Let
(
P
n
(
x
)
)
n
=
0
∞
{\displaystyle (P_{n}(x))_{n=0}^{\infty }}
be a sequence of orthogonal polynomials on the interval
[
a
,
b
]
{\displaystyle [a,b]}
with respect to weight function
w
(
x
)
{\displaystyle w(x)}
. That is, they have degrees
deg
(
P
n
)
=
n
{\displaystyle \deg(P_{n})=n}
, satisfy the orthogonality condition
∫
a
b
P
m
(
x
)
P
n
(
x
)
w
(
x
)
d
x
=
K
n
δ
m
,
n
{\displaystyle \int _{a}^{b}P_{m}(x)P_{n}(x)w(x)\,dx=K_{n}\delta _{m,n}}
where
K
n
{\displaystyle K_{n}}
are nonzero constants depending on
n
{\displaystyle n}
, and
δ
m
,
n
{\displaystyle \delta _{m,n}}
is the Kronecker delta . The interval
[
a
,
b
]
{\displaystyle [a,b]}
may be infinite in one or both ends.
Rodrigues' type formula — If
w
(
x
)
=
W
(
x
)
/
B
(
x
)
,
W
′
(
x
)
W
(
x
)
=
A
(
x
)
B
(
x
)
,
{\displaystyle w(x)=W(x)/B(x),\quad {\frac {W'(x)}{W(x)}}={\frac {A(x)}{B(x)}},}
where
A
(
x
)
{\displaystyle A(x)}
is a polynomial with degree at most 1 and
B
(
x
)
{\displaystyle B(x)}
is a polynomial with degree at most 2, and
lim
x
→
a
x
k
W
(
x
)
=
0
,
lim
x
→
b
x
k
W
(
x
)
=
0.
{\displaystyle \lim _{x\to a}x^{k}W(x)=0,\qquad \lim _{x\to b}x^{k}W(x)=0.}
for any
k
=
0
,
1
,
2
,
…
{\displaystyle k=0,1,2,\dots }
.
Then, if
d
n
d
x
n
[
B
(
x
)
n
w
(
x
)
]
≠
0
{\displaystyle {\frac {d^{n}}{dx^{n}}}\!\left[B(x)^{n}w(x)\right]\neq 0}
for all
n
=
0
,
1
,
2
,
…
{\displaystyle n=0,1,2,\dots }
, then
P
n
(
x
)
=
c
n
w
(
x
)
d
n
d
x
n
[
B
(
x
)
n
w
(
x
)
]
,
{\displaystyle P_{n}(x)={\frac {c_{n}}{w(x)}}{\frac {d^{n}}{dx^{n}}}\!\left[B(x)^{n}w(x)\right],}
for some constants
c
n
{\displaystyle c_{n}}
.
Proof[ 1]
Let
F
k
:=
1
w
D
x
k
(
B
n
w
)
{\textstyle F_{k}:={\frac {1}{w}}D_{x}^{k}(B^{n}w)}
, then
F
k
=
B
n
−
k
p
k
{\textstyle F_{k}=B^{n-k}p_{k}}
for all
k
∈
0
:
n
{\textstyle k\in 0:n}
for some polynomials
p
k
{\textstyle p_{k}}
, such that
d
e
g
(
p
k
)
≤
k
{\textstyle deg(p_{k})\leq k}
. Proven by induction on
k
{\textstyle k}
:
F
k
+
1
=
B
n
−
k
−
1
(
B
p
k
′
+
(
n
−
k
)
B
′
p
k
+
(
A
−
B
′
)
p
k
)
{\displaystyle F_{k+1}=B^{n-k-1}(Bp_{k}'+(n-k)B'p_{k}+(A-B')p_{k})}
Let
Q
n
:=
1
w
D
x
n
(
B
n
w
)
{\textstyle Q_{n}:={\frac {1}{w}}D_{x}^{n}(B^{n}w)}
. We have shown that
Q
n
{\textstyle Q_{n}}
is a polynomial of degree
≤
n
{\displaystyle \leq n}
. With integration by parts, we have for all
n
>
m
{\textstyle n>m}
,
∫
a
b
Q
m
Q
n
w
d
x
=
∫
a
b
B
n
w
(
D
x
n
Q
m
)
d
x
=
0
{\displaystyle \int _{a}^{b}Q_{m}Q_{n}wdx=\int _{a}^{b}B^{n}w(D_{x}^{n}Q_{m})dx=0}
since
D
x
n
Q
m
=
0
{\textstyle D_{x}^{n}Q_{m}=0}
. Thus,
Q
0
,
Q
1
,
…
{\textstyle Q_{0},Q_{1},\dots }
make up an orthogonal polynomial series with respect to
w
{\textstyle w}
. Thus,
P
n
=
c
n
Q
n
{\textstyle P_{n}=c_{n}Q_{n}}
for some constants
c
n
{\textstyle c_{n}}
.
