84 G. Schmidt (1.3) T : '! u := uj ; where u is the solution of (1.1{2). It is well known (cf. [7], [1]) that the linear operator T : H ( )! e H ( ) i

Size: px
Start display at page:

Download "84 G. Schmidt (1.3) T : '! u := uj ; where u is the solution of (1.1{2). It is well known (cf. [7], [1]) that the linear operator T : H ( )! e H ( ) i"

Transcription

1 Numer. Math. 69: 83{11 (1994) Numerische Mathematik c Springer-Verlag 1994 Electronic Edition Boundary element discretization of Poincare{Steklov operators Gunther Schmidt Institut fur Angewandte Analysis und Stochastik, Mohrenstrae 39, D-1117 Berlin, Germany Received October 1992 / Revised version received April 14, 1994 Summary. The construction of a discretization of Poincare{Steklov operators with the boundary element method is given providing the same mapping properties as the nite element discretization of these operators. Mathematics Subject Classication (1991): 65N38, 65N12, 65N55, 65N3, 31A1, 31B1 1. Introduction Poincare{Steklov (PS) operators are natural mathematical tools for the investigation of boundary value problems and their numerical solution with domain decomposition (DD) methods based on the nite element (FE) solution of the subproblems ([1], [9]). The purpose of this paper is to construct discretizations of PS operators for elliptic boundary value problems with the boundary element method (BEM). We will show that the discretizations obtained via a direct Galerkin BEM possess the same mapping properties as the FE discretizations if the boundary elements satisfy some natural conditions. Hence the given construction provides a base for the analysis of dierent DD methods using the BE solution of subproblems, of the coupling of FE and BE methods and related problems. As model problem we consider the mixed boundary value problem for the Laplacian. Let R d (d = 2; 3) be a bounded domain with piecewise smooth boundary such that angles at corners and edges do not degenerate. The boundary is partitioned into 3 nonintersecting domains = D [ N [ and suppose that D 6= ;, \ N = ;. Let ' 2 H ( ) be given and consider the problem: Find u 2 H 1 () such that (1.1) u = ; in ; uj D = ; n uj N = ; (1.2) n uj = ' : Here n u denotes the derivative with respect to the outer normal. The PS operator T is dened as the mapping page 83 of Numer. Math. 69: 83{11 (1994)

2 84 G. Schmidt (1.3) T : '! u := uj ; where u is the solution of (1.1{2). It is well known (cf. [7], [1]) that the linear operator T : H ( )! e H ( ) is bounded and invertible, symmetric with respect to the duality between H ( ) and e H ( ), induced by the scalar product of L 2 ( ), and positive denite (T '; ) = ('; T ) = T '(x) (x)d x ; (T '; ') ck'k 2 H ( ) : Here H s denotes the usual Sobolev spaces: For H s () = fuj : u 2 H s (R d )g ; s 2 R ; H s () = 8 < : fuj : u 2 H s+ (R d )g ; s > ; L 2 () ; s = ; (H s ()) ; s < : H s ( ) = fuj : u 2 H s ()g ; s ; eh s ( ) = (H s ( )) ; s < ; eh s ( ) = fu 2 H s ( ) : ~u 2 H s ()g ; s ; H s ( ) = ( e H s ( )) ; s < ; where ~u denotes the extension of u by zero to. We note that T 1 = S, where (1.4) S : 2 e H ( )! 1 u := n uj and u solves the problem (1.1) together with the Dirichlet boundary condition on (1.5) u = The FE discretization of the PS operator T and its inverse S follows immediately from the variational formulation of (1.1). Let us suppose that we are given a space V h H 1 () of nite element functions on vanishing on D Denote by V h = fv h : v h j D = g ; dim V h < 1 : o V h = fv h 2 V h : v h = g ; ex h = f v h : v h 2 V h g e H ( ) : For given h 2 e Xh we determine u h 2 V h such that u h = h page 84 of Numer. Math. 69: 83{11 (1994)

3 Boundary element discretization of Poincare{Steklov operators 85 and (1.6) a(u h ; v h ) := ru h rv h d! = ; 8 v h 2 o V h : Then a(u h ; v h ), v h 2 V h, denes a linear functional of v h 2 e Xh bounded in the e H ( ){norm. Let P h : e H ( )! e Xh be a bounded projection onto Xh e, by Ph we denote its L2 ( ){adjoint projection and Y h := Im Ph H ( ) can be identied with the dual space of Xh e. Hence there exists h 2 Y h ; k h k H ( ) c such that a(u h ; v h ) = ( h ; v h ) : It can be easily seen (cf. [1]) that the mapping is linear and invertible, S h : e Xh! Y h dened by S h h = h (1.7) ks h h k H ( ) ck hk eh ( ) ; 8 h 2 e Xh ; symmetric (S h h ; h) = ( h ; S h h) = a(u h ; u h) ; where u h solves (1.6) with u h = h, and positive denite Moreover (1.8) (S h h ; h ) = a(u h ; u h ) c 1 ku h k 2 H 1 () ck hk 2 eh ( ) : k(s S h ) h k H ( ) k(p hs S h ) h k H ( ) + k(i P h)s h k H ( ) c c ((S S h ) h ; v h ) sup v h 2V h k v hk eh ( ) + inf ' h 2Y h k 1 u ' h k H ( ) inf ku v h k H v h 2V 1 () + inf k 1 u ' h k H h ' h 2Y ( ) h where u is the exact solution of (1.1) with u = h. We note that the constants (1.7) and (1.8) depend on the norm of P h. Thus, if kp hk is uniformly bounded with respect to h, then estimates (1.7) and eh ( ) (1.8) hold with constants independent of h and h. Now it is clear how to dene the FE discretization of T. Choose Y h and h 2 Y h, nd u h 2 V h such that a(u h ; v h ) = ( h ; v h ) ; 8 v h 2 V h ; and dene T h h := u h. Since T h = S h 1, this operator has analogous properties as S h and T, moreover k(t T h ) hk eh ( ) c inf v h 2V h ku v h k H 1 () ; ;! page 85 of Numer. Math. 69: 83{11 (1994)

4 86 G. Schmidt where u solves (1.1) and 1 u = h. The mentioned mapping properties of PS operators and their FE discretizations were essentially used (sometimes unknowingly) for the formulation and analysis of various DD methods, which are in fact equivalent to the iterative solution of operator equations with PS operators of dierent subdomains. In the following we will prove that a direct Galerkin BE solution of problem (1.1) (considered in Sect. 2) yields discretizations S h and T h of the operators S and T with the same mapping properties mentioned above (proved in Sect. 3). Therefore convergence results for many DD methods remain valid if the FE solution of subproblems is replaced by their BE solution whenever it is possible. 2. A direct Galerkin BEM In this section we discuss some results concerning a boundary integral method for solving mixed boundary value problems for the Laplacian. Let u 2 H 1 (), R d, be a solution of (2.1) u = in : Then we have the representation (2.2) u(x) = 1 2 G(x; y) n u(y)d y 1 2 with the fundamental solution G(x; y) = 8 >< >: 1 lnjx yj ; d = jx yj1 ; d = 3 : ny G(x; y)u(y)d y ; x 2 ; Let us dene the boundary integral operators for x 2 V '(x) := K '(x) := G(x; y)'(y)d y ; K'(x) := nx G(x; y)'(y)d y ; D'(x) := nx The following properties are well known (cf. Costabel (1988)): ny G(x; y)'(y)d y ; ny G(x; y)'(y)d y : V : H + ()! H + () ; D : H + ()! H + () ; K : H + ()! H + () ; K : H + ()! H + () (2.3) are continuous for jj. The operator K is the adjoint of K with respect to the duality between H () and H () induced by the scalar product (; ) L 2 (). Moreover, page 86 of Numer. Math. 69: 83{11 (1994)

5 Boundary element discretization of Poincare{Steklov operators 87 (Dv; v) L 2 () c jvj H () ; c > ; for all v 2 H () ; j j H () denotes the seminorm ; (V ; ) L 2 () c k k H () ; c > ; for all 2 H () if d = 3 ; (2:4) and for all 2 H () with ( ; 1) L 2 () = if d = 2 : To ensure the solvability of boundary integral equations we need that the operator V is invertible. Therefore if d = 2 then we assume in the following that cap () 6= 1, i.e. R 2 has the property (P) If 2 H () solves V = then =. Note that for any R 2 the domains m = fmx : x 2 g, m >, satisfy (P) with the exception of one value m e (cf. Symm (1967)). Now the representation (2.2) and the jump relations for single and double layer potentials lead to the equalities on (2.5) u = 1 2 ((I K)u + V nu) ; n u = 1 2 (Du + (I + K ) n u) : Hence, if we consider the mixed boundary value problem for (2.1) (2.6) uj 1 = g 1 ; n uj 2 = g 2 ; where g 1 2 H ( 1 ), g 2 2 H ( 2 ), 1 [ 2 =, 1 6= ;, and take the limits of u(x) for x 2 2 and of the normal derivative n u for x 2 1, then we get the equation (2.7) D22 K 21 v K 12 V 11 D21 I K 22 g1 = ; I + K 11 V 12 g 2 where the subscripts in D jk, etc., mean integration over k and evaluation on j. Here v := uj 2, := n uj 1 denote the unknown boundary values of the solution u of (2.1), (2.6). If we substitute in (2.7) v = v + lg 1, = + lg 2 with arbitrary extensions lg 1 2 H (), lg 2 2 H () then we obtain a system of boundary integral equations v (2.8) A := D22 K 21 K 12 V 11 v D2 I K lg1 f2 = 2 =: : I + K 1 V 1 lg 2 f 1 Here D 2 for example denotes integration over and evaluation on 2. If g 1 2 H s ( 1 ), g 2 2 H s1 ( 2 ) and s 2 (; 1) then f 2 2 H s1 ( 2 ), f 1 2 H s ( 1 ). Moreover, the operator A : eh s ( 2 ) H s1 ( 2 )! eh s1 ( 1 ) H s ( 1 ) is continuous and satises a Garding inequality in e H ( 2 ) e H ( 1 ), for U = (w; ') 2 e H ( 2 ) e H ( 1 ) we have page 87 of Numer. Math. 69: 83{11 (1994)

6 88 G. Schmidt hau; Ui = (D 22 w + K '; w) 21 L 2 ( 2) + (K 12 w + V 11 '; ') L 2 ( 1) = (D 22 w; w) L 2 ( 2) + (V 11 '; ') L 2 ( 1) = (D ~w; ~w) L 2 () + (V ~'; ~') L 2 () ; where ~w; ~' denote the extension by zero to. Relations (2.4) imply the existence of a compact operator C such that (2.9) h(a + C)U; Ui c kwk 2 eh + ( k'k2 : 2) eh ( 1) Since for d = 3 we can set in (2.9) C = and for d = 2 because of property (P) the system (2.8) has no eigensolutions we derive that (2.1) is invertible. A : eh ( 2 ) H ( 2 )! eh ( 1 ) H ( 1 ) Theorem 2.1. (Costabel and Stephan (1985a), von Petersdorf (1989)). Let g 1 2 H ( 1 ), g 2 2 H ( 2 ) with arbitrary extensions lg 1 2 H (), lg 2 2 H (). Then there exists exactly one solution (v ; ) 2 e H ( 2 ) e H ( 1 ) of (2.8) and v := v + lg 1 j 2 2 H ( 2 ), := + lg 2 j 1 2 H ( 1 ) solve (2.7). Theorem 2.2. Let g 1 2 e H ( 1 ), g 2 2 e H ( 2 ). Then (2.7) is uniquely solvable and c 1 kg 1k eh ( 1) + kg 2k eh ( 2) where c 1, c 2 do not depend on g 1 and g 2. kvk eh ( 2) + k k eh ( 1) c 2 kg 1k eh ( 1) + kg 2k eh ( 2) Proof. The boundary values of the solution of the mixed boundary value problem (2.1), (2.6) satisfy v 2 e H ( 2 ), 2 e H ( 1 ). If we extend g 1 and g 2 by zero, lg 1 j 2 = lg 2 j 1 =, then (v; ) is the unique solution of (2.8) and kvk eh ( 2) + k k eh ( 1) c kf 1 k H ( 1) + kf 2 k H ( 2) c 2 kg 1k eh ( 1) + kg 2k eh ( 2) where we have used (2.3) and (2.1). Since the same considerations can be applied to the boundary value problem for (2.1) with the data uj 2 = v ; n uj 1 = ; ; ; the twosided estimate follows immediately. ut page 88 of Numer. Math. 69: 83{11 (1994)

7 Boundary element discretization of Poincare{Steklov operators 89 The Garding inequality (2.9) yields the convergence of Galerkin methods for the approximate solution of system (2.8). We choose on 2 and 1 nite dimensional sets of approximating functions M h H ( 2 ), N h H e ( 1 ) with lim dim M h = lim dim N h = 1. Furthermore we suppose that g 1 can h! h! be extended by some lg 1 j 2 2 M h and g 2 can be extended by some lg 2 j 1 2 N h which implies that g 2 2 e H ( 2 ). We denote f Mh = M h \ e H ( 2 ) and consider the Galerkin method: Find (v h ; h ) 2 f Mh N h such that (2.11) A v h h ; wh ' h = f2 f 1 ; wh ' h ; 8 (w h ; ' h ) 2 f Mh N h : Under the assumption that [ h fm h and [ h N h are dense in e H ( 2 ) and eh ( 1 ), resp., by standard methods the following results can be obtained. Theorem 2.3. (Costabel and Stephan (1985a)). For any g 1 2 H ( 1 ), g 2 2 eh ( 2 ) and all (suciently small if d = 2) h the Galerkin equations (2.11) are uniquely solvable and kv h k eh ( 2) + k h k eh ( 1) c klg 1 k H () + klg 2 k H () : Moreover, the functions v h := v h + lg 1j 2 and h := h + lg 2j 1 approximate the unknown boundary values quasioptimally kv v h k H ( 2) + k hk eh ( 1) c inf k ' hk ' h 2N eh h ( + inf kv w 1) hk eh w h 2 em ( 2) h where the constant c does not depend on h. Using the regularity of the solution (v ; ) for the case d = 2 (cf. Costabel and Stephan (1985)), the structure of the mapping A and the Aubin{Nitsche Lemma one can estimate the convergence of the approximate solutions in Sobolev norms of lower order. Theorem 2.4. Let d = 2. Then there exist > 1=4 and c > such that kv v h k H (2) + k hk eh (1) c " (h) kv v h k H ( 2) + k hk eh ( 1)! ; where " (h) := sup w2eh + ( 2) '2eH + ( 1) inf w n2 em h ' h 2N h kw w hk eh ( 2) kwk eh + ( 2) + k' ' hk eh ( 1) k'k eh+ (1)! : page 89 of Numer. Math. 69: 83{11 (1994)

8 9 G. Schmidt From the density of [ h fm h and [ h N h we have " (h)! as h!. Note that depends on the inner angles at the corners of and at the points where boundary conditions change (Costabel and Stephan (1985)). We remark that under the condition g 2 2 H e ( 2 ) one can choose lg 2 j 1 =, but for some considerations in Sect. 3 it is useful to admit lg 2 j 1 6=. Now we dene a positive bilinear form for the boundary values of the BE solutions. First of all we note that the Galerkin method (2.11) admits the following equivalent formulation: Find v h 2 M h and h 2 N h such that (2.12) ~v h := g1 v h 2 H (); ~ h := and the boundary values of the function (2.13) u h (x) := 1 2 satisfy the equations (2.14) 1 (u h g 1 )' h d + G(x; y) ~ h(y)d y 1 2 h g 2 2 H () ny G(x; y)~v h (y)d y ; x 2 ; 2 ( n u h g 2 ) w h d = ; 8 (w h ; ' h ) 2 f Mh N h : Under the assumptions of Theorem 2.3 the solution u h exists. From (2.2) and (2.5) we conclude that u h solves (2.1) and on it holds (2.15) u h = 1 2 ((I K)~v h + V ~ h) ; n u h = 1 2 (D~v h + (I + K ) ~ h) : Therefore Theorem 2.3 yields the estimate on the boundary (2.16) ku h uk H () + k n u h n uk H () c inf kv w hk eh w h 2 em ( + inf k ' 2) hk : ' h 2N eh h ( 1) h! Let us now suppose that u h and u h are of the form (2.13) with the densities (~v h, ~ h ) and (~v h, ~ h ), resp., and are solutions of equations (2.14) for corresponding boundary data g 1, g 1 2 e H ( 1 ), g 2, g 2 2 e H ( 2 ). Then we may choose lg 1 j 2 = lg 1j 2 =, whereas lg 2 j 1 ; lg 2j 1 2 N h are arbitrary. We consider the bilinear form (2.17) b(u h ; u h) := g 1 n u h d + g u 2 hd 1 2 In view of (2.14) we get page 9 of Numer. Math. 69: 83{11 (1994)

9 Boundary element discretization of Poincare{Steklov operators 91 b(u h ; u h) = u h n u h d + (g 1 u h ) n u h d + (g 2 nu h) u h d 1 2 = u h n u h d + (g 1 u h ) ( n u h ' h) d + (g 2 nu h) (u h w h )d 1 2 for all (w h ; ' h ) 2 Mh f N h. If we choose w h = v h, ' h = h (cf. (2.12)) as solutions of the corresponding equations (2.11) then b(u h ; u h) = u h n u hd (2.18) = 1 2 (~v h u h ) ~ h n u h (D~v h ; ~v h) L 2 () + (V ~ h; ~ h ) L 2 () here we used (2.15) and the symmetry of D and V. d = b(u h ; u h) ; (2.19) We remark that ~v h u h = 1 (I + K)~v 2 h V ~ h ; ~ h n u h = 1 D~v 2 h + (I K ) ~ h are the boundary values on of the function u a h(x) := 1 2 G(x; y) ~ h(y)d y ny G(x; y)~v h (y)d y ; x 2 R d n : Now we use (2.4), Theorems 2.2 and 2.3 to estimate for the solution u h of (2.13{14) (D~v h ; ~v h ) L 2 () c 1 k~v h k 2 H c () 2 kvh k2 eh + ( kg 2) 1k 2 eh ( 1) c kg 1 k 2 eh + klg ( 1) 2k 2 ; H () (V ~ h; ~ h) L 2 () c 1 k ~ hk 2 H () c kg 1 k 2 eh ( 1) + klg 2k 2 H () (D~v h ; ~v h ) L 2 () c 1 j~v h j 2 H () c jv hj 2 H ( 2) + jg 1j 2 H ( 1) and if d = 3 (V ~ h; ~ h) L 2 () c 1 k ~ hk 2 H () ckg 2k 2 H ( 2) : ; cjg 1 j 2 H ( 1) ; If d = 2 then we represent ~ h = h +!e, where V e = 1 and! = ( ~ h; 1) L 2 ()= (e; 1) L 2 (). Then ( h ; 1) L 2 () = and (V ~ h; ~ h) L 2 () = V h ; h L 2 () +!2 (e; 1) L 2 () :! 2 can be estimated using Theorem 2.4, since page 91 of Numer. Math. 69: 83{11 (1994)

