Uniform estimate of the constant in the strengthened CBS inequality for anisotropic non-conforming FEM systems

Size: px
Start display at page:

Download "Uniform estimate of the constant in the strengthened CBS inequality for anisotropic non-conforming FEM systems"

Transcription

1 Uniform estimate of te constant in te strengtened CBS inequality for anisotropic non-conforming FEM systems R. Blaeta S. Margenov M. Neytceva Version of November 0, 00 Abstract Preconditioners based on various multilevel extensions of two-level finite element metods (FEM) lead to iterative metods wic ave an optimal order computational complexity wit respect to te size of te system. Suc metods were first presented in [0, ], and are based on (recursive) two-level splittings of te finite element space. Te key role in te derivation of optimal convergence rate estimates plays te constant γ in te so-called Caucy-Bunyakowski-Scwarz (CBS) inequality, associated wit te angle between te two subspaces of te splitting. It turns out tat only existence of uniform estimates for tis constant is not enoug and accurate quantitative bounds for γ ave to be found as well. More precisely, te value of te upper bound for γ (0, ) is a part of te construction of various multilevel extensions of te related two-level metods. In tis paper an algebraic two-level preconditioning algoritm for second order elliptic boundary value problems is constructed, were te discretization is done using Crouzeix- Raviart non-conforming linear finite elements on triangles. An important point to make is tat in tis case te finite element spaces corresponding to two successive levels of mes refinements are not nested. To andle tis, a proper two-level basis is considered, wic enables us to fit te general framework for te construction of two-level preconditioners for conforming finite elements and to generalize te metod to te multilevel case. Te major contribution of tis paper is te derived estimates of te related constant γ in te strengtened CBS inequality. Tese estimates are uniform wit respect to bot coefficient and mes anisotropy. Up to our knowledge, te results presented in te paper are te first for non-conforming FEM systems. Institute of Geonics, Czec Academy of Sciences, Studentska 768, Ostrava-Poruba, Te Czec Republic, blaeta@ugn.cas.cz Central Laboratory for Parallel Processing, Bulgarian Academy of Sciences, Acad. G. Boncev str, bl. 5A, Sofia, Bulgaria, margenov@parallel.bas.bg Department of Scientific Computing, Institute of Information Tecnology, Uppsala University, Box 0, SE Uppsala, Sweden, maya@tdb.uu.se

2 KEY WORDS: multilevel preconditioners, ierarcical basis, CBS constant. Introduction. Two-level preconditioners for conforming finite element discretizations In tis paper we consider te elliptic boundary value problem Lu (a(x) u(x)) = f(x) in Ω, u = 0 on Γ D, (a(x) u(x)) n = 0 on Γ N, () were Ω is a convex polygonal domain in R, f(x) is a given function in L (Ω), a(x) = [a ij (x)] i,j= is a symmetric and uniformly positive definite matrix in Ω, n is te outward unit vector normal to te boundary Γ = Ω, and Γ = Γ D Γ N. We assume tat te entries a ij (x) are piece-wise smoot functions on Ω. Te weak formulation of te above problem reads as follows: given f L (Ω) find u V HD (Ω) = {v H (Ω) : v = 0 on Γ D }, satisfying A(u, v) = (f, v) v HD(Ω), were A(u, v) = a(x) u(x) v(x)dx. () We assume tat te domain Ω is discretized using triangular elements. Te partition is denoted by T and is assumed to be obtained by a proper refinement of a given coarser triangulation T H. Te partition T H is aligned wit te discontinuities of te coefficient a(x) so tat over eac element E T H te function a(x) is smoot. Te variational problem () is ten discretized using te finite element metod, i.e., te continuous space V is replaced by a finite dimensional subspace V. Ten te finite element formulation is: find u V, satisfying A (u, v ) = (f, v ) v V, were A (u, v ) = a(e) u v dx. () e T Here a(e) is a piece-wise constant coefficient matrix, defined by te integral averaged values of a(x) over eac triangle from te coarser triangulation T H. We note tat in tis way strong coefficient jumps across te boundaries between adjacent finite elements from T H are allowed. Piece-wise linear finite elements are considered in te paper. For te standard conforming FEM, te nodal basis is associated wit te vertices of te triangles wile for te case of nonconforming Crouzeix-Raviart finite elements, te interpolation nodes are te mid points of te sides. Te resulting discrete problem to be solved is ten a linear system of equations Ω A u = F, (4) e

3 θ 6 θ 4 θ θ 5 θ θ (a) Refined mes (b) One macro-element (c) One element Figure : Conforming linear finite elements. wit A being te corresponding global stiffness matrix and being te discretization (messize) parameter for te underlying triangulation T of Ω. Te aim of tis paper is to investigate two-level preconditioners for solving te system (4). Te general setting and some well-known results for te case of conforming finite elements are summarized in te rest of tis section. Te next two sections are devoted to te study of two-level preconditioners for te case of non-conforming Crouzeix-Raviart finite elements. In Section we consider a uniform estimate of te strengtened CBS constant for te twolevel splitting, firstly introduced in [0]. Section analyses anoter two-level splitting, introduced recently in [4], wic allows for an easier extension to te multi-level case. Some concluding remarks are given at te end.. Te two-level setting We are concerned wit te construction of a two-level preconditioner M for A, suc tat te spectral condition number κ(m A ) of te preconditioned matrix M A is uniformly bounded wit respect to te messize parameter, te sape of triangular finite elements and arbitrary coefficient anisotropy. Te classical teory for constructing optimal order two-level preconditioners was first developed in [7, ], see also []. Te general framework requires to define two nested finite element spaces V H V, tat correspond to two consecutive (regular) mes refinements, as illustrated in Figure (c) and (b). Let T H and T be two successive mes refinements of te domain Ω, wic correspond to V H and V. Let {φ (k) H, k =,,, N H} and {φ (k), k =,,, N } be te corresponding standard finite element basis functions. We split te mespoints from T into two groups, {N H } T H and {N \H } T \T H, (5) wic latter are te newly added node-points. Next we define te so-called ierarcical basis functions { φ (k), k =,,, N } = {φ (l) H on T H} {φ (m) on T \T H }. (6) Let ten à be te corresponding ierarcical stiffness matrix.

4 Under te splitting (5) bot matrices A and à admit in a natural way a two-by-two block structure [ ] ] A A A = }N\H [à Ã, à A A }N = }N \H. (7) H à à }N H [ ] I 0 As is well-known, tere exists a transformation matrix J =, wic relates te nodal J I point vectors for te standard and te ierarcical basis functions as follows, ] [ ] [ṽ v ṽ ṽ = = J, = v. ṽ = J v + v ṽ v Remark. Clearly, te ierarcical stiffness matrix à is more dense tan A and terefore its action on a vector is computationally more expensive. Te transformation matrix J, owever, enables us in practical implementations to work wit A, since à = JA J T.. Two-level preconditioners and te strengtened Caucy-Bunyakowski- Scwarz inequality Consider a general matrix A, wic is assumed to be symmetric positive definite and partitioned as in (7). Te quality of tis partitioning is caracterized by te corresponding CBS inequality constant: γ = sup v R n n, v R n were n = N and n = N H. Consider now two preconditioners to A. v T A v, (8) (v T A v ) / (v T / A v ) (a) A preconditioner of block-diagonal (additive) form. Consider te following symmetric block-diagonal matrix as a preconditioner to A, [ ] B 0 M B = 0 B under te additional assumptions α A B α A and β A B β A for some positive constants α i, β i, i =,. Te latter inequalities are in a positive semidefinite sense. Under te above assumptions, tere olds te following general estimate of te spectral condition number of M B A : κ(m B A) α α ( γ ) [ ] (+θ ) + 4 ( θ ) + θ γ [ ] (+θ ) + 4 ( θ ) + θ γ, 4 (9)

5 were γ is te CBS constant in (8) and θ i = α i β i, i =,. Wen B = A and B = A, ten estimate (9) reduces to κ(m B A) + γ γ. (0) (b) A full block-matrix factorization preconditioner (multiplicative, or of block Gauss-Seidel form). Tis type of preconditioning is based on te exact block-matrix factorization of A, [ ] [ ] A 0 I A A A =, A S 0 I were S = A A A A is te exact Scur complement of A. Te multiplicative preconditioner is ten of te form [ ] [ ] A 0 I A A M F =, () A M S 0 I were M S is properly cosen. Assuming tat tere olds δ 0 A M S δ A, were γ < δ 0 δ, ten κ(m F A) δ γ δ 0 γ δ δ 0 γ. () Detailed proofs of (9), (0) and (), and analysis of oter versions of constructing M B and M F are found, for instance, in []. In te ierarcical bases context V and V are subspaces of te finite element space V spanned, respectively, by te basis functions at te new nodes {N \H } and by te basis functions at te old nodes {N H }. For te strengtened CBS inequality constant, tere olds tat γ = cos(v, V ) = A(u, v) sup () u V, v V A(u, u)a(v, v) were A(, ) is te bilinear form wic appears in te variational formulation of te original problem. Wen V V = {0}, te constant γ is strictly less tan one. As sown in [7], te constant γ can be estimated locally over eac finite element E T H, wic means tat γ = max E γ E, were γ E = sup u V (E), v V (E) A E (u, v) AE (u, u)a E (v, v), v const. Te spaces V k (E) above contain te functions from V k restricted to E and A E (u, v) corresponds to A(u, v) restricted over te element E of T H (see also [6]). 5

