[dp][图论]Node Pairs Codeforces1763E

Let's call an ordered pair of nodes (u,v)(u,v) in a directed graph unidirectional if u≠vu≠v, there exists a path from uu to vv, and there are no paths from vv to uu.

A directed graph is called pp-reachable if it contains exactly pp ordered pairs of nodes (u,v)(u,v) such that u<vu<v and uu and vv are reachable from each other. Find the minimum number of nodes required to create a pp-reachable directed graph.

Also, among all such pp-reachable directed graphs with the minimum number of nodes, let GG denote a graph which maximizes the number of unidirectional pairs of nodes. Find this number.

Input

The first and only line contains a single integer pp (0≤p≤2⋅1050≤p≤2⋅105) — the number of ordered pairs of nodes.

Output

Print a single line containing two integers — the minimum number of nodes required to create a pp-reachable directed graph, and the maximum number of unidirectional pairs of nodes among all such pp-reachable directed graphs with the minimum number of nodes.

Examples

input

3

output

3 0

input

4

output

5 6

input

0

output

0 0

题意： 给出一个参数p，要求能构成p可达图的最少点数，以及在这个点数下单向连通点对的最大值。其中p可达即为u < v且从u能到v从v也能到u的(u, v)个数，单向连通即u < v且从u能到v或v能到u的(u, v)个数。

分析： 这道题的关键在于考虑整幅图是由什么构成的，其实任何一个有向图都可以看作由若干强连通分量连接而成，这里的连接指的是一些单向边，也就是连接以后还是这些个强连通分量，它们不会增大也不会减小。其实知道了p可达图最少点数就知道了最多的单向连通数了，因为(u, v)之间的关系就三种，要么是双向连通，要么是单向连通，要么是不连通，而(u, v)的总数一共是n*(n-1)/2个，现在已知其中有p个是双向连通，那么单向连通+不连通个数就是n*(n-1)/2-p个，由于要单向连通数尽量多，那么自然就是n*(n-1)/2-p个了。

每个强连通分量对于答案的贡献是size*(size-1)/2，size是该强连通分量中点的个数，最后p就是由若干强连通分量的贡献加和得到，求解过程可以用dp处理，设dp[i]表示i可达图的最少点数，dp[1] = 2，然后枚举其中某个连通分量的点个数来进行更新，也就是dp[i] = min(dp[i], dp[i-j*(j-1)/2]+j)。

具体代码如下：

#include <iostream>
#include <cstdio>
#include <algorithm>
#include <cstring>
#include <cmath>
#include <string>
#define inf 0x3f3f3f3f3f3f3f3f
#define int long long
using namespace std;

int dp[200005];

signed main(){
	int p;
	cin >> p;
	dp[1] = 2;
	for(int i = 2; i <= p; i++){
		dp[i] = inf;
		for(int j = 1; j*(j-1)/2 <= i; j++)
			dp[i] = min(dp[i], dp[i-j*(j-1)/2]+j);
	}
	printf("%lld %lld", dp[p], dp[p]*(dp[p]-1)/2-p);
	
    return 0;
}