結果

問題 No.3305 Shift Sort
ユーザー siganai
提出日時 2025-10-06 02:01:32
言語 C++17
(gcc 13.3.0 + boost 1.87.0)
結果
AC  
実行時間 136 ms / 2,000 ms
コード長 9,177 bytes
コンパイル時間 3,548 ms
コンパイル使用メモリ 323,776 KB
実行使用メモリ 7,808 KB
最終ジャッジ日時 2025-10-06 02:01:43
合計ジャッジ時間 11,145 ms
ジャッジサーバーID
（参考情報） 		judge5 / judge4
このコードへのチャレンジ
（要ログイン）
ファイルパターン 結果
other AC * 20
権限があれば一括ダウンロードができます

ソースコード

diff # 

#line 1 "main.cpp"
#include<bits/stdc++.h>
using namespace std;
#ifdef LOCAL
#include <debug.hpp>
#define debug(...) debug_print::multi_print(#__VA_ARGS__, __VA_ARGS__)
#else
#define debug(...) (static_cast<void>(0))
#endif
//#pragma GCC target("avx,avx2")
//#pragma GCC optimize("O3")
//#pragma GCC optimize("unroll-loops")
using ll = long long;
using ull = unsigned long long;
using ld = long double;
using pll = pair<ll,ll>;
using pii = pair<int,int>;
using vi = vector<int>;
using vvi = vector<vi>;
using vvvi = vector<vvi>;
using vl = vector<ll>;
using vvl = vector<vl>;
using vvvl = vector<vvl>;
using vul = vector<ull>;
using vpii = vector<pii>;
using vvpii = vector<vpii>;
using vpll = vector<pll>;
using vs = vector<string>;
template<class T> using pq = priority_queue<T,vector<T>, greater<T>>;
#define overload4(_1, _2, _3, _4, name, ...) name
#define overload3(a,b,c,name,...) name
#define rep1(n) for (ll UNUSED_NUMBER = 0; UNUSED_NUMBER < (n); ++UNUSED_NUMBER)
#define rep2(i, n) for (ll i = 0; i < (n); ++i)
#define rep3(i, a, b) for (ll i = (a); i < (b); ++i)
#define rep4(i, a, b, c) for (ll i = (a); i < (b); i += (c))
#define rep(...) overload4(__VA_ARGS__, rep4, rep3, rep2, rep1)(__VA_ARGS__)
#define rrep1(n) for(ll i = (n) - 1;i >= 0;i--)
#define rrep2(i,n) for(ll i = (n) - 1;i >= 0;i--)
#define rrep3(i,a,b) for(ll i = (b) - 1;i >= (a);i--)
#define rrep4(i,a,b,c) for(ll i = (a) + (((b)-(a)-1) / (c) - (((b)-(a)-1) % (c) && (((b)-(a)-1) ^ c) < 0)) * (c);i >= (a);i -= c)
#define rrep(...) overload4(__VA_ARGS__, rrep4, rrep3, rrep2, rrep1)(__VA_ARGS__)
#define all1(i) begin(i) , end(i)
#define all2(i,a) begin(i) , begin(i) + a
#define all3(i,a,b) begin(i) + a , begin(i) + b
#define all(...) overload3(__VA_ARGS__, all3, all2, all1)(__VA_ARGS__)
#define sum(...) accumulate(all(__VA_ARGS__),0LL)
template<class T> bool chmin(T &a, const T &b){ if(a > b){ a = b; return 1; } else return 0; }
template<class T> bool chmax(T &a, const T &b){ if(a < b){ a = b; return 1; } else return 0; }
template<class T> auto min(const T& a){return *min_element(all(a));}
template<class T> auto max(const T& a){return *max_element(all(a));}
template<class... Ts> void in(Ts&... t);
#define INT(...) int __VA_ARGS__; in(__VA_ARGS__)
#define LL(...) ll __VA_ARGS__; in(__VA_ARGS__)
#define STR(...) string __VA_ARGS__; in(__VA_ARGS__)
#define CHR(...) char __VA_ARGS__; in(__VA_ARGS__)
#define DBL(...) double __VA_ARGS__; in(__VA_ARGS__)
#define LD(...) ld __VA_ARGS__; in(__VA_ARGS__)
#define VEC(type, name, size) vector<type> name(size); in(name)
#define VV(type, name, h, w) vector<vector<type>> name(h, vector<type>(w)); in(name)
ll intpow(ll a, ll b){ll ans = 1; while(b){if(b & 1) ans *= a; a *= a; b /= 2;} return ans;}
ll modpow(ll a, ll b, ll p){ ll ans = 1; a %= p;if(a < 0) a += p;while(b){ if(b & 1) (ans *= a) %= p; (a *= a) %= p; b /= 2; } return ans; }
bool is_clamp(ll val,ll low,ll high) {return low <= val && val < high;}
void Yes() {cout << "Yes\n";return;}
void No() {cout << "No\n";return;}
void YES() {cout << "YES\n";return;}
void NO() {cout << "NO\n";return;}
template <typename U,typename T>
U floor(U a, T b) {return a / b - (a % b && (a ^ b) < 0);}
template <typename U,typename T>
U ceil(U x, T y) {return floor(x + y - 1, y);}
template <typename U,typename T>
T bmod(U x, T y) {return x - y * floor(x, y);}
template <typename U,typename T>
pair<U, T> divmod(U x, T y) {U q = floor(x, y);return {q, x - q * y};}
namespace IO{
#define VOID(a) decltype(void(a))
struct setting{ setting(){cin.tie(nullptr); ios::sync_with_stdio(false);fixed(cout); cout.precision(15);}} setting;
template<int I> struct P : P<I-1>{};
template<> struct P<0>{};
template<class T> void i(T& t){ i(t, P<3>{}); }
void i(vector<bool>::reference t, P<3>){ int a; i(a); t = a; }
template<class T> auto i(T& t, P<2>) -> VOID(cin >> t){ cin >> t; }
template<class T> auto i(T& t, P<1>) -> VOID(begin(t)){ for(auto&& x : t) i(x); }
template<class T, size_t... idx> void ituple(T& t, index_sequence<idx...>){in(get<idx>(t)...);}
template<class T> auto i(T& t, P<0>) -> VOID(tuple_size<T>{}){ituple(t, make_index_sequence<tuple_size<T>::value>{});} 
#undef VOID
}
#define unpack(a) (void)initializer_list<int>{(a, 0)...}
template<class... Ts> void in(Ts&... t){ unpack(IO :: i(t)); }
#undef unpack
constexpr long double PI = 3.141592653589793238462643383279L;
template <class F> struct REC {
    F f;
    REC(F &&f_) : f(forward<F>(f_)) {}
    template <class... Args> auto operator()(Args &&...args) const { return f(*this, forward<Args>(args)...); }};

