CF 191C Fools and Roads lca 或者 树链剖分
They say that Berland has exactly two problems, fools and roads. Besides, Berland has n cities, populated by the fools and connected by the roads. All Berland roads are bidirectional. As there are many fools in Berland, between each pair of cities there is a path (or else the fools would get upset). Also, between each pair of cities there is no more than one simple path (or else the fools would get lost).
But that is not the end of Berland's special features. In this country fools sometimes visit each other and thus spoil the roads. The fools aren't very smart, so they always use only the simple paths.
A simple path is the path which goes through every Berland city not more than once.
The Berland government knows the paths which the fools use. Help the government count for each road, how many distinct fools can go on it.
Note how the fools' paths are given in the input.
The first line contains a single integer n (2 ≤ n ≤ 105) — the number of cities.
Each of the next n - 1 lines contains two space-separated integers ui, vi (1 ≤ ui, vi ≤ n, ui ≠ vi), that means that there is a road connecting cities ui and vi.
The next line contains integer k (0 ≤ k ≤ 105) — the number of pairs of fools who visit each other.
Next k lines contain two space-separated numbers. The i-th line (i > 0) contains numbers ai, bi (1 ≤ ai, bi ≤ n). That means that the fool number 2i - 1 lives in city ai and visits the fool number 2i, who lives in city bi. The given pairs describe simple paths, because between every pair of cities there is only one simple path.
Print n - 1 integer. The integers should be separated by spaces. The i-th number should equal the number of fools who can go on the i-th road. The roads are numbered starting from one in the order, in which they occur in the input.
5
1 2
1 3
2 4
2 5
2
1 4
3 5
2 1 1 1
5
3 4
4 5
1 4
2 4
3
2 3
1 3
3 5
3 1 1 1
In the first sample the fool number one goes on the first and third road and the fool number 3 goes on the second, first and fourth ones.
In the second sample, the fools number 1, 3 and 5 go on the first road, the fool number 5 will go on the second road, on the third road goes the fool number 3, and on the fourth one goes fool number 1.
给出一棵树,树有边权,边权初始都为0,给出k个操作,每一个操作给出u v
表示u,v路径上的所有边权都加1
所有操作完后,按照边输入的顺序输出每一条边的边权
一看,树链剖分裸题
但是想想,其实是可以不用树链剖分做的
可以用lca
我们令root=1
dis[i]表示节点i到root的距离
对于每一个操作u v,相当于
dis[u]++
dis[v]++
dis[lca(u,v)]-=2
所有操作完后,我们再一次dfs(root),自底向上累加所有dis,就可以了
dis[u]+=dis[son(u)]
然后按照输入的顺序输出边权就好啦
#include<cstdio>
#include<cstring>
#include<algorithm>
#include<iostream>
#include<map>
#include<set> #define LL long lon
#define pb push_back using namespace std; const int inf=0x3f3f3f3f;
const int maxn=1e5+; int dis[maxn];
int dep[maxn];
int dp[maxn][];
int e[maxn][]; struct Edge
{
int to,next;
};
Edge edge[maxn<<];
int head[maxn];
int tot; void init_edge()
{
memset(head,-,sizeof head);
tot=;
} void addedge(int u,int v)
{
edge[tot].to=v;
edge[tot].next=head[u];
head[u]=tot++;
} void solve(int ); int main()
{
init_edge();
int n;
scanf("%d",&n);
for(int i=;i<n;i++){
scanf("%d %d",&e[i][],&e[i][]);
addedge(e[i][],e[i][]);
addedge(e[i][],e[i][]);
}
solve(n); return ;
} void dfs0(int u,int pre)
{
dep[u]=dep[pre]+;
for(int i=head[u];~i;i=edge[i].next){
int v=edge[i].to;
if(v==pre)
continue;
dp[v][]=u;
dfs0(v,u);
}
} void init_dp(int n)
{
for(int j=;(<<j)<=n;j++){
for(int i=;i<=n;i++){
if(dp[i][j-]!=-)
dp[i][j]=dp[dp[i][j-]][j-];
}
}
} int query_lca(int a,int b)
{
if(dep[a]<dep[b])
swap(a,b);
int cnt;
for(cnt=;(<<cnt)<=dep[a];cnt++)
;
cnt--;
for(int j=cnt;j>=;j--){
if(dep[a]-(<<j)>=dep[b])
a=dp[a][j];
}
if(a==b)
return a;
for(int j=cnt;j>=;j--){
if(dp[a][j]!=- && dp[a][j]!=dp[b][j]){
a=dp[a][j];
b=dp[b][j];
}
}
return dp[a][];
} void dfs1(int u)
{
for(int i=head[u];~i;i=edge[i].next){
int v=edge[i].to;
if(v==dp[u][])
continue;
dfs1(v);
dis[u]+=dis[v];
}
} void solve(int n)
{
memset(dp,-,sizeof dp);
memset(dep,,sizeof dep);
memset(dis,,sizeof dis);
dfs0(,);
init_dp(n);
int q;
scanf("%d",&q);
for(int i=;i<=q;i++){
int u,v;
scanf("%d %d",&u,&v);
dis[u]++;
dis[v]++;
dis[query_lca(u,v)]-=;
}
dfs1(); for(int i=;i<n;i++){
int cur=e[i][];
if(dep[e[i][]]<dep[e[i][]])
cur=e[i][];
printf("%d",dis[cur]);
i==n-?puts(""):printf(" ");
}
return ;
}
CF 191C Fools and Roads lca 或者 树链剖分的更多相关文章
- [CF 191C]Fools and Roads[LCA Tarjan算法][LCA 与 RMQ问题的转化][LCA ST算法]
参考: 1. 郭华阳 - 算法合集之<RMQ与LCA问题>. 讲得很清楚! 2. http://www.cnblogs.com/lazycal/archive/2012/08/11/263 ...
