Codeforces Round #456 (Div. 2) D. Fishes (期望 贪心)

D. Fishes
time limit per test
1 second
memory limit per test
256 megabytes
input
standard input
output
standard output

While Grisha was celebrating New Year with Ded Moroz, Misha gifted Sasha a small rectangular pond of size n × m, divided into cells of size 1 × 1, inhabited by tiny evil fishes (no more than one fish per cell, otherwise they'll strife!).

The gift bundle also includes a square scoop of size r × r, designed for fishing. If the lower-left corner of the scoop-net is located at cell (x, y), all fishes inside the square (x, y)...(x + r - 1, y + r - 1) get caught. Note that the scoop-net should lie completely inside the pond when used.

Unfortunately, Sasha is not that skilled in fishing and hence throws the scoop randomly. In order to not frustrate Sasha, Misha decided to release k fishes into the empty pond in such a way that the expected value of the number of caught fishes is as high as possible. Help Misha! In other words, put k fishes in the pond into distinct cells in such a way that when the scoop-net is placed into a random position among (n - r + 1)·(m - r + 1) possible positions, the average number of caught fishes is as high as possible.

Input

The only line contains four integers n, m, r, k (1 ≤ n, m ≤ 1051 ≤ r ≤ min(n, m)1 ≤ k ≤ min(n·m, 105)).

Output

Print a single number — the maximum possible expected number of caught fishes.

You answer is considered correct, is its absolute or relative error does not exceed 10 - 9. Namely, let your answer be a, and the jury's answer be b. Your answer is considered correct, if .

Examples
input
3 3 2 3
output
2.0000000000
input
12 17 9 40
output
32.8333333333
Note

In the first example you can put the fishes in cells (2, 1)(2, 2)(2, 3). In this case, for any of four possible positions of the scoop-net (highlighted with light green), the number of fishes inside is equal to two, and so is the expected value.





求期望,很显然,中心的那个得到的概率是最高的,然后其次的从外扩散。。。
我们就先从中间的开始,求这个点一共能有几种捕获方法,然后bfs向外扩散,可用优先队列来对其求最优解。
然后sum / 网面积

#include <iostream>
#include <algorithm>
#include <string.h>
#include <utility>
#include <stdio.h>
#include <vector>
#include <queue>
#include <map>
using namespace std;
using ll = long long;
const int dir[4][2] = {{-1, 0}, {0, 1}, {1, 0}, {0, -1}};
ll n, m, r, k;
struct Node {
    ll x, y;
    Node(ll _x, ll _y) : x(_x), y(_y){};
    Node() = default;
};

ll get_vs(Node const& n) {
    return (min(::n - r + 1, n.x) - max(1ll, n.x - r + 1) + 1) * (min(m - r + 1, n.y) - max(1ll, n.y - r + 1) + 1);
}

bool operator < (Node const& n1, Node const& n2) {
    return get_vs(n2) > get_vs(n1);
}
int main() {
    cin >> n >> m >> r >> k;
    map<pair<ll, ll>, bool> um;
    priority_queue<Node> que;
    que.emplace((n + 1) / 2, (m + 1) / 2);
    um[{(n + 1) / 2, (m + 1) / 2}] = 1;
    ll sum = 0;
    while (!que.empty()) {
        Node u = que.top();
        que.pop();
        sum += get_vs(u);
        k--;
        if (!k)
            break;
        for (int i = 0; i < 4; ++i) {
            ll tx = u.x + dir[i][0];
            ll ty = u.y + dir[i][1];
            if (tx >= 1 && tx <= n && ty >= 1 && ty <= m && um.find({tx, ty}) == um.end()) {
                um[{tx, ty}] = 1;
                que.emplace(tx, ty);
            }
        }
    }
    printf("%.10lf\n", sum * 1.0 / (n - r + 1) / (m - r + 1));
}



  • 0
    点赞
  • 0
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值