Introduction
edit
Combinatorial class = C
Combinatorial parameter = \chi or \zeta
P(X = k) = #(y \in C | X(y) = k) / #C
So X, \chi, \zeta is both a variable and function...
Or, P(E) = \mu(E) = \sum_{e \in E} p(e)
f(x) (pdf)
F(x) = P(X \leq x) = \int_{-\infty}^x f(u)du (cdf)
For continuous variable:
Fourier/characteristic: \lambda_X(s) = \int_{\R} e^{sx} f(x) dx
Laplace: \phi_X(t) = \int_{\R} e^{itx} f(x) dx
(I don't necessarily believe the above...)
X* = \zeta - E(\zeta) / \sqrt{V(\zeta)}
If Y = X - \mu / \sigma then
\phi_Y(t) = e^{-\mu it} \phi_X(t/\sigma)
\lambda_Y(s) = e^{-\mu s} \lambda_X(s/\sigma)
Characteristic function of standard normal:
1/\sqrt{2\pi} \int_R e^{itz} e^{-z^2/2} dz
Discrete limit law
edit
This applies when the mean and standard deviation of the random variable remain finite.
If
Ω
{\displaystyle \Omega }
is a subset of the unit disc with at least one point of accumulation in the interior of the disc. If
p
n
(
u
)
=
∑
k
≥
0
p
n
,
k
u
k
{\displaystyle p_{n}(u)=\sum _{k\geq 0}p_{n,k}u^{k}}
,
q
(
u
)
=
∑
k
≥
0
q
k
u
k
{\displaystyle q(u)=\sum _{k\geq 0}q_{k}u^{k}}
and pointwise for each
u
∈
Ω
{\displaystyle u\in \Omega }
lim
n
→
+
∞
p
n
(
u
)
=
q
(
u
)
{\displaystyle \lim _{n\to +\infty }p_{n}(u)=q(u)}
then
lim
n
→
+
∞
p
n
,
k
=
q
k
{\displaystyle \lim _{n\to +\infty }p_{n,k}=q_{k}}
and
lim
n
→
+
∞
∑
j
≤
k
p
n
,
j
=
∑
j
≤
k
q
j
{\displaystyle \lim _{n\to +\infty }\sum _{j\leq k}p_{n,j}=\sum _{j\leq k}q_{j}}
[1]
Applying Vitali's theorem , we take the sequence to be
{
p
n
(
u
)
}
{\displaystyle \{p_{n}(u)\}}
and
S
{\displaystyle S}
to be the unit disc. All
p
n
(
u
)
{\displaystyle p_{n}(u)}
are analytic and bounded on
S
{\displaystyle S}
(due to
p
n
(
1
)
=
1
{\displaystyle p_{n}(1)=1}
). The theorem assumes that
{
p
n
(
u
)
}
{\displaystyle \{p_{n}(u)\}}
converges to
q
(
u
)
{\displaystyle q(u)}
in
Ω
⊂
S
{\displaystyle \Omega \subset S}
which has an accumulation point in
S
{\displaystyle S}
.
Vitali's theorem states
{
p
n
(
u
)
}
{\displaystyle \{p_{n}(u)\}}
uniformly converges to
q
(
u
)
{\displaystyle q(u)}
on any compact subset of
S
{\displaystyle S}
, which we will take as the disc
|
u
|
≤
1
/
2
{\displaystyle |u|\leq 1/2}
and use Cauchy's coefficient formula :
q
k
=
1
2
i
π
∫
|
u
|
=
1
/
2
q
(
u
)
d
u
u
k
+
1
=
lim
n
→
∞
1
2
i
π
∫
|
u
|
=
1
/
2
p
n
(
u
)
d
u
u
k
+
1
=
lim
n
→
∞
p
n
,
k
{\displaystyle q_{k}={\frac {1}{2i\pi }}\int _{|u|=1/2}q(u){\frac {du}{u^{k+1}}}=\lim _{n\to \infty }{\frac {1}{2i\pi }}\int _{|u|=1/2}p_{n}(u){\frac {du}{u^{k+1}}}=\lim _{n\to \infty }p_{n,k}}
[2]
Vitali's theorem
edit
Theorem
Let
{
f
n
(
z
)
}
{\displaystyle \{f_{n}(z)\}}
be a sequence of functions, all analytic on an open connected set
S
{\displaystyle S}
and
|
f
n
(
z
)
|
≤
M
{\displaystyle |f_{n}(z)|\leq M}
for all
n
{\displaystyle n}
and
z
{\displaystyle z}
in
S
{\displaystyle S}
.
