Gegenbauer functions and applications (esp. circular envelope special case)?

417 Views Asked by At

Playing around for a solver for orthogonal polynomials in differential equations I stumbled upon the Gegenbauer polynomials described on Wikipedia the other day. They are the family of polynomials $y(x)$ satisfying the following differential equation: $$(1-x^2)y'' - (2\alpha+1) xy' +n(n+2\alpha)=0$$

I am curious in which applications they show up. In particular the one where $\alpha = 3/2$ which seems to correspond to the constant polynomial have it's function become a semi-circle. The function being constructed as : $$f_k(x) = \sqrt{w(x)}\cdot P_k(x)$$

In other words, building the functions so that they become orthogonal under the trivial (weight=1) scalar product, because

$$<f_i,f_j> = \int_{-1}^1 f_i(x)f_j(x)dx = \int_{-1}^1 P_i(x)P_j(x)w(x)dx = \delta_{i-j}$$

and since $w(x) = (1-x^2)^{\alpha-1/2}$, for $\alpha=3/2$ we get $$\sqrt{w(x)} = (1-x^2)^{(3/2-1/2)/2} = (1-x^2)^{1/2}$$ which is the curve of a circle (in fact the unit circle centered on origo) we can see in the plot for the function corresponding to the constant (bleu) polynomial below.enter image description here

In what kind of "real-world" application do these show up? Any branch of science or engineering is of interest.

1

There are 1 best solutions below

0
On BEST ANSWER

Gegenbauer are generalizations to the Legendre Polynomials, this is easy to see from their generating function:

$$ \frac{1}{(1 - 2xt + t^2)^{\alpha}} = \sum_{n=0}^{+\infty}C_n^{(\alpha)}t^n $$

which should be compared with the generating function for Legendre polynomials

$$ \frac{1}{(1 - 2xt + t^2)^{1/2}} = \sum_{n=0}^{+\infty}P_n(x)t^n $$

As such Gegenbauer polynomials provide a natural extension to spherical harmonics in higher dymensions: hyperspherical harmonics. In physics the set $P_n(x)$ is used to assemble the basis for solving the Hydrogen atom (two body problem), it is but expected that the set $C_n^{(\alpha)}$ can be used to solve a Coulomb problem in higher dimensions: and they do!

They are used for describing a basis for a many problem. Have a look at this reference.