ON THE REGULARITY OF SOLUTIONS TO THE BELTRAMI EQUATION IN THE PLANE
|
|
- Justina Bruce
- 6 years ago
- Views:
Transcription
1 ON THE REGULARITY OF SOLUTIONS TO THE BELTRAMI EQUATION IN THE PLANE by Bindu K Veetel B. Sc. in Mathematics, Bangalore University, India M. Sc. in Mathematics, Bangalore University, India A dissertation submitted to The Faculty of the College of Science of Northeastern University in partial fulllment of the requirements for the Degree of Doctor of Philosophy April 17, 2014 Dissertation directed by Robert McOwen Professor of Mathematics
2 Acknowledgements First of all, I would like to thank my advisor, Prof. Robert McOwen, with whom I had a great learning experience. His encouraging advice and guidance kept me motivated through the ups and downs of research and thesis work. Thank you for the patience and understanding throughout the course of this research work. Thanks to Prof. Maxim Braverman, Prof. Gordana Todorov, Prof. Jerzy Weyman, Prof. Alex Suciu, Prof. Christopher King and Ms. Rajini Jesudason for all the help and guidance that enriched my learning and teaching experiences. I am also thankful to all the professors who helped me better understand various aspects of Mathematics through the graduate courses offered. Special thanks to Gabriel Cunningham for all the discussions that made my initial years of research work enriching, easier and fruitful. I would also like to thank all the faculty members, office staff and all my friends for making my journey through PhD a pleasant one. I am grateful to my parents and brother for their continuing support and love throughtout these years. None of my efforts through these years would have been fruitful but for the relentless support, patience and encouragement of my family- my husband Saji Nair and my sons Nandan and Nikhil. ii
3 Abstract of Dissertation We obtain estimates on regularity of solutions to the inhomogeneous Beltrami equation in the plane namely f µ f ν f = h when h and the coefficients µ and ν are Dini continuous; the coefficients satisfy an ellipticity condition µ(z) + ν(z) κ < 1. In the case when h and the coefficients µ and ν are Hölder continuous with exponent α, 0 < α < 1, it is well known (cf.[3]) that the solutions and their first order derivatives are Hölder continuous. In our case, we find that although the solutions are in C 1, their derivatives are less regular than the coefficients. Nevertheless, we are able to get a priori estimates. One application of this work is to stability analysis of the inverse problem of Calderon in two dimensions. iii
4 Table of Contents Acknowledgements ii Abstract of Dissertation iii Table of Contents iv 1. Introduction 1 2. Dini Continuous Functions 9 3. Constant coefficients in bounded domain The space C (λ) on the torus Non constant coefficients in a bounded domain 53 References 57 iv
5 1. Introduction In this thesis, we study the regularity of solutions to the inhomogeneous Beltrami equation in the plane, namely (1) f µ f ν f = h, where = z = 1 2 ( x + i y ), = z = 1 2 ( x i y ). We assume the ellipticity condition, µ(z) + ν(z) κ < 1 for all z Ω C where Ω is a bounded domain in C. When the coefficients µ and ν and the function h are Holder continuous with exponent α (0, 1) i.e. in the Holder space C α (Ω), it is well known (cf.[3]) that the first order derivatives of a solution to the Beltrami equation (1) also lie in the space C α (Ω). In this thesis, we consider the coefficients µ and ν and h to have modulus of continuity ω satisfying the Dini condition ɛ 0 ω(r) r dr <. The result that we obtain is that the first order derivatives of a solution of (1) will have modulus of continuity σ that could be weaker than ω, in particular, need not satisfy the Dini condition. Nevertheless, we are able to obtain interior estimates of the form (2) f C σ (D) + f C σ (D) C( h C ω (U) + f C 0 (U) ) where the constant C depends on µ, ν and the domains D, U where D is compactly contained in U and U is compactly contained in Ω. (In (2), we have used C ω to denote functions continuous with respect to the modulus of continuity ω, even though this is inconsistent with the notation C α for Hölder continuity; cf. Section 3). The techniques we use are mostly based on Fourier analysis. Littlewood-Paley theory and 1
6 Fourier multiplier operators on the torus are used to describe and work with functions that we encounter. Our main result is given as Theorem 5 in section 6. One of the applications of (2) is to stability of the inverse problem of Calderon in two dimensions. An inverse problem assumes a direct problem that is well-posed: a solution exists, is unique and stable under perturbation of parameters. It uses information about the solutions to obtain a parameter in the problem. To describe the Calderon problem, let u be the unique solution to the Dirichlet problem (3) γ u = 0 in a bounded domain U R 2, u = f on U. If we let Λ γ (f) = γ u ν then the map Λ γ where ν is the exterior unit normal on U, : H 1/2 ( U) H 1/2 ( U) is called the the Dirichlet to Neumann map. If γ represents the electrical conductivity of U, then Λ γ is the current induced at the boundary by applying a voltage f at the boundary of U. Hence experimental data involving current and voltage measurements at the boundary determines Λ γ. The inverse problem of Calderon is the determination of conductivity γ in U from the boundary measurements, i.e. from Λ γ. In particular, solving the inverse problem requires showing that the Dirichlet to Neumann map Λ γ uniquely determines γ. Stability is a stronger statement than 2
7 uniqueness; it says that (4) γ 1 γ 2 L V ( Λ γ1 Λ γ2 ), where denotes the operator norm H 1/2 ( U) H 1/2 ( U) and V (ρ) is a stability function that satisfies V (ρ) 0 as ρ 0. Hence stability implies that small changes in the Dirichlet to Neumann map will only correspond to small changes in the conductivity. This inverse problem has important real world applications like in medical imaging to detect tumors in the human body, reconstruction of the interior of human body using exterior measurements. L Astala and Päivärinta [5] showed that Λ γ uniquely determines γ by reducing the problem to a Beltrami equation. In fact, they recover the complex solutions of (3) by letting (5) u γ = R(f ν ) + i I(f ν ) where f ν is the complex geometric optics solution of the R-linear Beltrami equation, (6) f = ν f, and ν is defined in terms of conductivity γ as ν = 1 γ. Notice that the 1+γ above equation is a special case of (1) where the coefficient µ and the function h are zero. In [5], it is shown that for each k C, there exits a unique solution f ν (z, k) of (6) of the form f ν (z, k) = e ikz M ν (z, k) 3
8 where M ν (z, k) = 1 + O(z 1 ) as z. In order to control the behaviour of M ν as z, [5] shows that f ν (z, k) = e ikφ(z,k) where for each fixed k, φ satisfies a Beltrami equation. The work of Barcelo, Faraco and Ruiz [4], stem from that of Astala and Päivärinta in [5]. They use regularity results for (6) to obtain stability in the case of Hölder continuous conductivity. In particular, they obtain that when the coefficients µ, ν and the function h are Hölder continuous with exponent α (0, 1), any solution f in the Sobolev space W 1,2 (Ω) satisfies the estimate f C α (D) + f C α (D) C f C 0 (U) where the constant C depends on µ, ν, α and the domains D, U such that D is compactly contained in U and U is compactly contained in Ω. This estimate is then used to obtain regularity results for (6). The interior estimate (2) obtained in this thesis can be used to obtain corresponding estimates needed for stability of the complex solutions of the inverse problem of Calderon in two dimensions in the Dini case. This will be further discussed in the upcoming paper [14]. The basic structure of our work is as follows: In section 2, we define Dini continuous functions and Banach spaces associated with these functions. In section 3, we obtain the regularity of solutions to the 4
9 inhomogeneous Beltrami equation with constant coefficients, namely (7) f µ 0 f ν 0 f = h where h is Dini continuous. This is done first for the C-R equation, the special case of the Beltrami equation where the coefficients µ and ν are zero, and then translated to that of the Beltrami equation (7). But these estimates do not easily generalize to variable coefficients. Hence in section 4, we follow [15] by introducing a space of functions C (λ) on a torus using Fourier multipliers and a Littlewood-Paley partition of unity; here the function λ is defined in terms of modulus of continuity ω. These functions need not be Dini continuous but are well behaved under the Beurling transform. We first obtain regularity estimates in terms of C (λ) on solutions of the Beltrami equation with constant coefficients µ 0 and ν 0 but with h C (λ). We then use these estimates to obtain similar estimates on solutions of the Beltrami equation (1) with variable coefficients in C (λ). We make use of tools like the Beurling transform, Fourier transform and paraproduct. In section 5, we translate our results on the torus to regularity estimates for (1) with Dini continuous coefficients in a bounded domain Ω. The following are some terms used in our work: Modulus of continuity: It is a continuous, increasing function on [0, ), usually denoted by ω such that ω(0) = 0. The definition of modulus of continuity was first given by H. Lebesgue in 1910 for functions of one real variable. It is a 5