If the sequence
{
f
n
(
z
)
}
{\displaystyle \{f_{n}(z)\}}
converges on a subset of
S
{\displaystyle S}
with a point of accumulation in
S
{\displaystyle S}
, then
{
f
n
(
z
)
}
{\displaystyle \{f_{n}(z)\}}
converges uniformly on every compact subset of
S
{\displaystyle S}
.[3]
Proof
Let
f
n
(
z
)
=
a
0
,
n
+
a
1
,
n
z
+
⋯
{\displaystyle f_{n}(z)=a_{0,n}+a_{1,n}z+\cdots }
We want to prove that each
a
v
,
n
{\displaystyle a_{v,n}}
converges to a limit.
We start with a disc centred at the origin of radius
R
{\displaystyle R}
and point of accumulation as the origin. Then
|
f
n
(
z
)
−
f
n
(
0
)
|
≤
|
f
n
(
z
)
|
+
|
f
n
(
0
)
|
≤
2
M
{\displaystyle |f_{n}(z)-f_{n}(0)|\leq |f_{n}(z)|+|f_{n}(0)|\leq 2M}
Let
z
0
≠
0
{\displaystyle z_{0}\neq 0}
be a point where the sequence converges. Then
|
f
n
(
0
)
−
f
n
+
m
(
0
)
|
≤
|
f
n
(
0
)
−
f
n
(
z
0
)
|
+
|
f
n
(
z
0
)
−
f
n
+
m
(
z
0
)
|
+
|
f
n
+
m
(
z
0
)
−
f
n
+
m
(
0
)
|
≤
4
M
|
z
0
|
R
+
|
f
n
(
z
0
)
−
f
n
+
m
(
z
0
)
|
{\displaystyle |f_{n}(0)-f_{n+m}(0)|\leq |f_{n}(0)-f_{n}(z_{0})|+|f_{n}(z_{0})-f_{n+m}(z_{0})|+|f_{n+m}(z_{0})-f_{n+m}(0)|\leq {\frac {4M|z_{0}|}{R}}+|f_{n}(z_{0})-f_{n+m}(z_{0})|}
because, by Schwarz's lemma ,
|
f
n
(
0
)
−
f
n
(
z
0
)
|
≤
2
M
|
z
0
|
/
R
{\displaystyle |f_{n}(0)-f_{n}(z_{0})|\leq 2M|z_{0}|/R}
and
|
f
n
+
m
(
z
0
)
−
f
n
+
m
(
0
)
|
≤
2
M
|
z
0
|
/
R
{\displaystyle |f_{n+m}(z_{0})-f_{n+m}(0)|\leq 2M|z_{0}|/R}
.
We choose
z
0
{\displaystyle z_{0}}
such that the first term is arbitrarily small and
n
{\displaystyle n}
large enough that the second term is arbitrarily small. Therefore,
f
n
(
0
)
=
a
0
,
n
{\displaystyle f_{n}(0)=a_{0,n}}
converges to a limit.
