#826340 (C++17) No.5011 Better Mo's Algorithm is Needed!! (Weighted)

• 問題: No.5011 Better Mo's Algorithm is Needed!! (Weighted)
• ユーザー: NyaanNyaan
• 提出日時: 2022-12-17 03:01:43
• 言語: C++17 (gcc 12.3.0 + boost 1.83.0)
• 結果: AC
• 実行時間: 4,918 ms / 5,000 ms
• コード長: 12,091 bytes
• コンパイル時間: 3,530 ms
• 実行使用メモリ: 19,340 KB
• スコア: 42,648,720,899
• 最終ジャッジ日時: 2022-12-17 03:13:20
• 合計ジャッジ時間: 637,783 ms
• ジャッジサーバーID （参考情報）: judge10 / judge13

入力	結果	実行時間; 実行使用メモリ
testcase_00	AC	4,918 ms; 17,364 KB
testcase_01	AC	4,918 ms; 17,372 KB
testcase_02	AC	4,917 ms; 17,364 KB
testcase_03	AC	4,918 ms; 17,368 KB
testcase_04	AC	4,917 ms; 17,372 KB
testcase_05	AC	4,918 ms; 17,312 KB
testcase_06	AC	4,917 ms; 17,532 KB
testcase_07	AC	4,918 ms; 17,492 KB
testcase_08	AC	4,917 ms; 17,308 KB
testcase_09	AC	4,918 ms; 17,364 KB
testcase_10	AC	4,917 ms; 17,380 KB
testcase_11	AC	4,918 ms; 17,384 KB
testcase_12	AC	4,917 ms; 17,384 KB
testcase_13	AC	4,918 ms; 17,580 KB
testcase_14	AC	4,917 ms; 17,364 KB
testcase_15	AC	4,918 ms; 17,488 KB
testcase_16	AC	4,917 ms; 17,316 KB
testcase_17	AC	4,918 ms; 17,420 KB
testcase_18	AC	4,917 ms; 17,380 KB
testcase_19	AC	4,918 ms; 17,316 KB
testcase_20	AC	4,917 ms; 17,380 KB
testcase_21	AC	4,918 ms; 17,364 KB
testcase_22	AC	4,917 ms; 17,464 KB
testcase_23	AC	4,918 ms; 17,368 KB
testcase_24	AC	4,917 ms; 17,584 KB
testcase_25	AC	4,918 ms; 17,380 KB
testcase_26	AC	4,917 ms; 17,492 KB
testcase_27	AC	4,917 ms; 17,488 KB
testcase_28	AC	4,917 ms; 17,492 KB
testcase_29	AC	4,917 ms; 17,356 KB
testcase_30	AC	4,917 ms; 17,524 KB
testcase_31	AC	4,918 ms; 17,424 KB
testcase_32	AC	4,917 ms; 17,304 KB
testcase_33	AC	4,917 ms; 18,340 KB
testcase_34	AC	4,918 ms; 17,424 KB
testcase_35	AC	4,918 ms; 17,512 KB
testcase_36	AC	4,918 ms; 17,244 KB
testcase_37	AC	4,918 ms; 17,952 KB
testcase_38	AC	4,918 ms; 17,308 KB
testcase_39	AC	4,917 ms; 18,356 KB
testcase_40	AC	4,918 ms; 17,244 KB
testcase_41	AC	4,917 ms; 17,520 KB
testcase_42	AC	4,918 ms; 17,360 KB
testcase_43	AC	4,917 ms; 17,304 KB
testcase_44	AC	4,918 ms; 17,304 KB
testcase_45	AC	4,917 ms; 17,464 KB
testcase_46	AC	4,917 ms; 17,492 KB
testcase_47	AC	4,917 ms; 17,336 KB
testcase_48	AC	4,918 ms; 18,456 KB
testcase_49	AC	4,917 ms; 17,512 KB
testcase_50	AC	4,918 ms; 17,352 KB
testcase_51	AC	4,917 ms; 18,176 KB
testcase_52	AC	4,917 ms; 18,072 KB
testcase_53	AC	4,917 ms; 18,048 KB