constexpr int mod = 998244353;
//constexpr int mod = 1000000007;
#line 2 "library/data-structure/wavelet-matrix.hpp"
#include <immintrin.h>

struct bit_vector {
  using u32 = uint32_t;
  using i64 = int64_t;
  using u64 = uint64_t;

  static constexpr u32 w = 64;
  vector<u64> block;
  vector<u32> count;
  u32 n, zeros;

  inline u32 get(u32 i) const { return u32(block[i / w] >> (i % w)) & 1u; }
  inline void set(u32 i) { block[i / w] |= 1LL << (i % w); }

  bit_vector() {}
  bit_vector(int _n) { init(_n); }
  __attribute__((optimize("O3", "unroll-loops"))) void init(int _n) {
    n = zeros = _n;
    block.resize(n / w + 1, 0);
    count.resize(block.size(), 0);
  }

  __attribute__((target("popcnt"))) void build() {
    for (u32 i = 1; i < block.size(); ++i)
      count[i] = count[i - 1] + _mm_popcnt_u64(block[i - 1]);
    zeros = rank0(n);
  }

  inline u32 rank0(u32 i) const { return i - rank1(i); }
  __attribute__((target("bmi2,popcnt"))) inline u32 rank1(u32 i) const {
    return count[i / w] + _mm_popcnt_u64(_bzhi_u64(block[i / w], i % w));
  }
};

template <typename T>
struct WaveletMatrix {
  using u32 = uint32_t;
  using i64 = int64_t;
  using u64 = uint64_t;

  int n, lg;
  vector<T> a;
  vector<bit_vector> bv;

  WaveletMatrix(u32 _n) : n(max<u32>(_n, 1)), a(n) {}
  WaveletMatrix(const vector<T>& _a) : n(_a.size()), a(_a) { build(); }

