递推式题解

本文在此处同步更新
题目

样例

4
2 3 100
1 4 1000
3 3 1000000000
9 123456 78987

54
4
299
12153

首先我们可以推出如下式子

1. $a_n=x{a}_{n-1}+y{a}_{n-2}$

2. $a_n^2=x^2{a}_{n-1}^2+2xy{a}_{n-1}{a}_{n-2}+y^2{a}_{n-2}^2$

3. $s_n=s_{n-1}+a_n^2$

然后发现，要求$s_n$，先要知道$a_n^2$，要知道$a_n^2$，还要知道$a_{n-1}{a}_{n-2}$

虽然我们可以先求出$s_n$与$a_n^2$以及$a_{n+1}^2$（前2个都告诉你了），可$a_{n+2}$呢？

那就得知道$a_n{a}_{n+1}$了

可得$a_n{a}_{n+1}=a_n(x{a}_n+y{a}_{n-1})=x{a}_n^2+y{a}_n{a}_{n-1}$

$a_n^2$已知

$a_n{a}_{n-1}$可以通过同样的式子（除了初始化时，不过那可以直接算）求出来

于是我们知道了怎么构造$ans$矩阵
{ s [ n − 1 ] , a 1 a [ n ] 2 , a 2 a [ n − 1 ] 2 , a 3 a [ n ] ∗ a [ n − 1 ] a 4 } \{\\ s[n-1], a1 \\ a[n]^2, a2 \\ a[n-1]^2, a3 \\ a[n]*a[n-1] a4 \\ \} {s[n−1],a1a[n]2,a2a[n−1]2,a3a[n]∗a[n−1]a4}
至于$base$矩阵，可以直接根据上面的算

第一项可以直接加法得到，第三项则直接由原来第二项变换过来

第二、第四项的简化式子如下

$a[n+1{]}^2=x^2\ast a2+y^2\ast a3+2xy\ast a4$
$a[n]a[n+1]=x\ast a[n{]}^2+y\ast a[n]\ast a[n-1]=x\ast a2+y\ast a4$
可得$base$为
{ 1     1     0     0 0   x 2   y 2   2 x y 0     1     0     0 0     x     0     y } \{\\ 1~~~1~~~0~~~0\\ 0~x^2~y^2~2xy\\ 0~~~1~~~0~~~0\\ 0~~~x~~~0~~~y\\ \} {1   1   0   00 x2 y2 2xy0   1   0   00   x   0   y}
剩下的主要是预处理与快速幂
还有 不开long long见祖宗

所有预处理记得都取模一下，不然 long long会炸

快速幂有点烦，其他还好

//写得有些丑，见谅
#pragma GCC optimize(1, "inline", "Ofast")
#pragma GCC optimize(2, "inline", "Ofast")
#pragma GCC optimize(3, "inline", "Ofast")
// #include <bits/extc++.h>
#include <bits/stdc++.h>
#define memery() std::cerr << "The memory is " << (&ed - &st) / 1024.0 / 1024 << "MB." << std::endl
//#include <"C:\Users\admin\Documents\myhead\fstio.hpp">
//#include <"C:\Users\admin\Documents\myhead\def.hpp">
//#include <"C:\Users\admin\Documents\myhead\help.h">
// using namespace __gnu_cxx;
// using namespace __gnu_pbds;
using namespace std;
#define int long long
//变量区
bool st;
const int N = 5e6 + 7;
//please write sth.
int tk, a[4][4], b[4][4], c[4][4], n, m, x;
bool ed;
//函数区
void input()
{
    //please write sth.
    cin >> x >> n >> m;
}
void mul1(int b[4][4])
{
    //5 4
    for (int i = 0; i < 4; i++)
        for (int j = 0; j < 1; j++)
            c[i][j] = 0;
    for (int k = 0; k < 4; k++)
        for (int i = 0; i < 4; i++) //0 0  0 x  x 0
            for (int j = 0; j < 1; j++)
                (c[i][j] += ((1ll * b[i][k] * a[k][j]) % m + m) % m) %= m;
    for (int i = 0; i < 4; i++)
        for (int j = 0; j < 1; j++)
            a[i][j] = c[i][j];
}
void mul2(int a[4][4], int b[4][4], int c[4][4])
{
    //5 4
    for (int i = 0; i < 4; i++)
        for (int j = 0; j < 4; j++)
            c[i][j] = 0;
    for (int k = 0; k < 4; k++)
        for (int i = 0; i < 4; i++) //0 0  0 x  x 0
            for (int j = 0; j < 4; j++)
                (c[i][j] += ((1ll * a[i][k] * b[k][j]) % m + m) % m) %= m;
}
void k(int a[4][4], int b[4][4])
{
    int xx[4][4];
    mul2(a, b, xx);
    for (int i = 0; i < 4; i++) //0 0  0 x  x 0
        for (int j = 0; j < 4; j++)
            a[i][j] = xx[i][j];
}
void work()
{
    //please write sth.
    a[0][0] = 1;
    a[1][0] = (1ll * x * x) % m;
    a[2][0] = 1;
    a[3][0] = x;
    x *= 2, x %= m;
    int y = -1;
    int bb[4][4] = { { 1, 1, 0, 0 },
        { 0, (1ll * x * x) % m, (1ll * y * y) % m, (2ll * x * y) % m },
        { 0, 1, 0, 0 },
        { 0, x, 0, y } };
    for (int i = 0; i < 4; i++)
        for (int j = 0; j < 4; j++)
            b[i][j] = bb[i][j];
    n--;
    for (; n; n >>= 1, k(b, b))
        if (n & 1)
            mul1(b);
}
void output()
{
    //please write sth.
    if (n == 1)
        cout << 1 << endl;
    else
        cout << a[0][0] << endl;
}
//主函数
signed main()
{
    //memory();
    std::ios::sync_with_stdio(0);
    std::cin.tie(), std::cout.tie();
    cin >> tk;
    while (tk--) {
        input();
        work();
        output();
    }
}