[SDOI2013] 随机数生成器
description
solution
肯定是非常想找一个通项公式来表示第
n
n
n个数的
依据形式，考虑化成等比数列
x i + 1 + k = a ( x i + k ) = a ⋅ x i + b + t ⇒ k = b a − 1 x_{i+1}+k=a(x_i+k)=a·x_i+b+t\Rightarrow k=\frac{b}{a-1} xi+1+k=a(xi+k)=a⋅xi+b+t⇒k=a−1b
⇒ x i + b a − 1 = a i − 1 ( x 1 + b a − 1 ) ( m o d p ) ⇒ a i − 1 ≡ ( x i + b ⋅ ( a − 1 ) − 1 ) ⋅ ( x 1 + b ⋅ ( a − 1 ) − 1 ) − 1 ( m o d p ) \Rightarrow x_i+\frac{b}{a-1}=a^{i-1}(x_1+\frac{b}{a-1})\pmod p\Rightarrow a^{i-1}\equiv (x_i+b·(a-1)^{-1})·(x_1+b·(a-1)^{-1})^{-1}\pmod p ⇒xi+a−1b=ai−1(x1+a−1b)(modp)⇒ai−1≡(xi+b⋅(a−1)−1)⋅(x1+b⋅(a−1)−1)−1(modp)
x
i
x_i
xi就是题目中的
t
t
t，所以式子只有
a
i
−
1
a^{i-1}
ai−1这个指数未知，bsgs求解最小指数（
p
p
p是质数）
但要先特判 a = 0 / 1 a=0/1 a=0/1两种情况，显然此情况下等比数列式子是不成立的
-
a=0x i = b x_i=b xi=b
-
a=1x i ≡ x 1 + b ( i − 1 ) ( m o d p ) ⇔ ( t − x 1 ) b − 1 ≡ i − 1 ( m o d p ) x_i\equiv x_1+b(i-1)\pmod p\Leftrightarrow (t-x_1)b^{-1}\equiv i-1\pmod p xi≡x1+b(i−1)(modp)⇔(t−x1)b−1≡i−1(modp)
code
#include <map>
#include <cmath>
#include <cstdio>
using namespace std;
#define int long long
map < int, int > mp;
int bsgs( int a, int r, int p ) {
if( a % p == 0 ) return -1;
mp.clear();
int m = ceil( sqrt( p * 1.0 ) );
int IS = 1;
for( int i = 1;i <= m;i ++ ) {
IS = IS * a % p;
mp[IS * r % p] = i;
}
int SI = 1;
for( int i = 1;i <= m;i ++ ) {
SI = SI * IS % p;
if( mp.count( SI ) )
return i * m - mp[SI] + 1;
}
return -1;
}
int T, p, a, b, x1, t;
int qkpow( int x, int y ) {
int ans = 1;
while( y ) {
if( y & 1 ) ans = ans * x % p;
x = x * x % p;
y >>= 1;
}
return ans;
}
int inv( int x ) {
return qkpow( x, p - 2 );
}
signed main() {
scanf( "%lld", &T );
while( T -- ) {
scanf( "%lld %lld %lld %lld %lld", &p, &a, &b, &x1, &t );
if( x1 == t ) {
printf( "1\n" );
continue;
}
if( a == 0 ) {
if( x1 == t ) printf( "1\n" );
else if( b == t ) printf( "2\n" );
else printf( "-1\n" );
continue;
}
if( a == 1 ) {
if( b == 0 ) printf( "-1\n" );
else printf( "%lld\n", ( ( t - x1 ) % p + p ) % p * inv( b ) % p + 1 );
continue;
}
printf( "%lld\n", bsgs( a, ( t + b * inv( a - 1 ) % p ) * inv( x1 + b * inv( a - 1 ) % p ) % p, p ) );
}
return 0;
}
[BZOJ4128]Matrix
description
solution
因为保证了 A A A矩阵有逆，所以可以选择求矩阵的逆
但其实还是选择两边同乘矩阵要好一些
