BZOJ5306 [HAOI2018]染色

Address

• BZOJ5306 洛谷P5306

Solution

• 暴力推式子。
• 枚举出现 $s$ 次的颜色种数，则剩余颜色出现次数不能为 $s$ 次，由容斥原理得： a n s = ∑ i = 0 min ⁡ { ⌊ n s ⌋ , m } C m i n ! W i ( s ! ) i ( n − i s ) ! ∑ j = 0 min ⁡ { ⌊ n s ⌋ , m } − i ( − 1 ) j C m − i j ( n − i s ) ! ( m − i − j ) n − i s − j s ( s ! ) j ( n − i s − j s ) ! ans = \sum \limits_{i = 0}^{\min\{\lfloor \frac{n}{s} \rfloor, m\}} C_{m}^{i} \frac{n!W_i}{(s!)^i (n - is)!} \sum \limits_{j = 0}^{\min\{\lfloor \frac{n}{s} \rfloor, m\} - i} (-1)^j C_{m - i}^j \frac{(n - is)!(m - i - j)^{n - is - js}}{(s!)^j(n - is - js)!} ans=i=0∑min{⌊sn​⌋,m}​Cmi​(s!)i(n−is)!n!Wi​​j=0∑min{⌊sn​⌋,m}−i​(−1)jCm−ij​(s!)j(n−is−js)!(n−is)!(m−i−j)n−is−js​
• 展开式子，把无关项前移： a n s = n ! m ! ∑ i = 0 min ⁡ { ⌊ n s ⌋ , m } W i i ! ( s ! ) i ∑ j = 0 min ⁡ { ⌊ n s ⌋ , m } − i ( − 1 ) j j ! ( s ! ) j ( m − i − j ) n − i s − j s ( m − i − j ) ! ( n − i s − j s ) ! ans = n!m!\sum \limits_{i = 0}^{\min\{\lfloor \frac{n}{s} \rfloor, m\}} \frac{W_i}{i!(s!)^i} \sum \limits_{j = 0}^{\min\{\lfloor \frac{n}{s} \rfloor, m\} - i} \frac{(-1)^j}{j!(s!)^j} \frac{(m - i - j)^{n - is - js}}{(m - i - j)!(n - is - js)!} ans=n!m!i=0∑min{⌊sn​⌋,m}​i!(s!)iWi​​j=0∑min{⌊sn​⌋,m}−i​j!(s!)j(−1)j​(m−i−j)!(n−is−js)!(m−i−j)n−is−js​
• 好像没什么头绪，尝试把枚举 $j$ 改为枚举 $i+j$：
  a n s = n ! m ! ∑ i = 0 min ⁡ { ⌊ n s ⌋ , m } W i i ! ( s ! ) i ∑ j = i min ⁡ { ⌊ n s ⌋ , m } ( − 1 ) j − i ( j − i ) ! ( s ! ) j − i ( m − j ) n − j s ( m − j ) ! ( n − j s ) ! ans = n!m!\sum \limits_{i = 0}^{\min\{\lfloor \frac{n}{s} \rfloor, m\}} \frac{W_i}{i!(s!)^i} \sum \limits_{j = i}^{\min\{\lfloor \frac{n}{s} \rfloor, m\}} \frac{(-1)^{j - i}}{(j - i)!(s!)^{j - i}} \frac{(m - j)^{n - js}}{(m - j)!(n - js)!} ans=n!m!i=0∑min{⌊sn​⌋,m}​i!(s!)iWi​​j=i∑min{⌊sn​⌋,m}​(j−i)!(s!)j−i(−1)j−i​(m−j)!(n−js)!(m−j)n−js​
• 已经简单很多了，再改变下 $i,j$ 的枚举顺序：
  a n s = n ! m ! ∑ j = 0 min ⁡ { ⌊ n s ⌋ , m } ( m − j ) n − j s ( m − j ) ! ( n − j s ) ! ( s ! ) j ∑ i = 0 j W i i ! ( − 1 ) j − i ( j − i ) ! ans = n!m!\sum \limits_{j = 0}^{\min\{\lfloor \frac{n}{s} \rfloor, m\}} \frac{(m - j)^{n - js}}{(m - j)!(n - js)!(s!)^j} \sum \limits_{i = 0}^{j} \frac{W_i}{i!} \frac{(-1)^{j - i}}{(j - i)!} ans=n!m!j=0∑min{⌊sn​⌋,m}​(m−j)!(n−js)!(s!)j(m−j)n−js​i=0∑j​i!Wi​​(j−i)!(−1)j−i​
• 后面是个卷积的形式，NTT 优化即可。
• 时间复杂度 $O(m\log m+n)$。

