711D

19992144

于 2016-08-30 13:53:39 发布

阅读量219

点赞数

分类专栏： tarjan

本文链接：https://blog.csdn.net/xhk2000/article/details/52368633

版权

tarjan 专栏收录该内容

1 篇文章 0 订阅

订阅专栏

ZS the Coder and Chris the Baboon has explored Udayland for quite some time. They realize that it consists of n towns numbered from 1 to n.

There are n directed roads in the Udayland. i-th of them goes from town i to some other town a_i (a_i ≠ i). ZS the Coder can flip the direction of any road in Udayland, i.e. if it goes from town A to town B before the flip, it will go from town B to town A after.

ZS the Coder considers the roads in the Udayland confusing, if there is a sequence of distinct towns A₁, A₂, ..., A_k (k > 1) such that for every 1 ≤ i < k there is a road from town A_i to town A_i + 1 and another road from town A_k to town A₁. In other words, the roads are confusing if some of them form a directed cycle of some towns.

Now ZS the Coder wonders how many sets of roads (there are 2ⁿ variants) in initial configuration can he choose to flip such that after flipping each road in the set exactly once, the resulting network will not be confusing.

Note that it is allowed that after the flipping there are more than one directed road from some town and possibly some towns with no roads leading out of it, or multiple roads between any pair of cities.

Input

The first line of the input contains single integer n (2 ≤ n ≤ 2·10⁵) — the number of towns in Udayland.

The next line contains n integers a₁, a₂, ..., a_n (1 ≤ a_i ≤ n, a_i ≠ i), a_i denotes a road going from town i to town a_i.

Output

Print a single integer — the number of ways to flip some set of the roads so that the resulting whole set of all roads is not confusing. Since this number may be too large, print the answer modulo 10⁹ + 7.

tarjan裸题

一个强连通分量（环）（1<<环的大小）-2；

其他为1<<k k=n-环的元素总和；

两者相乘

tarjan还是不太懂。。。

代码：

#include<stack>
#include <set>
#include <map>
#include <queue>
#include <deque>
#include <cstdio>
#include <string>
#include <vector>
#include <math.h>
#include <time.h>
#include <utility>
#include <cstdlib>
#include <sstream>
#include <cstring>
#include <stdio.h>
#include <iomanip>
#include <iostream>
#include <algorithm>
using namespace std;
const int inf=1e9+7,SIZE=200010;
long long n,stamp,ans=1,sum,j;
int used[SIZE],dfn[SIZE],low[SIZE];
stack<int>s;
vector<int>graph[SIZE];
void tarjan(int u)
{
   dfn[u]=low[u]=++stamp;
   used[u]=1;
   s.push(u);
   for(int i=0;i<graph[u].size();i++)
   {
       int v=graph[u][i];
       if(used[v]==0)
       {
           tarjan(v);
           low[u]=min(low[u],low[v]);
       }
       else if(used[v]==1)
       {
           low[u]=min(low[u],low[v]);
       }
   }
   if(low[u]==dfn[u])
   {
       int temp=0;
       while(s.top()!=u)
       {
           used[s.top()]=-1;
           s.pop();
           temp++;
       }
       used[s.top()]=-1;
       s.pop();
       temp++;
       if(temp>1)
       {
           int num=1;
           for(int i=1;i<=temp;i++)
           {
               num*=2;
               num%=inf;
           }
           num=(num+inf-2)%inf;
           ans*=num%inf;
           ans%=inf;
       }
       else if(temp==1)
       {
           ans*=2;
           ans%=inf;
       }
   }
}
int main()
{
   cin>>n;
   for(int i=1;i<=n;i++)
   {
       scanf("%d",&j);
       graph[i].push_back(j);
   }
   for(int i=1;i<=n;i++)
   {
       if(!used[i])
       {
           tarjan(i);
       }
   }
   cout<<ans<<endl;
   return 0;
}