10 precise way to measure smoothness of a function. Beltrami Equation: Beltrami equations play an important part in complex function theory and theory of differential equations. Various theorems like the measurable Riemann mapping theorem can be proved using the Beltrami equation. The first global solution in C of the Beltrami equation of the form f = µ f was given by Venkua [17] for compactly supported µ. The L p properties of the operators associated with the Beltrami equation have been used to prove the measurable Riemann mapping theorem in [3]. Littlewood-Paley partition of unity: This is a tool in analysis which allows us to decompose a function into pieces such that the frequency supports of these pieces are almost disjoint. On R n, it can be defined as a collection {ψ k (ξ)} of smooth functions with compact support such that for k > 0, each ψ k (ξ) has support in 2 k 1 < ξ < 2 k+1, ψ 0 (ξ) has support in ξ < 2 and k=0 ψ k(ξ) = 1. Fourier Transform: Let S(R 2 ) denote Schwartz space i.e. the space of smooth functions such that all derivatives decay rapidly at infinity. The Fourier transform F : S(R 2 ) S(R 2 ) is defined by (8) (Ff)(ξ) = ˆf(ξ) = 1 f(z) e iz.ξ dz for ξ R 2 (2π) 2 R 2 6
11 and the inverse Fourier transform F 1 g is given by (9) (F 1 g)(z) = ǧ(z) = g(ξ) e iz.ξ dξ. R 2 If S (R 2 ) denote the dual of the Schwartz space, i.e. the space of tempered distributions, the Fourier transform can be extended to F : S (R 2 ) S (R 2 ) by defining (10) û(φ) = u ˆ (φ) for φ S(R 2 ). If f, g L 1 (R 2 ) and one of them has compact support, then (fg) (z) = (f g )(z) = f (z z)g ( z)d z. R 2 Fourier Multiplier Operator: Using the Fourier transform, we can define Fourier multiplier operators M(D) for any bounded, measurable function m(ξ) as (M(D)f) (ξ) = m(ξ) ˆf(ξ). The function m(ξ) is called its symbol or multiplier. Examples of Fourier multiplier operators include differential operators. Fourier multiplier operators are a special case of the pseudodifferential operators which are extensions of the concept of differential operators and that of singular integral operators and are used extensively in the field of partial differential equations and quantum field theory. Paraproduct: According to S. Jansen and J. Peetre [11], The name paraproduct denotes an idea rather than a unique definition; several definitions exist 7
12 and can be used for the same purpose. The term paraproduct means beyond product. It can be considered as a bilinear, non-commutative operator that is used to reconstruct a product such that it generally has better properties than the usual product of functions. The first version of a paraproduct appeared in A. P. Calderon s work on commutators [7]. Several versions have appeared since then. J.-M. Bony introduced a paraproduct in his work, Theory of paradifferential operators [6], which came to be known as Bony s paraproduct. It is defined using a Littlewood-Paley partition of unity and helps in reconsructing a product fg based on the frequency supports of f and g. It separates out the parts where frequency supports of f and g are disjoint and hence enables analysis of each part of the product appropriately. The part where the supports are not disjoint is considered as the error term and hence plays the role of the best part of the product reconstruction. Smoothing operator: It is an operator that maps non smooth functions or even distibutions to smooth functions. Any Fourier multiplier operator ψ(d) L 1 (R n ) with compact support can be considered as a smoothing operator. The Fourier multiplier operator ψ 0 (D) associated with the Littlewood-Paley partition of unity {ψ j (ξ)} is an example of a smoothing operator used in this work. Fredholm operators: A bounded linear operator T : X Y between two Banach spaces X and Y is said to be Fredholm if it has finite dimensional kernel and 8
13 cokernel. The index of the Fredholm operator T is given by ind(t )= nul(t ) - def(t ) where nul(t ) is the dimension of kernel of T and def(t ) is the dimension of cokernel of T. The composition of Fredholm operators is also Fredholm with index of the composition equal to the sum of index of each operator. If K : X X is a compact operator, then the operator T = I + K is Fredholm with ind(t )=0. A bounded linear operator T : X Y between two Banach spaces X and Y is said to be semi-fredholm if the range of T is a closed subspace of Y and atleast one of kernel T or cokernel T is of finite dimension. It is well known that if an operator T satisfies an a priori estimate of the form f X C( T f Y + f Y ) where X is compactly contained in Y, then T has a finite dimensional kernel and closed range, i.e. is a semi-fredholm operator. 2. Dini Continuous Functions Consider a function ω with the following properties: ω(0) = 0, ω(r) is continuous and strictly increasing for 0 r 1 and satisfies the Dini condition (11) ɛ 0 ω(r) r dr < for some ɛ > 0. For technical reasons, also assume (12) 0 < ω(r) 1 and that (13) r β ω(r) is decreasing on (0, 1] for some β (0, 1). 9
14 When convenient, we extend ω to r > 1 by letting ω(r) = ω(1). This extension of ω will still satisfy (13) on (0, ). As a consequence of (13), it is easy to see that ω(2r) 2 β ω(r). Hence there exists constants C 1 and C 2 such that C 1 ω(2r) ω(r) C 2 ω(2r). For 0 r 1, define a function σ by (14) σ(r) := Using (13), it is easy to see that r 0 ω(s) s ds. (15) C 1 σ(2r) σ(r) C 2 σ(2r) for some constants C 1, C 2 and ω(r) βσ(r) certainly implies (16) ω(r) σ(r). We shall use ω and σ as modulii of continuity for functions. Let U be a bounded domain in R 2. Definition 1. For any modulus of continuity ω, let C ω (U) denote those functions f C(U) for which f(x) f(y) C ω( x y ) for all x, y U. C ω (U) is a Banach space under the norm (17) f C ω (U) f(x) f(y) := sup f(x) + sup x U x,y U ω( x y ). x y Definition 2. For any domain Ω, the collection of all functions f such that f C ω (U) for all U compactly contained in Ω is denoted by C ω (Ω). 10
15 Definition 3. Let C 1,ω (U) denote those functions f C 1 (U) whose first order derivatives f/ z and f/ z are in C ω (U). C 1,ω (U) is a Banach space under the norm (18) f C 1,ω (U) := f C ω (U) + f C ω (U) + f C 0 (U). Definition 4. For any domain Ω, the collection of all functions f such that f C 1,ω (U) for all U compactly contained in Ω is denoted by C 1,ω (Ω). When ω satisfies the Dini condition (11), the functions in C ω (Ω) are called Dini continuous. Throughout this paper, we shall use ω to denote the modulus of continuity with the properties (11), (12) and (13). Using (14), C σ (U) and C 1,σ (U) are Banach spaces defined as above using σ as the modulus of continuity. In fact, using (16), we see that C ω (Ω) is contained in C σ (Ω). When ω(r) = r α with α (0, 1), we get the special case of Holder continuous functions. In this case, σ(r) is so equivalent to r α and hence C ω (Ω) coincides with C σ (Ω). In this work, we also consider the collection of all functions f whose first-order derivatives are square-integrable in every U compactly contained in any domain Ω, which is generally denoted by W 1,2 loc (Ω). 11
16 3. Constant coefficients in bounded domain Throughout this work, we shall denote an open disk with center z 0 and radius r as D r (z 0 ) and an open disk with center as the origin and radius r as D r. We shall consider Ω to be a bounded domain in R 2. Theorem 1. Let h C ω (Ω) and µ 0 and ν 0 be constants satisfying µ 0 + ν 0 κ < 1. Let f W 1,2 loc (Ω) satisfy (19) f µ 0 f ν 0 f = h. Then f C 1,σ (Ω) and satisfies (20) f C σ (D) + f C σ (D) C(κ, D, U) ( h C ω (U) + f C ω (U) ) for any domain D, U such that D is compactly contained in U and U is compactly contained in Ω. We first consider the special case of the Beltrami equation where the coefficients µ and ν are zero. This is the inhomogeneous Cauchy- Riemann equation of the form (21) p = q. If q is integrable on a domain Ω, then the domain potential of q is given by (22) p(z) = 1 π Ω q(τ) (z τ) da τ and is a distribution solution of the equation (21) in Ω since 1 πz fundamental solution for the C-R operator. is the 12
17 An important result that we use in this section is that for any disk D R (z 0 ), (23) D R (z 0 ) 1 (z 1 τ) 2 da τ = 0 for z 1 D R (z 0 ). Using polar coordinates, this can easily be seen for the case when z 1 = z 0. When z 1 z 0, we can write D R (z 0 ) 1 (z 1 τ) 2 da τ = lim ɛ 0 D R (z 0 )\D ɛ(z 1 ) Now Green s theorem can be used to get (23). 1 (z 1 τ) 2 da τ. We first prove results which are necessary to prove the above theorem. Lemma 1. Consider a function φ C ω (D 1 ). Then φ(τ) (24) z (z τ) da φ(τ) τ = D 1 (z τ) da 2 τ for z D 1. D 1 Proof: Let w(z) := D 1 φ(τ) (z τ) da φ(τ) φ(z) τ and u(z) := da D 1 (z τ) 2 τ. Note that using (23), u(z) = D 1 φ(τ) (z τ) 2 da τ. Consider a function η C (R) such that 0 η 1, 0 η 2 and 0, for t 1 η(t) = 1, for t 2 Denote η( z τ ) as η ɛ ɛ (z τ). For ɛ > 0, define η ɛ (z τ) φ(τ) w ɛ (z) := da τ. D 1 (z τ) 13
18 Note that w ɛ (z) is smooth as it is the convolution with the smooth η( z function 1 ). Now w is the uniform limit of w z ɛ ɛ since w ɛ (z) w(z) = 1 η ɛ (z τ) φ(τ) da τ z τ <2ɛ (z τ) 1 η ɛ (z τ) φ(τ) da τ z τ z τ <2ɛ C(ɛ) sup φ(τ) τ D 1 where C(ɛ) tends to 0 as ɛ tends to 0. We now prove that u is the uniform limit of z w ɛ. Consider [ ] ηɛ (z τ) z w ɛ (z) = z (φ(τ) φ(z)) da τ D 1 (z τ) (25) [ ] ηɛ (z τ) + φ(z) da τ (z τ) D 1 z The second integral can be written as D 1 z [ ] ηɛ (z τ) da τ = (z τ) = i 2 = i 2 D 1 τ D 1 [ ] ηɛ (z τ) da τ (z τ) [ ] ηɛ (z τ) d τ (z τ) D 1 1 (z τ) d τ since ɛ can be chosen so that η ɛ (z τ) = 1 for τ on D 1 for fixed z. Now using polar coordinates, it can be seen that D 1 1 (z τ) d τ = 0 Hence (26) D 1 z [ ] ηɛ (z τ) da τ = 0. (z τ) 14
