HDU 5521 Meeting(虚拟节点+最短路)
Meeting
Time Limit: 12000/6000 MS (Java/Others) Memory Limit: 262144/262144 K (Java/Others)
Total Submission(s): 1358 Accepted Submission(s): 435
fences they were separated into different blocks. John's farm are divided into n blocks
labelled from 1 to n.
Bessie lives in the first block while Elsie lives in the n-th
one. They have a map of the farm
which shows that it takes they ti minutes
to travel from a block in Ei to
another block
in Ei where Ei (1≤i≤m) is
a set of blocks. They want to know how soon they can meet each other
and which block should be chosen to have the meeting.
the number of test cases. Then T test
cases
follow.
The first line of input contains n and m. 2≤n≤105.
The following m lines
describe the sets Ei (1≤i≤m).
Each line will contain two integers ti(1≤ti≤109)and Si (Si>0) firstly.
Then Si integer
follows which are the labels of blocks in Ei.
It is guaranteed that ∑mi=1Si≤106.
Otherwise, output two lines. The first line contains an integer, the time it takes for they to meet.
The second line contains the numbers of blocks where they meet. If there are multiple
optional blocks, output all of them in ascending order.
In the first case, it will take Bessie 1 minute travelling to the 3rd block, and it will take Elsie 3 minutes travelling to the 3rd block. It will take Bessie 3 minutes travelling to the 4th block, and it will take Elsie 3 minutes travelling to the 4th block. In the second case, it is impossible for them to meet.
题目链接:HDU 5521
题意:给定点数n和集合个数m,然后给你m个集合,每一个集合有si个点,两两之间的到达时间都是ti,一个人在1,一个人在n,求两人同时出发的相遇的最短时间
由于每一个集合的点有很多,若集合两两之间连边,边数非常大,一开始这样就超时了……然后正确做法是对每一个集合再虚拟一个节点(范围是[n+1,n+m]),给每一个集合内的点连边权为ti的双向边到本集合对应虚拟节点,集合内其他点不连边,这样就可以通过虚拟的节点来到达其他地方从而减少边数(方法真是太巧妙了),然后求相遇的最短时间,显然现在无法得知到底选哪个地点作为见面地点,那就对1跑一遍单源最短路,对n跑一边单源最短路,然后统计1~n中每一个点的可能最短时间(一个人早到一个人晚到,显然用max取时间长的那个数),然后选出1~n中的最短时间mndx,再遍历一下看哪些点的最短时间为mndx并记录输出,最后记得把最短时间除以2,因为连的边是ti,进入虚拟节点又出来会多算一次
点数为1e5,题目中说了所有集合大小之和不会超过1e6,每一个集合都有2*|Si|条边,那就是2*1e6条边因此N可设为1e5+10,M可设为2*1e6+10。
以上是以前的解法,昨天计蒜客被惨虐之后仔细看了一下D题发现其实跟这道题是同一个原理,这题是群内的点之间两两距离为ti,那不妨把Block点拆成入口和出口,然后这样连边:
<u, Block入口, 0>,<Block出口, u , 0>,<Block入口, Block出口, ti>
然后从S和T各跑一遍SPFA然后记录下Max更新答案即可,也不用像上面的解法一样除以2了,点数最差情况下一个点作为Block,应该是3e5,边数应该为2sum{Si}+m,应该是2e6+1e5
代码:
#include <bits/stdc++.h>
using namespace std;
#define INF 0x3f3f3f3f
#define LC(x) (x<<1)
#define RC(x) ((x<<1)+1)
#define MID(x,y) ((x+y)>>1)
#define CLR(arr,val) memset(arr,val,sizeof(arr))
#define FAST_IO ios::sync_with_stdio(false);cin.tie(0);
typedef pair<int, int> pii;
typedef long long LL;
const double PI = acos(-1.0);
const int N = 3e5 + 7;
const int M = 2e6 + 1e5 + 7;
struct edge
{
int to, nxt;
LL w;
edge() {}
edge(int _to, int _nxt, LL _w): to(_to), nxt(_nxt), w(_w) {}
};
edge E[M];
int head[N], tot;
int vis[N];
LL ds[N], de[N], Mindist[N]; void init()
{
CLR(head, -1);
tot = 0;
}
inline void add(int s, int t, LL d)
{
E[tot] = edge(t, head[s], d);
head[s] = tot++;
}
void spfa(int s, int flag, LL d[])
{
CLR(vis, 0);
if (flag)
CLR(ds, INF), ds[s] = 0;
else
CLR(de, INF), de[s] = 0;
queue<int>Q;
Q.push(s);
while (!Q.empty())
{
int u = Q.front();
Q.pop();
vis[u] = 0;
for (int i = head[u]; ~i; i = E[i].nxt)
{
int v = E[i].to;
if (d[v] > d[u] + E[i].w)
{
d[v] = d[u] + E[i].w;
if (!vis[v])
{
vis[v] = 1;
Q.push(v);
}
}
}
}
}
int main(void)
{
int tcase, n, m, si, u, i;
scanf("%d", &tcase);
for (int q = 1; q <= tcase; ++q)
{
init();
scanf("%d%d", &n, &m);
LL ti;
for (i = 1; i <= m; ++i)
{
scanf("%I64d%d", &ti, &si);
add(i, i + m, ti); //m
while (si--)
{
scanf("%d", &u);
add(u + (m << 1), i, 0LL); //si
add(i + m, u + (m << 1), 0LL); //si
}
}
spfa(1 + (m << 1), 1, ds);
spfa(n + (m << 1), 0, de);
LL ans = 0x3f3f3f3f3f3f3f3f;
printf("Case #%d: ", q);
vector<int>pos;
for (i = 2 * m + 1; i <= 2 * m + n; ++i)
{
Mindist[i] = max<LL>(ds[i], de[i]);
if (Mindist[i] < ans)
ans = Mindist[i];
}
if (ans == 0x3f3f3f3f3f3f3f3f)
puts("Evil John");
else
{
printf("%I64d\n", ans);
for (i = 2 * m + 1; i <= 2 * m + n; ++i)
if (Mindist[i] == ans)
pos.push_back(i - (m << 1));
int sz = pos.size();
for (i = 0; i < sz; ++i)
printf("%d%c", pos[i], " \n"[i == sz - 1]);
}
}
return 0;
}
HDU 5521 Meeting(虚拟节点+最短路)的更多相关文章
- HDU 5521.Meeting 最短路模板题
Meeting Time Limit: 12000/6000 MS (Java/Others) Memory Limit: 262144/262144 K (Java/Others)Total ...