6 (a) Refined mes II 7 θ 6 5 θ θ 8 III 9 4 (b) One macro-element I θ θ θ (c) One element Figure : Crouzeix-Raviart non-conforming linear finite elements. Using te local estimates, it is possible to sow tat te value of γ depends on te type of te basis functions cosen for V and V but is independent of (for te -version of te ierarcical bases metod). Furter, it does not depend on te geometry of te domain Ω. It is also easily seen tat γ is independent of any discontinuities of te coefficients of te bilinear form A(, ), as long as tey do not occur witin any element of te coarse triangulation used. Te -independence means tat if we ave a ierarcy of refinements of te domain wic preserve te properties of te initial triangulation (refinement by congruent triangles, for example), ten γ is independent of te level of te refinement as well. For certain implementations, it is sown tat γ is independent of anisotropy (see [, 9, 9, 5, 5]). Hence, as long as te rate of convergence is bounded by some function of γ, it is independent of various problem and discretization parameters, suc as te ones mentioned above. We stress ere, tat te above tecnique is developed and straigtforwardly applicable for conforming finite elements and nested finite element spaces, i.e., wen V H V. First reduce (FR) two-level preconditioning for Crouzeix- Raviart systems. Te FR algoritm We consider now a finite element discretization, done using Crouzeix-Raviart non-conforming linear finite elements on triangles. Figure (b) illustrates a macro-element obtained after one regular mes-refinement step. We see tat in tis case NH E = {I, II, III} and N E = {,,, 9}, tus, obviously V H and V are not nested. Let us recall now, tat, as noted in [4], see also [8], it suffices to provide a local analysis for a reference (isosceles rigt-angled) triangle and arbitrary coefficients in te bilinear form, or, equivalently, for te model Laplace problem on arbitrary saped triangles. Te latter strategy is adopted in tis section, wile te alternative approac is used in Section. For te rest, te analysis will old for a general sape triangle wit angles θ, θ and θ = π (θ + θ ). Let a = cot θ, b = cot θ and c = cot θ. Witout loss of generality we can assume tat 6

7 θ θ θ. Remarkably enoug, te standard nodal basis element stiffness matrix for Crouzeix-Raviart non-conforming linear elements (see Figure (c)) A CR e coincides wit tat for te conforming linear elements A cl e (see Figure (c)), up to a scalar factor 4, namely, A CR e = 4A cl e. Terefore, we will furter work wit te following element stiffness matrix for an arbitrary saped triangle, b + c c b c a + c a = c + β β α + α b a a + b β α α + β were α = a/c and β = b/c. It is sown (cf. Lemma in [8]), tat te following relations old. (i) a = ( bc)/(b + c); (ii) if θ θ θ ten a b c; (iii) a + b > 0 (iv) Te true region of definition for α = a/c and β = b/c is D = {(α, β) R : } < α, 0 < β, α + β > 0, α β. β D 0.5 α=β α 0.5 α+β=0 Figure : Te true region of definition (α, β) D. Next, we follow [0] to define an algebraic two-level preconditioner. Let ϕ E = {φ i (x, y)} 9 i= be te macro-element vector of te nodal basis functions and A E be te macroelement stiffness matrix corresponding to E T. Te global stiffness matrix A can be written as A = E T A E. 7

8 were te summation is understood as te FEM assembly procedure. Let us introduce te following macroelement level transformation matrix J E =, (4) wic defines locally a two level ierarcical basis ϕ E, namely, ϕ E = J E ϕ E. Te ierarcical two-level macro-element stiffness matrix is ten obtained as and te related global stiffness matrix reads as à E = J E A E J T E, à = E T à E. We split now te two-level stiffness matrix à into block form à = à à à Ã, (5) were à corresponds to interior nodal unknowns wit respect to te macro-elements E T. Te first step of First Reduce (FR) algoritm is to eliminate tese unknowns. For tis purpose we factor Ã, i.e., à = à 0 à B à à I 0 I, (6) were B = à Ãà à stands for te Scur complement of tis elimination step. Next we consider a two level splitting of te matrix B in te block form B = B B B B, (7) were te first block corresponds to te alf-difference basis functions. Now it is easily seen tat te matrix B can be associated wit te coarse grid. It is important to note tat kerb E; = kera e = span{,, } wic allows us to apply a local analysis to estimate te constant γ corresponding te splitting defined by te block partition (7). For our analysis we proceed as follows: 8

9 Step : We observe tat te top-left block of à is block-diagonal (it corresponds to te interior points,,, cf. Figure (b) wic are not connected to nodes in oter macroelements), tus, it can be eliminated exactly and tis can be done locally. Terefore, we first eliminate tat block and obtain te (6 6) matrix B E. Next we split B E as [ ] BE, B B E = E, }two-level alf-difference basis functions B E, B E, }two-level alf-sum basis functions written again in two-by-two block form wit blocks of order ( ). Step : We are now in a position to estimate te CBS constant γ locally. We will utilize te leftand side inequality of te following result (given, for instance, in Lemma 9. (b) from []), namely, γ vt Sv v T B v for all v, were S = B B B B. To estimate γ, it suffices to find an upper bound for λ min (B E, S E).. Uniform estimate of te constant in te strengtened CBS inequality For te estimation of γ we ave used symbolic computations wit MAPLE. Wit te latter, all matrices A E, ÃE, B E and S E can be manipulated as functions of te parameters α and β. Let us consider te local eigenvalue problem S E v = λb E, v, v const = (c, c, c) T. (8) Te minimum eigenvalue of B E, S E is found to be of te form λ min (B E, S E) = 5σ σ(σ 8αβ), were σ = (α + )(β + )(α + β). (9) 8σ It is easily seen (cf. (i)-(iv)) tat σ > 0. Using (9), we will next prove tat λ min (B E, S E) 4, (0) wic was inted by corresponding numerical experiments in MATLAB. Indeed, it is readily seen tat inequality (0) is equivalent to αβ σ. () Relation () is obviously true for α 0. To sow tat same olds for α < 0, we introduce te auxiliary function Ψ(α, β) = (α + )(β + )(α + β) + αβ. () 9

10 Figure 4: Surface of te MATLAB computed values of γ varying (α, β) D. If Ψ as an extremum in a point ( α, β) in te interior of D, ten Ψ/ α = 0 at tis point and α = β + β +. (β + ) Consider now Ψ( α, β), 0 < β. After some formula manipulations we find tat Ψ ( α, β) = (β + β + )(β + 5β + ) 4 (β + ) wic is negative. Tus, Ψ( α, β) is a strictly decreasing function of β and attends its minimum on te boundary of D, at β =. Now, a simple analysis of Ψ(α, ) = α +α+ sows tat te minimum of Ψ for α ( /, 0) as to be taken for α = /. We find ten Ψ( /, ) = 0, tus, our proposition is confirmed. We find tus, tat We collect te above results in a teorem. γ λ min (B E, S E) 4. Teorem. Let te assumptions of te introduced FR two-level algoritm are fulfilled. Ten, te related constant in te strengtened CBS inequality is uniformly bounded wit respect to bot coefficient and mes anisotropy by /4, namely, γ 4. () Tis result is also independent of te size of te problem (or te mes parameter ) and of possible coefficient jumps aligned wit te finite element partition T H. 0

11 Te obtained upper bound for γ is exact wic is also seen from te plot of te surface given in Figure 4. Having sown estimate (), from (9), (0) and () we can ten estimate te condition numbers of te preconditioned systems using te preconditioner M B and M F. For instance, estimate (0) becomes κ(m B A ) 7. (4) Two-level splitting by differentiation and aggregation (DA) for te Crouzeix - Raviart finite elements Splitting (4) is not te only one possible for discretizations wit Crouzeix - Raviart finite elements. Oter variants of two-level splittings are discussed in [4], referred to as differentiation and aggregation (DA) splittings. Te most important case of tese is defined and analysed in tis section. Te splitting is easily described for one macroelement E, see Figure (b). If φ,..., φ 9 are te standard nodal nonconforming linear finite element basis functions on te macroelement, ten we define V (E) = span {φ,..., φ 9 } = V (E) V (E), (5) V (E) = span {φ, φ, φ, φ 4 φ 5, φ 6 φ 7, φ 8 φ 9 }, (6) V (E) = span {φ + φ 4 + φ 5, φ + φ 6 + φ 7, φ + φ 8 + φ 9 }. (7) Using te transformation matrix J E, J E =, (8) te vector of te macroelement basis functions ϕ E = {φ i } 9 i= is transformed to a new ierarcical basis ϕ E = { φ i } 9 i = J E ϕ E. Accordingly, J transforms te macroelement stiffness matrix into a ierarcical form à E = J E A E J T E = [ ÃE, à E, à E, à E, ] φi V (E) φ i V (E). (9) For te wole finite element space V wit te standard nodal finite element basis ϕ = {φ (i) : i =,..., N }, we can similarly construct a new ierarcical basis ϕ = ϕ ϕ ϕ and a