19 Using (26) in (25),we get We now have z w ɛ (z) = (27) u(z) z w ɛ (z) = D 1 z z τ <2ɛ z τ <2ɛ [ ] ηɛ (z τ) (φ(τ) φ(z)) da τ (z τ) But ( ( ) 1 η ɛ (z τ) z z τ 1 z τ ɛ z τ η ɛ ( ) 2 ɛ z τ + 1 z τ 2 Substituting the above in (27), we get [ ] 1 ηɛ (z τ) z (φ(τ) φ(z)) da τ (z τ) 1 η ɛ (z τ) z z τ φ(τ) φ(z) da τ + 1 η ) ɛ( z τ ) z τ 2 u(z) z w ɛ (z) ( ) 2 z τ <2ɛ ɛ z τ + 1 φ(τ) φ(z) da z τ 2 τ ( ) 2 φ C ω (D 1 ) ɛ z τ + 1 ω( τ z ) da z τ 2 τ z τ <2ɛ φ C ω (D 1 )(4 ω(2ɛ) + 2π σ(2ɛ)) which tends to 0 as ɛ tends to 0. Hence we can conclude that z w(z) = u(z) which is our required result. In the case of Holder continuous functions, it is well known (cf.[3]) that the Holder norm of derivative of the solution p to the inhomogeneous C-R equation (21) is bounded by the Holder norm of q. We now obtain 15
20 similar bounds in the case of (21) for Dini continuous functions. This is where we get to see the role of the C σ norm. Lemma 2. For q C ω (D ɛ ), the domain potential p(z) = 1 π lies in C 1,σ (D ɛ/2 ) and satisfies the estimate D ɛ q(τ) (z τ) da τ (28) p C σ (D ɛ/2 ) + p C σ (D ɛ/2 ) K(ɛ) q C ω (D ɛ). Proof: For z D ɛ/2, using (24) and (23), we can write p(z) = 1 π D ɛ q(τ) (z τ) da 2 τ = 1 q(τ) q(z) da π D ɛ (z τ) 2 τ. Since p(z) = q(z), we have p C σ (D ɛ/2 ) K q C ω (D ɛ). Now let us estimate p C σ (D ɛ/2 ). For any z 1, z 2 D ɛ/2, we get p(z 1 ) p(z 2 ) = 1 q(τ) q(z 1 ) q(τ) q(z 2 ) π da D ɛ (z 1 τ) 2 τ da D ɛ (z 2 τ) 2 τ Let z 1 z 2 = δ < ɛ, z 3 = z 1+z 2 2 so that D δ (z 3 ) is contained in D ɛ. The above equation can now be estimated as (29) p(z 1 ) p(z 2 ) 1 q(τ) q(z 1 ) π da D δ (z 3 ) (z 1 τ) 2 τ + 1 π + 1 q(τ) q(z 1 ) π da (z 1 τ) 2 τ + D ɛ D δ (z 3 ) D δ (z 3 ) D ɛ D δ (z 3 ) q(τ) q(z 2 ) da (z 2 τ) 2 τ q(τ) q(z 2 ) da (z 2 τ) 2 τ Now q(τ) q(z 1 q C ω (D ɛ) ω( τ z 1 ) gives q(τ) q(z 1 ) da D δ (z 3 ) (z 1 τ) 2 τ q C ω (D ɛ) (30) q C ω (D ɛ) D δ (z 3 ) D 2δ (z 1 ) = 2π q C ω (D ɛ) σ(2δ). ω( τ z 1 ) z 1 τ 2 da τ ω( τ z 1 ) z 1 τ 2 da τ 16
21 Using (15), (30) can be written as (31) q(τ) q(z 1 ) da (z 1 τ) 2 τ C q C ω (D σ(δ). ɛ) Similarly (32) D δ (z 3 ) D δ (z 3 ) q(τ) q(z 2 ) da (z 2 τ) 2 τ C q C ω (D σ(δ). ɛ) Now we estimate the last term in (29). First observe that using (23), we get = D ɛ D δ (z 3 ) D ɛ D δ (z 3 ) q(τ) q(z 1 ) q(τ) q(z 2 ) da (z 1 τ) 2 τ da D ɛ D δ (z 3 ) (z 2 τ) 2 τ [ ] 1 (z 1 τ) 1 (q(τ) q(z 2 (z 2 τ) 2 2 )) da τ Now by the mean value theorem, there exists z τ between z 1 and z 2 such that (33) 1 (z 1 τ) 1 2 (z 2 τ) = (z z 2 ). ( z τ τ). 3 Hence we get [ ] 1 D ɛ D δ (z 3 ) (z 1 τ) 1 (q(τ) q(z 2 (z 2 τ) 2 2 )) da τ (34) q C ω (D ɛ) z 1 z 2 2 D ɛ D δ (z 3 ) ( z τ τ) 3 ω( τ z 2 ) da τ ω( τ z 2 ) = 2δ q C ω (D ɛ) da z τ τ 3 τ. But D ɛ D δ (z 3 ) z 2 τ z 2 z 3 τ z 3 2ɛ 3δ 2 δ/2 = z 1 z 2. 2 Using property (13) of ω(r), we get ( ) β z 2 τ β z1 z 2 ω( τ z 2 ) ω( z 1 z 2 ). 2 17
22 Using these we get D ɛ D δ (z 3 ) ω( τ z 2 ) z τ τ 3 da τ D ɛ D δ (z 3 ) 2 β δ β ω(δ) ω( z 1 z 2 ( z 1 z 2 ) β 2 z τ 3 z 2 τ β D ɛ D δ (z 3 ) Using (33), it can be seen that z 3 τ 2 z τ τ. Also z 2 τ δ 2 + z 3 τ 3 z 3 τ. 2 Using these, the above integral can be written as (35) D ɛ D δ (z 3 ) ω( τ z 2 ) z τ τ 3 dτ 2 β δ β ω(δ) β δ β ω(δ) D ɛ D δ (z 3 ) z 3 τ δ = β δ β ω(δ) π δ 1+β 1 β Cδ 1 ω(δ). da τ 1 z τ τ 3 z 2 τ β da τ. 3 β 2 3 β z 3 τ 3 β da τ 1 z 3 τ 3 β da τ Now using (35) in (34) we get [ ] 1 (36) D ɛ D δ (z 3 ) (z 1 τ) 1 (q(z 2 (z 2 τ) 2 2 ) q(τ)) da τ C 1 q C ω (D ɛ) ω(δ). Using (31),(32), (36) and (16) in (29), we get p(z 1 ) p(z 2 ) K q C ω (D σ(δ) ɛ) or p C σ (D ɛ/2 ) K(ɛ) q C ω (D. ɛ) 18
23 Hence we get our required result (28). To further understand the regularity properties of solutions to the Cauchy- Riemann equation and the Beltrami equation, it is convenient to make use of analytic properties of operators that are defined in the whole complex plane. For any bounded function φ with compact support on C, the Cauchy transform is defined by (37) (Cφ)(z) := 1 φ(τ) π (z τ) da τ and the Beurling transform is the singular integral operator defined by (38) (Sφ)(z) := 1 π C C φ(τ) (z τ) 2 da τ. When φ is integrable in a domain Ω C with compact support in Ω and if φ is extended by zero outside Ω, then the domain potential of φ coincides with the Cauchy transform of φ. The Cauchy transform p = Cq solves the C-R equation (21) when q is bounded and has compact support. Note that using (23) in Lemma 1, we can conclude that for any φ C ω (U) with compact support in K U and extended by zero outside K, (39) z Cφ(z) = Sφ(z) for z C. Also for any φ C 1,ω (U) with compact support in K U, (40) φ(z) = C( φ(z)) for z C. 19
24 Proposition 1. For any domain Ω and q C ω (Ω), every solution p W 1,2 loc (Ω) of the equation (21) lies in C1,σ (Ω) and satisfies the estimate (41) p C σ (D ɛ/2 ) + p C σ (D ɛ/2 ) C( q C ω (D ɛ) + p C ω (D ɛ) ) where C = C(ɛ) and D ɛ, D ɛ/2 are concentric disks of radius ɛ and ɛ/2 respectively that are compactly contained in Ω. Proof: We first prove that the solution p lies in C 1,σ (Ω). Note that it is sufficient to prove that p lies in C 1,σ (D ɛ/2 ). Consider a smooth function φ with compact support in D ɛ such that φ 1 in D ɛ/2. Let q 1 = φ q. Now since q 1 has compact support in D ɛ then let p 1 = Cq 1, which is also the domain potential, so we know by Lemma 2 that p 1 C 1,σ (D ε/2 ). Hence for z D ε/2, (p p1 )(z) = (q q 1 )(z) = 0. So (p p 1 ) is holomorphic in D ε/2 and hence is smooth in D ε/2. Thus p C 1,σ (D ε/2 ). Consider the smooth function φ introduced above. Define p = φ p. Then (42) p = φ p + φ q. Since φ p + φ q C ω (D ɛ ) with compact support in D ɛ, using (40), p = C( p) solves (42). For z D ɛ/2, p(z) can be written as p(z) = p(z) = C p(τ) = 1 p(τ) π (z τ) da τ = 1 ( φ p + φ q)(τ) da τ. π D ɛ (z τ) Using Lemma 2, p satisfies the estimate D ɛ p C σ (D ɛ/2 ) + p C σ (D ɛ/2 ) K φ p + φ q C ω (D ɛ). But p = p in D ɛ/2. Hence (43) p C σ (D ɛ/2 ) + p C σ (D ɛ/2 ) K( φ p C ω (D ɛ) + φ q C ω (D ɛ) ). 20
25 For z 1, z 2 D ɛ, We now get ( φ p)(z 1 ) ( φ p)(z 2 ) ω( z 1 z 2 ) = φ(z 1 )(p(z 1 ) p(z 2 )) + p(z 2 )( φ(z 1 ) φ(z 2 ) ω( z 1 z 2 ) φ(z 1 ) p(z 1) p(z 2 ) ω( z 1 z 2 ) + p(z 2 ) φ(z 1 ) φ(z 2 ). ω( z 1 z 2 ) (44) φ p C ω (D ɛ) φ C 0 (D ɛ) p C ω (D ɛ) + p C 0 (D ɛ) φ C ω (D ɛ) C 1 p C ω (D ɛ) where C 1 = C 1 (ɛ). Similarly we can see that φ q C ω (D ɛ) C 2 q C ω (D ɛ) where C 2 = C 2 (ɛ). Using this and (44) in (43), we get the desired result (41). Proof of Theorem 1: As in [3],we first reduce the Beltrami equation (19) to an inhomogeneous Cauchy-Riemann equation and then try to obtain the bounds for the solution of (19) using the solution of the inhomogeneous Cauchy-Riemann equation. For any constants a, b, define (45) p(z) := f(ξ) + b f(ξ), where ξ = z + a z which can be written as (46) f(ξ) = p(z) b p(z) 1 b 2 where z = ξ a ξ 1 a 2. 21
26 Let u(z) := h(ξ) = h(z + a z). Choose constants a, b such that f(ξ) satisfying (19) can be reduced to the form (47) p(z) = u(z) + ab u(z). We choose the pair of values for a and b as (48) a = 2 µ µ 0 2 ν (1 + µ 0 2 ν 0 2 ) 2 4 µ 0 2 ) 1/2, (49) b = This gives a, b κ < 1. 2 ν ν 0 2 µ (1 ν µ 0 2 ) 2 4 µ 0 2 ) 1/2. Using Proposition 1, any solution p W 1,2 loc (Ω) of (47) lies in C1,σ (D ɛ/2 ) and satisfies p C σ (D ɛ/2 ) + p C σ (D ɛ/2 ) K(ɛ) ( u C ω (D ɛ) + p C ω (D ɛ) ) Using the definition of u(z), the above relation can be written as (50) p C σ (D ɛ/2 ) + p C σ (D ɛ/2 ) K 1(ɛ)( h C ω (D ɛ) + p C ω (D ɛ) ). Using (85) and Proposition 1, we can see that any solution f W 1,2 loc (Ω) of (19) lies in C 1,σ (D ɛ/2 ). Now we estimate f C σ (D ɛ/2 ) + f C σ (D ɛ/2 ) in terms of p C σ (D ɛ/2 ) + p C σ (D ɛ/2 ). We first note that the map from z to ξ takes unit circles to ellipses which are contained in circles with double the radius. Since ξ = z + a z, we get ξ 1 ξ 2 z 1 z 2 (1 + a ) 2 z 1 z 2. 22
27 This gives (51) ω( ξ 1 ξ 2 ) ω(2 z 1 z 2 ) Cω( z 1 z 2 ). Again since z = ξ b ξ, we get z 1 b 2 1 z 2 ξ 1 ξ 2. Hence 1 κ (52) σ( z 1 z 2 ) σ( ξ 1 ξ 2 1 κ ) (1 κ) β σ( ξ 1 ξ 2 ) using the definition of sigma. Now let us obtain a relation between sigma norm of p and sigma norm of f. Let ξ = φ(z). Hence (85) can be written as p(z) = f(φ(z)) + bf(φ(z)). Now the C ω norm of p can be written as (53) p C ω (D ɛ/2 ) f φ C ω (D ɛ/2 ) + b f φ C ω (D ɛ/2 ) C 1 (1 + b ) f C ω (D ɛ) where the last inequality is obtained using (51) and the fact that φ(d ɛ/2 ) is contained in D ɛ. Now we get a relation between sigma norm of the derivatives of f and p. Let the map from ξ to z be denoted by ψ. Now using (46), we get f ξ(ξ) = (1 + ab) p z(z) (a + b) p z (z) (1 b 2 )(1 a 2 ) = (1 + ab) p zψ(ξ) (a + b) p z ψ(ξ) (1 b 2 )(1 a 2 ) where we have used z = ψ(ξ) in the last equality. Using z = ξ a ξ 1 a 2, we get (1 κ) z ξ. Choose δ such that D δ is contained in φ(d ɛ/2 ). 23