- hdu 5521 Meeting(最短路)
题目链接:http://acm.hdu.edu.cn/showproblem.php?pid=5521 题意:有1-n共n个点,给出m个块(完全图),并知道块内各点之间互相到达花费时间均为ti.已知两 ...
- HDU 5521 Meeting【最短路】
今天旁观了Angry_Newbie的模拟区域赛(2015shenyang) 倒着看最先看的M题,很明显的最短路问题,在我看懂的时候他们已经开始敲B了. 后来听说D过了很多人.. D题一看是个博弈,给了 ...
- HDU 5521 Meeting (最短路,dijstra)
题意:有N个点,两个人,其中一个人住在点1,另一个人住在点n,有M个点集,集合内的数表示任意两点的距离为dis ,现在问,如果两个人要见面, 需要最短距离是多少,有哪几个点能被当成见面点. 析:分别对 ...
- Ural 1741 Communication Fiend(隐式图+虚拟节点最短路)
1741. Communication Fiend Time limit: 1.0 second Memory limit: 64 MB Kolya has returned from a summe ...
- HDU 5521 Meeting
2015 ACM / ICPC 沈阳站现场赛 M题 最短路 设置N+M个节点,前N个节点是Block,后M个节点是Set,每一组Set中的点向该Set连边,从1和n开始分别求最短路.注意爆int. # ...
- HDU - 5521 Meeting (Dijkstra)
思路: 看了好久才看懂题意,文中给了n个点,有m个集合,每个集合有s个点,集合内的每两个点之间有一个权值为t的边,现在有两个人,要从1号点,和n号点,走到同一个顶点,问最少花费以及花费最少的点. 那就 ...
- HDU 5521:Meeting(最短路)
http://acm.hdu.edu.cn/showproblem.php?pid=5521 Meeting Problem Description Bessie and her friend E ...
- hdu 5521 最短路
Meeting Time Limit: 12000/6000 MS (Java/Others) Memory Limit: 262144/262144 K (Java/Others)Total ...
随机推荐
- ORACLE清除某一字段重复的数据(选取重复数据中另一个字段时期最大值)
需求:资产维修表中同一资产可能维修完继续申请维修,这时候维修状态需要根据最近的维修时间去判断维修状态,所以同一资产ID下会出现重复的数据(维修审批通过,维修审批未通过),或者可能不出现(未申请维修), ...
- MySQL 查询最大最小值优化
1. 假设你使用了Innodb存储引擎2. 假设你在innodb设定了主键(聚集索引) 3. 因为聚集索引页面之间是通过双向链表链接,页按照主键的顺序排序 每个页中的记录也是通过双向链表维护.聚集索引 ...
- window常用软件
ftpserver QQ asc pan 屏保 view putty 迅雷 teamviewer绿色 teamviewer单文件 魔方 chrome winscp WinRAR xshell 鲁大师 ...
- fedora 损坏,开机出现GRUB 2 启动提示符
如果设置不当,GRUB 2 可能会加载失败,随后进入启动提示符中.要解决此问题,按如下步骤执行: 1. 列出 GRUB 2 识别的驱动器: grub2> ls 2. 对于一个含有三个分区的 do ...
- Android_adb shell am/pm使用
转自:http://blog.sina.com.cn/s/blog_51335a0001017ux5.html adb shell am instrument [options] <COMP ...
- wp8 入门到精通 LINQ to SQL
http://msdn.microsoft.com/zh-cn/library/bb397924.aspx LINQ 查询操作中的类型关系 (C#) 使用一个人类发明快速检索的方法 // Northw ...
- Ubuntu小点汇总,更新中...
转自:http://blog.csdn.net/zxz_tsgx/article/details/39713627 昨天重装了Ubuntu14.04 64位版,又被一些基础操作/设置给搞怕了,以前安装 ...
- 一个快速、高效的Levenshtein算法实现
转自:http://www.cnblogs.com/ymind/archive/2012/03/27/fast-memory-efficient-Levenshtein-algorithm.html ...
- 从维度理解dp问题
对于dp,我目前的理解就是,干成题目中的那件事需要作出若干次决策,然后你要取其中最优的结果,我们可以用深搜来递归地找最优解,然后我们来观察一下这个递归树的形状,如果它能从底往上直接递推的话,就不用递归 ...
- jq生成目录文件树jQuery Ztree基本用法
转自:http://www.cnblogs.com/linjiqin/p/4547452.html 1.首先在页面上有<ul/>标签 ? 1 <ul id="tree&qu ...