Gym - 101147J Whistle's New Car 树上差分
15 seconds
512 megabytes
car.in
standard output
Whistle has bought a new car, which has an infinite fuel tank capacity.
He discovered an irregular country since it has n cities and there are exactly n - 1 roads between them, of course, all cities are connected. He is so much clever, he realized that the country is like a rooted tree of n nodes and node 1 is the root. Each city i has only one filling station by which he can fill his car's fuel tank in no more than Xi liter. Whistle liked the country very much, and he wants to know what the most attractive city in the country is. The attractiveness of the city i is defined by how much it’s reachable from other cities, in other words the attractiveness of city is the number of cities j that satisfies these condition:
- City j is in the subtree of city i (except for city i itself).
- Whistle will start at city j and will only fill his car’s fuel tank with Xj liters and never fill it again until he reach city i.
- Whistle should be able to reach city i with non-negative fuel.
He knows the length of every road and that 1 Km will take exactly 1 liter on any road.
As you know, Whistle is very clever, but he is not that good at programming, so he asked you to help him. He wants to know the attractiveness of each city, so that he can decide which city to live in.
The first line of input contains one integer T, the number of test cases.
The next line contains one integer (1 ≤ n ≤ 500, 000), The number of cities in the country.
The next line contains n integers (1 ≤ Xi ≤ 1, 000, 000, 000).
Each one of the next n - 1 line contains three integers A, B, C (1 ≤ A, B ≤ n and 1 ≤ C ≤ 1, 000, 000, 000), that means there is a road between city A and city B of length C.
For each test case, output a line containing n integers, the attractiveness of each city.
1
4
5 10 5 10
1 2 100
2 3 5
3 4 5
0 2 1 0
Large I/O files. Please consider using fast input/output methods.
设dis[i]表示从根节点1走到第 i个节点的路径长度。
那么任意的路径u-->v(前提是u是v的祖先,不然要用lca,这里不需要)可以表示为dis[v] - dis[u]
那么如果v能走回到u,则需要a[v] >= dis[v] - dis[u],即是dis[v] - a[v] <= dis[u],其中dis[v] - a[v]是一个确定值。
那么题意转换成,对于每一个节点u,有一个值dis[u],问其子树中,有多少个节点,满足dis[sonNode] - a[sonNode] <= dis[u]的。
这里可以用dfs序,然后用区间第k大的线段树或者分块、离线BIT之类的数据结构维护。
但是注意到dis[]是关于深度递增的,就是越走到下面,dis越大,也就是单调。
那么dfs的时候,用个vector记录一下路径长度和节点id。二分找到第一个祖先,使得满足dis[sonNode] - a[sonNode] <= dis[u]
那么可以知道ans[那个祖先节点 ---- 当前节点的爸爸],这条路径上的贡献都要加1
其实也就是打标记。
和一维差不多,L....R中加上1等价于sum[L] += 1, sum[R + 1] -= 1
这里一样,sum[当前节点的爸爸] += 1, sum[第一个满足的祖先的爸爸] -= 1
不能从上到下打标记,因为儿子数目不确定,而爸爸是确定的。
#include <cstdio>
#include <cstdlib>
#include <cstring>
#include <cmath>
#include <algorithm>
#include <assert.h>
#define IOS ios::sync_with_stdio(false)
using namespace std;
#define inf (0x3f3f3f3f)
typedef long long int LL; #include <iostream>
#include <sstream>
#include <vector>
#include <set>
#include <map>
#include <queue>
#include <string>
const int maxn = + ;
struct Edge {
int u, v, tonext, len;
}e[maxn * ];
int first[maxn], num;
int a[maxn];
void addEdge(int u, int v, int w) {
++num;
e[num].u = u, e[num].v = v, e[num].len = w;
e[num].tonext = first[u];
first[u] = num;
}
int ans[maxn];
vector< pair<LL, int> > vc;
int vis[maxn], DFN;
void dfs(int cur, LL nowLen, int fa) {
ans[fa]++;
int pos = lower_bound(vc.begin(), vc.end(), make_pair(nowLen - a[cur], )) - vc.begin();
pos--; //去到第一个满足的的祖先的爸爸
if (pos >= ) {
ans[vc[pos].second]--;
}
vc.push_back(make_pair(nowLen, cur));
for (int i = first[cur]; i; i = e[i].tonext) {
int v = e[i].v;
if (v == fa) continue;
dfs(v, nowLen + e[i].len, cur);
ans[cur] += ans[v];
}
vc.pop_back();
}
void work() {
num = ;
++DFN;
memset(first, false, sizeof first);
memset(ans, , sizeof ans);
int n;
scanf("%d", &n);
for (int i = ; i <= n; ++i) cin >> a[i];
for (int i = ; i <= n - ; ++i) {
int u, v, w;
scanf("%d%d%d", &u, &v, &w);
addEdge(u, v, w);
addEdge(v, u, w);
}
dfs(, , );
for (int i = ; i <= n; ++i) {
printf("%d ", ans[i]);
}
printf("\n");
} int main() {
#ifdef local
freopen("data.txt", "r", stdin);
// freopen("data.txt", "w", stdout);
#endif
freopen("car.in", "r", stdin);
int t;
scanf("%d", &t);
while (t--) work();
return ;
}
Gym - 101147J Whistle's New Car 树上差分的更多相关文章
- Codeforces Gym - 101147J Whistle's New Car
Discription Statements Whistle has bought a new car, which has an infinite fuel tank capacity. He di ...