Differential equation[ 2] —
B
(
x
)
d
2
d
x
2
P
n
(
x
)
+
A
(
x
)
d
d
x
P
n
(
x
)
+
λ
n
P
n
(
x
)
=
0
{\displaystyle B(x){\frac {d^{2}}{dx^{2}}}P_{n}(x)+A(x){\frac {d}{dx}}P_{n}(x)+\lambda _{n}P_{n}(x)=0}
λ
n
=
−
1
2
n
(
n
−
1
)
B
″
−
n
A
′
{\displaystyle \lambda _{n}=-{\frac {1}{2}}n(n-1)B''-nA'}
Proof[ 3]
When
n
=
0
{\displaystyle n=0}
, it is trivial. When
n
=
1
{\displaystyle n=1}
, it simplifies to
A
P
1
′
=
A
′
P
1
{\displaystyle AP_{1}'=A'P_{1}}
, which is true since
P
1
=
c
1
w
(
B
w
)
′
=
c
1
A
{\displaystyle P_{1}={\frac {c_{1}}{w}}(Bw)'=c_{1}A}
. So assume
n
≥
2
{\displaystyle n\geq 2}
. Define
I
n
(
x
)
=
d
n
d
x
n
(
B
n
(
x
)
w
(
x
)
)
{\displaystyle I_{n}(x)={\frac {d^{n}}{dx^{n}}}(B^{n}(x)w(x))}
, then by direct computation and simplification, the equation to be proven is equivalent to
d
2
d
x
2
(
B
(
x
)
I
n
(
x
)
)
−
d
d
x
(
A
(
x
)
I
n
(
x
)
)
+
λ
n
I
n
(
x
)
=
0
{\displaystyle {\frac {d^{2}}{dx^{2}}}(B(x)I_{n}(x))-{\frac {d}{dx}}(A(x)I_{n}(x))+\lambda _{n}I_{n}(x)=0}
By Leibniz differentiation rule, we have
B
(
x
)
d
n
d
x
n
y
=
d
n
d
x
n
(
B
(
x
)
y
)
−
n
d
n
−
1
d
x
n
−
1
(
B
′
(
x
)
y
)
+
n
(
n
−
1
)
2
d
n
−
2
d
x
n
−
2
(
B
″
y
)
{\displaystyle B(x){\frac {d^{n}}{dx^{n}}}y={\frac {d^{n}}{dx^{n}}}(B(x)y)-n{\frac {d^{n-1}}{dx^{n-1}}}(B'(x)y)+{\frac {n(n-1)}{2}}{\frac {d^{n-2}}{dx^{n-2}}}(B''y)}
A
(
x
)
d
n
d
x
n
y
=
d
n
d
x
n
(
A
(
x
)
y
)
−
n
d
n
−
1
d
x
n
−
1
(
A
′
y
)
{\displaystyle A(x){\frac {d^{n}}{dx^{n}}}y={\frac {d^{n}}{dx^{n}}}(A(x)y)-n{\frac {d^{n-1}}{dx^{n-1}}}(A'y)}
for arbitrary
y
{\displaystyle y}
. This allows us to move
A
(
x
)
,
B
(
x
)
{\displaystyle A(x),B(x)}
to the other side of the
n
{\displaystyle n}
-th derivative. Set
y
=
B
n
(
x
)
w
(
x
)
{\displaystyle y=B^{n}(x)w(x)}
, and define
J
(
x
)
=
d
2
d
x
2
(
B
(
x
)
y
(
x
)
)
−
n
d
d
x
(
B
′
(
x
)
y
(
x
)
)
+
