Inverse Scattering Theory

Size: px
Start display at page:

Download "Inverse Scattering Theory"

Transcription

1 Chapter 1 Inverse Scattering Theory In this introductory chapter we provide an overview of the basic ideas of inverse scattering theory for the special case when the scattering object is an isotropic inhomogeneous medium of compact support and the focus is on the inverse problem of determining the support of the scattering object from a knowledge of the scattering data. The results presented here are basic to the chapters that follow which develop in more detail the role of the transmission eigenvalues in inverse scattering theory including the case of anisotropic media. 1.1 The Helmholtz Equation The starting point of any discussion of classical scattering theory is the Helmholtz equation and in particular spherical Bessel functions and spherical harmonics which arise when separation of variables is implemented in spherical coordinates. More specifically, we look for solutions of the Helmholtz equation in R 3 u + k 2 u =0 for k>0 in the form u(x) =f(k x )Y m n (ˆx) where x 2 R 3,ˆx := x/ x, Yn m (ˆx) isaspherical harmonic defined by s Yn m 2n +1(n m )! (, ):= 4 (n + m )! P n m (cos )e im, m = n,..., n, n =0, 1, 2,...,(, ) are the spherical angles of ˆx and P m n is an associated Legendre polynomial. We note here that {Y m n } is a complete 1

2 CHAPTER 1. INVERSE SCATTERING THEORY 2 orthonormal system in L 2 (S 2 )where S 2 := {x : x =1} and Y 0 0 = 1 p 4.Thenf is a solution of the spherical Bessel equation t 2 f 00 (t)+2tf 0 (t)+ t 2 n(n 1) f(t) = 0 (1.1) with two linearly independent solutions j n (t) := 1X p=0 y n (t) := (2n)! 2 n n! ( 1) p t n+2p 2 p p!1 3 (2n +2p + 1) 1X p=0 ( 1) p t 2p n 1 2 p p!( 2n + 1)( 2n + 3) ( 2n +2p 1) (1.2) called, respectively, the spherical Bessel function and the spherical Neumann function of order n. We note that The functions j 0 (t) = sin t t, y 0 (t) = cos t. (1.3) t h (1) n (t) :=j n (t)+iy n (t) h (2) n (t) :=j n (t) iy n (t) are called, respectively, the spherical Hankel functions of the first and second kind of order n. From (1.2) and (1.3) we have that for f n = j n or f n = y n that d f n+1 (t) = t n t n f n (t) dt for n =0, 1, 2,... and h (1) 0 (t) =eit it, h (2) e it 0 (t) = it. From this we see that the spherical Hankel functions have the asymptotic behavior h (1) n (t) = 1 t ei(t n 2 2 ) 1 1+O t h (2) n (t) = 1 (1.4) t e i(t n 2 2 ) 1 1+O t

3 CHAPTER 1. INVERSE SCATTERING THEORY 3 as t tends to infinity. In particular, h (1) n (kr) satisfies the Sommerfield radiation lim r iku =0, i.e. if u(x) =h (1) n (kr)yn m (ˆx) thenu(x)e i!t (where! is the frequency and t is time) is and outgoing wave. Solutions of the Helmholtz equation satisfying the Sommerfield radiation condition uniformly in ˆx are called radiating. The Wronskian of h (1) n (t) and h (2) n (t) is given by W h (1) n,h (2) n :=h (1) n (t)h (2)0 n (t) h (2) n (t)h (1)0 n (t) = 2i t 2. (1.5) Now let be a bounded domain such that R 3 \ is connected and assume is of class C 2 with unit outward normal directed into the exterior of. Let (x, y) := 1 eik x y 4 x y, x 6= y be the radiating fundamental solution to the Helmholtz equation. using Green s second identity (u v v u) dx = Then we can deduce Green s formula for functions u 2 C 2 () \ C 1 u(x) = (x, y) u (x, (1.6) u + k 2 u (x, y) dy, x 2. Theorem 1.1. Let u 2 C 2 () be a solution of the Helmholtz equation in. Then u is analytic in, i.e. u can be locally expanded in a power series for each point x 2. Proof. Let x 2 and choose a closed ball contained in with center x. Apply Green s formula to the ball and note that for x 6= y we have that (x, y) is real analytic in x.

4 CHAPTER 1. INVERSE SCATTERING THEORY 4 Theorem 1.2 (Holmgren s Theorem). Let u 2 C 2 ()\C 1 () be a solution to the Helmholtz equation in such that u =0on for some open Then u is identically zero in. Proof. Using (1.6) we can extend u u(x) := (x, y) u (x, y) ds(y) for x 2 (R 3 \ ) [. By Green s second identity applied to u and (x, ) we see that u =0inR 3 \. Butu is a solution of the Helmholtz equation in (R 3 [ and hence by the analyticity of u we have that u =0in. We now derive a representation formula analogous to (1.6) for radiating solutions of the Helmholtz equation in R 3 \. Part of the proof of this theorem will also be use at the end of this section in order to provide a uniqueness theorem for radiating solutions of the Helmholtz equation. Theorem 1.3. Let u 2 C 2 (R 3 \ ) \ C 1 (R 3 \ ) be a radiating solution to the Helmholtz equation. Then we have Green s (x, u(x) = u (y) (x, y) ds(y), x 2 R3 Proof. Let S r := {x : x = r}. Then the Sommerfeld radiation condition implies that ( + k 2 u 2 +2kIm ) ds S r = S iku 2 ds! 0 (1.7) as r tends to infinity. We now assume that r is large enough such that is contained in the ball bounded by S r and apply Green s first identity (u v + ru rv) dx =

5 CHAPTER 1. INVERSE SCATTERING THEORY 5 to r := x 2 R 3 \ : x <r to u ds = ds k2 u 2 dy + r S r r!1 S r ru 2 dy. (1.8) Taking the imaginary part of (1.8) and substituting into (1.7) gives ( ) lim + k 2 u 2 ds = 2k Im ds (1.9) which implies that S r u 2 ds = O(1),r!1. Using the Cauchy Schwarz inequality and the Sommerfeld radiation condition we now have (x, u (x, y) ds(y) (y) S (x, y) = u ik (x, y) ds(y) (y) S (x, y) iku ds(y)! 0 S r as r tends to infinity. Hence, applying Green s formula (1.6) to r and letting r tend to infinity gives the theorem. Corollary 1.4. An entire solution to the Helmholtz equation satisfying the Sommerfeld radiation condition must vanish identically. Proof. This follows immediately from Green s formula and Green s second identity. Corollary 1.5. Every radiating solution u to the Helmholtz equation has the asymptotic behavior of an outgoing spherical wave u(x) = eik x 1 u 1 (ˆx)+O x x 2, x!1 uniformly in all directions ˆx = x/ x. The function u 1 defined on the unit sphere S 2 is called the far field pattern of

6 CHAPTER 1. INVERSE SCATTERING THEORY 6 Proof. From we obtain and x y (y) q x 2 2x y + y 2 1 = x ˆx y +O x eik x y x y = eik x 1 e ikˆx y +O x x eik x y x y = eik x (y) e ikˆx y +O 1 x as x!1uniformly for all y The Corollary now follows by substituting into Green s formula. The next result is a cornerstone of scattering theory and will be used repeatedly in the sequel. Theorem 1.6 (Rellich s Lemma). Let u 2 C 2 (R 3 \ ) be a solution to the Helmholtz equation satisfying lim u(x) 2 dx =0. Then u =0in R 3 \. r!1 x =r Proof. For x su ciently large we have that where and x =r u(x) = 1X a m n (r) = nx n=0 m= n u(x) 2 ds = r 2 a m n (r)y m n (ˆx) S 2 u(rˆx)y m n (ˆx) ds(ˆx) (1.10) 1 X nx n=0 m= n The assumption of the theorem implies that a m n (r) 2. lim r!1 r2 a m n (r) 2 =0. (1.11)

7 CHAPTER 1. INVERSE SCATTERING THEORY 7 But from (1.10) and the fact that u is a solution of the Helmholtz equation we can deduce that the a m n (r) are solutions of the spherical Bessel equation (1.1), i.e. a m n (r) = n m h (1) n (kr)+ n m h (2) n (kr) (1.12) where n m and n m are constants. Substituting (1.12) into (1.11) and using the asymptotic formulae (1.4) now implies that n m = n m = 0 for all n, m and hence u = 0 outside a su ciently large ball. This implies that u =0in R 3 \ by analyticity (Theorem 1.1). Corollary 1.7. Assume u 2 C 2 (R 3 \)\C 1 (R 3 \) is a radiating solution to the Helmholtz equation such that Im ds 0. Then u =0in R 3 Proof. From (1.9) and the assumption of the theorem we have that the assumption of Rellich s lemma is valid. 1.2 Scattering by an Inhomogeneous Medium We will now present the scattering problem that will be the main focus of these introductory lectures on inverse scattering theory. We begin by considering the propagation of sound waves of small amplitude in R 3 viewed as a problem in fluid dynamics. Let v(x, t), x 2 R 3, be the velocity potential of a fluid particle in an inviscid fluid and let p(x, t) bethepressure, (x, t) the density and S(x, t) the specific entropy. Then, if there are no external forces, we +(v r)v + 1 r + r( v) = 0 p = + v rs = 0 (Euler s equation) (equation of continuity) (equation of state) (adiabatic hypothesis) where f is a function depending on the fluid. Assuming that v(x, t), p(x, t), (x, t) and S(x, t) are small we perturb around the static case v = 0, p =

8 CHAPTER 1. INVERSE SCATTERING THEORY 8 p 0 = constant, = 0 (x), S = S 0 (x) withp 0 = f( 0,S 0 ): v(x, t) = v 1 (x, t)+o( 2 ) p(x, t) =p 0 + p 1 (x, t)+o( 2 ) (x, t) = 0 (x)+ 1 (x, t)+o( 2 ) S(x, t) =S 0 (x)+ S 1 (x, t)+o( 2 ) where 0 < 1. Substituting the above into the equations of motion and equating the coe cients of we arrive + 1 rp 1 =0 + r( 0v 1 + c2 (x) where the sound speed c is defined by 2 p + v 1 r 0 c 2 f ( 0(x),S 0 (x)). 1 = c2 (x) 0 (x)r 0 (x) rp 1. If p 1 (x, t) =Re u(x)e iwt we have that u satisfies 1 0 (x)r 0 (x) ru + w2 c 2 (x) u =0. Making the further assumption that r 0 can be ignored, we arrive at u + w2 c 2 u =0. (1.13) (x) We now assume that the slowly varying inhomogeneous medium is of compact support and is embedded in R 3 where the sound speed is c(x) =c 0 = constant. If the wave motion is caused by an incident field u i satisfying (1.13) with c(x) =c 0, we arrive at the scattering problem of determining u such that u + k 2 n(x)u =0 inr 3 (1.14)