28 Now the sigma norm can be obtained as f ξ C σ (D δ ) 1 + ab p z ψ C σ (D δ ) + a + b p z ψ C σ (D δ ) (1 b 2 )(1 a 2 ) 1 + ab p z ψ C σ (φ(d ɛ/2 )) + a + b p z ψ C σ (φ(d ɛ/2 )). (1 b 2 )(1 a 2 ) Using (52), the above inequality can be written as (54) Similarly we can get f ξ C σ (D δ ) 1 + ab (1 κ) β p z C σ (D ɛ/2 ) (1 b 2 )(1 a 2 ) + a + b (1 κ) β p z C σ (D ɛ/2 ). (1 b 2 )(1 a 2 ) (55) f ξ C σ (D δ ) a + b (1 κ) β p z C σ (D ɛ/2 ) (1 b 2 )(1 a 2 ) ab (1 κ) β p z C σ (D ɛ/2 ). (1 b 2 )(1 a 2 ) Adding (54) and (55), we get f ξ C σ (D δ ) + f ξ C σ (D δ ) (1 κ) β ( p z C σ (D ɛ/2 ) + p z C σ (D ɛ/2 ) ) (1 + b )(1 + a ) (1 κ) β K 1 (ɛ)( h C ω (D + p ɛ) C ω (D ) ɛ). (1 + b )(1 + a ) The last inequality is obtained using (50). Using (53), we now get Hence f ξ C σ (D δ ) + f ξ C σ (D δ ) K 2(ɛ)( h C ω (D + f ɛ) C ω (D 2ɛ ) ). (1 κ) β (1 + b )(1 + a ) (56) f ξ C σ (D δ ) + f ξ C σ (D δ ) C(κ, ɛ))( h C ω (D 2ɛ ) + f C ω (D 2ɛ ) ). For any D compactly contained in U, we now need to get a similar inequality with norms in D and U. Consider any point ξ D. Choose 24
29 ɛ such that D 2ɛ (ξ) is contained in U. Now for ξ 1, ξ 2 D such that ξ = (ξ 1 + ξ 2 )/2 and ξ 1 ξ 2 < δ, we get f(ξ 1 ) f(ξ 2 ) σ( ξ 1 ξ 2 ) f(η) f(ζ) sup η,ζ D δ (ξ) σ( η ζ ) f(ξ) sup f(ζ). ζ D δ (ξ) Combining the above two inequalities, we get that for any ξ D, Similarly, f(ξ 1 ) f(ξ 2 ) σ( ξ 1 ξ 2 ) f(ξ 1 ) f(ξ 2 ) σ( ξ 1 ξ 2 ) + f(ξ) f C σ (D δ (ξ)). + f(ξ) f C σ (D δ (ξ)). Adding the above two inequalities and using (56), we get f(ξ 1 ) f(ξ 2 ) σ( ξ 1 ξ 2 ) + f(ξ) + f(ξ 1 ) f(ξ 2 ) σ( ξ 1 ξ 2 ) + f(ξ) f C σ (D δ (ξ)) + f C σ (D δ (ξ)) C(κ, ɛ)( h C ω (D 2ɛ (ξ)) + f C ω (D 2ɛ (ξ)) ) C(κ, ɛ)( h C ω (U) + f C ω (U) ). Taking the supremum over ξ and dependant ξ 1, ξ 2, we get f C σ (D) + f C σ (D) C(κ, D, U)( h C ω (U) + f C ω (U) ). In case of the Beltrami equation with variable coefficients, to obtain the estimates on domains D, U compactly contained in Ω, the standard method for generalising from constant to variable coefficients is the following: We first freeze the coefficients at a point in Ω and consider functions with support in a small disk around this point: the disk is 25
30 chosen such that the difference between the variable coefficients and value of the coefficient at the center of the disk are sufficiently small. We now look at the Beltrami equation satisfied by this function and then obtain estimates similar to (20). In obtaining this estimate, the standard absorption trick cannot be used since it requires having the same norms on the right hand side as the left hand side of the inequality. The norms on the right hand side and left hand side of (20) are not the same. We have the C 1,σ norm of the solution bounded by the C ω norm of the solution. Hence the above mentioned absorption cannot be done in this case. To overcome this problem, we now look at functions that are less regular than the functions in the C ω space but are more regular than the functions in the C σ space. The idea is that when the coefficients of the Beltrami equation are functions with this kind of regularity, we can obtain a result where the norm of the derivative of solution to the Beltrami equation is bounded by the same norm of the solution. This space of functions is defined using Fourier multipliers. We first work with functions on a torus, obtain the required estimates to solutions of the Beltrami equation and then use this to get estimates on any bounded domain. We work with functions on a torus rather than the whole complex plane since this gives us boundedness of the Beurling transform which will be a key tool used in the next section. 4. The space C (λ) on the torus In [15], the function space C (λ) (R n ) with (57) λ(j) = ω(2 j ) 26
31 was defined for a general modulus of continuity ω using a Littlewood- Paley decomposition {ψ j } as the collection of all f R n such that (ψ j (D)f) L (R n ) Cλ(j). where ψ j (D) is the Fourier multiplier associated with the Littlewood- Paley partition of unity ψ j (ξ). C (λ) (R n ) is a Banach space under the norm f C (λ) (R n ) := sup j (ψ j (D)f) L (R n ). λ(j) The classical results in Fourier analysis on R n were used in [15] to analyse these functions and their properties. We can use these results to study functions on a torus by considering them as periodic functions on R 2. Let T 2 = (R/2πZ) 2 denote the 2-dimensional torus. We identify T 2 with [ π, π) [ π, π) R 2 and functions on T 2 with functions on R 2 that are 2π -periodic in each of the coordinate directions. For functions on a torus, the toroidal Fourier transform is defined from C (T 2 ) to S(Z 2 ) where S(Z 2 ) denotes the space of rapidly decaying functions on Z 2. However, in our work we will be using the Fourier transform on R 2. But since periodic functions on R 2 are not L 1, while taking the Fourier transform, we consider them as tempered distributuions, for which the Fourier transform is given by (10). In order to define the C (λ) space, we first need a smooth Littlewood- Paley decomposition which is obtained as follows: Choose a compactly 27
32 supported function Ψ 0 such that 1 for ξ 1 Ψ 0 (ξ) = 0 for ξ 2 for ξ R 2. For k 1, set (58) Ψ k (ξ) = Ψ 0 (2 k ξ). Let (59) ψ 0 (ξ) = Ψ 0 (ξ) and ψ k (ξ) = Ψ k (ξ) Ψ k 1 (ξ) for k 1 so that (60) Ψ k (ξ) = ψ 0 (ξ) ψ k (ξ) and ψ k (ξ) = 1. The collection {ψ k (ξ)} forms a Littlewood-Paley partition of unity. Note that for each k, Ψ k (ξ) is supported on a disk ξ < 2 k+1. For k > 0, each ψ k (ξ) is supported on the annulus 2 k 1 < ξ < 2 k+1 and ψ 0 (ξ) is supported on ξ < 2. k=0 Since ψ j (ξ) has compact support, we consider ψ j (D) : S (R 2 ) C (R 2 ) as a Fourier multiplier operator associated with the function ψ j (ξ). Here since we want to define ψ j (D) on periodic functions on R 2, we again consider the functions as tempered distributions. It can also be written as a convolution given by (ψ j (D)f)(z) = ψ j ( z)f(z z)d z. R 2 Note that a Fourier multiplier operator preserves periodicity, as can be seen from the following argument: Let f be a function on R 2 with 28
33 period 2π and P (D) be a Fourier multiplier operator with symbol p(ξ). Then we have (P (D)f)(z) = p ( z)f(z z)d z R 2 = p ( z)f(z z + 2π)d z R 2 = p ( z)f((z + 2π) z)d z R 2 = (P (D)f)(z + 2π). Hence ψ j (D) can be considered as an operator on the torus. We are now ready to define the function space C (λ) (T 2 ). We assume that ω is a Dini modulus of continuity, i.e. satisfies the Dini condition (11). Definition 5. C (λ) (T 2 ) is the collection of all periodic functions f R 2 such that (61) (ψ j (D)f) L (R 2 ) Cλ(j) and is a Banach space under the norm (62) f C (λ) (T 2 ) := sup j (ψ j (D)f) L (R 2 ). λ(j) The space C (λ) (T 2 ) is identified with C (λ) P (R2 ) where the subscript P refers to periodic functions on R 2. Definition 6. We denote by C 1,(λ) (T 2 ), those functions f C 1 (T 2 ) whose first order derivatives f/ z and f/ z are in C (λ) (T 2 ). C 1,(λ) (T 2 ) is a Banach space under the norm (63) f C 1,(λ) (T 2 ) := f C (λ) (T 2 ) + f C (λ) (T 2 ) + f C 0 (T 2 ). 29
34 The relation between the spaces C (λ), C ω and C σ that holds on R n (as shown in [15]) also holds in the case of a torus T 2 i.e. (64) C ω (T 2 ) C (λ) (T 2 ) C σ (T 2 ). This is true since the spaces C ω (T 2 ), C (λ) (T 2 ) and C σ (T 2 ) are just the periodic functions on R 2 which belong to the corresponding spaces on R 2. We observe here that λ(j) has the following properties: Lemma 3. The positive decreasing sequence λ(j) given by λ(j) = ω(2 j ) satisfies the following properties: (65) (66) λ(j) <, j λ 2 (j) cλ(l), j l and λ(j) is slowly varying, i.e. (67) λ(j) Cλ(j + 1). Proof: Proof of (65): 1 0 ω(t) dt t j=0 ω(2 j ) 2 j (2 j 2 (j+1) ) = 1 2 ω(2 j ) = 1 2 λ(j). j Using (11), we get j=0 λ(j) <. Proof of (66): Since λ(j) is a decreasing sequence, we can write Σ j l λ 2 (j) λ(l)σ j l λ(j). 30
35 Using (65), we get the final result. Proof of (67): Using (13), we get λ(j) = ω(2 j ) 2 β ω(2 (j+1) ) Cλ(j + 1). A tool that we will use extensively in this section is the Beurling transform. It is defined for functions on a torus as a Fourier multiplier operator by (68) (Ŝf)(ξ) = m(ξ) ˆf(ξ) where m(ξ) = ξ/ ξ for ξ R 2. In working with Fourier multiplier operators, convolution of functions will be an operation that we will use frequently in this section. We shall make use of the following property of convolution of functions: If f L 1 (R 2 ), g L (R 2 ) and one of them has compact support, then (69) f g L (R 2 ) f L 1 (R 2 ) g L (R 2 ). We now prove the boundedness of the Beurling transform which plays a vital role in obtaining our estimates. Lemma 4. If h C (λ) (T 2 ), then (70) Sh C (λ) (T 2 ) C h C (λ) (T 2 ). Proof: We first observe that the symbol of S has a singularity at ξ = 0. We handle this singularity by writing S in terms of a smoothing operator. Using the Littlewood-Paley partition of unity, we get S = ψ k (D)S = ψ 0 (D)S + ψ k (D)S. k=0 k=1 31
36 Let S := k=1 ψ k(d)s. We now have (71) Sh C (λ) (T 2 ) (ψ 0 (D)S)h C (λ) (T 2 ) + Sh C (λ) (T 2 ) Observe that as the symbol of ψ 0 (D)S has compact support, ψ 0 (D)S is a smoothing operator which takes a periodic function h C (λ) (R 2 ) to a smooth periodic function and so is bounded from C (λ) (T 2 ) to C (λ) (T 2 ), i.e. (72) ψ 0 (D)Sh C (λ) (T 2 ) C 2 h C (λ) (T 2 ). Let us now look at the boundedness of S. To estimate Sh C (λ) (T 2 ), we first consider ψ l (D) Sh L (R 2 ) = ψ l (D) ψ k (D)Sh L (R 2 ) k=1 ( Fh) = F 1 ψ l (ξ) ψ k (ξ) ξ ξ k=1 L (R 2 ). Using the fact that ψ l (ξ) is supported only in 2 l 1 ξ 2 l+1, we get ( ψ l (D) Sh k=l+1 L (R 2 ) = F Fh) 1 ψ l (ξ) ψ k (ξ) ξ ξ k=l 1 L (R 2 ) ( k=l+1 = F 1 ψ k (ξ) ξ l(ξ)fh) ξ ψ = F 1 We now use (69) to get k=l 1 ( k=l+1 k=l 1 L (R 2 ) ) ψ k (ξ) ξ F 1 (ψ l (ξ)fh) ξ L (R 2 ). (73) ψ l (D) Sh L (R 2 ) ( k=l+1 ) F 1 ψ k (ξ) ξ ξ k=l 1 L 1 (R 2 ) F 1 (ψ l (ξ)fh) L (R 2 ). 32