n
(
n
−
1
)
2
B
″
y
(
x
)
{\displaystyle J(x)={\frac {d^{2}}{dx^{2}}}(B(x)y(x))-n{\frac {d}{dx}}(B'(x)y(x))+{\frac {n(n-1)}{2}}B''y(x)}
K
(
x
)
=
−
d
d
x
(
A
(
x
)
y
(
x
)
)
+
n
A
′
y
(
x
)
{\displaystyle K(x)=-{\frac {d}{dx}}(A(x)y(x))+nA'y(x)}
L
(
x
)
=
λ
n
y
(
x
)
{\displaystyle L(x)=\lambda _{n}y(x)}
Then the equation simplifies to
d
n
d
x
n
(
J
+
K
+
L
)
=
0
{\displaystyle {\frac {d^{n}}{dx^{n}}}(J+K+L)=0}
J
(
x
)
{\displaystyle J(x)}
has three terms, call them in order
J
1
(
x
)
,
J
2
(
x
)
,
J
3
(
x
)
{\displaystyle J_{1}(x),J_{2}(x),J_{3}(x)}
.
K
(
x
)
{\displaystyle K(x)}
has two terms, call them in order
K
1
(
x
)
,
K
2
(
x
)
{\displaystyle K_{1}(x),K_{2}(x)}
.
J
3
(
x
)
+
K
2
(
x
)
+
L
(
x
)
=
(
λ
n
+
n
(
n
−
1
)
2
B
″
+
n
A
′
)
y
=
0
{\displaystyle J_{3}(x)+K_{2}(x)+L(x)=(\lambda _{n}+{\frac {n(n-1)}{2}}B''+nA')y=0}
.
That
J
1
(
x
)
+
J
2
(
x
)
+
K
1
(
x
)
=
0
{\displaystyle J_{1}(x)+J_{2}(x)+K_{1}(x)=0}
. follows from first writing
J
1
(
x
)
{\displaystyle J_{1}(x)}
as
J
1
(
x
)
=
d
2
d
x
2
(
B
n
(
x
)
∫
exp
(
A
(
x
)
B
(
x
)
)
d
x
)
{\displaystyle J_{1}(x)={\frac {d^{2}}{dx^{2}}}\left(B^{n}(x)\int \exp \left({\frac {A(x)}{B(x)}}\right)dx\right)}
and then taking the innermost first derivative to obtain
J
1
(
x
)
=
d
d
x
[
(
n
B
′
(
x
)
B
n
−
1
(
x
)
+
A
(
x
)
B
n
−
1
(
x
)
)
∫
exp
(
A
(
x
)
B
(
x
)
)
d
x
]
{\displaystyle J_{1}(x)={\frac {d}{dx}}\left[{\bigg (}nB'(x)B^{n-1}(x)+A(x)B^{n-1}(x){\bigg )}\int \exp \left({\frac {A(x)}{B(x)}}\right)dx\right]}
and then rewriting this as
J
1
(
x
)
=
d
d
x
(
n
B
′
(
x
)
B
n
(
x
)
w
(
x
)
+
A
(
x
)
B
n
(
x
)
w
(
x
)
)
{\displaystyle J_{1}(x)={\frac {d}{dx}}{\Big (}nB'(x)B^{n}(x)w(x)+A(x)B^{n}(x)w(x){\Big )}}
The first term is the negative of
J
2
(
x
)
{\displaystyle J_{2}(x)}
and the second term is the negative of
K
1
(
x
)
{\displaystyle K_{1}(x)}
.