- 【BZOJ3626】LCA(树链剖分,Link-Cut Tree)
[BZOJ3626]LCA(树链剖分,Link-Cut Tree) 题面 Description 给出一个n个节点的有根树(编号为0到n-1,根节点为0).一个点的深度定义为这个节点到根的距离+1. ...
- uva 12655 Trucks [LCA](树链剖分+MST)
The Subtle Balloons Company (SBC) is the main balloon provider for programming contests; it hashuge ...
- LCA 倍增||树链剖分
方法1:倍增 1498ms #include <iostream> #include <cstdio> #include <algorithm> #include ...
- 从lca到树链剖分 bestcoder round#45 1003
bestcoder round#45 1003 题,给定两个点,要我们求这两个点的树上路径所经过的点的权值是否出现过奇数次.如果是一般人,那么就是用lca求树上路径,然后判断是否出现过奇数次(用异或) ...
- P3379 【模板】最近公共祖先(LCA)(树链剖分)版
#include <bits/stdc++.h> #define read read() #define up(i,l,r) for(register int i = (l);i < ...
- 2018.09.16 bzoj3626: [LNOI2014]LCA(树链剖分)
传送门 树链剖分好题. 对于每个点维护一个值vi" role="presentation" style="position: relative;"&g ...
- BZOJ 3626 LCA(离线+树链剖分)
首先注意到这样一个事实. 树上两个点(u,v)的LCA的深度,可以转化为先将u到根路径点权都加1,然后求v到根路径上的总点权值. 并且该题支持离线.那么我们可以把一个区间询问拆成两个前缀和形式的询问. ...
- JZYZOJ1454 NOIP2015 D2T3_运输计划 二分 差分数组 lca tarjan 树链剖分
http://172.20.6.3/Problem_Show.asp?id=1454 从这道题我充分认识到我的脑子里好多水orz. 如果知道了这个要用二分和差分写,就没什么思考上的难点了(屁咧你写了一 ...
随机推荐
- scala言语基础学习十一
隐式转换 使用隐式转换加强现有的类型的功能-类似于设计模式的装饰模式
- POJ1087 A Plug for UNIX(网络流)
A Plug for UNIX Time Limit: 1000MS Memory Limit: 65536K Total S ...
- poj2375 强连通
题意:有一个 l * w 大小的滑雪场,每个格子都有一个高度,每个格子可以直接通到上下左右四个格子中高度小于等于自己的格子,现在要建立通道,能够连通任意两个格子,问最少建多少通道能够使所有格子能够互相 ...
- C++@类的静态成员变量和静态成员函数
参考: http://blog.csdn.net/morewindows/article/details/6721430 http://www.cnblogs.com/lzjsky/archive/2 ...
- lua堆栈操作常用函数学习二
/* ** basic stack manipulation */ LUA_API int <strong> (lua_gettop) (lua_State *L); </str ...
- MySQL删除重复记录只保留一条
删除表中重复记录,只保留一条: delete from 表名 where 字段ID in (select * from (select max(字段ID) from 表名 group by 重复的字段 ...
- OpenJudge计算概论-单词排序
/*===================================== 单词排序 总时间限制:1000ms 内存限制: 65536kB 描述 输入一行单词序列,相邻单词之间由1个或多个空格间隔 ...
- OpenJudge计算概论-矩阵归零消减序列和
矩阵归零消减序列和 总时间限制: 1000ms 内存限制: 65536kB 描述 给定一个n*n的矩阵( <= n <= ,元素的值都是非负整数).通过n-1次实施下述过程,可把这个矩阵转 ...
- angular.element方法汇总以及AngularJS 动态添加元素和删除元素
addClass()-为每个匹配的元素添加指定的样式类名after()-在匹配元素集合中的每个元素后面插入参数所指定的内容,作为其兄弟节点append()-在每个匹配元素里面的末尾处插入参数内容att ...
- .htaccess中301强制跳转到带www前缀或不带www的域名
相信很多站长朋友都有这样的的问题出现.即带www前缀的域名与不带www前缀的顶级域名收录情况是不同的.这个问题主要是由于搜索引擎对于顶级域名与二级域名权重判定不同造成的.毫无疑问地,唯一的域名能够给你 ...