The next application I want to talk about here of Fourier analysis is to (a basic case of) ellipic regularity. Later we will use refinements of these techniques to obtain all kinds of estimates. Anyway, for now, a partial differential operator
is called elliptic if the homogeneous polynomial
has no zeros outside the origin. For instance, the Laplace operator is elliptic. Later I will discuss how this generalizes to other PDEs, and how this polynomial becomes the symbol of the operator. For the moment, though let’s define . The definition of
such that
and we know that for
large enough. This is a very important fact, because it shows that the Fourier transform of
exerts control on that of
. However, we cannot quite solve for
by dividing
by
because
is going to have zeros. So define a smoothing function
which vanishes outside a large disk
. Outside this disk, an estimate
will be assumed to hold.
A parametrix for
We’re going to start by finding a parametrix for the operator ; this is not quite a fundamental solution, but it is close enough. We could formally get a fundamental solution (since
) by considering
, but this is nonsense. Rather, consider the distribution
with
which is indeed well-defined as a tempered distribution by the hypothesis on . It is then clear that
, so
where the is at least suitably controlled (e.g. in
). So we have something close to a fundamental solution, and
is called the parametrix. Since we convolve things with a Moreover, I claim—and this is crucial—that
comes very close to being a function of
as well; the only problem occurs because of the not-that-fast decrease at infinity. In particular, the singular locus of
is the origin. This will require some discussion of intermediate concepts.
How to convolve two distributions
We already know how to convolve a distribution and a function
. Take
defined by
and set
It is easy to check that this coincides with the old definition of convolution when . This is always
. When
is compactly supported, which is to say that
when
vanishes inside a sufficiently large compact set
, then we have
for some . In this case it follows that
in fact. So given another
, not necessarily of compact support, we can heuristically use the properties of convolution to “write”
Make this a definition. Then in particular, we have a way of talking about as a distribution in itself—just evaluate the convolution at
.
Let’s go back to what I just said about compact support. We can generalize this: say that a distribution vanishes on an open set
if
for
supported in
. Then a partition of unity argument shows that there is a largest open set
on which
vanishes; the complement is called the support
. This is a generalization of the notion for functions, as is easily seen. Anyway, it is a well-known fact about functions that
. That this is true for distributions when one is compactly supported follows by regularizing each: given
, we convolve each with an approximation to the identity to get smooth functions, one of which is compactly supported, that approxiamte
arbitrarily closely (in the weak* topology).
The singular locus of a distribution
Say that a distribution is regular in an open set
if for
smooth and supported in
, we have
for
smooth. Basically, this means that when restricted to
,
behaves just like a smooth function.
Using a partition of unity, we see that if is regular in
and
, then it is regular in
, and moreover for infinite unions. In particular, there is a maximal open set on which
is regular. The complement of this set is written
. For instance,
. This behaves well with respect to convolution, if one is compactly supported :
Lemma 1
The reason is that we write where
have supports barely outside the singular loci of
and
are smooth. Then
and all but the first term are smooth. The first term is supported in a small neighborhood of , which we can make arbitrarily small.
January 17, 2010 at 4:21 pm
[…] Climbing Mount Bourbaki Thoughts on mathematics « Towards constant-coefficient elliptic regularity: parametrices and generalities on the singular … […]