More abstractly, this can be viewed through Sturm–Liouville theory . Define an operator
L
f
:=
−
1
w
(
W
f
′
)
′
{\displaystyle Lf:=-{\frac {1}{w}}(Wf')'}
, then the differential equation is equivalent to
L
P
n
=
λ
n
P
n
{\displaystyle LP_{n}=\lambda _{n}P_{n}}
. Define the functional space
X
=
L
2
(
[
a
,
b
]
,
w
(
x
)
d
x
)
{\displaystyle X=L^{2}([a,b],w(x)dx)}
as the Hilbert space of functions over
[
a
,
b
]
{\displaystyle [a,b]}
, such that
⟨
f
,
g
⟩
:=
∫
a
b
f
g
w
{\displaystyle \langle f,g\rangle :=\int _{a}^{b}fgw}
. Then the operator
L
{\displaystyle L}
is self-adjoint on functions satisfying certain boundary conditions, allowing us to apply the spectral theorem .
Generating function [ edit ]
A simple argument using Cauchy's integral formula shows that the orthogonal polynomials obtained from the Rodrigues formula have a generating function of the form
G
(
x
,
u
)
=
∑
n
=
0
∞
u
n
P
n
(
x
)
G(x,u)=\sum _{n=0}^{\infty }u^{n}P_{n}(x)
The
P
n
(
x
)
{\displaystyle P_{n}(x)}
functions here may not have the standard normalizations. But we can write this equivalently as
G
(
x
,
u
)
=
∑
n
=
0
∞
u
n
N
n
N
n
P
n
(
x
)
G(x,u)=\sum _{n=0}^{\infty }{\frac {u^{n}}{N_{n}}}N_{n}P_{n}(x)
where the
N
n
{\displaystyle N_{n}}
are chosen according to the application so as to give the desired normalizations. The variable u may be replaced by a constant multiple of u so that
G
(
x
,
α
u
)
=
∑
n
=
0
∞
α
n
u
n
N
n
N
n
P
n
(
x
)
G(x,\alpha u)=\sum _{n=0}^{\infty }{\frac {\alpha ^{n}u^{n}}{N_{n}}}N_{n}P_{n}(x)
This gives an alternate form of the generating function.
By Cauchy's integral formula , Rodrigues’ formula is equivalent to
P
n
(
x
)
=
n
!
2
π
i
c
n
w
(
x
)
∮
C
B
n
(
t
)
w
(
t
)
(
t
−
x
)
n
+
1
d
t
{\displaystyle P_{n}(x)={\frac {n!}{2\pi i}}{\frac {c_{n}}{w(x)}}\oint _{C}{\frac {B^{n}(t)w(t)}{(t-x)^{n+1}}}\,dt}
where the integral is along a counterclockwise closed loop around
x
{\displaystyle x}
. Let
u
=
t
−
x
B
(
t
)
u={\frac {t-x}{B(t)}}
Then the complex path integral takes the form
P
n
(
x
)
=
n
!
2
π
i
c
n
∮
C
G
(
x
,
u
)
u
n
+
1
d
u
P_{n}(x)={\frac {n!}{2\pi i}}c_{n}\oint _{C}{\frac {G(x,u)}{u^{n+1}}}\,du
G
(
x
,
u
)
=
w
(
t
)
d
t
d
u
w
(
x
)
B
(
t
)
G(x,u)={\frac {w(t){\frac {dt}{du}}}{w(x)B(t)}}
where now the closed path C encircles the origin. In the equation for
G
(
x
,
u
)
{\displaystyle G(x,u)}
,
t
{\displaystyle t}
is an implicit function of
u
{\displaystyle u}
. Expanding
G
(
x
,
u
)
{\displaystyle G(x,u)}
in the power series given earlier gives
1
2
π
i
∮
C
G
(
x
,
u
)
u
n
+
1
d
u
=
1
2
π
i
∮
C
∑
m
=
0
∞
u
m
P
m
(
x
)
u
n
+
1
d
u
=
P
n
(
x
)
{\displaystyle {\frac {1}{2\pi i}}\oint _{C}{\frac {G(x,u)}{u^{n+1}}}\,du={\frac {1}{2\pi i}}\oint _{C}{\frac {\sum _{m=0}^{\infty }u^{m}P_{m}(x)}{u^{n+1}}}\,du=P_{n}(x)}
Only the
m
=
n
{\displaystyle m=n}
term has a nonzero residue, which is
P
n
(
x
)
{\displaystyle P_{n}(x)}
. The
n
!
