hdu 5489 Removed Interval（线段树+LIS）

最新推荐文章于 2026-06-15 16:36:57 发布

原创最新推荐文章于 2026-06-15 16:36:57 发布 · 1.1k 阅读

1 ·

本内容遵循CC 4.0 BY-SA版权协议

GRADE:C 同时被 3 个专栏收录

737 篇文章

订阅专栏

HDU

582 篇文章

订阅专栏

数据结构-线段树

77 篇文章

订阅专栏

这篇博客介绍了如何利用线段树和最长递增子序列（LIS）解决HDU 5489问题。作者首先正序遍历求解LIS，然后逆序遍历过程中通过线段树查询特定位置前的LIS信息，从而找出符合条件的子序列。

题目链接：hdu 5489 Removed Interval

解题思路

正序一遍LIS，逆序一遍LIS。在正序的时候维护一棵线段树，逆序做的时候可以查询当前位置向前间隔L之后，并且终止val小于当前位置的LIS。

代码

#include <cstdio>
#include <cstring>
#include <algorithm>

using namespace std;
const int maxn = 100005;
#define lson(x) ((x)<<1)
#define rson(x) (((x)<<1)|1)

struct SegTree {
    int lc[maxn<<2], rc[maxn<<2], s[maxn<<2];

    void pushup(int u) { s[u] = max(s[lson(u)], s[rson(u)]); }

    void modify(int u, int x, int v) {
        if (lc[u] == rc[u]) {
            s[u] = v;
            return;
        }

        int mid = (lc[u] + rc[u]) >> 1;
        if (x <= mid) modify(lson(u), x, v);
        else modify(rson(u), x, v);
        pushup(u);
    }

    int query(int u, int l, int r) {
        if (l <= lc[u] && rc[u] <= r) return s[u];

        int mid = (lc[u] + rc[u]) >> 1, ret = 0;
        if (l <= mid) ret = max(ret, query(lson(u), l, r));
        if (r > mid) ret = max(ret, query(rson(u), l, r));
        return ret;
    }

    void build (int u, int l, int r) {
        lc[u] = l, rc[u] = r, s[u] = 0;
        if (l == r) return;

        int mid = (l + r) >> 1;
        build(lson(u), l, mid);
        build(rson(u), mid+1, r);
        pushup(u);
    }
}T;

struct State {
    int idx, val, pos;
    State (int idx = 0, int val = 0, int pos = 0):idx(idx), val(val), pos(pos) {}
}S[maxn];

inline bool cmpval(const State& a, const State& b) {
    if (a.val != b.val) return a.val < b.val;
    return a.idx < b.idx;
}

inline bool cmpidx(const State& a, const State& b) {
    return a.idx < b.idx;
}

int N, M, L, X[maxn], P[maxn];

void  init () {
    scanf("%d%d", &N, &L);
    for (int i = 0; i < N; i++) {
        scanf("%d", &X[i]);
        S[i] = State(i, X[i]);
    }
    sort(X, X+N);
    M = unique(X, X+N) - X;

    for (int i = 0; i < N; i++)
        S[i].val = lower_bound(X, X+M, S[i].val) - X;
    sort(S, S+N, cmpval);
    for (int i = 0; i < N; i++)
        S[i].pos = i + 1;
    sort(S, S+N, cmpidx);

    memset(P, 0, sizeof(P));
    for (int i = 0; i < N; i++)
        P[S[i].val]++;

    P[0]++;
    for (int i = 1; i < M; i++)
        P[i] += P[i-1];
}

struct Stack {
    int n, s[maxn];
    int search(int x) {
        return lower_bound(s, s + n, x) - s;
    }

    void set(int pos, int x) {
        n = max(n, pos+1);
        s[pos] = x;
    }
}sta;

int solve () {
    T.build(1, 0, N);
    sta.n = 0;

    for (int i = 0; i < N; i++) {
        int t = sta.search(S[i].val);
        T.modify(1, S[i].pos, t+1);
        sta.set(t, S[i].val);
    }

    for (int i = 1; i <= L; i++)
        T.modify(1, S[N-i].pos, 0);

    int ret = T.s[1];
    sta.n = 0;
    for (int i = N-1; i >= L; i--) {
            T.modify(1, S[i-L].pos, 0);
        int t = sta.search(M-S[i].val);
        //printf("%d-%d\n", t+1, T.query(1, 0, P[S[i].val]-1));
        ret = max(ret, t + 1 + T.query(1, 0, P[S[i].val]-1));
        sta.set(t, M-S[i].val);
    }
    return ret;
}

int main () {
    int cas;
    scanf("%d", &cas);
    for (int kcas = 1; kcas <= cas; kcas++) {
        init();
        printf("Case #%d: %d\n", kcas, solve());
    }
    return 0;
}