Weyl's formula for a class of pseudodifferential operators with negative
Recommend Documents
Oct 4, 2007 - which commutes with every pseudodifferential operator with symbol F(x+Jξ), ... which induces the norm ||f||2 = ||ãf,fã||1/2, and we denote by En its Banach space ... The elements of H are the smooth vectors for the action of the ..
May 9, 1994 - operators is carried over to matrix symbols and to operators acting on vector- ... implying a special sort of calculus of operator-valued pseu-.
CHARACTERIZATIONS OF PSEUDODIFFERENTIAL. OPERATORS ON THE CIRCLE. SEVERINO T. MELO. (Communicated by Palle E. T. Jorgensen). Abstract.
poO~qo, k
Compact magnetic pseudodifferential operators. Viorel Iftimie. To Professor Ion
Colojoar˘a on the occasion of his 80th birthday. Abstract - In previous papers, ...
Feb 28, 2018 - (MÏTx)(MÏâ² Txâ² ) = eâ2ÏixÏâ². MÏ+Ïâ² Tx+xâ² . Let f â Sâ²(Rd). We define the short-time Fourier transform of f as. (8). Vgf(x, Ï) = ãf,Ï(x, ...
Apr 14, 2018 - First, we give a new proof for the Beals commutator criterion for non-magnetic ... by their commutation properties with fundamental objects like ...
tization of a given cont act transformation $\phi$ is an extension of $\phi$ to a ring ... $A(x_{i} \xi)\circ B(x, \xi)=\Sigma\frac{1}{\alpha!}\partial_{\zeta}^{\alpha}A(x, ...
Oct 10, 2007 - Vgf(x, ξ) = â«Rd f(t)g(t â x)eâ2Ïit·ξ dt. = ãf,MξTxgã = ãf,Ï(z)gã . The short-time Fourier transform of a symbol Ï(x, ξ), for x, ξ â Rd, is a function on ...
Dec 20, 2011 - DG] 20 Dec 2011. PSEUDODIFFERENTIAL OPERATORS ON MANIFOLDS WITH. FIBRED CORNERS. CLAIRE DEBORD, JEAN-MARIE ...
ABSTRACT This chapter illustrates the use of Gabor frame analysis to derive results on the spectral properties of integral and pseudodifferential operators.
Now, the Lie algebra of vector fields on the circle is included naturally ... Our aim in this work is to study central extensions of Lie algebras of formal ... technical and it requires a very careful and critical reading, and so we ...... extension
operators on M0. An important ingredient in Melrose's program is to define a suitable pseudodifferential calculus Ψ. â. V. (M0) on M0 adapted in a certain sense.
Nov 22, 2000 - Introduction Analytical index formulas play an important part in the study of topological characteristics of elliptic operators. They complement ...
Nov 29, 2015 - AP] 29 Nov 2015. Characterization of Non-Smooth Pseudodifferential. Operators. Helmut Abels and Christine Pfeuffer. December 4, 2015.
arXiv:hep-th/9410016v1 4 Oct 1994. Göteborg ITP 94-23 .... The Schwinger term for ΨDO's in glres lies in the same cohomology class as the so-called. âtwistedâ ...
theorem. Our approach is based on an unpublished paper by Melrose and Nistor âHomology of pseudo-differential operators I. Manifolds with boundaryâ [39].
Mar 29, 2017 - Vgf(z) = F[fTxg](ξ) = â«Rd f(y) g(y â x) eâ2Ïiyξ dy for z = (x, ξ) â Rd à Rd. For description of decay properties, we use weight functions on the ...
Jan 8, 2008 - noninteger order ΨDOs is called the canonical trace. This is a ... that any non-integer order (resp. parity class) ΨDO is a sum of commutators of.
Jun 19, 1998 - We show that if a symbol Ï lies in the modulation space ... in the decomposition of the operator or its symbol into simpler parts that are easier to.
quantizes V in Melrose's sense. We give some concrete Fredholmness conditions for pseudodifferential operators in 묉. V. (M) for a large class of Lie manifolds.
Jun 22, 1993 - operators on L^^X}. We will treat all the elements of this space as "pseudodifferential operators" and we will define their symbols. Let A e S\X} ...
Apr 30, 2010 - The key to these constructions is the notion of a bi-submersion, which we are going to use here as well. In a broad sense this may be thought of ...
Weyl's formula for a class of pseudodifferential operators with negative
OPERATORS WITH NEGATIVE ORDER ON L2(R n). Monique .... for a class of hypoelliptic operators (3.12) and we shall deduce from that a Weyl's formula of ...
