Codeforces 459E. Pashmak and Graph (DP)

Description
Pashmak’s homework is a problem about graphs. Although he always tries to do his homework completely, he can’t solve this problem. As you know, he’s really weak at graph theory; so try to help him in solving the problem.

You are given a weighted directed graph with n vertices and m edges. You need to find a path (perhaps, non-simple) with maximum number of edges, such that the weights of the edges increase along the path. In other words, each edge of the path must have strictly greater weight than the previous edge in the path.

Help Pashmak, print the number of edges in the required path.

Input
The first line contains two integers $n,m(2\le n\le 3\cdot 10^5;1\le m\le min(n\cdot (n-1),3\cdot 10^5))$. Then, m lines follows. The i-th line contains three space separated integers: $u_i,v_i,w_i(1\le u_i,v_i\le n;1\le w_i\le 10^5)$ which indicates that there’s a directed edge with weight $w_i$ from vertex $u_i$ to vertex $v_i$.

It’s guaranteed that the graph doesn’t contain self-loops and multiple edges.

Output
Print a single integer — the answer to the problem.

Examples
Input
3 3
1 2 1
2 3 1
3 1 1
Output
1

Input
3 3
1 2 1
2 3 2
3 1 3
Output
3

Input
6 7
1 2 1
3 2 5
2 4 2
2 5 2
2 6 9
5 4 3
4 3 4
Output
6

Solution
定义$edp[i]$：以边i结尾的最长路径，$ndp[i]$ ：以点i结尾的最长路径
因为路径上的边权需要递增，所以需要按边权从小到大的顺序做DP转移

	//非严格递增的转移方程（按边权顺序）
    int u = e.u,v = e.v,id = e.id;// 分别为有向边的两端点和边的编号
	edp[id] = ndp[u] + 1;
	ndp[v] = max(ndp[v],edp[id]);

然而这里需要严格递增，所以对于边权相同的边，他们的“拓扑序”是相同的，不应该出现状态转移的先后
我的方法是先 做$Nod{e}_u->Edg{e}_{id}$ 转移，全部转移完后再做 $Edg{e}_{id}->Nod{e}_v$ 转移
这样就不会受到相同权值边的影响

Hint
对于一些问题，重点就是要找到其“拓扑序”，从而满足无后效性的特点，才能使用DP进行转移

Code

#include <bits/stdc++.h>
#define pb push_back
using namespace std;
const int maxn = 3e5 + 5;
const int maxv = 1e5 + 7;
int lim;
int ndp[maxn],edp[maxn];
struct Edge{
	int u,v,id;
};
vector<Edge>val[maxv];
int main(){
	int n,m;scanf("%d%d",&n,&m);
	for(int i = 1;i <= m;++i){
		int u,v,w;scanf("%d%d%d",&u,&v,&w);
		val[w].pb((Edge){u,v,i});
		lim = max(lim,w);
	}
	int res = 0;
	for(int w = 1;w <= lim;++w){
		for(auto e : val[w]){
			int u = e.u,v = e.v,id = e.id;
			edp[id] = ndp[u] + 1;
		}
		for(auto e : val[w]){
			int u = e.u, v = e.v,id = e.id;
			ndp[v] = max(ndp[v],edp[id]);
		}
	}
	for(int i = 1;i <= n;++i) res = max(res,ndp[i]);
	printf("%d\n", res);
    return 0;
}