testcase_54	AC	4,918 ms; 18,672 KB
testcase_55	AC	4,916 ms; 18,624 KB
testcase_56	AC	4,917 ms; 18,424 KB
testcase_57	AC	4,917 ms; 17,640 KB
testcase_58	AC	4,917 ms; 18,444 KB
testcase_59	AC	4,917 ms; 18,716 KB
testcase_60	AC	4,917 ms; 17,696 KB
testcase_61	AC	4,916 ms; 17,892 KB
testcase_62	AC	4,917 ms; 17,412 KB
testcase_63	AC	4,917 ms; 17,532 KB
testcase_64	AC	4,917 ms; 17,620 KB
testcase_65	AC	4,917 ms; 17,632 KB
testcase_66	AC	4,917 ms; 19,032 KB
testcase_67	AC	4,917 ms; 17,904 KB
testcase_68	AC	4,917 ms; 18,556 KB
testcase_69	AC	4,917 ms; 18,596 KB
testcase_70	AC	4,916 ms; 18,648 KB
testcase_71	AC	4,917 ms; 19,004 KB
testcase_72	AC	4,917 ms; 17,284 KB
testcase_73	AC	4,917 ms; 19,004 KB
testcase_74	AC	4,917 ms; 18,952 KB
testcase_75	AC	4,918 ms; 18,904 KB
testcase_76	AC	4,918 ms; 18,156 KB
testcase_77	AC	4,916 ms; 18,688 KB
testcase_78	AC	4,918 ms; 18,464 KB
testcase_79	AC	4,917 ms; 18,684 KB
testcase_80	AC	4,918 ms; 18,888 KB
testcase_81	AC	4,917 ms; 18,896 KB
testcase_82	AC	4,918 ms; 18,932 KB
testcase_83	AC	4,917 ms; 18,880 KB
testcase_84	AC	4,918 ms; 18,472 KB
testcase_85	AC	4,917 ms; 18,952 KB
testcase_86	AC	4,917 ms; 18,792 KB
testcase_87	AC	4,917 ms; 19,132 KB
testcase_88	AC	4,918 ms; 18,508 KB
testcase_89	AC	4,917 ms; 18,244 KB
testcase_90	AC	4,918 ms; 18,860 KB
testcase_91	AC	4,916 ms; 17,912 KB
testcase_92	AC	4,918 ms; 18,696 KB
testcase_93	AC	4,917 ms; 18,040 KB
testcase_94	AC	4,918 ms; 18,404 KB
testcase_95	AC	4,917 ms; 18,984 KB
testcase_96	AC	4,917 ms; 17,820 KB
testcase_97	AC	4,917 ms; 18,560 KB
testcase_98	AC	4,918 ms; 19,168 KB
testcase_99	AC	4,917 ms; 18,136 KB
testcase_100	AC	4,918 ms; 18,864 KB
testcase_101	AC	4,918 ms; 18,600 KB
testcase_102	AC	4,917 ms; 17,616 KB
testcase_103	AC	4,918 ms; 19,340 KB
testcase_104	AC	4,916 ms; 17,776 KB
testcase_105	AC	4,918 ms; 19,108 KB
testcase_106	AC	4,917 ms; 18,896 KB
testcase_107	AC	4,918 ms; 17,516 KB
testcase_108	AC	4,916 ms; 17,640 KB
testcase_109	AC	4,918 ms; 18,528 KB
testcase_110	AC	4,917 ms; 18,416 KB
testcase_111	AC	4,917 ms; 18,760 KB
testcase_112	AC	4,917 ms; 18,800 KB
testcase_113	AC	4,918 ms; 18,832 KB
testcase_114	AC	4,917 ms; 18,000 KB
testcase_115	AC	4,918 ms; 18,984 KB
testcase_116	AC	4,917 ms; 18,336 KB
testcase_117	AC	4,918 ms; 18,232 KB
testcase_118	AC	4,916 ms; 19,196 KB
testcase_119	AC	4,918 ms; 18,796 KB

