POJ 3159 Candies 【差分约束系统 dijkstra+heap/spfa+stack】

POJ 3159 Candies

题目链接：vjudge传送门

题目大意：
给n个人分糖果，m组数据a，b，c；意思是a比b少的糖果个数绝对不超过c个，也就是d(b)-d(a) < c，求1比n少的糖果数的最大值。

具体思路：
差分约束系统第一次接触，我也不太懂，见博客
据说spfa+queue会超时，看好多人都用spfa+stack优化【第一次知道能用stack优化，但是我仍不知两者的差别】
故贴出dijkstra+优先队列的代码

其中关于优先队列的元素间的比较，用自定义比较结构体，重载()，一直wa，最后才用的重载<号运算符，至今不知为何错误。重载()慎用，最后我发现了pair有默认的比较函数，默认，比较first，first相同则比较second，对于dij+优先队列很方便code

具体代码：

//重载< ac code
#include
#include
#include
#include
using namespace std;
const int N = 3e4 + 5, M = 15e4 + 5;
struct Node {
     
	int v, w, next;
}edg[M];
int visit[N];
int h[N];
int d[N];
int n, m;
int cnt = 1;

struct cmp {
     
	int v, d;
	cmp(int d, int v) : d(d), v(v) {
     }
	bool operator < (const cmp &w) const {
     
		return d > w.d;
	}
};

void add(int u, int v, int w)
{
     
	edg[cnt].v = v;
	edg[cnt].w = w;
	edg[cnt].next = h[u];
	h[u] = cnt++;
}
void dijkstra()
{
     
	memset(d, 0x3f, sizeof(d));
	d[1] = 0;
	priority_queue<cmp> q;
	q.push(cmp(0, 1));
	while (q.size())
	{
     
		int t = q.top().v;
		q.pop();
		if (visit[t])continue;
		visit[t] = 1;
		for (int i = h[t]; i; i = edg[i].next)
		{
     
			int v = edg[i].v;
			int w = edg[i].w;
			if (d[v] > d[t] + w) {
     
				d[v] = d[t] + w;
				q.push(cmp(d[v], v));
			}
		}
	}
}
int main()
{
     
	scanf("%d%d", &n, &m);
	for (int i = 1; i <= m; i++)
	{
     
		int u, v, w;	//u-v<=w
		scanf("%d%d%d", &u, &v, &w);
		add(u, v, w);
	}
	dijkstra();
	printf("%d", d[n]);
	return 0;
}

//pair ac code
#include
#include
#include
#include
using namespace std;
typedef pair<int, int> P;
const int N = 3e4 + 5, M = 15e4 + 5;
struct Node {
     
	int v, w, next;
}edg[M];
int visit[N];
int h[N];
int d[N];
int n, m;
int cnt = 1;
void add(int u, int v, int w)
{
     
	edg[cnt].v = v;
	edg[cnt].w = w;
	edg[cnt].next = h[u];
	h[u] = cnt++;
}
void dijkstra()
{
     
	memset(d, 0x3f, sizeof(d));
	d[1] = 0;
	priority_queue<P, vector<P>, greater<P>> q;
	q.push(P (0,1));
	while (q.size())
	{
     
		int t = q.top().second;
		q.pop();
		if (visit[t])continue;
		visit[t] = 1;
		for (int i = h[t]; i; i = edg[i].next)
		{
     
			int v = edg[i].v;
			int w = edg[i].w;
			if (d[v] > d[t] + w) {
     
				d[v] = d[t] + w;
				q.push(P(d[v], v));
			}
		}
	}
}
int main()
{
     
	scanf("%d%d", &n, &m);
	for (int i = 1; i <= m; i++)
	{
     
		int u, v, w;	//u-v<=w
		scanf("%d%d%d", &u, &v, &w);
		add(u, v, w);
	}
	dijkstra();
	printf("%d", d[n]);
	return 0;
}

分割线

---

贴一下一直wa的代码，留个坑

//WA code
#include
#include
#include
#include
using namespace std;
const int N = 3e4 + 5, M = 15e4 + 5;
struct Node {
     
	int v, w, next;
}edg[M];
int visit[N];
int h[N];
int d[N];
int n, m;
int cnt = 1;
struct cmp {
     
	bool operator()(int a, int b)
	{
     
		return d[a] > d[b];
	}
};
void add(int u, int v, int w)
{
     
	edg[cnt].v = v;
	edg[cnt].w = w;
	edg[cnt].next = h[u];
	h[u] = cnt++;
}
void dijkstra()
{
     
	memset(d, 0x3f, sizeof(d));
	d[1] = 0;
	priority_queue<int, vector<int>, cmp> q;
	q.push(1);
	while (q.size())
	{
     
		int t = q.top();
		q.pop();
		if (visit[t])continue;
		visit[t] = 1;
		for (int i = h[t]; i; i = edg[i].next)
		{
     
			int v = edg[i].v;
			int w = edg[i].w;
			if (d[v] > d[t] + w) {
     
				d[v] = d[t] + w;
				q.push(v);
			}
		}
	}
}
int main()
{
     
	scanf("%d%d", &n, &m);
	for (int i = 1; i <= m; i++)
	{
     
		int u, v, w;	//u-v<=w
		scanf("%d%d%d", &u, &v, &w);
		add(u, v, w);
	}
	dijkstra();
	printf("%d", d[n]);
	return 0;
}