12 corresponding splitting V = V V, (i) V = span{ φ ϕ (i) ϕ }, V = span{ φ ϕ } (0) Te new basis can be described in te following way. Denote te edges of te coarse triangular elements from T H as C edges and tese edges of te fine triangles from T, wic lie outside C edges, as F edges. Ten ϕ consists of te basis functions φ (i) corresponding to te nodes on F edges, ϕ consists from te differences φ (i) φ(j) corresponding to te pairs of nodes on C edges and ϕ consists from te aggregates φ (i) + φ(j) + φ(k) or φ (i) + φ(j) + φ(k) + φ (e) corresponding to te pairs of nodes on C edges and nodes on te F edges inside te coarse triangles adjacent to te given C edge, see Figure 5. Te transformation J suc tat ϕ = Jϕ, can be used for transformation of te stiffness matrix A to ierarcical form à = JA J T, wic allows preconditioning by te two-level preconditioners based on te splitting (0) ϕ = {φ (k) : k = 5, 6, 8, 9,, }, ϕ = {φ (i) φ(j) : (i, j) = (, ), (7, 4), (4, ), (, 0), (5, 6)}, ϕ = {φ (i) + φ(j) + φ(k) : (i, j, k) = (,, 9), (7, 4, 5), (, 0, ), (5, 6, 8)} {φ (4) + φ() + φ (6) + φ() }. Figure 5: Te ierarcical basis wit aggregations Now, we are in a position to analyze te constant γ = cos(v, V ) for te splitting (0). Again, as in te previous section, tis analysis is performed locally, by considering te corresponding problems on macroelements. We follow a procedure (cf. [5], [5]), wic sligtly differs from tat one used in te previous section in te sense tat we consider first te case of te reference rigt angle macroelement, see Figure 6. Let V (Ê), V(Ê) be te two-level splitting (5) - (7) for te reference macroelement Ê and for u V (Ê), v V(Ê) denote d(k) = d (k) (u) = u Tk, δ (k) = δ (k) (v) = v Tk. Ten te

13 6 4 T 7 T T T Figure 6: Te reference coarse grid triangle and te macroelement Ê. relations between te function values in some nodal points, namely u(p 4 ) = u(p 5 ), u(p 6 ) = u(p 7 ), u(p 8 ) = u(p 9 ) and v(p ) = v(p 4 ) = v(p 5 ), v(p ) = v(p 6 ) = v(p 7 ), v(p ) = v(p 8 ) = v(p 9 ), imply tat d () + d () + d () + d (4) = 0, () δ () = δ () = δ () = δ (4) = δ. () Hence, 4 4 A, Ê (u, v) = a u vdx = aδ (k), d (k) k= T k= k = aδ, d () + d () + d () d (4) () = aδ, d (4) δ a d (4) a were = area(t k ), x, y = x T y denotes te inner product in R, x a = ax, x. Furter, leads to and d (4) a = d () + d () + d () a 4 A, Ê (u, u) = k= d (k) a d (k) a k= ( + ) d (4) a (4) A, Ê (v, v) = 4 δ a. (5)

14 Tus, A, Ê (u, v) A 4,Ê(u, u) = 4 A 4,Ê(v, v) A (u, u),ê A, Ê (v, v). (6) In te case of arbitrary saped macroelement E we can use te affine mapping F : Ê E for transformation of te problem to te reference macroelement, for more details see e.g. [5], [5] and also te proof of te following teorem. Tis transformation canges te anisotropy will still old since te result (6) for te reference 4 of te problem, but te estimate γ E macroelement does not depend on anisotropy. Te obtained results are summarized in te following teorem, wic is analogous to Teorem.. Teorem. Let us consider te two level splitting wit aggregations (0). Ten te corresponding strengtened C.B.S. inequality constant γ is uniformly bounded wit respect to bot coefficients and mes anisotropy, γ /4. Te latter estimate is independent on te discretization (mes) parameter and possible coefficient jumps aligned wit te finite element partitioning T H. Te following teorem is useful for extending te two-level to multilevel preconditioners. Teorem. Let à be te stiffness matrix corresponding to te space V wit te basis ϕ from te splitting (0) and let A H be te stiffness matrix corresponding to te finite element space V H, corresponding to te coarse discretization T H, equipped wit te standard nodal finite element basis {φ (k) H : k =,..., N H}. Ten Proof Let x, y R N H. Ten were u H A H x, y = = i x i φ (i) H, v H à = 4 A H. (7) E T H A E (u H, v H ), Ãx, y = = i y i φ (i) H, u = Now, we sall sow tat 4A E (u H, v H ) = φ (i) ϕ e T, e E E T H x i φ(i), v = A e (u, v). e T, e E φ (i) ϕ y i φ(i). A e (u, v), 4

15 For te reference macroelement Ê, we get AÊ(u H, v H ) = 4 aδ(u H ), δ(v H ) 4 A Tk (u, v) = ad (k) (u), d (k) (v) T k were = area(t k ), δ(u H ) = u H, δ(v H ) = v H, d (k) (u) = u Tk, d (k) (v) = v Tk. Here T k are obtained by subdividing of Ê into four congruent triangles. It is easy to sow tat k= d () (u) = d () (u) = d () (u) = d (4) (u) = δ(u H ), d () (v) = d () (v) = d () (v) = d (4) (v) = δ(v H ). Tis gives 4AÊ(u H, v H ) = T k Ê A Tk (u, v). (8) For an arbitrary saped macroelement E, we exploit te existing affine mapping F : Ê E wit constant Jacobian DF (x) = G for a transformation of te problem. A E (u H, v H ) = ÂÊ (û H, ˆv H ) = â û H ˆv H dx were â = det(g) G ag T, û H = u H F, ˆv H = v H F are from V H (Ê), Ê A e (u, v) = A Tk (û, ˆv) = 4 â û ˆv dx, e T, e E T k Ê k= T k were â is te same as above, û = u F, ˆv = v F are from V (Ê). As (8) olds for any anisotropy coefficients on Ê, a similar identity olds for any macroelement E. 4 Concluding remarks Tis study is strongly motivated by te expanding interest in nonconforming finite elements, wic are very elpful for solving problems, were te standard conforming elements suffer from te so-called locking effects. Te success of te Crouzeix-Raviart and oter non-conforming finite elements can be explained e.g. by te fact tat tey produce algebraic systems tat are equivalent to te Scur complement system for te Lagrange multipliers arising from te mixed finite element metod for Raviart-Tomas elements (see [6]). Tere are also oter advantages of te non-conforming Crouzeix-Raviart finite elements, as less density of te stiffness matrix 5

16 etc. In our study, te class of robust AMLI algoritms (see e.g. [, 9, 9]) for conforming linear elements forms te background for generalizations addressed to tis important non-conforming case. In tis paper, we presented new estimates for te constant in te strengtened CBS inequality for Crouzeix-Raviart FEM systems. Te obtained uniform estimates wit respect to bot mes and coefficient anisotropy are te first results of tis kind for non-conforming finite elements. Note tat te obtained value of te strengtened CBS inequality constant is te same as in te case of conforming linear triangular finite elements. At tis stage of te study we underline te following issues. Te DA algoritm allows for a direct extension of te γ estimate to te multi-level case, see also Teorem.. For te FR algoritm, te obtained estimates of te CBS constant γ for te considered two-level algoritms are not directly applicable to te multilevel case. Te estimate of te CBS constant γ guarantees uniform convergence rate of te related PCG algoritm but tis is not enoug to ave a robust algoritm. Te next important step is to construct optimal order preconditioners for te systems wit te first diagonal block matrix corresponding to te current two level splitting. Tis block is well-conditioned for model type problems but becomes increasingly ill-conditioned wen te problem coefficients become strongly anisotropic or, equivalently, wen te mes aspect ratio increases. Crouzeix-Raviart non-conforming finite elements are also known as a stable discretization tool for pure displacement elasticity problem in te almost incompressible case. In te conforming case, te same estimates for γ old for bot te scalar elliptic and te elasticity problem (see e.g. [, 8, 5, 5]). Tis is not generally true in te non-conforming case. As it was sown in [7], for te model case of isosceles triangles and Crouzeix-Raviart nonconforming linear basis functions γ < 4 uniformly wit respect to te Poisson ratio ν [0, /). Tis result was obtained for FR two-level algoritm but its extention to te case of a general triangulation seems to be muc arder from tecnical point of view. We expect tat te ere presented DA algoritm based on a direct aggregation and differentiation will be better suited for generalizations to te case of pure displacement and oter elasticity problems, wic is to be addressed in a fortcoming work. 5 Acknowledgments Tis work as been supported by te Center of Excellence BIS- grant ICA Te second autor was also supported in part by te Bulgarian NSF grant I-00/000. 6