c
n
{\displaystyle n!\,c_{n}}
coefficient was dropped since normalizations are conventions which can be inserted afterwards as discussed earlier.
By expressing t in terms of u in the general formula just given for
G
(
x
,
u
)
{\displaystyle G(x,u)}
, explicit formulas for
G
(
x
,
u
)
{\displaystyle G(x,u)}
may be found. As a simple example, let
B
(
x
)
=
1
{\displaystyle B(x)=1}
and
A
(
x
)
=
−
x
{\displaystyle A(x)=-x}
(Hermite polynomials) so that
w
(
x
)
=
exp
(
−
x
2
2
)
{\displaystyle w(x)=\exp \left(-{\frac {x^{2}}{2}}\right)}
,
t
=
u
+
x
{\displaystyle t=u+x}
,
w
(
t
)
=
exp
(
−
(
u
+
x
)
2
2
)
{\displaystyle w(t)=\exp \left(-{\frac {(u+x)^{2}}{2}}\right)}
and so
G
(
x
,
u
)
=
exp
(
−
x
u
−
u
2
2
)
{\displaystyle G(x,u)=\exp \left(-xu-{\frac {u^{2}}{2}}\right)}
.
Family
[
a
,
b
]
{\displaystyle [a,b]}
w
{\displaystyle w}
W
{\displaystyle W}
A
{\displaystyle A}
B
{\displaystyle B}
c
n
{\displaystyle c_{n}}
Legendre
P
n
{\displaystyle P_{n}}
[
−
1
,
+
1
]
{\displaystyle [-1,+1]}
1
{\displaystyle 1}
1
−
x
2
{\displaystyle 1-x^{2}}
−
2
x
{\displaystyle -2x}
1
−
x
2
{\displaystyle 1-x^{2}}
(
−
1
)
n
2
n
n
!
{\displaystyle {\frac {(-1)^{n}}{2^{n}n!}}}
Chebyshev (of the first kind)
T
n
{\displaystyle T_{n}}
[
−
1
,
+
1
]
{\displaystyle [-1,+1]}
1
/
1
−
x
2
{\displaystyle 1/{\sqrt {1-x^{2}}}}
1
−
x
2
{\displaystyle {\sqrt {1-x^{2}}}}
−
x
{\displaystyle -x}
1
−
x
2
{\displaystyle 1-x^{2}}
(
−
1
)
n
(
2
n
−
1
)
!
!
{\displaystyle {\frac {(-1)^{n}}{(2n-1)!!}}}
Chebyshev (of the second kind)
U
n
{\displaystyle U_{n}}
[
−
1
,
+
1
]
{\displaystyle [-1,+1]}
1
−
x
2
{\displaystyle {\sqrt {1-x^{2}}}}
(
1
−
x
2
)
3
/
2
{\displaystyle (1-x^{2})^{3/2}}
−
3
x
{\displaystyle -3x}
1
−
x
2
{\displaystyle 1-x^{2}}
(
−
1
)
n
(
n
+
1
)
(
2
n
+
1
)
!
!
