Ch’s gift

Time Limit: 6000/3000 MS (Java/Others)    Memory Limit: 65536/65536 K (Java/Others)
Total Submission(s): 2534    Accepted Submission(s): 887

题目链接

http://acm.hdu.edu.cn/showproblem.php?pid=6162

Problem Description

Mr.
Cui is working off-campus and he misses his girl friend very much.
After a whole night tossing and turning, he decides to get to his girl
friend's city and of course, with well-chosen gifts. He knows neither
too low the price could a gift be since his girl friend won't like it,
nor too high of it since he might consider not worth to do. So he will
only buy gifts whose price is between [a,b].
There are n cities in
the country and (n-1) bi-directional roads. Each city can be reached
from any other city. In the ith city, there is a specialty of price ci
Cui could buy as a gift. Cui buy at most 1 gift in a city. Cui starts
his trip from city s and his girl friend is in city t. As mentioned
above, Cui is so hurry that he will choose the quickest way to his girl
friend(in other words, he won't pass a city twice) and of course, buy as
many as gifts as possible. Now he wants to know, how much money does he
need to prepare for all the gifts?
 

Input

There are multiple cases.

For each case:
The first line contains tow integers n,m(1≤n,m≤10^5), representing the number of cities and the number of situations.
The second line contains n integers c1,c2,...,cn(1≤ci≤10^9), indicating the price of city i's specialty.
Then n-1 lines follows. Each line has two integers x,y(1≤x,y≤n), meaning there is road between city x and city y.
Next
m line follows. In each line there are four integers
s,t,a,b(1≤s,t≤n;1≤a≤b≤10^9), which indicates start city, end city, lower
bound of the price, upper bound of the price, respectively, as the
exact meaning mentioned in the description above

 

Output

Output m space-separated integers in one line, and the ith number should be the answer to the ith situation.
 

Sample Input

5 3
1 2 1 3 2
1 2
2 4
3 1
2 5
4 5 1 3
1 1 1 1
3 5 2 3

Sample Output

7 1 4
 

Source

 

题意

给你一棵树,每个点有点权,每次求一条路径点权大于等于a小于等于b的点权和,即求vi的和(a<=vi<=b)
 

题解

我在上次南昌网络赛遇到几乎一模一样的题,当时直接树链剖分+主席数过了,结果这次T了,而且这题数据暴力都能过,我还很不服气的拿暴力对拍,结果真的真的比暴力慢了三四倍吧,想哭了。
在这题上耗了两小时,终于选择向大佬低头,去看题解了,果然我还是太弱了。

以上来自以为蒟蒻的内心独白,接下来假装自己想出来的,讲波题解:
 
我们令sum(x)表示小于等于x的权值和,那么ans=sum(b)-sum(a-1)。当然对于每条路径,sum(x)都不一样,真的吗。
对于一条路径(l,r,a,b)先拆成两个路径,即(l,r,a-1,-1)和(l,r,b,1),并统称为(l,r,val,id),分别求答案,最后相减即可。
然后我们问题转化成了求路径小于等于val的权值和。
对两条路径(l1,r1,val1,id1) 和(l2,r2,val2,id2) ,如果val1<val2那么第一条路径的答案被第二条路径的答案包含。
接下来定义get_sum(u,v)为点u到v路径的权值和,一开始每个点都等于零。
这样不难想到将所有路径按val 从小到大排序,然后扫描路径数组,每次将树上节点权值小于等于val的所有点加上该点的权值,此路径的ans=get_sum(l,r)。
 