Next, we define
g
n
(
z
)
=
f
n
(
z
)
−
a
0
,
n
z
=
a
1
,
n
+
a
2
,
n
z
+
⋯
≤
2
M
R
{\displaystyle g_{n}(z)={\frac {f_{n}(z)-a_{0,n}}{z}}=a_{1,n}+a_{2,n}z+\cdots \leq {\frac {2M}{R}}}
For
|
z
|
<
R
{\displaystyle |z|<R}
.
g
n
(
z
)
{\displaystyle g_{n}(z)}
converges to a limit at
z
0
{\displaystyle z_{0}}
as both
f
n
(
z
)
{\displaystyle f_{n}(z)}
and
a
0
,
n
{\displaystyle a_{0,n}}
converge to a limit. We repeat the argument above, which proved that
f
n
(
0
)
=
a
0
,
n
{\displaystyle f_{n}(0)=a_{0,n}}
converges to a limit, to prove that
g
n
(
0
)
=
a
1
,
n
{\displaystyle g_{n}(0)=a_{1,n}}
also converges to a limit. We keep repeating this to prove that
a
v
,
n
{\displaystyle a_{v,n}}
converges for all
v
{\displaystyle v}
.
Since each term of
f
n
(
z
)
{\displaystyle f_{n}(z)}
converges to a limit,
f
n
(
z
)
{\displaystyle f_{n}(z)}
also converges to a limit for
z
<
R
{\displaystyle z<R}
. If we repeat the argument with another disc centred at any limit point (e.g.
z
0
{\displaystyle z_{0}}
), we can extend to any region bounded by
S
{\displaystyle S}
.[4]
Schwarz's lemma
edit
Lemma
If
f
(
z
)
{\displaystyle f(z)}
is analytic and regular(?) for
|
z
|
≤
R
{\displaystyle |z|\leq R}
,
|
f
(
z
)
|
≤
M
{\displaystyle |f(z)|\leq M}
for
|
z
|
=
R
{\displaystyle |z|=R}
and
f
(
0
)
=
0
{\displaystyle f(0)=0}
, then
f
(
r
e
i
θ
≤
M
r
R
(
0
≤
r
≤
R
)
{\displaystyle f(re^{i\theta }\leq {\frac {Mr}{R}}\quad (0\leq r\leq R)}
[5]
Proof
f
(
z
)
/
z
{\displaystyle f(z)/z}
is regular for
|
z
|
≤
R
{\displaystyle |z|\leq R}
and
|
f
(
z
)
/
z
|
≤
M
/
R
{\displaystyle |f(z)/z|\leq M/R}
on the circle
|
z
|
=
R
{\displaystyle |z|=R}
. By the maximum modulus theorem(?) this inequality holds also inside the circle. Because
|
f
(
z
)
/
z
|
=
|
f
(
z
)
|
/
r
{\displaystyle |f(z)/z|=|f(z)|/r}
we multiply both sides of the inequality by
r
{\displaystyle r}
which gives us our lemma.[6]
The number
k
{\displaystyle k}
of singleton cycles in a permutation of length
n
{\displaystyle n}
is given by
P
(
z
,
u
)
=
e
z
(
u
−
1
)
1
−
z
{\displaystyle P(z,u)={\frac {e^{z(u-1)}}{1-z}}}
.
For each
u
{\displaystyle u}
we have a singularity at
z
=
1
{\displaystyle z=1}
of order 1. If we treat
u
{\displaystyle u}
as a constant and apply our estimate for meromorphic functions
p
n
(
u
)
=
[
z
n
]
P
(
z
,
u
)
∼
(
−
1
)
e
(
u
−
1
)
−
1
=
e
u
−
1
{\displaystyle p_{n}(u)=[z^{n}]P(z,u)\sim {\frac {(-1)e^{(u-1)}}{-1}}=e^{u-1}}
as
n
→
∞
{\displaystyle n\to \infty }
.
Therefore,
p
n
,
k
∼
[
u
k
]
e
u
−
1
=
1
k
!
e
{\displaystyle p_{n,k}\sim [u^{k}]e^{u-1}={\frac {1}{k!e}}}
as
n
→
∞
{\displaystyle n\to \infty }
.[7]
Continuous limit law
edit
This applies when the mean and standard deviation of the random variable tend to infinity.