9 CHAPTER 1. INVERSE SCATTERING THEORY 9 u = u i + u s s lim r iku s = 0 (1.16) where n(x) = 1 outside the inhomogeneous medium, n(x) = c2 0 c 2 (x) inside the inhomogeneous medium, r = x, the radiation condition (1.16) is valid uniformly with respect to ˆx = x/ x, k = w/c 0 > 0isthewave number, u i is an entire solution of the Helmholtz equation u + k 2 u = 0, u s is the scattered field and we refer to the function n(x) as the refractive index (In the engineering and physics literature c 0 /c(x) is the refractive index). The scattering problem (1.14)-(1.16) is the simplest model in which to introduce the basic ideas of inverse scattering theory. However, we shall later consider more physically realistic models in which we no longer ignore r 0 and allow u to have jump discontinuities across the boundary of the inhomogeneous media (c.f. Section 1.9). We now assume that n(x) is piecewise continuous such that n(x) > 0 and set m := 1 n and := x 2 R 3 : m(x) =0 with the assumption that R 3 \ is connected with Lipschitz We again let (x, y) := 1 eik x y 4 x y,x6= y. A proof of the following theorem can be found in [10]. Theorem 1.8. Given two bounded domains and G, thevolume potential (V')(x) := (x, y)'(y) dy, x 2 R 3 defines a bounded operator V : L 2 ()! H 2 (G) where H 2 (G) denotes a Sobolev space. We now show that the scattering problem (1.14)-(1.16) is equivalent to solving the Lippmann Schwinger integral equation u(x) =u i (x) k 2 (x, y)m(y)u(y) dy, x 2 R 3. (1.17) R 3

10 CHAPTER 1. INVERSE SCATTERING THEORY 10 Theorem 1.9. If u 2 H 2 loc (R3 ) is a solution of (1.14)-(1.16) then u is a solution of (1.17). Conversely, if u 2 C(R 3 ) is a solution of (1.17) then u 2 H 2 loc (R3 ) and u is a solution of (1.14)-(1.16). Proof. Let u 2 Hloc 2 (R3 ) be a solution of (1.14)-(1.16). Let x 2 R 3 and B a ball containing x and. Then Green s formula u(x) = (x, y) u (x, y) ds(y) k 2 (x, y)m(y)u(y) dy and u i (x) i (x, y) u (y) (x, y) ds(y). Furthermore, Green s second identity and the Sommerfeld radiation condition implies s (x, y) u (x, y) ds(y) =0. Adding these equations together gives the Lippmann Schwinger integral equation (1.17), noting that since m has compact support the integral over B can be replace by an integral over R 3. Conversely, let u 2 C(R 3 ) be a solution of (1.17) and define u s (x) := k 2 (x, y)m(y)u(y) dy, x 2 R 3. R 3 Then u s satisfies the Sommerfeld radiation condition and u s 2 Hloc 2 (R3 ) satisfies u s + k 2 u s = k 2 mu. Since u i + k 2 u i = 0 we have that u = u i + u s satisfies u + k 2 nu =0inR 3. The existence of a unique solution to the scattering problem (1.14)- (1.16) is now equivalent to showing the existence of a unique solution to the Lippmann Schwinger integral equation. For the wave number k su ciently small, this can be done by the method of successive approximations. Theorem Suppose that m(x) = 0 for x a and k 2 < 2/M a 2 where M := max x applea m(x). Then there exists a unique solution to the Lippmann Schwinger integral equation.

11 CHAPTER 1. INVERSE SCATTERING THEORY 11 Proof. It su ces to solve (1.17) in C(B) withb := x 2 R 3 : x <a.on C(B) define (T m )(x) := (x, y)m(y)u(y) dy, x 2 B. B By the method of successive approximations, the theorem will be proved if kt m k 1 apple Ma 2 /2. To this end, we have Now note that (T m )(x) apple M kuk 1 4 h(x) := B B dy x y dy x y,x2 B,x2 B. satisfies h = 4 and is a function only of r = x. Hence 1 d r 2 r 2 dh = 4 dr dr and thus h(r) = 2 3 r2 + c 1 r + c 2 where c 1 and c 2 are constants. Since h is continuous at the origin, c 1 = 0 and letting r tend to zero shows that c 2 = h(0) = B dy a y =4 0 d =2 a 2. Hence h(r) =2 (a 2 r 2 /3) and thus khk 1 =2 a 2. We now have that and the theorem follows. (T m u)(x) apple Ma2 2 kuk 1,x2 B From (1.17) we see that and hence u s (x) = k 2 R 3 (x, y)m(y)u(y) dy, x 2 R u s (x) = eik x 1 u 1 (ˆx)+O x x 2 3, x!1

12 CHAPTER 1. INVERSE SCATTERING THEORY 12 where the far field pattern u 1 is given by u 1 (ˆx) = k2 e ikˆx y m(y)u(y) dy 4 R 3, ˆx = x x. Assuming that k is su ciently small and replacing u by the first term in solving (1.17) by iteration (the weak scattering assumption) gives the Born approximation to the far field pattern u 1 (ˆx) k2 e ikˆx y m(y)u i (y) dy. 4 R 3 The Born approximation has been used extensively in inverse scattering where the weak scattering assumption is valid and for details of such an approach see [15]. The proof of the existence of a unique solution to the Lippmann Schwinger integral equation for arbitrary k>0 is more delicate than for k>0su - ciently small and is based on the unique construction principle. This principle is a basic result in the theory of linear elliptic partial di erential equations and in the case of elliptic equations in R 3 dates back to Mülcer [23], [24]. Unique Continuation Principle. Let G be a domain in R 3 and suppose u 2 H 2 (G) is a solution of u+k 2 n(x)u =0in G where n is piecewise continuous in G. Then if u vanishes in a neighborhood of some point in G, u is identically zero in G. For a proof of the above unique continuation principle see [10]. We can now use this principle to show that for each k>0 there exists a unique solution u 2 H 2 loc (R3 ) to the scattering problem (1.14)-(1.16) (or equivalently the Lippmann Schwinger integral equation). Theorem For each k>0 there exists a unique solution u 2 H 2 loc (R3 ) to the scattering problem (1.14)-(1.16). Proof. The integral operator appearing in the Lippmann Schwinger integral equation has a weakly singular kernel and hence this operator is compact on C(B) whereb is a ball containing the support of m. Hence by the Riesz Fredholm theory it su ces to show the uniqueness of a solution to (1.17),

13 CHAPTER 1. INVERSE SCATTERING THEORY 13 i.e. that the only solution of lim r r!1 u + k 2 n(x)u =0 inr 3 iku = 0 is u = 0. To this end, Green s first identity and (1.18) imply that n ds = ru 2 k 2 n u 2o dx and B u u =0. By Corollary 1.7 u(x) = 0 for x 2 R 3 \ B and hence by the unique continuation principle u(x) = 0 for all x 2 R The Far Field Operator The far field operator plays a central role in inverse scattering theory and will appear in many of the remaining chapters of this monograph. Hence in this section we will introduce this operator and derive its most important analytic properties. In the course of our analysis we will also encounter the transmission eigenvalue problem which will be seen to play an important role in all of our subsequent investigations. In order to proceed we will need to be more specific on the nature of the incident field u i. In particular, from now on we will assume that u i (x) = e ikx d where d = 1. Then the solution of the scattering problem u + k 2 n(x)u = 0 (1.20) u(x) =e ikx d + u s (x) s lim r iku s = 0 (1.22) will depend on d and in particular the far field pattern u 1 (ˆx) =u 1 (ˆx, d) defined by u s (x) = eik x 1 u 1 (ˆx, d)+o x x 2 now depends on d. The following reciprocity principle is basic to our investigations.

14 CHAPTER 1. INVERSE SCATTERING THEORY 14 Theorem 1.12 (Reciprocity Principle). Let u 1 (ˆx, d) be the far field pattern corresponding to (1.20)-(1.22). Then u 1 (ˆx, d) =u 1 ( d, ˆx). Proof. Let {x: x < a} where again := {x: m(x) 6= 0}. Green s second identity implies that u i (y, ui (y, ˆx) u i (y, ui (y, d) ds(y) =0 y =a y =a u s (y, us (y, ˆx) u s (y, us (y, d) ds(y) =0 Then where u i (ˆx, d) =e ikx d. Corollary 1.5 shows that u s (y, ui (y, ˆx) u i (y, us (y, d) ds(y) =4 u 1 (ˆx, d) y =a y =a u s ( y, ui (y, d) u i (y, us (y, ˆx) ds(y) =4 u 1 ( d, ˆx). Subtracting the last of these equations from the sum of the first three gives 4 [u 1 (ˆx, d) u 1 ( d, ˆx)] = u(y, ˆx) u(y, ˆx) u(y, d) ds(y) =0 y =a by Green s second identity. We now define the far field operator F : L 2 (S 2 )! L 2 (S 2 )by (F g)(ˆx) := u 1 (ˆx, d)g(d) ds(d). S 2 Since u 1 (ˆx, d) is infinitely di erentiable with respect to each of its variables, F is clearly compact. The corresponding scattering operator S : L 2 (S 2 )! L 2 (S 2 )isdefinedby S := I + ik 4 F.

15 CHAPTER 1. INVERSE SCATTERING THEORY 15 We now want to prove some properties of these operators. To this end we define a Herglotz wave function to be a function of the form v(x) = e ikx d g(d) ds(y),x2 R 3 S 2 where g 2 L 2 (S 2 ). The function g is called the Herglotz kernel of v. Herglotz wave functions are clearly entire solutions of the Helmholtz equation. We note that for a given g 2 L 2 (S 2 ) the function v(x) = e ikx d g(d) ds(d),x2 R 3 S 2 is also a Herglotz wave function. Furthermore, if a Herglotz wave function vanishes in some open subset of R 3 then its kernel must be identically zero [10]. In what follows (, ) istheinnerproductinl 2 (S 2 ). Theorem Let g, h 2 L 2 (S 2 ) and let v i and w i be the Herglotz wave functions with kernels g and h respectively. Then if v and w are the solutions of the scattering problem (1.20)-(1.22) corresponding to the incident field e ikx d being replace by the incident fields v i and w i respectively we have that 0=2 (F g, h) 2 (g, F h) ik(f g, F h). Proof. ([9],[10]) Let v s = v v i and w s = w w i denote the scattered fields with far field patterns v 1 and w 1 respectively. Then by linearity v 1 = F g and w 1 = F h and Green s second identity implies that and x =a x =a ( v Furthermore, for R>awe have that v w ds = x =a x =R ds = 0 (1.23) ) w ds =0. v 2ik w ds S 2 v 1 w 1 ds = 2ik(F g, F h)

16 CHAPTER 1. INVERSE SCATTERING THEORY 16 as R tends to infinity. Finally, we have that x =a v = S 2 g(d) = 4 w x =a ds ikx e w eikx d S 2 g(d)w 1 (d) ds(d) = 4 (g, F h) ds(x)ds(d) and similarly x =a ( v ) w ds =4 (F g, h). Substituting the above identities into (1.23) now implies the theorem. Theorem The far field operator is normal, i.e. F F = FF,and the scattering operator S is unitary, i.e. S S = S S =I. Proof. Theorem 1.10 implies that ik(f g, F h) =2 [(F g, h) (g, F h)] (1.24) for g, h 2 L 2 (S 2 ). By reciprocity we have that (F g)(ˆx) = u 1 (d, ˆx)g(d) ds(d) S 2 = u 1 ( S 2 ˆx, d)g(d) ds(d) = S 2 i.e. F g = R F R g where (R h)(ˆx) :=h( ˆx). Since (R g, R h) =(g, h) =

