time limit per test2 seconds

memory limit per test256 megabytes

inputstandard input

outputstandard output

Local authorities have heard a lot about combinatorial abilities of Ostap Bender so they decided to ask his help in the question of urbanization. There are n people who plan to move to the cities. The wealth of the i of them is equal to ai. Authorities plan to build two cities, first for n1 people and second for n2 people. Of course, each of n candidates can settle in only one of the cities. Thus, first some subset of candidates of size n1 settle in the first city and then some subset of size n2 is chosen among the remaining candidates and the move to the second city. All other candidates receive an official refuse and go back home.

To make the statistic of local region look better in the eyes of their bosses, local authorities decided to pick subsets of candidates in such a way that the sum of arithmetic mean of wealth of people in each of the cities is as large as possible. Arithmetic mean of wealth in one city is the sum of wealth ai among all its residents divided by the number of them (n1 or n2 depending on the city). The division should be done in real numbers without any rounding.

Please, help authorities find the optimal way to pick residents for two cities.

Input

The first line of the input contains three integers n, n1 and n2 (1 ≤ n, n1, n2 ≤ 100 000, n1 + n2 ≤ n) — the number of candidates who want to move to the cities, the planned number of residents of the first city and the planned number of residents of the second city.

The second line contains n integers a1, a2, …, an (1 ≤ ai ≤ 100 000), the i-th of them is equal to the wealth of the i-th candidate.

Output

Print one real value — the maximum possible sum of arithmetic means of wealth of cities’ residents. You answer will be considered correct if its absolute or relative error does not exceed 10 - 6.

Namely: let’s assume that your answer is a, and the answer of the jury is b. The checker program will consider your answer correct, if .

Examples

input

2 1 1

1 5

output

6.00000000

input

4 2 1

1 4 2 3

output

6.50000000

Note

In the first sample, one of the optimal solutions is to move candidate 1 to the first city and candidate 2 to the second.

In the second sample, the optimal solution is to pick candidates 3 and 4 for the first city, and candidate 2 for the second one. Thus we obtain (a3 + a4) / 2 + a2 = (3 + 2) / 2 + 4 = 6.5

【题目链接】:http://codeforces.com/contest/735/problem/B

【题解】



贪心;

把分母小的放在那个分子大的财富和的上面;

剩下的分配在分母大的那个财富和上面;

即把ai排序下;

然后从大到小分配两个部分出来;

a[n-min(n1,n2)+1..n]分配给min(n1,n2);

剩下的从大到小选max(n1,n2)个给max(n1,n2);

我也不知道;反正就是贪呗.



【完整代码】

#include <cstdio>
#include <cstdlib>
#include <cmath>
#include <set>
#include <map>
#include <iostream>
#include <algorithm>
#include <cstring>
#include <queue>
#include <vector>
#include <stack>
#include <string>
using namespace std;
#define lson l,m,rt<<1
#define rson m+1,r,rt<<1|1
#define LL long long
#define rep1(i,a,b) for (int i = a;i <= b;i++)
#define rep2(i,a,b) for (int i = a;i >= b;i--)
#define mp make_pair
#define pb push_back
#define fi first
#define se second typedef pair<int,int> pii;
typedef pair<LL,LL> pll; void rel(LL &r)
{
r = 0;
char t = getchar();
while (!isdigit(t) && t!='-') t = getchar();
LL sign = 1;
if (t == '-')sign = -1;
while (!isdigit(t)) t = getchar();
while (isdigit(t)) r = r * 10 + t - '0', t = getchar();
r = r*sign;
} void rei(int &r)
{
r = 0;
char t = getchar();
while (!isdigit(t)&&t!='-') t = getchar();
int sign = 1;
if (t == '-')sign = -1;
while (!isdigit(t)) t = getchar();
while (isdigit(t)) r = r * 10 + t - '0', t = getchar();
r = r*sign;
} const int MAXN = 1e5+100;
const int dx[5] = {0,1,-1,0,0};
const int dy[5] = {0,0,0,-1,1};
const double pi = acos(-1.0); int n,n1,n2;
int a[MAXN]; int main()
{
//freopen("F:\\rush.txt","r",stdin);
rei(n);rei(n1);rei(n2);
rep1(i,1,n)
rei(a[i]);
sort(a+1,a+1+n);
double ans = 0;
int t = min(n1,n2);
rep1(i,n-t+1,n)
ans+=a[i];
ans=ans/(t*1.0);
double ans2= 0;
int t1 = max(n1,n2);
rep1(i,n-t-t1+1,n-t)
ans2+=a[i];
ans2 = ans2/(t1*1.0);
ans+=ans2;
printf("%.6lf\n",ans);
return 0;
}

【50.88%】【Codeforces round 382B】Urbanization的更多相关文章

  1. 【57.97%】【codeforces Round #380A】Interview with Oleg

    time limit per test1 second memory limit per test256 megabytes inputstandard input outputstandard ou ...

