hdu 6102 GCDispower 树状数组+容斥


GCDispower

Time Limit: 10000/5000 MS (Java/Others)    Memory Limit: 65536/65536 K (Java/Others)
Total Submission(s): 175    Accepted Submission(s): 87


Problem Description
There is a mysterious organization called GCD. They get a permutation P, permutation is a sequence of length N, only consists of number 1 to N and each number appears only once.
They want to gain magic power from this permutation. For the interval [L, R] they can get power:
i=LRj=i+1Rk=j+1R(gcd(P[i],P[j])==P[k])P[k]
Now they have M queries, for each query can you help them calculate how many power they can get?
 

Input
The first line is an integer T which indicates the case number.
And as for each case, first line is two integer N and M which indicates the length of permutation and number of query.
Next line is N integer which indicate the permutation P
Next M line, for each line is two integer L and R which indicate each query.

Limit
1T100
1N,M100000
1PiN , for any pair  (i,j) PiPj
1LiRiN

About 95% test case:  1N1000
Other test case:  1N100000
 

Output
As for each case, you need to output M integer which indicate the answer of each query.
 

Sample Input
  
  
2 3 1 3 2 1 1 3 6 3 6 5 4 3 2 1 1 6 2 6 3 5
 

Sample Output
  
  
1 8 5 0
 

Source
 

Recommend
liuyiding   |   We have carefully selected several similar problems for you:   6107  6106  6105  6104  6103 
参考的博客:http://blog.csdn.net/creatorx/article/details/77106506

题意:给你n个数,m次询问(l,r)的价值,l到r的价值定义为gcd(p[i],p[j])==p[k]的对数*p[k](i,j,k均在范围内且互不相等)。

思路:根据官方题解,题目可以转化为求l到r,p[i]/p[k]和p[j]/p[k]互质的对数*p[k],所以我们可以固定右端点r,求出所有左端点的答案。做法如下:先枚举右端点r,再用O(logn)的时间记录坐标<r并且是p[r]的倍数的位置,再用O(logn*logn)的时间从小到大排序。再把这些数从右往左扫一遍,假设当前位置为i,每次可以通过容斥求出i+1~r-1有多少个数和他互质,从而更新某段区间的答案,这里用树状数组或者线段树更新,具体方法看代码,不好说明。最后对每个r操作完即可查询右端点为r的询问的答案,记录一下就好了。

下面给代码:

#include <bits/stdc++.h>   
using namespace std;  
#define maxn 100005
#define inf 0x3f3f3f3f
typedef long long LL;
const int mod=1e9+7;
vector<pair<int,int> >status[maxn];
vector<pair<int,int> >query[maxn];
vector<int>prime[maxn];
int vis[maxn],p[maxn],index[maxn],temp[maxn],now[maxn];
LL tree[maxn],ans[maxn];
int n,m;
#define lowerbit(x) x&(-x)
void add(int pos,LL val)
{
	while(pos<=n)
	{
		tree[pos]+=val;
		pos+=lowerbit(pos);
	}
}
void update(int l,int r,LL val){
	if(l-1>0)
		add(l-1,-val);
	add(r,val);
}
LL get(int l,int r)
{
	LL ans=0;
	while(r)
	{
		ans+=tree[r];
		r-=lowerbit(r);
	}
	l--;
	while(l){
		ans-=tree[l];
		l-=lowerbit(l);
	}
	return ans;
}
LL getans(int x){
	LL cnt=0;
	for(int i=0;i<status[x].size();i++)
		cnt+=now[status[x][i].first]*status[x][i].second;
	return cnt;
}
void insert(int x,int val){
	for(int i=0;i<status[x].size();i++)
		now[status[x][i].first]+=val;
}
void init(){
	for(int i=2;i<maxn;i++)
		if(!vis[i])
			for(int j=i;j<maxn;j+=i){
				vis[j]=1;
				prime[j].push_back(i);
			}
	for(int i=1;i<maxn;i++)
		for(int j=0;j<(1<<prime[i].size());j++){
			status[i].push_back(make_pair(1,1));
			for(int k=0;k<prime[i].size();k++){
				if((1<<k)&j){
					status[i][j].first*=prime[i][k];
					status[i][j].second*=-1;
				}
			}			
		}
}
int main(){
	init();
	int t;
	scanf("%d",&t);
	while(t--){
		scanf("%d%d",&n,&m);
		memset(tree,0,sizeof(tree));
		for(int i=1;i<=n;i++){
			scanf("%d",&p[i]);
			index[p[i]]=i;
			query[i].clear();
		}
		for(int i=0;i<m;i++){
			int l,r;
			scanf("%d%d",&l,&r);
			query[r].push_back(make_pair(i,l));
		}
		for(int i=1;i<=n;i++){
			int length=0;
			for(int j=p[i]*2;j<=n;j+=p[i])
				if(index[j]<i)
					temp[length++]=index[j];
			sort(temp,temp+length);
			LL sum=0;
			for(int j=length-1;j>=0;j--){
				sum+=getans(p[temp[j]]/p[i])*p[i];
				insert(p[temp[j]]/p[i],1);
				if(j>0)
					update(temp[j-1]+1,temp[j],sum);
				else
					update(1,temp[0],sum);
			}
			for(int j=0;j<length;j++)
				insert(p[temp[j]]/p[i],-1);
			for(int j=0;j<query[i].size();j++)
				ans[query[i][j].first]=get(query[i][j].second,i);	
		}
		for(int i=0;i<m;i++)
			printf("%lld\n",ans[i]);
	}
}

  • 0
    点赞
  • 1
    收藏
    觉得还不错? 一键收藏
  • 0
    评论

“相关推荐”对你有帮助么?

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
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