BSGS模板，有优质代码是写了非常迅速的hash
但是偶直接map里面塞矩阵，重载一下map的小于等于比较挺不戳的！
code
#include <map>
#include <cmath>
#include <cstdio>
#include <cstring>
using namespace std;
int n, mod;
struct matrix {
int c[71][71];
void clear() {
memset( c, 0, sizeof( c ) );
}
matrix() {
clear();
}
void init() {
for( int i = 1;i <= n;i ++ ) {
for( int j = 1;j <= n;j ++ )
c[i][j] = 0;
c[i][i] = 1;
}
}
matrix operator * ( const matrix &t ) const {
matrix ans;
for( int i = 1;i <= n;i ++ )
for( int j = 1;j <= n;j ++ )
for( int k = 1;k <= n;k ++ )
ans.c[i][j] = ( ans.c[i][j] + c[i][k] * t.c[k][j] ) % mod;
return ans;
}
friend bool operator == ( const matrix &s, const matrix &t ) {
for( int i = 1;i <= n;i ++ )
for( int j = 1;j <= n;j ++ )
if( s.c[i][j] ^ t.c[i][j] ) return 0;
return 1;
}
friend bool operator < ( const matrix &s, const matrix &t ) {
for( int i = 1;i <= n;i ++ )
for( int j = 1;j <= n;j ++ )
if( s.c[i][j] ^ t.c[i][j] ) return s.c[i][j] < t.c[i][j];
return 0;
}
}A, B, g, h;
map < matrix, int > mp;
int main() {
scanf( "%d %d", &n, &mod );
for( int i = 1;i <= n;i ++ )
for( int j = 1;j <= n;j ++ )
scanf( "%d", &A.c[i][j] );
for( int i = 1;i <= n;i ++ )
for( int j = 1;j <= n;j ++ )
scanf( "%d", &B.c[i][j] );
int m = ceil( sqrt( mod * 1.0 ) );
g.init();
for( int i = 1;i <= m;i ++ ) {
g = g * A;
h = g * B;
mp[h] = i;
}
h.init();
for( int i = 1;i <= m;i ++ ) {
h = h * g;
if( mp.count( h ) ) return ! printf( "%d\n", i * m - mp[h] );
}
return 0;
}
CF1106F Lunar New Year and a Recursive Sequence
problem
solution
f i = ∏ j = 1 k f i − j b j ( m o d p ) f_{i}=\prod_{j=1}^kf_{i-j}^{b_j}\pmod p fi=∏j=1kfi−jbj(modp)
p p p是非常特殊的质数，其原根为 3 3 3，可以用原根改写上式，令 3 g i = f i 3^{g_i}=f_i 3gi=fi
g i = ∑ j = 1 k g i − j ⋅ b j ( m o d φ ( p ) ) g_i=\sum_{j=1}^kg_{i-j}·b_j\pmod {\varphi(p)} gi=∑j=1kgi−j⋅bj(modφ(p))
此时的
g
g
g可以用矩阵加速，初始化
f
i
=
0
(
i
<
k
)
,
f
k
=
1
f_i=0(i<k),f_k=1
fi=0(i<k),fk=1，转移
n
−
k
n-k
n−k次变成
[
f
n
−
k
+
1
,
.
.
.
,
f
n
]
[f_{n-k+1},...,f_n]
[fn−k+1,...,fn]
[
h
i
−
k
+
1
,
.
.
.
,
h
i
⏟
k
]
×
[
0
0
.
.
.
0
b
k
1
0
.
.
.
0
b
k
−
1
0
1
.
.
.
0
b
k
−
2
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
0
0
.
.
.
0
b
2
0
0
.
.
.
1
b
1
]
=
[
h
i
+
k
+
2
,
.
.
.
