一般情况下如果
f
(
n
)
=
∑
i
=
0
n
a
n
,
i
g
(
i
)
⟺
g
(
n
)
=
∑
i
=
0
n
b
n
,
i
f
(
i
)
f(n)=\sum_{i=0}^{n}a_{n,i}g(i)\Longleftrightarrow g(n)=\sum_{i=0}^{n}b_{n,i}f(i)
f(n)=i=0∑nan,ig(i)⟺g(n)=i=0∑nbn,if(i)
成立，则
f
(
n
)
=
∑
i
=
0
n
a
n
,
i
g
(
i
+
k
)
⟺
g
(
n
+
k
)
=
∑
i
=
0
n
b
n
,
i
f
(
i
)
f(n)=\sum_{i=0}^{n}a_{n,i}g(i+k)\Longleftrightarrow g(n+k)=\sum_{i=0}^{n}b_{n,i}f(i)
f(n)=i=0∑nan,ig(i+k)⟺g(n+k)=i=0∑nbn,if(i)
也成立
二项式反演
∑
i
=
j
n
(
−
1
)
i
−
j
(
n
i
)
(
i
j
)
=
[
n
−
j
=
0
]
①
\sum_{i=j}^{n}(-1)^{i-j}{n\choose i}{i\choose j}=[n-j=0]\quad ①
i=j∑n(−1)i−j(in)(ji)=[n−j=0]①
证明
∑
i
=
j
n
(
−
1
)
i
−
j
(
n
i
)
(
i
j
)
=
∑
i
=
j
n
(
−
1
)
i
−
j
(
n
j
)
(
n
−
j
i
−
j
)
=
(
n
j
)
∑
i
=
j
n
(
−
1
)
i
−
j
(
n
−
j
i
−
j
)
=
(
n
j
)
∑
i
=
0
n
−
j
(
−
1
)
i
(
n
−
j
i
)
=
(
n
j
)
(
1
−
1
)
n
−
j
=
(
n
j
)
[
n
−
j
=
0
]
=
[
n
−
j
=
0
]
\begin{aligned} \sum_{i=j}^{n}(-1)^{i-j}{n\choose i}{i\choose j}&=\sum_{i=j}^{n}(-1)^{i-j}{n\choose j}{n-j\choose i-j}\\ &={n\choose j}\sum_{i=j}^{n}(-1)^{i-j}{n-j\choose i-j}\\ &={n\choose j}\sum_{i=0}^{n-j}(-1)^i{n-j\choose i}\\ &={n\choose j}(1-1)^{n-j}\\ &={n\choose j}[n-j=0]\\ &=[n-j=0] \end{aligned}
i=j∑n(−1)i−j(in)(ji)=i=j∑n(−1)i−j(jn)(i−jn−j)=(jn)i=j∑n(−1)i−j(i−jn−j)=(jn)i=0∑n−j(−1)i(in−j)=(jn)(1−1)n−j=(jn)[n−j=0]=[n−j=0]
f
(
n
)
=
∑
i
=
0
n
(
n
i
)
g
(
i
)
⟺
g
(
n
)
=
∑
i
=
0
n
(
−
1
)
n
−
i
(
n
i
)
f
(
i
)
\begin{aligned} f(n)=\sum_{i=0}^{n}{n \choose i}g(i)&\Longleftrightarrow g(n)=\sum_{i=0}^{n}(-1)^{n-i}{n\choose i}f(i)\\ \end{aligned}
f(n)=i=0∑n(in)g(i)⟺g(n)=i=0∑n(−1)n−i(in)f(i)
证明
f
(
n
)
=
∑
i
=
0
n
(
n
i
)
g
(
i
+
k
)
=
∑
i
=
0
n
∑
j
=
0
i
(
−
1
)
i
−
j
(
n
i
)
(
i
j
)
f
(
j
)
=
∑
j
=
0
n
f
(
j
)
∑
i
=
j
n
(
−
1
)
i
−
j
(
n
i
)
(
i
j
)
=
∑
j
=
0
n
f
(
j
)
[
n
−
j
=
0
]
根
据
公
式
①
=
f
(
n
)
\begin{aligned} f(n)&=\sum_{i=0}^{n}{n \choose i}g(i+k)\\ &=\sum_{i=0}^{n}\sum_{j=0}^{i}(-1)^{i-j}{n\choose i}{i\choose j}f(j)\\ &=\sum_{j=0}^{n}f(j)\sum_{i=j}^{n}(-1)^{i-j}{n\choose i}{i\choose j}\\ &=\sum_{j=0}^{n}f(j)[n-j=0]\quad 根据公式①\\ &=f(n) \end{aligned}