WEYL'S FORMULA FOR A CLASS OF PSEUDODIFFERENTIAL OPERATORS WITH NEGATIVE ORDER ON L2(R n) Monique Dauge and Didier Robert U.A.C.N.R.S. 75g, D~partement de Math~matiques et d'lnformat~que 2, Rue de la Houssini~re F. 44072 NANTES C~dex, FRANCE
Introduction The starting point o* this work is a paper by Birman and Solomjak [BI-SO I] in which they study the eigenvaiues asymptotlcs of a class o[ integral compact operators on L~(Rn). These operators are pseudodifferential, with symbols that are homogeneous (or quasi-homogeneous) of negative order with respect to the phase variable ; there are very few regularity assumptions with respect to the space variable. If such an operator A is sel:[-adjoint, (X~ (A)) denoting the positive eigenvalues decreasing sequence of A, [BI-SO I] gives an equivalent to the X+ (A), by powers of /
j when j + + o~ , and the corresponding result for negative eigenvalues (cf § 2.A). The method of [BI-SO t ] is based on Courant's mini-max principle and consists in reducing to a model problem on the torus'[ 'n. In this work, we intend to extend and precise [BI-SO 1] results in two directions : (l) relaxing the homogeneity assumption (2) getting a remainder estimate. As the symbol may vanish and change of sign, there are some difficulties we will overcome by making
an hypoelliptic regularization of the operators and then using
a Mellin functionnal pseudodifferential calculus. Here is the structure of our paper §l : Assumptions and the main result. §2 : Special cases and examples. §3 : Further information about spectral theory of globally hypoeiliptic pseudodifferential operators on R n, §4 : Proof of the main result. Annex A : Weyl-Ky Fan inequalities. Annex B : Composition formula with precised remainder. Annex C : Computation for an example.
92
The main results of this paper have been announced in the authors'
note [ D A - R O ] .
§1 : Assumptions and the main r e s u l t
Our operators are described by the properties of their Weyl's symbols, according to L. H6rmander's
formalization
[HO
l]. For the sake of s i m p l i c i t y ,
we r e s t r i c t our-
selves to the diagonal metrics. L e t us recall some d e f i m t i o n s . On T -~ R n, i d e n t i f i e d to R 2n : Rnx x R ~n , let qb and cp be two w e i g h t f u n c t i o n s ,
with thezr values in ]0,+oo[. (1.1) D e f i n i t i o n L e t m be a w e i g h t f u n c t i o n , m : R 2n ÷ ] 0 , + m [ . (1)
If t h e r e exists C O , C 1 > 0 such t h a t : [y[ ¢(x,g) + Inl ~(x,g) , 0
on
R2n
such t h a t K' m Y' , ) ~
-
For the functions a decreasing function
dx 0' YI ' Y2 >0 such that : f
I is derivable on ]0,t 0]
(T) ~ ¥ 1 f(t) 0 : ql a 4
(ii2)
3 61,82 > 0 : .~
IVal
w i t h (it 2) and (2.3), t h a t yields c o n d i t i o n (T) a b o u t V+0~ ; a). Note t h a t
if a is quasi-homogeneous, with weights k t on x i and h i on
~j , we
may take : F(x,~) = -(k I x I , ... , kn xn ; h I ~I ' "'" ' hn ~n)
2.C : S c h r ~ d i n g e r o p e r a t o r s L e t us c o n s i d e r : A -- (-A+E) 0/2 V(-A+E) °/2 w i t h A t h e L a p i a c e ' s o p e r a t o r on R n, E >0, o
(2.#)
< 0 , and t h e p o t e n t i a l V such t h a t t h e r e is s < 0 :
!a~ V(x)[ .R i=l i i x i
In t h a t case, we have : N+& ; A) = V+(~ ; a 0) + 0(~ ~ fs,o(X)) where fs,o00 = ~n/2s if s < o and fs,a(~) = t n / 2 s L o g t i f
s = O.
As an a p p l i c a t i o n , tet us c o n s i d e r t h e s t a t i o n n a r y p r o b l e m for S c h r 6 d i n g e r e q u a t i o n : (2.6)
(A+gV) ~= C~
where E > 0 is the energy and g e R is a coupling constant. For a fixed E, we search values of g for which (2.6) has a non-null solution ~F in D(A)
n D(V). For V v e r i f y i n g
(2.#) and g ~ 0, (2.6) is e q u i v a l e n t to (2.7) ~F ~ LZ(R n) \ {0} , (-D+E) -1/2 V(-A+E) -1/2 ~= _l ~F g
It is t h e a b o v e case,
with
a = - [ . Thus, (2.6) has n o n - t r i v m l solutions for t w o
s e q u e n c e s (gk+ (E)) and (gk (E)) w h e r e ( gk+ (E)) is positive) i n c r e a s i n g and (gk (E)) is n e g a tive,
decreasing.