17 References [] B. Accab and J.F. Maitre, Estimate of te constant in two strengtened CBS inequalities for FEM systems of D elasticity: Application to multilevel metods and a posteriori error estimates, Numer. Linear Algebra Appl., (996), [] D.N. Arnold and F. Brezzi, Mixed and nonconforming finite element metods: implementation, postprocessing and error estimates, RAIRO Modélisation Matematique et Analyse Numérique, 9 (985), 7-. [] O. Axelsson, On Algebraic Multilevel Iteration Metods for Selfadjoint Elliptic Problems wit Anisotropy, Rend. Sem. Mat. Univers. Politec. Torino, Fascicolo Speciale, Numerical Metods, 99, -6. [4] O. Axelsson, Iterative solution metods. Cambridge University Press, 994. [5] O. Axelsson, Stabilization of algebraic multilevel iteration metods; additive metods, Numerical Algoritms, (999), -47. [6] O. Axelsson, R. Blaeta, Two simple derivations of universal bounds for te C.B.S. inequality constant, Rep. 0, Dept. of Mat., University of Nijmegen, December 00. To appear in Applications of Matematics, 00. [7] O. Axelsson and I. Gustafsson, Preconditioning and two-level multigrid metods of arbitrary degree of approximations, Matematics of Computation, 40(98), [8] O. Axelsson, S. Margenov, An optimal order multilevel preconditioner wit respect to problem and discretization parameters, Minev, Wong and Lin eds., Advances in Computations, Teory and Practice, Nova Science, Vol. 7 (00), -8. [9] O. Axelsson, A. Padiy, On te additive version of te algebraic multilevel iteration metod for anisotropic elliptic problems, SIAM Journal on Scientific Computing 0, No.5 (999), [0] O. Axelsson and P.S. Vassilevski, Algebraic Multilevel Preconditioning Metods I, Numerisce Matematik, 56 (989), [] O. Axelsson and P.S. Vassilevski, Algebraic Multilevel Preconditioning Metods II, SIAM Journal on Numerical Analysis, 7 (990), [] R. Bank and T. Dupont, An Optimal Order Process for Solving Finite Element Equations, Matematics of Computation, 6 (98), [] R. Bank, T. Dupont and H. Yserentant, Te Hierarcical Basis Multigrid Metod, Numerisce Matematik, 5 (988),

18 [4] R. Blaeta, Investigation of multi-level iterative metods and te strengtened CBS inequality, Tecnical Report BIS-, CLPP BAS Sofia 00. [5] R. Blaeta, Nested tetraedral grids and strengtened C.B.S. inequality, submitted to Numerical Linear Algebra wit Applications. [6] V. Eijkout and P.S. Vassilevski, Te Role of te Strengtened Caucy-Bunyakowski- Scwarz Inequality in Multilevel Metods, SIAM Review, (99), [7] Tz. Kolev, S. Margenov, Two-level preconditioning of pure displacement nonconforming FEM systems, Numerical Linear Algebra wit Applications, 6 (999), [8] S. Margenov, Upper bound of te constant in te strengtened C.B.S. inequality for FEM D elasticity equations, Numerical Linear Algebra wit Applications, (994), [9] S. Margenov, P.S. Vassilevski, Algebraic multilevel preconditioning of anisotropic elliptic problems, SIAM Journal on Scientific Computing, V.5(5) (994), [0] S. Margenov, P.S. Vassilevski, Two-level preconditioning of non-conforming FEM systems, Griebel, Iliev, Margenov, Vassilevski, eds., Large-Scale Scientific Computations of Engineering and Environmental Problems, Notes on Numerical Fluid Mecanics, V 6, VIEWEG (998),

Optimal multilevel preconditioning of strongly anisotropic problems.part II: non-conforming FEM. p. 1/36

Optimal multilevel preconditioning of strongly anisotropic problems.part II: non-conforming FEM. p. 1/36 Optimal multilevel preconditioning of strongly anisotropic problems. Part II: non-conforming FEM. Svetozar Margenov margenov@parallel.bas.bg Institute for Parallel Processing, Bulgarian Academy of Sciences,

More information

Multilevel Preconditioning of Graph-Laplacians: Polynomial Approximation of the Pivot Blocks Inverses

Multilevel Preconditioning of Graph-Laplacians: Polynomial Approximation of the Pivot Blocks Inverses Multilevel Preconditioning of Graph-Laplacians: Polynomial Approximation of the Pivot Blocks Inverses P. Boyanova 1, I. Georgiev 34, S. Margenov, L. Zikatanov 5 1 Uppsala University, Box 337, 751 05 Uppsala,

More information

Preconditioning in H(div) and Applications

Preconditioning in H(div) and Applications 1 Preconditioning in H(div) and Applications Douglas N. Arnold 1, Ricard S. Falk 2 and Ragnar Winter 3 4 Abstract. Summarizing te work of [AFW97], we sow ow to construct preconditioners using domain decomposition

More information

arxiv: v1 [math.na] 28 Apr 2017

arxiv: v1 [math.na] 28 Apr 2017 THE SCOTT-VOGELIUS FINITE ELEMENTS REVISITED JOHNNY GUZMÁN AND L RIDGWAY SCOTT arxiv:170500020v1 [matna] 28 Apr 2017 Abstract We prove tat te Scott-Vogelius finite elements are inf-sup stable on sape-regular

More information

Differentiation in higher dimensions

Differentiation in higher dimensions Capter 2 Differentiation in iger dimensions 2.1 Te Total Derivative Recall tat if f : R R is a 1-variable function, and a R, we say tat f is differentiable at x = a if and only if te ratio f(a+) f(a) tends

More information

A Hybrid Mixed Discontinuous Galerkin Finite Element Method for Convection-Diffusion Problems

A Hybrid Mixed Discontinuous Galerkin Finite Element Method for Convection-Diffusion Problems A Hybrid Mixed Discontinuous Galerkin Finite Element Metod for Convection-Diffusion Problems Herbert Egger Joacim Scöberl We propose and analyse a new finite element metod for convection diffusion problems

More information

Comparative Analysis of Mesh Generators and MIC(0) Preconditioning of FEM Elasticity Systems

Comparative Analysis of Mesh Generators and MIC(0) Preconditioning of FEM Elasticity Systems Comparative Analysis of Mesh Generators and MIC(0) Preconditioning of FEM Elasticity Systems Nikola Kosturski and Svetozar Margenov Institute for Parallel Processing, Bulgarian Academy of Sciences Abstract.

More information

The Laplace equation, cylindrically or spherically symmetric case

The Laplace equation, cylindrically or spherically symmetric case Numerisce Metoden II, 7 4, und Übungen, 7 5 Course Notes, Summer Term 7 Some material and exercises Te Laplace equation, cylindrically or sperically symmetric case Electric and gravitational potential,

More information

lecture 26: Richardson extrapolation

lecture 26: Richardson extrapolation 43 lecture 26: Ricardson extrapolation 35 Ricardson extrapolation, Romberg integration Trougout numerical analysis, one encounters procedures tat apply some simple approximation (eg, linear interpolation)

More information

1. Introduction. We consider the model problem: seeking an unknown function u satisfying

1. Introduction. We consider the model problem: seeking an unknown function u satisfying A DISCONTINUOUS LEAST-SQUARES FINITE ELEMENT METHOD FOR SECOND ORDER ELLIPTIC EQUATIONS XIU YE AND SHANGYOU ZHANG Abstract In tis paper, a discontinuous least-squares (DLS) finite element metod is introduced

More information

Polynomial Interpolation

Polynomial Interpolation Capter 4 Polynomial Interpolation In tis capter, we consider te important problem of approximatinga function fx, wose values at a set of distinct points x, x, x,, x n are known, by a polynomial P x suc

More information

New Streamfunction Approach for Magnetohydrodynamics

New Streamfunction Approach for Magnetohydrodynamics New Streamfunction Approac for Magnetoydrodynamics Kab Seo Kang Brooaven National Laboratory, Computational Science Center, Building 63, Room, Upton NY 973, USA. sang@bnl.gov Summary. We apply te finite

More information

Poisson Equation in Sobolev Spaces

Poisson Equation in Sobolev Spaces Poisson Equation in Sobolev Spaces OcMountain Dayligt Time. 6, 011 Today we discuss te Poisson equation in Sobolev spaces. It s existence, uniqueness, and regularity. Weak Solution. u = f in, u = g on

More information

A Mixed-Hybrid-Discontinuous Galerkin Finite Element Method for Convection-Diffusion Problems

A Mixed-Hybrid-Discontinuous Galerkin Finite Element Method for Convection-Diffusion Problems A Mixed-Hybrid-Discontinuous Galerkin Finite Element Metod for Convection-Diffusion Problems Herbert Egger Joacim Scöberl We propose and analyse a new finite element metod for convection diffusion problems

More information

Crouzeix-Velte Decompositions and the Stokes Problem

Crouzeix-Velte Decompositions and the Stokes Problem Crouzeix-Velte Decompositions and te Stokes Problem PD Tesis Strauber Györgyi Eötvös Loránd University of Sciences, Insitute of Matematics, Matematical Doctoral Scool Director of te Doctoral Scool: Dr.

