Ural 1560 牛顿恒等式 + 线段树

SHOHOKUKU

已于 2022-03-31 19:58:15 修改

阅读量605

点赞数

分类专栏：数学数据结构文章标签：算法

于 2022-03-31 19:40:36 首次发布

本文链接：https://blog.csdn.net/neweryyy/article/details/123881045

版权

数据结构同时被 2 个专栏收录

203 篇文章 1 订阅

订阅专栏

数学

164 篇文章 0 订阅

订阅专栏

该博客介绍了如何运用牛顿恒等式解决关于基本对称多项式的计算问题。通过维护线段树，实现了在O(kqlogn)的时间复杂度内求解指定次数项的和，并递推出所有基本对称多项式。文章包含详细的算法实现及代码示例。

摘要由CSDN通过智能技术生成

题意

传送门 Ural 1560 Elementary Symmetric Functions

题解

根据牛顿恒等式
$\sum\limits_{i=1}^{k} S_ib_{k-i} + k\times b_k = 0, k\in N^+$ 其中 $S_k = \sum_{l\leq i\leq r} a_i^k$ ，且 $b_k = (-1)^kf_k$
$f_k = \sum\limits_{l\leq i_1<\dots <i_k\leq r}\Big(\prod\limits_{m=1}^{k}x_{i_m}\Big)$ 基本对称多项式 $f$ 的边界情况 $f_0 = 1, f_k = 0(k > r - l + 1)$ 。

已知 $k$ 次项求和结果，则可以递推出 $f$ 。维护 $k$ 棵线段树即可。总时间复杂度 $O(kq\log n)$ 。

#include <bits/stdc++.h>
using namespace std;
using ll = long long;
constexpr int MAXN = 8E4 + 5, SZ = 1 << 18;
int N, M, P;
ll A[MAXN];

struct ST
{
    int lim;
    ll dat[SZ];
    void set(int x) { lim = x; }
    void init(int k = 0, int l = 0, int r = N)
    {
        if (r - l == 1)
        {
            ll t = 1;
            for (int i = 0; i < lim; ++i)
                t = t * A[l] % P;
            dat[k] = t;
            return;
        }
        int m = (l + r) / 2, chl = k * 2 + 1, chr = k * 2 + 2;
        init(chl, l, m), init(chr, m, r);
        dat[k] = (dat[chl] + dat[chr]) % P;
    }
    void change(int a, int b, ll x, int k = 0, int l = 0, int r = N)
    {
        if (r <= a || b <= l)
            return;
        if (a <= l && r <= b)
        {
            ll t = 1;
            for (int i = 0; i < lim; ++i)
                t = t * x % P;
            dat[k] = t;
            return;
        }
        int m = (l + r) / 2, chl = k * 2 + 1, chr = k * 2 + 2;
        change(a, b, x, chl, l, m), change(a, b, x, chr, m, r);
        dat[k] = (dat[chl] + dat[chr]) % P;
    }
    ll ask(int a, int b, int k = 0, int l = 0, int r = N)
    {
        if (a <= l && r <= b)
            return dat[k];
        int m = (l + r) / 2, chl = k * 2 + 1, chr = k * 2 + 2;
        if (b <= m)
            return ask(a, b, chl, l, m);
        else if (m <= a)
            return ask(a, b, chr, m, r);
        return (ask(a, b, chl, l, m) + ask(a, b, chr, m, r)) % P;
    }
} tr[4];

ll inv(ll x)
{
    ll res = 1, n = P - 2, mod = P;
    while (n > 0)
    {
        if (n & 1)
            res = res * x % mod;
        x = x * x % mod, n >>= 1;
    }
    return res;
}

int main()
{
    ios::sync_with_stdio(0), cin.tie(0), cout.tie(0);
    cin >> N >> M >> P;
    for (int i = 0; i < N; ++i)
        cin >> A[i];
    for (int i = 0; i < 4; ++i)
    {
        tr[i].set(i + 1);
        tr[i].init();
    }
    while (M--)
    {
        char op;
        cin >> op;
        if (op == 'I')
        {
            int p, d;
            cin >> p >> d;
            --p;
            A[p] = (A[p] + d) % P;
            for (int i = 0; i < 4; ++i)
                tr[i].change(p, p + 1, A[p]);
        }
        else
        {
            int l, r, lim;
            cin >> l >> r >> lim;
            --l;
            vector<ll> s(lim + 1), f(lim + 1);
            for (int i = 0; i < lim; ++i)
                s[i + 1] = tr[i].ask(l, r);
            f[0] = 1;
            ll mod = P;
            for (int k = 1; k <= lim; ++k)
            {
                ll sum = 0;
                for (int i = 1; i <= k; ++i)
                {
                    ll d = s[i] * f[k - i] % mod;
                    if ((k - i) & 1)
                        d = mod - d;
                    sum += d;
                }
                if ((k + 1) & 1)
                    sum = mod - sum;
                sum = sum % mod * inv(k) % mod;
                f[k] = (sum + mod) % mod;
            }
            for (int i = 0; i <= lim; ++i)
                cout << f[i] << (i == lim ? '\n' : ' ');
        }
    }
    return 0;
}