f(n)=i=0∑n(in)g(i+k)=i=0∑nj=0∑i(−1)i−j(in)(ji)f(j)=j=0∑nf(j)i=j∑n(−1)i−j(in)(ji)=j=0∑nf(j)[n−j=0]根据公式①=f(n)
子集反演
∑
T
⊆
S
(
−
1
)
∣
T
∣
=
[
S
=
∅
]
②
\sum_{T\subseteq S}(-1)^{|T|}=[S=\emptyset]\quad ②
T⊆S∑(−1)∣T∣=[S=∅]②
证明
枚举集合
T
T
T可以变成枚举集合
T
T
T的元素个数
i
i
i，然后这样的集合有
(
∣
S
∣
i
)
{|S|\choose i}
(i∣S∣)种
∑
T
⊆
S
(
−
1
)
∣
T
∣
=
∑
i
=
0
∣
S
∣
(
−
1
)
i
(
∣
S
∣
i
)
=
[
∣
S
∣
=
0
]
=
[
S
=
∅
]
\begin{aligned} \sum_{T\subseteq S}(-1)^{|T|}&=\sum_{i=0}^{|S|}(-1)^{i}{|S|\choose i}\\ &=[|S|=0]\\ &=[S=\emptyset] \end{aligned}
T⊆S∑(−1)∣T∣=i=0∑∣S∣(−1)i(i∣S∣)=[∣S∣=0]=[S=∅]
f
(
S
)
=
∑
T
⊆
S
g
(
T
)
⟺
g
(
S
)
=
∑
T
⊆
S
(
−
1
)
∣
S
∣
−
∣
T
∣
f
(
T
)
f(S)=\sum_{T\subseteq S}g(T)\Longleftrightarrow g(S)=\sum_{T\subseteq S}(-1)^{|S|-|T|}f(T)
f(S)=T⊆S∑g(T)⟺g(S)=T⊆S∑(−1)∣S∣−∣T∣f(T)
证明
f
(
S
)
=
∑
T
⊆
S
g
(
T
)
=
∑
T
⊆
S
∑
Q
⊆
T
(
−
1
)
∣
T
∣
−
∣
Q
∣
f
(
Q
)
=
∑
Q
⊆
S
f
(
Q
)
∑
Q
⊆
T
⊆
S
(
−
1
)
∣
T
∣
−
∣
Q
∣
=
∑
Q
⊆
S
f
(
Q
)
∑
∅
⊆
T
⊆
S
−
Q
(
−
1
)
∣
T
∣
=
∑
Q
⊆
S
f
(
Q
)
[
S
−
Q
=
∅
]
根
据
公
式
②
=
f
(
S
)
\begin{aligned} f(S)&=\sum_{T\subseteq S}g(T)\\ &=\sum_{T\subseteq S}\sum_{Q\subseteq T}(-1)^{|T|-|Q|}f(Q)\\ &=\sum_{Q\subseteq S}f(Q)\sum_{Q\subseteq T\subseteq S}(-1)^{|T|-|Q|}\\ &=\sum_{Q\subseteq S}f(Q)\sum_{\emptyset\subseteq T\subseteq S-Q}(-1)^{|T|}\\ &=\sum_{Q\subseteq S}f(Q)[S-Q=\emptyset]\quad 根据公式②\\ &=f(S) \end{aligned}
f(S)=T⊆S∑g(T)=T⊆S∑Q⊆T∑(−1)∣T∣−∣Q∣f(Q)=Q⊆S∑f(Q)Q⊆T⊆S∑(−1)∣T∣−∣Q∣=Q⊆S∑f(Q)∅⊆T⊆S−Q∑(−1)∣T∣=Q⊆S∑f(Q)[S−Q=∅]根据公式②=f(S)
多重子集反演
现在考虑的是多重集合
φ
(
S
)
=
{
0
S
是
多
重
集
合
(
−
1
)
∣
S
∣
S
不
是
多
重
集
合
\varphi(S)=\begin{cases} 0\quad S是多重集合\\ (-1)^{|S|}\quad S不是多重集合 \end{cases}
φ(S)={0S是多重集合(−1)∣S∣S不是多重集合
∑
T
⊆
S
φ
(
T
)
=
[
S
=
∅
]
\sum_{T\subseteq S}\varphi(T)=[S=\emptyset]
T⊆S∑φ(T)=[S=∅]
证明
T
T
T如果是多重集合，则不会对式子有贡献，所以只要考虑不是多重集合，那么证明和上面一样，枚举集合
T
T
T可以变成枚举集合的元素个数
i
i