There are three cases/theorems:
Meromorphic
Singularity
Saddle
Meromorphic schema
edit
Theorem
Let
F
(
z
,
u
)
{\displaystyle F(z,u)}
be bivariate analytic at
(
z
,
u
)
=
(
0
,
0
)
{\displaystyle (z,u)=(0,0)}
with non-negative coefficients. Let
F
(
z
,
1
)
{\displaystyle F(z,1)}
be meromorphic in
z
≤
r
{\displaystyle z\leq r}
with only a simple pole
ρ
{\displaystyle \rho }
for positive
ρ
<
r
{\displaystyle \rho <r}
. Assume:
Meromorphic perturbation : there exists
ϵ
>
0
{\displaystyle \epsilon >0}
and
r
>
ρ
{\displaystyle r>\rho }
such that in the domain
D
=
{
|
z
|
≤
r
}
×
{
|
u
−
1
|
<
ϵ
}
{\displaystyle D=\{|z|\leq r\}\times \{|u-1|<\epsilon \}}
:
F
(
z
,
u
)
=
B
(
z
,
u
)
C
(
z
,
u
)
{\displaystyle F(z,u)={\frac {B(z,u)}{C(z,u)}}}
where
B
(
z
,
u
)
{\displaystyle B(z,u)}
and
C
(
z
,
u
)
{\displaystyle C(z,u)}
are analytic in
D
{\displaystyle D}
and
B
(
ρ
,
1
)
≠
0
{\displaystyle B(\rho ,1)\neq 0}
and
ρ
{\displaystyle \rho }
is a simple zero of
C
(
z
,
1
)
{\displaystyle C(z,1)}
.
Non-degeneracy :
∂
z
C
(
ρ
,
1
)
∂
u
C
(
ρ
,
1
)
≠
0
{\displaystyle \partial _{z}C(\rho ,1)\partial _{u}C(\rho ,1)\neq 0}
.
Variability :
v
(
ρ
(
1
)
ρ
(
u
)
)
≠
0
{\displaystyle v\left({\frac {\rho (1)}{\rho (u)}}\right)\neq 0}
.
Then, the standardised random variable
X
n
{\displaystyle X_{n}}
converges to the Gaussian variable, with speed of convergence
O
(
n
−
1
/
2
)
{\displaystyle O(n^{-1/2})}
and mean and standard deviation asymptotically linear in
n
{\displaystyle n}
.[8]
Proof
Construct annular region composed of two concentric circles
C
s
{\displaystyle C_{s}}
and
C
r
{\displaystyle C_{r}}
of radius
s
{\displaystyle s}
(
ρ
<
s
<
r
{\displaystyle \rho <s<r}
) and
r
{\displaystyle r}
, respectively. By the global Cauchy formula[9] and residue theorem, for
z
{\displaystyle z}
in the annular region and
u
{\displaystyle u}
in a small enough domain around 1,
f
n
(
u
)
=
1
2
π
i
∫
C
s
B
(
z
,
u
)
C
(
z
,
u
)
d
z
z
n
+
1
+
1
2
π
i
∫
C
r
F
(
z
,
u
)
d
z
z
n
+
1
=
R
e
s
(
B
(
z
,
u
)
C
(
z
,
u
)
z
n
+
1
;
z
=
ρ
(
u
)
)
+
1
2
π
i
∫
C
r
F
(
z
,
u
)
d
z
z
n
+
1
{\displaystyle f_{n}(u)={\frac {1}{2\pi i}}\int _{C_{s}}{\frac {B(z,u)}{C(z,u)}}{\frac {dz}{z^{n+1}}}+{\frac {1}{2\pi i}}\int _{C_{r}}F(z,u){\frac {dz}{z^{n+1}}}=Res\left({\frac {B(z,u)}{C(z,u)z^{n+1}}};z=\rho (u)\right)+{\frac {1}{2\pi i}}\int _{C_{r}}F(z,u){\frac {dz}{z^{n+1}}}}
By Cauchy's inequality ,
∫
C
r
F
(
z
,
u
)
d
z
z
n
+
1
≤
max
|
z
|
=
r
,
|
u
−
1
|
≤
ϵ
|
F
(
z
,
u
)
|
r
n
=
sup
|
z
|
=
r
,
|
u
−
1
|
≤
ϵ
|
B
(
z
,
u
)
|
inf
|
z
|
=
r
,
|
u
−
1
|
≤
ϵ
|
C
(
z
,
u
)
|
r
n
=
K
r
n
=
O
(
r
−
n
)
{\displaystyle \int _{C_{r}}F(z,u){\frac {dz}{z^{n+1}}}\leq {\frac {\max _{|z|=r,|u-1|\leq \epsilon }|F(z,u)|}{r^{n}}}={\frac {\sup _{|z|=r,|u-1|\leq \epsilon }|B(z,u)|}{\inf _{|z|=r,|u-1|\leq \epsilon }|C(z,u)|r^{n}}}={\frac {K}{r^{n}}}=O(r^{-n})}
where
K
{\displaystyle K}
is a finite constant, due to the fact that
C
(
z
,
u
)
{\displaystyle C(z,u)}
is non-zero on
|
z
|
=
r
{\displaystyle |z|=r}
and
B
(
z
,
u
)
{\displaystyle B(z,u)}
is analytic and therefore bounded...
R
e
s
(
B
(
z
,
u
)
C
(
z
,
u
)
z
n
+
1
;
z
=
ρ
(
u
)
)
=
B
(
ρ
(
u
)
,
u
)
C
z
′
(
ρ
(
u
)
,
u
)
ρ
(
u
)
n
+
1
{\displaystyle Res\left({\frac {B(z,u)}{C(z,u)z^{n+1}}};z=\rho (u)\right)={\frac {B(\rho (u),u)}{C_{z}^{'}(\rho (u),u)\rho (u)^{n+1}}}}
Therefore,
f
n
(
u
)
=
B
(
ρ
(
u
)
,
u
)
C
z
′
(
ρ
(
u
)
,
u
)
ρ
(
u
)
n
+
1
+
O
(
r
−
n
)
{\displaystyle f_{n}(u)={\frac {B(\rho (u),u)}{C_{z}^{'}(\rho (u),u)\rho (u)^{n+1}}}+O(r^{-n})}
meaning it meets the conditions of the Quasi-powers theorem .[10]
Singularity
edit
Saddle-point
edit
Theorem
Assume
p
n
(
u
)
=
e
h
n
(
u
)
(
1
+
o
(
1
)
)
{\displaystyle p_{n}(u)=e^{h_{n}(u)}(1+o(1))}
uniformly for
u
{\displaystyle u}
in a fixed neighbourhood
Ω
{\displaystyle \Omega }
of
1
{\displaystyle 1}
and each
h
n
(
u
)
{\displaystyle h_{n}(u)}
analytic in
Ω
{\displaystyle \Omega }
. Assume also
h
n
′
(
1
)
+
h
n
″
(
1
)
→
∞
{\displaystyle h_{n}'(1)+h_{n}''(1)\to \infty }
and
h
n
‴
(
u
)
(
h
n
′
(
1
)
+
h
n
″
(
1
)
)
3
/
2
→
0
{\displaystyle {\frac {h_{n}'''(u)}{(h_{n}'(1)+h_{n}''(1))^{3/2}}}\to 0}
uniformly for
u
∈
Ω
{\displaystyle u\in \Omega }
. Then
X
n
∗
=
X
n
−
h
n
′
(
1
)
(
h
n
′
(
1
)
+
h
n
″
(
1
)
)
1
/
2
{\displaystyle X_{n}^{*}={\frac {X_{n}-h_{n}'(1)}{(h_{n}'(1)+h_{n}''(1))^{1/2}}}}
converges in distribution to a Gaussian with mean 0 and variance 1.[11]
Proof
Quasi-powers theorem
edit
Theorem
Let the
X
n
{\displaystyle X_{n}}
be non-negative discrete random variables and
p
n
(
u
)
{\displaystyle p_{n}(u)}
be probability generating functions. Assume uniformly in a fixed complex neighbourhood of
u
=
1
{\displaystyle u=1}
for
β
n
,
κ
n
→
+
∞
{\displaystyle \beta _{n},\kappa _{n}\to +\infty }
p
n
(
u
)
=
A
(
u
)
.