More information

COMPUTATIONAL COMPARISON BETWEEN THE TAYLOR HOOD AND THE CONFORMING CROUZEIX RAVIART ELEMENT

COMPUTATIONAL COMPARISON BETWEEN THE TAYLOR HOOD AND THE CONFORMING CROUZEIX RAVIART ELEMENT Proceedings of ALGORITMY 5 pp. 369 379 COMPUTATIONAL COMPARISON BETWEEN E TAYLOR HOOD AND E CONFORMING OUZEIX RAVIART ELEMENT ROLF KRAHL AND EBERHARD BÄNSCH Abstract. Tis paper is concerned wit te computational

More information

Polynomial Interpolation

Polynomial Interpolation Capter 4 Polynomial Interpolation In tis capter, we consider te important problem of approximating a function f(x, wose values at a set of distinct points x, x, x 2,,x n are known, by a polynomial P (x

More information

AN ANALYSIS OF THE EMBEDDED DISCONTINUOUS GALERKIN METHOD FOR SECOND ORDER ELLIPTIC PROBLEMS

AN ANALYSIS OF THE EMBEDDED DISCONTINUOUS GALERKIN METHOD FOR SECOND ORDER ELLIPTIC PROBLEMS AN ANALYSIS OF THE EMBEDDED DISCONTINUOUS GALERKIN METHOD FOR SECOND ORDER ELLIPTIC PROBLEMS BERNARDO COCKBURN, JOHNNY GUZMÁN, SEE-CHEW SOON, AND HENRYK K. STOLARSKI Abstract. Te embedded discontinuous

More information

arxiv: v1 [math.na] 6 Dec 2010

arxiv: v1 [math.na] 6 Dec 2010 MULTILEVEL PRECONDITIONERS FOR DISCONTINUOUS GALERKIN APPROXIMATIONS OF ELLIPTIC PROBLEMS WITH JUMP COEFFICIENTS BLANCA AYUSO DE DIOS, MICHAEL HOLST, YUNRONG ZHU, AND LUDMIL ZIKATANOV arxiv:1012.1287v1

More information

A method of Lagrange Galerkin of second order in time. Une méthode de Lagrange Galerkin d ordre deux en temps

A method of Lagrange Galerkin of second order in time. Une méthode de Lagrange Galerkin d ordre deux en temps A metod of Lagrange Galerkin of second order in time Une métode de Lagrange Galerkin d ordre deux en temps Jocelyn Étienne a a DAMTP, University of Cambridge, Wilberforce Road, Cambridge CB3 0WA, Great-Britain.

More information

LEAST-SQUARES FINITE ELEMENT APPROXIMATIONS TO SOLUTIONS OF INTERFACE PROBLEMS

LEAST-SQUARES FINITE ELEMENT APPROXIMATIONS TO SOLUTIONS OF INTERFACE PROBLEMS SIAM J. NUMER. ANAL. c 998 Society for Industrial Applied Matematics Vol. 35, No., pp. 393 405, February 998 00 LEAST-SQUARES FINITE ELEMENT APPROXIMATIONS TO SOLUTIONS OF INTERFACE PROBLEMS YANZHAO CAO

More information

On convexity of polynomial paths and generalized majorizations

On convexity of polynomial paths and generalized majorizations On convexity of polynomial pats and generalized majorizations Marija Dodig Centro de Estruturas Lineares e Combinatórias, CELC, Universidade de Lisboa, Av. Prof. Gama Pinto 2, 1649-003 Lisboa, Portugal

More information

Numerical Experiments Using MATLAB: Superconvergence of Nonconforming Finite Element Approximation for Second-Order Elliptic Problems

Numerical Experiments Using MATLAB: Superconvergence of Nonconforming Finite Element Approximation for Second-Order Elliptic Problems Applied Matematics, 06, 7, 74-8 ttp://wwwscirporg/journal/am ISSN Online: 5-7393 ISSN Print: 5-7385 Numerical Experiments Using MATLAB: Superconvergence of Nonconforming Finite Element Approximation for

More information

M. Croci (Oxford), M. B. Giles (Oxford), P. E. Farrell (Oxford), M. E. Rognes (Simula) MCQMC July 4, 2018

M. Croci (Oxford), M. B. Giles (Oxford), P. E. Farrell (Oxford), M. E. Rognes (Simula) MCQMC July 4, 2018 Efficient wite noise sampling and coupling for multilevel Monte Carlo M. Croci (Oxford), M. B. Giles (Oxford), P. E. Farrell (Oxford), M. E. Rognes (Simula) MCQMC2018 - July 4, 2018 Overview Introduction

More information

Dedicated to the 70th birthday of Professor Lin Qun

Dedicated to the 70th birthday of Professor Lin Qun Journal of Computational Matematics, Vol.4, No.3, 6, 4 44. ACCELERATION METHODS OF NONLINEAR ITERATION FOR NONLINEAR PARABOLIC EQUATIONS Guang-wei Yuan Xu-deng Hang Laboratory of Computational Pysics,

More information

A MULTILEVEL PRECONDITIONER FOR THE INTERIOR PENALTY DISCONTINUOUS GALERKIN METHOD

A MULTILEVEL PRECONDITIONER FOR THE INTERIOR PENALTY DISCONTINUOUS GALERKIN METHOD A MULTILEVEL PRECONDITIONER FOR THE INTERIOR PENALTY DISCONTINUOUS GALERKIN METHOD KOLJA BRIX, MARTIN CAMPOS PINTO, AND WOLFGANG DAHMEN Abstract. In tis article we present a multilevel preconditioner for

More information

arxiv: v2 [math.na] 5 Jul 2017

arxiv: v2 [math.na] 5 Jul 2017 Trace Finite Element Metods for PDEs on Surfaces Maxim A. Olsanskii and Arnold Reusken arxiv:1612.00054v2 [mat.na] 5 Jul 2017 Abstract In tis paper we consider a class of unfitted finite element metods

More information

ROBUST MULTISCALE ITERATIVE SOLVERS FOR NONLINEAR FLOWS IN HIGHLY HETEROGENEOUS MEDIA

ROBUST MULTISCALE ITERATIVE SOLVERS FOR NONLINEAR FLOWS IN HIGHLY HETEROGENEOUS MEDIA ROBUST MULTISCALE ITERATIVE SOLVERS FOR NONLINEAR FLOWS IN HIGHLY HETEROGENEOUS MEDIA Y. EFENDIEV, J. GALVIS, S. KI KANG, AND R.D. LAZAROV Abstract. In tis paper, we study robust iterative solvers for

More information

232 Calculus and Structures

232 Calculus and Structures 3 Calculus and Structures CHAPTER 17 JUSTIFICATION OF THE AREA AND SLOPE METHODS FOR EVALUATING BEAMS Calculus and Structures 33 Copyrigt Capter 17 JUSTIFICATION OF THE AREA AND SLOPE METHODS 17.1 THE

More information

THE IDEA OF DIFFERENTIABILITY FOR FUNCTIONS OF SEVERAL VARIABLES Math 225

THE IDEA OF DIFFERENTIABILITY FOR FUNCTIONS OF SEVERAL VARIABLES Math 225 THE IDEA OF DIFFERENTIABILITY FOR FUNCTIONS OF SEVERAL VARIABLES Mat 225 As we ave seen, te definition of derivative for a Mat 111 function g : R R and for acurveγ : R E n are te same, except for interpretation:

More information

MIXED DISCONTINUOUS GALERKIN APPROXIMATION OF THE MAXWELL OPERATOR. SIAM J. Numer. Anal., Vol. 42 (2004), pp

MIXED DISCONTINUOUS GALERKIN APPROXIMATION OF THE MAXWELL OPERATOR. SIAM J. Numer. Anal., Vol. 42 (2004), pp MIXED DISCONTINUOUS GALERIN APPROXIMATION OF THE MAXWELL OPERATOR PAUL HOUSTON, ILARIA PERUGIA, AND DOMINI SCHÖTZAU SIAM J. Numer. Anal., Vol. 4 (004), pp. 434 459 Abstract. We introduce and analyze a

More information

Analysis of A Continuous Finite Element Method for H(curl, div)-elliptic Interface Problem

Analysis of A Continuous Finite Element Method for H(curl, div)-elliptic Interface Problem Analysis of A Continuous inite Element Metod for Hcurl, div)-elliptic Interface Problem Huoyuan Duan, Ping Lin, and Roger C. E. Tan Abstract In tis paper, we develop a continuous finite element metod for

