[codeforces538E]Demiurges Play Again

试题描述

Demiurges Shambambukli and Mazukta love to watch the games of ordinary people. Today, they noticed two men who play the following game.

There is a rooted tree on n nodes, m of which are leaves (a leaf is a nodes that does not have any children), edges of the tree are directed from parent to children. In the leaves of the tree integers from 1 to m are placed in such a way that each number appears exactly in one leaf.

Initially, the root of the tree contains a piece. Two players move this piece in turns, during a move a player moves the piece from its current nodes to one of its children; if the player can not make a move, the game ends immediately. The result of the game is the number placed in the leaf where a piece has completed its movement. The player who makes the first move tries to maximize the result of the game and the second player, on the contrary, tries to minimize the result. We can assume that both players move optimally well.

Demiurges are omnipotent, so before the game they can arbitrarily rearrange the numbers placed in the leaves. Shambambukli wants to rearrange numbers so that the result of the game when both players play optimally well is as large as possible, and Mazukta wants the result to be as small as possible. What will be the outcome of the game, if the numbers are rearranged by Shambambukli, and what will it be if the numbers are rearranged by Mazukta? Of course, the Demiurges choose the best possible option of arranging numbers.

输入

The first line contains a single integer n — the number of nodes in the tree (1 ≤ n ≤ 2·105).

Each of the next n - 1 lines contains two integers ui and vi (1 ≤ ui, vi ≤ n) — the ends of the edge of the tree; the edge leads from node ui to node vi. It is guaranteed that the described graph is a rooted tree, and the root is the node 1.

输出

Print two space-separated integers — the maximum possible and the minimum possible result of the game.

输入示例


输出示例

 

数据规模及约定

见“输入”

题解

树形 dp。设 mn(u, 0) 表示对于以节点 u 为根的子树中所有叶子从 1 开始编号,当前轮到后手(即希望最终答案最小的人)走所能得到的最小编号;mn(u, 1) 表示轮到先手走所能得到的最小编号。那么显然,因为后手一定是选择下一步最小的儿子走;,我们要让先手选到最大值最小,可以把每个子树 v 中小于等于 mn(v, 0) 的编号拿出来密密地排列,大于 mn(v, 0) 的部分统统扔到后面以免它们占位置。

对于求最大值的情况,做法类似。

#include <iostream>
#include <cstdio>
#include <algorithm>
#include <cmath>
#include <stack>
#include <vector>
#include <queue>
#include <cstring>
#include <string>
#include <map>
#include <set>
using namespace std; const int BufferSize = 1 << 16;
char buffer[BufferSize], *Head, *Tail;
inline char Getchar() {
if(Head == Tail) {
int l = fread(buffer, 1, BufferSize, stdin);
Tail = (Head = buffer) + l;
}
return *Head++;
}
int read() {
int x = 0, f = 1; char c = Getchar();
while(!isdigit(c)){ if(c == '-') f = -1; c = Getchar(); }
while(isdigit(c)){ x = x * 10 + c - '0'; c = Getchar(); }
return x * f;
} #define maxn 200010
#define maxm 400010
#define oo 2147483647 int n, m, head[maxn], next[maxm], to[maxm];
void AddEdge(int a, int b) {
to[++m] = b; next[m] = head[a]; head[a] = m;
swap(a, b);
to[++m] = b; next[m] = head[a]; head[a] = m;
return ;
}
int cntl, mn[2][maxn], mx[2][maxn];
void calc(int u, int fa) {
bool has = 0;
for(int e = head[u]; e; e = next[e]) if(to[e] != fa)
has = 1, calc(to[e], u);
if(!has) cntl++, mn[0][u] = mn[1][u] = mx[0][u] = mx[1][u] = 1;
return ;
} void dp(int u, int fa) {
if(mn[1][u]) return ;
mn[1][u] = 0; mn[0][u] = oo;
mx[1][u] = oo; mx[0][u] = 0;
for(int e = head[u]; e; e = next[e]) if(to[e] != fa) {
dp(to[e], u);
mn[1][u] += mn[0][to[e]],
mn[0][u] = min(mn[0][u], mn[1][to[e]]),
mx[1][u] = min(mx[1][u], mx[0][to[e]]),
mx[0][u] += mx[1][to[e]];
}
// printf("%d: %d %d %d %d\n", u, mn[1][u], mn[0][u], mx[1][u], mx[0][u]);
return ;
} int main() {
n = read();
for(int i = 1; i < n; i++) {
int a = read(), b = read();
AddEdge(a, b);
} calc(1, 0);
dp(1, 0); printf("%d %d\n", cntl + 1 - mx[1][1], mn[1][1]); return 0;
}