Code

• 注意阶乘要预处理到 max ⁡ { n , m } \max\{n, m\} max{n,m}。

#include <algorithm>
#include <iostream>
#include <cstring>
#include <cstdlib>
#include <cctype>
#include <cstdio>
#include <cmath>
#include <ctime>

template <class T>
inline void read(T &res)
{
	char ch;
	while (ch = getchar(), !isdigit(ch));
	res = ch ^ 48;
	while (ch = getchar(), isdigit(ch))
		res = res * 10 + ch - 48;
}

const int N = 1e7 + 5;
const int M = 1e5 + 5;
const int M4 = 4e5 + 5;
const int mod = 1004535809;
const int inv3 = (mod + 1) / 3;
int fra[N], inv[N], w[M], f[M4], g[M4], rev[M4]; 
int n, m, s, tm, ans;

inline void add(int &x, int y)
{
	x += y;
	x >= mod ? x -= mod : 0;
}

inline int quick_pow(int x, int k)
{
	int res = 1;
	while (k)
	{
		if (k & 1) res = 1ll * res * x % mod;
		x = 1ll * x * x % mod; k >>= 1;
	}
	return res;
}

inline void NTT(int *f, int fm, int opt)
{
	int g = opt == 1 ? 3 : inv3;
	for (int i = 0; i < fm; ++i)
		if (i < rev[i])
			std::swap(f[i], f[rev[i]]);
	for (int k = 1; k < fm; k <<= 1)
	{
		int w = quick_pow(g, (mod - 1) / (k << 1));
		for (int i = 0; i < fm; i += k << 1)
		{
			int res = 1;
			for (int j = 0; j < k; ++j)
			{
				int u = f[i + j],
					v = 1ll * res * f[i + j + k] % mod;
				f[i + j] = f[i + j + k] = u;
				add(f[i + j], v);
				add(f[i + j + k], mod - v);
				res = 1ll * res * w % mod;
			}
		}
	}
	if (opt == -1)
	{
		for (int i = 0, inv = quick_pow(fm, mod - 2); i < fm; ++i)
			f[i] = 1ll * f[i] * inv % mod;
	}
}

template <class T>
inline T Min(T x, T y) {return x < y ? x : y;}
template <class T>
inline T Max(T x, T y) {return x > y ? x : y;}

int main()
{
	read(n); read(m); read(s);
	for (int i = 0; i <= m; ++i)
		read(w[i]);
	
	tm = Max(n, m);
	fra[0] = 1;
	for (int i = 1; i <= tm; ++i)
		fra[i] = 1ll * fra[i - 1] * i % mod;
	inv[tm] = quick_pow(fra[tm], mod - 2);
	for (int i = tm; i >= 1; --i)
		inv[i - 1] = 1ll * inv[i] * i % mod;

	tm = Min(n / s, m);
	for (int i = 0; i <= tm; ++i)
	{
		g[i] = 1ll * w[i] * inv[i] % mod;
		f[i] = inv[i];
		if (i & 1) f[i] = mod - f[i];
	}

	int fm, k = 0; tm <<= 1;
	for (fm = 1; fm <= tm; fm <<= 1)
		++k; --k;
	for (int i = 1; i < fm; ++i)
		rev[i] = (rev[i >> 1] >> 1) | ((i & 1) << k);
	NTT(f, fm, 1); NTT(g, fm, 1);
	for (int i = 0; i < fm; ++i)
		f[i] = 1ll * f[i] * g[i] % mod;
	NTT(f, fm, -1);

	int mul_s = 1; tm >>= 1;
	for (int i = 0; i <= tm; ++i)
	{
		int tx = m - i, ty = n - i * s;
		ans = (1ll * quick_pow(tx, ty) * inv[tx] % mod * inv[ty] % mod * mul_s % mod * f[i] + ans) % mod;
		mul_s = 1ll * mul_s * inv[s] % mod;
	}
	std::cout << 1ll * fra[n] * fra[m] % mod * ans % mod << std::endl;
}