Re: > Eigenvalue problem for sufficiently well-behaved functions
- From: Robert Israel <israel@xxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
- Date: Mon, 06 Oct 2008 12:21:35 -0500
"k.hofmann" <boquiqui@xxxxxxxxxxxxxxxx> writes:
Suppose f is a sufficiently well-behaved real-valued function (say, a
smooth function with compact support) defined on the set of real numbers),
and define
T(f) = \integal _{-\infty}^x y f(y) dy
What are the eigenvalues for T ?
None, on that domain.
Which eigenvalues lead to square-integrable functions?
All c with Re(c) < 0. That is, these are eigenvalues for the unbounded
operator on L^2(R) with domain
{f in L^2: int_{-infty}^x y f(y) dy in L^2(R)}.
I tried differentiating both sides of
T(f) = cf
and obtained a differential equation
cf'(x) - x f(x) = 0,
equivalently, f'/f = x/c
provided c is not 0, in which case f will have the form
Ke^{x^2/2c}
For T to be defined on this (removing the restriction of compact support),
you need Re(c) < 0.
I'm not terribly sure how to obtain the eigenvalues directly, though.
That looks pretty direct to me. Note, however, that you may also want to look
out for continuous spectrum. I think the spectrum of this unbounded operator
consists of the whole complex plane.
--
Robert Israel israel@xxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Department of Mathematics http://www.math.ubc.ca/~israel
University of British Columbia Vancouver, BC, Canada
.
- References:
- > Eigenvalue problem for sufficiently well-behaved functions
- From: k.hofmann
- > Eigenvalue problem for sufficiently well-behaved functions
- Prev by Date: Re: CERN concern
- Next by Date: Re: Out-of-print math books: An Update
- Previous by thread: > Eigenvalue problem for sufficiently well-behaved functions
- Next by thread: partition function
- Index(es):
Relevant Pages
|