{\displaystyle {\frac {(-1)^{n}(n+1)}{(2n+1)!!}}}
Gegenbauer/ultraspherical
C
n
(
α
)
(
x
)
{\displaystyle C_{n}^{(\alpha )}(x)}
[
−
1
,
+
1
]
{\displaystyle [-1,+1]}
(
1
−
x
)
α
−
1
/
2
(
1
+
x
)
α
−
1
/
2
{\displaystyle (1-x)^{\alpha -1/2}(1+x)^{\alpha -1/2}}
(
1
−
x
)
α
+
1
/
2
(
1
+
x
)
α
+
1
/
2
{\displaystyle (1-x)^{\alpha +1/2}(1+x)^{\alpha +1/2}}
−
(
2
α
+
1
)
x
{\displaystyle -(2\alpha +1)x}
1
−
x
2
{\displaystyle 1-x^{2}}
(
−
1
)
n
(
2
α
)
n
(
α
+
1
2
)
n
2
n
n
!
{\displaystyle {\frac {(-1)^{n}(2\alpha )_{n}}{(\alpha +{\frac {1}{2}})_{n}2^{n}n!}}}
Jacobi
P
n
(
α
,
β
)
{\displaystyle P_{n}^{(\alpha ,\beta )}}
[
−
1
,
+
1
]
{\displaystyle [-1,+1]}
(
1
−
x
)
α
(
1
+
x
)
β
{\displaystyle (1-x)^{\alpha }(1+x)^{\beta }}
(
1
−
x
)
α
+
1
(
1
+
x
)
β
+
1
{\displaystyle (1-x)^{\alpha +1}(1+x)^{\beta +1}}
(
β
−
α
)
−
(
α
+
β
+
2
)
x
{\displaystyle (\beta -\alpha )-(\alpha +\beta +2)x}
1
−
x
2
{\displaystyle 1-x^{2}}
(
−
1
)
n
2
n
n
!
{\displaystyle {\frac {(-1)^{n}}{2^{n}n!}}}
associated Laguerre
L
n
(
α
)
{\displaystyle L_{n}^{(\alpha )}}
[
0
,
∞
)
{\displaystyle [0,\infty )}
x
α
e
−
x
{\displaystyle x^{\alpha }e^{-x}}
x
α
+
1
e
−
x
{\displaystyle x^{\alpha +1}e^{-x}}
α
+
1
−
x
{\displaystyle \alpha +1-x}
x
{\displaystyle x}
1
n
!
{\displaystyle {\frac {1}{n!}}}
physicist's Hermite
H
n
{\displaystyle H_{n}}
(
−
∞
,
+
∞
)
{\displaystyle (-\infty ,+\infty )}
e
−
x
2
{\displaystyle e^{-x^{2}}}
e
−
x
2
{\displaystyle e^{-x^{2}}}
−
2
x
{\displaystyle -2x}
1
{\displaystyle 1}
(
−
1
)
n
{\displaystyle (-1)^{n}}
These formulae
[ 4]
[ 5] are for the classical orthogonal polynomials . Similar formulae hold for many other sequences of orthogonal functions arising from Sturm–Liouville equations , and these are also called the Rodrigues formula (or Rodrigues' type formula), especially when the resulting sequence is polynomial.
Source:[ 6]
Rodrigues stated his formula for Legendre polynomials
P
n
{\displaystyle P_{n}}
:
P
n
(
x
)
=
1
2
n
n
!
d
n
d
x
n
[
(
x
2
−
1
)
n
]
.
{\displaystyle P_{n}(x)={\frac {1}{2^{n}n!}}{\frac {d^{n}}{dx^{n}}}\!\left[(x^{2}-1)^{n}\right]\!.}
(
1
−
x
2
)
P
n
″
(
x
)
−
2
x
P
n
′
(
x
)
+
n
(
n
+
1
)
P
n
(
x
)
=
0
{\displaystyle (1-x^{2})P_{n}''(x)-2xP_{n}'(x)+n(n+1)P_{n}(x)=0}
For Legendre polynomials, the generating function is defined as
G
(
x
,
u
)
=
∑
n
=
0
∞
u
n
P
n
(
x
)
G(x,u)=\sum _{n=0}^{\infty }u^{n}P_{n}(x)
.
The contour integral gives the Schläfli integral [ 7] for Legendre polynomials:
P
n
(
x
)
=
1
2
π
i
2
n
∮
C
(
t
2
−
1
)
n
(
t
−
x
)
n
+
1
d
t
{\displaystyle P_{n}(x)={\frac {1}{2\pi i2^{n}}}\oint _{C}{\frac {(t^{2}-1)^{n}}{(t-x)^{n+1}}}dt}
Summing up the integrand
G
(
x
,
u
)
=
1
1
−
2
u
x
+
u
2
1
2
π
i
∮
C
(
1
t
−
t
−
−
1
t
−
t
+
)
d
t
{\displaystyle G(x,u)={\frac {1}{\sqrt {1-2ux+u^{2}}}}{\frac {1}{2\pi i}}\oint _{C}\left({\frac {1}{t-t_{-}}}-{\frac {1}{t-t_{+}}}\right)dt}
where
t
±
=
1
u
(
1
±
1
−
2
u
x
+
u
2
)
{\displaystyle t_{\pm }={\frac {1}{u}}(1\pm {\sqrt {1-2ux+u^{2}}})}
. For small
u
{\displaystyle u}
, we have
t
−
≈
x
,
t
+
→
∞
{\displaystyle t_{-}\approx x,t_{+}\to \infty }
, which heuristically suggests that the integral should be the residue around
t
−
{\displaystyle t_{-}}
, thus giving
G
(
x
,
u
)
=
1
1
−
2
u
x
+
u
2
{\displaystyle G(x,u)={\frac {1}{\sqrt {1-2ux+u^{2}}}}}
Source:[ 8]
Physicist's Hermite polynomials :
H
n
(
x
)
=
(
−
1
)
n
e
x
2
d
n
d
x
n
[
e
−
x
2
]
=
(
2
x
−
d
d
x
)
n
⋅
1.
{\displaystyle H_{n}(x)=(-1)^{n}e^{x^{2}}{\frac {d^{n}}{dx^{n}}}\!\left[e^{-x^{2}}\right]=\left(2x-{\frac {d}{dx}}\right)^{n}\cdot 1.}
H
n
″
−
2
x
H
n
′
+
2
n
H
n
=
0
{\displaystyle H_{n}''-2xH_{n}'+2nH_{n}=0}
The generating function is defined as
G
(
x
,
u
)
=
∑
n
=
0
∞
H
n
(
x
)
n
!
u
n
.
{\displaystyle G(x,u)=\sum _{n=0}^{\infty }{\frac {H_{n}(x)}{n!}}\,u^{n}.}
The contour integral gives
H
n
(
x
)
=
(
−
1
)
n
e
x
2
n
!
2
π
i
∮
C
e
−
t
2
(
t
−
x
)
n
+
1
d
t
.
{\displaystyle H_{n}(x)=(-1)^{n}e^{x^{2}}{\frac {n!}{2\pi i}}\oint _{C}{\frac {e^{-t^{2}}}{(t-x)^{n+1}}}\,dt.}
G
(
x
,
u
)
=
∑
n
=
0
∞
(
−
1
)
n
e
x
2
n
!
n
!
2
π
i
u
n
∮
C
e
−
t
2
(
t
−
x
)
n
+
1
d
t
=
e
x
2
1
2
π
i
∮
C
e
−
t
2
(
∑
n
=
0
∞
(
−
1
)
n
u
n
(
t
−
x
)
n
+
1
)
d
t
=
e
x
2
1
2
π
i
∮
C
e
−
t
2
1
t
−
x
+
u
=
e
x
2
e
−
(
x
−
u
)
2
=
e
2
x
u
−
u
2
{\displaystyle {\begin{aligned}G(x,u)&=\sum _{n=0}^{\infty }{\frac {(-1)^{n}e^{x^{2}}}{n!}}{\frac {n!}{2\pi i}}\,u^{n}\oint _{C}{\frac {e^{-t^{2}}}{(t-x)^{n+1}}}\,dt\\&=e^{x^{2}}{\frac {1}{2\pi i}}\oint _{C}e^{-t^{2}}\left(\sum _{n=0}^{\infty }{\frac {(-1)^{n}u^{n}}{(t-x)^{n+1}}}\right)dt\\&=e^{x^{2}}{\frac {1}{2\pi i}}\oint _{C}e^{-t^{2}}{\frac {1}{t-x+u}}\\&=e^{x^{2}}\,e^{-(x-u)^{2}}\\&=e^{2xu-u^{2}}\end{aligned}}}
Source:[ 9]
For associated Laguerre polynomials
L
n
(
α
)
(
x
)
=
x
−
α
e
x
n
!
d
n
d
x
n
(
e
−
x
x
n
+
α
)
=
x
−
α
n
!