i，然后这样的集合有
(
∣
S
∣
i
)
{|S|\choose i}
(i∣S∣)种，
∣
S
∣
|S|
∣S∣的含义变成
S
S
S去重后的大小，后面就和上面一样。
反演式子也和上面类似。
f
(
S
)
=
∑
T
⊆
S
g
(
T
)
⟺
g
(
S
)
=
∑
T
⊆
S
φ
(
S
−
T
)
f
(
T
)
f(S)=\sum_{T\subseteq S}g(T)\Longleftrightarrow g(S)=\sum_{T\subseteq S}\varphi(S-T)f(T)
f(S)=T⊆S∑g(T)⟺g(S)=T⊆S∑φ(S−T)f(T)
反方向子集反演
f
(
S
)
=
∑
S
⊆
T
g
(
T
)
⟺
g
(
S
)
=
∑
S
⊆
T
(
−
1
)
∣
T
∣
−
∣
S
∣
f
(
T
)
f(S)=\sum_{S\subseteq T}g(T)\Longleftrightarrow g(S)=\sum_{S\subseteq T}(-1)^{|T|-|S|}f(T)
f(S)=S⊆T∑g(T)⟺g(S)=S⊆T∑(−1)∣T∣−∣S∣f(T)
证明和上面类似
单位根反演
1
n
∑
i
=
0
n
−
1
ω
n
m
i
=
[
m
m
o
d
n
=
0
]
③
\frac{1}{n}\sum_{i=0}^{n-1}\omega^{mi}_n=[m\ mod\,n=0]\quad ③
n1i=0∑n−1ωnmi=[m modn=0]③
证明
f
(
m
)
=
∑
i
=
0
n
−
1
ω
n
m
i
g
(
i
)
⟺
g
(
m
)
=
1
n
∑
i
=
0
n
−
1
ω
n
−
m
i
f
(
i
)
0
≤
m
≤
n
−
1
\begin{aligned} f(m)=\sum_{i=0}^{n-1}\omega_n^{mi}g(i)\Longleftrightarrow g(m)=\frac{1}{n}\sum_{i=0}^{n-1}\omega_n^{-mi}f(i)\quad 0\le&m\le n-1\\ \end{aligned}
f(m)=i=0∑n−1ωnmig(i)⟺g(m)=n1i=0∑n−1ωn−mif(i)0≤m≤n−1
证明
f
(
m
)
=
∑
i
=
0
n
−
1
ω
n
m
i
g
(
i
)
0
≤
m
≤
n
−
1
=
∑
i
=
0
n
−
1
ω
n
m
i
1
n
∑
j
=
0
n
−
1
ω
n
−
i
j
f
(
j
)
=
∑
j
=
0
n
−
1
f
(
j
)
1
n
∑
i
=
0
n
−
1
ω
n
i
(
m
−
j
)
根
据
公
式
③
=
∑
j
=
0
n
−
1
f
(
j
)
[
m
−
j
m
o
d
n
=
0
]
=
∑
j
=
0
n
−
1
f
(
j
)
[
m
−
j
=
0
]
=
f
(
m
)
\begin{aligned} f(m)&=\sum_{i=0}^{n-1}\omega_n^{mi}g(i)\quad 0\le m\le n-1\\ &=\sum_{i=0}^{n-1}\omega_n^{mi}\frac{1}{n}\sum_{j=0}^{n-1}\omega_n^{-ij}f(j)\\ &=\sum_{j=0}^{n-1}f(j)\frac{1}{n}\sum_{i=0}^{n-1}\omega_n^{i(m-j)}\quad 根据公式③\\ &=\sum_{j=0}^{n-1}f(j)[m-j\ mod\,n=0]\\ &=\sum_{j=0}^{n-1}f(j)[m-j=0]\\ &=f(m) \end{aligned}
f(m)=i=0∑n−1ωnmig(i)0≤m≤n−1=i=0∑n−1ωnmin1j=0∑n−1ωn−ijf(j)=j=0∑n−1f(j)n1i=0∑n−1ωni(m−j)根据公式③=j=0∑n−1f(j)[m−j modn=0]=j=0∑n−1f(j)[m−j=0]=f(m)
min-max反演
m
a
x
(
S
)
=
∑
T
⊆
S
(
−
1
)
∣
T
∣
−
1
m
i
n
(
T
)
m
i
n
(
S
)
=
∑
T
⊆
S
(
−
1
)
∣
T
∣
−
1
m
a
x
(
T
)
\begin{aligned} max(S)&=\sum_{T\subseteq S}(-1)^{|T|-1}min(T)\\ min(S)&=\sum_{T\subseteq S}(-1)^{|T|-1}max(T) \end{aligned}
