DFS(G, i);
}
}
}
int main() {
GraAdList G;
Creat(G);
myprint(G);
for (int i = 0; i < G.vexnum; i++)
{
visited[i] = 0;
}
cout << endl << "深度遍历: " << endl;
DFS(G, 0);
return 0;
}
执行结果:
我创建的图:
版权归原作者 smart_jackli 所有, 如有侵权,请联系我们删除。