Taxi Cab Scheme
时间限制: 1 Sec 内存限制: 64 MB
题目描述
Running a taxi station is not all that simple. Apart from the obvious demand for a centralised coordination of the cabs in order to pick up the customers calling to get a cab as soon as possible,there is also a need to schedule all the taxi rides which have been booked in advance.Given a list of all booked taxi rides for the next day, you want to minimise the number of cabs needed to carry out all of the rides. For the sake of simplicity, we model a city as a rectangular grid. An address in the city is denoted by two integers: the street and avenue number. The time needed to get from the address a, b to c, d by taxi is |a - c| + |b - d| minutes. A cab may carry out a booked ride if it is its first ride of the day, or if it can get to the source address of the new ride from its latest,at least one minute before the new ride's scheduled departure. Note that some rides may end after midnight.
给你N个出租车的预定单表,有初始时间,起点和终点。问最少用多少辆出租车可以满足这N个预订单。
输入
On the first line of the input is a single positive integer N, telling the number of test scenarios to follow. Each scenario begins with a line containing an integer M, 0 < M < 500, being the number of booked taxi rides. The following M lines contain the rides. Each ride is described by a departure time on the format hh:mm (ranging from 00:00 to 23:59), two integers a b that are the coordinates of the source address and two integers c d that are the coordinates of the destination address. All coordinates are at least 0 and strictly smaller than 200. The booked rides in each scenario are sorted in order of increasing departure time.
输出
For each scenario, output one line containing the minimum number of cabs required to carry out all the booked taxi rides.
样例输入
2
08:00 10 11 9 16
08:07 9 16 10 11
2
08:00 10 11 9 16
08:06 9 16 10 11
样例输出
2
提示
同样的转化为图G=(V,E),则问题转化为:
在图G中选取尽可能少的点,使得图中每一条边至少有一个端点被选中。
这个问题在二分图问题中被称为最小点覆盖问题。即用最少的点去覆盖所有的边。
结论:由König定理可知最小点覆盖的点数 = 二分图最大匹配
匈牙利算法需要我们从右边的某个没有匹配的点,走出一条使得“一条没被匹配、一条已经匹配过,再下一条又没匹配这样交替地出现”的路(交错轨,增广路)。但是,现在我们已经找到了最大匹配,已经不存在这样的路了。换句话说,我们能寻找到很多可能的增广路,但最后都以找不到“终点是还没有匹配过的点”而失败。我们给所有这样的点打上记号:从右边的所有没有匹配过的点出发,按照增广路的“交替出现”的要求可以走到的所有点(最后走出的路径是很多条不完整的增广路)。那么这些点组成了最小覆盖点集:右边所有没有打上记号的点,加上左边已经有记号的点。看图,右图中展示了两条这样的路径,标记了一共6个点(用 “√”表示)。那么,用红色圈起来的三个点就是我们的最小覆盖点集。
首先,为什么这样得到的点集点的个数恰好有M个呢?答案很简单,因为每个点都是某个匹配边的其中一个端点。如果右边的哪个点是没有匹配过的,那么它早就当成起点被标记了;如果左边的哪个点是没有匹配过的,那就走不到它那里去(否则就找到了一条完整的增广路)。而一个匹配边又不可能左端点是标记了的,同时右端点是没标记的(不然的话右边的点就可以经过这条边到达了)。因此,最后我们圈起来的点与匹配边一一对应。
其次,为什么这样得到的点集可以覆盖所有的边呢?答案同样简单。不可能存在某一条边,它的左端点是没有标记的,而右端点是有标记的。原因如下:如果这条边不属于我们的匹配边,那么左端点就可以通过这条边到达(从而得到标记);如果这条边属于我们的匹配边,那么右端点不可能是一条路径的起点,于是它的标记只能是从这条边的左端点过来的(想想匹配的定义),左端点就应该有标记。
最后,为什么这是最小的点覆盖集呢?这当然是最小的,不可能有比M还小的点覆盖集了,因为要覆盖这M条匹配边至少就需要M个点(再次回到匹配的定义)。
证完了。
这道题乍一看不知道怎么做,但是仔细想一想,就可以发现如果一辆车可以在另外一个乘客出发之前赶到出发点,就可以将这两个乘客之间连一条单向边,从而可以用二分图来解决。
#include<iostream>
#include<cstdio>
#include<cstring>
#include<cstdlib>
#include<algorithm>
#include<cmath>
#include<queue>
#include<stack>
#include<ctime>
using namespace std;
int n,t;
struct node
{
int time,time2;
int x1,y1,x2,y2;
}a[];
int f[][],match[],vis[],dfscnt;
int suan(int x1,int y1,int x2,int y2)
{
return abs(x2-x1)+abs(y2-y1);
}
bool dfs(int root)
{
int i;
for(i=;i<=n;i++)
{
if(f[root][i])
{
if(vis[i]!=dfscnt)
{
vis[i]=dfscnt;
if(!match[i]||dfs(match[i]))
{
match[i]=root;
return ;
}
}
}
}
return ;
}
int main()
{
int i,j;
scanf("%d",&t);
while(t--)
{
memset(f,,sizeof(f));
memset(match,,sizeof(match));
memset(vis,,sizeof(vis));
dfscnt=;
scanf("%d",&n);
for(i=;i<=n;i++)
{
int hour,minute;
scanf("%d:%d",&hour,&minute);
a[i].time=hour*+minute;
scanf("%d%d%d%d",&a[i].x1,&a[i].y1,&a[i].x2,&a[i].y2);
a[i].time2=a[i].time+suan(a[i].x1,a[i].y1,a[i].x2,a[i].y2);
}
for(i=;i<=n;i++)
{
for(j=i;j<=n;j++)
{
if(a[i].time2+suan(a[i].x2,a[i].y2,a[j].x1,a[j].y1)<a[j].time)
f[i][j]=;
}
}
int ans=;
for(i=;i<=n;i++)
{
dfscnt++;
if(dfs(i))ans++;
}
printf("%d\n",n-ans);
}
}