17 CHAPTER 1. INVERSE SCATTERING THEORY 17 (g, h), we have from (1.24) that ik(f h, F g)=ik(r F R g, R F R h) = ik(f R g, F R h) =2 (F R g, R h) 2 (R g, F R h) =2 (R F R g, h) 2 (g, R F R h) =2 (h, F g) 2 (F h, g) =2 (F h, g) 2 (h, F g) = ik(f h, F g) and hence F F = FF. Finally, (1.24) implies that (g, ik F F h) =2 (g, (F F )h), i.e. ik F F =2 (F F ). This, together with F F = FF, implies that S S = SS = I by direct substitution. We now introduce the transmission eigenvalue problem: eterminek>0 and v, w 2 L 2 (), v w 2 H0 2 (), such that v 6= 0, w 6= 0 and w + k 2 n(x)w =0 in v + k 2 v =0 in v = Such values of k are called transmission eigenvalues. Recall that := {x: n(x) 6= 0} and it is assumed that is bounded with Lipschitz such that R 3 \ is connected. Theorem Let F be the far field operator corresponding to the scattering problem (1.20)-(1.22). Then F is injective if k is not a transmission eigenvalue. Proof. ([12],[17]) Suppose F g = 0. Then the far field pattern w 1 of the scattering field w s corresponding to the incident field w i (x) := e ikx d g(d) ds(d) S 2

18 CHAPTER 1. INVERSE SCATTERING THEORY 18 vanishes. By Rellich s lemma w s = w w i vanishes outside. Then w = w i + w s satisfies w + k 2 nw =0inR 3 and w w i = 0 (w wi ) = 0 Ifk is not a transmission eigenvalue then w i = w =0 and hence g = 0, i.e. F is injective. Corollary Let F be the far field operator corresponding to the scattering problem (1.20)-(1.22). Then F has dense range if k is not a transmission eigenvalue. Proof. ([12],[17]) From a well known theorem in functional analysis, the orthogonal complement of the range of F is equal to the null space of its adjoint F. Hence we must show that if F h =0thenh = 0. To this end, we have that if F h =0i.e. S 2 u 1 (d, ˆx)h(d) ds(d) =0 then and hence, using reciprocity, S 2 u 1 ( ˆx, d)h(d) ds(d) =0 S 2 u 1 (ˆx, d)h( d) ds(d) =0. Since F is injective by Theorem 1.15, we can now conclude that h = 0 as desired. 1.4 Inverse Scattering We again consider the scattering problem (1.20)-(1.22). It has previously been shown that u s (x, d) = eik x 1 u 1 (ˆx, d)+o x x 2 as x!1where u 1 (ˆx, d) = k2 4 R 3 e ikˆx y m(y)u(y) dy

19 CHAPTER 1. INVERSE SCATTERING THEORY 19 and m := 1 n. The inverse scattering problem is to determine n(x) (or some properties of n(x)) from u 1 (ˆx, d). We begin with uniqueness. As motivation we first prove a simple result for harmonic functions. Theorem The set of products h 1 h 2 of entire harmonic functions h 1 and h 2 is complete in L 2 () for any bounded domain R 3. Proof. [7] Given y 2 R 3 choose a vector b 2 R 3 with b y = 0 and b = y. Then for z := y = ib 2 C 3 we have z z = 0 which implies that h z := e iz x, x 2 R 3, is harmonic. Now assume ' 2 L 2 () is such that 'h 1 h 2 dx =0 for all pairs of entire harmonic function h 1 and h 2. Our theorem will be proved if we can show that ' = 0. But for h 1 = h z, h 2 = h z we have that '(x)e 2iy x dx =0 for y 2 R 3 which implies that ' = 0 almost everywhere by the Fourier integral theorem. To prove uniqueness for the inverse scattering problem of determining n(x) from u 1 (ˆx, d) we need a property corresponding to the above theorem for products v 1 v 2 of solutions to v 1 + k 2 n 1 v 1 = 0 and v 2 + k 2 n 2 v 2 = 0 for two di erent refractive indices n 1 and n 2. Such a result was first established by Sylvester and Uhlmann [30]. The proofs of the following two theorems can be found in [10] and [19]. Theorem Let B be and open ball centered at the origin and containing the support of m := 1 n. Then there exists a constant C>0such that for each z 2 C 3 with z z =0and Re z > 2k 2 knk 1 there exists a solution v 2 H 2 () of v + k 2 nv =0in B of the form v(x) =e iz x [1 w(x)] where kwk L 2 () apple C Re z.

20 CHAPTER 1. INVERSE SCATTERING THEORY 20 Theorem Let B and B 0 be two open balls centered at the origin and containing the support of m := 1 n such that B B 0. Then the set of total fuelds u(,d): d 2 S 2 satisfying (1.20)-(1.22) is complete in the closure of with respect to the L 2 (B) norm. v 2 H 2 (): v + k 2 nv =0in B 0 We are now ready to prove the following uniqueness result for the inverse scattering probeem due to Nachman [25], Novikov [26] and Ramm [27]. Theorem The index of refraction n is uniquely determined by a knowledge of the far field pattern u 1 (ˆx, d) for x, d 2 S 2 and a fixed wave number k. Proof. Assume that n 1 and n 2 are two refractive indices such that u 1,1 (,d)=u 2,1 (,d),d2 S 2 and let B and B 0 be two open balls centered at the origin and containing the support of 1 n 1 and 1 n 2 such that B B 0.ByRellich slemmawe have that u 1 (,d)=u 2 (,d)inr 3 \ B for all d 2 S 2. Hence u := u 1 u 2 satisfies u = 0 (1.25) and u + k 2 n 1 u = k 2 (n 2 n 2 )u 2 in B. From this and the partial di erential equation for ũ 1 := u 1 (, d) wehave that k 2 ũ 1 u 2 (n 2 n 1 )=ũ 1 ( u + k 2 n 1 u)=ũ 1 u u ũ 1. Green s second identity and (1.25) now imply that u 1 (, d)u 2 (,d)(n 1 n 1 ) dx =0 B for all d, d 2 S 2. Hence, from Theorem 1.15, it follows that v 1 v 2 (n 1 n 2 ) dx = 0 (1.26) B for all solutions v 1,v 2 2 H 2 () of v 1 + k 2 n 1 v 1 = 0, v 2 + k 2 n 2 v 2 =0in B 0.

21 CHAPTER 1. INVERSE SCATTERING THEORY 21 Given y 2 R 3 \{0} and >0 we now choose vectors a, b 2 R 3 such that {y, a, b} is an orthogonal basis in R 3 and a = 1, b 2 = y Then for z 1 := y + a + ib, z 2 := y a ib we have that and z j z j = Re z j 2 = y b 2 =0 Im z j 2 +2i Re z j Im z j Re z j 2 = y In (1.26) we now insert the solutions v 1 and v 2 constructed in Theorem 1.14 for the indices of refraction n 1 and n 2 and the vectors z 1 and z 2 respectively. Since z 1 + z 2 =2y this yields e 2iy x [1 + w 1 (x)] [1 + w 2 (x)] [n 1 (x) n 2 (x)] dx =0 B and passing to the limit as tends to infinity gives e 2iy x [n 1 (x) n 2 (x)] dx =0. B By the Fourier integral theorem we now have that n 1 = n 2. We will now show how, in principle, n(x) can be constructed from u 1 (ˆx, d) through the use of nonlinear optimization methods. To this end, we define the operator F : m 7! u 1 for u 1 = u 1 (ˆx, d). Letting B be a ball containing the (unknown) support of m, weinterpretf as an operator from L 2 (B) intol 2 (S 2 S 2 ). From The Lippmann Schwinger integral equation we can write (F m)(ˆx, d) = k2 e ikˆx y m(y)u(y) dy (1.27) 4 B where u(,d) is the unique solution of u(x, d)+k 2 (x, y)m(y)u(y, d) dy = e ikx d (1.28) B where again (x, y) := 1 eik x y 4 x y,x6= y.

22 CHAPTER 1. INVERSE SCATTERING THEORY 22 Note that F is a nonlinear operator. Recall now that a mapping T : X! Y of a normal space X into a normal space Y is called Fréchet di erentiable if there exists a bounded linear operator A: X! Y such that lim h!0 1 kt (x + h) T (x) A hk =0 khk and we write T 0 (x) =A. In particular, from (1.28) it can be seen that the Fréchet derivative v := u 0 mh of u with respect to m (in the direction h) satisfies the Lippmann Schwinger integral equation v(x, d)+k 2 (x, y)[m(y)v(y, d)+h(y)u(y, d)] dy =0,x2 B (1.29) B and from (1.27) we have that (F 0 m h)(ˆx) = k2 e ikˆx y [m(y)v(y, d)+h(y)u(y, d)] dy 4 B for ˆx, d 2 S 2. Hence (F 0 m h)(ˆx) coincides with the far field pattern of the solution v(,d) 2 Hloc 2 (R3 ) of (1.29). Note also that F 0 m : L 2 (B)! L 2 (S 2 S 2 ) is compact. We have proven the following theorem [10]: Theorem The operator F : m 7! u 1 is Fréchet di erentiable. The derivative is given by F 0 m h = v 1 where v 1 is the far field pattern of the radiating solution v 2 H 2 loc (R3 ) to v + k 2 nv = k 2 uh in R 3. Theorem The operator F 0 m : L 2 (B)! L 2 (S 2 S 2 ) is injective. Proof. [10] Assume that h 2 L 2 (B) satisfies F 0 m h = 0. We want to show that h = 0. Since F 0 m h = 0 we have that for each d 2 S 2 the far field pattern of the solution v of (1.29) vanishes and Rellich s lemma implies that v(,d) = 0 Hence Green s second identity implies that k 2 hu(,d)wdx=0 for all d 2 S 2 and any solution w 2 H 2 (B) of Theorem 1.15 we can now conclude that hn wdx=0 B B w + k 2 nw =0inB 0. By for all w, w satisfying w + k 2 nw = 0 and w + k 2 n w =0inB 0 B.The proof can now be completed as in the proof of Theorem 1.20.

23 CHAPTER 1. INVERSE SCATTERING THEORY 23 We can now apply Newton s method to the nonlinear equation F(m) = u 1. However to implement this procedure we must solve a direct scattering problem at each step of the iteration procedure. We furthermore have the possible problem of local minima and need to solve an ill-posed compact operator equation of the first kind at each step. How to solve this last problem will be dealt with in the next section. 1.5 Ill-Posed Problems In the previous sections we have introduced two di erent methods for solving the inverse scattering problem: the Born approximation and Newton s method applied to the nonlinear equation F(m) =u 1. Both methods involve the solution of an integral equation of the first kind over a bounded region with a smooth kernel. In particular, in both cases the integral operator is compact. As we shall see shortly, the problem of inversion of such an operator is ill-posed in the sense that the solution does not depend continuously on the given (measured) data. The same problem will also arise later when we use the factorization method or the linear sampling method to determine the support of the scattering object. In short, all the available methods for solving the inverse scattering problem involve the solution of ill-posed integral equations of the first kind. Hence in this section we shall give a brief survey of how to solve such equations. For a more comprehensive study we refer the reader to [16], [18] and [21]. efinition Let A: X! V Y be an operator from a normal space X into a subset V of a normal space Y. The equation A ' = f is called well-posed if A: X! V is bijective and the inverse operator A 1 : V! X is continuous. Otherwise the equation is called ill-posed. Theorem Let A: X! V Y be a linear compact operator. Then A ' = f is ill-posed if X is not finite dimensional. Proof. If A 1 : V! X exists and is continuous then I = A 1 A is compact which implies that X is finite dimensional. We now assume that A is a linear compact operator and wish to approximate the solution ' to A ' = f from a knowledge of a perturbed right hand side f with a known error level f f apple. We will always assume that A: X! Y is injective and want the approximate solution ' to depend continuously on f.

