Jacobi polynomials explained

In mathematics, Jacobi polynomials (occasionally called hypergeometric polynomials)

(\alpha,\beta)
P
n

(x)

are a class of classical orthogonal polynomials. They are orthogonal with respect to the weight

(1-x)\alpha(1+x)\beta

on the interval

[-1,1]

. The Gegenbauer polynomials, and thus also the Legendre, Zernike and Chebyshev polynomials, are special cases of the Jacobi polynomials.[1]

The Jacobi polynomials were introduced by Carl Gustav Jacob Jacobi.

Definitions

Via the hypergeometric function

The Jacobi polynomials are defined via the hypergeometric function as follows:

(\alpha,\beta)
P(z)=
n
(\alpha+1)n
n!

{}2F1\left(-n,1+\alpha+\beta+n;\alpha+1;\tfrac{1}{2}(1-z)\right),

where

(\alpha+1)n

is Pochhammer's symbol (for the falling factorial). In this case, the series for the hypergeometric function is finite, therefore one obtains the following equivalent expression:
(\alpha,\beta)
P
n

(z)=

\Gamma(\alpha+n+1)
n!\Gamma(\alpha+\beta+n+1)
n
\sum
m=0

{n\choosem}

\Gamma(\alpha+\beta+n+m+1)\left(
\Gamma(\alpha+m+1)
z-1
2

\right)m.

Rodrigues' formula

An equivalent definition is given by Rodrigues' formula:[1]

(\alpha,\beta)
P
n

(z)=

(-1)n
2nn!

(1-z)-\alpha(1+z)-\beta

dn
dzn

\left\{(1-z)\alpha(1+z)\beta\left(1-z2\right)n\right\}.

If

\alpha=\beta=0

, then it reduces to the Legendre polynomials:

Pn(z)=

1
2nn!
dn
dzn

(z2-1)n.

Alternate expression for real argument

For real

x

the Jacobi polynomial can alternatively be written as
(\alpha,\beta)
P
n

(x)=

n
\sum
s=0

{n+\alpha\choosen-s}{n+\beta\chooses}\left(

x-1
2

\right)s\left(

x+1
2

\right)n-s

and for integer

n

{z\choosen}=\begin{cases}

\Gamma(z+1)
\Gamma(n+1)\Gamma(z-n+1)

&n\geq0\ 0&n<0\end{cases}

where

\Gamma(z)

is the gamma function.

In the special case that the four quantities

n

,

n+\alpha

,

n+\beta

,

n+\alpha+\beta

are nonnegative integers, the Jacobi polynomial can be written as

The sum extends over all integer values of

s

for which the arguments of the factorials are nonnegative.

Special cases

(\alpha,\beta)
P
0

(z)=1,

(\alpha,\beta)
P
1

(z)=(\alpha+1)+(\alpha+\beta+2)

z-1
2

,

(\alpha,\beta)
P
2

(z)=

(\alpha+1)(\alpha+2)
2

+(\alpha+2)(\alpha+\beta+3)

z-1
2

+

(\alpha+\beta+3)(\alpha+\beta+4)\left(
2
z-1
2

\right)2.

Basic properties

Orthogonality

The Jacobi polynomials satisfy the orthogonality condition

1
\int
-1

(1-x)\alpha(1+x)\beta

(\alpha,\beta)
P
m
(\alpha,\beta)
(x)P
n

(x)dx=

2\alpha+\beta+1
2n+\alpha+\beta+1
\Gamma(n+\alpha+1)\Gamma(n+\beta+1)
\Gamma(n+\alpha+\beta+1)n!

\deltanm,    \alpha,\beta>-1.

As defined, they do not have unit norm with respect to the weight. This can be corrected by dividing by the square root of the right hand side of the equation above, when

n=m

.

Although it does not yield an orthonormal basis, an alternative normalization is sometimes preferred due to its simplicity:

(\alpha,\beta)
P
n

(1)={n+\alpha\choosen}.

Symmetry relation

The polynomials have the symmetry relation

(\alpha,\beta)
P
n

(-z)=(-1)n

(\beta,\alpha)
P
n

(z);

thus the other terminal value is

(\alpha,\beta)
P
n

(-1)=(-1)n{n+\beta\choosen}.

Derivatives

The

k

th derivative of the explicit expression leads to
dk
dzk
(\alpha,\beta)
P
n

(z)=

\Gamma(\alpha+\beta+n+1+k)
2k\Gamma(\alpha+\beta+n+1)
(\alpha+k,\beta+k)
P
n-k

(z).

Differential equation

The Jacobi polynomial

(\alpha,\beta)
P
n
is a solution of the second order linear homogeneous differential equation[1]

\left(1-x2\right)y''+(\beta-\alpha-(\alpha+\beta+2)x)y'+n(n+\alpha+\beta+1)y=0.

