Re: Fractional Transforms
- From: "rancid moth" <rancidmoth@xxxxxxxxx>
- Date: Mon, 15 Dec 2008 09:36:03 +1100
"amy666" <tommy1729@xxxxxxxxxxx> wrote in message
news:25975231.1229169752229.JavaMail.jakarta@xxxxxxxxxxxxxxxxxxxxxxxxx
Being intrested in both fractional iterations and many transforms like[cut]
fourier and laplace transforms it naturally leads to the following idea :
" Fractional Transforms "
let T be an integral transform
let L be the Laplace integral transform
T(T(f(x)) = L(f(x)) for all f(x).
Let T be defined by
T(f) = g(x) = int(0,oo) f(y)k(xy) dy
Let
T^-1(g)= f(x) = int(0,oo) g(y)h(xy) dy
then
H(s)*K(1-s)=1
where H and K are the mellin transforms of h and k. This is a nessecary and
sufficient condition that T(T^1)f =f once sutable restrictions are place on
the space for f.
Let f = L(v) for some function v. you want
T^1(L(v)) = T(v)
Take the mellin tranform of both sides
Gamma(1-s)V(s)H(s) = K(s)V(1-s) where V is the mellin tranform of v
This becomes
K(s) / (K(1-s)* Gamma(1-s)) = V(s) / V(1-s)
i think the above must show that a transform having a kernal of the form
k(xy) can not satisfy what you want.
of course this doesnt clear other integral transforms where the kernel takes
a different form say k(x-y) for example.
.
- Follow-Ups:
- Re: Fractional Transforms
- From: David C . Ullrich
- Re: Fractional Transforms
- References:
- Fractional Transforms
- From: amy666
- Fractional Transforms
- Prev by Date: Re: not anymore ?
- Next by Date: Re: Quasi repeating numbers
- Previous by thread: Re: Fractional Transforms
- Next by thread: Re: Fractional Transforms
- Index(es):
Relevant Pages
|
Loading