(
d
d
x
−
1
)
n
x
n
+
α
.
{\displaystyle L_{n}^{(\alpha )}(x)={x^{-\alpha }e^{x} \over n!}{d^{n} \over dx^{n}}\left(e^{-x}x^{n+\alpha }\right)={\frac {x^{-\alpha }}{n!}}\left({\frac {d}{dx}}-1\right)^{n}x^{n+\alpha }.}
x
L
n
(
α
)
(
x
)
″
+
(
α
+
1
−
x
)
L
n
(
α
)
(
x
)
′
+
n
L
n
(
α
)
(
x
)
=
0
.
{\displaystyle xL_{n}^{(\alpha )}(x)''+(\alpha +1-x)L_{n}^{(\alpha )}(x)'+nL_{n}^{(\alpha )}(x)=0~.}
The generating function is defined as
G
(
x
,
u
)
:=
∑
n
=
0
∞
u
n
L
n
(
α
)
(
x
)
{\displaystyle G(x,u):=\sum _{n=0}^{\infty }u^{n}L_{n}^{(\alpha )}(x)}
By the same method, we have
G
(
x
,
u
)
=
1
(
1
−
u
)
α
+
1
e
−
u
x
1
−
u
{\displaystyle G(x,u)={\frac {1}{(1-u)^{\alpha +1}}}e^{-{\frac {ux}{1-u}}}}
.
Source:[ 10]
P
n
(
α
,
β
)
(
x
)
=
(
−
1
)
n
2
n
n
!
(
1
−
x
)
−
α
(
1
+
x
)
−
β
d
n
d
x
n
{
(
1
−
x
)
α
(
1
+
x
)
β
(
1
−
x
2
)
n
}
.
{\displaystyle P_{n}^{(\alpha ,\beta )}(x)={\frac {(-1)^{n}}{2^{n}n!}}(1-x)^{-\alpha }(1+x)^{-\beta }{\frac {d^{n}}{dx^{n}}}\left\{(1-x)^{\alpha }(1+x)^{\beta }\left(1-x^{2}\right)^{n}\right\}.}
(
1
−
x
2
)
P
n
(
α
,
β
)
″
+
(
β
−
α
−
(
α
+
β
+
2
)
x
)
P
n
(
α
,
β
)
′
+
n
(
n
+
α
+
β
+
1
)
P
n
(
α
,
β
)
=
0.
{\displaystyle \left(1-x^{2}\right)P_{n}^{(\alpha ,\beta )}{}''+(\beta -\alpha -(\alpha +\beta +2)x)P_{n}^{(\alpha ,\beta )}{}'+n(n+\alpha +\beta +1)P_{n}^{(\alpha ,\beta )}=0.}
∑
n
=
0
∞
P
n
(
α
,
β
)
(
x
)
u
n
=
2
α
+
β
R
−
1
(
1
−
u
+
R
)
−
α
(
1
+
u
+
R
)
−
β
,
{\displaystyle \sum _{n=0}^{\infty }P_{n}^{(\alpha ,\beta )}(x)u^{n}=2^{\alpha +\beta }R^{-1}(1-u+R)^{-\alpha }(1+u+R)^{-\beta },}
where
R
=
1
−
2
u
x
+
u
2
{\textstyle R={\sqrt {1-2ux+u^{2}}}}
, and the branch of square root is chosen so that
R
(
x
,
0
)
=
1
{\displaystyle R(x,0)=1}
.