题目链接:
http://acm.hust.edu.cn/vjudge/problem/356795
Racing Gems
输入
The input file contains several test cases, each of them as described below.
The first line of input contains four space-separated integers n, r, w, and h (1 ≤ n ≤ 105
, 1 ≤ r ≤ 10,
1 ≤ w, h ≤ 109
). Each of the following n lines contains two space-separated integers xi and yi
, denoting
the coordinate of the i-th gem (0 ≤ xi ≤ w, 0 < yi ≤ h). There will be at most one gem per location.
The input does not include a value for v.
输出
For each case, print, on a single line, the maximum number of gems that can be collected during the
race.
样例
sample input
5 1 10 10
8 8
5 1
4 6
4 7
7 9
5 1 100 100
27 75
79 77
40 93
62 41
52 45
10 3 30 30
14 9
2 20
3 23
15 19
13 5
17 24
6 16
21 5
14 10
3 6sample output
3
3
4
题意
以辆赛车可以从x轴上任意点出发,他的水平速度允许他向每向上移动1个单位,就能向左或向右移动1/r个单位(也就是它的辐射范围是个等腰三角形)
现在赛车从x轴出发,问它在到达终点前能吃到的最多钻石。
题解
考虑每个钻石的向下辐射范围,并且将其投影到x轴上的两个点,(辐射范围与x轴的两个焦点),然后我们就把题目转化成了一个区间覆盖问题,我们用x<y表示x区间被y区间覆盖,即求二维最长的上升子序列:a1<=a2<=...<=an。
我们按每个钻石的左端点排序,然后跑右端点的最长不下降子序列就可以了。由于数据比较大,用二分处理成nlogn的复杂度。
代码
#include<map>
#include<queue>
#include<vector>
#include<cstdio>
#include<cstring>
#include<string>
#define X first
#define Y second
#include<iostream>
#include<algorithm>
#define mkp make_pair
#define lson (o<<1)
#define rson ((o<<1)|1)
#define M (l+(r-l)/2)
#define bug(a) cout<<#a<<" = "<<a<<endl
using namespace std;
typedef __int64 LL;
const int maxn=1e5+10;
struct Node{
LL a,b;
bool operator <(const Node& tmp) const {
return a>tmp.a||a==tmp.a&&b<tmp.b;
}
}nds[maxn];
int n,w,h;
LL r;
LL ra[maxn];
int dp[maxn];
int main() {
memset(dp,0,sizeof(dp));
scanf("%d%I64d%d%d",&n,&r,&w,&h);
for(int i=1;i<=n;i++){
int x,y;
scanf("%d%d",&x,&y);
//映射到x轴的左端点
nds[i].a=r*x-y;
//映射到x轴的右端点
nds[i].b=r*x+y;
}
//按左端点降序排,左端点相同时右端点升序排
sort(nds+1,nds+n+1);
//初始的ra数组为空
int tot=1,ans=1;
for(int i=1;i<=n;i++){
//二分查找
int pos=upper_bound(ra+1,ra+tot,nds[i].b)-ra;
dp[i]=pos;
ra[pos]=nds[i].b;
if(tot==pos) tot++;
ans=max(ans,dp[i]);
}
printf("%d
",ans);
return 0;
}