37 ( k=l+1 ) Let us now estimate the term F 1 k=l 1 ψ k(ξ) ξ L ξ 1 (R ). 2 Using the definition of the inverse Fourier transform, (58) and (59), we get ( k=l+1 F 1 k=l 1 ) ψ k (ξ) ξ = (Ψ 0 (2 (l+1) ξ) Ψ 0 (2 (l 2) ξ)) ξ eiz ξ dξ ξ R ξ 2 = (Ψ 0 (2 (l+1) ξ) Ψ 0 (2 (l 2) ξ)) ξ eiz ξ R ξ 2 Let Ψ(η) := Ψ 0 (2 1 η) Ψ 0 (2 2 η). Observe that Ψ(η) is supported in 2 1 < η < 2 2. The above expression can now be written as ( k=l+1 ) F 1 ψ k (ξ) ξ Ψ(2 l ξ) ξ eiz ξ dz ξ ξ k=l 1 = 2 2l F 1 (m Ψ)(2 l z) R 2 R 2 2 2l where m(η) = η. η ( k=l+1 ) Now the L 1 norm of F 1 k=l 1 ψ k(ξ) ξ ξ F 1 ( k=l+1 k=l 1 ψ k (ξ) ξ ξ ) L 1 (R 2 ) Ψ(η) η η eiz 2lη 2 2l dη R 2 Ψ(η) η η ei2l z η dη can be written as = 2 2l F 1 (m Ψ)(2 l z) = R 2 2 2l L 1 (R 2 ) F 1 (m Ψ)(2 l z) dz = 2 2l F 1 (m Ψ)( z) 2 2l d z R 2 = F 1 (m Ψ) C L 1 (R 2 ) 33
38 where C is independent of l. This is true due to the fact that as m Ψ is smooth and has compact support, it is in Schwartz class and hence in L 1 (R 2 ). Hence which gives ψ l (D) Sh L (R 2 ) C 1 ψ l (D)h L (R 2 ) (74) Sh C (λ) (T 2 ) C 2 h C (λ) (T 2 ). Using (72) and (74 ) in (71), we get (75) Sh C (λ) (T 2 ) C h C (λ) (T 2 ). Taylor had shown in [15] that C (λ) (R 2 ) is a Banach algebra. The following lemma uses a slightly different proof to show that C (λ) (T 2 ) is a Banach algebra. Lemma 5. Assume f, g C (λ) (T 2 ). Then (76) fg C (λ) (T 2 ) C f C (λ) (T 2 ) g C (λ) (T 2 ). Proof: Here we use Bony s paraproduct decomposition. For the product fg, it is given by fg = T f g + T g f + R f g where T f is Bony s paraproduct defined by T f g = k 5 Ψ k 5 (D)f ψ k+1 (D)g 34
39 and R f g, which is used to denote the remainder is given by R f g = j φ j (D)f ψ j (D)g where φ j (D)f = j k 3 ψ k (D)f Hence (77) fg C (λ) (T 2 ) T f g C (λ) (T 2 ) + T g f C (λ) (T 2 ) + R f g C (λ) (T 2 ). Let us first estimate R f g C (λ) (T 2 ). We have (ψ l (D)R f g) L (R 2 ) = ψ l ( j φ j (D)f ψ j (D)g) L (R 2 ) = ψ l (D) j ( j k 3 ψ k (D)f) ψ j (D)g L (R 2 ). Now let us consider the term in the right hand side of the above equation. The transform of the j th term of R f g is F( ψ k (D)f ψ j (D)g)(ξ) j k 3 = F((ψ j ψ j+3 )(D)f.ψ j (D)g)(ξ) = [(ψ j ψ j+3 ) f) (ψ j ĝ)](ξ) = ψ l (ξ) j [((ψ j ψ j+3 ) f) (ψ j ĝ)](ξ)) The term j (ψ j ψ j 3 )(D)f is supported in 2 j 4 ξ 2 j+4 and the support of ψ j (D)g is in 2 j 1 ξ 2 j+1. Hence the support of transform of the j th term of R f g is in ξ C2 j+4. Now ψ l (ξ) is supported in 2 l 1 < ξ 2 l+1. Hence for the annulus to intersect the disc ξ 2 j+4, we need C2 j+4 2 l 1. We can now say that there 35
40 exists some integer N such that j l N. Hence we can write ψ l (D)R f g L (R 2 ) = (ψ l (D) ( ψ k (D)f) ψ j (D)g L (R 2 ) Now we can use (69) to get j l N j k 3 = F 1 (ψ l (ξ)) (F ψ k (D)f ψ j (D)g) L (R 2 ). j l N j k 3 ψ l (D)R f g L (R 2 ) ψ l L 1 (R 2 ) ( (ψ j ψ j+3 )(D)f) L (R 2 ) ψ j (D)g L (R 2 ) j l N C f C (λ) (T 2 ) g C (λ) (T 2 ) C f C (λ) (T 2 ) g C (λ) (T 2 ) C f C (λ) (T 2 ) g C (λ) (T 2 ) (λ(j 3) λ(j + 3))λ(j) j l N j l N j l N (k 1 λ(j) + k 2 λ(j) + k 2 λ(j) + 4λ(j))λ(j) C 1 λ 2 (j). The last inequality is obtained using (67) and the property that λ(j) is a positive decreasing sequence. Now j l N (k 1λ(j)+k 2 λ(j)+k 2 λ(j)+ 4λ(j))λ(j) C 1 λ 2 (j). We now use (66) to get (ψ l (D)R f g L (R 2 ) C 2 f C (λ) (T 2 ) g C (λ) (T 2 )λ(l). Using the above inequality we get (78) R f g C (λ) (T 2 ) K 1 f C (λ) (T 2 ) g C (λ) (T 2 ). Now let us estimate T f g C (λ) (T 2 ). We have (ψ l (D)T f g) L (R 2 ) = ψ l (D) k 5 Ψ k 5 (D)f ψ k+1 (D)g L (R 2 ) = ψ l (D) k 5(ψ ψ k 5 )(D)f ψ k+1 (D)g L (R 2 ) 36
41 Since ψ l is supported in the annulus 2 l 1 ξ 2 l+1, we must have k 5 l 1 or k l + 4. Let m = min(k 5, l + 1). Now we have (ψ l (D)T f g) L (R 2 ) = ψ l (D) ψ l L 1 (T 2 ) m k l+4 j=l 1 k l+4 C f C (λ) (T 2 ) g C (λ) (T 2 ) C f C (λ) (T 2 ) g C (λ) (T 2 ) C f C (λ) (T 2 ) g C (λ) (T 2 ) ψ j (D)f ψ k+1 (D)g L (R 2 ) ψ j (D)f L (R 2 ) ψ k+1 (D)g L (R 2 ) ( k l+4 k l+4 k l+4 C 2 f C (λ) (T 2 ) g C (λ) (T 2 )λ(l) m j=l 1 λ(j))λ(k + 1) λ(l 1) + λ(l)) + λ(l + 1)λ(k + 1) C 1 λ(l)λ(k + 1) The last inequality is obtained by using k l+4 λ(k + 1) C for some constant C. We now get (79) T f g C (λ) (T 2 ) K 2 f C (λ) (T 2 ) g C (λ) (T 2 ). Similarly we can get (80) T g f C (λ) (T 2 ) K 3 f C (λ) (T 2 ) g C (λ) (T 2 ). Using (78),(79) and (80) in (77), we get (76). We now obtain a result on the operator which will be useful in analysing the Beltrami equation. Proposition 2. The operator : C 1,(λ) C (λ) is Fredholm of index zero with kernel and cokernel consisting of constants. 37
42 Proof: Define a parametrix P for by P h = (φ(ξ) 2i ξ ĥ(ξ)) where φ is a smooth function satisfying φ(0) = 0 and φ(ξ) = 1 for ξ 1. Using a similar proof as in lemma 4 for the Beurling transform, it can be shown that for h C (λ) (T 2 ), P h C 1,(λ) (T 2 ) C h C (λ) (T 2 ). Now P h = (φ(ξ)ĥ(ξ)) = ((1 Φ(ξ))ĥ(ξ)) where Φ is a smooth function with compact support defined by 1 Φ(ξ) = φ(ξ). Hence P = P = I K where K = Φ(D). But Φ(D) : C (λ) (T 2 ) C (T 2 ) is a smoothing operator. Since C (T 2 ) is compactly contained in C (λ) (T 2 ), we get that K is a compact operator on C (λ) (T 2 ). Hence : C 1,(λ) C (λ) is invertible modulo compact operators and is a Fredholm operator of index zero. Let f = 0 on T 2. Hence f is analytic and periodic. This implies that f is a constant. Hence kernel consists of constants. Let f = h. Using integration by parts, we see that h = f = 0. T 2 T 2 This shows that h is orthogonal to the constants. But the cokernel is one-dimensional (since the map has index zero and one-dimensional kernel), so the cokernel is exactly the constants. We now have the necessary tools to to obtain estimates for the constant coefficient Beltrami equation. 38
43 Theorem 2. Let h C (λ) (T 2 ) and µ 0 and ν 0 be constants satisfying µ 0 + ν 0 κ < 1. Let L 0 denote the operator given by L 0 f = f µ 0 f ν 0 f. Then L 0 : C 1,(λ) (T 2 ) C (λ) (T 2 ) is Fredholm of index zero with kernel and cokernel consisting of constants. f C 1,(λ) (T 2 ) satisfies the equation Moreover, if (81) f µ 0 f ν 0 f = h, then we have the estimate (82) f C (λ) (T 2 ) + f C (λ) (T 2 ) K L 0 f C (λ) (T ). 2 where K is a constant depending on µ 0 and ν 0. Proof: We first observe that as in Section 4, for functions on R 2, the equation L 0 f = h can be reduced to an inhomogeneous Cauchy- Riemann equation of the form (47), i.e. we have (83) A 1 L 0 Ap = p. Here A is a transformation given by (84) Ap(z) := p(z) b p(z) 1 b 2 = f(ξ) where z = ξ a ξ 1 a 2 and A 1 is given by (85) A 1 f(ξ) := f(ξ) + b f(ξ) = p(z), where ξ = z + a z. Note that the constants a and b are chosen as in (48) and (49) such that a, b κ < 1. We first prove that there exists constants c 1 and c 2 such that (86) c 1 Ap C (λ) (R 2 ) p C (λ) (R 2 ) c 2 Ap C (λ) (R ). 2 39
44 This can be proved as follows: Consider Hence we obtain p(z) b p(z) f C (λ) (R 2 ) = 1 b 2 C (λ) (R 2 ) 1 + b 1 b 2 p C (λ) (R 2 ) 1 1 b p C (λ) (R 2 ) 1 1 κ p C (λ) (R 2 ) (87) Ap C (λ) (R 2 ) 1 1 κ p C (λ) (R 2 ) or c 1 Ap C (λ) (R 2 ) p C (λ) (R 2 ). Now consider p C (λ) (R 2 ) = f(ξ) + b f(ξ) C (λ) (R 2 ) This gives the estimate (1 + b ) f C (λ) (R 2 ) (1 + κ) f C (λ) (R 2 ). (88) p C (λ) (R 2 ) (1 + κ) Ap C (λ) (R 2 ) c 2 Ap C (λ) (R 2 ). Hence we have proved (86). Similarly we can prove (89) c 3 A 1 f C (λ) (R 2 ) f C (λ) (R 2 ) c 4 A 1 f C (λ) (R 2 ). 40
45 We now obtain the estimates (82) on R 2. Lemma 4, if we let h = p and use S =, we obtain First observe that in p C (λ) (R 2 ) C p C (λ) (R 2 ). Using (86), we can now obtain p C (λ) (R 2 ) + p C (λ) (R 2 ) C p C (λ) (R 2 ) Hence we have = C A 1 L 0 Ap C (λ) (R 2 ) C 1 L 0 Ap C (λ) (R 2 ). (90) p C (λ) (R 2 ) + p C (λ) (R 2 ) C 1 L 0 f C (λ) (R 2 ). Now need to get a relation between the C (λ) norms of the first order derivatives of f and p. Now using (46), we get f ξ(ξ) = (1 + ab) p z(z) (a + b) p z (z) (1 b 2 )(1 a 2 ) Now the C (λ) norm can be obtained as (91) f ξ C (λ) (R 2 ) 1 + ab p z C (λ) (R 2 ) + a + b p z C (λ) (R 2 ) (1 b 2 )(1 a 2 ) Similarly we can get (92) f ξ C (λ) (R 2 ) a + b p z C (λ) (R 2 )) ab p z C (λ) (R 2 ). (1 b 2 )(1 a 2 ) Adding (91) and (92), we obtain f ξ C (λ) (R 2 ) + f ξ C (λ) (R 2 ) ( p z C (λ) (R 2 ) + p z C (λ) (R 2 ) (1 + b )(1 + a ) C 1 L 0 f C (λ) (R 2 ) (1 + b )(1 + a ). 41
46 The last inequality is obtained using (90). Hence (93) f ξ C (λ) (R 2 ) + f ξ C (λ) (R 2 ) K L 0 f C (λ) (R 2 ). In particular, for periodic functions on R 2, we get the estimate f C (λ) (T 2 ) + f C (λ) (T 2 ) K( L 0 f C (λ) (R 2 ). which is our required estimate. Now adding the term f C 0 (T 2 ) on both sides of the above estimate, we obtain (94) f C 1,(λ) (T 2 ) K ( L 0 f C (λ) (T 2 ) + f C 0 (T 2 )) K ( L 0 f C (λ) (T 2 ) + f C (λ) (T 2 )). Now C 1,(λ) (T 2 ) is compactly contained in C (λ) (T 2 ). Since L 0 satisfies the above estimate, we can conclude that dimension of kernel of L 0 is finite and the range is closed, i.e. L 0 is semi-fredholm. But as we have seen, A 1 L 0 A =. Consider the homotopy G t = (1 t) + tl 0 for 0 t 1. Since we know from Proposition (2) that is an operator with index 0, using G t, we can conclude that L 0 also has index 0. Hence we have proved that L 0 : C 1,(λ) (T 2 ) C (λ) (T 2 ) is Fredholm of index zero. We use the estimate (82) to prove that the kernel consists of constants. Let L 0 f = 0. Then using (82), we obtain f C (λ) (T 2 ) + f C (λ) (T 2 ) 0. This shows that f has to be a constant and hence the kernel of L 0 consists of constants. 42