  2. 【42.86%】【Codeforces Round #380D】Sea Battle

    time limit per test1 second memory limit per test256 megabytes inputstandard input outputstandard ou ...

  3. 【26.83%】【Codeforces Round #380C】Road to Cinema

    time limit per test1 second memory limit per test256 megabytes inputstandard input outputstandard ou ...

  4. 【21.21%】【codeforces round 382D】Taxes

    time limit per test2 seconds memory limit per test256 megabytes inputstandard input outputstandard o ...

  5. 【Codeforces Round 1137】Codeforces #545 (Div. 1)

    Codeforces Round 1137 这场比赛做了\(A\).\(B\),排名\(376\). 主要是\(A\)题做的时间又长又交了两次\(wa4\)的. 这两次错误的提交是因为我第一开始想的求 ...

  6. 【Codeforces Round 1132】Educational Round 61

    Codeforces Round 1132 这场比赛做了\(A\).\(B\).\(C\).\(F\)四题,排名\(89\). \(A\)题\(wa\)了一次,少考虑了一种情况 \(D\)题最后做出来 ...

  7. 【Codeforces Round 1120】Technocup 2019 Final Round (Div. 1)

    Codeforces Round 1120 这场比赛做了\(A\).\(C\)两题,排名\(73\). \(A\)题其实过的有点莫名其妙...就是我感觉好像能找到一个反例(现在发现我的算法是对的... ...

  8. 【Codeforces Round 1129】Alex Lopashev Thanks-Round (Div. 1)

    Codeforces Round 1129 这场模拟比赛做了\(A1\).\(A2\).\(B\).\(C\),\(Div.1\)排名40. \(A\)题是道贪心,可以考虑每一个站点是分开来的,把目的 ...

  9. 【Codeforces Round 1117】Educational Round 60

    Codeforces Round 1117 这场比赛做了\(A\).\(B\).\(C\).\(D\).\(E\),\(div.2\)排名\(31\),加上\(div.1\)排名\(64\). 主要是 ...

随机推荐

  1. tomcat的classloader机制

    本系列博客打算分析一下tomcat7.x的源码,其中可能会穿插一些java基础知识的介绍  读tomcat的源码的时候,我建议和官方的User Guide一起阅读,明白tomcat做某件事情的目的之后 ...

  2. Fragment Summary 1/2

    转自:http://blog.csdn.net/lmj623565791/article/details/37970961 自从Fragment出现,曾经有段时间,感觉大家谈什么都能跟Fragment ...

  3. [Vue] Load components when needed with Vue async components

    In large applications, dividing the application into smaller chunks is often times necessary. In thi ...

  4. XML Parser Errors See Details for more Information XML Parser Error on line 1: Document root ele

    1.错误描写叙述 XML Parser Errors See Details for more Information XML Parser Error on line 1: Document roo ...

  5. 链表(三)——链表删除冗余结点&amp;插入结点到有序链表

    1.一个以递增方式排列的链表,去掉链表中的冗余值. 思路一:设有两个指针p和q.使p不动,q依次往后循环直到p->data不等于q->data,再将中间的冗余数据删除. 思路二:设有两个指 ...

  6. html的meta标签的charset应该用UTF-8还是utf-8?

    之前我也纠结过写html的时候是用<meta charset="UTF-8"/> 或者是 <meta charset="utf-8"/> ...

  7. python类继承时构造函数的运行问题

    假设子类定义了自己的__init__构造方法函数.当子类的实例对象被创建时,子类仅仅会运行自己的__init__方法函数.假设子类没有定义自己的构造方法函数.会沿着搜索树找到父类的构造方法函数去运行父 ...

  8. [RxJS] Implement pause and resume feature correctly through RxJS

    Eventually you will feel the need for pausing the observation of an Observable and resuming it later ...

  9. QQ号快速登录漏洞及被盗原理

    web安全:QQ号快速登录漏洞及被盗原理   为什么你什么都没干,但QQ空间中却发了很多小广告?也许你的QQ账号已经被盗.本文将讲解一个QQ的快速登录的漏洞. 我前阵子在论坛上看到一个QQ的快速登录的 ...

  10. 【29.70%】【codeforces 723D】Lakes in Berland

    time limit per test2 seconds memory limit per test256 megabytes inputstandard input outputstandard o ...