10 92 G. Schmidt j ~ h d j = j k h k eh (1 So for d = 2 we obtain (V ~ h; ~ h) L 2 () c ~h n u d j = j 1 ( h )d j 1 d c " (h) kg 1k eh ( + klg 1) 2k H () kg 2 k 2 H ( "2 2) (h) kg 1 k 2 eh + ( klg 1) 2k 2 H () Now (2.19) and the derived estimates let us conclude Theorem 2.5. Let u h, u h be the BE solutions of (2.1), (2.5) for the boundary data g 1 ; g 1 2 e H ( 1 ) and g 2 ; g 2 2 e H ( 2 ), resp., obtained via (2.13{14). Then the form b(u h ; u h ) dened in (2.17) is symmetric and bilinear. Furthermore, for any extension lg 2 of g 2 with lg 2 j 1 2 N h b(u h ; u h ) c 1 kg 1 k 2 eh ( 1) + klg 2k 2 H () : : and b(u h ; u h ) c 2 jg1 j 2 H ( 1) + kg 2k 2 H ( 2) " 2 (h) kg 1 k 2 eh ( 1) + klg 2k 2 H () ; with constants independent of g 1, g 2 and h. The second term of the right hand side in the last inequality appears only if d = BE discretization of PS operators Now we are in the position to construct discretizations of the operators S and T using the Galerkin BEM. According to the denition of this method we are given spaces of trial functions M h H ( N ) and N h 2 e H ( D ), fm h = M h \ e H ( N ). Furthermore, on we have nite dimensional spaces ex h e H ( ) to approximate the boundary values u and Y h e H ( ) for the approximation of the normal derivatives 1 u. In order to obtain mappings between the BE approximations of u and 1 u providing the same properties as the FE discretizations of S and T we need some relations between the spaces of trial functions. For the existence of an isomorphism S n : e Xh! Y h with (S h h ; h ) 6= ; 6= h 2 e Xh ; or T n : Y h! e Xh with (T h h ; h ) 6= ; 6= h 2 Y h ; it is necessary that e X h \ Y h = ;. Since the spaces of trial functions are nitedimensional this is equivalent to 1. For any h there exists a bounded projection P h : e H ( )! e Xh onto e Xh, such that the L 2 ( ){adjoint projection P h maps onto Y h, Im P h = Y h, and (3.1) kp h fk H ( ) ckfk H ( ) ; 8 f 2 H ( ) : page 92 of Numer. Math. 69: 83{11 (1994)

11 Boundary element discretization of Poincare{Steklov operators 93 In order to prove the required properties of our BE discretization of the operator T we need also the following technical assumption which is fullled in any reasonable setting. 2. Since \ N = ; it is natural to introduce h := f h : h j D 2 N h, h j 2 Y h g. We require that Y h h possess the extension property: For any h there exists a linear extension operator E h : Y h! h ; E h ' h j = ' h, (3.2) ke h ' hk eh ( [ D) ck' hk H ( ) ; 8 ' h 2 Y h : Note that in general we do not assume that the constants in (3.1) and (3.2) are independent of h. But it will be clear that for some estimates concerning the BE discretizations S h and T h as h! we have to require the uniform boundedness of kp hk eh ( ) and ke hk H ( )!eh ( [ D). Due to the boundary conditions uj D =, n uj N = of (1.1) the BE solutions u h are of the form (2.13) with densities (3.3) ~v h 2 H () such that ~v h j D = ; ~v h j N 2 Mh f ; ~v h j 2 Xh e ; ~ h 2 H () such that ~ hj D 2 N h ; ~ h j N = ; ~ h j 2 Y h : Because of \ N = ; the conditions ~v h j N 2 f Mh, ~v h j 2 e Xh are natural. Hence the solutions u h of (1.1) with the direct Galerkin BEM belong to the linear space V h := ( u h of the form (2.13), the densities ~v h ; ~ h satisfy (3.3) and u h ' h d + D N n u h w h d = ; 8 (w h ; ' h ) 2 f Mh N h BE solutions of (1.1) with Dirichlet data h 2 e Xh we denote by R D;h h, with Neumann data h 2 Y h by R N;h h, i.e. h 2 Xh e R! R D;h h 2 V h with ( (R D;h h ) h )' h d = ; 8 ' h 2 Y h ; R h 2 Y h! R N;h h 2 V h with ( 1 (R N;h h ) h )w h d = ; 8 w h 2 Xh e : (3.4) From Theorem 2.3 we deduce that for all (suciently small if d = 2) h the functions R D;h h and R N;h h are uniquely determined. Now we construct the BE discretization of the operator S. Let h 2 Xh e, then due to Theorem 2.4 with b(r D;h h ; R D;h h ) = [ D ~ h n R D;h h d ~ h (x) = h (x) ; x 2 ; ; x 2 D ; ) : page 93 of Numer. Math. 69: 83{11 (1994)

12 94 G. Schmidt and b(r D;h h ; R D;h h ) c 1 k ~ h k 2 eh ( [ D) ; b(r D;h h ; R D;h h ) c 2 j ~ h j 2 eh ( [ D) "2 (h)k ~ h k 2 eh ( [ D) where in the last inequality the second term of the right hand side appears only for d = 2. But ; k ~ hk eh ( [ D) and j ~ hj eh ( [ D) c 3 k hk eh ( ) c 4 k hk eh ( ) with constants independent of h and h. Hence for all (suciently small if d = 2) h we have (3.5) c 1 k h k 2 eh ( ) b(r D;h h ; R D;h h ) c 2 k h k 2 eh ( ) Using (3.1) we dene (3.6) such that (3.7) b(r D;h h ; R D;h ' h ) = S h h := P h 1 (R D;h h ) h S h ' h d = ( h ; S h ' h ) ; h ; ' h 2 e Xh : Theorem 3.1. For all h (< h if d = 2) the BE discretization S h (3.6) of the operator S has the following properties: (i) S h : e Xh! Y h is bounded, (3.8) ks h h k H ( ) ck hk eh ( ) ; 8 h 2 e Xh : If the projections P h are uniformly bounded then the constant c in (3.8) is independent of h. (ii) c 1 k h k 2 eh ( ) (S h h ; h ) c 2 k h k 2 eh ( ), c 1, c 2 do not depend on h 2 Xh e and h. (iii) (S h h ; ' h ) = ( h ; S h ' h ) ; 8 h ; ' h 2 Xh e. (iv) k(s h S) h k H ( ) c inf ku w hk eh w h 2 em ( + inf k N) n u hk, h h 2 eh h ( [ D) where u solves (1.1) with the Dirichlet condition u = h on. If the projections P h are uniformly bounded, then c is independent of h. Proof. (i) From (3.1) and Theorem 2.4 we have ks h h k H ( ) = kp h 1(R D;h h )k H ( ) c 1 k 1 (R D;h h )k H ( ) ck hk eh ( ) : (ii) and (iii) follow immediately from (3.5), (3.7) and Theorem 2.5. page 94 of Numer. Math. 69: 83{11 (1994)

13 Boundary element discretization of Poincare{Steklov operators 95 (iv) k(s h S) h k H ( ) k(s h P h S) hk H ( ) + k(i P h)s h k H ( ) c k 1 (R D;h h u)k H ( ) + inf k 1 u ' h k H ' h 2Y ( ) h The application of Theorem 2.3 and the estimate inf k 1 u ' h k H ' h 2Y ( ) inf k n u hk h h 2 eh h ( [ D) yield the assertion. ut Next we dene the BE discretization of T. Let h 2 Y h, R N;h h 2 V h exists for all (suciently small if d = 2) h and we set (3.9) T h h := P h R N;h h ; such that from Theorem 2.5 and (3.1) (3.1) (T h h ; ' h ) = b(r N;h h ; R N;h ' h ) ; h ; ' h 2 Y h : But in general T h 6= S h 1, as can be seen from the following. Let h 2 ex h. If we denote the densities dening R D;h h via (2.13) by ~v h, ~ h and the corresponding densities of R N;h (S h h ) by ~v h, ~ h then Hence and (3.11) b(r D;h h ; R N;h (S h h )) = = = R D;h h n R N;h (S h h )d (~v h R D;h h ) ~ h n R N;h (S h h ) ~v h n R N;h (S h h )d + h 1 (R N;h (S h h ))d + = (S h h ; h ) = b(r D;h h ; R D;h h ) d ~ h R D;h h d ~v h ~ h d S h h (R D;h h )d b(r D;h h R N;h (S h h ); R D;h h R N;h (S h h )) = (S h h ; h ) 2(S h h ; h ) + (T h S h h ; S h h ) ((T h S h I) h ; S h h ) h S h h d = 1 (D(~v h ~v 2 h); ~v h ~v h) L 2 () + (V ( ~ h ~ h ); ~ h ~ h ) L 2 () c j~v h ~v h j2 H + k ~ () h ~ h k 2 : H () But using the extension property (3.2) we can prove that T h is spectrally equivalent to S 1 h. page 95 of Numer. Math. 69: 83{11 (1994)

14 96 G. Schmidt Theorem 3.2. Suppose that estimates (3.1) and (3.2) are satised uniformly with respect to h. Then for any h(< h if d = 2) the BE discretization T h (3.9) of the PS operator T has the following properties: (i) T h : Y h! e Xh is bounded, kt h hk eh ( ) ck hk H ( ) ; 8 h 2 Y h : (ii) c 1 k h k 2 H ( ) (T h h ; h ) c 2 k h k 2 H ( ), the constants do not depend on h 2 Y h and h. (iii) (T h h ; ' h ) = ( h ; T h ' h ), h ; ' h 2 Y h. (iv) ((T h S h I) h ; S h h ), 8 h 2 e Xh. (v) k(t h T ) hk eh ( ) c inf w h 2eX h ku w hk eh ( ) + inf k n u ' hk ' h 2N eh h ( + inf ku w D) hk eh w h 2 em ( N) h where u solves (1.1) with the Neumann condition 1 u = h. Proof. (i) kt h hk eh ( ) j(t h h ; ')j = sup '6= k'k H ( ) j(t h h ; Ph = sup ')j '6= kph 'kh ( ) Using (3.2), Theorem 2.5 and (3.1) we obtain! kph 'kh ( ) : k'k H ( ) j(t h h ; Ph')j = jb(r N;h h ; R N;h (Ph'))j c 1 ke h hk eh ( [ ke D) h(ph')k eh ( [ D) From (3.1) we get the assertion. (ii) Theorem 2.5 and (3.2) imply ck h k H ( )kp h 'k H ( ) : (T h h ; h ) = b(r N;h h ; R N;h h ) c 1 ke h h k 2 eh ( [ D) ck hk 2 H ( ) :, If d = 2 then (T h h ; ) c 1 k h k 2 H ( ) "2 (h)ke h h k 2 eh ( [ D) c 1 (1 " 2 (h))k h k 2 H ( ) : Note that the boundedness of P h is not used. (iii) and (iv) follow from Theorem 2.5, (3.1) and (3.11). (v) k(t h T ) hk eh ( ) k(t h P h T ) hk eh ( ) + k(i P h)t hk eh ( ) sup '6= j((t h P h T ) h ; ')) k'k H ( ) + c inf w h 2eX h kw h T hk eh ( ) : page 96 of Numer. Math. 69: 83{11 (1994)

