Re: Adjoint of Integral operator.



In article <1127370447.074213.125950@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
Stephen <stevethemathsman@xxxxxxxxx> wrote:
>I have come across the following operator in my work (written in
>LaTeX),

>\int_t^\infty \tau^{\mu} f(\tau ) d \tau, where f is zero at
>infinity.

>I was wondering if anyone knows how to go about finding the adjoint of
>this operator.

I assume what you really mean is the following: you're working on the
Hilbert space H = L^2(0,infty), mu is a nonnegative constant, and your
operator is T where Tf(t) is given by your formula for all f in H
with compact support. Then for suitable g in H,

(g, Tf) = int_0^infty conjugate(g(t)) int_t^infty tau^mu f(tau) dtau dt
= int_0^infty tau^mu f(tau) int_0^tau conjugate(g(t)) dt dtau
= (T* g, f)

where T*g(tau) = tau^mu int_0^tau g(t) dt. The domain of T* is all
g in H for which tau^mu int_0^tau g(t) dt is in H (note in particular
that this will require int_0^infty g(t) dt = 0).

Robert Israel israel@xxxxxxxxxxx
Department of Mathematics http://www.math.ubc.ca/~israel
University of British Columbia Vancouver, BC, Canada

.



Relevant Pages

  • Adjoint of Integral operator.
    ... LaTeX), ... infinity. ... I was wondering if anyone knows how to go about finding the adjoint of ... Prev by Date: ...
    (sci.math)
  • Hint on This Limit Problem Please
    ... * For those of you with latex: ... The first thing I tried was to use Schwarz Inequality. ... But after a whole mess of algebra, the second integral goes to infinity. ... Prev by Date: ...
    (sci.math)
  • Re: Preferred format for math?
    ... I have seen infinity typed as "oo", but I have not yet seen ... any complex formulas involving summations or integrals... ... Would LaTeX be acceptable? ... You don't really need braces to tell us that t*e^-t*dt goes together, ...
    (sci.math)
  • Re: ECC point of inf question
    ... >> The point at infinity is defined precisely because the slope is ... I blame this on the ... > fact they've never heard of LaTeX and decide to do all their equations ...
    (sci.crypt)