24 CHAPTER 1. INVERSE SCATTERING THEORY 24 efinition Let A: X! Y be an injective compact linear operator. Then a family of bounded linear operators R : Y! X with the property that R f! A 1 f,! 0 (1.30) for all f 2 A(X) is called a regularization scheme for A. The parameter is called the regularization parameter. It is easily verified that if X is infinite dimensional then the operator R cannot be uniformly bounded with respect to and the operators R A cannot be norm convergent as! 0 [10]. A regularization scheme approximates the solution ' of A ' = f by the regularized solution ' := R f. Hence ' ' =R f R f +R A ' ' which implies that ' ' apple kr k + kr A ' 'k. The error consists of two parts. The first term reflects the error in the data and the second term the error between R and A 1. From the above discussion we see that the first term will be increasing as! 0duetothe ill-posed nature of the problem whereas the second term will be decreasing as '! 0 according to (1.30). efinition A strategy for a regularization scheme R, >0, i.e. the choice of the regularization parameter = (, f ), is called regular if for all f 2 A(X) and f 2 Y with f f apple we have that R (,f ) f! A 1 f, > 0. A natural strategy is the Marozov discrepancy principle which is based on the idea that the residual should not be smaller than the accuracy of the measurements, i.e. AR f f apple for some parameter apple 1. From now on let X and Y be Hilbert spaces and A: X! Y be a compact linear operator with adjoint A : Y! X. The nonnegative square roots of the eigenvalues of A A: X! X are called the singular values of A. We always assume that A 6= 0. For a proof of the following theorem see [5] or [10]. Theorem Let (µ n ), µ 1 µ 2 be the singular values of A. Then there exists orthonormal sequences (' n ) in X and (g n ) in Y such that A ' n = µ n g n, A g n = µ n ' n

25 CHAPTER 1. INVERSE SCATTERING THEORY 25 and for all ' 2 X ' = A ' = 1X (', ' n )' n +Q' n=1 1X µ n (', ' n )g n n=1 where Q: X! N(A) is the orthogonal projection operator. (µ n,' n,g n ) is called a singular system of A. The system Theorem 1.28 (Picard s Theorem). Let A: X! Y be a compact linear operator with singular system (µ n,' n,g n ). Then A ' = f is solvable if and only if f 2 N(A )? and satisfies 1X 1 µ 2 n=1 n In this case a solution is given by ' = 1X n=1 (f,g n ) 2 < 1. (1.31) 1 µ n (f,g n )' n. (1.32) Proof. The necessity of f 2 N(A )? follows from N(A )? = A(X). If A ' = f then µ n (', ' n )=(', A g n )=(A', g n )=(f,g n ) and hence 1X 1 µ 2 n=1 n (f,g n ) 2 = 1X (', ' n ) 2 applek'k 2 n=1 and the necessity of (1.31) follows. Conversely, if f 2 N(A )? and (1.31) is satisfied then (1.32) converges in X. Applying A to (1.32) gives since f 2 N(A )?. A ' = 1X (f,g n )g n = f n=1 Picard s theorem shows that the ill-posedness of A ' = f comes from the fact that µ n! 0. This suggests filtering out the influence of 1/µ n in the solution of (1.32). To this end we have the following theorem.

26 CHAPTER 1. INVERSE SCATTERING THEORY 26 Theorem Let A: X! Y be an injective compact linear operator with singular system (µ n,' n,g n ) and let q :(0, 1) (0, kak)! R be a bounded function such that for each '>0 there exists a positive constant c( ) with and q(, µ) applec( )µ,0 <µapplekak (1.33) lim q(, µ) =1, 0 <µapplekak. (1.34)!0 Then the bounded operators R : Y! X, >0, defined by R f : 1X n=1 1 µ n q(, µ n )(f,g n )' n,f 2 Y describes a regularization scheme with kr kapplec( ). Proof. Since for all f 2 Y we have that we have from (1.33) that kfk 2 = kr fk 2 = 1X (f,g n ) 2 + kq fk 2 n=1 1X 1 µ 2 n=1 n X 1 apple c( ) 2 q(, µ n ) 2 (f,g n ) 2 n=1 apple c( ) 2 kfk 2 (f,g n ) 2 form all f 2 Y and hence kr kapplec( ). With the aid of (R A ', ' n )= 1 µ n q(, µ n )(A ', g n ) = q(, µ n )(', ' n ) and the singular value decomposition for R A ' kr A ' 'k 2 = = 1X (R A ' ', ' n ) 2 n=1 ' we obtain (1.35) 1X [q(, µ n ) 1] 2 (', ' n ) 2 n=1

27 CHAPTER 1. INVERSE SCATTERING THEORY 27 where we have used the fact that A is injective. Now let ' 2 X with ' 6= 0 and let >0 be given. Let q(, µ) applem. Then there exists N = N( ) such that 1X n=n+1 (', ' n ) 2 < 2(M + 1) 2 By (1.34) there exists 0 = 0 ( ) > 0 such that [q(, µ n ) 1] 2 < 2 k'k 2 for all n =1, 2,,N and 0 < < 0. Splitting the series (1.35) into two parts now yields kr A ' 'k 2 < 2 k'k 2 NX (', ' n ) apple for 0 < apple 0. Hence R A '! ' as! 0 for all ' 2 X and the proof is complete. The special choice n=1 q(, µ) = µ2 + µ 2 leads to Tikhonov regularization with is arguably the most popular method for solving ill-posed operator equations of the first kind. Theorem Let A: X! Y be a compact linear operator. Then for each >0 the operator I+A A: X! X is bijective and has a bounded inverse. Furthermore, if A is injective then R := ( I+A A) 1 A describes a regularization scheme with kr kapple 1 2 p. Proof. From k'k 2 apple ( ' +A A ', ') for all ' 2 X we conclude that for >0 the operator I+A A is injective. Let (µ n,' n,g n ) be a singular system for A and Q: X! N(A) denote the orthogonal projection operator. Then T: X! X defined by T ' := 1X 1 + µ 2 (', ' n )' n + 1 n Q(') n=1 is bounded and ( I+A A) T = T( I+A A) = I, i.e. T = ( I+A A) 1.

28 CHAPTER 1. INVERSE SCATTERING THEORY 28 If A is injective then for the unique solution ' of ' +A A ' =A f we deduce from the above expression for ( I+A A) 1 and the identity (A f,' n )=µ n (f,g n ) that Hence R f = ' = 1X n=1 1X µ n + µ 2 (f,g n )' n. n n=1 1 µ n q(, µ n )(f,g n )' n,f 2 Y with q(, µ) = µ2 +µ 2. The function q satisfies the conditions of Theorem 1.29 with c( ) =1/2 p due to the fact that p + µ 2 µ apple. 2 The proof of the theorem is now complete. It can be shown that the Morozov discrepancy principle is a regular strategy for choosing [10], [21]. Regularization methods can also be developed for the case when the operator A is perturbed with a known error level [10]. 1.6 The Factorization Method We now present a method for determining the support of m := 1 n which is not based on either the weak scattering approximation or the use of nonlinear optimization techniques. Such a method is an example of the qualitative approach to inverse catering theory [5]. The method for determining discussed in this lecture is called the factorization method [18], [20] and is based on the following theorem from functional analysis [5], [10], [20]: Theorem Let X and H be Hilbert spaces and assume that F : H! H, B: X! H and T: X! X be bounded linear operators that satisfy where B is the adjoint of B, F =BTB Im(T f,f) 6= 0

29 CHAPTER 1. INVERSE SCATTERING THEORY 29 for all f 2 B (H) with f 6= 0and T=T 0 +C where C is compact such that (T 0 f,f) 2 R (T 0 f,f) c kfk 2 for all f 2 B (H) and some c > 0. In addition, let the operator F be compact, injective and assume that I+i Fisunitaryforsome >0. Then the ranges B(X) and (F F ) 1/4 (H) coincide. We now want to determine the support of m := 1 n from a knowledge of the far field pattern u 1 (ˆx, d) of the scattered field corresponding to u + k 2 n(x)u =0 inr 3 (1.36) u(x, d) =e ikx d + u s (x, d) s lim r iku s =0. (1.38) We assume that n(x) > 1 for x 2, is piecewise continuous in and rewrite (1.36) as u s + k 2 nu s = k 2 mu i in R 3 where u i (x, d) =e ikx d. More generally, we consider u s + k 2 nu s = mf (1.39) in R 3 where f 2 L 2 () and u s satisfies the Sommerfeld radiation condition (1.38). The existence of a unique solution to (1.39) in Hloc 2 (R3 ) follows from the existence of a unique solution to the Lippmann Schwinger integral equation in L 2 () and Theorem 1.8. We now definite the operator G: L 2 ()! L 2 (S 2 ) which maps f 2 L 2 () onto the far field pattern of the solution to (1.39). We again let u 1 2 L 2 (S 2 ) be the far field pattern of the scattering problem (1.36)-(1.38) and define the far field operator F : L 2 (S 2 )! L 2 (S 2 )by (F g)(ˆx) := u 1 (ˆx, d)g(d) ds(d), ˆx 2 S 2. S 2 Theorem Let F and G be defined as above. Then F =4 k 2 G S G

30 CHAPTER 1. INVERSE SCATTERING THEORY 30 where S is the adjoint of S : L 2 ()! L 2 () defined by (x) (S )(x) := k 2 (x, y) (y) dy, x 2 m(x) and G : L 2 (S 2 )! L 2 () is the adjoint of G. Proof. ([10], [19]) We have that u 1 = k 2 G u i. We now define the Herglotz operator H: L 2 (S 2 )! L 2 () by (H g)(x) := e ikx d g(d) ds(d),x2 S 2 and note that F g is the far field pattern corresponding to the incident field H g, i.e. F = k 2 GH. Since (H )(ˆx) = e ikˆx y (y) dy, ˆx 2 S 2, we have that H =4 w 1 where w 1 is the far field pattern of w(x) := (x, y) (y) dy, x 2 R 3. But and hence w + k 2 nw = m m + k2 w H =4 w 1 = 4 G m + k2 w =4 G S, i.e. H =4 G S. Thus H = 4 S G and since F = k 2 G H the theorem follows. Lemma For z 2 R 3 let 1(ˆx, z) = 1 4 e ikˆx z be the far field pattern of (x, z). Then z 2 if and only if 1(,z) is in the range of G.

31 CHAPTER 1. INVERSE SCATTERING THEORY 31 Proof. For z 2 choose a cut-o function 2 C 1 (R 3 ) which vanishes near z and equals one in R 3 \. Thenv(x) = (x) (x, z) has 1(ˆx, z) as its far field pattern. Hence 1(,z)=Gf where f := 1 m ( v + k2 nv) 2 L 2 (). Now assume that z /2 and 1(,z) = Gf for some f 2 L 2 (). By Rellich s lemma (,z)=u s in the exterior of [{z} and this is a contradiction since u s is smooth near z but (,z) is singular at z Theorem Let S : L 2 ()! L 2 () be defined as in Theorem 1.32 and let S 0 : L 2 ()! L 2 () be given by Then S 0 := m. 1. S 0 is bounded, self-adjoint and satisfies (S 0, ) 1 kmk 1 k k 2, 2 L 2 (). 2. S S 0 : L 2 ()! L 2 () is compact. 3. S is an isomorphism from L 2 () onto L 2 (). 4. Im(S, ) apple 0 for all 2 L 2 () with strict inequality holding for all 2 G (L 2 ()) with 6= 0. Proof. ([10], [19]) 1. This follows since m(x) > 0 for x This follows from Theorem 1.8 and the fact that H 2 () is compactly embedded in L 2 (). 3. From its definition, S 0 clearly has a bounded inverse. Hence by 2. and the Riesz Fredholm theorem it su ces to show that S is injective. Suppose S = 0. Then ' = /m satisfies the homogeneous Lippmann Schwinger integral equation and hence ' = 0 and thus =0 4. Let 2 L 2 () and define f 2 L 2 () by f(x) := (x)+k 2 m(x) (x, y) (y) dy x 2.

32 CHAPTER 1. INVERSE SCATTERING THEORY 32 Then S = f/m and setting w(x) := (x, y) (y) dy, x 2 R 3 we see that (S, ) = = f m f k2 mw dx 1 m f 2 dx + k 2 fwdx. Since w + k 2 nw = k 2 mw = f in, bygreen sfirstidentity we have that fwdx= w w + k 2 nw dx and hence = = n grad w Im(S, ) = ds k 2 n w 2o dx ds. (1.40) From the identity (1.9) we now have that Im(S, ) apple 0sincew satisfies w +k 2 w =0inR 3 \ and the Sommerfeld radiation condition. We now prove the last part of 4. Let 2 L 2 () being (L 2 ()) such that Im(S, ) = 0. We want to show that = 0. From (1.40) and Corollary 1.7 we have that w =0inR 3 \. Since from the proof of Theorem 1.32 we have that H =4 S G and S (and hence S )is an isomorphism from L 2 () onto L 2 () we have that the range of G coincides with the range of H. Hence there exists a sequence j =Hg j with j! 2 G (L 2 ()). efine v j by v j (x) := e ikx d g(d) ds(d),x2 R 3. S 2

33 CHAPTER 1. INVERSE SCATTERING THEORY 33 Then v j = j and v j + k 2 v j =0inR 3. From the definition of w we have that w + k 2 w = in and by Green s second identity we have that j dx = v j dx = w + k 2 w v j dx = =0 v j + k 2 v j v j since w =0inR 3 \. Hence, letting j!1, we have that 2 dx =0 and hence =0inL 2 (). ds We can now derive the factorization method for determining the support of from v 1. To this end, assume that k is not a transmission eigenvalue. Then the far field operator F is injective and normal. In particular, I +i k 2 F is unitary. Hence we can now apply Theorem 1.31 to obtain the following [10], [19]: Theorem Assume that n(x) > 1 for x 2 and k > 0 is not a transmission eigenvalue. Then z 2 is and only if 1(,z) is in the range of (F F ) 1/4. We note that this theorem is also true if we assume that 0 <n(x) < 1 for x 2 with the proof being exactly the same. A method to determine the support of m =1 n using Theorem 1.35 is to use Tikhonov regularization to find a regularized solution of (F F ) 1/4 = 1 (,z) (1.41) and note that the regularized solution gz of (1.41) converges in L 2 (S 2 ) as! 0 if and only if z 2. An alternative method to construct is to let n and n be the eigenvalues and eigenfunctions of F and note that

34 CHAPTER 1. INVERSE SCATTERING THEORY 34 (F F ) 1/4 has the singular system ( p n, n, n ). Then by Picard s theorem (Theorem 1.28) and Theorem 1.35, z 2 if and only if 1X ( n, 1(,z)) 2 < 1. n n=1 For details of the numerical implementation of the factorization method we refer the reader to [20]. 1.7 The Linear Sampling Method The linear sampling method [8], [14] is another method for determining the support of m =1 n from the far field pattern u 1 that is closely related to the factorization method. To present this method we begin by showing the existence of a unique solution u, w 2 L 2 (), v w 2 H 2 () of the inhomogeneous interior transmission problem w + k 2 n(x)w =0 in (1.42) v + k 2 v =0 in (1.43) w v = (,z) (1.44) (,z) (1.45) where for the sake of simplicity we assume that is connected with a connected C 2 and z 2. efinition Let H be the linear space of all Herglotz wave functions and H the closure of H in L 2 (). For ' 2 L 2 () define the volume potential by (T m ')(x) := )(x, y)m(y)'(y) dy, x 2 R 3. Then a pair v, w with v 2 H and w 2 L 2 () is said to be a solution of the inhomogeneous interior transmission problem (1.42)-(1.45) with point source z 2 if v and w satisfy the integral equation w + k 2 T m w = v in and the boundary condition k 2 T m w = (,z) where B is an open ball centered at the origin with B.

35 CHAPTER 1. INVERSE SCATTERING THEORY 35 Note that if v, w is a solution of the inhomogeneous interior transmission problem with point source z 2 then (k 2 T m w)(x) = (x, z) for all x 2 R 3 \. We assume that m(x) =1 n(x) > 0 for x 2 and define an equivalent norm on L 2 () viatheinnerproduct (', ) m := m(y)'(y) (y) dy. Note here we could also assume that m(x) < 0 for x 2 and define (', ) m as above with a minus sign before the integral sign. The following theorem follows from the definition of transmission eigenvalues: Theorem Assume that k is not a transmission eigenvalue. Then for every source point z 2 there exists at most one solution of the inhomogeneous interior transmission problem. Theorem Assume that k is not a transmission eigenvalue. Then for every source point z 2 there exists a solution to the inhomogeneous interior transmission problem. Proof. [10], [14] Assume without loss of generality that z = 0. efine H 0 1 := span j p (kr)y q p (ˆx): p =1, 2,, p apple q apple p and let H 1 be the closure of H1 0 in L2 (). Let H 1? denote the orthogonal complement of H 1 with respect to the inner product (, ) m. Then there exists a nontrivial 2 H 1? \ H and (j 0, ) m 6= 0. Let P: L 2 ()! H? be the orthogonal projection operator with respect to (, ) m. Now consider the equation u + k 2 PT m u = k 2 PT m (1.46) in L 2 (). Since T m is compact and P is bounded, P T m is compact in L 2 (). We want to apply the Riesz Fredholm theory. To this end, assume that w 2 L 2 () satisfies w + k 2 PT m w =0. Then w 2 H? and satisfies v := k 2 (I P) T m w 2 H w + k 2 T m w = v.

36 CHAPTER 1. INVERSE SCATTERING THEORY 36 Since (w, ') m = 0 for all ' 2 H, the addition formula for Bessel functions (x, y) =ik 1X nx n=0 m= n h (1) n (k x )Y m n (ˆx)j n (k x )Y m n (ŷ), x > y, (1.47) implies that T m w = 0 Since k is not a transmission eigenvalue we have that v = w = 0 and by the Riesz Fredholm theory we have that I+k 2 PT m is invertible. Now let u be the solution of (1.46) and note that u 2 H?. efine the constant c and function w 2 L 2 () by c := 1 k 2 (, j 0 ) m,w := c(u ). Then and hence where Since for all h 2 H 1 and w + k 2 PT m w = c w + k 2 T m w = v v = k 2 (I P) T m w c 2 H. (w, h) m = c(u, h) m =0 (w, j 0 ) m = c(u, j 0 ) m = 1 k 2 we have from the addition formula (1.47) that k 2 (T m w)(x) = ik and the proof is complete. 4 h(1) 0 (k x ) = (x, 0), x We can now derive the basic result of the linear sampling method [10], [14]: Theorem Assume that n(x) > 1 for x 2 or 0 <n(x) < 1 for x 2. Let F be the far field operator and assume that k is not a transmission eigenvalue. Then

37 CHAPTER 1. INVERSE SCATTERING THEORY For z 2 and given >0 there existts a function gz 2 L 2 (S 2 ) such that kf gz 1(,z)k L 2 (S 2 ) < (1.48) and the Herglotz wave function with kernel g z converges to the solution v 2 H of the inhomogeneous interior transmission problem as! 0 2. For z/2 every g z 2 L 2 (S 2 ) that satisfies (1.48) for a given >0 is such that lim!0 v g z L 2 () = 1 where v g z is the Herglotz wave function with kernel g z. Proof. Let v(,z), w(,z) be the unique solution to the inhomogeneous interior transmission problem with source point z 2. By the definition of H we have that for every >0 and z 2 there exists g 2 L 2 (S 2 ) such that kv(,z) v g k L 2 () apple where v g is the Herglotz wave function with kernel g. Hence for u g := (I +k 2 T m ) 1 v g we have that kw(,z) u g k L 2 () apple c for some constant c>0. Since T m : L 2 ()! C(@B) is bounded we have that k 2 T m u g + (,z) C(@B) apple c 0 for some constant c 0 > 0 and since for the far field pattern k 2 T 1 m u g of k 2 T m u g we have that k 2 (T 1 m u g )(ˆx) = u 1 (ˆx, d)g(d) ds(d). S 2 We have by the well-posedness of the exterior irichlet problem [10] that k 2 T 1 m u g + (,z) L apple c00 2 (S 2 ) for some constant c 00 > 0. Noting that F g = k 2 T 1 m u g and letting! 0 establishes 1. Now let z /2 and assume that there exist Herglotz wave functions v j with kernels g j = g j z such that kv j k L 2 () remains bounded as j! 0. Without loss of generality we can assume that v j *v2 L 2 () as j!1.

38 CHAPTER 1. INVERSE SCATTERING THEORY 38 Let v s be the scattered field corresponding to the incident field v and let v 1 be its far field pattern. Since F g j is the far field pattern corresponding to the incident field v j then if u j + k 2 T m u j = v j we have that F g j = k 2 T 1 m u j. Letting j!1and using (1.48) shows that v 1 = k 2 T 1 m u = 1 (,z) where u =(I+k 2 T m ) 1 v. But k 2 T m u 2 H 1 loc (R3 \ ) and (,z) is not and by Rellich s lemma this is a contradiction. A problem with the linear sampling method is that in general there does not exist a solution to F g = (,z) for noise free data and hence it is not clear what solution is obtained by using Tikhonov regularization to solve this equation. This problem can be resolve by using the theory of the factorization method [2], [3]. Theorem Let F be the far field operator and assume that k is not a transmission eigenvalue. For z 2 denote by g z the solution of (F F ) 1/4 g z = 1(,z) and for >0 and z 2 R 3 let gz denote the solution of the far field equation F g = 1 (,z) obtained by Tikhonov regularization, i.e. the solution of gz + F F gz = F 1(,z), and let v g z denote the Herglotz wave function with kernel gz.ifz2then lim!0 v g z (z) exists and c kg z k 2 apple lim!0 v g z (z) applekg z k 2 for some positive constant c depending only on. Ifz/2 then lim v gz!0 (z) = An Inverse Spectral Problem The transmission eigenvalue problem for an isotropic spherically stratified medium in R 3 is to find a nontrivial solution v, w 2 L 2 (B a ), v w 2 H 2 0 (B a) to w + k 2 n(r)w =0 inb v + k 2 v =0 inb a (1.50) v w = 0 = 0 a (1.52)

Inverse Obstacle Scattering

Inverse Obstacle Scattering , Göttingen AIP 2011, Pre-Conference Workshop Texas A&M University, May 2011 Scattering theory Scattering theory is concerned with the effects that obstacles and inhomogenities have on the propagation

More information

A coupled BEM and FEM for the interior transmission problem

A coupled BEM and FEM for the interior transmission problem A coupled BEM and FEM for the interior transmission problem George C. Hsiao, Liwei Xu, Fengshan Liu, Jiguang Sun Abstract The interior transmission problem (ITP) is a boundary value problem arising in