15 Boundary element discretization of Poincare{Steklov operators 97 We have j((t h P h T ) h ; ')j k'k H ( ) c j((t h T ) h ; P h ')j kp h 'k H ( ) and since ((T h T ) h ; P h') = [ D (R N;h h u) E h (P h')d ' h R N;h h d =, 8 ' h 2 N h, and u denotes the solution of (1.1) with D 1 u = h. Therefore, using (3.2) k(t h T ) hk eh ( ) c sup '6= [ D (R N;h h u) E h (P h') d ke h (P h ')k eh ( [ D) c kr N;h h uk H ( [ D) and Theorem 2.3 yields the assertion. ut 4. Remarks In these concluding remarks we give an example of BE functions satisfying (3.1) and (3.2) and consider the BE discretization of PS operators in the case =. Let d = 2 and suppose that u and 1 u shall be approximated by linear boundary elements, i.e. e Xh and Y h consist of piecewise linear functions. It is natural to use ex h = o S1 () ; the set of continuous piecewise linear functions subordinate to a mesh = fx k g n k= on and vanishing at the endpoints x and x n of. There are many dierent possibilities to choose a projection P h onto Xh e bounded in H e ( ) and so to determine Y h = Im Ph. For example, denote by w k, k = 1; : : : ; n 1, the hat functions, i.e. w k 2 o S1 (), w k (x j ) = kj, k; j = 1; : : : ; n 1 ; and set X n1 w (x) = 1 k=1 w k (x) ; x 2 : Then S 1 () = span(w k ; k = ; : : : ; n 1) is the set of periodic linear splines, let P : L 2 ( )! S 1 () page 97 of Numer. Math. 69: 83{11 (1994)

16 98 G. Schmidt be the orthoprojection. By f' k g n1 S k= 1() we denote the biorthogonal base of fw k g n1 k=. We introduce P u := n1 X k=1 (u; ' k ) w k : Clearly P is bounded in e H ( ), Im P = o S1 () and Im P = span(' k ; k = 1; : : : ; n 1) = Y h S 1 (). The projections P are uniformly bounded in e H ( ) for any sequence of quasiuniform meshes, i.e. := max jx k x k1 j c min jx k x k1 j 1kn 1kn for any mesh = fx k g n k= with a constant c independent of. Indeed, for quasiuniform meshes we have and k' k L 2 ( ) kw k L 2 ( ) c kp uk L 2 ( ) kp uk L 2 ( ) + k(u; ' )w k L 2 ( ) ckuk L 2 ( ) : For u 2 e H 1 ( ) the interpolating spline Q u = which implies n1 X k=1 ku Q uk L 2 ( ) c kuk eh 1 ( ) ; u(x k ) w k j(u; ' )j = j(u Q u; ' )j c kuk eh 1 ( ) k' k L 2 ( ) : satises So the uniform boundedness of P and the inverse property of splines yield kp uk eh 1 ( ) kp uk eh 1 ( ) + k(u; ' ) w k H 1 ( ) Hence by interpolation kuk eh 1 ( ) + c kw k H 1 ( ) kuk eh 1 ( ) k' k L 2 ( ) c kuk eh 1 ( ) : kp uk eh ( ) c kuk eh ( ) ; 8 u 2 e H ( ). The extension property (3.2) holds for a wide class of piecewise polynomial functions, as shown in [11] for the case of Sobolev spaces H m ( ), m nonnegative integer, and mentioned in [6] for arbitrary H s ( ). Especially, (3.2) holds for the piecewise linear functions in H ( ). Now we apply the results of Sects. 2 and 3 in order to construct S h and T h if =, i.e. we consider the BE discretizations of the operators which map the boundary value of a harmonic function onto its normal derivative and vice versa. page 98 of Numer. Math. 69: 83{11 (1994)

17 Boundary element discretization of Poincare{Steklov operators 99 Since the Neumann problem for the Laplace equation in a bounded domain is not uniquely solvable these operators are not bijective and some modications are necessary. Let us denote by R() the set of constant functions on. It is well known that ker (I+K) = R() ; Im (I+K) = fv 2 H () : (v; V 1 1) L 2 () = g : (4.1) Then Im V 1 (I + K) = f 2 H () : ( ; 1) L 2 () = g =: H () ; and from (2.5) we obtain Su := n u = V 1 (I + K)u is a bijective mapping from the factor space H ()=R() onto H (). Remark that the two spaces are in duality with respect to the L 2 (){scalar product, (H ()) = H ()=R(). Therefore (2.4) implies that D : H ()=R()! H () is bijective, by D (1) we denote its inverse. From (4.1) we see that I K : H ()! H () ; hence the second equality of (2.5) leads to T n u := u = D (1) (I K ) n u : The relations DK = K D and KV = V K (cf. [5]) show that the above mentioned properties of the PS operators remain valid if T and S are considered as mappings acting between H ()=R() and H (). Under the natural assumption that R() X h where X h is the space of traces of nite elements v h 2 V h H 1 () we can construct as in Sect. 1 the FE discretizations of T and S which are symmetric isomorphisms between X h =R() and its dual. We show that the BE discretizations S h and T h constructed analogously to Sect. 3 preserve these properties. Suppose we are given spaces of trial functions X h H () for the approximation of uj and Y h H () for the approximation of n uj. As mentioned in Sect. 3 we have to assume that dim X h = dim Y h ; R() X h and X h \ Y h = ; : Then there exists a bounded projection onto X h, P h : H ()! X h such that Im P h = Y h, i.e. Y h can be identied with the dual of X h. For given h 2 X h the density ~ h of (2.13) is determined from page 99 of Numer. Math. 69: 83{11 (1994)

18 1 G. Schmidt (V ~ h; ' h ) L 2 () = ((I + K) h ; ' h ) L 2 () ; 8 ' h 2 Y h ; or, equivalently, from the projection equation (4.2) P h V ~ h = P h (I + K) h : Since R() = ker (I + K) X h it suces to restrict to h 2 e Xh := X h =R() = P h H ()=R() : On the other hand it follows that ker Ph (I + K )j Yh of (4.2) belongs to the set 6= ; and the solution ~ h Y h := f h 2 Y h : V h ker P h(i + K )j Yh g : But in general P h V e h 62 R() for some e h 2 ker P h (I + K )j Yh, such that ~ h 62 e Xh = fh 2 Y h : ( h ; 1) L 2 () = g = P h (H ()) =: e Yh : Therefore the mapping h 2 Xh e! ~ 2 Y h does not provide the properties of PS operators, it can happen that this mapping is not bijective, in general. But dening as in Sect. 3 we see that S h h = Ph(D h + (I + K ) ~ h) = (PhD + Ph(I + K )Ph(P h V Ph) 1 P h (I + K)) h S h e Xh = e Yh = e Xh : Applying the ideas of the proof of Theorem 3.1 we conclude that for S h : e Xh! ey h the assertions of this theorem are valid. For h 2 e Yh the construction analogous to Sect. 3 leads to T h h = (P h V + P h (I K)P h (P hdp h ) (1) P h(i K )) h ; which can be considered as mapping from e Yh to e Xh and the assertions of Theorem 3.2 remain true. Here (P h DP h) (1) denotes the inverse of P h DP h : e Xh! ey h. Finally we mention the matrix representation for S h and T h. Let fw k g n1 X k= h, f' k g n1 Y k= h be biorthogonal bases, (w j ; ' k ) L 2 () = jk, then where S h = D h + (I h + K h)v 1 h (I h + K h ) ; T h = V h + (I h K h )D (1) h (I h K h) ; (V h ) j;k = (V ' j ; ' k ) L 2 () ; (K h ) j;k = (Kw j ; ' k ) L 2 () = (Kh ) k;j ; (D h ) j;k = (Dw j ; w k ) L 2 () ; (I h ) j;k = (w j ; ' k ) L 2 () = jk : page 1 of Numer. Math. 69: 83{11 (1994)

