需要快速求出
C
k
=
∑
i
⨂
j
=
k
A
i
∗
B
j
C_k=\sum_{i\bigotimes j=k}A_i*B_j
Ck=∑i⨂j=kAi∗Bj
考虑构造一个可逆的矩阵满足
(
f
∗
A
)
∗
(
f
∗
B
)
=
(
f
∗
C
)
(f*A)*(f*B)=(f*C)
(f∗A)∗(f∗B)=(f∗C)
那么只需要快速求出
f
∗
A
,
f
∗
B
,
f
−
1
∗
(
f
∗
C
)
f*A,f*B,f^{-1}*(f*C)
f∗A,f∗B,f−1∗(f∗C) 即可
推导矩阵的系数:
(
f
∗
A
)
[
x
]
∗
(
f
∗
B
)
[
x
]
=
(
f
∗
C
)
[
x
]
(f*A)[x]*(f*B)[x]=(f*C)[x]
(f∗A)[x]∗(f∗B)[x]=(f∗C)[x]
∑
i
=
0
n
f
x
,
i
∗
a
i
∗
∑
i
=
0
n
f
x
,
i
b
i
=
∑
i
=
0
n
f
x
,
i
c
i
\sum_{i=0}^nf_{x,i}*a_i*\sum_{i=0}^nf_{x,i}b_i=\sum_{i=0}^nf_{x,i}c_i
∑i=0nfx,i∗ai∗∑i=0nfx,ibi=∑i=0nfx,ici
需要满足
f
x
,
k
=
∑
i
⨂
j
=
k
f
x
,
i
∗
f
x
,
j
f_{x,k}=\sum_{i\bigotimes j=k}f_{x,i}*f_{x,j}
fx,k=∑i⨂j=kfx,i∗fx,j
这样的矩阵显然太大了
每一位都是相同的,考虑构造一个
k
k
k 位的矩阵
设一个数 x 的
m
m
m 位分别是
x
0
,
x
1
,
.
.
.
,
x
m
−
1
x_0,x_1,...,x_{m-1}
x0,x1,...,xm−1
g
x
,
y
=
∏
i
x
i
y
i
g_{x,y}=\prod_i x_iy_i
gx,y=∏ixiyi,这样子矩阵的大小就是 k
考虑如果知道
g
g
g 如何求
f
∗
A
,
f
∗
B
f*A,f*B
f∗A,f∗B
设
x
^
\hat x
x^ 为 x 的最高位,
x
′
x'
x′ 为 x 除去最高位后剩下的
有
(
f
∗
A
)
[
x
]
=
∑
i
=
0
m
−
1
f
x
,
i
a
i
(f*A)[x]=\sum_{i=0}^{m-1}f_{x,i}a_i
(f∗A)[x]=∑i=0m−1fx,iai
=
∑
i
=
0
k
−
1
g
x
^
,
i
∑
j
^
=
i
f
x
′
,
j
′
a
j
=\sum_{i=0}^{k-1}g_{\hat x,i}\sum_{\hat j=i}f_{x',j'}a_j
=∑i=0k−1gx^,i∑j^=ifx′,j′aj
=
∑
i
=
0
k
−
1
f
x
^
,
i
(
f
∗
A
′
)
[
x
′
]
=\sum_{i=0}^{k-1} f_{\hat x,i}(f*A')[x']
=∑i=0k−1fx^,i(f∗A′)[x′]
A
′
A'
A′ 就是原序列把最高位是
i
i
i 的那一段截出来的序列
然后就可以通过分治求得这一个东西
以下定义卷积符号是 k 进制的不进位加法
如果要让
f
f
f 满足条件,也就是每一位都满足条件,对应到
g
g
g 上面就是
g
x
,
k
=
∑
K
∣
i
+
j
−
k
g
x
,
i
∗
g
x
,
j
g_{x,k}=\sum_{K|i+j-k}g_{x,i}*g_{x,j}
gx,k=∑K∣i+j−kgx,i∗gx,j
想到了单位根反演,并且循环有意义,那么可以按如下方式构造
g = 1 1 1 . . . 1 w K 1 w K 2 . . . 1 w K 2 w K 4 . . . . . . . . . . . . . . . 1 w K K − 1 w K 2 ∗ ( K − 1 ) . . . g=\begin{matrix} 1 & 1 & 1&... \\ 1 & w_K^1 & w_K^2 &...\\ 1 & w_K^2 & w_K^4 & ... \\ ... &...&...&...\\ 1 & w_K^{K-1} & w_K^{2*(K-1)} & ...\end{matrix} g=111...11wK1wK2...wKK−11wK2wK4...wK2∗(K−1)...............
g
−
1
=
1
K
1
1
1
.