B
(
u
)
β
n
(
1
+
O
(
1
κ
n
)
)
{\displaystyle p_{n}(u)=A(u).B(u)^{\beta _{n}}\left(1+O\left({\frac {1}{\kappa _{n}}}\right)\right)}
where
A
(
u
)
,
B
(
u
)
{\displaystyle A(u),B(u)}
are analytic at
u
=
1
{\displaystyle u=1}
and
A
(
u
)
=
B
(
u
)
=
1
{\displaystyle A(u)=B(u)=1}
. Assume
B
(
u
)
{\displaystyle B(u)}
satisfies the "variability condition"
v
(
B
(
u
)
)
≡
B
″
(
1
)
+
B
′
(
1
)
−
B
′
(
1
)
2
≠
0
{\displaystyle v(B(u))\equiv B''(1)+B'(1)-B'(1)^{2}\neq 0}
Then the mean and variance of
X
n
{\displaystyle X_{n}}
satisfy
μ
n
≡
E
(
X
n
)
=
β
n
m
(
B
(
u
)
)
+
m
(
A
(
u
)
)
+
O
(
κ
n
−
1
)
{\displaystyle \mu _{n}\equiv E(X_{n})=\beta _{n}m(B(u))+m(A(u))+O(\kappa _{n}^{-1})}
σ
n
2
≡
V
(
X
n
)
=
β
n
v
(
B
(
u
)
)
+
v
(
A
(
u
)
)
+
O
(
κ
n
−
1
)
{\displaystyle \sigma _{n}^{2}\equiv V(X_{n})=\beta _{n}v(B(u))+v(A(u))+O(\kappa _{n}^{-1})}
and
P
(
X
n
−
E
(
X
n
)
V
(
X
n
)
≤
x
)
=
Φ
(
x
)
+
O
(
1
κ
n
+
1
β
n
)
{\displaystyle P\left({\frac {X_{n}-E(X_{n})}{\sqrt {V(X_{n})}}}\leq x\right)=\Phi (x)+O\left({\frac {1}{\kappa _{n}}}+{\frac {1}{\sqrt {\beta _{n}}}}\right)}
where
Φ
(
x
)
=
1
2
π
∫
−
∞
x
e
−
w
2
/
2
d
w
{\displaystyle \Phi (x)={\frac {1}{\sqrt {2\pi }}}\int _{-\infty }^{x}e^{-w^{2}/2}dw}
.[12]
Proof
Convert to a Laplace transform
λ
n
(
s
)
=
e
β
n
log
B
(
s
)
+
log
A
(
s
)
(
1
+
O
(
κ
n
−
1
)
)
{\displaystyle \lambda _{n}(s)=e^{\beta _{n}\log B(s)+\log A(s)}(1+O(\kappa _{n}^{-1}))}
therefore
log
λ
n
(
s
)
=
β
n
log
B
(
s
)
+
log
A
(
s
)
+
O
(
κ
n
−
1
)
{\displaystyle \log \lambda _{n}(s)=\beta _{n}\log B(s)+\log A(s)+O(\kappa _{n}^{-1})}
d
d
s
log
λ
n
(
s
)
=
β
n
d
d
s
log
B
(
s
)
+
d
d
s
log
A
(
s
)
+
O
(
κ
n
−
1
)
=
β
n
m
(
B
(
s
)
)
+
m
(
A
(
s
)
)
+
O
(
κ
n
−
1
)
{\displaystyle {\frac {d}{ds}}\log \lambda _{n}(s)=\beta _{n}{\frac {d}{ds}}\log B(s)+{\frac {d}{ds}}\log A(s)+O(\kappa _{n}^{-1})=\beta _{n}m(B(s))+m(A(s))+O(\kappa _{n}^{-1})}
d
2
d
s
2
log
λ
n
(
s
)
=
β
n
d
2
d
s
2
log
B
(
s
)
+
d
2
d
s
2
log
A
(
s
)
+
O
(
κ
n
−
1
)
=
β
n
v
(
B
(
s
)
)
+
v
(
A
(
s
)
)
+
O
(
κ
n
−
1
)
{\displaystyle {\frac {d^{2}}{ds^{2}}}\log \lambda _{n}(s)=\beta _{n}{\frac {d^{2}}{ds^{2}}}\log B(s)+{\frac {d^{2}}{ds^{2}}}\log A(s)+O(\kappa _{n}^{-1})=\beta _{n}v(B(s))+v(A(s))+O(\kappa _{n}^{-1})}
[13]
Continuity of integral transforms
edit
Theorem from Flajolet and Sedgewick[14] .