19 Boundary element discretization of Poincare{Steklov operators 11 Especially we note the case X h = Y h which is often used. For a given basis fw k g n1 k= X h the biorthogonal basis is given by f' k g n1 k= = M 1 h fw kg n1 k= ; where M h denotes the mass matrix, (M h ) j;k = (w j ; w k ) L 2 (). Hence V h = M h 1 V hm h 1 with (V h ) j;k = (V w j ; w k ) L 2 () ; K h = K h M h 1 with (K h ) j;k = (Kw j ; w k ) L 2 () ; which implies the formulas S h = D h + (M h + M h 1 K h M h)v h 1 (M h + M h K h M h 1 ) ; T h = M h 1 Vh + (M h M h K h M h 1 )D(1) h (M h M h 1 K h M h) M 1 Acknowledgement. Thanks are due to Dr. H. Strese whose numerical experiments with a number of DD algorithms based on the BEM showed that naive BE discretizations of PS operators lead to nonstable iterative procedures and so stimulated the theoretical investigations. The author thanks also Prof. W.L. Wendland and Dr. B.N. Khoromskij for stimulating discussions and remarks. This work has been partially supported by the German Research Foundation under Grant Ko 182/2{1. h : References 1. Agoshkow, V.I. (1988): Poincare{Steklov's operators and domain decomposition methods in nite dimensional spaces. In: Glowinski, R. et. al., eds., Domain Decomposition Methods for Partial Dierential Equations I, 73{112, SIAM, Philadelphia 2. Costabel, M. (1988): Boundary integral operators on Lipschitz domains: elementary results. SIAM J. Math. Anal. 19, 613{ Costabel, M., Stephan, E.P. (1985): Boundary integral equations for mixed boundary value problems in polygonal domains and Galerkin approximations. In: Fiszdon, W. et. al., eds., Mathematical Models and Methods in Mechanics (1981), PWN, Warsaw, 175{ Costabel, M., Stephan, E.P. (1985a): The method of Mellin transformation for boundary integral equations on curves with corners. In: Gerasoulis, A. et. al., eds., Numerical Solutions of Singular Integral Equations, IMACS, 95{12 5. Costabel, M., Stephan, E.P. (1985b): A direct boundary integral equation method for transmission problems. J. Math. Anal. Appl. 16, 367{ Hebeker, F.K. (1991): On the parallel Schwarz algorithm for symmetric strongly elliptic integral equations. In: Glowinski, R. et. al., eds., Domain Decomposition Methods for Partial Dierential Equations IV, SIAM, Philadelphia, 382{ Lebedev, V.I., Agoshkow, V.I. (1983): Operatory Poincare{Steklova i ich primenenija v analize. OVM, Moskva 8. von Petersdor, T. (1989): Randwertprobleme der Elastizitatstheorie fur Polyeder{Singularitaten und Approximation mit Randelementmethoden. Dissertation, TH Darmstadt 9. Quarteroni, A., Valli, A. (1991): Theory and application of Steklov{Poincare operators for boundary value problems. In: Spigler R., ed., Applied and Industrial Mathematics, pp. 179{23, Kluwer, Dordrecht 1. Symm, G.T. (1967): Numerical mapping of exterior domains. Numer. Math. 1, 437{ Widlund, O.B. (1987): An extension theorem for nite element spaces with three applications. In: Hackbusch, W. et. al., eds., Numerical Techniques in Continuum Mechanics, pp. 11{122, Vieweg, Braunschweig This article was processed by the author using the LaT E X style le cljour1 from Springer-Verlag. page 11 of Numer. Math. 69: 83{11 (1994)

Technische Universität Graz

Technische Universität Graz Technische Universität Graz Stability of the Laplace single layer boundary integral operator in Sobolev spaces O. Steinbach Berichte aus dem Institut für Numerische Mathematik Bericht 2016/2 Technische

More information

Preconditioned space-time boundary element methods for the heat equation

Preconditioned space-time boundary element methods for the heat equation W I S S E N T E C H N I K L E I D E N S C H A F T Preconditioned space-time boundary element methods for the heat equation S. Dohr and O. Steinbach Institut für Numerische Mathematik Space-Time Methods

More information

WALTER LITTMAN BO LIU kind of problem is related to the deformation of a membrane with strings attached to several sides of the polygon. In other word

WALTER LITTMAN BO LIU kind of problem is related to the deformation of a membrane with strings attached to several sides of the polygon. In other word THE REGULARITY AND SINGULARITY OF SOLUTIONS OF CERTAIN ELLIPTIC PROBLEMS ON POLYGONAL DOMAINS Walter Littman Bo Liu School of Math., University of Minnesota, Minneapolis, MN55455 Abstract. The regularity

More information

It is known that Morley element is not C 0 element and it is divergent for Poisson equation (see [6]). When Morley element is applied to solve problem

It is known that Morley element is not C 0 element and it is divergent for Poisson equation (see [6]). When Morley element is applied to solve problem Modied Morley Element Method for a ourth Order Elliptic Singular Perturbation Problem Λ Wang Ming LMAM, School of Mathematical Science, Peking University Jinchao u School of Mathematical Science, Peking

More information

A Finite Element Method for an Ill-Posed Problem. Martin-Luther-Universitat, Fachbereich Mathematik/Informatik,Postfach 8, D Halle, Abstract

A Finite Element Method for an Ill-Posed Problem. Martin-Luther-Universitat, Fachbereich Mathematik/Informatik,Postfach 8, D Halle, Abstract A Finite Element Method for an Ill-Posed Problem W. Lucht Martin-Luther-Universitat, Fachbereich Mathematik/Informatik,Postfach 8, D-699 Halle, Germany Abstract For an ill-posed problem which has its origin

More information

Domain Decomposition Algorithms for an Indefinite Hypersingular Integral Equation in Three Dimensions

Domain Decomposition Algorithms for an Indefinite Hypersingular Integral Equation in Three Dimensions Domain Decomposition Algorithms for an Indefinite Hypersingular Integral Equation in Three Dimensions Ernst P. Stephan 1, Matthias Maischak 2, and Thanh Tran 3 1 Institut für Angewandte Mathematik, Leibniz

More information

Variational Formulations

Variational Formulations Chapter 2 Variational Formulations In this chapter we will derive a variational (or weak) formulation of the elliptic boundary value problem (1.4). We will discuss all fundamental theoretical results that

More information

Technische Universität Graz

Technische Universität Graz Technische Universität Graz A note on the stable coupling of finite and boundary elements O. Steinbach Berichte aus dem Institut für Numerische Mathematik Bericht 2009/4 Technische Universität Graz A

More information

/00 $ $.25 per page

/00 $ $.25 per page Contemporary Mathematics Volume 00, 0000 Domain Decomposition For Linear And Nonlinear Elliptic Problems Via Function Or Space Decomposition UE-CHENG TAI Abstract. In this article, we use a function decomposition

More information

A priori error analysis of the BEM with graded meshes for the electric eld integral equation on polyhedral surfaces

A priori error analysis of the BEM with graded meshes for the electric eld integral equation on polyhedral surfaces A priori error analysis of the BEM with graded meshes for the electric eld integral equation on polyhedral surfaces A. Bespalov S. Nicaise Abstract The Galerkin boundary element discretisations of the

More information

The Mortar Wavelet Method Silvia Bertoluzza Valerie Perrier y October 29, 1999 Abstract This paper deals with the construction of wavelet approximatio

The Mortar Wavelet Method Silvia Bertoluzza Valerie Perrier y October 29, 1999 Abstract This paper deals with the construction of wavelet approximatio The Mortar Wavelet Method Silvia Bertoluzza Valerie Perrier y October 9, 1999 Abstract This paper deals with the construction of wavelet approximation spaces, in the framework of the Mortar method. We

More information

Domain decomposition methods via boundary integral equations

Domain decomposition methods via boundary integral equations Domain decomposition methods via boundary integral equations G. C. Hsiao a O. Steinbach b W. L. Wendland b a Department of Mathematical Sciences, University of Delaware, Newark, Delaware 19716, USA. E

More information

LECTURE # 0 BASIC NOTATIONS AND CONCEPTS IN THE THEORY OF PARTIAL DIFFERENTIAL EQUATIONS (PDES)

LECTURE # 0 BASIC NOTATIONS AND CONCEPTS IN THE THEORY OF PARTIAL DIFFERENTIAL EQUATIONS (PDES) LECTURE # 0 BASIC NOTATIONS AND CONCEPTS IN THE THEORY OF PARTIAL DIFFERENTIAL EQUATIONS (PDES) RAYTCHO LAZAROV 1 Notations and Basic Functional Spaces Scalar function in R d, d 1 will be denoted by u,

More information

ON TRIVIAL GRADIENT YOUNG MEASURES BAISHENG YAN Abstract. We give a condition on a closed set K of real nm matrices which ensures that any W 1 p -grad

ON TRIVIAL GRADIENT YOUNG MEASURES BAISHENG YAN Abstract. We give a condition on a closed set K of real nm matrices which ensures that any W 1 p -grad ON TRIVIAL GRAIENT YOUNG MEASURES BAISHENG YAN Abstract. We give a condition on a closed set K of real nm matrices which ensures that any W 1 p -gradient Young measure supported on K must be trivial the

More information

QUASI-UNIFORMLY POSITIVE OPERATORS IN KREIN SPACE. Denitizable operators in Krein spaces have spectral properties similar to those

QUASI-UNIFORMLY POSITIVE OPERATORS IN KREIN SPACE. Denitizable operators in Krein spaces have spectral properties similar to those QUASI-UNIFORMLY POSITIVE OPERATORS IN KREIN SPACE BRANKO CURGUS and BRANKO NAJMAN Denitizable operators in Krein spaces have spectral properties similar to those of selfadjoint operators in Hilbert spaces.