If i n f i n i t e ,
each
of
those
sequences
a r e n o t bounded.
(2.5), w i t h :
N+(g,E) = # {k/g k (E) >. -+g} we have, when g + + oo :
(g >0)
I~ V v e r i f i e s
97
(2.g)
N+(g,E) - _
(21I)-nn l'n ;
(g V(x)-E)+n/2- dx + 0 ( g - % I ( g ( l + l x t 2 ) Q E ) n / 2 dx
where Y is the volume of the unit bali of R n. n Many works have given formulas of (2.8) type ([SI],[MA]). 2.D : Equation A~ = ~B¥.
L e t A = Op W a and B = Op W b be two s e l f - a d j o i n t operators on L2(Rn), possibly unbounded w i t h domains D(A) and D(B). With (Gq0) weight functions v e r i f y i n g ( H I ) ~ (H2) , we suppose that :
A is positive and invertibte,
(i)
a = a 0 + a t where a 0 is a t e m p e r a t e weight such t h a t a 0 ~ S(a 0 ; O,O and a 1 ~ S(ao(~q~) -~ ; Gq~) with E > 0
b e S(q ; c),qo) where q is a t e m p e r a t e weight such t h a t (ii)
q ~ S(q ; #,q~) ; -1 m :: a 0 q v e r i f i e s the hypotheses (W) and (N) • And we consider the s p e c t r a l problem :
(2.9)
A T = ~,B~Y ,
~
D(A) r i D ( B ) \ { 0 }
.
By mean of c o m p o s i t i o n by the o p e r a t o r C : A -I/2, (2.9) is e q u i v a l e n t to :
~] = I C B C ~
,
$ ~ L2(R n ) \ { O }
Thanks to (n), t h e o r e m (1.3) can be appiied to C B C, and if the volume functions satisfy the condition (T), we obtain, for a
¢ >0, when X + + ¢o •
(2.10) N+(I ; A,B) = (2K)-n ] ]
dx de + 0(~-¢ a(x,~)< + X b(x,~)
where N+~
jj
dx d¢)
a(x,~)< X q(x,~)
; A,B) (resp. N (~ ; A,B)) is the n u m b e r oi eigenvalues of (2.9) belonging
to ]0~l] (resp. l-l,0[). Many papers are devoted to Weyl's formulas of (2.10) type, w l t h various hypotheses (see [BI-SO 3] and [ F L - L A ] ,
as w e l l as their bibliographies). Most of them suppose that
A is an elliptic d i f f e r e n t i a l o p e r a t o r and B is the multiplication by a function O. in particular,
for Fleckinger and Lapidus in [FL-LA], A is of Schr6dinger's type and 0
may be discontinuous (for instance~ the c a r a c t e r i s t i c function of a c o m p a c t set). However, those works generally only give an equivalent. L e t us note t h a t we find t h e same result as the one announced by Boitmatov and Kostyuchenko in [BO-KO] under assumptions t h a t seem to us more general and more natural.
g8 §3 F u r t h e r i n f o r m a t i o n
about spectral theory of globally hypoelliptic pseudodifferential
o p e r a t o r s on R n. 3.A : Introduction
Let (~,~0) be weight functions satisfying (HI). Let p be a symbol satisfying the condition (W), i.e : p ~ S(p ; O~c~), and the further condition (S) which is inverse of the condition (N) : (S) 3 C , C' >0, ~ ~' >0 such that C p6< ((~c~) ~~y>0. 3.B : Functionnal calculus : parametrix. For z ~ C \ R+ , we have to study (P-z) -I pK with K positive integer (the reason of the introduction oi this parameter
K will appear
in section 3.C). Thus, we build
a parametrix for the equation : (3.1) V ~(K) z o (P-z) = pK ,
(3.2)
As a first approximation, we get the symbol : (K) = pK (p_z)-I qz ; 0 Then, by recurrence over j, we define the symbols :
(3.3)
(K)
qz ; j
=
(p_z)-I [p(K) - j~l J
where F(cc,g) = ((~ ~ BI) -1
k=0
Z
r(~,6)(a~
0~+8--j-k
2-1~1 (- 2)-1~1
Dg x
p)(a~ o ~x q z ;(K) k)]
and pj(K) is the jth term in the asymptotics of
the symbol of p K More precisely, pj(K) ~ s(pK(~bq) )-j ; ~b~c~)and is a polynomial expression of the ~
x p for I~.g[. i , we get :
J
= ~ (_])k d(K) k=0 jk (P-z)-k-I
(K beiongs where the djk) are universal polynomial functions of the 3( 36x p for [~+6[~~2 k~ Tr(Pk ~ Z Op W Pk ~) f(k)(0) ~:0 We have showed that the contribution for k:2 is zero and that the others are generally non zero.
