Graph Reconstruction

Time Limit: 2 Seconds      Memory Limit: 65536 KB      Special Judge

Let there be a simple graph with N vertices but we just know the degree of each vertex. Is it possible to reconstruct the graph only by these information?

A simple graph is an undirected graph that has no loops (edges connected at both ends to the same vertex) and no more than one edge between any two different vertices. The degree of a vertex is the number of edges that connect to it.

Input

There are multiple cases. Each case contains two lines. The first line contains one integer N (2 ≤ N ≤ 100), the number of vertices in the graph. The second line conrains N integers in which the ith item is the degree of ith vertex and each degree is between 0 and N-1(inclusive).

Output

If the graph can be uniquely determined by the vertex degree information, output "UNIQUE" in the first line. Then output the graph.

If there are two or more different graphs can induce the same degree for all vertices, output "MULTIPLE" in the first line. Then output two different graphs in the following lines to proof.

If the vertex degree sequence cannot deduced any graph, just output "IMPOSSIBLE".

The output format of graph is as follows:

N E
u

1

 u

2

 ... u

E

v

1

 v

2

 ... v

EWhere N is the number of vertices and E is the number of edges, and {ui,vi} is the ith edge the the graph. The order of edges and the order of vertices in the edge representation is not important since we would use special judge to verify your answer. The number of each vertex is labeled from 1 to N. See sample output for more detail.

Sample Input

1
0
6
5 5 5 4 4 3
6
5 4 4 4 4 3
6
3 4 3 1 2 0

Sample Output

UNIQUE
1 0 UNIQUE
6 13
3 3 3 3 3 2 2 2 2 1 1 1 5
2 1 5 4 6 1 5 4 6 5 4 6 4
MULTIPLE
6 12
1 1 1 1 1 5 5 5 6 6 2 2
5 4 3 2 6 4 3 2 4 3 4 3
6 12
1 1 1 1 1 5 5 5 6 6 3 3
5 4 3 2 6 4 3 2 4 2 4 2
IMPOSSIBLE

havel定理的简介:

给定一个非负整数序列{d1,d2,...dn},若存在一个无向图使得图中各点的度与此序列一一对应,则称此序列可图化。进一步,若图为简单图,则称此序列可简单图化。

可图化的判定比较简单:d1+d2+...dn=0(mod2)。关于具体图的构造,我们可以简单地把奇数度的点配对,剩下的全部搞成自环。

可简单图化的判定,有一个Havel定理,是说: 我们把序列排成不增序,即d1>=d2>=...>=dn,则d可简单图化当且仅当d'=(d2-1, d3-1, ... d(d1+1)-1, d(d1+2), d(d1+3), ... dn)可简单图化。这个定理写起来麻烦,实际上就是说,我们把d排序以后,找出度最大的点(设度为d1),把它和度次大的d1个点之间连边,然后这个点就可以不管了,一直继续这个过程,直到建出完整的图,或出现负度等明显不合理的情况。

定理的简单证明如下:

(<=)若d'可简单图化,我们只需把原图中的最大度点和d'中度最大的d1个点连边即可,易得此图必为简单图。

(=>)若d可简单图化,设得到的简单图为G。分两种情况考虑:

(a)若G中存在边(V1,V2), (V1,V3), ...(V1,V(d1+1)),则把这些边除去得简单图G',于是d'可简单图化为G'

(b)若存在点Vi,Vj使得i<j, (V1,Vi)不在G中,但(V1,Vj)在G中。这时,因为di>=dj,必存在k使得(Vi, Vk)在G中但(Vj,Vk)不在G中。这时我们可以令GG=G-{(Vi,Vk),(V1,Vj)}+{(Vk,Vj),(V1,Vi)}。GG的度序列仍为d,我们又回到了情况(a)。