More information

Research Collection. On the definition of nonlinear stability for numerical methods. Working Paper. ETH Library. Author(s): Pirovino, Magnus

Research Collection. On the definition of nonlinear stability for numerical methods. Working Paper. ETH Library. Author(s): Pirovino, Magnus Research Collection Working Paper On the definition of nonlinear stability for numerical methods Author(s): Pirovino, Magnus Publication Date: 1991 Permanent Link: https://doi.org/10.3929/ethz-a-000622087

More information

the sum of two projections. Finally, in Section 5, we apply the theory of Section 4 to the case of nite element spaces. 2. Additive Algorithms for Par

the sum of two projections. Finally, in Section 5, we apply the theory of Section 4 to the case of nite element spaces. 2. Additive Algorithms for Par ON THE SPECTRA OF SUMS OF ORTHOGONAL PROJECTIONS WITH APPLICATIONS TO PARALLEL COMPUTING PETTER E. BJRSTAD y AND JAN MANDEL z Abstract. Many parallel iterative algorithms for solving symmetric, positive

More information

Rearrangements and polar factorisation of countably degenerate functions G.R. Burton, School of Mathematical Sciences, University of Bath, Claverton D

Rearrangements and polar factorisation of countably degenerate functions G.R. Burton, School of Mathematical Sciences, University of Bath, Claverton D Rearrangements and polar factorisation of countably degenerate functions G.R. Burton, School of Mathematical Sciences, University of Bath, Claverton Down, Bath BA2 7AY, U.K. R.J. Douglas, Isaac Newton

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

A MULTIGRID ALGORITHM FOR. Richard E. Ewing and Jian Shen. Institute for Scientic Computation. Texas A&M University. College Station, Texas SUMMARY

A MULTIGRID ALGORITHM FOR. Richard E. Ewing and Jian Shen. Institute for Scientic Computation. Texas A&M University. College Station, Texas SUMMARY A MULTIGRID ALGORITHM FOR THE CELL-CENTERED FINITE DIFFERENCE SCHEME Richard E. Ewing and Jian Shen Institute for Scientic Computation Texas A&M University College Station, Texas SUMMARY In this article,

More information

MARY ANN HORN be decoupled into three wave equations. Thus, we would hope that results, analogous to those available for the wave equation, would hold

MARY ANN HORN be decoupled into three wave equations. Thus, we would hope that results, analogous to those available for the wave equation, would hold Journal of Mathematical Systems, Estimation, and Control Vol. 8, No. 2, 1998, pp. 1{11 c 1998 Birkhauser-Boston Sharp Trace Regularity for the Solutions of the Equations of Dynamic Elasticity Mary Ann

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

5.3. Compactness results for Dirichlet spaces Compactness results for manifolds 39 References 39 sec:intro 1. Introduction We mean by a spectr

5.3. Compactness results for Dirichlet spaces Compactness results for manifolds 39 References 39 sec:intro 1. Introduction We mean by a spectr CONVERGENCE OF SPECTRAL STRUCTURES: A FUNCTIONAL ANALYTIC THEORY AND ITS APPLICATIONS TO SPECTRAL GEOMETRY KAZUHIRO KUWAE AND TAKASHI SHIOYA Abstract. The purpose of this paper is to present a functional

More information

BOUNDARY CONTROLLABILITY IN PROBLEMS JOHN E. LAGNESE. linear hyperbolic systems having piecewise constant coecients in

BOUNDARY CONTROLLABILITY IN PROBLEMS JOHN E. LAGNESE. linear hyperbolic systems having piecewise constant coecients in ESAIM: Control, Optimisation and Calculus of Variations URL: http://www.emath.fr/cocv/,, BOUNDARY CONTROLLABILITY IN PROBLEMS OF TRANSMISSION FOR A CLASS OF SECOND ORDER HYPERBOLIC SYSTEMS JOHN E. LAGNESE

More information

ANDREA TOSELLI. Abstract. Two-level overlapping Schwarz methods are considered for nite element problems

ANDREA TOSELLI. Abstract. Two-level overlapping Schwarz methods are considered for nite element problems OVERLAPPING SCHWARZ METHODS FOR MAXWELL'S EQUATIONS IN THREE DIMENSIONS ANDREA TOSELLI Abstract. Two-level overlapping Schwarz methods are considered for nite element problems of 3D Maxwell's equations.

More information

Chapter 1 Foundations of Elliptic Boundary Value Problems 1.1 Euler equations of variational problems

Chapter 1 Foundations of Elliptic Boundary Value Problems 1.1 Euler equations of variational problems Chapter 1 Foundations of Elliptic Boundary Value Problems 1.1 Euler equations of variational problems Elliptic boundary value problems often occur as the Euler equations of variational problems the latter

More information

Title: Localized self-adjointness of Schrödinger-type operators on Riemannian manifolds. Proposed running head: Schrödinger-type operators on

Title: Localized self-adjointness of Schrödinger-type operators on Riemannian manifolds. Proposed running head: Schrödinger-type operators on Title: Localized self-adjointness of Schrödinger-type operators on Riemannian manifolds. Proposed running head: Schrödinger-type operators on manifolds. Author: Ognjen Milatovic Department Address: Department

More information

arxiv: v1 [math.na] 27 Jan 2016

arxiv: v1 [math.na] 27 Jan 2016 Virtual Element Method for fourth order problems: L 2 estimates Claudia Chinosi a, L. Donatella Marini b arxiv:1601.07484v1 [math.na] 27 Jan 2016 a Dipartimento di Scienze e Innovazione Tecnologica, Università

More information

Scientific Computing WS 2018/2019. Lecture 15. Jürgen Fuhrmann Lecture 15 Slide 1

Scientific Computing WS 2018/2019. Lecture 15. Jürgen Fuhrmann Lecture 15 Slide 1 Scientific Computing WS 2018/2019 Lecture 15 Jürgen Fuhrmann juergen.fuhrmann@wias-berlin.de Lecture 15 Slide 1 Lecture 15 Slide 2 Problems with strong formulation Writing the PDE with divergence and gradient

More information

Inequalities of Babuška-Aziz and Friedrichs-Velte for differential forms

Inequalities of Babuška-Aziz and Friedrichs-Velte for differential forms Inequalities of Babuška-Aziz and Friedrichs-Velte for differential forms Martin Costabel Abstract. For sufficiently smooth bounded plane domains, the equivalence between the inequalities of Babuška Aziz

More information

Space time finite and boundary element methods

Space time finite and boundary element methods Space time finite and boundary element methods Olaf Steinbach Institut für Numerische Mathematik, TU Graz http://www.numerik.math.tu-graz.ac.at based on joint work with M. Neumüller, H. Yang, M. Fleischhacker,

More information

XIAO-CHUAN CAI AND MAKSYMILIAN DRYJA. strongly elliptic equations discretized by the nite element methods.

XIAO-CHUAN CAI AND MAKSYMILIAN DRYJA. strongly elliptic equations discretized by the nite element methods. Contemporary Mathematics Volume 00, 0000 Domain Decomposition Methods for Monotone Nonlinear Elliptic Problems XIAO-CHUAN CAI AND MAKSYMILIAN DRYJA Abstract. In this paper, we study several overlapping

More information

A Finite Element Method Using Singular Functions for Poisson Equations: Mixed Boundary Conditions

A Finite Element Method Using Singular Functions for Poisson Equations: Mixed Boundary Conditions A Finite Element Method Using Singular Functions for Poisson Equations: Mixed Boundary Conditions Zhiqiang Cai Seokchan Kim Sangdong Kim Sooryun Kong Abstract In [7], we proposed a new finite element method

More information

Abstract. 1. Introduction

Abstract. 1. Introduction Journal of Computational Mathematics Vol.28, No.2, 2010, 273 288. http://www.global-sci.org/jcm doi:10.4208/jcm.2009.10-m2870 UNIFORM SUPERCONVERGENCE OF GALERKIN METHODS FOR SINGULARLY PERTURBED PROBLEMS

More information

Neumann problem: p u = juj N() : p 2 u; in ; u n = ; on : Robin problem: R() : Steklov problem: S() : p u = juj p 2 u; in ; p 2 u jruj n + jujp 2 u =

Neumann problem: p u = juj N() : p 2 u; in ; u n = ; on : Robin problem: R() : Steklov problem: S() : p u = juj p 2 u; in ; p 2 u jruj n + jujp 2 u = EIGENVALUE PROBLEMS FOR THE p-laplacian AN L^E DEPARTMENT OF MATHEMATICS, UNIVERSITY OF UTAH 55 SOUTH 4 EAST SALT LAKE CITY, UT 842, USA Abstract. We study nonlinear eigenvalue problems for the p-laplace

More information

Chapter 5 A priori error estimates for nonconforming finite element approximations 5.1 Strang s first lemma

Chapter 5 A priori error estimates for nonconforming finite element approximations 5.1 Strang s first lemma Chapter 5 A priori error estimates for nonconforming finite element approximations 51 Strang s first lemma We consider the variational equation (51 a(u, v = l(v, v V H 1 (Ω, and assume that the conditions

More information

290 J.M. Carnicer, J.M. Pe~na basis (u 1 ; : : : ; u n ) consisting of minimally supported elements, yet also has a basis (v 1 ; : : : ; v n ) which f

290 J.M. Carnicer, J.M. Pe~na basis (u 1 ; : : : ; u n ) consisting of minimally supported elements, yet also has a basis (v 1 ; : : : ; v n ) which f Numer. Math. 67: 289{301 (1994) Numerische Mathematik c Springer-Verlag 1994 Electronic Edition Least supported bases and local linear independence J.M. Carnicer, J.M. Pe~na? Departamento de Matematica

More information

Scientific Computing WS 2017/2018. Lecture 18. Jürgen Fuhrmann Lecture 18 Slide 1

Scientific Computing WS 2017/2018. Lecture 18. Jürgen Fuhrmann Lecture 18 Slide 1 Scientific Computing WS 2017/2018 Lecture 18 Jürgen Fuhrmann juergen.fuhrmann@wias-berlin.de Lecture 18 Slide 1 Lecture 18 Slide 2 Weak formulation of homogeneous Dirichlet problem Search u H0 1 (Ω) (here,

More information

Introduction to the Boundary Element Method

Introduction to the Boundary Element Method Introduction to the Boundary Element Method Salim Meddahi University of Oviedo, Spain University of Trento, Trento April 27 - May 15, 2015 1 Syllabus The Laplace problem Potential theory: the classical

More information

on! 0, 1 and 2 In the Zienkiewicz-Zhu SPR p 1 and p 2 are obtained by solving the locally discrete least-squares p

on! 0, 1 and 2 In the Zienkiewicz-Zhu SPR p 1 and p 2 are obtained by solving the locally discrete least-squares p Analysis of a Class of Superconvergence Patch Recovery Techniques for Linear and Bilinear Finite Elements Bo Li Zhimin Zhang y Abstract Mathematical proofs are presented for the derivative superconvergence

More information

On the relationship of local projection stabilization to other stabilized methods for one-dimensional advection-diffusion equations

