HDU 5840 This world need more Zhu

$Source:$ 2016中国大学生程序设计竞赛 - 网络选拔赛

$Problem:$
$\bullet$一棵树，$n$个结点各有一个值
$\bullet$$q$个询问，问$u\rightarrow v$这条路径上第$k, 2k..mk$个点的最大值。
$\bullet$类似于计蒜客 $17120$ $Xor$, 但是求的是这条路径上的异或值，而最大值只能通过数据结构来维护。

$Idea:$
$\bullet$$small-big$分块+树链剖分+RMQ/线段树/树状数组。
$\bullet$较大的$k$通过轻重链暴力往上跳，复杂度是$O(q*n/divk*log(divk))$
$\bullet$较小的$k$全部存起来，对每一个$k$都建一颗线段树。对于一个$u$点往上跳，无非是找到所有$(dep$ $mod$ $k)$相同的点比较大小。
$\bullet$这颗线段树的区间是以深度$(dep$ $mod$ $k)$为第一关键字，$dfs$序为第二关键字排序的。保证了同一深度下$dfs$序是有序的。这样对于每一条轻重链都可以在二分之后直接查询了。 预处理复杂度是$O(divk*n*log(n))$，查询复杂度是$O(q*log(n))$
$\bullet$大概是RMQ预处理时多出来的$log(n)$太多了… $divk$取了20才过

$Code:$

#include<bits/stdc++.h>
using namespace std;

#define fi first
#define se second
#define pb push_back
#define CLR(A, X) memset(A, X, sizeof(A))
#define bitcount(X) __builtin_popcountll(X)
typedef long long LL;
typedef pair<int, int> PII;
const double eps = 1e-10;
const int MOD = 255;
const LL INF = 1e18;
int dcmp(double x) { if(fabs(x) < eps) return 0; return x<0?-1:1; }

int BL = 20;
const int MAXN = 1e5+5;

struct Node {
    int u, fa, id;
};

int sz[MAXN], son[MAXN], p[MAXN], dep[MAXN], top[MAXN], dfn[MAXN], lab[MAXN], dfs_clock;
int dp[MAXN][25], a[MAXN], pos[MAXN], L[MAXN], R[MAXN], c[MAXN], ans[MAXN];
vector<int> G[MAXN];
vector<Node> H[MAXN];

void init(int n) {
    p[1] = 0, dep[1] = 1, top[1] = 1;
    dfs_clock = 0;
    for(int i = 1; i <= n; i++) G[i].clear(), H[i].clear();
}

void RMQ_init(int n, int *A) {
    for(int i = 0; i < n; i++) dp[i][0] = A[lab[pos[i+1]]];
    for(int j = 1; (1<<j) <= n; j++)
        for(int i = 0; i+(1<<j)-1 < n; i++)
            dp[i][j] = max(dp[i][j-1], dp[i+(1<<(j-1))][j-1]);
}

int RMQ(int L, int R) {
    int k = 0;
    while((1<<(k+1)) <= R-L+1) k++;
    return max(dp[L][k], dp[R-(1<<k)+1][k]);
}

int dfs_0(int u, int fa) {
    sz[u] = 1; son[u] = 0;
    for(int v:G[u]) if(v != fa) {
        p[v] = u;
        dep[v] = dep[u]+1;
        dfs_0(v, u);
        sz[u] += sz[v];
        if(sz[v] > sz[son[u]]) son[u] = v;
    }
}

int dfs_1(int u, int fa) {
    dfn[u] = ++dfs_clock;
    lab[dfs_clock] = u;
    if(son[u]) {
        top[son[u]] = top[u];
        dfs_1(son[u], u);
    }
    for(int v:G[u]) if(v!=fa && v!=son[u]) {
        top[v] = v;
        dfs_1(v, u);
    }
}

int lca(int u, int v) {
    while(top[u] != top[v]) {
        if(dep[top[u]] < dep[top[v]]) swap(u, v);
        u = p[top[u]];
    }
    if(dep[u] > dep[v]) swap(u, v);
    return u;
}

int kfa(int u, int k) {
    while(dep[u]-dep[top[u]] < k) {
        k -= (dep[u]-dep[top[u]]+1);
        u = p[top[u]];
        if(u == 0) return 0;
    }
    return lab[dfn[u]-k];
}

int cal(int u, int fa, int k) {
    int d = dep[u]%k, mx = 0;
    while(dep[u] >= dep[fa]) {
        int idr = upper_bound(pos+L[d], pos+R[d]+1, dfn[u])-pos-1;
        int l = dfn[top[u]];
        if(dep[lab[l]] < dep[fa]) l = dfn[fa];
        int idl = lower_bound(pos+L[d], pos+R[d]+1, l)-pos;
        if(idl <= idr) mx = max(mx, RMQ(idl-1, idr-1));
        u = p[top[u]];
    }
    return mx;
}

int bigcase(int u, int fa, int k) {
    int mx = 0;
    while(dep[u] >= dep[fa]) {
        mx = max(mx, a[u]);
        u = kfa(u, k);
    }
    return mx;
}

void smallcase(int k) {
    for(int i = 0; i < k; i++) c[i] = 0;
    for(int i = 1; i <= dfs_clock; i++) c[dep[i]%k]++;
    for(int i = 1; i < k; i++) c[i] += c[i-1];
    for(int i = dfs_clock; i >= 1; i--) pos[c[dep[lab[i]]%k]--] = i;
    RMQ_init(dfs_clock, a);
    for(int i = 1; i <= dfs_clock; i++) {
        int x = dep[lab[pos[i]]]%k;
        if(i==1 || x!=dep[lab[pos[i-1]]]%k) L[x] = i;
        R[x] = i;
    }
    for(Node t:H[k]) {
        int id = t.id, u = t.u, fa = t.fa;
        ans[id] = max(ans[id], cal(u, fa, k));
    }
}

int main() {
    int X;
    scanf("%d", &X);
    for(int cas = 1; cas <= X; cas++) {
        int n, m, u, v, k;
        scanf("%d%d", &n, &m);
        init(n);
        for(int i = 1; i <= n; i++) scanf("%d", &a[i]);
        for(int i = 1; i < n; i++) {
            scanf("%d%d", &u, &v);
            G[u].pb(v); G[v].pb(u);
        }
        dfs_0(1, -1); dfs_1(1, -1);
        for(int i = 1; i <= m; i++) {
            scanf("%d%d%d", &u, &v, &k);
            int fa = lca(u, v);
            int d = (dep[u]+dep[v]-2*dep[fa]+1)%k;
            v = kfa(v, d);
            u = kfa(u, k-1);
            if(k > BL) {
                ans[i] = max(bigcase(u, fa, k), bigcase(v, fa, k));
            }
            else {
                ans[i] = 0;
                H[k].pb((Node){u, fa, i});
                H[k].pb((Node){v, fa, i});
            }
        }
        for(int i = 1; i <= BL; i++) if(H[i].size()) {
            smallcase(i);
        }
        printf("Case #%d:\n", cas);
        for(int i = 1; i <= m; i++) printf("%d\n", ans[i]);
    }
    return 0;
}