对于sigma_{l<=i<=r}dep[LCA(i,z)],其(num[i]表示:l到r,有num[i]个点在i子树中)
= deep[z] * num[z] + deep[ fa(z) ] * ( num[ fa(z) ] - num[z] ) + .... + deep[root] * ( num[root] - num[ ?? ] )
= deep[ z ]*num[z] -0
+deep[ fa(z) ]*num[ fa(z) ] - deep[ fa(z) ]*num[z]
+...
+deep[root]*num[root] - deep[ fa(fa(fa(...))) ]*num[ fa(fa(fa(...))) ]
= num[z] + num[ fa(z) ] + num[ fa(fa(z)) ] + ... + num[root]
而对于询问区间[l,r],我们可以将其插为[1,l-1]和[1,r]两个区间,sigma_{l<=i<=r}dep[LCA(i,z)] = sigma_{1<=i<=r}dep[LCA(i,z)] - sigma_{1<=i<=l-1}dep[LCA(i,z)],
这样的话,我们就可以离线统计所有分拆后的区间。
而剩下的,则采用树链剖分求解
//#pragma comment(linker, "/STACK:102400000,102400000")
#include<cstdio>
#include<cstring>
#include<vector>
#include<queue>
#include<cmath>
#include<cctype>
#include<string>
#include<algorithm>
#include<iostream>
#include<ctime>
#include<map>
#include<set>
using namespace std;
#define MP(x,y) make_pair((x),(y))
#define PB(x) push_back(x)
typedef long long LL;
//typedef unsigned __int64 ULL;
/* ****************** */
const int INF=100011122;
const double INFF=1e100;
const double eps=1e-8;
const int mod=1000000009;
const int NN=50010;
const int MM=1000010;
/* ****************** */
struct Q
{
int r,x,id,ad;
Q(){}
Q(int a,int b,int c,int d)
{
r=a;
x=b;
id=c;
ad=d;
}
bool operator<(const Q &tt)const
{
return r<tt.r;
}
}q[NN*2];
LL ans[NN];
struct TR
{
int l,r,fg;
LL sum;
int mid()
{
return (l+r)>>1;
}
}tr[NN*4];
//c_dfs1
int anc[NN];//anc[i] i的父亲
int per[NN];//pre[i] -1:i下面没有重边;其他x表示通过重边连到的儿子
int si[NN]; //si[i] i子树有多少个儿子
//c_dfs2
int pos[NN];//pos[i] i在oula序中的位置
int top[NN];//top[i] i向上通过重边连到的最高点;top[]相同,说明在一条链上面
int tsp;
struct G
{
int v,next;
}E[NN];
int p[NN],T;
void add(int u,int v)
{
E[T].v=v;
E[T].next=p[u];
p[u]=T++;
}
void push_up(int R)
{
tr[R].sum=tr[R<<1].sum+tr[R<<1|1].sum;
}
void down(int R,int col)
{
tr[R].sum+=(LL)(tr[R].r-tr[R].l+1)*col;
tr[R].fg+=col;
}
void push_down(int R)
{
if(tr[R].fg!=0)
{
down(R<<1,tr[R].fg);
down(R<<1|1,tr[R].fg);
tr[R].fg=0;
}
}
void build(int l,int r,int R)
{
tr[R].l=l;
tr[R].r=r;
tr[R].fg=0;
tr[R].sum=0;
if(l==r)return;
int mid=tr[R].mid();
build(l,mid,R<<1);
build(mid+1,r,R<<1|1);
}
LL query(int l,int r,int R)
{
if(l<=tr[R].l && tr[R].r<=r)
return tr[R].sum;
push_down(R);
LL t=0;
int mid=tr[R].mid();
if(l<=mid)
t=t+query(l,r,R<<1);
if(r>=mid+1)
t=t+query(l,r,R<<1|1);
return t;
}
void update(int l,int r,int R,int col)
{
if(l<=tr[R].l && tr[R].r<=r)
{
down(R,col);
return;
}
push_down(R);
int mid=tr[R].mid();
if(l<=mid)
update(l,r,R<<1,col);
if(r>=mid+1)
update(l,r,R<<1|1,col);
push_up(R);
}
//root->v的和
LL c_query(int v)
{
LL s=0;
while(v!=-1)
{
s=s+query(pos[top[v]],pos[v],1);
v=top[v];
v=anc[v];
}
return s;
}
//root->v +1
void c_update(int v)
{
while(v!=-1)
{
update(pos[top[v]],pos[v],1,1);
v=top[v];
v=anc[v];
}
}
void c_dfs1(int u,int fa)
{
int i,v;
si[u]=1;
per[u]=-1;
anc[u]=fa;
for(i=p[u];i+1;i=E[i].next)
{
v=E[i].v;
if(v==fa)continue;
c_dfs1(v,u);
si[u]+=si[v];
if(per[u]==-1 || si[v]>si[ per[u] ])
per[u]=v;
}
}
void c_dfs2(int u,int fa,int now_head)
{
pos[u]=++tsp;
top[u]=now_head;
if(per[u]!=-1)c_dfs2(per[u],u,now_head);
int i,v;
for(i=p[u];i+1;i=E[i].next)
{
v=E[i].v;
if(v==fa || v==per[u])continue;
c_dfs2(v,u,v);
}
}
int main()
{
int n,m,i,j,v;
int l,r,x,tol;
while(scanf("%d%d",&n,&m)!=EOF)
{
memset(p,-1,sizeof(p));
T=0;
build(1,n,1);
tol=0;
for(i=2;i<=n;i++)
{
scanf("%d",&v);
v++;
add(v,i);
}
for(i=0;i<m;i++)
{
scanf("%d%d%d",&l,&r,&x);
q[tol++]=Q(l,x+1,i,-1);
q[tol++]=Q(r+1,x+1,i,1);
}
c_dfs1(1,-1);
tsp=0;
c_dfs2(1,-1,1);
sort(q,q+tol);
j=0;
memset(ans,0,sizeof(ans));
for(i=0;i<=n;i++)
{
if(i)
{
c_update(i);
}
while(j<tol && q[j].r==i)
{
LL t=c_query(q[j].x);
t=t*q[j].ad;
ans[ q[j].id ]+=t;
j++;
}
}
for(i=0;i<m;i++)
{
printf("%lld\n",ans[i]%201314);
}
}
return 0;
}