More information

Variational Localizations of the Dual Weighted Residual Estimator

Variational Localizations of the Dual Weighted Residual Estimator Publised in Journal for Computational and Applied Matematics, pp. 192-208, 2015 Variational Localizations of te Dual Weigted Residual Estimator Tomas Ricter Tomas Wick Te dual weigted residual metod (DWR)

More information

Order of Accuracy. ũ h u Ch p, (1)

Order of Accuracy. ũ h u Ch p, (1) Order of Accuracy 1 Terminology We consider a numerical approximation of an exact value u. Te approximation depends on a small parameter, wic can be for instance te grid size or time step in a numerical

More information

MANY scientific and engineering problems can be

MANY scientific and engineering problems can be A Domain Decomposition Metod using Elliptical Arc Artificial Boundary for Exterior Problems Yajun Cen, and Qikui Du Abstract In tis paper, a Diriclet-Neumann alternating metod using elliptical arc artificial

More information

LECTURE 14 NUMERICAL INTEGRATION. Find

LECTURE 14 NUMERICAL INTEGRATION. Find LECTURE 14 NUMERCAL NTEGRATON Find b a fxdx or b a vx ux fx ydy dx Often integration is required. However te form of fx may be suc tat analytical integration would be very difficult or impossible. Use

More information

A CLASS OF EVEN DEGREE SPLINES OBTAINED THROUGH A MINIMUM CONDITION

A CLASS OF EVEN DEGREE SPLINES OBTAINED THROUGH A MINIMUM CONDITION STUDIA UNIV. BABEŞ BOLYAI, MATHEMATICA, Volume XLVIII, Number 3, September 2003 A CLASS OF EVEN DEGREE SPLINES OBTAINED THROUGH A MINIMUM CONDITION GH. MICULA, E. SANTI, AND M. G. CIMORONI Dedicated to

More information

Mass Lumping for Constant Density Acoustics

Mass Lumping for Constant Density Acoustics Lumping 1 Mass Lumping for Constant Density Acoustics William W. Symes ABSTRACT Mass lumping provides an avenue for efficient time-stepping of time-dependent problems wit conforming finite element spatial

More information

Multigrid Methods for Obstacle Problems

Multigrid Methods for Obstacle Problems Multigrid Metods for Obstacle Problems by Cunxiao Wu A Researc Paper presented to te University of Waterloo in partial fulfillment of te requirement for te degree of Master of Matematics in Computational

More information

A SHORT INTRODUCTION TO BANACH LATTICES AND

A SHORT INTRODUCTION TO BANACH LATTICES AND CHAPTER A SHORT INTRODUCTION TO BANACH LATTICES AND POSITIVE OPERATORS In tis capter we give a brief introduction to Banac lattices and positive operators. Most results of tis capter can be found, e.g.,

More information

Different Approaches to a Posteriori Error Analysis of the Discontinuous Galerkin Method

Different Approaches to a Posteriori Error Analysis of the Discontinuous Galerkin Method WDS'10 Proceedings of Contributed Papers, Part I, 151 156, 2010. ISBN 978-80-7378-139-2 MATFYZPRESS Different Approaces to a Posteriori Error Analysis of te Discontinuous Galerkin Metod I. Šebestová Carles

More information

Robust approximation error estimates and multigrid solvers for isogeometric multi-patch discretizations

Robust approximation error estimates and multigrid solvers for isogeometric multi-patch discretizations www.oeaw.ac.at Robust approximation error estimates and multigrid solvers for isogeometric multi-patc discretizations S. Takacs RICAM-Report 2017-32 www.ricam.oeaw.ac.at Robust approximation error estimates

More information

Journal of Computational and Applied Mathematics

Journal of Computational and Applied Mathematics Journal of Computational and Applied Matematics 94 (6) 75 96 Contents lists available at ScienceDirect Journal of Computational and Applied Matematics journal omepage: www.elsevier.com/locate/cam Smootness-Increasing

More information

LIMITS AND DERIVATIVES CONDITIONS FOR THE EXISTENCE OF A LIMIT

LIMITS AND DERIVATIVES CONDITIONS FOR THE EXISTENCE OF A LIMIT LIMITS AND DERIVATIVES Te limit of a function is defined as te value of y tat te curve approaces, as x approaces a particular value. Te limit of f (x) as x approaces a is written as f (x) approaces, as

More information

5 Ordinary Differential Equations: Finite Difference Methods for Boundary Problems

5 Ordinary Differential Equations: Finite Difference Methods for Boundary Problems 5 Ordinary Differential Equations: Finite Difference Metods for Boundary Problems Read sections 10.1, 10.2, 10.4 Review questions 10.1 10.4, 10.8 10.9, 10.13 5.1 Introduction In te previous capters we

More information

A = h w (1) Error Analysis Physics 141

A = h w (1) Error Analysis Physics 141 Introduction In all brances of pysical science and engineering one deals constantly wit numbers wic results more or less directly from experimental observations. Experimental observations always ave inaccuracies.

More information

Numerical Differentiation

Numerical Differentiation Numerical Differentiation Finite Difference Formulas for te first derivative (Using Taylor Expansion tecnique) (section 8.3.) Suppose tat f() = g() is a function of te variable, and tat as 0 te function

More information

WYSE Academic Challenge 2004 Sectional Mathematics Solution Set

WYSE Academic Challenge 2004 Sectional Mathematics Solution Set WYSE Academic Callenge 00 Sectional Matematics Solution Set. Answer: B. Since te equation can be written in te form x + y, we ave a major 5 semi-axis of lengt 5 and minor semi-axis of lengt. Tis means

More information

arxiv: v1 [math.na] 17 Jul 2014

arxiv: v1 [math.na] 17 Jul 2014 Div First-Order System LL* FOSLL* for Second-Order Elliptic Partial Differential Equations Ziqiang Cai Rob Falgout Sun Zang arxiv:1407.4558v1 [mat.na] 17 Jul 2014 February 13, 2018 Abstract. Te first-order

More information

A SADDLE POINT LEAST SQUARES APPROACH TO MIXED METHODS

A SADDLE POINT LEAST SQUARES APPROACH TO MIXED METHODS A SADDLE POINT LEAST SQUARES APPROACH TO MIXED METHODS CONSTANTIN BACUTA AND KLAJDI QIRKO Abstract. We investigate new PDE discretization approaces for solving variational formulations wit different types

More information

Chapter 5 FINITE DIFFERENCE METHOD (FDM)

Chapter 5 FINITE DIFFERENCE METHOD (FDM) MEE7 Computer Modeling Tecniques in Engineering Capter 5 FINITE DIFFERENCE METHOD (FDM) 5. Introduction to FDM Te finite difference tecniques are based upon approximations wic permit replacing differential

More information

arxiv: v1 [math.na] 20 Jul 2009

arxiv: v1 [math.na] 20 Jul 2009 STABILITY OF LAGRANGE ELEMENTS FOR THE MIXED LAPLACIAN DOUGLAS N. ARNOLD AND MARIE E. ROGNES arxiv:0907.3438v1 [mat.na] 20 Jul 2009 Abstract. Te stability properties of simple element coices for te mixed

More information

POLYNOMIAL AND SPLINE ESTIMATORS OF THE DISTRIBUTION FUNCTION WITH PRESCRIBED ACCURACY

POLYNOMIAL AND SPLINE ESTIMATORS OF THE DISTRIBUTION FUNCTION WITH PRESCRIBED ACCURACY APPLICATIONES MATHEMATICAE 36, (29), pp. 2 Zbigniew Ciesielski (Sopot) Ryszard Zieliński (Warszawa) POLYNOMIAL AND SPLINE ESTIMATORS OF THE DISTRIBUTION FUNCTION WITH PRESCRIBED ACCURACY Abstract. Dvoretzky

More information

Notes on Multigrid Methods

Notes on Multigrid Methods Notes on Multigrid Metods Qingai Zang April, 17 Motivation of multigrids. Te convergence rates of classical iterative metod depend on te grid spacing, or problem size. In contrast, convergence rates of

More information

How to Find the Derivative of a Function: Calculus 1

How to Find the Derivative of a Function: Calculus 1 Introduction How to Find te Derivative of a Function: Calculus 1 Calculus is not an easy matematics course Te fact tat you ave enrolled in suc a difficult subject indicates tat you are interested in te

More information

Overlapping domain decomposition methods for elliptic quasi-variational inequalities related to impulse control problem with mixed boundary conditions

Overlapping domain decomposition methods for elliptic quasi-variational inequalities related to impulse control problem with mixed boundary conditions Proc. Indian Acad. Sci. (Mat. Sci.) Vol. 121, No. 4, November 2011, pp. 481 493. c Indian Academy of Sciences Overlapping domain decomposition metods for elliptic quasi-variational inequalities related