47 Now we prove that the cokernel of L 0 consists of constants. Consider L 0 f = h. Since µ 0 and ν 0 are constants, we can use integration by parts to obtain h dz = ( f µ 0 f ν 0 f) dz = 0. T 2 T 2 This shows that h is orthogonal to the constants. But we have already proved that the map L 0 has index zero and that the kernel is one dimensional. This indicates that the cokernel of L 0 should also be one-dimensional and hence it is exactly the constants. We shall hereafter use L to denote the Beltrami operator defined by (95) Lf := f µ f ν f. We need the following two lemmas in the proof of Theorem 3 below: Lemma 6. Let µ, ν C ω (T 2 ) satisfy µ(ξ) + ν(ξ) κ < 1, for all ξ T 2. If f W 1,2 (T 2 ) satisfies Lf = 0, then f is a constant. Proof: If f W 1,2 (T 2 ) satisfies Lf = 0, then f = µ f + ν f. Taking an inner product with f on both sides, we obtain (96) ( f, f) = (µ f, f) + (ν f, f). 43
48 But using the definition of inner product, we have ( f, f) = f(z) f(z)dz T 2 = 1 ( x + i y )f(x + iy) ( x i y )f(x + iy)dxdy 4 T 2 = 1 x f 2 + y f 2 + i( y f x f x f y f)dxdy 4 T 2 = 1 4 f 2 L 2 (T 2 ) where the term ( T 2 y f x f x f y f)dxdy = 0, using integration by parts. Hence we can write (96) as f 2 L 2 (T 2 ) = µ(z) f(z) f(z)dz + ν(z) f(z) f(z)dz T 2 T 2 = µ(z) f(z) f(z)dz + ν(z) f(z) f(z)dz T 2 T 2 = µ(x + iy)( x i y )f(x + iy) ( x i y ) f(x + iy)dxdy T 2 + ν(x + iy)( x + i y ) f(x + iy) ( x i y ) f(x + iy)dxdy T 2 We now obtain f 2 L 2 (T 2 ) ( µ L (T 2 ) + ν L (T 2 )) f 2 L 2 (T 2 ) κ f 2 L 2 (T 2 ). The above inequality implies that f=0, i.e. f = 0. Lemma 7. Let g C ω (T 2 ) such that g g 0 C ω (D ɛ) < C(ɛ) where g 0 = g at the center of D ɛ. Then we can find an extension g of g g 0 restricted to D ɛ such that g has support only in D 2ɛ and (97) g C (λ) (T 2 ) g g 0 C ω (D 2ɛ ). 44
49 Proof: We obtain the extension g with the desired properties by reflecting the values of g g 0 on D 2ɛ \ D ɛ i.e for each z D 2ɛ (ξ 0 ), choose ξ = 2ɛ z z z on D 2ɛ \D ɛ so that as z ɛ, ξ z and as z approaches the origin, ξ approaches the boundary of D 2ɛ. We define g by (g g 0 )(ξ) for ξ D ɛ g(ξ) = (g g 0 )(z) for ξ D 2ɛ \ D ɛ where ξ is the reflection of z in D 2ɛ (ξ 0 ). We extend g by zero outside D 2ɛ. First we need to show that this extension is C ω at the boundary of each of the disks D ɛ and D 2ɛ. First we consider the boundary of D ɛ. Let z 0 be a point on the boundary of D ɛ. For δ < ɛ and sufficiently small, consider a small disk D δ (z 0 ). We just need to consider points on either side of the boundary. For z, ξ 1 D δ (z 0 ) such that z D ɛ and ξ 1 D 2ɛ \ D ɛ, we obtain g C ω (D δ (z 0 )) = g(ξ 1 ) g(z) sup g( ξ) + sup ξ D δ (z 0 ) z,ξ 1 D δ (z 0 ) ω( ξ 1 z ) g(z 1 ) g(z) = sup g( z) + sup z D δ (z 0 ) z,ξ 1 D δ (z 0 ) ω( ξ 1 z ) where z 1 is the reflection of ξ 1 and z is the reflection of ξ in D ɛ. As δ tends to 0, z and ξ 1 approach the boundary of D ɛ and z 1 approaches the boundary of D ɛ. Now ξ 1 z = 2ɛ z 1 z 1 z 1 z = z 1 (2ɛ 1 z 1 1) z. But z 1 ɛ and z 1 as ɛ as δ 0. Hence 2ɛ 1 z We now obtain ξ 1 z z 1 z which gives ω( ξ 1 z ) ω( z 1 z ). Hence we conclude that g C ω (D δ (z 0 )) g C ω (D ɛ) 45
The double layer potential
The double layer potential In this project, our goal is to explain how the Dirichlet problem for a linear elliptic partial differential equation can be converted into an integral equation by representing
More informationWe denote the space of distributions on Ω by D ( Ω) 2.
Sep. 1 0, 008 Distributions Distributions are generalized functions. Some familiarity with the theory of distributions helps understanding of various function spaces which play important roles in the study
More informationA new class of pseudodifferential operators with mixed homogenities
A new class of pseudodifferential operators with mixed homogenities Po-Lam Yung University of Oxford Jan 20, 2014 Introduction Given a smooth distribution of hyperplanes on R N (or more generally on a
More informationOn a class of pseudodifferential operators with mixed homogeneities
On a class of pseudodifferential operators with mixed homogeneities Po-Lam Yung University of Oxford July 25, 2014 Introduction Joint work with E. Stein (and an outgrowth of work of Nagel-Ricci-Stein-Wainger,
More information8 Singular Integral Operators and L p -Regularity Theory
8 Singular Integral Operators and L p -Regularity Theory 8. Motivation See hand-written notes! 8.2 Mikhlin Multiplier Theorem Recall that the Fourier transformation F and the inverse Fourier transformation
More informationDeterminant lines and determinant line bundles
CHAPTER Determinant lines and determinant line bundles This appendix is an exposition of G. Segal s work sketched in [?] on determinant line bundles over the moduli spaces of Riemann surfaces with parametrized
More informationReminder Notes for the Course on Distribution Theory
Reminder Notes for the Course on Distribution Theory T. C. Dorlas Dublin Institute for Advanced Studies School of Theoretical Physics 10 Burlington Road, Dublin 4, Ireland. Email: dorlas@stp.dias.ie March
More informationSTABILITY OF CALDERÓN S INVERSE CONDUCTIVITY PROBLEM IN THE PLANE FOR DISCONTINUOUS CONDUCTIVITIES. Albert Clop. Daniel Faraco.
Inverse Problems and Imaging Volume 4, No., 00, 49 9 doi:0.3934/ipi.00.4.49 STABILITY OF CALDERÓN S INVERSE CONDUCTIVITY PROBLEM IN THE PLANE FOR DISCONTINUOUS CONDUCTIVITIES Albert Clop Department of
More informationChapter One. The Calderón-Zygmund Theory I: Ellipticity
Chapter One The Calderón-Zygmund Theory I: Ellipticity Our story begins with a classical situation: convolution with homogeneous, Calderón- Zygmund ( kernels on R n. Let S n 1 R n denote the unit sphere
More informationDIEUDONNE AGBOR AND JAN BOMAN
ON THE MODULUS OF CONTINUITY OF MAPPINGS BETWEEN EUCLIDEAN SPACES DIEUDONNE AGBOR AND JAN BOMAN Abstract Let f be a function from R p to R q and let Λ be a finite set of pairs (θ, η) R p R q. Assume that
More informationSingular Integrals. 1 Calderon-Zygmund decomposition
Singular Integrals Analysis III Calderon-Zygmund decomposition Let f be an integrable function f dx 0, f = g + b with g Cα almost everywhere, with b
More informationAlgebras of singular integral operators with kernels controlled by multiple norms
Algebras of singular integral operators with kernels controlled by multiple norms Alexander Nagel Conference in Harmonic Analysis in Honor of Michael Christ This is a report on joint work with Fulvio Ricci,
More informationJUHA KINNUNEN. Harmonic Analysis
JUHA KINNUNEN Harmonic Analysis Department of Mathematics and Systems Analysis, Aalto University 27 Contents Calderón-Zygmund decomposition. Dyadic subcubes of a cube.........................2 Dyadic cubes
More informationMATH 205C: STATIONARY PHASE LEMMA
MATH 205C: STATIONARY PHASE LEMMA For ω, consider an integral of the form I(ω) = e iωf(x) u(x) dx, where u Cc (R n ) complex valued, with support in a compact set K, and f C (R n ) real valued. Thus, I(ω)
More informationQuasi-conformal maps and Beltrami equation
Chapter 7 Quasi-conformal maps and Beltrami equation 7. Linear distortion Assume that f(x + iy) =u(x + iy)+iv(x + iy) be a (real) linear map from C C that is orientation preserving. Let z = x + iy and
More informationare Banach algebras. f(x)g(x) max Example 7.4. Similarly, A = L and A = l with the pointwise multiplication
7. Banach algebras Definition 7.1. A is called a Banach algebra (with unit) if: (1) A is a Banach space; (2) There is a multiplication A A A that has the following properties: (xy)z = x(yz), (x + y)z =
More information2 Infinite products and existence of compactly supported φ
415 Wavelets 1 Infinite products and existence of compactly supported φ Infinite products.1 Infinite products a n need to be defined via limits. But we do not simply say that a n = lim a n N whenever the
More informationTraces, extensions and co-normal derivatives for elliptic systems on Lipschitz domains
Traces, extensions and co-normal derivatives for elliptic systems on Lipschitz domains Sergey E. Mikhailov Brunel University West London, Department of Mathematics, Uxbridge, UB8 3PH, UK J. Math. Analysis
More informationComplex Analysis, Stein and Shakarchi Meromorphic Functions and the Logarithm
Complex Analysis, Stein and Shakarchi Chapter 3 Meromorphic Functions and the Logarithm Yung-Hsiang Huang 217.11.5 Exercises 1. From the identity sin πz = eiπz e iπz 2i, it s easy to show its zeros are
More informationINTRODUCTION TO REAL ANALYTIC GEOMETRY
INTRODUCTION TO REAL ANALYTIC GEOMETRY KRZYSZTOF KURDYKA 1. Analytic functions in several variables 1.1. Summable families. Let (E, ) be a normed space over the field R or C, dim E
More informationFourier Transform & Sobolev Spaces
Fourier Transform & Sobolev Spaces Michael Reiter, Arthur Schuster Summer Term 2008 Abstract We introduce the concept of weak derivative that allows us to define new interesting Hilbert spaces the Sobolev
More informationNotes on Complex Analysis
Michael Papadimitrakis Notes on Complex Analysis Department of Mathematics University of Crete Contents The complex plane.. The complex plane...................................2 Argument and polar representation.........................