AC代码

 #include<bits/stdc++.h>
using namespace std;
#define ll long long
#define N 100050
#define INF 123456789
int n,m;
int tot,last[N];
ll ans[N];
int cnt,fa[N],dp[N],size[N],son[N],rk[N],kth[N],top[N];
struct Query
{
int l,r,id; ll val;
bool operator <(const Query&b)const
{return val<b.val;}
}a[N],que[N<<];
struct Edge{int from,to,s;}edges[N<<];
struct Tree{int l,r;ll sum;}tr[N<<];
template<typename T>void read(T&x)
{
ll k=; char c=getchar();
x=;
while(!isdigit(c)&&c!=EOF)k^=c=='-',c=getchar();
if (c==EOF)exit();
while(isdigit(c))x=x*+c-'',c=getchar();
x=k?-x:x;
}
void read_char(char &c)
{while(!isalpha(c=getchar())&&c!=EOF);}
void AddEdge(int x,int y)
{
edges[++tot]=Edge{x,y,last[x]};
last[x]=tot;
}
void dfs1(int x,int pre)
{
fa[x]=pre;
dp[x]=dp[pre]+;
size[x]=;
son[x]=;
for(int i=last[x];i;i=edges[i].s)
{
Edge &e=edges[i];
if (e.to==pre)continue;
dfs1(e.to,x);
size[x]+=size[e.to];
if (size[e.to]>size[son[x]])son[x]=e.to;
}
}
void dfs2(int x,int y)
{
rk[x]=++cnt;
kth[cnt]=x;
top[x]=y;
if (son[x]==)return;
dfs2(son[x],y);
for(int i=last[x];i;i=edges[i].s)
{
Edge &e=edges[i];
if (e.to==fa[x]||e.to==son[x])continue;
dfs2(e.to,e.to);
}
}
void bt(int x,int l,int r)
{
tr[x].l=l; tr[x].r=r; tr[x].sum=;
if (l==r)return;
int mid=(l+r)>>;
bt(x<<,l,mid);
bt(x<<|,mid+,r);
}
void update(int x,int p,ll tt)
{
if (p<=tr[x].l&&tr[x].r<=p)
{
tr[x].sum+=tt;
return;
}
int mid=(tr[x].l+tr[x].r)>>;
if (p<=mid)update(x<<,p,tt);
if (mid<p)update(x<<|,p,tt);
tr[x].sum=tr[x<<].sum+tr[x<<|].sum;
}
ll query(int x,int l,int r)
{
if (l<=tr[x].l&&tr[x].r<=r)
return tr[x].sum;
int mid=(tr[x].l+tr[x].r)>>; ll ans=;
if (l<=mid)ans+=query(x<<,l,r);
if (mid<r)ans+=query(x<<|,l,r);
return ans;
}
ll get_sum(int x,int y)
{
int fx=top[x],fy=top[y];ll ans=;
while(fx!=fy)
{
if (dp[fx]<dp[fy])swap(x,y),swap(fx,fy);
ans+=query(,rk[fx],rk[x]);
x=fa[fx]; fx=top[x];
}
if (dp[x]<dp[y])swap(x,y);
ans+=query(,rk[y],rk[x]);
return ans;
}
void work()
{
read(n); read(m);
for(int i=;i<=n;i++)read(a[i].val),a[i].id=i;
for(int i=;i<=n-;i++)
{
int x,y;
read(x); read(y);
AddEdge(x,y);
AddEdge(y,x);
}
int num=;
for(int i=;i<=m;i++)
{
int l,r,x,y;
read(l); read(r); read(x);read(y);
que[++num]=Query{l,r,-i,x-};
que[++num]=Query{l,r,i,y};
}
sort(a+,a+n+);
sort(que+,que+num+);
dfs1(,);
dfs2(,);
bt(,,n);
int ds=;
for(int i=;i<=num;i++)
{
while(ds<=n&&a[ds].val<=que[i].val)
{
update(,rk[a[ds].id],a[ds].val);
ds++;
}
ll sum=get_sum(que[i].l,que[i].r);
if (que[i].id<) ans[-que[i].id]-=sum;
else ans[que[i].id]+=sum;
}
printf("%lld",ans[]);
for(int i=;i<=m;i++)printf(" %lld",ans[i]);
printf("\n");
}
void clear()
{
tot=; cnt=;
memset(last,,sizeof(last));
memset(ans,,sizeof(ans));
}
int main()
{
#ifndef ONLINE_JUDGE
freopen("aa.in","r",stdin);
//freopen("my.out","w",stdout);
#endif
while()
{
clear();
work();
}
}

TLE代码(树链剖分+主席树)