判定过程:
(1)对当前数列排序,使其呈非递增序列
(2)从第二个数开始对其后d[1]个数字减1,d[1]代表排序后第1个数的值
(3)然后删除第一个之后对剩下的数继续排序
(3)一直循环直到当前序列出现负数(即不是可图的情况)或者当前序列全为0 (可图)时退出。

这题不知道为什么一直wa。。

附一个别人的AC代码,和自己的代码

code(self):

#include<iostream>
#include<cstdio>
#include<cstring>
#include<vector>
#include<queue>
#include<algorithm>
#include<map>
#define INF 0x3f3f3f3f
#define M(a,b) memset(a,b,sizeof(a)) using namespace std; int N;
int line1[1006];
int line2[1006]; struct nod
{
int du;
int du1;
int num;
bool operator < (const nod & rhs) const
{
return du>rhs.du;
}
}point[1006]; bool cmp(nod a,nod b){return a.du1>b.du1;} int main()
{
int all = 0;
while(scanf("%d",&N)==1)
{
all = 0;
int ft = 0;
for(int i = 1;i<=N;i++)
{
scanf("%d",&point[i].du);
point[i].du1 = point[i].du;
point[i].num = i;
all+=point[i].du;
if(point[i].du<0) ft = 1;
}
if(ft == 1||all%2==1||point[1].du>N-1) {puts("IMPOSSIBLE");continue;}
int flag = 1;
int ct = 0;
for(int i = 1;i<=N;i++)
{
sort(point+1,point+N+1);
//for(int j = 1;j<=N;j++) cout<<point[j].du<<' ';
//cout<<endl;
if(point[1].du>N-i) {puts("IMPOSSIBLE");flag = 0;break;} if(point[1].du==0&&flag == 2)
{
puts("MULTIPLE");
printf("%d %d\n",N,all/2);
for(int j = 0;j<ct-1;j++)
printf("%d ",line1[j]);
printf("%d\n",line1[ct-1]);
for(int j = 0;j<ct-1;j++)
printf("%d ",line2[j]);
printf("%d\n",line2[ct-1]);
break;
} if(point[1].du==0)
{
puts("UNIQUE"); printf("%d %d\n",N,all/2);
if(all == 0) {puts("");puts("");break;}
for(int j = 0;j<ct-1;j++)
printf("%d ",line1[j]);
printf("%d\n",line1[ct-1]);
for(int j = 0;j<ct-1;j++)
printf("%d ",line2[j]);
printf("%d\n",line2[ct-1]);
break;
} if(point[point[1].du+2].du==point[1+point[1].du].du) {flag = 2;} for(int j = 2;j<=point[1].du+1;j++)
{
point[j].du--;
line1[ct] = point[1].num;
line2[ct] = point[j].num;
ct++;
if(point[j].du<0) {puts("IMPOSSIBLE");flag = 0; break;}
} point[1].du = 0;
if(flag == 0) break;
} if(flag==2)
{
ct = 0;
int fg = 0;
for(int i = 1;i<=N;i++)
{
sort(point+1,point+N+1,cmp);
//for(int j = 1;j<=N;j++) cout<<point[j].du1<<' ';
//cout<<endl;
if(point[1].du1==0)
{
printf("%d %d\n",N,all/2);
for(int j = 0;j<ct-1;j++)
printf("%d ",line1[j]);
printf("%d\n",line1[ct-1]);
for(int j = 0;j<ct-1;j++)
printf("%d ",line2[j]);
printf("%d\n",line2[ct-1]);
break;
} for(int j = 2;j<=point[1].du1+1;j++)
{
point[j].du1--;
line1[ct] = point[1].num;
line2[ct] = point[j].num;
ct++;
} if(point[point[1].du1+2].du1==point[1+point[1].du1].du1+1)
{
point[point[1].du1+2].du1--;
//line1[ct-1] = point[i].num;
line2[ct-1] = point[1+point[1].du1+1].num;
point[1+point[1].du1].du1++;
}
point[1].du1 = 0;
}
}
}
return 0;
}

 code(others):