More informationGLOBAL WELL-POSEDNESS IN SPATIALLY CRITICAL BESOV SPACE FOR THE BOLTZMANN EQUATION
GLOBAL WELL-POSEDNESS IN SPATIALLY CRITICAL BESOV SPACE FOR THE BOLTZMANN EQUATION RENJUN DUAN, SHUANGQIAN LIU, AND JIANG XU Abstract. The unique global strong solution in the Chemin-Lerner type space
More informationParaproducts and the bilinear Calderón-Zygmund theory
Paraproducts and the bilinear Calderón-Zygmund theory Diego Maldonado Department of Mathematics Kansas State University Manhattan, KS 66506 12th New Mexico Analysis Seminar April 23-25, 2009 Outline of
More informationFollow links Class Use and other Permissions. For more information, send to:
COPYRIGHT NOTICE: Kari Astala, Tadeusz Iwaniec & Gaven Martin: Elliptic Partial Differential Equations and Quasiconformal Mappings in the Plane is published by Princeton University Press and copyrighted,
More informationMATH 722, COMPLEX ANALYSIS, SPRING 2009 PART 5
MATH 722, COMPLEX ANALYSIS, SPRING 2009 PART 5.. The Arzela-Ascoli Theorem.. The Riemann mapping theorem Let X be a metric space, and let F be a family of continuous complex-valued functions on X. We have
More informationSolution to Homework 4, Math 7651, Tanveer 1. Show that the function w(z) = 1 2
Solution to Homework 4, Math 7651, Tanveer 1. Show that the function w(z) = 1 (z + 1/z) maps the exterior of a unit circle centered around the origin in the z-plane to the exterior of a straight line cut
More informationEXPOSITORY NOTES ON DISTRIBUTION THEORY, FALL 2018
EXPOSITORY NOTES ON DISTRIBUTION THEORY, FALL 2018 While these notes are under construction, I expect there will be many typos. The main reference for this is volume 1 of Hörmander, The analysis of liner
More informationTopics in Harmonic Analysis Lecture 6: Pseudodifferential calculus and almost orthogonality
Topics in Harmonic Analysis Lecture 6: Pseudodifferential calculus and almost orthogonality Po-Lam Yung The Chinese University of Hong Kong Introduction While multiplier operators are very useful in studying
More informationCOMPLEX ANALYSIS Spring 2014
COMPLEX ANALYSIS Spring 204 Cauchy and Runge Under the Same Roof. These notes can be used as an alternative to Section 5.5 of Chapter 2 in the textbook. They assume the theorem on winding numbers of the
More informationMath The Laplacian. 1 Green s Identities, Fundamental Solution
Math. 209 The Laplacian Green s Identities, Fundamental Solution Let be a bounded open set in R n, n 2, with smooth boundary. The fact that the boundary is smooth means that at each point x the external
More informationCalderon s inverse conductivity problem in the plane
Calderon s inverse conductivity problem in the plane Kari Astala and Lassi Päivärinta Abstract We show that the Dirichlet to Neumann map for the equation σ u = 0 in a two dimensional domain uniquely determines
More informationDuality of multiparameter Hardy spaces H p on spaces of homogeneous type
Duality of multiparameter Hardy spaces H p on spaces of homogeneous type Yongsheng Han, Ji Li, and Guozhen Lu Department of Mathematics Vanderbilt University Nashville, TN Internet Analysis Seminar 2012
More informationTools from Lebesgue integration
Tools from Lebesgue integration E.P. van den Ban Fall 2005 Introduction In these notes we describe some of the basic tools from the theory of Lebesgue integration. Definitions and results will be given
More information1 Math 241A-B Homework Problem List for F2015 and W2016
1 Math 241A-B Homework Problem List for F2015 W2016 1.1 Homework 1. Due Wednesday, October 7, 2015 Notation 1.1 Let U be any set, g be a positive function on U, Y be a normed space. For any f : U Y let
More informationA local estimate from Radon transform and stability of Inverse EIT with partial data
A local estimate from Radon transform and stability of Inverse EIT with partial data Alberto Ruiz Universidad Autónoma de Madrid U. California, Irvine.June 2012 w/ P. Caro (U. Helsinki) and D. Dos Santos
More informationMicrolocal Methods in X-ray Tomography
Microlocal Methods in X-ray Tomography Plamen Stefanov Purdue University Lecture I: Euclidean X-ray tomography Mini Course, Fields Institute, 2012 Plamen Stefanov (Purdue University ) Microlocal Methods
More informationCOMPLEX ANALYSIS Spring 2014
COMPLEX ANALYSIS Spring 24 Homework 4 Solutions Exercise Do and hand in exercise, Chapter 3, p. 4. Solution. The exercise states: Show that if a
More informationMichael Lacey and Christoph Thiele. f(ξ)e 2πiξx dξ
Mathematical Research Letters 7, 36 370 (2000) A PROOF OF BOUNDEDNESS OF THE CARLESON OPERATOR Michael Lacey and Christoph Thiele Abstract. We give a simplified proof that the Carleson operator is of weaktype
More informationWEIERSTRASS THEOREMS AND RINGS OF HOLOMORPHIC FUNCTIONS
WEIERSTRASS THEOREMS AND RINGS OF HOLOMORPHIC FUNCTIONS YIFEI ZHAO Contents. The Weierstrass factorization theorem 2. The Weierstrass preparation theorem 6 3. The Weierstrass division theorem 8 References
More information11 COMPLEX ANALYSIS IN C. 1.1 Holomorphic Functions
11 COMPLEX ANALYSIS IN C 1.1 Holomorphic Functions A domain Ω in the complex plane C is a connected, open subset of C. Let z o Ω and f a map f : Ω C. We say that f is real differentiable at z o if there
More information4 Riesz Kernels. Since the functions k i (ξ) = ξ i. are bounded functions it is clear that R
4 Riesz Kernels. A natural generalization of the Hilbert transform to higher dimension is mutiplication of the Fourier Transform by homogeneous functions of degree 0, the simplest ones being R i f(ξ) =
More informationEuler Equations: local existence
Euler Equations: local existence Mat 529, Lesson 2. 1 Active scalars formulation We start with a lemma. Lemma 1. Assume that w is a magnetization variable, i.e. t w + u w + ( u) w = 0. If u = Pw then u
More informationPROBLEMS. (b) (Polarization Identity) Show that in any inner product space
1 Professor Carl Cowen Math 54600 Fall 09 PROBLEMS 1. (Geometry in Inner Product Spaces) (a) (Parallelogram Law) Show that in any inner product space x + y 2 + x y 2 = 2( x 2 + y 2 ). (b) (Polarization
More informationInvertibility of elliptic operators
CHAPTER 6 Invertibility of elliptic operators Next we will use the local elliptic estimates obtained earlier on open sets in R n to analyse the global invertibility properties of elliptic operators on
More informationPOLYNOMIALS WITH COEFFICIENTS FROM A FINITE SET
TRANSACTIONS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 00, Number 0, Pages 000 000 S 0002-9947(XX)0000-0 POLYNOMIALS WITH COEFFICIENTS FROM A FINITE SET PETER BORWEIN, TAMÁS ERDÉLYI, FRIEDRICH LITTMANN
More informationFourier transform of tempered distributions
Fourier transform of tempered distributions 1 Test functions and distributions As we have seen before, many functions are not classical in the sense that they cannot be evaluated at any point. For eample,
More informationMORE NOTES FOR MATH 823, FALL 2007
MORE NOTES FOR MATH 83, FALL 007 Prop 1.1 Prop 1. Lemma 1.3 1. The Siegel upper half space 1.1. The Siegel upper half space and its Bergman kernel. The Siegel upper half space is the domain { U n+1 z C
More informationA LOWER BOUND ON BLOWUP RATES FOR THE 3D INCOMPRESSIBLE EULER EQUATION AND A SINGLE EXPONENTIAL BEALE-KATO-MAJDA ESTIMATE. 1.
A LOWER BOUND ON BLOWUP RATES FOR THE 3D INCOMPRESSIBLE EULER EQUATION AND A SINGLE EXPONENTIAL BEALE-KATO-MAJDA ESTIMATE THOMAS CHEN AND NATAŠA PAVLOVIĆ Abstract. We prove a Beale-Kato-Majda criterion
More informationQualifying Exams I, 2014 Spring
Qualifying Exams I, 2014 Spring 1. (Algebra) Let k = F q be a finite field with q elements. Count the number of monic irreducible polynomials of degree 12 over k. 2. (Algebraic Geometry) (a) Show that
More informationLittlewood-Paley theory
Chapitre 6 Littlewood-Paley theory Introduction The purpose of this chapter is the introduction by this theory which is nothing but a precise way of counting derivatives using the localization in the frequency
More informationIV. Conformal Maps. 1. Geometric interpretation of differentiability. 2. Automorphisms of the Riemann sphere: Möbius transformations
MTH6111 Complex Analysis 2009-10 Lecture Notes c Shaun Bullett 2009 IV. Conformal Maps 1. Geometric interpretation of differentiability We saw from the definition of complex differentiability that if f
More informationRudin Real and Complex Analysis - Harmonic Functions
Rudin Real and Complex Analysis - Harmonic Functions Aaron Lou December 2018 1 Notes 1.1 The Cauchy-Riemann Equations 11.1: The Operators and Suppose f is a complex function defined in a plane open set
More informationComposition operators: the essential norm and norm-attaining
Composition operators: the essential norm and norm-attaining Mikael Lindström Department of Mathematical Sciences University of Oulu Valencia, April, 2011 The purpose of this talk is to first discuss the
More informationCalderón s inverse problem in 2D Electrical Impedance Tomography
Calderón s inverse problem in 2D Electrical Impedance Tomography Kari Astala (University of Helsinki) Joint work with: Matti Lassas, Lassi Päivärinta, Samuli Siltanen, Jennifer Mueller and Alan Perämäki
More informationThe purpose of this lecture is to present a few applications of conformal mappings in problems which arise in physics and engineering.
Lecture 16 Applications of Conformal Mapping MATH-GA 451.001 Complex Variables The purpose of this lecture is to present a few applications of conformal mappings in problems which arise in physics and
More informationFRAMES AND TIME-FREQUENCY ANALYSIS
FRAMES AND TIME-FREQUENCY ANALYSIS LECTURE 5: MODULATION SPACES AND APPLICATIONS Christopher Heil Georgia Tech heil@math.gatech.edu http://www.math.gatech.edu/ heil READING For background on Banach spaces,
More informationThe NLS on product spaces and applications
October 2014, Orsay The NLS on product spaces and applications Nikolay Tzvetkov Cergy-Pontoise University based on joint work with Zaher Hani, Benoit Pausader and Nicola Visciglia A basic result Consider
More informationIntroduction to Real Analysis Alternative Chapter 1
Christopher Heil Introduction to Real Analysis Alternative Chapter 1 A Primer on Norms and Banach Spaces Last Updated: March 10, 2018 c 2018 by Christopher Heil Chapter 1 A Primer on Norms and Banach Spaces
More informationWavelets and modular inequalities in variable L p spaces
Wavelets and modular inequalities in variable L p spaces Mitsuo Izuki July 14, 2007 Abstract The aim of this paper is to characterize variable L p spaces L p( ) (R n ) using wavelets with proper smoothness
More informationSCALE INVARIANT FOURIER RESTRICTION TO A HYPERBOLIC SURFACE
SCALE INVARIANT FOURIER RESTRICTION TO A HYPERBOLIC SURFACE BETSY STOVALL Abstract. This result sharpens the bilinear to linear deduction of Lee and Vargas for extension estimates on the hyperbolic paraboloid
More informationQualifying Exam Complex Analysis (Math 530) January 2019
Qualifying Exam Complex Analysis (Math 53) January 219 1. Let D be a domain. A function f : D C is antiholomorphic if for every z D the limit f(z + h) f(z) lim h h exists. Write f(z) = f(x + iy) = u(x,
More informationMotivation towards Clifford Analysis: The classical Cauchy s Integral Theorem from a Clifford perspective
Motivation towards Clifford Analysis: The classical Cauchy s Integral Theorem from a Clifford perspective Lashi Bandara November 26, 29 Abstract Clifford Algebras generalise complex variables algebraically
More informationCHAPTER 6. Differentiation
CHPTER 6 Differentiation The generalization from elementary calculus of differentiation in measure theory is less obvious than that of integration, and the methods of treating it are somewhat involved.