  __attribute__((optimize("O3"))) void build() {
    lg = __lg(max<T>(*max_element(begin(a), end(a)), 1)) + 1;
    bv.assign(lg, n);
    vector<T> cur = a, nxt(n);
    for (int h = lg - 1; h >= 0; --h) {
      for (int i = 0; i < n; ++i)
        if ((cur[i] >> h) & 1) bv[h].set(i);
      bv[h].build();
      array<decltype(begin(nxt)), 2> it{begin(nxt), begin(nxt) + bv[h].zeros};
      for (int i = 0; i < n; ++i) *it[bv[h].get(i)]++ = cur[i];
      swap(cur, nxt);
    }
    return;
  }

  void set(u32 i, const T& x) { 
    assert(x >= 0);
    a[i] = x; 
  }

  inline pair<u32, u32> succ0(int l, int r, int h) const {
    return make_pair(bv[h].rank0(l), bv[h].rank0(r));
  }

  inline pair<u32, u32> succ1(int l, int r, int h) const {
    u32 l0 = bv[h].rank0(l);
    u32 r0 = bv[h].rank0(r);
    u32 zeros = bv[h].zeros;
    return make_pair(l + zeros - l0, r + zeros - r0);
  }

  // return a[k]
  T access(u32 k) const {
    T ret = 0;
    for (int h = lg - 1; h >= 0; --h) {
      u32 f = bv[h].get(k);
      ret |= f ? T(1) << h : 0;
      k = f ? bv[h].rank1(k) + bv[h].zeros : bv[h].rank0(k);
    }
    return ret;
  }

  // k-th (0-indexed) smallest number in a[l, r)
  T kth_smallest(u32 l, u32 r, u32 k) const {
    T res = 0;
    for (int h = lg - 1; h >= 0; --h) {
      u32 l0 = bv[h].rank0(l), r0 = bv[h].rank0(r);
      if (k < r0 - l0)
        l = l0, r = r0;
      else {
        k -= r0 - l0;
        res |= (T)1 << h;
        l += bv[h].zeros - l0;
        r += bv[h].zeros - r0;
      }
    }
    return res;
  }

  // k-th (0-indexed) largest number in a[l, r)
  T kth_largest(int l, int r, int k) {
    return kth_smallest(l, r, r - l - k - 1);
  }

  // count i s.t. (l <= i < r) && (v[i] < upper)
  int range_freq(int l, int r, T upper) {
    if (upper >= (T(1) << lg)) return r - l;
    int ret = 0;
    for (int h = lg - 1; h >= 0; --h) {
      bool f = (upper >> h) & 1;
      u32 l0 = bv[h].rank0(l), r0 = bv[h].rank0(r);
      if (f) {
        ret += r0 - l0;
        l += bv[h].zeros - l0;
        r += bv[h].zeros - r0;
      } else {
        l = l0;
        r = r0;
      }
    }
    return ret;
  }

  int range_freq(int l, int r, T lower, T upper) {
    return range_freq(l, r, upper) - range_freq(l, r, lower);
  }

  // max v[i] s.t. (l <= i < r) && (v[i] < upper)
  T prev_value(int l, int r, T upper) {
    int cnt = range_freq(l, r, upper);
    return cnt == 0 ? T(-1) : kth_smallest(l, r, cnt - 1);
  }

  // min v[i] s.t. (l <= i < r) && (lower <= v[i])
  T next_value(int l, int r, T lower) {
    int cnt = range_freq(l, r, lower);
    return cnt == r - l ? T(-1) : kth_smallest(l, r, cnt);
  }
};
/*
 * @brief Wavelet Matrix
 * @docs docs/data-structure-2d/wavelet-matrix.md
 */
#line 99 "main.cpp"
void solve() {
    INT(n,q);
    VEC(int,a,n);
    vi st;
    vi L(n,0);
    rrep(i,n) {
        while(!st.empty() && a[st.back()] < a[i]) {
            L[st.back()] = i + 1;
            st.pop_back();
        }
        st.emplace_back(i);
    }
    WaveletMatrix<int> WM(L);
    debug(L);
    rep(i,q) {
        INT(l,r);
        l--;
        int ans = (r - l) - WM.range_freq(l,r,l + 1);
        cout << ans << '\n';
    }
}   


int main() {
    //INT(TT);
    int TT = 1;
    rep(i,TT) solve();
}
0