[codeforces538E]Demiurges Play Again的更多相关文章

  1. Codeforces Round #300 E - Demiurges Play Again

    E - Demiurges Play Again 感觉这种类型的dp以前没遇到过... 不是很好想.. dp[u] 表示的是以u为子树进行游戏得到的值是第几大的. #include<bits/s ...

  2. Codeforces 538E Demiurges Play Again(博弈DP)

    http://codeforces.com/problemset/problem/538/E 题目大意: 给出一棵树,叶子节点上都有一个值,从1-m.有两个人交替从根选择道路,先手希望到达的叶子节点尽 ...

  3. 【codeforces 538E】Demiurges Play Again

    [题目链接]:http://codeforces.com/problemset/problem/538/E [题意] 给你一棵树; 有两个人,分别从根节点开始,往叶子节点的方向走; 每个人每次只能走一 ...

  4. Codeforces 刷水记录

    Codeforces-566F 题目大意:给出一个有序数列a,这个数列中每两个数,如果满足一个数能整除另一个数,则这两个数中间是有一条边的,现在有这样的图,求最大联通子图. 题解:并不需要把图搞出来, ...

  5. Codeforces Round #300 解题报告

    呜呜周日的时候手感一直很好 代码一般都是一遍过编译一遍过样例 做CF的时候前三题也都是一遍过Pretest没想着去检查... 期间姐姐提醒说有Announcement也自信不去看 呜呜然后就FST了 ...

随机推荐

  1. salt-stack系列报错

    master启动报错 实验环境为: [root@master salt]# cat /proc/version Linux version 3.10.0-327.el7.x86_64 (mockbui ...

  2. SQL 列拼接使用

    一个产品收藏表 Collection , 把该产品被收藏的人拼接在一列中如下: SQL SERVER SELECT ProjectID, UserIDs = ','+(STUFF((SELECT ', ...

  3. P1664 每日打卡心情好

    题目背景 在洛谷中,打卡不只是一个简单的鼠标点击动作,通过每天在洛谷打卡,可以清晰地记录下自己在洛谷学习的足迹.通过每天打卡,来不断地暗示自己:我又在洛谷学习了一天,进而帮助自己培养恒心.耐心.细心. ...

  4. P1720 月落乌啼算钱

    题目背景 (本道题目木有以藏歌曲……不用猜了……) <爱与愁的故事第一弹·heartache>最终章. 吃完pizza,月落乌啼知道超出自己的预算了.为了不在爱与愁大神面前献丑,只好还是硬 ...

  5. 用vue做一个酷炫的menu

    写在前面 最近看到一个非常酷炫的menu插件,一直想把它鼓捣成vue形式,谁让我是vue的死灰粉呢,如果这都不算爱

  6. Xilinx HLS

    Xilinx 的高层次综合(High Level Synthesis, HLS)技术是将C/C++/SystemC软件语言转换成Verilog或VHDL硬件描述语言的技术.现已应用在SDAccel,S ...

  7. Javaweb学习笔记10—文件上传与下载

    今天来讲javaweb的第10阶段学习.文件的上传与下载,今天主要说的是这个功能的实现,不用说了,听名字就是外行人也知道肯定很重要啦. 老规矩,首先先用一张思维导图来展现今天的博客内容.       ...

  8. Android-Emulator使用

    1.查看当前android支持的avd版本 2.创建Emulator avd: android create avd -n magicyu -t 2 -n后面接需要创建avd的名字,-t后面接需要创建 ...

  9. 纯css滚动公告栏目

    <!DOCTYPE html> <html> <head> <meta charset="UTF-8"> <title> ...

  10. 大数据开发学习之构建Hadoop集群-(0)

    有多种方式来获取hadoop集群,包括从其他人获取或是自行搭建专属集群,抑或是从Cloudera Manager 或apach ambari等管理工具来构建hadoop集群等,但是由自己搭建则可以了解 ...