More information

HYBRIDIZED GLOBALLY DIVERGENCE-FREE LDG METHODS. PART I: THE STOKES PROBLEM

HYBRIDIZED GLOBALLY DIVERGENCE-FREE LDG METHODS. PART I: THE STOKES PROBLEM MATHEMATICS OF COMPUTATION Volume 75, Number 254, Pages 533 563 S 0025-5718(05)01804-1 Article electronically publised on December 16, 2005 HYBRIDIZED GLOBALLY DIVERGENCE-FREE LDG METHODS. PART I: THE

More information

3 Parabolic Differential Equations

3 Parabolic Differential Equations 3 Parabolic Differential Equations 3.1 Classical solutions We consider existence and uniqueness results for initial-boundary value problems for te linear eat equation in Q := Ω (, T ), were Ω is a bounded

More information

MATH745 Fall MATH745 Fall

MATH745 Fall MATH745 Fall MATH745 Fall 5 MATH745 Fall 5 INTRODUCTION WELCOME TO MATH 745 TOPICS IN NUMERICAL ANALYSIS Instructor: Dr Bartosz Protas Department of Matematics & Statistics Email: bprotas@mcmasterca Office HH 36, Ext

More information

arxiv: v1 [math.na] 3 Nov 2011

arxiv: v1 [math.na] 3 Nov 2011 arxiv:.983v [mat.na] 3 Nov 2 A Finite Difference Gost-cell Multigrid approac for Poisson Equation wit mixed Boundary Conditions in Arbitrary Domain Armando Coco, Giovanni Russo November 7, 2 Abstract In

More information

c 2004 Society for Industrial and Applied Mathematics

c 2004 Society for Industrial and Applied Mathematics SIAM J NUMER ANAL Vol 4, No, pp 86 84 c 004 Society for Industrial and Applied Matematics LEAST-SQUARES METHODS FOR LINEAR ELASTICITY ZHIQIANG CAI AND GERHARD STARKE Abstract Tis paper develops least-squares

More information

arxiv: v1 [math.na] 27 Jan 2014

arxiv: v1 [math.na] 27 Jan 2014 L 2 -ERROR ESTIMATES FOR FINITE ELEMENT APPROXIMATIONS OF BOUNDARY FLUXES MATS G. LARSON AND ANDRÉ MASSING arxiv:1401.6994v1 [mat.na] 27 Jan 2014 Abstract. We prove quasi-optimal a priori error estimates

More information

More on generalized inverses of partitioned matrices with Banachiewicz-Schur forms

More on generalized inverses of partitioned matrices with Banachiewicz-Schur forms More on generalized inverses of partitioned matrices wit anaciewicz-scur forms Yongge Tian a,, Yosio Takane b a Cina Economics and Management cademy, Central University of Finance and Economics, eijing,

More information

ADAPTIVE MULTILEVEL INEXACT SQP METHODS FOR PDE CONSTRAINED OPTIMIZATION

ADAPTIVE MULTILEVEL INEXACT SQP METHODS FOR PDE CONSTRAINED OPTIMIZATION ADAPTIVE MULTILEVEL INEXACT SQP METHODS FOR PDE CONSTRAINED OPTIMIZATION J CARSTEN ZIEMS AND STEFAN ULBRICH Abstract We present a class of inexact adaptive multilevel trust-region SQP-metods for te efficient

More information

Gradient Descent etc.

Gradient Descent etc. 1 Gradient Descent etc EE 13: Networked estimation and control Prof Kan) I DERIVATIVE Consider f : R R x fx) Te derivative is defined as d fx) = lim dx fx + ) fx) Te cain rule states tat if d d f gx) )

More information

4. The slope of the line 2x 7y = 8 is (a) 2/7 (b) 7/2 (c) 2 (d) 2/7 (e) None of these.

4. The slope of the line 2x 7y = 8 is (a) 2/7 (b) 7/2 (c) 2 (d) 2/7 (e) None of these. Mat 11. Test Form N Fall 016 Name. Instructions. Te first eleven problems are wort points eac. Te last six problems are wort 5 points eac. For te last six problems, you must use relevant metods of algebra

More information

Lecture 15. Interpolation II. 2 Piecewise polynomial interpolation Hermite splines

Lecture 15. Interpolation II. 2 Piecewise polynomial interpolation Hermite splines Lecture 5 Interpolation II Introduction In te previous lecture we focused primarily on polynomial interpolation of a set of n points. A difficulty we observed is tat wen n is large, our polynomial as to

More information

arxiv: v1 [math.na] 11 May 2018

arxiv: v1 [math.na] 11 May 2018 Nitsce s metod for unilateral contact problems arxiv:1805.04283v1 [mat.na] 11 May 2018 Tom Gustafsson, Rolf Stenberg and Jua Videman May 14, 2018 Abstract We derive optimal a priori and a posteriori error

More information

Combining functions: algebraic methods

Combining functions: algebraic methods Combining functions: algebraic metods Functions can be added, subtracted, multiplied, divided, and raised to a power, just like numbers or algebra expressions. If f(x) = x 2 and g(x) = x + 2, clearly f(x)

More information

arxiv: v1 [math.na] 7 Mar 2019

arxiv: v1 [math.na] 7 Mar 2019 Local Fourier analysis for mixed finite-element metods for te Stokes equations Yunui He a,, Scott P. MacLaclan a a Department of Matematics and Statistics, Memorial University of Newfoundland, St. Jon

More information

Convergence and Descent Properties for a Class of Multilevel Optimization Algorithms

Convergence and Descent Properties for a Class of Multilevel Optimization Algorithms Convergence and Descent Properties for a Class of Multilevel Optimization Algoritms Stepen G. Nas April 28, 2010 Abstract I present a multilevel optimization approac (termed MG/Opt) for te solution of

More information

A UNIFORM INF SUP CONDITION WITH APPLICATIONS TO PRECONDITIONING

A UNIFORM INF SUP CONDITION WITH APPLICATIONS TO PRECONDITIONING A UNIFORM INF SUP CONDIION WIH APPLICAIONS O PRECONDIIONING KEN ANDRE MARDAL, JOACHIM SCHÖBERL, AND RAGNAR WINHER Abstract. A uniform inf sup condition related to a parameter dependent Stokes problem is

More information

Error estimates for a semi-implicit fully discrete finite element scheme for the mean curvature flow of graphs

Error estimates for a semi-implicit fully discrete finite element scheme for the mean curvature flow of graphs Interfaces and Free Boundaries 2, 2000 34 359 Error estimates for a semi-implicit fully discrete finite element sceme for te mean curvature flow of graps KLAUS DECKELNICK Scool of Matematical Sciences,

More information

arxiv: v1 [math.na] 11 Jul 2011

arxiv: v1 [math.na] 11 Jul 2011 Multigrid Preconditioner for Nonconforming Discretization of Elliptic Problems with Jump Coefficients arxiv:07.260v [math.na] Jul 20 Blanca Ayuso De Dios, Michael Holst 2, Yunrong Zhu 2, and Ludmil Zikatanov

More information

AN ANALYSIS OF NEW FINITE ELEMENT SPACES FOR MAXWELL S EQUATIONS

AN ANALYSIS OF NEW FINITE ELEMENT SPACES FOR MAXWELL S EQUATIONS Journal of Matematical Sciences: Advances and Applications Volume 5 8 Pages -9 Available at ttp://scientificadvances.co.in DOI: ttp://d.doi.org/.864/jmsaa_7975 AN ANALYSIS OF NEW FINITE ELEMENT SPACES

More information

THE INF-SUP STABILITY OF THE LOWEST ORDER TAYLOR-HOOD PAIR ON ANISOTROPIC MESHES arxiv: v1 [math.na] 21 Oct 2017

THE INF-SUP STABILITY OF THE LOWEST ORDER TAYLOR-HOOD PAIR ON ANISOTROPIC MESHES arxiv: v1 [math.na] 21 Oct 2017 THE INF-SUP STABILITY OF THE LOWEST ORDER TAYLOR-HOOD PAIR ON ANISOTROPIC MESHES arxiv:1710.07857v1 [mat.na] 21 Oct 2017 GABRIEL R. BARRENECHEA AND ANDREAS WACHTEL Abstract. Uniform LBB conditions are

More information

APPROXIMATION OF CRYSTALLINE DENDRITE GROWTH IN TWO SPACE DIMENSIONS. Introduction

APPROXIMATION OF CRYSTALLINE DENDRITE GROWTH IN TWO SPACE DIMENSIONS. Introduction Acta Mat. Univ. Comenianae Vol. LXVII, 1(1998), pp. 57 68 57 APPROXIMATION OF CRYSTALLINE DENDRITE GROWTH IN TWO SPACE DIMENSIONS A. SCHMIDT Abstract. Te pase transition between solid and liquid in an

More information

Multigrid Methods for Discretized PDE Problems

Multigrid Methods for Discretized PDE Problems Towards Metods for Discretized PDE Problems Institute for Applied Matematics University of Heidelberg Feb 1-5, 2010 Towards Outline A model problem Solution of very large linear systems Iterative Metods