More information

ON THE MATHEMATICAL BASIS OF THE LINEAR SAMPLING METHOD

ON THE MATHEMATICAL BASIS OF THE LINEAR SAMPLING METHOD Georgian Mathematical Journal Volume 10 (2003), Number 3, 411 425 ON THE MATHEMATICAL BASIS OF THE LINEAR SAMPLING METHOD FIORALBA CAKONI AND DAVID COLTON Dedicated to the memory of Professor Victor Kupradze

More information

The Helmholtz Equation

The Helmholtz Equation The Helmholtz Equation Seminar BEM on Wave Scattering Rene Rühr ETH Zürich October 28, 2010 Outline Steklov-Poincare Operator Helmholtz Equation: From the Wave equation to Radiation condition Uniqueness

More information

LECTURE 5 APPLICATIONS OF BDIE METHOD: ACOUSTIC SCATTERING BY INHOMOGENEOUS ANISOTROPIC OBSTACLES DAVID NATROSHVILI

LECTURE 5 APPLICATIONS OF BDIE METHOD: ACOUSTIC SCATTERING BY INHOMOGENEOUS ANISOTROPIC OBSTACLES DAVID NATROSHVILI LECTURE 5 APPLICATIONS OF BDIE METHOD: ACOUSTIC SCATTERING BY INHOMOGENEOUS ANISOTROPIC OBSTACLES DAVID NATROSHVILI Georgian Technical University Tbilisi, GEORGIA 0-0 1. Formulation of the corresponding

More information

An eigenvalue method using multiple frequency data for inverse scattering problems

An eigenvalue method using multiple frequency data for inverse scattering problems An eigenvalue method using multiple frequency data for inverse scattering problems Jiguang Sun Abstract Dirichlet and transmission eigenvalues have important applications in qualitative methods in inverse

More information

Finite-dimensional spaces. C n is the space of n-tuples x = (x 1,..., x n ) of complex numbers. It is a Hilbert space with the inner product

Finite-dimensional spaces. C n is the space of n-tuples x = (x 1,..., x n ) of complex numbers. It is a Hilbert space with the inner product Chapter 4 Hilbert Spaces 4.1 Inner Product Spaces Inner Product Space. A complex vector space E is called an inner product space (or a pre-hilbert space, or a unitary space) if there is a mapping (, )

More information

The Imaging of Anisotropic Media in Inverse Electromagnetic Scattering

The Imaging of Anisotropic Media in Inverse Electromagnetic Scattering The Imaging of Anisotropic Media in Inverse Electromagnetic Scattering Fioralba Cakoni Department of Mathematical Sciences University of Delaware Newark, DE 19716, USA email: cakoni@math.udel.edu Research

More information

Inverse Scattering Theory: Transmission Eigenvalues and Non-destructive Testing

Inverse Scattering Theory: Transmission Eigenvalues and Non-destructive Testing Inverse Scattering Theory: Transmission Eigenvalues and Non-destructive Testing Isaac Harris Texas A & M University College Station, Texas 77843-3368 iharris@math.tamu.edu Joint work with: F. Cakoni, H.

More information

STEKLOFF EIGENVALUES AND INVERSE SCATTERING THEORY

STEKLOFF EIGENVALUES AND INVERSE SCATTERING THEORY STEKLOFF EIGENVALUES AND INVERSE SCATTERING THEORY David Colton, Shixu Meng, Peter Monk University of Delaware Fioralba Cakoni Rutgers University Research supported by AFOSR Grant FA 9550-13-1-0199 Scattering

More information

Transmission Eigenvalues in Inverse Scattering Theory

Transmission Eigenvalues in Inverse Scattering Theory Transmission Eigenvalues in Inverse Scattering Theory David Colton Department of Mathematical Sciences University of Delaware Newark, DE 19716, USA email: colton@math.udel.edu Research supported by a grant

More information

RECENT DEVELOPMENTS IN INVERSE ACOUSTIC SCATTERING THEORY

RECENT DEVELOPMENTS IN INVERSE ACOUSTIC SCATTERING THEORY RECENT DEVELOPMENTS IN INVERSE ACOUSTIC SCATTERING THEORY DAVID COLTON, JOE COYLE, AND PETER MONK Abstract. We survey some of the highlights of inverse scattering theory as it has developed over the past

More information

Partial Differential Equations

Partial Differential Equations Part II Partial Differential Equations Year 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2015 Paper 4, Section II 29E Partial Differential Equations 72 (a) Show that the Cauchy problem for u(x,

More information

Proceedings of the 5th International Conference on Inverse Problems in Engineering: Theory and Practice, Cambridge, UK, 11-15th July 2005

Proceedings of the 5th International Conference on Inverse Problems in Engineering: Theory and Practice, Cambridge, UK, 11-15th July 2005 Proceedings of the 5th International Conference on Inverse Problems in Engineering: Theory and Practice, Cambridge, UK, 11-15th July 2005 SOME INVERSE SCATTERING PROBLEMS FOR TWO-DIMENSIONAL SCHRÖDINGER

More information

Chapter 7: Bounded Operators in Hilbert Spaces

Chapter 7: Bounded Operators in Hilbert Spaces Chapter 7: Bounded Operators in Hilbert Spaces I-Liang Chern Department of Applied Mathematics National Chiao Tung University and Department of Mathematics National Taiwan University Fall, 2013 1 / 84

More information

Estimation of transmission eigenvalues and the index of refraction from Cauchy data

Estimation of transmission eigenvalues and the index of refraction from Cauchy data Estimation of transmission eigenvalues and the index of refraction from Cauchy data Jiguang Sun Abstract Recently the transmission eigenvalue problem has come to play an important role and received a lot

More information

Transmission Eigenvalues in Inverse Scattering Theory

Transmission Eigenvalues in Inverse Scattering Theory Transmission Eigenvalues in Inverse Scattering Theory Fioralba Cakoni Department of Mathematical Sciences University of Delaware Newark, DE 19716, USA email: cakoni@math.udel.edu Jointly with D. Colton,

More information

Recall that any inner product space V has an associated norm defined by

Recall that any inner product space V has an associated norm defined by Hilbert Spaces Recall that any inner product space V has an associated norm defined by v = v v. Thus an inner product space can be viewed as a special kind of normed vector space. In particular every inner

More information

Inverse scattering problem with underdetermined data.

Inverse scattering problem with underdetermined data. Math. Methods in Natur. Phenom. (MMNP), 9, N5, (2014), 244-253. Inverse scattering problem with underdetermined data. A. G. Ramm Mathematics epartment, Kansas State University, Manhattan, KS 66506-2602,

More information

NEW RESULTS ON TRANSMISSION EIGENVALUES. Fioralba Cakoni. Drossos Gintides

NEW RESULTS ON TRANSMISSION EIGENVALUES. Fioralba Cakoni. Drossos Gintides Inverse Problems and Imaging Volume 0, No. 0, 0, 0 Web site: http://www.aimsciences.org NEW RESULTS ON TRANSMISSION EIGENVALUES Fioralba Cakoni epartment of Mathematical Sciences University of elaware

More information

The Factorization Method for Inverse Scattering Problems Part I

The Factorization Method for Inverse Scattering Problems Part I The Factorization Method for Inverse Scattering Problems Part I Andreas Kirsch Madrid 2011 Department of Mathematics KIT University of the State of Baden-Württemberg and National Large-scale Research Center

More information

The Interior Transmission Eigenvalue Problem for Maxwell s Equations

The Interior Transmission Eigenvalue Problem for Maxwell s Equations The Interior Transmission Eigenvalue Problem for Maxwell s Equations Andreas Kirsch MSRI 2010 epartment of Mathematics KIT University of the State of Baden-Württemberg and National Large-scale Research

More information

u xx + u yy = 0. (5.1)

u xx + u yy = 0. (5.1) Chapter 5 Laplace Equation The following equation is called Laplace equation in two independent variables x, y: The non-homogeneous problem u xx + u yy =. (5.1) u xx + u yy = F, (5.) where F is a function

More information

Inverse scattering problem from an impedance obstacle

Inverse scattering problem from an impedance obstacle Inverse Inverse scattering problem from an impedance obstacle Department of Mathematics, NCKU 5 th Workshop on Boundary Element Methods, Integral Equations and Related Topics in Taiwan NSYSU, October 4,

More information

Factorization method in inverse

Factorization method in inverse Title: Name: Affil./Addr.: Factorization method in inverse scattering Armin Lechleiter University of Bremen Zentrum für Technomathematik Bibliothekstr. 1 28359 Bremen Germany Phone: +49 (421) 218-63891

More information

The Factorization Method for Maxwell s Equations

The Factorization Method for Maxwell s Equations The Factorization Method for Maxwell s Equations Andreas Kirsch University of Karlsruhe Department of Mathematics 76128 Karlsruhe Germany December 15, 2004 Abstract: The factorization method can be applied

More information

PART IV Spectral Methods

PART IV Spectral Methods PART IV Spectral Methods Additional References: R. Peyret, Spectral methods for incompressible viscous flow, Springer (2002), B. Mercier, An introduction to the numerical analysis of spectral methods,

More information

Chapter 12. Partial di erential equations Di erential operators in R n. The gradient and Jacobian. Divergence and rotation

Chapter 12. Partial di erential equations Di erential operators in R n. The gradient and Jacobian. Divergence and rotation Chapter 12 Partial di erential equations 12.1 Di erential operators in R n The gradient and Jacobian We recall the definition of the gradient of a scalar function f : R n! R, as @f grad f = rf =,..., @f

More information

PDEs, part 1: Introduction and elliptic PDEs

PDEs, part 1: Introduction and elliptic PDEs PDEs, part 1: Introduction and elliptic PDEs Anna-Karin Tornberg Mathematical Models, Analysis and Simulation Fall semester, 2013 Partial di erential equations The solution depends on several variables,

More information

Laplace s Equation. Chapter Mean Value Formulas

Laplace s Equation. Chapter Mean Value Formulas Chapter 1 Laplace s Equation Let be an open set in R n. A function u C 2 () is called harmonic in if it satisfies Laplace s equation n (1.1) u := D ii u = 0 in. i=1 A function u C 2 () is called subharmonic

More information

Hilbert Spaces. Hilbert space is a vector space with some extra structure. We start with formal (axiomatic) definition of a vector space.

Hilbert Spaces. Hilbert space is a vector space with some extra structure. We start with formal (axiomatic) definition of a vector space. Hilbert Spaces Hilbert space is a vector space with some extra structure. We start with formal (axiomatic) definition of a vector space. Vector Space. Vector space, ν, over the field of complex numbers,

More information

Global minimization. Chapter Upper and lower bounds

Global minimization. Chapter Upper and lower bounds Chapter 1 Global minimization The issues related to the behavior of global minimization problems along a sequence of functionals F are by now well understood, and mainly rely on the concept of -limit.