ソースコード

/**
 *  date : 2022-12-17 03:01:23
 */

#define NDEBUG
using namespace std;

// intrinstic
#include <immintrin.h>

#include <algorithm>
#include <array>
#include <bitset>
#include <cassert>
#include <cctype>
#include <cfenv>
#include <cfloat>
#include <chrono>
#include <cinttypes>
#include <climits>
#include <cmath>
#include <complex>
#include <cstdarg>
#include <cstddef>
#include <cstdint>
#include <cstdio>
#include <cstdlib>
#include <cstring>
#include <deque>
#include <fstream>
#include <functional>
#include <initializer_list>
#include <iomanip>
#include <ios>
#include <iostream>
#include <istream>
#include <iterator>
#include <limits>
#include <list>
#include <map>
#include <memory>
#include <new>
#include <numeric>
#include <ostream>
#include <queue>
#include <random>
#include <set>
#include <sstream>
#include <stack>
#include <streambuf>
#include <string>
#include <tuple>
#include <type_traits>
#include <typeinfo>
#include <unordered_map>
#include <unordered_set>
#include <utility>
#include <vector>

// utility
namespace Nyaan {
using ll = long long;
using i64 = long long;
using u64 = unsigned long long;
using i128 = __int128_t;
using u128 = __uint128_t;

template <typename T>
using V = vector<T>;
template <typename T>
using VV = vector<vector<T>>;
using vi = vector<int>;
using vl = vector<long long>;
using vd = V<double>;
using vs = V<string>;
using vvi = vector<vector<int>>;
using vvl = vector<vector<long long>>;

template <typename T, typename U>
struct P : pair<T, U> {
  template <typename... Args>
  P(Args... args) : pair<T, U>(args...) {}

  using pair<T, U>::first;
  using pair<T, U>::second;

  P &operator+=(const P &r) {
    first += r.first;
    second += r.second;
    return *this;
  }
  P &operator-=(const P &r) {
    first -= r.first;
    second -= r.second;
    return *this;
  }
  P &operator*=(const P &r) {
    first *= r.first;
    second *= r.second;
    return *this;
  }
  template <typename S>
  P &operator*=(const S &r) {
    first *= r, second *= r;
    return *this;
  }
  P operator+(const P &r) const { return P(*this) += r; }
  P operator-(const P &r) const { return P(*this) -= r; }
  P operator*(const P &r) const { return P(*this) *= r; }
  template <typename S>
  P operator*(const S &r) const {
    return P(*this) *= r;
  }
  P operator-() const { return P{-first, -second}; }
};

using pl = P<ll, ll>;
using pi = P<int, int>;
using vp = V<pl>;

constexpr int inf = 1001001001;
constexpr long long infLL = 4004004004004004004LL;

template <typename T>
int sz(const T &t) {
  return t.size();
}

template <typename T, typename U>
inline bool amin(T &x, U y) {
  return (y < x) ? (x = y, true) : false;
}
template <typename T, typename U>
inline bool amax(T &x, U y) {
  return (x < y) ? (x = y, true) : false;
}

template <typename T>
inline T Max(const vector<T> &v) {
  return *max_element(begin(v), end(v));
}
template <typename T>
inline T Min(const vector<T> &v) {
  return *min_element(begin(v), end(v));
}
template <typename T>
inline long long Sum(const vector<T> &v) {
  return accumulate(begin(v), end(v), 0LL);
}

template <typename T>
int lb(const vector<T> &v, const T &a) {
  return lower_bound(begin(v), end(v), a) - begin(v);
}
template <typename T>
int ub(const vector<T> &v, const T &a) {
  return upper_bound(begin(v), end(v), a) - begin(v);
}

constexpr long long TEN(int n) {
  long long ret = 1, x = 10;
  for (; n; x *= x, n >>= 1) ret *= (n & 1 ? x : 1);
  return ret;
}

template <typename T, typename U>
pair<T, U> mkp(const T &t, const U &u) {
  return make_pair(t, u);
}

template <typename T>
vector<T> mkrui(const vector<T> &v, bool rev = false) {
  vector<T> ret(v.size() + 1);
  if (rev) {
    for (int i = int(v.size()) - 1; i >= 0; i--) ret[i] = v[i] + ret[i + 1];
  } else {
    for (int i = 0; i < int(v.size()); i++) ret[i + 1] = ret[i] + v[i];
  }
  return ret;
};

template <typename T>
vector<T> mkuni(const vector<T> &v) {
  vector<T> ret(v);
  sort(ret.begin(), ret.end());
  ret.erase(unique(ret.begin(), ret.end()), ret.end());
  return ret;
}

template <typename F>
vector<int> mkord(int N,F f) {
  vector<int> ord(N);
  iota(begin(ord), end(ord), 0);
  sort(begin(ord), end(ord), f);
  return ord;
}

template <typename T>
vector<int> mkinv(vector<T> &v) {
  int max_val = *max_element(begin(v), end(v));
  vector<int> inv(max_val + 1, -1);
  for (int i = 0; i < (int)v.size(); i++) inv[v[i]] = i;
  return inv;
}

vector<int> mkiota(int n) {
  vector<int> ret(n);
  iota(begin(ret), end(ret), 0);
  return ret;
}

template <typename T>
T mkrev(const T &v) {
  T w{v};
  reverse(begin(w), end(w));
  return w;
}

template <typename T>
bool nxp(vector<T> &v) {
  return next_permutation(begin(v), end(v));
}

template <typename T>
using minpq = priority_queue<T, vector<T>, greater<T>>;