.
.
1
w
K
−
1
w
K
−
2
.
.
.
1
w
K
−
2
w
K
−
4
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
w
K
−
(
K
−
1
)
w
K
−
2
∗
(
K
−
1
)
.
.
.
g^{-1}=\frac{1}{K}\begin{matrix} 1 & 1 & 1&... \\ 1 & w_K^{-1} & w_K^{-2} &...\\ 1 & w_K^{-2} & w_K^{-4} & ... \\ ... &...&...&...\\ 1 & w_K^{-(K-1)} & w_K^{-2*(K-1)} & ...\end{matrix}
g−1=K1111...11wK−1wK−2...wK−(K−1)1wK−2wK−4...wK−2∗(K−1)...............
【清华集训2016】石家庄的工人阶级队伍比较坚强
#include<bits/stdc++.h>
#define cs const
using namespace std;
int read(){
int cnt = 0, f = 1; char ch = 0;
while(!isdigit(ch)){ ch = getchar(); if(ch == '-') f = -1; }
while(isdigit(ch)) cnt = cnt*10 + (ch-'0'), ch = getchar();
return cnt * f;
}
cs int N = 1e6 + 5, M = 15;
int Mod, n, m, T, pw[M];
int add(int a, int b){ return a + b >= Mod ? a + b - Mod : a + b; }
int mul(int a, int b){ return 1ll * a * b % Mod; }
int dec(int a, int b){ return a - b < 0 ? a - b + Mod : a - b; }
int ksm(int a, int b){ int ans = 1; for(;b;b>>=1,a=mul(a,a)) if(b&1) ans=mul(ans,a); return ans; }
void exgcd(int a, int b, int &x, int &y){
if(!b){ x = 1; y = 0; return; } exgcd(b,a-a/b*b,x,y); int t=x; x=y;y=t-a/b*y;
}
int inv(int a){ int x,y,b=Mod; exgcd(a,b,x,y); return (x%b+b)%b; }
struct data{
int x, y;
data(int _x = 0, int _y = 0){ x = _x, y = _y; }
data operator + (cs data &a){ return data(add(x,a.x),add(y,a.y)); }
data operator - (cs data &a){ return data(dec(x,a.x),dec(y,a.y)); }
data operator * (cs data &a){ return data(dec(mul(x,a.x),mul(y,a.y)), dec(add(mul(x,a.y), mul(y,a.x)), mul(y,a.y))); }
bool operator < (cs data &a)cs{ return x^a.x ? x<a.x : y<a.y;}
}w[3], f[N], g[N];
data power(data a, int b){ data ans(1,0); for(;b;b>>=1,a=a*a) if(b&1) ans=ans*a; return ans; }
int b[M][M];
void dfs(int u, int vl, int ct1, int ct2){
if(u == m){ g[vl].x = b[ct1][ct2]; return; }
dfs(u+1, vl, ct1, ct2);
dfs(u+1, vl+pw[u], ct1+1, ct2);
dfs(u+1, vl+pw[u]+pw[u], ct1, ct2+1);
}
void dft(data *a, int typ){
for(int i = 1; i < n; i *= 3){
for(int j = 0, len = i * 3; j < n; j += len){
for(int k = 0; k < i; k++){
data a0 = a[j + k], a1 = a[j + k + i], a2 = a[j + k + i + i];
a[j + k] = a0 + a1 + a2;
a[j + k + i] = a0 + a1 * w[1] + a2 * w[2];
a[j + k + i + i] = a0 + a1 * w[2] + a2 * w[1];
if(typ == -1) swap(a[j + k + i], a[j + k + i + i]);
}
}
}
}
map<data, data> S;
data calc(data x){ if(S.count(x)) return S[x];return S[x] = power(x, T); }
int main(){
scanf("%d%d%d", &m, &T, &Mod);
pw[0] = 1; for(int i = 1; i <= m; i++) pw[i] = pw[i-1] * 3;
w[0] = data(1, 0); w[1] = data(0, 1); w[2] = data(Mod-1,Mod-1);
n = pw[m];
for(int i = 0; i < n; i++) f[i].x = read();
for(int i = 0; i <= m; i++) for(int j = 0; j <= m-i; j++) b[i][j] = read();
dfs(0, 0, 0, 0);
dft(f, 1); dft(g, 1);
for(int i = 0; i < n; i++) f[i] = f[i] * calc(g[i]);
dft(f, -1); int iv = inv(n);
for(int i = 0; i < n; i++) cout << mul(iv, f[i].x) << '\n';
return 0;
}