链接
题解
枚举最小公倍数 x x x,乘法原理统计答案,显然这种情况对答案的贡献是一个连乘的式子,这个其实好统计,因为这个连乘式子的每一项都不超过 d ( x ) d(x) d(x)( d ( x ) d(x) d(x)表示 x x x的约数个数),我只要同类合并用一个快速幂搞搞就行了
然后还要减去那些不包含 x x x的
复杂度是 O ( n l o g 2 n ) O(nlog^2n) O(nlog2n),其中一个 l o g log log来自调和级数、另一个来自快速幂
代码
#include <bits/stdc++.h>
#include <ext/pb_ds/assoc_container.hpp>
#include <ext/pb_ds/tree_policy.hpp>
#define mod 1000000007ll
#define iinf 0x3f3f3f3f
#define linf (1ll<<60)
#define eps 1e-8
#define maxn 1000010
#define maxe 1000010
#define cl(x) memset(x,0,sizeof(x))
#define rep(_,__) for(_=1;_<=(__);_++)
#define em(x) emplace(x)
#define emb(x) emplace_back(x)
#define emf(x) emplace_front(x)
#define fi first
#define se second
#define de(x) cerr<<#x<<" = "<<x<<endl
using namespace std;
using namespace __gnu_pbds;
typedef long long ll;
typedef pair<int,int> pii;
typedef pair<ll,ll> pll;
ll read(ll x=0)
{
ll c, f(1);
for(c=getchar();!isdigit(c);c=getchar())if(c=='-')f=-f;
for(;isdigit(c);c=getchar())x=x*10+c-0x30;
return f*x;
}
struct EasyMath
{
ll fastpow(ll a, ll b, ll c)
{
ll t(a%c), ans(1ll);
for(;b;b>>=1,t=t*t%c)if(b&1)ans=ans*t%c;
return ans;
}
}em;
ll s[maxn], n, ans, lim;
vector<ll> f[100010];
int main()
{
ll i, j, x;
n=read();
rep(i,n)
{
s[x=read()]++;
lim=max(x,lim);
}
rep(i,lim)for(j=i;j<=lim;j+=i)f[j].emb(i);
for(i=lim;i;i--)s[i]+=s[i+1];
rep(i,lim)
{
ll t=1, last;
for(j=f[i].size()-1;~j;j--)
{
if(j==f[i].size()-1)
(t*=em.fastpow(j+1,s[f[i][j]],mod)-em.fastpow(j,s[f[i][j]],mod))%=mod;
else
(t*=em.fastpow(j+1,s[f[i][j]]-last,mod))%=mod;
last=s[f[i][j]];
}
(ans+=t)%=mod;
}
cout<<(ans+mod)%mod;
return 0;
}