LCA最近公共祖先 RMQ法 洛谷p3379

dfs序上两点之间深度最小的点就是最近公共祖先节点，用rmq在深度数组求区间最值即可

//#pragma GCC optimize(3,"Ofast","inline")
#include<unordered_map>
//#include<unordered_set>
#include<cstdio>
#include<iostream>
#include<cmath>
#include<functional>
#include<cstring>
#include<string>
#include<cstdlib>
#include<queue>
#include<map>
#include<algorithm>
#include<set>
#include<stack>
#include<vector>
#include<sstream>
#include<list>
using namespace std;
typedef long long ll;
const int INF=0x3f3f3f3f;
int mod=1000000007;
const int N=2e3+10;
const int maxn=5e5+7;
int num;
int f[2*maxn][25],d[maxn],vis[maxn],head[maxn],id[maxn],dfsx[2*maxn];
int n,m,root;
struct node
{
    int to,next;
} p[2*maxn];
void add(int from,int to)
{
    p[num].to=to;
    p[num].next=head[from];
    head[from]=num++;
}
void dfs(int x,int ceng)
{
    dfsx[num]=x;
    id[x]=num;
    d[x]=ceng;
    num++;
    vis[x]=1;
    for(int i=head[x]; i!=-1; i=p[i].next)
        if(!vis[p[i].to])
        {
            dfs(p[i].to,ceng+1);
            dfsx[num]=x;
            num++;
        }
}
void rmq()
{
    for(int i=1; i<=num; i++)f[i][0]=dfsx[i];
    for(int j=1; (1<<j)<=num; j++)
        for(int i=1; i+(1<<j)-1<=num; i++)
        {
            if(d[f[i][j-1]]<d[f[i+(1<<j-1)][j-1]])f[i][j]=f[i][j-1];
            else f[i][j]=f[i+(1<<j-1)][j-1];
        }
}
int lca(int x,int y)
{
    if(id[x]>id[y])swap(x,y);
    int len=id[y]-id[x]+1;
    int cnt=0,ans=1;
    while(ans<=len)
    {
        cnt++;
        ans<<=1;
    }
    cnt--;
    ans>>=1;
    if(d[f[id[x]][cnt]]<d[f[id[y]-ans+1][cnt]]) return f[id[x]][cnt];
    else return f[id[y]-ans+1][cnt];
}
int main()
{
    memset(head,-1,sizeof head);
    cin>>n>>m>>root;
    int x,y;
    for(int i=1; i<n; i++)
    {
        cin>>x>>y;
        add(x,y);
        add(y,x);
    }
    num=1;
    dfs(root,0);
    num--;
    rmq();
    while(m--)
    {
        cin>>x>>y;
        cout<<lca(x,y)<<endl;
    }
    return 0;
}