Fourier
Also known as Lévy's continuity theorem.
If
p
n
(
u
)
{\displaystyle p_{n}(u)}
and
q
(
u
)
{\displaystyle q(u)}
have Fourier transforms
ϕ
n
(
t
)
{\displaystyle \phi _{n}(t)}
and
ϕ
(
t
)
{\displaystyle \phi (t)}
, respectively, and
q
(
u
)
{\displaystyle q(u)}
has a continuous distribution function, then
lim
n
→
+
∞
p
n
(
u
)
=
q
(
u
)
{\displaystyle \lim _{n\to +\infty }p_{n}(u)=q(u)}
if and only if, pointwise for each real
t
{\displaystyle t}
lim
n
→
+
∞
ϕ
n
(
t
)
=
ϕ
(
t
)
{\displaystyle \lim _{n\to +\infty }\phi _{n}(t)=\phi (t)}
Laplace
Assume
p
n
(
u
)
{\displaystyle p_{n}(u)}
and
q
(
u
)
{\displaystyle q(u)}
have Laplace transforms
λ
n
(
s
)
{\displaystyle \lambda _{n}(s)}
and
λ
(
s
)
{\displaystyle \lambda (s)}
, respectively, defined on
s
∈
[
−
s
0
,
s
0
]
{\displaystyle s\in [-s_{0},s_{0}]}
. If pointwise for each real
s
∈
[
−
s
0
,
s
0
]
{\displaystyle s\in [-s_{0},s_{0}]}
lim
n
→
+
∞
λ
n
(
u
)
=
λ
(
u
)
{\displaystyle \lim _{n\to +\infty }\lambda _{n}(u)=\lambda (u)}
then
lim
n
→
+
∞
p
n
(
u
)
=
q
(
u
)
{\displaystyle \lim _{n\to +\infty }p_{n}(u)=q(u)}
Appendix
edit
↑ Flajolet and Sedgewick 2009, pp. 624.
↑ Flajolet and Sedgewick 2009, pp. 624.
↑ Flajolet and Sedgewick 2009, pp. 624. Titchmarsh 1939, pp. 168-169.
↑ Titchmarsh 1939, pp. 168-169.
↑ Titchmarsh 1939, pp. 168.
↑ Titchmarsh 1939, pp. 168.
↑ Flajolet and Sedgewick 2009, pp. 625-626.
↑ Flajolet and Sedgewick 2009, pp. 656.
↑ Lang 1999, pp. 145.
↑ Flajolet and Sedgewick 2009, pp. 656-657.
↑ Flajolet and Sedgewick 2009, pp. 690.
↑ Flajolet and Sedgewick 2009, pp. 645-646.
↑ Flajolet and Sedgewick 2009, pp. 646-647.
↑ Flajolet and Sedgewick 2009, pp. 639-640.
References
edit
Flajolet, Philippe; Sedgewick, Robert (2009). Analytic Combinatorics (PDF) . Cambridge University Press.
Lang, Serge (1999). Complex Analysis (4th ed.). Springer Science+Business Media, LLC.
Titchmarsh, E. C. (1939). The Theory of Functions (2nd ed.). Oxford University Press.