Description

Misha trains several ACM teams at the university. He is an experienced coach, and he does not underestimate the meaning of friendly and collaborative atmosphere during training sessions. It used to be that way, but one of the teams happened to win contests a little bit more often than others, and hence became slightly too big for their boots. That does not contribute to positive spirit which is essential for successful training. But Misha knows what to do!
Representatives of k teams participate in Misha’s training sessions, each team has three members. Alas, teams rarely attend en masse, but at least one member per team is always present, of course. During the next training session Misha is going to split everyone into npairs, so that each pair will include representatives of different teams. Players will play a mini-contest against each other in each pair.
A situation when no two mini-contests are won by representatives of one team is the one that suits Misha’s goals best. He may be somewhat cunning when selecting winner in each pair in order to achieve such situation. Find out whether Misha will succeed.

Input

The first line contains two numbers — n and k (1 ≤ n ≤ 10 5, 2 ≤ k ≤ 10 5). n lines follow. i-th of these contains two numbers x i, y i (1 ≤ x i, y i ≤ k; x i ≠ y i) — the numbers of teams, whose representatives are in pair number i.
It is guaranteed that each team is represented in no less than one and no more than three pairs.

Output

If Misha is to succeed, output Yes in the first line. In each of the following n lines output one number — the number of team that is to win in the corresponding pair. If there are several answers, output any.
If Misha is to fail, output No in the only line.

Sample Input

input output
3 4
1 2
2 3
1 4
Yes
2
3
4
6 4
1 2
1 3
1 4
2 3
2 4
3 4
No

题意:给出n对点a,b  要求从没对点中选出一个,且最终选出的点n个数不能存在相同的。输入数据满足每种数最多出现3次,最少出现1次

思路:第i对点的编号2*i, 2*i+1,   因为每个数最多出现3次,那么完全可以枚举每个数,然后相同的数之间的编号连一条边,表示这两个编号不能同时选,这样跑完twosat就能得到一个满足情况的解或无解。

#include <cstdio>
#include <cstring>
#include <iostream>
#include <algorithm>
#include <queue>
#include <stack>
#include <cstdlib>
#include <map>
#include <set>
#include <cmath>
using namespace std;
const int N = 2e6 + ;
struct Edge {
int to, nex;
}e[N];
int head[N], tot;
void init() {
tot = ; memset(head, -, sizeof head);
}
void add(int u, int v) {
e[tot].to = v;
e[tot].nex = head[u];
head[u] = tot++;
}
int Low[N], DFN[N], Stack[N], Belong[N];
int Index, top;
int scc;
bool Instack[N];
int num[N]; void Tarjan(int u) {
int v;
Low[u] = DFN[u] = ++Index;
Stack[top++] = u;
Instack[u] = true; for(int i = head[u]; ~i; i = e[i].nex) {
v = e[i].to;
if(!DFN[v]) {
Tarjan(v);
if(Low[u] > Low[v]) Low[u] = Low[v];
}else if(Instack[v] && Low[u] > DFN[v]) Low[u] = DFN[v];
}
if(Low[u] == DFN[u]) {
scc++;
do
{
v = Stack[--top];
Instack[v] = false;
Belong[v] = scc;
num[scc]++;
}while(v != u);
}
}
bool solvable(int n) {
memset(DFN, , sizeof DFN);
memset(Instack, false, sizeof Instack);
memset(num, , sizeof num);
Index = scc = top = ;
for(int i = ; i < n; ++i) if(!DFN[i]) Tarjan(i); for(int i = ; i < n; i += ) {
if(Belong[i] == Belong[i ^ ]) return false;
}
return true;
} queue<int> q1, q2;
vector<vector<int> >dag;
int vis[N];
int indeg[N];
int cf[N];
void solve(int n) {
dag.assign(scc+, vector<int>());
memset(indeg, , sizeof indeg);
memset(vis, , sizeof vis);
for(int u = ; u < n; ++u) {
for(int i = head[u]; ~i; i = e[i].nex) {
int v = e[i].to;
if(Belong[u] != Belong[v]) {
dag[ Belong[v] ].push_back(Belong[u]);
indeg[ Belong[u] ]++;
}
}
}
for(int i = ; i < n; i += ) {
cf[ Belong[i] ] = Belong[i ^ ];
cf[ Belong[i ^ ] ] = Belong[i];
}
while(!q1.empty()) q1.pop();
while(!q2.empty()) q2.pop();
for(int i = ; i <= scc; ++i) if(indeg[i] == ) q1.push(i); while(!q1.empty()) {
int u = q1.front();
q1.pop();
if(vis[u] == ) {
vis[u] = ;
vis[ cf[u] ] = ;
}
int sz = dag[u].size();
for(int i = ; i < sz; ++i) {
indeg[ dag[u][i] ]--;
if(indeg[ dag[u][i] ] == ) q1.push(dag[u][i]);
}
}
}
int r[N];
vector<int> g[N];
int main() {
#ifdef LOCAL
freopen("in", "r", stdin);
#endif
int n, m;
while(~scanf("%d%d", &m, &n)) {
init();
int tot = ; int u, v;
for(int i = ; i < m; ++i) {
scanf("%d%d", &u, &v);
r[tot++] = u;
r[tot++] = v;
g[u].push_back( * i);
g[v].push_back( * i + );
} for(int i = ; i <= n; ++i) {
int sx = g[i].size();
for(int j1 = ; j1 < sx; ++j1) {
for(int j2 = j1 + ; j2 < sx; ++j2) {
int v1 = g[i][j1];
int v2 = g[i][j2];
add(v1, v2 ^ );
add(v2, v1 ^ );
}
}
}
if(solvable( * m)) {
solve( * m);
puts("Yes");
for(int i = ; i < m; ++i) {
if(vis[ Belong[ * i] ]) printf("%d\n", r[ * i + ]);
else printf("%d\n", r[ * i]);
}
}else puts("No");
}
return ;
}