On the relationship of local projection stabilization to other stabilized methods for one-dimensional advection-diffusion equations On the relationship of local projection stabilization to other stabilized methods for one-dimensional advection-diffusion equations Lutz Tobiska Institut für Analysis und Numerik Otto-von-Guericke-Universität

More information

MIXED FINITE ELEMENT METHODS FOR PROBLEMS WITH ROBIN BOUNDARY CONDITIONS

MIXED FINITE ELEMENT METHODS FOR PROBLEMS WITH ROBIN BOUNDARY CONDITIONS MIXED FINITE ELEMENT METHODS FOR PROBLEMS WITH ROBIN BOUNDARY CONDITIONS JUHO KÖNNÖ, DOMINIK SCHÖTZAU, AND ROLF STENBERG Abstract. We derive new a-priori and a-posteriori error estimates for mixed nite

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

Numerische Mathematik

Numerische Mathematik Numer. Math. (997) 76: 479 488 Numerische Mathematik c Springer-Verlag 997 Electronic Edition Exponential decay of C cubic splines vanishing at two symmetric points in each knot interval Sang Dong Kim,,

More information

The mortar element method for quasilinear elliptic boundary value problems

The mortar element method for quasilinear elliptic boundary value problems The mortar element method for quasilinear elliptic boundary value problems Leszek Marcinkowski 1 Abstract We consider a discretization of quasilinear elliptic boundary value problems by the mortar version

More information

Contents Introduction 3 Johnson's paradigm 4 3 A posteriori error analysis for steady hyperbolic PDEs 7 3. Symmetric hyperbolic systems : : : : : : :

Contents Introduction 3 Johnson's paradigm 4 3 A posteriori error analysis for steady hyperbolic PDEs 7 3. Symmetric hyperbolic systems : : : : : : : Report no. 96/09 Finite element methods for hyperbolic problems: a posteriori error analysis and adaptivity a Endre Suli Paul Houston This paper presents an overview of recent developments in the area

More information

Convergence of A Galerkin Method for 2-D Discontinuous Euler Flows Jian-Guo Liu 1 Institute for Physical Science andtechnology and Department of Mathe

Convergence of A Galerkin Method for 2-D Discontinuous Euler Flows Jian-Guo Liu 1 Institute for Physical Science andtechnology and Department of Mathe Convergence of A Galerkin Method for 2-D Discontinuous Euler Flows Jian-Guo Liu 1 Institute for Physical Science andtechnology and Department of Mathematics University of Maryland College Park, MD 2742

More information

Lecture Note III: Least-Squares Method

Lecture Note III: Least-Squares Method Lecture Note III: Least-Squares Method Zhiqiang Cai October 4, 004 In this chapter, we shall present least-squares methods for second-order scalar partial differential equations, elastic equations of solids,

More information

Math 413/513 Chapter 6 (from Friedberg, Insel, & Spence)

Math 413/513 Chapter 6 (from Friedberg, Insel, & Spence) Math 413/513 Chapter 6 (from Friedberg, Insel, & Spence) David Glickenstein December 7, 2015 1 Inner product spaces In this chapter, we will only consider the elds R and C. De nition 1 Let V be a vector

More information

Numerical Integration for Multivariable. October Abstract. We consider the numerical integration of functions with point singularities over

Numerical Integration for Multivariable. October Abstract. We consider the numerical integration of functions with point singularities over Numerical Integration for Multivariable Functions with Point Singularities Yaun Yang and Kendall E. Atkinson y October 199 Abstract We consider the numerical integration of functions with point singularities

More information

Space-time sparse discretization of linear parabolic equations

Space-time sparse discretization of linear parabolic equations Space-time sparse discretization of linear parabolic equations Roman Andreev August 2, 200 Seminar for Applied Mathematics, ETH Zürich, Switzerland Support by SNF Grant No. PDFMP2-27034/ Part of PhD thesis

More information

Pointwise convergence rate for nonlinear conservation. Eitan Tadmor and Tao Tang

Pointwise convergence rate for nonlinear conservation. Eitan Tadmor and Tao Tang Pointwise convergence rate for nonlinear conservation laws Eitan Tadmor and Tao Tang Abstract. We introduce a new method to obtain pointwise error estimates for vanishing viscosity and nite dierence approximations

More information

Ole Christensen 3. October 20, Abstract. We point out some connections between the existing theories for

Ole Christensen 3. October 20, Abstract. We point out some connections between the existing theories for Frames and pseudo-inverses. Ole Christensen 3 October 20, 1994 Abstract We point out some connections between the existing theories for frames and pseudo-inverses. In particular, using the pseudo-inverse

More information

Max-Planck-Institut fur Mathematik in den Naturwissenschaften Leipzig H 2 -matrix approximation of integral operators by interpolation by Wolfgang Hackbusch and Steen Borm Preprint no.: 04 200 H 2 -Matrix

More information

1. Introduction. We consider the model problem that seeks an unknown function u = u(x) satisfying

1. Introduction. We consider the model problem that seeks an unknown function u = u(x) satisfying A SIMPLE FINITE ELEMENT METHOD FOR LINEAR HYPERBOLIC PROBLEMS LIN MU AND XIU YE Abstract. In this paper, we introduce a simple finite element method for solving first order hyperbolic equations with easy

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

Traces, extensions and co-normal derivatives for elliptic systems on Lipschitz domains

Traces, 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 information

Stanford Mathematics Department Math 205A Lecture Supplement #4 Borel Regular & Radon Measures

Stanford Mathematics Department Math 205A Lecture Supplement #4 Borel Regular & Radon Measures 2 1 Borel Regular Measures We now state and prove an important regularity property of Borel regular outer measures: Stanford Mathematics Department Math 205A Lecture Supplement #4 Borel Regular & Radon

More information

A Domain Decomposition Method for Quasilinear Elliptic PDEs Using Mortar Finite Elements

A Domain Decomposition Method for Quasilinear Elliptic PDEs Using Mortar Finite Elements W I S S E N T E C H N I K L E I D E N S C H A F T A Domain Decomposition Method for Quasilinear Elliptic PDEs Using Mortar Finite Elements Matthias Gsell and Olaf Steinbach Institute of Computational Mathematics

More information

2 Ben Schweizer As Hocking describes in [5], a discussion focused on the following question: Can one assume that the dynamic contact angle is a consta

2 Ben Schweizer As Hocking describes in [5], a discussion focused on the following question: Can one assume that the dynamic contact angle is a consta A well-posed model for dynamic contact angles Ben Schweizer February 27, 1998 Abstract We consider uid systems with a free boundary and with a point of contact of the free boundary with a solid wall. We

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

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

Matrix square root and interpolation spaces

Matrix square root and interpolation spaces Matrix square root and interpolation spaces Mario Arioli and Daniel Loghin m.arioli@rl.ac.uk STFC-Rutherford Appleton Laboratory, and University of Birmingham Sparse Days,CERFACS, Toulouse, 2008 p.1/30

More information

Finite Elements. Colin Cotter. February 22, Colin Cotter FEM

Finite Elements. Colin Cotter. February 22, Colin Cotter FEM Finite Elements February 22, 2019 In the previous sections, we introduced the concept of finite element spaces, which contain certain functions defined on a domain. Finite element spaces are examples of

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

Fact Sheet Functional Analysis

Fact Sheet Functional Analysis Fact Sheet Functional Analysis Literature: Hackbusch, W.: Theorie und Numerik elliptischer Differentialgleichungen. Teubner, 986. Knabner, P., Angermann, L.: Numerik partieller Differentialgleichungen.

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

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

Boundary Integral Equations on the Sphere with Radial Basis Functions: Error Analysis

Boundary Integral Equations on the Sphere with Radial Basis Functions: Error Analysis Boundary Integral Equations on the Sphere with Radial Basis Functions: Error Analysis T. Tran Q. T. Le Gia I. H. Sloan E. P. Stephan Abstract Radial basis functions are used to define approximate solutions

More information

with Applications to Elasticity and Compressible Flow Daoqi Yang March 20, 1997 Abstract

with Applications to Elasticity and Compressible Flow Daoqi Yang March 20, 1997 Abstract Stabilized Schemes for Mixed Finite Element Methods with Applications to Elasticity and Compressible Flow Problems Daoqi Yang March 20, 1997 Abstract Stabilized iterative schemes for mixed nite element

More information

A Priori Error Analysis for Space-Time Finite Element Discretization of Parabolic Optimal Control Problems

A Priori Error Analysis for Space-Time Finite Element Discretization of Parabolic Optimal Control Problems A Priori Error Analysis for Space-Time Finite Element Discretization of Parabolic Optimal Control Problems D. Meidner and B. Vexler Abstract In this article we discuss a priori error estimates for Galerkin

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

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

1 Dieses Skriptum ist fur eine summer{school entstanden, die im August 1993 an der Universitat Jyvaskyla, Finnland, stattgefunden hat. Es wird in die

1 Dieses Skriptum ist fur eine summer{school entstanden, die im August 1993 an der Universitat Jyvaskyla, Finnland, stattgefunden hat. Es wird in die Linear Elliptic Boundary Value Problems of Second Order Karl{Josef Witsch Fachbereich 6 Mathematik und Informatik der Universitat GH Essen 45117 Essen August 1993 1 Dieses Skriptum ist fur eine summer{school

More information

if <v;w>=0. The length of a vector v is kvk, its distance from 0. If kvk =1,then v is said to be a unit vector. When V is a real vector space, then on

if <v;w>=0. The length of a vector v is kvk, its distance from 0. If kvk =1,then v is said to be a unit vector. When V is a real vector space, then on Function Spaces x1. Inner products and norms. From linear algebra, we recall that an inner product for a complex vector space V is a function < ; >: VV!C that satises the following properties. I1. Positivity:

More information

CONDITION-DEPENDENT HILBERT SPACES FOR STEEPEST DESCENT AND APPLICATION TO THE TRICOMI EQUATION Jason W. Montgomery

CONDITION-DEPENDENT HILBERT SPACES FOR STEEPEST DESCENT AND APPLICATION TO THE TRICOMI EQUATION Jason W. Montgomery CONDITION-DEPENDENT HILBERT SPACES FOR STEEPEST DESCENT AND APPLICATION TO THE TRICOMI EQUATION Jason W. Montgomery Dissertation Prepared for the Degree of DOCTOR OF PHILOSOPHY UNIVERSITY OF NORTH TEXAS

More information

AN EXISTENCE AND UNIQUENESS RESULT 21

AN EXISTENCE AND UNIQUENESS RESULT 21 AN EXISTENCE AND UNIQUENESS RESULT 21 research was supported by the National Science Foundation under Grant No. DMS 9206244. Bibliography [1] Agmon, S., Douglis, A., and Nirenberg, L., Estimates near the

More information

Numerical Solutions to Partial Differential Equations

Numerical Solutions to Partial Differential Equations Numerical Solutions to Partial Differential Equations Zhiping Li LMAM and School of Mathematical Sciences Peking University Variational Problems of the Dirichlet BVP of the Poisson Equation 1 For the homogeneous

More information

Technische Universität Graz

Technische Universität Graz Technische Universität Graz A non-symmetric coupling of the finite volume method and the boundary element method C. Erath, G. Of, F. J. Sayas Berichte aus dem Institut für Numerische Mathematik Bericht

More information

RIESZ BASES AND UNCONDITIONAL BASES

RIESZ BASES AND UNCONDITIONAL BASES In this paper we give a brief introduction to adjoint operators on Hilbert spaces and a characterization of the dual space of a Hilbert space. We then introduce the notion of a Riesz basis and give some

More information

PARTITION OF UNITY FOR THE STOKES PROBLEM ON NONMATCHING GRIDS

PARTITION OF UNITY FOR THE STOKES PROBLEM ON NONMATCHING GRIDS PARTITION OF UNITY FOR THE STOES PROBLEM ON NONMATCHING GRIDS CONSTANTIN BACUTA AND JINCHAO XU Abstract. We consider the Stokes Problem on a plane polygonal domain Ω R 2. We propose a finite element method

More information

t x 0.25

t x 0.25 Journal of ELECTRICAL ENGINEERING, VOL. 52, NO. /s, 2, 48{52 COMPARISON OF BROYDEN AND NEWTON METHODS FOR SOLVING NONLINEAR PARABOLIC EQUATIONS Ivan Cimrak If the time discretization of a nonlinear parabolic

More information

A PROPERTY OF SOBOLEV SPACES ON COMPLETE RIEMANNIAN MANIFOLDS

A PROPERTY OF SOBOLEV SPACES ON COMPLETE RIEMANNIAN MANIFOLDS Electronic Journal of Differential Equations, Vol. 2005(2005), No.??, pp. 1 10. ISSN: 1072-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu ftp ejde.math.txstate.edu (login: ftp) A PROPERTY

More information

Splines which are piecewise solutions of polyharmonic equation

Splines which are piecewise solutions of polyharmonic equation Splines which are piecewise solutions of polyharmonic equation Ognyan Kounchev March 25, 2006 Abstract This paper appeared in Proceedings of the Conference Curves and Surfaces, Chamonix, 1993 1 Introduction

More information

PREPRINT 2010:25. Fictitious domain finite element methods using cut elements: II. A stabilized Nitsche method ERIK BURMAN PETER HANSBO

PREPRINT 2010:25. Fictitious domain finite element methods using cut elements: II. A stabilized Nitsche method ERIK BURMAN PETER HANSBO PREPRINT 2010:25 Fictitious domain finite element methods using cut elements: II. A stabilized Nitsche method ERIK BURMAN PETER HANSBO Department of Mathematical Sciences Division of Mathematics CHALMERS

More information

The Single Layer Heat Potential and. Galerkin Boundary Element Methods for. Patrick James Noon

The Single Layer Heat Potential and. Galerkin Boundary Element Methods for. Patrick James Noon The Single Layer Heat Potential and Galerkin Boundary Element Methods for the Heat Equation by Patrick James Noon Dissertation submitted to the Faculty of the Graduate School of The University of Maryland

More information

THE FORM SUM AND THE FRIEDRICHS EXTENSION OF SCHRÖDINGER-TYPE OPERATORS ON RIEMANNIAN MANIFOLDS

THE FORM SUM AND THE FRIEDRICHS EXTENSION OF SCHRÖDINGER-TYPE OPERATORS ON RIEMANNIAN MANIFOLDS THE FORM SUM AND THE FRIEDRICHS EXTENSION OF SCHRÖDINGER-TYPE OPERATORS ON RIEMANNIAN MANIFOLDS OGNJEN MILATOVIC Abstract. We consider H V = M +V, where (M, g) is a Riemannian manifold (not necessarily

More information

Max-Planck-Institut fur Mathematik in den Naturwissenschaften Leipzig Uniformly distributed measures in Euclidean spaces by Bernd Kirchheim and David Preiss Preprint-Nr.: 37 1998 Uniformly Distributed

More information

Numerische Mathematik

Numerische Mathematik Numer. Math. (2003) 94: 195 202 Digital Object Identifier (DOI) 10.1007/s002110100308 Numerische Mathematik Some observations on Babuška and Brezzi theories Jinchao Xu, Ludmil Zikatanov Department of Mathematics,

More information

SUPERCONVERGENCE PROPERTIES FOR OPTIMAL CONTROL PROBLEMS DISCRETIZED BY PIECEWISE LINEAR AND DISCONTINUOUS FUNCTIONS

SUPERCONVERGENCE PROPERTIES FOR OPTIMAL CONTROL PROBLEMS DISCRETIZED BY PIECEWISE LINEAR AND DISCONTINUOUS FUNCTIONS SUPERCONVERGENCE PROPERTIES FOR OPTIMAL CONTROL PROBLEMS DISCRETIZED BY PIECEWISE LINEAR AND DISCONTINUOUS FUNCTIONS A. RÖSCH AND R. SIMON Abstract. An optimal control problem for an elliptic equation

More information

Second order sucient optimality condition for a. nonlinear elliptic control problem. Eduardo Casas and Fredi Troltzsch and Andreas Unger

Second order sucient optimality condition for a. nonlinear elliptic control problem. Eduardo Casas and Fredi Troltzsch and Andreas Unger Second order sucient optimality condition for a nonlinear elliptic control problem by Eduardo Casas and Fredi Troltzsch and Andreas Unger Technical University of Chemnitz wickau Faculty of Mathematics

More information

On Nonlinear Dirichlet Neumann Algorithms for Jumping Nonlinearities

On Nonlinear Dirichlet Neumann Algorithms for Jumping Nonlinearities On Nonlinear Dirichlet Neumann Algorithms for Jumping Nonlinearities Heiko Berninger, Ralf Kornhuber, and Oliver Sander FU Berlin, FB Mathematik und Informatik (http://www.math.fu-berlin.de/rd/we-02/numerik/)

More information

Math 5520 Homework 2 Solutions

Math 5520 Homework 2 Solutions Math 552 Homework 2 Solutions March, 26. Consider the function fx) = 2x ) 3 if x, 3x ) 2 if < x 2. Determine for which k there holds f H k, 2). Find D α f for α k. Solution. We show that k = 2. The formulas

More information

Discontinuous Galerkin Time Discretization Methods for Parabolic Problems with Linear Constraints

Discontinuous Galerkin Time Discretization Methods for Parabolic Problems with Linear Constraints J A N U A R Y 0 1 8 P R E P R N T 4 7 4 Discontinuous Galerkin Time Discretization Methods for Parabolic Problems with Linear Constraints gor Voulis * and Arnold Reusken nstitut für Geometrie und Praktische

More information

The Mortar Boundary Element Method

The Mortar Boundary Element Method The Mortar Boundary Element Method A Thesis submitted for the degree of Doctor of Philosophy by Martin Healey School of Information Systems, Computing and Mathematics Brunel University March 2010 Abstract

More information

University of Cyprus. Abstract. singularities is analyzed. In this method, the solution is approximated by the

University of Cyprus. Abstract. singularities is analyzed. In this method, the solution is approximated by the A Singular Function Boundary Integral Method for Elliptic Problems with Singularities hristos Xenophontos, Miltiades Elliotis and Georgios Georgiou Department of Mathematics and Statistics University of

More information

Error analysis and fast solvers for high-frequency scattering problems

Error analysis and fast solvers for high-frequency scattering problems Error analysis and fast solvers for high-frequency scattering problems I.G. Graham (University of Bath) Woudschoten October 2014 High freq. problem for the Helmholtz equation Given an object Ω R d, with

More information

Supraconvergence of a Non-Uniform Discretisation for an Elliptic Third-Kind Boundary-Value Problem with Mixed Derivatives

Supraconvergence of a Non-Uniform Discretisation for an Elliptic Third-Kind Boundary-Value Problem with Mixed Derivatives Supraconvergence of a Non-Uniform Discretisation for an Elliptic Third-Kind Boundary-Value Problem with Mixed Derivatives Etienne Emmrich Technische Universität Berlin, Institut für Mathematik, Straße

More information

An explicit nite element method for convection-dominated compressible viscous Stokes system with inow boundary

An explicit nite element method for convection-dominated compressible viscous Stokes system with inow boundary Journal of Computational and Applied Mathematics 156 (2003) 319 343 www.elsevier.com/locate/cam An explicit nite element method for convection-dominated compressible viscous Stokes system with inow boundary

More information

Element diameter free stability parameters. for stabilized methods applied to uids

Element diameter free stability parameters. for stabilized methods applied to uids Element diameter free stability parameters for stabilized methods applied to uids by Leopoldo P. Franca Laboratorio Nacional de Computac~ao Cientica (LNCC/CNPq) Rua Lauro Muller 455 22290 Rio de Janeiro,

More information