#include <cstdio>
#include <cstring>
#include <algorithm>
using namespace std;
const int maxn=111;
int sum,n;
typedef pair<int ,int > P;
P deg[maxn];
P tmpdeg[maxn];
int ash1[maxn*maxn*2],ash2[maxn*maxn*2],alen; bool cmp(P A,P B)
{
if(A.first==B.first) return A.second<B.second;
return A.first>B.first;
}
void prints(){
printf("%d %d\n",n,sum>>1); if(alen>0)printf("%d",ash1[0]);
for(int i=1;i<alen;i++){
printf(" %d",ash1[i]);
}
puts("");
if(alen>0)printf("%d",ash2[0]);
for(int i=1;i<alen;i++){
printf(" %d",ash2[i]);
}
puts("");
} void cpy(){
for(int i=0;i<n;i++)tmpdeg[i]=deg[i];
sort(tmpdeg,tmpdeg+n,cmp);
} bool single;
bool rebuild(){
cpy();
alen=0;
single=true;
for(;tmpdeg[0].first>0;){
int amount=0;
for(int j=1;j<n;j++){
if(tmpdeg[j].first>0){
amount++;
}
}
if(amount<tmpdeg[0].first)return false;
if(single&&(tmpdeg[tmpdeg[0].first].first)==(tmpdeg[tmpdeg[0].first+1].first)){
single=false;
}
for(int j=1;j<=tmpdeg[0].first;j++){
tmpdeg[j].first--;
ash1[alen]=tmpdeg[0].second;
ash2[alen++]=tmpdeg[j].second;
} tmpdeg[0].first=0;
sort(tmpdeg,tmpdeg+n,cmp);
}
return true;
} bool rebuild2(){
cpy();
alen=0;
for(;tmpdeg[0].first>0;){
//if(amount<tmpdeg[0].first)return false;
if((tmpdeg[tmpdeg[0].first].first)==(tmpdeg[tmpdeg[0].first+1].first)){
swap(tmpdeg[tmpdeg[0].first].second,tmpdeg[tmpdeg[0].first+1].second);
}
for(int j=1;j<=tmpdeg[0].first;j++){
tmpdeg[j].first--;
ash1[alen]=tmpdeg[0].second;
ash2[alen++]=tmpdeg[j].second;
}
tmpdeg[0].first=0;
sort(tmpdeg,tmpdeg+n,cmp);
}
return true;
} int main(){
while(scanf("%d",&n)==1){
bool failed =false;
for(int i=1;i<=n;i++){
scanf("%d",&(deg[i-1].first));
deg[i-1].second=i;
if(deg[i-1].first<0)failed=true;
}
sort(deg,deg+n,cmp);
sum=0;
for(int i=0;i<n;i++){
if(deg[i].first>=n)failed=true;
sum+=deg[i].first;
}
if(sum&1)failed=true;
if(failed||!rebuild()){
puts("IMPOSSIBLE");
}
else if(single){
puts("UNIQUE");
prints();
}
else {
puts("MULTIPLE");
prints();
rebuild2();
prints();
}
}
return 0;
}

  

2013长沙 G Graph Reconstruction (Havel-Hakimi定理)的更多相关文章

  1. POJ1659 Frogs' Neighborhood(Havel–Hakimi定理)

    题意 题目链接 \(T\)组数据,给出\(n\)个点的度数,问是否可以构造出一个简单图 Sol Havel–Hakimi定理: 给定一串有限多个非负整数组成的序列,是否存在一个简单图使得其度数列恰为这 ...

  2. 2013亚洲区域赛长沙站 ZOJ 3732 Graph Reconstruction

    题目链接 Graph Reconstruction 题意 给你无向图每个点的度数, 问是否存在唯一解, 存在输出唯一解, 多解输出两个, 无解输出IMPOSSIBLE 思路 这里用到了 Havel-H ...