max(S)min(S)=T⊆S∑(−1)∣T∣−1min(T)=T⊆S∑(−1)∣T∣−1max(T)
证明
枚举第
i
i
i大的值
a
i
a_i
ai作为
m
i
n
(
T
)
min(T)
min(T)，然后枚举哪些
T
T
T满足
a
i
=
m
i
n
(
T
)
a_i=min(T)
ai=min(T)
m
a
x
(
S
)
=
a
1
=
∑
i
=
1
∣
S
∣
∑
j
=
1
i
(
i
−
1
j
−
1
)
(
−
1
)
j
−
1
a
i
=
∑
i
=
1
∣
S
∣
a
i
∑
j
=
0
i
−
1
(
i
−
1
j
)
(
−
1
)
j
=
∑
i
=
1
∣
S
∣
a
i
[
i
−
1
=
0
]
=
a
1
\begin{aligned} max(S)=a_{1}&=\sum_{i=1}^{|S|}\sum_{j=1}^{i}{i-1\choose j-1}(-1)^{j-1}a_i\\ &=\sum_{i=1}^{|S|}a_i\sum_{j=0}^{i-1}{i-1\choose j}(-1)^j\\ &=\sum_{i=1}^{|S|}a_i[i-1=0]\\ &=a_1 \end{aligned}
max(S)=a1=i=1∑∣S∣j=1∑i(j−1i−1)(−1)j−1ai=i=1∑∣S∣aij=0∑i−1(ji−1)(−1)j=i=1∑∣S∣ai[i−1=0]=a1
第二个证明类似
kth min-max反演
假设容斥系数为
f
(
∣
T
∣
)
f(|T|)
f(∣T∣)
k
t
h
_
m
a
x
(
S
)
=
a
k
=
∑
T
⊆
S
f
(
∣
T
∣
)
m
i
n
(
T
)
=
∑
i
=
1
∣
S
∣
a
i
∑
j
=
1
i
(
i
−
1
j
−
1
)
f
(
j
)
\begin{aligned} kth\_max(S)=a_k&=\sum_{T\subseteq S}f(|T|)min(T)\\ &=\sum_{i=1}^{|S|}a_i\sum_{j=1}^{i}{i-1\choose j-1}f(j)\\ \end{aligned}
kth_max(S)=ak=T⊆S∑f(∣T∣)min(T)=i=1∑∣S∣aij=1∑i(j−1i−1)f(j)
向上面一样，只要找到
f
(
j
)
f(j)
f(j)，使得
∑
j
=
1
i
(
i
−
1
j
−
1
)
f
(
j
)
=
[
i
−
k
=
0
]
∑
j
=
0
i
−
1
(
i
−
1
j
)
f
(
j
+
1
)
=
[
i
−
k
=
0
]
\begin{aligned} \sum_{j=1}^{i}{i-1\choose j-1}f(j)&=[i-k=0]\\ \sum_{j=0}^{i-1}{i-1\choose j}f(j+1)&=[i-k=0] \end{aligned}
j=1∑i(j−1i−1)f(j)j=0∑i−1(ji−1)f(j+1)=[i−k=0]=[i−k=0]
式子就能成立。根据二项式反演得到
f
(
i
)
=
∑
j
=
0
i
−
1
(
−
1
)
i
−
1
−
j
(
i
−
1
j
)
[
j
−
k
+
1
=
0
]
=
(
−
1
)
i
−
k
(
i
−
1
k
−
1
)
\begin{aligned} f(i)&=\sum_{j=0}^{i-1}(-1)^{i-1-j}{i-1\choose j}[j-k+1=0]\\ &=(-1)^{i-k}{i-1\choose k-1} \end{aligned}
f(i)=j=0∑i−1(−1)i−1−j(ji−1)[j−k+1=0]=(−1)i−k(k−1i−1)
所以得到
k
t
h
_
m
a
x
(
S
)
=
∑
T
⊆
S
(
−
1
)
∣
T
∣
−
k
(
∣
T
∣
−
1
k
−
1
)
m
i
n
(
T
)
kth\_max(S)=\sum_{T\subseteq S}(-1)^{|T|-k}{|T|-1\choose k-1}min(T)
kth_max(S)=T⊆S∑(−1)∣T∣−k(k−1∣T∣−1)min(T)
2537
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