,
h
i
+
1
⏟
k
]
\begin{bmatrix} \underbrace{h_{i-k+1},...,h_i}_{k} \end{bmatrix}\times \begin{bmatrix} 0&0&...&0&b_k\\ 1&0&...&0&b_{k-1}\\ 0&1&...&0&b_{k-2}\\ ...&...&...&...&...\\ 0&0&...&0&b_2\\ 0&0&...&1&b_1 \end{bmatrix}= \begin{bmatrix} \underbrace{h_{i+k+2},...,h_{i+1}}_{k} \end{bmatrix}
[k
hi−k+1,...,hi]×⎣⎢⎢⎢⎢⎢⎢⎡010...00001...00..................000...01bkbk−1bk−2...b2b1⎦⎥⎥⎥⎥⎥⎥⎤=[k
hi+k+2,...,hi+1]
同时，有
f
k
t
≡
f
n
(
m
o
d
p
)
,
3
g
k
=
f
k
f_k^t\equiv f_n\pmod p,3^{g_k}=f_k
fkt≡fn(modp),3gk=fk，这里的
t
t
t矩等于加速矩阵自乘后的
h
k
h_k
hk
f k T = 3 g k ⋅ T ≡ f n = 3 g n ( m o d p ) f_k^T=3^{g_k·T}\equiv f_n=3^{g_n}\pmod p fkT=3gk⋅T≡fn=3gn(modp)
对于
3
g
n
≡
f
n
(
m
o
d
p
)
3^{g_n}\equiv f_n\pmod p
3gn≡fn(modp)的
g
n
g_n
gn未知，用BSGS求得
对于指数用exgcd计算，
g
k
×
t
≡
g
n
(
m
o
d
p
−
1
)
⇔
t
×
g
k
−
(
p
−
1
)
×
y
=
g
n
g_k\times t\equiv g_n\pmod {p-1}\Leftrightarrow t\times g_k-(p-1)\times y=g_n
gk×t≡gn(modp−1)⇔t×gk−(p−1)×y=gn
未知数为 g k , y g_k,y gk,y，求出后 f k = 3 g k f_k=3^{g_k} fk=3gk
code
#include <map>
#include <cmath>
#include <cstdio>
#include <cstring>
using namespace std;
#define int long long
#define mod 998244353
struct matrix {
int n, m;
int c[101][101];
matrix() {
memset( c, 0, sizeof( c ) );
}
matrix operator * ( matrix &t ) const {
matrix ans;
ans.n = n, ans.m = t.m;
for( int i = 1;i <= n;i ++ )
for( int j = 1;j <= t.m;j ++ )
for( int k = 1;k <= m;k ++ )
ans.c[i][j] = ( ans.c[i][j] + c[i][k] * t.c[k][j] ) % ( mod - 1 );
return ans;
}
}h, t;
matrix qkpow( matrix x, int y ) {
matrix ans;
ans.n = ans.m = x.n;
for( int i = 1;i <= ans.n;i ++ )
ans.c[i][i] = 1;
while( y ) {
if( y & 1 ) ans = ans * x;
x = x * x;
y >>= 1;
}
return ans;
}
map < int, int > mp;
int bsgs( int a, int r ) {
int m = ceil( sqrt( mod * 1.0 ) );
int IS = 1;
for( int i = 1;i <= m;i ++ ) {
IS = IS * a % mod;
mp[IS * r % mod] = i;
}
int MS = 1;
for( int i = 1;i <= m;i ++ ) {
MS = MS * IS % mod;
if( mp.count( MS ) )
return i * m - mp[MS];
}
return -1;
}
int exgcd( int a, int b, int &x, int &y ) {
if( ! b ) {
x = 1, y = 0;
return a;
}
else {
int d = exgcd( b, a % b, y, x );
y -= a / b * x;
return d;
}
}
int qkpow( int x, int y ) {
int ans = 1;
while( y ) {
if( y & 1 ) ans = ans * x % mod;
x = x * x % mod;
y >>= 1;
}
return ans;
}
signed main() {
int k, n, fn;
scanf( "%lld", &k );
h.n = h.m = t.m = k, t.n = t.c[1][k] = 1;