  3. Codeforces Round #192 (Div. 1) C. Graph Reconstruction 随机化

    C. Graph Reconstruction Time Limit: 20 Sec Memory Limit: 256 MB 题目连接 http://codeforces.com/contest/3 ...

  4. 2013长沙网络赛H题Hypersphere (蛋疼的题目 神似邀请赛A题)

    Hypersphere Time Limit: 1 Second       Memory Limit: 32768 KB In the world of k-dimension, there's a ...

  5. zoj3732&& hdu4797 Graph Reconstruction

    Graph Reconstruction Time Limit: 2 Seconds      Memory Limit: 65536 KB      Special Judge Let there ...

  6. ZOJ3732 Graph Reconstruction Havel-Hakimi定理

    分析: 给定一个非负整数序列{dn},若存在一个无向图使得图中各点的度与此序列一一对应,则称此序列可图化. 进一步,若图为简单图,则称此序列可简单图化 (来自百度百科) 可简单图化的判定可以用Have ...

  7. HDU 2454 Degree Sequence of Graph G——可简单图化&&Heavel定理

    题意 给你一个度序列,问能否构成一个简单图. 分析 对于可图化,只要满足度数之和是偶数,即满足握手定理. 对于可简单图化,就是Heavel定理了. Heavel定理:把度序列排成不增序,即 $deg[ ...

  8. 【2013长沙区域赛】部分题解 hdu4791—4801

    1001: 签到题,二分一下即可 代码: #include <set> #include <map> #include <cmath> #include <c ...

  9. The Ninth Hunan Collegiate Programming Contest (2013) Problem G

    Problem G Good Teacher I want to be a good teacher, so at least I need to remember all the student n ...

随机推荐

  1. Guava 学习计划

    Guava这个框架,我想参加过工作了Java程序员都应该不会陌生,这个框架甚至比Apache Commons Lang3框架更受Java程序员的喜欢.但是遗憾的是怎么优秀和成熟的框架,在国内我居然没有 ...

  2. POJ 1064 Cable master (二分)

    题目链接: 传送门 Cable master Time Limit: 1000MS     Memory Limit: 65536K 题目描述 有N条绳子,它们长度分别为Li.如果从它们中切割出K条长 ...

  3. Beta版本——项目测试

    前端测试 一.测试用例(tutor_distribution_0001) 测试内容 获取下拉框的输入测试 测试代码 $("#sub-confirm").click(function ...

  4. jsp分页功能

    http://blog.csdn.net/xiazdong/article/details/6857515

  5. 添加css的方式:link与@import区别

    如何在html中添加css? 在html中设置css共有三种方式,分别是: 行内式 内嵌式 导入式-link 导入式-@import 1.行内式.即在html标签中的style属性中设置css,值得注 ...

  6. python Scrapy安装和介绍

    python Scrapy安装和介绍 Windows7下安装1.执行easy_install Scrapy Centos6.5下安装 1.库文件安装yum install libxslt-devel ...

  7. nginx 学习笔记(9) 配置HTTPS服务器--转载

    HTTPS服务器优化SSL证书链合并HTTP/HTTPS主机基于名字的HTTPS主机带有多个主机名的SSL证书主机名指示兼容性 配置HTTPS主机,必须在server配置块中打开SSL协议,还需要指定 ...

  8. winform 窗体传值

    在Form1中:我们要把一个文本传递给Form2窗体,假定为passText               public string passText               {          ...

  9. c#优化

    优化反射 http://kb.cnblogs.com/page/172119/ http://www.2cto.com/kf/201301/182765.html http://blog.sina.c ...

  10. My latest news (--2016.10)

    2016.10.31 22:44 一个“程序”,打代码占40%.思考占60% 2016.10.30 20:53 周末,话说今天有晚上讲座,还点名,了,悲催.之前学习的Qt有点问题,悲催.推荐个博文:h ...