104 We can also derive a trace formula for short range Schr6dinger operators, which may be compared with the results of [CVJ and [GU]. Let V ~ C~°(Rn) be the potential, such that there is s X ;
f~,0(p)= { oo
g~,,O ~ CO (30~,0)) and g)~,O --I on IOn,O) ; (3.23) ~/k e N, [(p 3p)k f~,0(l~)l+l(p ap)k gx,e(u)l < c k x ke . Then~ we have : (3.24) IN~ ; P) - Tr fk,0(P)[ .< Tr li0,,0) (P) (3.25) Tr li(k,O) (P) ~
1, t h e r e is pj ~ S(p(~ q0)-j ; qb, £0) such t h a t for e a c h N >~1 : N Z pj) = Op w r N P0 - OpW(p + j=l and e a c h
semi-norm
with
ol a pj d e p e n d
r N ~ S(p((~q0) - N - I ; do, ¢p)
only on a f i n i t e n u m b e r of s e m i - n o r m s
otb
and
b I - it is t h e m e r e c o n s t r u c t i o n of a p a r a m e t r i x Ior B 0. Thus, we a r e in a slightly m o r e g e n e r a l is w r i t t e n
as an a s y m p t o t i c
expansion
(just
s i t u a t i o n t h a n in §3 : t h e s y m b o l of P0 like c l a s s i c a l
symbols).
Nevertheless,
the
whole r e s u l t s may be a d a p t e d , by s t a r t i n g f r o m ; (K)
: (p_z)-I [p(jK)
qz ;j
- j -EI
k:O
Ej ~=0
E
Io~+6]=j-k-.~
F(c~,6) (3~ D B p~) (~ x
D c~ q(K) k) x
z ;
i n s t e a d of (3.3) (with P0 = p) Then, i n s t e a d of (3.22), we a r r i v e to :
[R(x
;
p0 ) - v(X,p)[
4 ( 2 . C 1 X0-6)[V(~.),I-0 ~ p)-~(x-xl-e
w i t h C 1 and C 2 depending only on a f i n i t e
;
p)]+C 2
number of s e m i - n o r m s of p , p t , . . . , p N ,
where
N is a fixed integer. Now, l e t us r e m a r k that, for 0 4 00 , ),8-6 may be r e p l a c e d by )-0 , and that, for any f i x e d c o n s t a n t c > 0, V(k+k 1-0 ; p) - V(' X'1-0 ; p) may be replaced by ~(X+c2, i - e ; p) _ V(X-ck 1-0 ; p).
On the other hand, we have, for la = k -1 : N(~ ; a O) = R(X ; PO) ; V(p ; b) : V(X ; p) and V0, -+ cA 1-0 ; P) = V(p(l+-cp0) -I ; b) We c h o o s e c=20-1 , and we h a v e
(l+clag) -1 > 1-10 (1-Cla0) -1 < l+k 0.
F r o m all t h a t , we d e d u c e (4.9). a L a s t l y , h e r e is a r e s u l t looking like G a r d i n g ' s inequality, It c o n s i s t s of a r e l a t i o n between
symbols positivity
and o p e r a t o r s
positivity
(that relation
is n o t s y s t e m a t i c
as
109
i t is in a n t i - w i c k q u a n t i z a t i o n - see [TU-SU]). (#.I0) L e m m a L e t b • S(m ; (b,~0)be a positive symbol. L e t be e > O. Then, there is d e ~ S(m(~ cO)-1 ; ¢,fl0) such t h a t : 02.
¢ >0, we denote, a e = a x(a/em) + emil - x(a/em)] •
We then have the obvious f o l l o w i n g properties :
(#.11) a e ~ S(m ; ¢@) (#.12)
ae
~
em
(#.13) a e co]'ncides w i t h a in the region {a >~2 em} (#.I#)