}  // namespace Nyaan

// bit operation
namespace Nyaan {
__attribute__((target("popcnt"))) inline int popcnt(const u64 &a) {
  return _mm_popcnt_u64(a);
}
inline int lsb(const u64 &a) { return a ? __builtin_ctzll(a) : 64; }
inline int ctz(const u64 &a) { return a ? __builtin_ctzll(a) : 64; }
inline int msb(const u64 &a) { return a ? 63 - __builtin_clzll(a) : -1; }
template <typename T>
inline int gbit(const T &a, int i) {
  return (a >> i) & 1;
}
template <typename T>
inline void sbit(T &a, int i, bool b) {
  if (gbit(a, i) != b) a ^= T(1) << i;
}
constexpr long long PW(int n) { return 1LL << n; }
constexpr long long MSK(int n) { return (1LL << n) - 1; }
}  // namespace Nyaan

// inout
namespace Nyaan {

template <typename T, typename U>
ostream &operator<<(ostream &os, const pair<T, U> &p) {
  os << p.first << " " << p.second;
  return os;
}
template <typename T, typename U>
istream &operator>>(istream &is, pair<T, U> &p) {
  is >> p.first >> p.second;
  return is;
}

template <typename T>
ostream &operator<<(ostream &os, const vector<T> &v) {
  int s = (int)v.size();
  for (int i = 0; i < s; i++) os << (i ? " " : "") << v[i];
  return os;
}
template <typename T>
istream &operator>>(istream &is, vector<T> &v) {
  for (auto &x : v) is >> x;
  return is;
}

istream &operator>>(istream &is, __int128_t &x) {
  string S;
  is >> S;
  x = 0;
  int flag = 0;
  for (auto &c : S) {
    if (c == '-') {
      flag = true;
      continue;
    }
    x *= 10;
    x += c - '0';
  }
  if (flag) x = -x;
  return is;
}

istream &operator>>(istream &is, __uint128_t &x) {
  string S;
  is >> S;
  x = 0;
  for (auto &c : S) {
    x *= 10;
    x += c - '0';
  }
  return is;
}

ostream &operator<<(ostream &os, __int128_t x) {
  if (x == 0) return os << 0;
  if (x < 0) os << '-', x = -x;
  string S;
  while (x) S.push_back('0' + x % 10), x /= 10;
  reverse(begin(S), end(S));
  return os << S;
}
ostream &operator<<(ostream &os, __uint128_t x) {
  if (x == 0) return os << 0;
  string S;
  while (x) S.push_back('0' + x % 10), x /= 10;
  reverse(begin(S), end(S));
  return os << S;
}

void in() {}
template <typename T, class... U>
void in(T &t, U &...u) {
  cin >> t;
  in(u...);
}

void out() { cout << "\n"; }
template <typename T, class... U, char sep = ' '>
void out(const T &t, const U &...u) {
  cout << t;
  if (sizeof...(u)) cout << sep;
  out(u...);
}

void outr() {}
template <typename T, class... U, char sep = ' '>
void outr(const T &t, const U &...u) {
  cout << t;
  outr(u...);
}

struct IoSetupNya {
  IoSetupNya() {
    cin.tie(nullptr);
    ios::sync_with_stdio(false);
    cout << fixed << setprecision(15);
    cerr << fixed << setprecision(7);
  }
} iosetupnya;

}  // namespace Nyaan

// debug

#ifdef NyaanDebug
#define trc(...) (void(0))
#else
#define trc(...) (void(0))
#endif

#ifdef NyaanLocal
#define trc2(...) (void(0))
#else
#define trc2(...) (void(0))
#endif

// macro
#define each(x, v) for (auto&& x : v)
#define each2(x, y, v) for (auto&& [x, y] : v)
#define all(v) (v).begin(), (v).end()
#define rep(i, N) for (long long i = 0; i < (long long)(N); i++)
#define repr(i, N) for (long long i = (long long)(N)-1; i >= 0; i--)
#define rep1(i, N) for (long long i = 1; i <= (long long)(N); i++)
#define repr1(i, N) for (long long i = (N); (long long)(i) > 0; i--)
#define reg(i, a, b) for (long long i = (a); i < (b); i++)
#define regr(i, a, b) for (long long i = (b)-1; i >= (a); i--)
#define fi first
#define se second
#define ini(...)   \
  int __VA_ARGS__; \
  in(__VA_ARGS__)
#define inl(...)         \
  long long __VA_ARGS__; \
  in(__VA_ARGS__)
#define ins(...)      \
  string __VA_ARGS__; \
  in(__VA_ARGS__)
#define in2(s, t)                           \
  for (int i = 0; i < (int)s.size(); i++) { \
    in(s[i], t[i]);                         \
  }
#define in3(s, t, u)                        \
  for (int i = 0; i < (int)s.size(); i++) { \
    in(s[i], t[i], u[i]);                   \
  }
#define in4(s, t, u, v)                     \
  for (int i = 0; i < (int)s.size(); i++) { \
    in(s[i], t[i], u[i], v[i]);             \
  }
#define die(...)             \
  do {                       \
    Nyaan::out(__VA_ARGS__); \
    return;                  \
  } while (0)

namespace Nyaan {
void solve();
}
int main() { Nyaan::solve(); }

//


struct Timer {
  chrono::high_resolution_clock::time_point st;