More information

ON THE CONVERGENCE OF A DUAL-PRIMAL SUBSTRUCTURING METHOD. January 2000 Revised April 2000

ON THE CONVERGENCE OF A DUAL-PRIMAL SUBSTRUCTURING METHOD. January 2000 Revised April 2000 ON THE CONVERGENCE OF A DUAL-PRIMAL SUBSTRUCTURING METHOD JAN MANDEL AND RADEK TEZAUR January 2000 Revised April 2000 Abstract In te Dual-Primal FETI metod, introduced by Farat et al [5], te domain is

More information

NUMERICAL DIFFERENTIATION. James T. Smith San Francisco State University. In calculus classes, you compute derivatives algebraically: for example,

NUMERICAL DIFFERENTIATION. James T. Smith San Francisco State University. In calculus classes, you compute derivatives algebraically: for example, NUMERICAL DIFFERENTIATION James T Smit San Francisco State University In calculus classes, you compute derivatives algebraically: for example, f( x) = x + x f ( x) = x x Tis tecnique requires your knowing

More information

Superconvergence of energy-conserving discontinuous Galerkin methods for. linear hyperbolic equations. Abstract

Superconvergence of energy-conserving discontinuous Galerkin methods for. linear hyperbolic equations. Abstract Superconvergence of energy-conserving discontinuous Galerkin metods for linear yperbolic equations Yong Liu, Ci-Wang Su and Mengping Zang 3 Abstract In tis paper, we study superconvergence properties of

More information

Computing eigenvalues and eigenfunctions of Schrödinger equations using a model reduction approach

Computing eigenvalues and eigenfunctions of Schrödinger equations using a model reduction approach Computing eigenvalues and eigenfunctions of Scrödinger equations using a model reduction approac Suangping Li 1, Ziwen Zang 2 1 Program in Applied and Computational Matematics, Princeton University, New

More information

A Finite Element Primer

A Finite Element Primer A Finite Element Primer David J. Silvester Scool of Matematics, University of Mancester d.silvester@mancester.ac.uk. Version.3 updated 4 October Contents A Model Diffusion Problem.................... x.

More information

Finite Element Methods for Linear Elasticity

Finite Element Methods for Linear Elasticity Finite Element Metods for Linear Elasticity Ricard S. Falk Department of Matematics - Hill Center Rutgers, Te State University of New Jersey 110 Frelinguysen Rd., Piscataway, NJ 08854-8019 falk@mat.rutgers.edu

More information

The Verlet Algorithm for Molecular Dynamics Simulations

The Verlet Algorithm for Molecular Dynamics Simulations Cemistry 380.37 Fall 2015 Dr. Jean M. Standard November 9, 2015 Te Verlet Algoritm for Molecular Dynamics Simulations Equations of motion For a many-body system consisting of N particles, Newton's classical

More information

arxiv: v1 [math.na] 9 Sep 2015

arxiv: v1 [math.na] 9 Sep 2015 arxiv:509.02595v [mat.na] 9 Sep 205 An Expandable Local and Parallel Two-Grid Finite Element Sceme Yanren ou, GuangZi Du Abstract An expandable local and parallel two-grid finite element sceme based on

More information

Solutions to the Multivariable Calculus and Linear Algebra problems on the Comprehensive Examination of January 31, 2014

Solutions to the Multivariable Calculus and Linear Algebra problems on the Comprehensive Examination of January 31, 2014 Solutions to te Multivariable Calculus and Linear Algebra problems on te Compreensive Examination of January 3, 24 Tere are 9 problems ( points eac, totaling 9 points) on tis portion of te examination.

More information

1 The concept of limits (p.217 p.229, p.242 p.249, p.255 p.256) 1.1 Limits Consider the function determined by the formula 3. x since at this point

1 The concept of limits (p.217 p.229, p.242 p.249, p.255 p.256) 1.1 Limits Consider the function determined by the formula 3. x since at this point MA00 Capter 6 Calculus and Basic Linear Algebra I Limits, Continuity and Differentiability Te concept of its (p.7 p.9, p.4 p.49, p.55 p.56). Limits Consider te function determined by te formula f Note

More information

University Mathematics 2

University Mathematics 2 University Matematics 2 1 Differentiability In tis section, we discuss te differentiability of functions. Definition 1.1 Differentiable function). Let f) be a function. We say tat f is differentiable at

More information

ERROR BOUNDS FOR THE METHODS OF GLIMM, GODUNOV AND LEVEQUE BRADLEY J. LUCIER*

ERROR BOUNDS FOR THE METHODS OF GLIMM, GODUNOV AND LEVEQUE BRADLEY J. LUCIER* EO BOUNDS FO THE METHODS OF GLIMM, GODUNOV AND LEVEQUE BADLEY J. LUCIE* Abstract. Te expected error in L ) attimet for Glimm s sceme wen applied to a scalar conservation law is bounded by + 2 ) ) /2 T

More information

Discontinuous Galerkin Methods for Relativistic Vlasov-Maxwell System

Discontinuous Galerkin Methods for Relativistic Vlasov-Maxwell System Discontinuous Galerkin Metods for Relativistic Vlasov-Maxwell System He Yang and Fengyan Li December 1, 16 Abstract e relativistic Vlasov-Maxwell (RVM) system is a kinetic model tat describes te dynamics

More information

Parameter Fitted Scheme for Singularly Perturbed Delay Differential Equations

Parameter Fitted Scheme for Singularly Perturbed Delay Differential Equations International Journal of Applied Science and Engineering 2013. 11, 4: 361-373 Parameter Fitted Sceme for Singularly Perturbed Delay Differential Equations Awoke Andargiea* and Y. N. Reddyb a b Department

More information

Volume 29, Issue 3. Existence of competitive equilibrium in economies with multi-member households

Volume 29, Issue 3. Existence of competitive equilibrium in economies with multi-member households Volume 29, Issue 3 Existence of competitive equilibrium in economies wit multi-member ouseolds Noriisa Sato Graduate Scool of Economics, Waseda University Abstract Tis paper focuses on te existence of

More information

1 Calculus. 1.1 Gradients and the Derivative. Q f(x+h) f(x)

1 Calculus. 1.1 Gradients and the Derivative. Q f(x+h) f(x) Calculus. Gradients and te Derivative Q f(x+) δy P T δx R f(x) 0 x x+ Let P (x, f(x)) and Q(x+, f(x+)) denote two points on te curve of te function y = f(x) and let R denote te point of intersection of

More information

Geometric Multigrid Methods

Geometric Multigrid Methods Geometric Multigrid Methods Susanne C. Brenner Department of Mathematics and Center for Computation & Technology Louisiana State University IMA Tutorial: Fast Solution Techniques November 28, 2010 Ideas

More information

ERROR ESTIMATES FOR THE DISCONTINUOUS GALERKIN METHODS FOR PARABOLIC EQUATIONS. 1. Introduction. We consider the parabolic PDE of the form,

ERROR ESTIMATES FOR THE DISCONTINUOUS GALERKIN METHODS FOR PARABOLIC EQUATIONS. 1. Introduction. We consider the parabolic PDE of the form, ERROR ESTIMATES FOR THE DISCONTINUOUS GALERKIN METHODS FOR PARABOLIC EQUATIONS K. CHRYSAFINOS AND NOEL. J. WALKINGTON Abstract. We analyze te classical discontinuous Galerkin metod for a general parabolic

More information

A proof in the finite-difference spirit of the superconvergence of the gradient for the Shortley-Weller method.

A proof in the finite-difference spirit of the superconvergence of the gradient for the Shortley-Weller method. A proof in te finite-difference spirit of te superconvergence of te gradient for te Sortley-Weller metod. L. Weynans 1 1 Team Mempis, INRIA Bordeaux-Sud-Ouest & CNRS UMR 551, Université de Bordeaux, France

More information

A Weak Galerkin Method with an Over-Relaxed Stabilization for Low Regularity Elliptic Problems

A Weak Galerkin Method with an Over-Relaxed Stabilization for Low Regularity Elliptic Problems J Sci Comput (07 7:95 8 DOI 0.007/s095-06-096-4 A Weak Galerkin Metod wit an Over-Relaxed Stabilization for Low Regularity Elliptic Problems Lunji Song, Kaifang Liu San Zao Received: April 06 / Revised:

More information

An approximation method using approximate approximations

An approximation method using approximate approximations Applicable Analysis: An International Journal Vol. 00, No. 00, September 2005, 1 13 An approximation metod using approximate approximations FRANK MÜLLER and WERNER VARNHORN, University of Kassel, Germany,

More information

Copyright c 2008 Kevin Long

Copyright c 2008 Kevin Long Lecture 4 Numerical solution of initial value problems Te metods you ve learned so far ave obtained closed-form solutions to initial value problems. A closedform solution is an explicit algebriac formula

More information