 #include<bits/stdc++.h>
using namespace std;
#define ll long long
#define N 100050
#define INF 123456789
int n,m,w[N];ll b[N];
int tot,last[N];
int tree_num,root[N];
int cnt,fa[N],dp[N],size[N],son[N],rk[N],kth[N],top[N];
struct Edge{int from,to,s;}edges[N<<];
struct Tree{int l,r,ls,rs;ll sum;}tr[];
template<typename T>void read(T&x)
{
ll k=; char c=getchar();
x=;
while(!isdigit(c)&&c!=EOF)k^=c=='-',c=getchar();
if (c==EOF)exit();
while(isdigit(c))x=x*+c-'',c=getchar();
x=k?-x:x;
}
void read_char(char &c)
{while(!isalpha(c=getchar())&&c!=EOF);}
void AddEdge(int x,int y)
{
edges[++tot]=Edge{x,y,last[x]};
last[x]=tot;
}
void dfs1(int x,int pre)
{
fa[x]=pre;
dp[x]=dp[pre]+;
size[x]=;
son[x]=;
for(int i=last[x];i;i=edges[i].s)
{
Edge &e=edges[i];
if (e.to==pre)continue;
dfs1(e.to,x);
size[x]+=size[e.to];
if (size[e.to]>size[son[x]])son[x]=e.to;
}
}
void dfs2(int x,int y)
{
rk[x]=++cnt;
kth[cnt]=x;
top[x]=y;
if (son[x]==)return;
dfs2(son[x],y);
for(int i=last[x];i;i=edges[i].s)
{
Edge &e=edges[i];
if (e.to==fa[x]||e.to==son[x])continue;
dfs2(e.to,e.to);
}
}
void bt(int &x,int l,int r)
{
x=++tree_num;
tr[x].l=l; tr[x].r=r; tr[x].sum=;
if (l==r)return;
int mid=(l+r)>>;
bt(tr[x].ls,l,mid);
bt(tr[x].rs,mid+,r);
}
void add(int &x,int last,int p)
{
x=++tree_num;
tr[x]=tr[last];
tr[x].sum+=b[p];
if (tr[x].l==tr[x].r)return;
int mid=(tr[x].l+tr[x].r)>>;
if(p<=mid)add(tr[x].ls,tr[last].ls,p);
else add(tr[x].rs,tr[last].rs,p);
}
ll ask(int x,int y,int p)
{
if (tr[x].r<=p)return tr[y].sum-tr[x].sum;
int mid=(tr[x].l+tr[x].r)>>;ll ans=;
if (<=mid)ans+=ask(tr[x].ls,tr[y].ls,p);
if (mid<p)ans+=ask(tr[x].rs,tr[y].rs,p);
return ans;
}
ll get_sum(int x,int y,int tt)
{
int fx=top[x],fy=top[y];ll ans=;
while(fx!=fy)
{
if (dp[fx]<dp[fy])swap(x,y),swap(fx,fy);
ans+=ask(root[rk[fx]-],root[rk[x]],tt);
x=fa[fx]; fx=top[x];
}
if (dp[x]<dp[y])swap(x,y);
ans+=ask(root[rk[y]-],root[rk[x]],tt);
return ans;
}
void work()
{
read(n); read(m);
int num=;
for(int i=;i<=n;i++)read(w[i]),b[++num]=w[i];
b[++num]=INF;
for(int i=;i<=n-;i++)
{
int x,y;
read(x); read(y);
AddEdge(x,y);
AddEdge(y,x);
}
sort(b+,b+num+);
num=unique(b+,b+num+)-b-;
dfs1(,);
dfs2(,);
bt(root[],,num);
for(int i=;i<=n;i++)
{
int tt=lower_bound(b+,b+num+,w[kth[i]])-b;
add(root[i],root[i-],tt);
}
for(int i=;i<=m;i++)
{
if (i>)printf(" ");
int x,y,l,r;
read(x); read(y); read(l); read(r);
l=lower_bound(b+,b+num+,l)-b-;
r=upper_bound(b+,b+num+,r)-b-;
ll ans=get_sum(x,y,r);
ans-=get_sum(x,y,l);
printf("%lld",ans);
}
printf("\n");
}
void clear()
{
tot=; cnt=; tree_num=;
memset(last,,sizeof(last));
}
int main()
{
#ifndef ONLINE_JUDGE
freopen("aa.in","r",stdin);
#endif
while()
{
clear();
work();
}
}

L - Ch’s gift HDU - 6162的更多相关文章

  1. 2017 Multi-University Training Contest - Team 9 1002&&HDU 6162 Ch’s gift【树链部分+线段树】

    Ch’s gift Time Limit: 6000/3000 MS (Java/Others)    Memory Limit: 65536/65536 K (Java/Others)Total S ...