  Timer() { reset(); }

  void reset() { st = chrono::high_resolution_clock::now(); }

  chrono::milliseconds::rep elapsed() {
    auto ed = chrono::high_resolution_clock::now();
    return chrono::duration_cast<chrono::milliseconds>(ed - st).count();
  }
};
using namespace Nyaan;

// ref : https://codeforces.com/blog/entry/61203
int64_t hilbertOrder(int x, int y, int pow, int rotate) {
  if (pow == 0) return 0;
  int hpow = 1 << (pow - 1);
  int seg = (x < hpow) ? ((y < hpow) ? 0 : 3) : ((y < hpow) ? 1 : 2);
  seg = (seg + rotate) & 3;
  const int rotateDelta[4] = {3, 0, 0, 1};
  int nx = x & (x ^ hpow), ny = y & (y ^ hpow);
  int nrot = (rotate + rotateDelta[seg]) & 3;
  int64_t subSquareSize = int64_t(1) << (2 * pow - 2);
  int64_t ans = seg * subSquareSize;
  int64_t add = hilbertOrder(nx, ny, pow - 1, nrot);
  ans += (seg == 1 || seg == 2) ? add : (subSquareSize - add - 1);
  return ans;
}

int rui[200003], p[200003], d[200003];

void q() {
  Timer timer;
  inl(N, Q, WT, ST);
  vl W(N);
  in(W);
  rep(i, N) rui[i + 1] = rui[i] + W[i];
  // cerr<<Max(W)<<endl;
  // cerr<<Sum(W)<<endl;

  vp xy(Q);
  in(xy);
  each2(x, y, xy) x--;
  // 0 番目 : (0, 0) (始点)
  constexpr int start = 0;
  xy.insert(begin(xy), pl{0, 0});

  // hilbert order でソート
  vp dat;
  rep(i, Q + 1) {
    ll x = xy[i].fi;
    ll y = xy[i].se;
    // ll h=hilbertOrder(x,y,30,0);
    ll h = hilbertOrder(rui[x], rui[y], 30, 0);
    dat.emplace_back(h, i);
  }
  sort(all(dat));
  rep(i, Q + 1) p[i] = dat[i].se;
  assert(p[0] == start);

  // 2-opt
  // ref : http://yosupo06.github.io/Algorithm/src/datastructure/sort_mo.hpp
  auto dist = [&](int i, int j) {
    auto& [x1, y1] = xy[i];
    auto& [x2, y2] = xy[j];
    return abs(rui[x2] - rui[x1]) + abs(rui[y2] - rui[y1]);
  };
  [[maybe_unused]] auto dsum = [&]() {
    ll ans = 0;
    rep(i, Q) ans += dist(p[i], p[i + 1]);
    return ans;
  };

  int TL = 4900;
  int dmax = 600;
  rep(i, Q) d[i] = dist(p[i], p[i + 1]);
  [[maybe_unused]] int rev_count = 0;

  while (true) {
    if (timer.elapsed() > TL) break;
    trc2(dmax, dsum(), timer.elapsed(), rev_count);
    for (int i = 1; i <= Q; i++) {
      if (i % 5 == 0 and timer.elapsed() > TL) break;
      int pre1 = d[i - 1];
      int js = i + 1, je = min<int>(i + dmax, Q);
      for (int j = je - 1; j >= js; j--) {
        // (i-1) - (i), (j) - (j+1)
        int pre2 = d[j];
        int now1 = dist(p[i - 1], p[j]);
        int now2 = dist(p[i], p[j + 1]);
        if (now1 + now2 < pre1 + pre2) {
          reverse(p + i, p + j + 1);
          reverse(d + i, d + j);
          d[i - 1] = pre1 = now1;
          d[j] = now2;
#ifdef NyaanLocal
          rev_count++;
#endif
        }
      }
    }
    dmax += 20;
  }
  trc2(dmax, dsum(), timer.elapsed(), rev_count);
  for (int i = 1; i <= Q; i++) cout << p[i] << " \n"[i == Q];
}

void Nyaan::solve() {
  int t = 1;
  // in(t);
  while (t--) q();
}