Recurrence relations

The recurrence relation for the Jacobi polynomials of fixed

\alpha

,

\beta

is:[1]

\begin{align} &2n(n+\alpha+\beta)(2n+\alpha+\beta-2)

(\alpha,\beta)
P
n

(z)\\ &    =(2n+\alpha+\beta-1)\{(2n+\alpha+\beta)(2n+\alpha+\beta-2)z+\alpha2-\beta2\}

(\alpha,\beta)
P
n-1

(z)-2(n+\alpha-1)(n+\beta-1)(2n+\alpha+\beta)

(\alpha,\beta)
P
n-2

(z), \end{align}

for

n=2,3,\ldots

.Writing for brevity

a:=n+\alpha

,

b:=n+\beta

and

c:=a+b=2n+\alpha+\beta

, this becomes in terms of

a,b,c

2n(c-n)(c-2)

(\alpha,\beta)
P
n

(z)=(c-1)\{c(c-2)z+(a-b)(c-2n)\}

(\alpha,\beta)
P
n-1

(z)-2(a-1)(b-1)c

(\alpha,\beta)
P
n-2

(z).

Since the Jacobi polynomials can be described in terms of the hypergeometric function, recurrences of the hypergeometric function give equivalent recurrences of the Jacobi polynomials. In particular, Gauss' contiguous relations correspond to the identities

\begin{align} (z-1)

d
dz
(\alpha,\beta)
P
n

(z)&=

1
2
(\alpha+1,\beta+1)
(z-1)(1+\alpha+\beta+n)P
n-1

\\ &=n

(\alpha,\beta)
P
n

-(\alpha+n)

(\alpha,\beta+1)
P
n-1

\\ &=(1+\alpha+\beta+n)\left(

(\alpha,\beta+1)
P
n

-

(\alpha,\beta)
P
n

\right)\\ &=(\alpha+n)

(\alpha-1,\beta+1)
P
n

-\alpha

(\alpha,\beta)
P
n

\\ &=

2(n+1)
(\alpha,\beta-1)
P
n+1
-\left(z(1+\alpha+\beta+n)+\alpha+1+n-\beta\right)
(\alpha,\beta)
P
n
1+z

\\ &=

(2\beta+n+nz)
(\alpha,\beta)
P
n
-2(\beta+n)
(\alpha,\beta-1)
P
n
1+z

\\ &=

1-z
1+z

\left(\beta

(\alpha,\beta)
P
n

-(\beta+n)

(\alpha+1,\beta-1)
P
n

\right). \end{align}

Generating function

The generating function of the Jacobi polynomials is given by

infty
\sum
n=0
(\alpha,\beta)
P
n

(z)tn=2\alphaR-1(1-t+R)-\alpha(1+t+R)-\beta,

where

R=R(z,t)=\left(1-2zt+t2\right)

1
2

~,

and the branch of square root is chosen so that

R(z,0)=1

.[1]

Asymptotics of Jacobi polynomials

For

x

in the interior of

[-1,1]

, the asymptotics of
(\alpha,\beta)
P
n
for large

n

is given by the Darboux formula[1]
(\alpha,\beta)
P
n

(\cos\theta)=

-1
2
n

k(\theta)\cos(N\theta+\gamma)+O\left

-3
2
(n

\right),

where

\begin{align} k(\theta)&=

-1
2
\pi
-\alpha-1
2
\sin

\tfrac{\theta}{2}

-\beta-1
2
\cos

\tfrac{\theta}{2},\\ N&=n+\tfrac{1}{2}(\alpha+\beta+1),\\ \gamma&=-\tfrac{\pi}{2}\left(\alpha+\tfrac{1}{2}\right),\\ 0<\theta&<\pi \end{align}

and the "

O

" term is uniform on the interval

[\varepsilon,\pi-\varepsilon]

for every

\varepsilon>0

.

The asymptotics of the Jacobi polynomials near the points

\pm1

is given by the Mehler - Heine formula

\begin{align} \limnn-\alpha

(\alpha,\beta)
P
n

\left(\cos\left(\tfrac{z}{n}\right)\right)&=\left(\tfrac{z}{2}\right)-\alphaJ\alpha(z)\ \limnn-\beta

(\alpha,\beta)
P
n

\left(\cos\left(\pi-\tfrac{z}{n}\right)\right)&=\left(\tfrac{z}{2}\right)-\betaJ\beta(z) \end{align}

where the limits are uniform for

z

in a bounded domain.

The asymptotics outside

[-1,1]

is less explicit.

Applications

Wigner d-matrix

j
d
m',m

(\phi)

(for

0\leq\phi\leq4\pi

)in terms of Jacobi polynomials:[2]
j
d
m'm

(\phi)

m-m'-|m-m'|
2
=(-1)

\left[

(j+M)!(j-M)!
(j+N)!(j-N)!
1
2
\right]

\left(\sin\tfrac{\phi}{2}\right)|m-m'|\left(\cos\tfrac{\phi}{2}\right)|m+m'|

(|m-m'|,|m+m'|)
P
j-m

(\cos\phi),

where

M=max(|m|,|m'|),N=min(|m|,|m'|)

.

See also

Notes and References

  1. Book: Szegő . Gábor . Orthogonal Polynomials . American Mathematical Society . Colloquium Publications . 978-0-8218-1023-1 . 0372517 . 1939 . XXIII. IV. Jacobi polynomials.. The definition is in IV.1; the differential equation  - in IV.2; Rodrigues' formula is in IV.3; the generating function is in IV.4; the recurrent relation is in IV.5; the asymptotic behavior is in VIII.2
  2. Book: Biedenharn, L.C.. Louck. J.D.. Angular Momentum in Quantum Physics. Addison-Wesley . Reading . 1981.