More informationA local estimate from Radon transform and stability of Inverse EIT with partial data
A local estimate from Radon transform and stability of Inverse EIT with partial data Alberto Ruiz Universidad Autónoma de Madrid ge Joint work with P. Caro (U. Helsinki) and D. Dos Santos Ferreira (Paris
More informationREAL ANALYSIS ANALYSIS NOTES. 0: Some basics. Notes by Eamon Quinlan. Liminfs and Limsups
ANALYSIS NOTES Notes by Eamon Quinlan REAL ANALYSIS 0: Some basics Liminfs and Limsups Def.- Let (x n ) R be a sequence. The limit inferior of (x n ) is defined by and, similarly, the limit superior of
More informationMicrolocal analysis and inverse problems Lecture 3 : Carleman estimates
Microlocal analysis and inverse problems ecture 3 : Carleman estimates David Dos Santos Ferreira AGA Université de Paris 13 Monday May 16 Instituto de Ciencias Matemáticas, Madrid David Dos Santos Ferreira
More informationMA3111S COMPLEX ANALYSIS I
MA3111S COMPLEX ANALYSIS I 1. The Algebra of Complex Numbers A complex number is an expression of the form a + ib, where a and b are real numbers. a is called the real part of a + ib and b the imaginary
More information5 Measure theory II. (or. lim. Prove the proposition. 5. For fixed F A and φ M define the restriction of φ on F by writing.
5 Measure theory II 1. Charges (signed measures). Let (Ω, A) be a σ -algebra. A map φ: A R is called a charge, (or signed measure or σ -additive set function) if φ = φ(a j ) (5.1) A j for any disjoint
More informationLECTURE-13 : GENERALIZED CAUCHY S THEOREM
LECTURE-3 : GENERALIZED CAUCHY S THEOREM VED V. DATAR The aim of this lecture to prove a general form of Cauchy s theorem applicable to multiply connected domains. We end with computations of some real
More informationRANDOM PROPERTIES BENOIT PAUSADER
RANDOM PROPERTIES BENOIT PAUSADER. Quasilinear problems In general, one consider the following trichotomy for nonlinear PDEs: A semilinear problem is a problem where the highest-order terms appears linearly
More informationA Critical Parabolic Sobolev Embedding via Littlewood-Paley Decomposition
International Mathematical Forum, Vol. 7, 01, no. 35, 1705-174 A Critical Parabolic Sobolev Embedding via Littlewood-Paley Decomposition H. Ibrahim Lebanese University, Faculty of Sciences Mathematics
More informationNotes. 1 Fourier transform and L p spaces. March 9, For a function in f L 1 (R n ) define the Fourier transform. ˆf(ξ) = f(x)e 2πi x,ξ dx.
Notes March 9, 27 1 Fourier transform and L p spaces For a function in f L 1 (R n ) define the Fourier transform ˆf(ξ) = f(x)e 2πi x,ξ dx. Properties R n 1. f g = ˆfĝ 2. δλ (f)(ξ) = ˆf(λξ), where δ λ f(x)
More informationThe Hilbert transform
The Hilbert transform Definition and properties ecall the distribution pv(, defined by pv(/(ϕ := lim ɛ ɛ ϕ( d. The Hilbert transform is defined via the convolution with pv(/, namely (Hf( := π lim f( t
More informationComplex geometrical optics solutions for Lipschitz conductivities
Rev. Mat. Iberoamericana 19 (2003), 57 72 Complex geometrical optics solutions for Lipschitz conductivities Lassi Päivärinta, Alexander Panchenko and Gunther Uhlmann Abstract We prove the existence of
More informationTHE MATTILA INTEGRAL ASSOCIATED WITH SIGN INDEFINITE MEASURES. Alex Iosevich and Misha Rudnev. August 3, Introduction
THE MATTILA INTEGRAL ASSOCIATED WITH SIGN INDEFINITE MEASURES Alex Iosevich and Misha Rudnev August 3, 005 Abstract. In order to quantitatively illustrate the role of positivity in the Falconer distance
More informationAn introduction to Birkhoff normal form
An introduction to Birkhoff normal form Dario Bambusi Dipartimento di Matematica, Universitá di Milano via Saldini 50, 0133 Milano (Italy) 19.11.14 1 Introduction The aim of this note is to present an
More information1.3.1 Definition and Basic Properties of Convolution
1.3 Convolution 15 1.3 Convolution Since L 1 (R) is a Banach space, we know that it has many useful properties. In particular the operations of addition and scalar multiplication are continuous. However,
More informationConservation law equations : problem set
Conservation law equations : problem set Luis Silvestre For Isaac Neal and Elia Portnoy in the 2018 summer bootcamp 1 Method of characteristics For the problems in this section, assume that the solutions
More informationClassical Fourier Analysis
Loukas Grafakos Classical Fourier Analysis Second Edition 4y Springer 1 IP Spaces and Interpolation 1 1.1 V and Weak IP 1 1.1.1 The Distribution Function 2 1.1.2 Convergence in Measure 5 1.1.3 A First
More information(x 1, y 1 ) = (x 2, y 2 ) if and only if x 1 = x 2 and y 1 = y 2.
1. Complex numbers A complex number z is defined as an ordered pair z = (x, y), where x and y are a pair of real numbers. In usual notation, we write z = x + iy, where i is a symbol. The operations of
More informationA review: The Laplacian and the d Alembertian. j=1
Chapter One A review: The Laplacian and the d Alembertian 1.1 THE LAPLACIAN One of the main goals of this course is to understand well the solution of wave equation both in Euclidean space and on manifolds
More informationMATH MEASURE THEORY AND FOURIER ANALYSIS. Contents
MATH 3969 - MEASURE THEORY AND FOURIER ANALYSIS ANDREW TULLOCH Contents 1. Measure Theory 2 1.1. Properties of Measures 3 1.2. Constructing σ-algebras and measures 3 1.3. Properties of the Lebesgue measure
More informationComplex Analysis, Stein and Shakarchi The Fourier Transform
Complex Analysis, Stein and Shakarchi Chapter 4 The Fourier Transform Yung-Hsiang Huang 2017.11.05 1 Exercises 1. Suppose f L 1 (), and f 0. Show that f 0. emark 1. This proof is observed by Newmann (published
More informationComplex Variables. Cathal Ormond
Complex Variables Cathal Ormond Contents 1 Introduction 3 1.1 Definition: Polar Form.............................. 3 1.2 Definition: Length................................ 3 1.3 Definitions.....................................
More informationIII. Consequences of Cauchy s Theorem
MTH6 Complex Analysis 2009-0 Lecture Notes c Shaun Bullett 2009 III. Consequences of Cauchy s Theorem. Cauchy s formulae. Cauchy s Integral Formula Let f be holomorphic on and everywhere inside a simple
More informationAn Introduction to Complex Analysis and Geometry John P. D Angelo, Pure and Applied Undergraduate Texts Volume 12, American Mathematical Society, 2010
An Introduction to Complex Analysis and Geometry John P. D Angelo, Pure and Applied Undergraduate Texts Volume 12, American Mathematical Society, 2010 John P. D Angelo, Univ. of Illinois, Urbana IL 61801.
More informationTHEOREMS, ETC., FOR MATH 515
THEOREMS, ETC., FOR MATH 515 Proposition 1 (=comment on page 17). If A is an algebra, then any finite union or finite intersection of sets in A is also in A. Proposition 2 (=Proposition 1.1). For every
More informationUSING FUNCTIONAL ANALYSIS AND SOBOLEV SPACES TO SOLVE POISSON S EQUATION
USING FUNCTIONAL ANALYSIS AND SOBOLEV SPACES TO SOLVE POISSON S EQUATION YI WANG Abstract. We study Banach and Hilbert spaces with an eye towards defining weak solutions to elliptic PDE. Using Lax-Milgram
More informationWave operators with non-lipschitz coefficients: energy and observability estimates
Wave operators with non-lipschitz coefficients: energy and observability estimates Institut de Mathématiques de Jussieu-Paris Rive Gauche UNIVERSITÉ PARIS DIDEROT PARIS 7 JOURNÉE JEUNES CONTRÔLEURS 2014
More informationl(y j ) = 0 for all y j (1)
Problem 1. The closed linear span of a subset {y j } of a normed vector space is defined as the intersection of all closed subspaces containing all y j and thus the smallest such subspace. 1 Show that
More informationOutline of Fourier Series: Math 201B
Outline of Fourier Series: Math 201B February 24, 2011 1 Functions and convolutions 1.1 Periodic functions Periodic functions. Let = R/(2πZ) denote the circle, or onedimensional torus. A function f : C
More informationSPRING 2008: POLYNOMIAL IMAGES OF CIRCLES
18.821 SPRING 28: POLYNOMIAL IMAGES OF CIRCLES JUSTIN CURRY, MICHAEL FORBES, MATTHEW GORDON Abstract. This paper considers the effect of complex polynomial maps on circles of various radii. Several phenomena
More informationA NEW CLASS OF OPERATORS AND A DESCRIPTION OF ADJOINTS OF COMPOSITION OPERATORS
A NEW CLASS OF OPERATORS AND A DESCRIPTION OF ADJOINTS OF COMPOSITION OPERATORS CARL C. COWEN AND EVA A. GALLARDO-GUTIÉRREZ Abstract. Starting with a general formula, precise but difficult to use, for
More informationSCHWARTZ SPACES ASSOCIATED WITH SOME NON-DIFFERENTIAL CONVOLUTION OPERATORS ON HOMOGENEOUS GROUPS
C O L L O Q U I U M M A T H E M A T I C U M VOL. LXIII 1992 FASC. 2 SCHWARTZ SPACES ASSOCIATED WITH SOME NON-DIFFERENTIAL CONVOLUTION OPERATORS ON HOMOGENEOUS GROUPS BY JACEK D Z I U B A Ń S K I (WROC
More informationTOOLS FROM HARMONIC ANALYSIS
TOOLS FROM HARMONIC ANALYSIS BRADLY STADIE Abstract. The Fourier transform can be thought of as a map that decomposes a function into oscillatory functions. In this paper, we will apply this decomposition
More informationComposition Operators on Hilbert Spaces of Analytic Functions
Composition Operators on Hilbert Spaces of Analytic Functions Carl C. Cowen IUPUI (Indiana University Purdue University Indianapolis) and Purdue University First International Conference on Mathematics
More informationCORRIGENDUM: THE SYMPLECTIC SUM FORMULA FOR GROMOV-WITTEN INVARIANTS
CORRIGENDUM: THE SYMPLECTIC SUM FORMULA FOR GROMOV-WITTEN INVARIANTS ELENY-NICOLETA IONEL AND THOMAS H. PARKER Abstract. We correct an error and an oversight in [IP]. The sign of the curvature in (8.7)
More informationWavelets and regularization of the Cauchy problem for the Laplace equation
J. Math. Anal. Appl. 338 008440 1447 www.elsevier.com/locate/jmaa Wavelets and regularization of the Cauchy problem for the Laplace equation Chun-Yu Qiu, Chu-Li Fu School of Mathematics and Statistics,
More informationSelected Solutions To Problems in Complex Analysis
Selected Solutions To Problems in Complex Analysis E. Chernysh November 3, 6 Contents Page 8 Problem................................... Problem 4................................... Problem 5...................................
More information