for( int i = 2;i <= k;i ++ )
h.c[i][i - 1] = 1;
for( int i = k;i;i -- )
scanf( "%lld", &h.c[i][k] );
scanf( "%lld %lld", &n, &fn );
h = qkpow( h, n - k );
t = t * h;
int g = bsgs( 3, fn );
int x, y, d = exgcd( t.c[1][k], mod - 1, x, y );
if( g % d ) return ! printf( "-1\n" );
else {
x = ( x * ( g / d ) % ( mod - 1 ) + ( mod - 1 ) ) % ( mod - 1 );
printf( "%lld\n", qkpow( 3, x ) );
}
return 0;
}
Fermat’s Last Theorem
description
solution
找出
p
p
p的原根
g
g
g，令
x
=
g
a
,
y
=
g
b
,
z
=
g
c
x=g^a,y=g^b,z=g^c
x=ga,y=gb,z=gc
则
g
a
n
+
g
b
n
≡
g
c
n
(
m
o
d
p
)
g^{an}+g^{bn}\equiv g^{cn}\pmod p
gan+gbn≡gcn(modp)
令
r
=
g
n
⇒
r
a
+
r
b
≡
r
c
(
m
o
d
p
)
r=g^n\Rightarrow r^a+r^b\equiv r^c\pmod p
r=gn⇒ra+rb≡rc(modp)
假设
a
≤
b
a\le b
a≤b
- a ≤ c ⇒ 1 + r b − a ≡ r c − a ( m o d p ) a\le c\Rightarrow 1+r^{b-a}\equiv r^{c-a}\pmod p a≤c⇒1+rb−a≡rc−a(modp)
- a > c ⇒ r a − c + r b − a ≡ 1 ( m o d p ) a>c\Rightarrow r^{a-c}+r^{b-a}\equiv 1\pmod p a>c⇒ra−c+rb−a≡1(modp)
枚举
r
k
r^k
rk，出现循环的时候就break
出现
1
+
r
k
1
≡
r
k
2
/
r
k
1
+
r
k
2
≡
1
(
m
o
d
p
)
1+r^{k_1}\equiv r^{k_2}/r^{k_1}+r^{k_2}\equiv 1\pmod p
1+rk1≡rk2/rk1+rk2≡1(modp)就是有解了
code
#include <cstdio>
#define int long long
#define maxn 1000005
int d[maxn], vis[maxn], mi[maxn];
int qkpow( int x, int y, int mod ) {
int ans = 1;
while( y ) {
if( y & 1 ) ans = ans * x % mod;
x = x * x % mod;
y >>= 1;
}
return ans;
}
int root( int p ) {
int cnt = 0, phi = p - 1;
for( int i = 2;i * i <= phi;i ++ )
if( phi % i == 0 ) {
d[++ cnt] = i;
if( i * i != phi ) d[++ cnt] = phi / i;
}
for( int g = 1, i;;g ++ ) {
for( i = 1;i <= cnt;i ++ )
if( qkpow( g, d[i], p ) == 1 )
break;
if( i > cnt ) return g;
}
}
signed main() {
int T, n, p;
scanf( "%lld", &T );
for( int Case = 1;Case <= T;Case ++ ) {
scanf( "%lld %lld", &n, &p );
int g = root( p ), r = qkpow( g, n, p );
int t = 1, x = -1, y, z;
for( int i = 0;i < p;i ++ ) {
if( i ) t = t * r % p;
if( vis[t] == Case ) break;
vis[t] = Case, mi[t] = i;
int MS = 1 + t;
if( MS >= p ) MS = 0;
int IS = 1 - t + p;
if( IS >= p ) IS = 0;
if( vis[MS] == Case ) {
x = 1, y = qkpow( g, i, p ), z = qkpow( g, mi[MS], p );
break;
}
if( vis[IS] == Case ) {
x = qkpow( g, i, p ), y = qkpow( g, mi[IS], p ), z = 1;
break;
}
}
if( ~ x ) printf( "%lld %lld %lld\n", x, y, z );
else printf( "-1\n" );
}
return 0;
}
2275
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
