[講義] RMT2018: Lecture 14
Retracing our steps we see that for in region , say, of 211.x,
載入超時,點擊重試
(219.1)
and we obtain, in particular, the asymptotics of as . There are, of course, analogous formulae for in the other 3 regions.
Remark 219+
A RHP can be converted into a problem of singular integral equations on in the following way.
Remark 219+
We see here a fundamentally new phenomena for the non-commutative steepest-descent method that is not present in the classical case: In the classical case, the leading order contribution to the integral comes from the stationary phase point, or a finite # of stationary phase points. But in the non-commutative case, we see that the leading order contribution to the RHP comes from a whole continuum of points, viz, .
Recall . Let
i.e. where is the Cauchy operator for . Then for the contours, we are considering (a finite union of smooth arcs, with a finite # of points of self-intersection), and certainly for far more general contours (Carleson curves!), we have for
almost everywhere as non-tangential limits.
Moreover, .
Also,
, almost everywhere .
(Supplemental information on RHPs in the sense of )
Remark 220+:
In order to make the asymptotic calculations for the RHP valid, we must define what it means to solve a RHP in the sense of .
(see, for example, P. Deift and X. Zhou CPAM 56, 2003, #8, 1029-1077, or ArXiv 0206224)
Let .
We say a pair of (matrix-valued) function in belong to if for some .
Here are the boundary values of the Cauchy transform on as before. Note that
so is uniquely determined by .
We say that solves the normalized RHP in the sense of if ,
(220+1.1)
for some , and
(220+1.2)
almost everywhere on .
We say that
is the (analytic) extension of off . Note that
- is analytic in ;
- almost everywhere on ;
- (in some sense) as .
so that above produces a solution of the RHP in the "natural" sense.
In order to solve the RHP with and , we proceed as follows: Let solve the equation
(220+2.1)
on .
Here and is the function identically on . Clearly .
We always assume is Carleson so that are bounded in . Hence .
More precisely, if we write , then (220+2.1) takes the form
(220+3.1)
which is an equation in . To solve (220+3.1), and hence (220+2.1), we need to know that
(220+3.2)
exists in .
Under this assumption, set
(220+3.3)
.
Now, using (220+2.1),
(220+3.4(i))
and
(220+3.4(ii))
and hence
(220+3.5)
.
But as
,
we see from (220+3.3) that and together with (220+3.4), we see that solve the normalized RHP in the sense of .
Thus the analysis of the RHP boils down to the analysis of the (singular integral) equation (220+2.1).
Now suppose that we have a sequence of normalized RHPs on a fixed contour , and suppose that as , at least pointwise almost everywhere on . How much more do we need to know about the convergence to ensure that the solutions of the RHPs converge to the solution of the (normalized) RHP ?
From (220+3.4(i)) (220+3.4(ii)), we have
;
.
So to obtain convergence in , we would certainly need to know that
(220+5.1)
in .
As , we have , so if
(a) in and by Dominated Convergence;
(b) .
Now from (220+3.1),
.
Together with the above assumption (a), we have
in
if
(c) .
In order for the second term to go to zero
in ,
we need, in addition to (c),
.
But
and so if
(d) .
Standard manipulations show that if
and , then for n sufficiently large. Assembling the above calculations we see that in order for in , we need
(220+7.1)
and
in .
This is why we cannot conclude that in Lecture 13! We do not have as .
Assume as before that and in addition,
(221.0)
.
Suppose that
(221.1)
is invertible in
and let
(221.2)
solve the equation
(221.3)
or more precisely,
.
Then
(221.4)
solves the normalized RHP . (Clearly is analytic in .) Indeed, for ,
.
Also .
Hence
.
Also, clearly, in some appropriate sense,
as ,
by (221.4). Thus solves the normalized RHP .
For , we see that
.
So we see now clearly that we would need in - to conclude that ! We know in , and point-wise almost everywhere on , but not in !
(For more details, see Remark 220+)
In order to show , we need more information about what is happening at the points . We proceed as follows.
Extend the RHP for on (212.x), to a RHP by adding in 2 small circles around
and set
(223.0)
Thus
for
and
for .
We then construct explicitly a parametrix with the following properties:
(223.1)
for outside the "dumbbell" region
(224.1)
where denote the interior of respectively:
(224.1)
where .
Note that on and so (224.1) is a condition on the construction of for .
It follows, in particular from the above that solves a normalized RHP where
Thus and have precisely the same jumps on . Hence solves a normalized RHP with jumps on
On
as is analytic across these arcs. But as points on these arcs are at a finite distance from , is exponentially decreasing, and hence the same is true foe .
On the other hand, on
, as is analytic across these arcs, and as , the same is true for .
It follows then that
and so
in
i.e.
.
As is known explicitly from the construction, this provides the asymptotics of as , and hence of , by (219.1).
We will show how the construct (using Airy functions!) in , and make the above argument precise, in the next lecture.
推薦閱讀: