Fast Algorithms For Boundary Integral Equations
Fast Algorithms for Boundary Integral EquationsLexing YingDepartment of Mathematics, University of Texas, Austin, TX 78712, USA,lexing@math.utexas.eduSummary. This article reviews several fast algorithms for boundary integral equations. Aftera brief introduction of the boundary integral equations for the Laplace and Helmholtz equations, we discuss in order the fast multipole method and its kernel independent variant, thehierarchical matrix framework, the wavelet based method, the high frequency fast multipolemethod, and the recently proposed multidirectional algorithm.Keywords. Boundary integral equations, Laplace equation, Helmholtz equation, fastalgorithms, fast multipole method, hierarchical matrices, wavelets, multiscale methods, multidirectional methods.AMS subject classifications. 45A05, 65R20.1 OutlineMany physical problems can be formulated as partial differential equations (PDEs)on certain geometric domains. For some of them, the PDEs can be reformulated usingthe so-called boundary integral equations (BIEs). These are integral equations whichonly involve quantities on the domain boundary. Some advantages of working withthe BIEs are automatic treatments of boundary condition at infinity, better conditionnumbers, and fewer numbers of unknowns in the numerical solution. On the otherhand, one of the major difficulties of the BIEs is that the resulting linear systems aredense, which is in direct contrast to the sparse systems of the PDEs. For large scaleproblems, direct solution of these dense linear systems becomes extremely timeconsuming. Hence, how to solve these dense linear systems efficiently has becomeone of the central questions. Many methods have been developed in the last twentyyears to address this question. In this article, we review some of these results.We start in Sect. 2 with a brief introduction of the boundary integral formulationwith the Laplace and Helmholtz equations as our examples. A major difference between these two equations is that the kernel of the Laplace equation is non-oscillatorywhile the one of the Helmholtz equation is oscillatory. For the non-oscillatory kernels, we discuss the fast multipole method (FMM) in Sect. 3 and its kernel indepen-
140L. Yingdent variant in Sect. 4, the hierarchical matrices frame in Sect. 5, and the waveletbased methods in Sect. 6. For the oscillatory kernels, we review the high frequencyfast multipole method (HF-FMM) in Sect. 7 and the recently developed multidirectional method in Sect. 8.The purpose of this article is to provide an introduction to these methods foradvanced undergraduate and graduate students. Therefore, our discussion mainly focuses on algorithmic ideas rather than theoretical estimates. For the same reason, wemostly refer only to the original papers of these methods and keep the size of thereference list to a minimum. Many important results are not discussed here due tovarious limitations and we apologize for that.2 Boundary Integral FormulationMany linear partial differential equation problems have boundary integral equationformulations. In this section, we focus on two of the most important examples anddemonstrate how to transform the PDE formulations into the BIE formulations.Our discussion mostly follows the presentation in [11, 18, 20]. We denote 1 with iand assume that all functions are sufficiently smooth.2.1 Laplace equationLet D be a bounded domain with smooth boundary in Rd (d 2, 3). n is the exteriornormal to D. The Laplace equation on D with Dirichlet boundary condition is u 0 in D(1)on D(2)u fwhere f is defined on D. The geometry of the problem is shown in Fig. 1. We seekto represent u(x) for x D in an integral form which uses only quantities on theboundary D.The Green’s function for the Laplace equation is(1ln 1(d 2)(3)G(x, y) 21π 1 x y (d 3)4π x y Some of the important properties of G(x, y) are G(x, y) is symmetric in x and y,G(x, y) is non-oscillatory, and x G(x, y) δy (x) and y G(x, y) δx (y)where x and y take the derivatives with respect x and y, respectively, and δx isthe Dirac function located at x. The following theorem is a simple consequence ofStokes’ theorem.
Fast Algorithms141Fig. 1. Domain of the Dirichlet boundary value problem of the Laplace equation.Theorem 1. Let u and v to be two sufficiently smooth functions on D̄. Then¶ZZ µ v(y) u(y)ds(y).u(u v v u)dx v n(y) n(y) DDA simple application of the previous theorem gives the following result.Theorem 2. Let u be a sufficiently smooth function on D̄ such that u 0 in D.For any x in D,¶Z µ u(y) G(x, y)u(x) G(x, y) u(y)ds(y). n(y) D n(y)Proof. Pick a small ball B at x that is contained in D (see Fig. 2). From the lasttheorem, we have G(x, y) G(x, y) u(y))ds(y) D\B (u(y) ³ G(x,y) u(y) (D\B) u(y) n(y) G(x, y) n(y) ds(y).RRSince u(y) 0 and G(x, y) 0 for y D \ B, the left hand side is equal tozero. Therefore, R ³ G(x,y) u(y) D u(y) n(y) G(x, y) n(y) ds(y) R ³ u(y) B u(y) G(x,y)n(y) G(x, y) n(y) ds(y)where n points towards x on B. Now let the radius of the ball B go to zero. The firstterm of the right hand side goes to u(x) while the second term approaches 0.From the last theorem, we see that u(x) for x in D can be represented as a sum oftwo boundary integrals. In the boundary integral formulation, we seek to representu(x) using only one of them. This degree of freedom gives rise to the following twoapproaches.
142L. YingFig. 2. Proof of Theorem 2.Method 1We represent u(x) for x D using the integral that contains G(x, y)u(x) Z Dϕ (y)G(x, y)ds(y)(4)where ϕ is an unknown density on D. This formulation is called the single layerform and ϕ is often called the single layer density. One can show that any sufficientlynice u(x) can be represented using the single layer form (see [20] for details). Lettingx approach z D, we getZf (z) u(z) Dϕ (y)G(z, y)ds(y),which is an integral equation that involves only boundary quantities ϕ and f . Therefore, the steps to solve the Laplace equation using the single layer form are:1. Find ϕ (z) on D such thatf (z) Z Dϕ (y)G(z, y)ds(y).(5)This equation is a Fredholm equation of the first kind (see [20]).2. For x in D, compute u(x) byu(x) Z Dϕ (y)G(x, y)ds(y).(6)
Fast Algorithms143Method 2We can also represent u(x) for x D using the integral that containsu(x) Z Dϕ (y) G(x, y)ds(y) n(y) G(x,y) n(y)(7)where ϕ is again an unknown density on D. This formulation is called the doublelayer form and ϕ is the double layer density. In fact, the double layer form is capableof representing any sufficiently nice u(x) in D [20]. If we now let x approach z D,we obtain the following equation on the boundary:1f (z) u(z) ϕ (z) 2Z D G(z, y)ϕ (y)ds(y). n(y)The extra 21 ϕ (z) term comes up because the integral (7) is not uniformly integrablenear z D. Hence, one cannot simply exchange the limit and integral signs. Sincethe boundary D is smooth, the integral operator with the kernel G(z,y)n(y) is a compactoperator. The steps to solve the Laplace equation using the double layer form are:1. Find ϕ (z) on D such that1f (z) ϕ (z) 2Z D G(z, y)ϕ (y)ds(y). n(y)(8)This equation is a Fredholm equation of the second kind.2. For x in D, compute u(x) withu(x) G(x, y)ϕ (y)ds(y). D n(y)Z(9)Between these two approaches, we often prefer to work with the double layerform (Method 2). The main reason is that the Fredholm equation of the second kindhas a much better condition number, thus dramatically reducing the number of iterations required in a typical iterative solver.2.2 Helmholtz equationWe now turn to the Helmholtz equation. Let D be a bounded domain with smoothboundary in Rd (d 2, 3) and n be the exterior normal to D. The unboundedHelmholtz equation on Rd \ D̄ (d 2, 3) with Dirichlet boundary condition describesthe scattering field of a sound soft object: u k2u 0 in Rd \ D̄u(x) uinc (x)for x D(10)(11)
144L. YingFig. 3. Domain of the Dirichlet boundary value problem of the Helmholtz equation.lim rr d–12µ¶ u iku 0 r(12)where k is the wave number, uinc is the incoming field and u is the scattering field.The last equation is called the Sommerfeld radiation condition which guarantees thatthe scattering field propagates to infinity. The geometry of the problem is describedin Fig. 3. Our goal is again to represent u(x) for x Rd \ D̄ in an integral form whichuses quantities defined on the boundary D.The Green’s function of the Helmholtz equation is(i 1H0 (k x y ) (d 2)G(x, y) 41 exp(ik x y )(13)(d 3)4π x y Some of the important properties of G(x, y) are G(x, y) is symmetric,G(x, y) is oscillatory,( x k2 )G(x, y) δy (x) and ( y k2 )G(x, y) δx (y).Theorem 3. Let C be a bounded domain with smooth boundary. Suppose that u issufficiently smooth in C̄ and satisfies ( k2 )u 0 in C. Then for any x in C¶Z µ u(y) G(x, y)u(x) G(x, y) u(y)ds(y). n(y) C n(y)Proof. Pick a small ball B centered at x. Then we haveR G(x, y) G(x, y) u(y))dy C\B (u(y) ³ G(x,y) u(y) (C\B) u(y) n(y) G(x, y) n(y) ds(y).RThe left hand side is equal toZC\B(u · ( G k2G) G( u k2u))dy 0.The rest of the proof is the same as the one of Theorem 2.
Fast Algorithms145Fig. 4. Proof of Theorem 4.The above theorem addresses a bounded domain C. However, what we are reallyinterested in is the unbounded domain Rd \ D̄.Theorem 4. Suppose that u is sufficiently smooth and satisfies ( k2 )u 0 inRd \ D. Then for any x in Rd \ D,¶Z µ u(y) G(x, y)ds(y).G(x, y) u(y)u(x) n(y) D n(y)Proof. Pick a large ball Γ that contains D. Consider the domain Γ \ D̄ (see Fig. 4).Let t be the exterior normal direction of Γ \ D̄. From the previous theorem, we have¶Z µ u(y) G(x, y)ds(y) G(x, y) u(y)u(x) t t Γ¶Z µ u(y) G(x, y)G(x, y) u(y)ds(y). t t DUsing the Sommerfeld condition at infinity, one can show that the integral over Γgoes to zero as one pushes the radius of Γ to infinity [11]. Noting that t n on D, we have¶Z µ G(x, y) u(y) G(x, y) ds(y).u(x) u(y) n(y) n(y) DFrom the last theorem, we see that u(x) for x in Rd \ D̄ can be represented as asum of two integrals. In the boundary integral formulation of the Helmholtz equation,one option is to represent u(x) by the double layer form:u(x) G(x, y)ϕ (y)ds(y) D n(y)Z
146L. YingDifferent from the double layer form of the Laplace equation, the double layer formof the Helmholtz equation is not capable of representing arbitrary field u(x) for x Rd \ D̄. If k is one of the internal resonant numbers such that the internal Neumannproblem with zero boundary condition has non-trivial solution, then this double layerform is singular (see [11]). In practice, we use¶Z µ G(x, y)u(x) iη G(x, y) ϕ (y)ds(y). n(y) Dwhere η is a real number (for example, η k). As we let x approach z on D, weget¶Z µ1 G(z, y) iη G(z, y) ϕ (y)ds(y) uinc (z) u(z) ϕ (z) 2 n(y) Dwhere the extra term 21 ϕ (z) is due to the fact that the integral is improper at z D.The steps to solve the Helmholtz equation using this double layer form are:1. Find a function ϕ (z) on D such that¶Z µ1 G(z, y) uinc (z) ϕ (z) iη G(z, y) ϕ (y)ds(y).2 n(y) D2. For point x in R3 \ D, compute u(x) with¶Z µ G(x, y) iη G(x, y) ϕ (y)ds(y).u(x) n(y) D(14)(15)We have seen the derivations of the BIEs for the interior Laplace Dirichlet boundary value problem and the exterior Helmholtz Dirichlet boundary value problem.Though both cases use the Green’s functions of the underlying equation and theStokes’ theorem, the derivation for the Helmholtz equation is complicated by the existence of the internal resonant numbers. For other elliptic boundary value problems,the derivations of the BIE formulations often differ from case to case.2.3 DiscretizationIn both BIEs discussed so far, we need to solve a problem of the following form: findϕ (x) on D such thatf (x) ϕ (x) orf (x) ZZ D DK(x, y)ϕ (y)ds(y),K(x, y)ϕ (y)ds(y),i.e.,i.e.,f (I K)ϕf Kϕ .where K(x, y) is either the Green’s function or its derivative of the underlying PDE.In order to solve these equations numerically, we often use one of the followingthree discretization methods: the Nyström method, the collocation method, and theGalerkin method. Let us discuss these methods briefly using the Fredholm equationof the second kind.
Fast Algorithms147Fig. 5. Nyström methodNyström methodThe idea of the Nyström method is to approximate integral operators with quadratureoperators. The steps are:1. Approximate the integral operator (K ϕ )(x) : K(x, y)ϕ (y)dy with the quadrature operatorRN(KN ϕ )(x) : K(x, x j )λ j ϕ (x j )j 1where {x j } are the quadrature points and {λ j } are the quadrature weights (seeFig. 5). Here we make the assumption that {λ j } are independent of x. In practice,{λ j } often depend on x when x j is in the neighborhood of x if the kernel K(x, y)has a singularity at x y.2. Find ϕ (x) such that ϕ KN ϕ f . We write down the equation at {xi }:nϕi K(xi , x j )λ j ϕ j fi ,j 1i 1, · · · , N(16)and solve for {ϕi }. Here fi f (xi ).3. The value of ϕ (x) at x D is computed usingnϕ (x) f (x) K(x, x j )λ j ϕ j .(17)j 1Collocation methodThe idea of the collocation method is to use subspace approximation. The steps are:1. Approximate ϕ (x) by Nj 1 c j ϕ j (x) where {ϕ j (x)} are basis functions on D.Let {x j } be a set of points on D (see Fig. 6), where x j is often the center ofsupp(ϕ j ).
148L. YingFig. 6. Collocation and Galerkin methods2. Find {c j } such that ϕ K ϕ f is satisfied at {x j }, i.e.,NN c j ϕ j (xi ) (K( c j ϕ j ))(xi ) f (xi ),j 1j 1i 1, · · · , N(18)Galerkin methodThe idea of the Galerkin method is to use space approximation with orthogonalization. The steps are:1. Approximate ϕ (x) by Nj 1 c j ϕ j (x) where {ϕ j (x)} are often localized basisfunctions on D.2. Find {c j } such that ϕ K ϕ f is orthogonal to all the subspace generated byϕ (x).jNNhϕi , c j ϕ j K( c j ϕ j ) f i 0,j 1j 1i 1, · · · , N(19)2.4 Iterative solutionThe following discussion is in the setting of the Nyström method. The situations forthe other methods are similar. In the matrix form, the linear system that one needs tosolve is(II K Λ )ϕ fwhere I is the identity matrix, K is the matrix with entries K(xi , x j ), Λ is the diagonalmatrix with the diagonal entries equal to {λ j }, ϕ is the vector of {ϕ j }, and f is thevector of { f j }.
Fast Algorithms149Since K is a dense matrix, the direct solution of this equation takes O(N 3 ) steps.For large N, this becomes extremely time-consuming and solving this system directlyis not feasible. Therefore, we need to resort to iterative solvers.Since the integral operator K is compact, its eigenvalues decay to zero. This isalso true for the discretized version, the matrix K . Therefore, the condition number ofI K Λ is small and independent of the number of quadrature points N. As a result,the number of iterations is also independent of N. In each iteration, one computesK ψ for a given vector ψ . Since K is dense, a naive implementation of this matrixvector multiplication takes O(N 2 ) steps, which can be still quite expensive for largevalues of N. How to compute the product K ψ is the question that we will address inthe following sections.Before we move on, let us compare the PDE and BIE formulations. For thePDE formulations, a numerical solution often requires O((1/h)d ) unknowns for agiven discretization size h. Special care is necessary for unbounded exterior problems. Since the resulting linear system is sparse, each iteration of the iterative solveris quite fast though the number of iterations might be large. Finally, the PDE formulations work for domains with arbitrary geometry and problems with variablecoefficients.For the BIE formulations, a numerical solution involves only O((1/h)d 1 ) unknowns on the domain boundary for a given discretization size h. No special careis needed for exterior domains. The resulting system is always dense, so fast algorithms are necessary for efficient iterative solutions of the BIE formulations. As wehave seen already, the Green’s functions are fundamental in deriving the integralequations. Since the Green’s functions are often unknown for problems with variable coefficients, most applications of the BIE formulations are for problems withconstant coefficient.3 Fast Multipole MethodIn each step of the iterative solution of a BIE formulation, we face the followingproblem. Given a set of charges { fi , 1 i N} located at points {pi , 1 i N}(see Fig. 7) and the Green’s function G(x, y) of the underlying equation, we want tocompute at each pi the potentialNui G(pi , p j ) f j .(20)j 1As we pointed earlier, a naive algorithm takes O(N 2 ) steps, which can be quite expensive for large values of N. In this section, we introduce the fast multipole methodby Greengard and Rokhlin [15, 16] for the Green’s function of the Laplace equation.This remarkable algorithm reduces the complexity from O(N 2 ) to O(N) for any fixedaccuracy ε .
150L. YingFig. 7. Distribution of quadrature points {pi } on the boundary of the domain D.3.1 Geometric partTwo sets A and B are said to be well-separated if the distance between A and B aregreater than their diameters. Let us consider the interaction from a set of points {y j }in B to a set of points {xi } in A, where both {y j } and {xi } are subsets of {pi }. Thegeometry is shown in Fig. 8.Fig. 8. Two boxes A and B are well-separated. Direct computation takes O(N 2 ) steps.Suppose that { f j } are the charges at {y j }. Let us consider the following approximation for the potential ui at each xiui u(cA ) G(cA , y j ) f j G(cA , cB ) f j .j(21)jThis approximation is quite accurate when A and B are far away from each other andis in fact used quite often in computational astrophysics to compute the interactionbetween distant galaxies. However, for two sets A and B which are merely wellseparated (the distance between them is comparable to their diameters), this approximation introduces significant error. Let us not worry too much about the accuracy
Fast Algorithms151Fig. 9. The three steps of the approximate procedure. The total number operations is O(N).at this moment and we will come back to this point later. A geometric description ofthis approximation is given in Fig. 9.We have introduced two representations in this simple approximation: fB , the far field representation of B that allows one to approximately reproducein the far field of B the potential generated by the source charges inside B.uA , the local field representation of A that allows one to approximately reproduceinside A the potential generated by the source charges in the far field of A.The computation of uA from fBuA G(cA , cB ) fBis called a far-to-local translation. Assuming both {y j } and {xi } contain O(n)points, the naive direct computation of the interaction takes O(n 2) steps. The proposed approximation is much more efficient: fB j f j takes O(n) steps.uA G(cA , cB ) fB takes O(1) steps.ui uA for all xi A takes O(n) steps as well.Hence, the complexity of this three step procedure is O(n). Viewing the interactionbetween A and B in a matrix form, we see that this interaction is approximately lowrank if A and B are well-separated. In fact, in the above approximation, a rank-1approximation is used.However, in the problem we want to address, all the points {pi } are mixed together and each pi is both a source and a target. Therefore, one cannot apply theabove procedure directly. The solution is to use an adaptive tree structure, namelythe octree in 3D or the quadtree in 2D (see Fig. 10). We first choose a box that contains all the points {pi }. Starting from this top level box, each box of the quadtree isrecursively partitioned unless the number of points inside it is less than a presc
2ϕ(z) term comes up because the integral (7) is not uniformly integrable near z D. Hence, one cannot simply exchange the limit and integral signs. Since the boundary D is smooth, the integral operator with the kernel G(z,y) n(y) is a compact operator. The steps to solve
Bruksanvisning för bilstereo . Bruksanvisning for bilstereo . Instrukcja obsługi samochodowego odtwarzacza stereo . Operating Instructions for Car Stereo . 610-104 . SV . Bruksanvisning i original
What are boundary integral equations? We can reformulate boundary value problems for PDEs in a domain as integral equations on the boundary of that domain. We typically use them for linear, elliptic, and homogeneous PDEs, but not always. Boundary integral equation methods refer to the numeric
10 tips och tricks för att lyckas med ert sap-projekt 20 SAPSANYTT 2/2015 De flesta projektledare känner säkert till Cobb’s paradox. Martin Cobb verkade som CIO för sekretariatet för Treasury Board of Canada 1995 då han ställde frågan
service i Norge och Finland drivs inom ramen för ett enskilt företag (NRK. 1 och Yleisradio), fin ns det i Sverige tre: Ett för tv (Sveriges Television , SVT ), ett för radio (Sveriges Radio , SR ) och ett för utbildnings program (Sveriges Utbildningsradio, UR, vilket till följd av sin begränsade storlek inte återfinns bland de 25 största
Hotell För hotell anges de tre klasserna A/B, C och D. Det betyder att den "normala" standarden C är acceptabel men att motiven för en högre standard är starka. Ljudklass C motsvarar de tidigare normkraven för hotell, ljudklass A/B motsvarar kraven för moderna hotell med hög standard och ljudklass D kan användas vid
LÄS NOGGRANT FÖLJANDE VILLKOR FÖR APPLE DEVELOPER PROGRAM LICENCE . Apple Developer Program License Agreement Syfte Du vill använda Apple-mjukvara (enligt definitionen nedan) för att utveckla en eller flera Applikationer (enligt definitionen nedan) för Apple-märkta produkter. . Applikationer som utvecklas för iOS-produkter, Apple .
‣ Problem formulation: PDE vs. boundary and volume integral equations ‣ Boundary integral equation method: Collocation, Galerkin Boundary Element Method (BEM) and Nyström methods for boundary integral equations,
1. Merancang aturan integral tak tentu dari aturan turunan, 2. Menghitung integral tak tentu fungsi aljabar dan trigonometri, 3. Menjelaskan integral tentu sebagai luas daerah di bidang datar, 4. Menghitung integral tentu dengan menggunakan integral tak tentu, 5. Menghitung integral dengan rumus integral substitusi, 6.
