Bessel Polynomials and Bessel Functions
Bessel Polynomials and Bessel Functions
1. The so called "modified" Bessel functions I)z) and K)z) - somet:j.mes improperly
designated as being of imaginary argument -, Kv(z) also known as the MacDonald func-
tion, satisfy the well-known (see e.g. [61] or [1]) differential equation
(1)
For fixed v with Re v > 0 the following asymptotic relations hold (see [1]):
For z + 0,
(2a)
-1 v -1 v-1
I)z) ; {r(v+l)} (z/2) , I~(z) ;; (2r(v)) (z/2) ;
These relations suggest that functions like ~(z), or 6(z), defined by w = z-v~. or
by w = z-ve-ze, with 6 = ez~. may exhibit a simpler behavior than the Bessel func-
tions themselves.
W
1
= (- 2v -1 + e)w
e 1
v e" e1 2 v e1 2
[-z + e- CeJ + c- z- -l + e) Jw
z
-1
Lw = (z6" - (2z+2v-1)6 1 + (2v-1)6)zw6 = 0.
By the general theory of linear differential equations (e.g., [11]) the origin
is a regular singular point and there exist (in general) two independent particular
solutions of the form e = za I c z
m
m
Here a is any one of the two (in general
m=O
5
c
2v+l-2m (2v-l) (2v-3) ... (2v-2m+l)
m (2v-m) m cm-1 (2v-l) (2v-2) ... (2v-m)m! cO •
so that
2 m
(2v-l) (2v-3) z
e (z, v) = co (l+z+ (2v-l) (2v-2) 2T + ••• +
(2v-l) (2v-3) ..• (2v-2m+l) z
(2v-l) (2v-2) ... (2v-m) m! + .. . ) .
Similarly,