D. Dirty Deeds Done Dirt Cheap(思维,排序）

https://codeforces.com/problemset/problem/1148/D

---

 

题目描述

You are given nn pairs of integers (a_1, b_1), (a_2, b_2), \ldots, (a_n, b_n)(a1​,b1​),(a2​,b2​),…,(an​,bn​) . All of the integers in the pairs are distinct and are in the range from 11 to 2 \cdot n2⋅n inclusive.

Let's call a sequence of integers x_1, x_2, \ldots, x_{2k}x1​,x2​,…,x2k​ good if either

• x_1 < x_2 > x_3 < \ldots < x_{2k-2} > x_{2k-1} < x_{2k}x1​<x2​>x3​<…<x2k−2​>x2k−1​<x2k​ , or
• x_1 > x_2 < x_3 > \ldots > x_{2k-2} < x_{2k-1} > x_{2k}x1​>x2​<x3​>…>x2k−2​<x2k−1​>x2k​ .

You need to choose a subset of distinct indices i_1, i_2, \ldots, i_ti1​,i2​,…,it​ and their order in a way that if you write down all numbers from the pairs in a single sequence (the sequence would be a_{i_1}, b_{i_1}, a_{i_2}, b_{i_2}, \ldots, a_{i_t}, b_{i_t}ai1​​,bi1​​,ai2​​,bi2​​,…,ait​​,bit​​ ), this sequence is good.

What is the largest subset of indices you can choose? You also need to construct the corresponding index sequence i_1, i_2, \ldots, i_ti1​,i2​,…,it​ .

输入格式

The first line contains single integer nn ( 2 \leq n \leq 3 \cdot 10^52≤n≤3⋅105 ) — the number of pairs.

Each of the next nn lines contain two numbers — a_iai​ and b_ibi​ ( 1 \le a_i, b_i \le 2 \cdot n1≤ai​,bi​≤2⋅n ) — the elements of the pairs.

It is guaranteed that all integers in the pairs are distinct, that is, every integer from 11 to 2 \cdot n2⋅n is mentioned exactly once.

输出格式

In the first line print a single integer tt — the number of pairs in the answer.

Then print tt distinct integers i_1, i_2, \ldots, i_ti1​,i2​,…,it​ — the indexes of pairs in the corresponding order.

题意翻译

题目描述

有 nn 个整数对 (a_1, b_1), (a_2, b_2), \cdots, (a_n, b_n)(a1​,b1​),(a2​,b2​),⋯,(an​,bn​). 保证 a_1, b_1, a_2, b_2, \cdots, a_n, b_na1​,b1​,a2​,b2​,⋯,an​,bn​ 两两不相等, 并且均在区间 [1, 2 \cdot n][1,2⋅n] 内.

好序列的定义:
对于一个序列 x_1, x_2, \cdots, x_{2k}x1​,x2​,⋯,x2k​, 满足

• x_1 < x_2 > x_3 < \cdots < x_{2k - 2} > x_{2k - 1} < x_{2k}x1​<x2​>x3​<⋯<x2k−2​>x2k−1​<x2k​ 或
• x_1 > x_2 < x_3 > \cdots > x_{2k - 2} < x_{2k - 1} > x_{2k}x1​>x2​<x3​>⋯>x2k−2​<x2k−1​>x2k​.

求一个序列 i_1, i_2, \cdots, i_ti1​,i2​,⋯,it​ 满足 a_{i_1}, b_{i_1}, a_{i_2}, b_{i_2}, \cdots, a_{i_t}, b_{i_t}ai1​​,bi1​​,ai2​​,bi2​​,⋯,ait​​,bit​​ 是好序列.

输出 tt 的最大值以及对应的序列 i_1, i_2, \cdots, i_ti1​,i2​,⋯,it​.

输入输出格式

输入格式

第一行包含一个整数 nn, 即数对的个数.

后面 nn 行为数对, 每行包含一个数对 (a_i, b_i)(ai​,bi​).

输出格式

第一行输出 tt, 即所求序列长度.

接着输出满足题意的序列 i_1, i_2, \cdots, i_ti1​,i2​,⋯,it​.

说明

数据范围

2 \leq n \leq 3 \cdot 10^52≤n≤3⋅105
1 \leq a_i, b_i \leq 2 \cdot n1≤ai​,bi​≤2⋅n
并且所有 a_i, b_iai​,bi​ 两两不相等.

样例解释

样例 1: 1 < 7 > 3 < 5 > 2 < 101<7>3<5>2<10.

样例 2: 6 > 1 < 3 > 2 < 5 > 46>1<3>2<5>4.

输入输出样例

输入 #1复制

5
1 7
6 4
2 10
9 8
3 5

输出 #1复制

3
1 5 3

输入 #2复制

3
5 4
3 2
6 1

输出 #2复制

3
3 2 1

说明/提示

The final sequence in the first example is 1 < 7 > 3 < 5 > 2 < 101<7>3<5>2<10 .

The final sequence in the second example is 6 > 1 < 3 > 2 < 5 > 46>1<3>2<5>4 .

---

题目想叉了模拟了半天心态炸裂quq

思路：最重要的是发现ai<bi和ai>bi这两种序列是不会同时出现在一个good里的。那么答案要不是第一种good，要不是第二种good，且在同一个good中，比如是ai<bi这种good，按照ai>aj去排序，这样必然能构造出整条都是good。[ai<bi>aj<bj]

然后看哪种good的数量多就好了

#include<iostream>
#include<vector>
#include<queue>
#include<cstring>
#include<cmath>
#include<map>
#include<set>
#include<cstdio>
#include<algorithm>
#define debug(a) cout<<#a<<"="<<a<<endl;
using namespace std;
const int maxn=3e5+100;
typedef long long LL;
LL n,cnt1,cnt2;
struct node{
	LL a,b,id;
}A[maxn],B[maxn];
bool cmpa(node A,node B){return A.a>B.a;}
bool cmpb(node A,node B){return A.b<B.b;}
int main(void)
{
  cin.tie(0);std::ios::sync_with_stdio(false);
  cin>>n;
  for(LL i=1;i<=n;i++){
  	LL u,v,id;cin>>u>>v;id=i;
  	if(u<v) A[++cnt1]={u,v,id};
  	else B[++cnt2]={u,v,id};
  }
  if(cnt1>cnt2){
  	cout<<cnt1<<endl;
  	sort(A+1,A+1+cnt1,cmpa);
  	for(LL i=1;i<=cnt1;i++){
  		cout<<A[i].id<<" ";	
	}
	cout<<endl;
  }
  else{
  	cout<<cnt2<<endl;
  	sort(B+1,B+1+cnt2,cmpb);
  	for(LL i=1;i<=cnt2;i++){
  		cout<<B[i].id<<" ";
	}
	cout<<endl;
  }
return 0;
}