More information

Green s Functions and Distributions

Green s Functions and Distributions CHAPTER 9 Green s Functions and Distributions 9.1. Boundary Value Problems We would like to study, and solve if possible, boundary value problems such as the following: (1.1) u = f in U u = g on U, where

More information

Notes on Transmission Eigenvalues

Notes on Transmission Eigenvalues Notes on Transmission Eigenvalues Cédric Bellis December 28, 2011 Contents 1 Scattering by inhomogeneous medium 1 2 Inverse scattering via the linear sampling method 2 2.1 Relationship with the solution

More information

Functional Analysis Review

Functional Analysis Review Functional Analysis Review Lorenzo Rosasco slides courtesy of Andre Wibisono 9.520: Statistical Learning Theory and Applications September 9, 2013 1 2 3 4 Vector Space A vector space is a set V with binary

More information

Some Old and Some New Results in Inverse Obstacle Scattering

Some Old and Some New Results in Inverse Obstacle Scattering Some Old and Some New Results in Inverse Obstacle Scattering Rainer Kress Abstract We will survey on uniqueness, that is, identifiability and on reconstruction issues for inverse obstacle scattering for

More information

Kirchhoff, Fresnel, Fraunhofer, Born approximation and more

Kirchhoff, Fresnel, Fraunhofer, Born approximation and more Kirchhoff, Fresnel, Fraunhofer, Born approximation and more Oberseminar, May 2008 Maxwell equations Or: X-ray wave fields X-rays are electromagnetic waves with wave length from 10 nm to 1 pm, i.e., 10

More information

Convergence for periodic Fourier series

Convergence for periodic Fourier series Chapter 8 Convergence for periodic Fourier series We are now in a position to address the Fourier series hypothesis that functions can realized as the infinite sum of trigonometric functions discussed

More information

Transmission Eigenvalues in Inverse Scattering Theory

Transmission Eigenvalues in Inverse Scattering Theory Transmission Eigenvalues in Inverse Scattering Theory Fioralba Cakoni, Houssem Haddar To cite this version: Fioralba Cakoni, Houssem Haddar. Transmission Eigenvalues in Inverse Scattering Theory. Gunther

More information

THE PROBLEMS FOR THE SECOND TEST FOR BRIEF SOLUTIONS

THE PROBLEMS FOR THE SECOND TEST FOR BRIEF SOLUTIONS THE PROBLEMS FOR THE SECOND TEST FOR 18.102 BRIEF SOLUTIONS RICHARD MELROSE Question.1 Show that a subset of a separable Hilbert space is compact if and only if it is closed and bounded and has the property

More information

Recent Developments in Inverse Acoustic Scattering Theory

Recent Developments in Inverse Acoustic Scattering Theory SIAM REVIEW Vol. 42, No. 3, pp. 369 414 c 2000 Society for Industrial and Applied Mathematics Recent Developments in Inverse Acoustic Scattering Theory David Colton Joe Coyle Peter Monk Abstract. We survey

More information

Infinite-dimensional Vector Spaces and Sequences

Infinite-dimensional Vector Spaces and Sequences 2 Infinite-dimensional Vector Spaces and Sequences After the introduction to frames in finite-dimensional vector spaces in Chapter 1, the rest of the book will deal with expansions in infinitedimensional

More information

Vector Spaces. Vector space, ν, over the field of complex numbers, C, is a set of elements a, b,..., satisfying the following axioms.

Vector Spaces. Vector space, ν, over the field of complex numbers, C, is a set of elements a, b,..., satisfying the following axioms. Vector Spaces Vector space, ν, over the field of complex numbers, C, is a set of elements a, b,..., satisfying the following axioms. For each two vectors a, b ν there exists a summation procedure: a +

More information

Looking Back on Inverse Scattering Theory

Looking Back on Inverse Scattering Theory Looking Back on Inverse Scattering Theory David Colton and Rainer Kress History will be kind to me for I intend to write it Abstract Winston Churchill We present an essay on the mathematical development

More information

Functional Analysis. Franck Sueur Metric spaces Definitions Completeness Compactness Separability...

Functional Analysis. Franck Sueur Metric spaces Definitions Completeness Compactness Separability... Functional Analysis Franck Sueur 2018-2019 Contents 1 Metric spaces 1 1.1 Definitions........................................ 1 1.2 Completeness...................................... 3 1.3 Compactness......................................

More information

The Inside-Outside Duality for Scattering Problems by Inhomogeneous Media

The Inside-Outside Duality for Scattering Problems by Inhomogeneous Media The Inside-Outside uality for Scattering Problems by Inhomogeneous Media Andreas Kirsch epartment of Mathematics Karlsruhe Institute of Technology (KIT) 76131 Karlsruhe Germany and Armin Lechleiter Center

More information

Spectral Properties of Elliptic Operators In previous work we have replaced the strong version of an elliptic boundary value problem

Spectral Properties of Elliptic Operators In previous work we have replaced the strong version of an elliptic boundary value problem Spectral Properties of Elliptic Operators In previous work we have replaced the strong version of an elliptic boundary value problem L u x f x BC u x g x with the weak problem find u V such that B u,v

More information

Second Order Elliptic PDE

Second Order Elliptic PDE Second Order Elliptic PDE T. Muthukumar tmk@iitk.ac.in December 16, 2014 Contents 1 A Quick Introduction to PDE 1 2 Classification of Second Order PDE 3 3 Linear Second Order Elliptic Operators 4 4 Periodic

More information

Analysis Preliminary Exam Workshop: Hilbert Spaces

Analysis Preliminary Exam Workshop: Hilbert Spaces Analysis Preliminary Exam Workshop: Hilbert Spaces 1. Hilbert spaces A Hilbert space H is a complete real or complex inner product space. Consider complex Hilbert spaces for definiteness. If (, ) : H H

More information

Real Analysis Qualifying Exam May 14th 2016

Real Analysis Qualifying Exam May 14th 2016 Real Analysis Qualifying Exam May 4th 26 Solve 8 out of 2 problems. () Prove the Banach contraction principle: Let T be a mapping from a complete metric space X into itself such that d(tx,ty) apple qd(x,

More information

Short note on compact operators - Monday 24 th March, Sylvester Eriksson-Bique

Short note on compact operators - Monday 24 th March, Sylvester Eriksson-Bique Short note on compact operators - Monday 24 th March, 2014 Sylvester Eriksson-Bique 1 Introduction In this note I will give a short outline about the structure theory of compact operators. I restrict attention

More information

arxiv: v1 [math.ap] 21 Dec 2018

arxiv: v1 [math.ap] 21 Dec 2018 Uniqueness to Inverse Acoustic and Electromagnetic Scattering From Locally Perturbed Rough Surfaces Yu Zhao, Guanghui Hu, Baoqiang Yan arxiv:1812.09009v1 [math.ap] 21 Dec 2018 Abstract In this paper, we

More information

ON THE EXISTENCE OF TRANSMISSION EIGENVALUES. Andreas Kirsch1

ON THE EXISTENCE OF TRANSMISSION EIGENVALUES. Andreas Kirsch1 Manuscript submitted to AIMS Journals Volume 3, Number 2, May 29 Website: http://aimsciences.org pp. 1 XX ON THE EXISTENCE OF TRANSMISSION EIGENVALUES Andreas Kirsch1 University of Karlsruhe epartment

More information

SPECTRAL PROPERTIES OF THE LAPLACIAN ON BOUNDED DOMAINS

SPECTRAL PROPERTIES OF THE LAPLACIAN ON BOUNDED DOMAINS SPECTRAL PROPERTIES OF THE LAPLACIAN ON BOUNDED DOMAINS TSOGTGEREL GANTUMUR Abstract. After establishing discrete spectra for a large class of elliptic operators, we present some fundamental spectral properties

More information

Inverse wave scattering problems: fast algorithms, resonance and applications

Inverse wave scattering problems: fast algorithms, resonance and applications Inverse wave scattering problems: fast algorithms, resonance and applications Wagner B. Muniz Department of Mathematics Federal University of Santa Catarina (UFSC) w.b.muniz@ufsc.br III Colóquio de Matemática

More information

MA5206 Homework 4. Group 4. April 26, ϕ 1 = 1, ϕ n (x) = 1 n 2 ϕ 1(n 2 x). = 1 and h n C 0. For any ξ ( 1 n, 2 n 2 ), n 3, h n (t) ξ t dt

MA5206 Homework 4. Group 4. April 26, ϕ 1 = 1, ϕ n (x) = 1 n 2 ϕ 1(n 2 x). = 1 and h n C 0. For any ξ ( 1 n, 2 n 2 ), n 3, h n (t) ξ t dt MA526 Homework 4 Group 4 April 26, 26 Qn 6.2 Show that H is not bounded as a map: L L. Deduce from this that H is not bounded as a map L L. Let {ϕ n } be an approximation of the identity s.t. ϕ C, sptϕ

More information

Discreteness of Transmission Eigenvalues via Upper Triangular Compact Operators

Discreteness of Transmission Eigenvalues via Upper Triangular Compact Operators Discreteness of Transmission Eigenvalues via Upper Triangular Compact Operators John Sylvester Department of Mathematics University of Washington Seattle, Washington 98195 U.S.A. June 3, 2011 This research

More information

Integral Representation Formula, Boundary Integral Operators and Calderón projection

Integral Representation Formula, Boundary Integral Operators and Calderón projection Integral Representation Formula, Boundary Integral Operators and Calderón projection Seminar BEM on Wave Scattering Franziska Weber ETH Zürich October 22, 2010 Outline Integral Representation Formula Newton

More information

Uniqueness in determining refractive indices by formally determined far-field data

Uniqueness in determining refractive indices by formally determined far-field data Applicable Analysis, 2015 Vol. 94, No. 6, 1259 1269, http://dx.doi.org/10.1080/00036811.2014.924215 Uniqueness in determining refractive indices by formally determined far-field data Guanghui Hu a, Jingzhi

More information

The Dirichlet-to-Neumann operator

The Dirichlet-to-Neumann operator Lecture 8 The Dirichlet-to-Neumann operator The Dirichlet-to-Neumann operator plays an important role in the theory of inverse problems. In fact, from measurements of electrical currents at the surface

More information

1. Subspaces A subset M of Hilbert space H is a subspace of it is closed under the operation of forming linear combinations;i.e.,

1. Subspaces A subset M of Hilbert space H is a subspace of it is closed under the operation of forming linear combinations;i.e., Abstract Hilbert Space Results We have learned a little about the Hilbert spaces L U and and we have at least defined H 1 U and the scale of Hilbert spaces H p U. Now we are going to develop additional

More information

Solutions: Problem Set 3 Math 201B, Winter 2007

Solutions: Problem Set 3 Math 201B, Winter 2007 Solutions: Problem Set 3 Math 201B, Winter 2007 Problem 1. Prove that an infinite-dimensional Hilbert space is a separable metric space if and only if it has a countable orthonormal basis. Solution. If

More information

Vectors in Function Spaces

Vectors in Function Spaces Jim Lambers MAT 66 Spring Semester 15-16 Lecture 18 Notes These notes correspond to Section 6.3 in the text. Vectors in Function Spaces We begin with some necessary terminology. A vector space V, also

More information

THE INTERIOR TRANSMISSION PROBLEM FOR REGIONS WITH CAVITIES

THE INTERIOR TRANSMISSION PROBLEM FOR REGIONS WITH CAVITIES THE INTERIOR TRANSMISSION PROBLEM FOR REGIONS WITH CAVITIES FIORALBA CAKONI, AVI COLTON, AN HOUSSEM HAAR Abstract. We consider the interior transmission problem in the case when the inhomogeneous medium

More information

Functional Analysis: Assignment Set # 11 Spring 2009 Professor: Fengbo Hang April 29, 2009

Functional Analysis: Assignment Set # 11 Spring 2009 Professor: Fengbo Hang April 29, 2009 Eduardo Corona Functional Analysis: Assignment Set # Spring 2009 Professor: Fengbo Hang April 29, 2009 29. (i) Every convolution operator commutes with translation: ( c u)(x) = u(x + c) (ii) Any two convolution

More information

Review and problem list for Applied Math I

Review and problem list for Applied Math I Review and problem list for Applied Math I (This is a first version of a serious review sheet; it may contain errors and it certainly omits a number of topic which were covered in the course. Let me know