URAL 2089 Experienced coach Twosat的更多相关文章

  1. 【2-SAT】URAL - 2089 - Experienced coach

    题意:给出n对点a,b  要求从没对点中选出一个,且最终选出的点n个数不能存在相同的.输入数据满足每种数最多出现3次,最少出现1次 思路:第i对点的编号2*i, 2*i+1,   因为每个数最多出现3 ...

  2. Ural2089:Experienced coach(二分图匹配)

    Misha trains several ACM teams at the university. He is an experienced coach, and he does not undere ...

  3. URAL 2078~2089

    URAL 2078~2089 A - Bowling game 题目描述:给出保龄球每一局击倒的球数,按照保龄球的规则,算出总得分的最小值和最大值. solution 首先是最小值:每一局第一球击倒\ ...

  4. URAL 1873. GOV Chronicles

    唔 神题一道 大家感受一下 1873. GOV Chronicles Time limit: 0.5 secondMemory limit: 64 MB A chilly autumn night. ...

  5. 【HDU 2089】不要62

    http://acm.hdu.edu.cn/showproblem.php?pid=2089 数位dp,参照了打野的博客 预处理出f数组,f[i][j]表示第i位为数字j时的可行的数字总数. 对于区间 ...

  6. 后缀数组 POJ 3974 Palindrome && URAL 1297 Palindrome

    题目链接 题意:求给定的字符串的最长回文子串 分析:做法是构造一个新的字符串是原字符串+反转后的原字符串(这样方便求两边回文的后缀的最长前缀),即newS = S + '$' + revS,枚举回文串 ...

  7. ural 2071. Juice Cocktails

    2071. Juice Cocktails Time limit: 1.0 secondMemory limit: 64 MB Once n Denchiks come to the bar and ...

  8. ural 2073. Log Files

    2073. Log Files Time limit: 1.0 secondMemory limit: 64 MB Nikolay has decided to become the best pro ...

  9. ural 2070. Interesting Numbers

    2070. Interesting Numbers Time limit: 2.0 secondMemory limit: 64 MB Nikolay and Asya investigate int ...

随机推荐

  1. xml解析技术

    本文总结Dom,sax解析,  使用Java作为工具解析xml文档. 1 Dom 综述:Dom解析xml通常也称为xmlDom (和htmlDom技术差不多),将xml文档封装成树,好处就是xml中的 ...

  2. [Data Structure] LCSs——最长公共子序列和最长公共子串

    1. 什么是 LCSs? 什么是 LCSs? 好多博友看到这几个字母可能比较困惑,因为这是我自己对两个常见问题的统称,它们分别为最长公共子序列问题(Longest-Common-Subsequence ...

  3. SQLServer日期函数用法

    --1.显示本月第一天 ,) ),)) --2.显示本月最后一天 ,),,))) ,,,)) --3.上个月的最后一天 ,,)) --4.本月的第一个星期一 , ) --5.本年的第一天 ,) --6 ...

  4. nginx tomcat session丢失的问题

    nginx反向代理tomcat,出现seesion获取不到的问题. 网上搜索到的解决方案大多是集群tomcat共享session共享的问题,但我这个只有一台tomcat服务器,不涉及到服务器集群问题. ...

  5. 使用stack的思考

    对于使用stack进行()的配对检查,不需要使用额外的空间对每个字符进行存储和push与pop的操作. 只使用size对()进行处理即可,因为只有一种括号,所以入栈为size加1,出栈为size-1. ...

  6. Interleaving String

    https://leetcode.com/problems/interleaving-string/ Given s1, s2, s3, find whether s3 is formed by th ...

  7. ng-repeat 嵌套访问父作用域里的属性

    在一个项目中,需要嵌套循环输出一个二维表的里的数据 数据结构 [ { id:1, list:[ { id:1, name:'li' } ] }, { id:2, list:[ { id:1, name ...

  8. oracle与mysql创建表时的区别

    oracle创建表时,不支持在建表时同时增加字段注释.故采用以下方式: #创建表CREATE TABLE predict_data as ( id integer ), mid ), time dat ...

  9. 前端JS常用字符串处理实例

    字符串处理常常用在处理服务器回传的数据.动态拼接生成html等,是前端面试的必考题. 我觉得字符串处理这种常用到的,一定要了然于心,不然用到时急急忙忙去翻手册费半天. 入正题,首先提出平常遇到的几个需 ...

  10. Java字节流和字符流区别

    1.字节流:直接操作文件本身. 2.字符流:通过缓冲区来操作文件. 所有的文件在硬盘或在传输时都是以字节的方式进行的,包括图片等都是按字节的方式存储的,而字符是只有在内存中才会形成,所以在开发中,字节 ...