  2. HDU 6162 - Ch’s gift | 2017 ZJUT Multi-University Training 9

    /* HDU 6162 - Ch’s gift [ LCA,线段树 ] | 2017 ZJUT Multi-University Training 9 题意: N节点的树,Q组询问 每次询问s,t两节 ...

  3. HDU 6162 Ch’s gift (树剖 + 离线线段树)

    Ch’s gift Time Limit: 6000/3000 MS (Java/Others)    Memory Limit: 65536/65536 K (Java/Others)Total S ...

  4. hdu6162 Ch’s gift

    地址:http://acm.split.hdu.edu.cn/showproblem.php?pid=6162 题目: Ch’s gift Time Limit: 6000/3000 MS (Java ...

  5. Ch’s gift

    Ch’s gift Time Limit: 6000/3000 MS (Java/Others)    Memory Limit: 65536/65536 K (Java/Others) Proble ...

  6. 2017多校第9场 HDU 6162 Ch’s gift 树剖加主席树

    题目链接:http://acm.hdu.edu.cn/showproblem.php?pid=6162 题意:给出一棵树的链接方法,每个点都有一个数字,询问U->V节点经过所有路径中l < ...

  7. 【HDU 6162】 Ch’s gift

    [题目链接] http://acm.hdu.edu.cn/showproblem.php?pid=6162 [算法] 离线树剖 我们知道,u到v路径上权值为[A,B]的数的和 = u到v路径上权值小于 ...

  8. HDU 6162 Ch’s gift

    Mr. Cui is working off-campus and he misses his girl friend very much. After a whole night tossing a ...

  9. HDU 6162 Ch's gift(树链剖分+线段树)

    题意: 已知树上的每个节点的值和节点之间的关系建成了一棵树,现在查询节点u到节点v的最短路径上的节点值在l到r之间的节点值的和. 思路: 用树链剖分将树映射到线段树上,线段树上维护3个值,max,mi ...

随机推荐

  1. linux中zookeeper

    linux中zookeeper 安装jdk tar -zxvf jdk-11.0.1_linux-x64_bin.tar.gz -C /usr/src sudo vim /etc/profile 输入 ...

  2. CNN感受野计算

    无痛理解CNN中的感受野receptive field CNN中感受野的计算 从直观上讲,感受野就是视觉感受区域的大小.在卷积神经网络中,感受野的定义是决定某一层输出结果中一个元素所对应的输入层的区域 ...

  3. 分布式爬虫搭建系列 之四---scrapy分布式框架

    带录入SAFCDS 

  4. Linux系统设置Samba共享笔记

    # "不允许一个用户使用一个以上用户名与服务器或共享资源的多重连接" C: 之前曾经与主机成功进行连接,连接断开后未将连接状态清除 $: net use */del /y # &q ...

  5. Python的Flask框架使用Redis做数据缓存的配置方法

    flask配置redis 首先得下载flask的缓存插件Flask-Cache,使用pip下载. sudo pip install flask_cache 为应用扩展flask_cache   app ...

  6. 01.webservice介绍

    不实现异构,实现远程系统之间的调用.

  7. codeforce452DIV2——E. Segments Removal

    题目 Vasya has an array of integers of length n. Vasya performs the following operations on the array: ...

  8. Unity shader saturate

    当你想将颜色值规范到0~1之间时,你可能会想到使用saturate函数(saturate(x)的作用是如果x取值小于0,则返回值为0.如果x取值大于1,则返回值为1.若x在0到1之间,则直接返回x的值 ...

  9. c++原型模式(Prototype)

    原型模式是通过已经存在的对象的接口快速方便的创建新的对象. #include <iostream> #include <string> using namespace std; ...

  10. 在Sqlserver中使用Try Catch

      创建错误日志表: CREATE TABLE ErrorLog(errNum INT,ErrSev NVARCHAR(1000),ErrState INT,ErrProc NVARCHAR(1000 ...