More information

A global solution curve for a class of free boundary value problems arising in plasma physics

A global solution curve for a class of free boundary value problems arising in plasma physics A global solution curve for a class of free boundary value problems arising in plasma physics Philip Korman epartment of Mathematical Sciences University of Cincinnati Cincinnati Ohio 4522-0025 Abstract

More information

1 Math 241A-B Homework Problem List for F2015 and W2016

1 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 information

1 Review of di erential calculus

1 Review of di erential calculus Review of di erential calculus This chapter presents the main elements of di erential calculus needed in probability theory. Often, students taking a course on probability theory have problems with concepts

More information

Ordinary Differential Equation Theory

Ordinary Differential Equation Theory Part I Ordinary Differential Equation Theory 1 Introductory Theory An n th order ODE for y = y(t) has the form Usually it can be written F (t, y, y,.., y (n) ) = y (n) = f(t, y, y,.., y (n 1) ) (Implicit

More information

The Factorization Method for a Class of Inverse Elliptic Problems

The Factorization Method for a Class of Inverse Elliptic Problems 1 The Factorization Method for a Class of Inverse Elliptic Problems Andreas Kirsch Mathematisches Institut II Universität Karlsruhe (TH), Germany email: kirsch@math.uni-karlsruhe.de Version of June 20,

More information

Inverse Scattering Theory and Transmission Eigenvalues

Inverse Scattering Theory and Transmission Eigenvalues Inverse Scattering Theory and Transmission Eigenvalues David Colton Department of Mathematical Sciences University of Delaware Newark, DE 19716, USA email: colton@udel.edu Research supported a grant from

More information

Elliptic Regularity. Throughout we assume all vector bundles are smooth bundles with metrics over a Riemannian manifold X n.

Elliptic Regularity. Throughout we assume all vector bundles are smooth bundles with metrics over a Riemannian manifold X n. Elliptic Regularity Throughout we assume all vector bundles are smooth bundles with metrics over a Riemannian manifold X n. 1 Review of Hodge Theory In this note I outline the proof of the following Fundamental

More information

1 Distributions (due January 22, 2009)

1 Distributions (due January 22, 2009) Distributions (due January 22, 29). The distribution derivative of the locally integrable function ln( x ) is the principal value distribution /x. We know that, φ = lim φ(x) dx. x ɛ x Show that x, φ =

More information

LECTURE 1: SOURCES OF ERRORS MATHEMATICAL TOOLS A PRIORI ERROR ESTIMATES. Sergey Korotov,

LECTURE 1: SOURCES OF ERRORS MATHEMATICAL TOOLS A PRIORI ERROR ESTIMATES. Sergey Korotov, LECTURE 1: SOURCES OF ERRORS MATHEMATICAL TOOLS A PRIORI ERROR ESTIMATES Sergey Korotov, Institute of Mathematics Helsinki University of Technology, Finland Academy of Finland 1 Main Problem in Mathematical

More information

MATH 124B Solution Key HW 05

MATH 124B Solution Key HW 05 7.1 GREEN S FIRST IENTITY MATH 14B Solution Key HW 05 7.1 GREEN S FIRST IENTITY 1. erive the 3-dimensional maximum principle from the mean value property. SOLUTION. We aim to prove that if u is harmonic

More information

CHAPTER VIII HILBERT SPACES

CHAPTER VIII HILBERT SPACES CHAPTER VIII HILBERT SPACES DEFINITION Let X and Y be two complex vector spaces. A map T : X Y is called a conjugate-linear transformation if it is a reallinear transformation from X into Y, and if T (λx)

More information

MAT 578 FUNCTIONAL ANALYSIS EXERCISES

MAT 578 FUNCTIONAL ANALYSIS EXERCISES MAT 578 FUNCTIONAL ANALYSIS EXERCISES JOHN QUIGG Exercise 1. Prove that if A is bounded in a topological vector space, then for every neighborhood V of 0 there exists c > 0 such that tv A for all t > c.

More information

Lecture Notes on PDEs

Lecture Notes on PDEs Lecture Notes on PDEs Alberto Bressan February 26, 2012 1 Elliptic equations Let IR n be a bounded open set Given measurable functions a ij, b i, c : IR, consider the linear, second order differential

More information

Volume and surface integral equations for electromagnetic scattering by a dielectric body

Volume and surface integral equations for electromagnetic scattering by a dielectric body Volume and surface integral equations for electromagnetic scattering by a dielectric body M. Costabel, E. Darrigrand, and E. H. Koné IRMAR, Université de Rennes 1,Campus de Beaulieu, 35042 Rennes, FRANCE

More information

Review of complex analysis in one variable

Review of complex analysis in one variable CHAPTER 130 Review of complex analysis in one variable This gives a brief review of some of the basic results in complex analysis. In particular, it outlines the background in single variable complex analysis

More information

Chapter 1. Preliminaries. The purpose of this chapter is to provide some basic background information. Linear Space. Hilbert Space.

Chapter 1. Preliminaries. The purpose of this chapter is to provide some basic background information. Linear Space. Hilbert Space. Chapter 1 Preliminaries The purpose of this chapter is to provide some basic background information. Linear Space Hilbert Space Basic Principles 1 2 Preliminaries Linear Space The notion of linear space

More information

A Direct Method for reconstructing inclusions from Electrostatic Data

A Direct Method for reconstructing inclusions from Electrostatic Data A Direct Method for reconstructing inclusions from Electrostatic Data Isaac Harris Texas A&M University, Department of Mathematics College Station, Texas 77843-3368 iharris@math.tamu.edu Joint work with:

More information

Chapter 8 Integral Operators

Chapter 8 Integral Operators Chapter 8 Integral Operators In our development of metrics, norms, inner products, and operator theory in Chapters 1 7 we only tangentially considered topics that involved the use of Lebesgue measure,

More information

Functional Analysis. Martin Brokate. 1 Normed Spaces 2. 2 Hilbert Spaces The Principle of Uniform Boundedness 32

Functional Analysis. Martin Brokate. 1 Normed Spaces 2. 2 Hilbert Spaces The Principle of Uniform Boundedness 32 Functional Analysis Martin Brokate Contents 1 Normed Spaces 2 2 Hilbert Spaces 2 3 The Principle of Uniform Boundedness 32 4 Extension, Reflexivity, Separation 37 5 Compact subsets of C and L p 46 6 Weak

More information

DIRECT SAMPLING METHODS FOR INVERSE SCATTERING PROBLEMS

DIRECT SAMPLING METHODS FOR INVERSE SCATTERING PROBLEMS Michigan Technological University Digital Commons @ Michigan Tech Dissertations, Master's Theses and Master's Reports 2017 DIRECT SAMPLING METHODS FOR INVERSE SCATTERING PROBLEMS Ala Mahmood Nahar Al Zaalig

More information

arxiv: v3 [math.ap] 4 Jan 2017

arxiv: v3 [math.ap] 4 Jan 2017 Recovery of an embedded obstacle and its surrounding medium by formally-determined scattering data Hongyu Liu 1 and Xiaodong Liu arxiv:1610.05836v3 [math.ap] 4 Jan 017 1 Department of Mathematics, Hong

More information

Math 350 Fall 2011 Notes about inner product spaces. In this notes we state and prove some important properties of inner product spaces.

Math 350 Fall 2011 Notes about inner product spaces. In this notes we state and prove some important properties of inner product spaces. Math 350 Fall 2011 Notes about inner product spaces In this notes we state and prove some important properties of inner product spaces. First, recall the dot product on R n : if x, y R n, say x = (x 1,...,

More information

Multivariable Calculus

Multivariable Calculus 2 Multivariable Calculus 2.1 Limits and Continuity Problem 2.1.1 (Fa94) Let the function f : R n R n satisfy the following two conditions: (i) f (K ) is compact whenever K is a compact subset of R n. (ii)

More information

3. Fourier decomposition of functions

3. Fourier decomposition of functions 22 C. MOUHOT 3.1. The Fourier transform. 3. Fourier decomposition of functions Definition 3.1 (Fourier Transform on L 1 (R d )). Given f 2 L 1 (R d ) define its Fourier transform F(f)( ) := R d e 2i x

More information

The Linear Sampling Method and the MUSIC Algorithm

The Linear Sampling Method and the MUSIC Algorithm CODEN:LUTEDX/(TEAT-7089)/1-6/(2000) The Linear Sampling Method and the MUSIC Algorithm Margaret Cheney Department of Electroscience Electromagnetic Theory Lund Institute of Technology Sweden Margaret Cheney

More information

RNDr. Petr Tomiczek CSc.

RNDr. Petr Tomiczek CSc. HABILITAČNÍ PRÁCE RNDr. Petr Tomiczek CSc. Plzeň 26 Nonlinear differential equation of second order RNDr. Petr Tomiczek CSc. Department of Mathematics, University of West Bohemia 5 Contents 1 Introduction

More information

7: FOURIER SERIES STEVEN HEILMAN

7: FOURIER SERIES STEVEN HEILMAN 7: FOURIER SERIES STEVE HEILMA Contents 1. Review 1 2. Introduction 1 3. Periodic Functions 2 4. Inner Products on Periodic Functions 3 5. Trigonometric Polynomials 5 6. Periodic Convolutions 7 7. Fourier

More information

LECTURE 7. k=1 (, v k)u k. Moreover r

LECTURE 7. k=1 (, v k)u k. Moreover r LECTURE 7 Finite rank operators Definition. T is said to be of rank r (r < ) if dim T(H) = r. The class of operators of rank r is denoted by K r and K := r K r. Theorem 1. T K r iff T K r. Proof. Let T

More information

Spectral Theory, with an Introduction to Operator Means. William L. Green

Spectral Theory, with an Introduction to Operator Means. William L. Green Spectral Theory, with an Introduction to Operator Means William L. Green January 30, 2008 Contents Introduction............................... 1 Hilbert Space.............................. 4 Linear Maps

More information

A new method for the solution of scattering problems

A new method for the solution of scattering problems A new method for the solution of scattering problems Thorsten Hohage, Frank Schmidt and Lin Zschiedrich Konrad-Zuse-Zentrum Berlin, hohage@zibde * after February 22: University of Göttingen Abstract We

More information

COMMON COMPLEMENTS OF TWO SUBSPACES OF A HILBERT SPACE

COMMON COMPLEMENTS OF TWO SUBSPACES OF A HILBERT SPACE COMMON COMPLEMENTS OF TWO SUBSPACES OF A HILBERT SPACE MICHAEL LAUZON AND SERGEI TREIL Abstract. In this paper we find a necessary and sufficient condition for two closed subspaces, X and Y, of a Hilbert

More information

4.2 Green s representation theorem

4.2 Green s representation theorem 4.2. REEN S REPRESENTATION THEOREM 57 i.e., the normal velocity on the boundary is proportional to the ecess pressure on the boundary. The coefficient χ is called the acoustic impedance of the obstacle

More information

Transmission eigenvalues in inverse scattering theory

Transmission eigenvalues in inverse scattering theory Inside Out II MSRI Publications Volume 6, 22 Transmission eigenvalues in inverse scattering theory FIORALBA CAKONI AN HOUSSEM HAAR In the past few years transmission eigenvalues have become an important

More information

Waves on 2 and 3 dimensional domains

Waves on 2 and 3 dimensional domains Chapter 14 Waves on 2 and 3 dimensional domains We now turn to the studying the initial boundary value problem for the wave equation in two and three dimensions. In this chapter we focus on the situation

More information