- Gym 101147J Whistle's New Car(dfs)
https://vjudge.net/problem/Gym-101147J 题意: 有n个城市,每个城市有一个权值,表示在这个城市的加油站可以加多少油. 现在要计算每个城市i,有多少个城市j可以到达 ...
- 【树状数组】Gym - 101147J - Whistle's New Car
题意就是对每个点i,统计在其子树内(不含自身),且depj-depi<=xj的点有多少个. 把点分别按照dep-x和dep进行排序,离线处理, 每次把dep-x小于等于当前dep值的点插入树状数 ...
- 【BZOJ-4326】运输计划 树链剖分 + 树上差分 + 二分
4326: NOIP2015 运输计划 Time Limit: 30 Sec Memory Limit: 128 MBSubmit: 703 Solved: 461[Submit][Status] ...
- [luogu P3128][USACO15DEC]Max Flow [LCA][树上差分]
题目描述 Farmer John has installed a new system of pipes to transport milk between the stalls in his b ...
- 树上差分 (瞎bb) [树上差分][LCA]
做noip2015的运输计划写了好久好久写不出来 QwQ 于是先来瞎bb一下树上差分 混积分 树上差分有2个常用的功能: (1)记录从点i到i的父亲这条路径走过几次 (2)将每条路径(s,t ...
- [填坑]树上差分 例题:[JLOI2014]松鼠的新家(LCA)
今天算是把LCA这个坑填上了一点点,又复习(其实是预习)了一下树上差分.其实普通的差分我还是会的,树上的嘛,也是懂原理的就是没怎么打过. 我们先来把树上差分能做到的看一下: 1.找所有路径公共覆盖的边 ...
- 【NOIP2016】【LCA】【树上差分】【史诗级难度】天天爱跑步
学弟不是说要出丧题吗>>所以我就研究了1天lca又研究了1天tj然后研究了一天天天爱跑步,终于写了出来.(最后的平均用时为240ms...比学弟快了1倍...) 题意:给你颗树,然后有m个 ...
- BZOJ_4238_电压_树上差分+dfs树
BZOJ_4238_电压_树上差分+dfs树 Description 你知道Just Odd Inventions社吗?这个公司的业务是“只不过是奇妙的发明(Just Odd Inventions)” ...
随机推荐
- codeforces B. Ilya and Queries 解题报告
题目链接:http://codeforces.com/problemset/problem/313/B 题目意思:给出一个只有 "." 和 "#" 组成的序 ...
- C3P0 配置
C3P0是一个开源的JDBC连接池. 在Spring中,C3P0的一些配置,介绍如下(只列了一部分,不是全部) <!-- c3p0连接池配置 --> <bean id="d ...
- Java(一)——认识Java语言
1.Java语言简介 Java是一种可以撰写跨平台应用程序的面向对象的程序设计语言,具有卓越的通用性.高效性.平台移植性和安全性.Sun 公司对 Java 编程语言的解释是:Java 编程语言是个简单 ...
- js 常见的小数取整问题
1.四舍五入取整 Math.round(5/2) // 3 2.直接去掉小数,取整 parseInt(5/2); // 2 3.向上取整,有小数整数部分就加1 Math.ceil(1/3 ...
- Js常见的六种报错
EvalError: raised when an error occurs executing code in eval() EvalError:当一个错误发生在()执行的代码RangeError: ...
- C语言解释器的实现--存储结构(一)
目录: 1. 内存池 2. 栈 3. Hash表 1.内存池 在一些小的程序里,没什么必要添加内存管理模块在里面.但是对于比较复杂的代码,如果需要很多的内存操作,那么加入自己的内存管理是有必要的.至 ...
- spark运行模式之二:Spark的Standalone模式安装部署
Spark运行模式 Spark 有很多种模式,最简单就是单机本地模式,还有单机伪分布式模式,复杂的则运行在集群中,目前能很好的运行在 Yarn和 Mesos 中,当然 Spark 还有自带的 Stan ...
- wampServer 设置
设置端口 打开 C:\wamp\bin\apache\apache2.4.9\conf\httpd.conf 文件 找到“Listen 80”和“ServerName localhost:80”,紧接 ...
- 1.oracle中decode的一些巧妙用法
1.符号函数sign在decode中的用法--比较大小 select decode(sign(变量1-变量2),-1,变量1,变量2) from dual; --取较小值sign()函数根据某个值是0 ...
- 使用Bootstrap模态框实现增删改查功能
模态框(Modal)是覆盖在父窗体上的子窗体.通常,目的是显示来自一个单独的源的内容,可以在不离开父窗体的情况下有一些互动.子窗体可提供信息.交互等. 本文实现的是使用模态框实现简单的增删改查的功能. ...