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

Size: px
Start display at page:

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

Transcription

1 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 equation. Symmetric error estimates for scemes of arbitrary order are presented. Te ideas we develop allow us to relax many assumptions freqently required in previous work. For example, we allow different discrete spaces to be used at eac time step and do not require te spatial operator to be self adjoint or independent of time. Our error estimates are posed in terms of projections of te exact solution onto te discrete spaces and are valid under te minimal regularity guaranteed by te natural energy estimate. Tese projections are local and enjoy optimal approximation properties wen te solution is sufficiently regular. 1. Introduction. We consider te parabolic PDE of te form, u t + Atu = F t, u = u. 1.1 Te operators act on Hilbert spaces related troug te standard pivot construction, U H H U, were eac embedding is continuous and dense. Ten, A. : U U is a linear map and F. U. Our goal is to analyze te classical discontinuous Galerkin DG sceme and derive fully-discrete error estimates under mimimal regularity assumptions. Te class of DG scemes we consider are classical in te sense tat te discrete solutions may be discontinuous in time but are conforming in space, i.e. are in a subspace of U at eac time. Our tecniques also apply to te more general implicit evolution equation [22, 23] Mtu t + Atu = F t, u = u, 1.2 were M. : H H is a self adjoint possitive definite operator. Te extension of our analysis to tis equation will be taken up seperatly. Te analysis below addresses te following issues wic ave not yet been adequately considered in te literature. Te operator A. may depend upon time and is not required to be self adjoint. To date te sarpest estimates for DG approximations exploit classical spectral teory for self adjoint positive definite operators, so require A to be suc an operator and to be independent of time. Wen A. is not self adjoint, multiplying 1.1 by u t does not give an estimate for te time derivative. Te subspaces of U used for te DG approximations may be different on eac time interval, t n ]. Tis adds a significant complication to te analysis Matematics subject classification: 65M6 Department of Matematics, Carnegie Mellon University, Pittsburg, PA Supported in part by National Science Foundation Grants DMS 28586, and ITR Tis work was also supported by te NSF troug te Center for Nonlinear Analysis. 1

2 wic is present even wen A =. Indeed, te first step in our analysis is to consider te DG sceme for ordinary differential equations in te Hilbert space H. Different subspaces are an essential ingredient of adaptive strategies used in conjunction wit a-posteriori error estimates to give guaranteed error bounds. Retriangulation is also necessary for many algoritms based upon a Lagrangian coordinate system; below we present suc an example. DG approximations of equations of te form 1.2 ave not been considered in te past. Below we sow tat equations of tis form arise wen Lagrangian scemes are constructed for te convection diffusion equation [6, 7, 8]. Te operator A. is not required to be strictly coercive; we only require an assumption of te form A.u, u c u 2 U C u 2 H. Here. U is a semi-norm suc tat. 2 U =. 2 U +. 2 H. Tis causes significant problems in te analysis of DG scems since te classical Gronwall argument, used for te continuous problem, fails in te discrete setting. Tis failure is due to te elemenary observation tat functions of te form χ u are not polynomial in time unless [,ˆt ˆt is a partition point, so are not available as test functions in te discrete setting. In te past tis problem as been circumvented by bounding temporal derivatives of te solution [5, 24] so tat te solution between te partition points can be controlled by te values at tese points. Tis line of argument fails for solutions aving minimal regularity. Below we circumvent tese issues by constructing polynomial approximations to te caracteristic functions χ. [,ˆt As stated above, our analysis does not require any regularity above and beyound te natural bounds tat follow from te usual energy estimate. Tis is essential for control problems were solutions of te dual problem typically will not enjoy any additional regularity. Our estimates sow tat te error can be bounded by te local truncation error of te ordinary differental equation obtained by setting A =. Tese local truncation errors can also be viewed as te approximation error of local projections of te solution onto te discrete subspaces. In our analysis we are careful to keep track of ow te various constants depend upon te coercivity constant on A.. Tis is important for te analysis of problems like te convection diffusion equation were te coercivity constant is small. We present and example of equations wic can be analyzed witin te general framework developed ere but fall outside of te teory developed, for example, in Tomée s text [24]. Convection Diffusion Equation: Te classical convection diffusion equation is ū t + V. ū ɛ ū =, and te problems tat arise wen ɛ is small are notorious. To address tese problem tis equation is sometimes considered in a Lagrangian variable. Specifically, let Ṽ be a Here χ [,ˆt is te caracteristic function equal to 1 on [, ˆt and zero oterwise. 2

3 numerical approximation of V and let x = χt, X be te cange of variables defined by te flow map associated wit Ṽ, i.e. If ut, X = ūt, xt, X ten ẋt, X = Ṽt, xt, X, x, X = X. u t + V Ṽ.F T X u ɛ1/jdiv X JF 1 F T X u =, were F ij = x i / X j is te Jacobian of te mapping and J = detf. Te natural weak problem for tis equation is ut + V Ṽ.F T X u v + ɛf T X u.f T X v J =. Ω Using te properties of determinants we find u t J = Ju t J u = Ju t J divṽu. If divṽ = ten J is constant te transformed problem takes te form of 1.1; oterwise, it takes te form of 1.2 wit M.u = Ju, and A.u = divṽuj + V Ṽ.F T X uj ɛ div X JF 1 F T X u. Tis statement of te problem generalizes te idea beind te caracteristic Galerkin sceme introduced by Douglas and Russel in [6] and DuPont in [7]. Tis cange of variables reduces te effective Peclet number, V Ṽ /ɛ, to order O1 if Ṽ is a sufficiently accurate approximation of V. Tis will eliminate many of te numerical difficulties encountered by algoritms based upon te classical statement; owever, oter problems arise. Wile te Jacobian of te transformation satisfies F, X = I, its condition number grows exponentially if Ṽ is anyting oter tan a rigid motion. In te context of a numerical sceme tis problem is circumvented by reinitalizing te transformation at eac or every few time steps. Tis reinitialization corresponds to canging te subspace for te numerical solution every few time steps. In essence, a trianglular mes in te X coordinate system will be a distorted mes in te x-coordinate system, and reinitializing te transform corresponds to projecting te solution onto a straigt sided triangular mes in te x coordinates. Tis gives rise to different subspaces at eac time step Related Results. Te discontinuous Galerkin metod was first introduced to model and simulate neutron transport by Lasaint and Raviart in [15]. Tere is an abundant literature concerning applications of te DG sceme in yperbolic problems, see e.g. [4, 14, 25] and references witin. Te DG metod for ordinary differential equations was considered by Delfour, Hager and Trocu in [5]. Tey sowed tat te DG sceme was super convergent at te partition points order 2k+2 for polynomials of degree k. Te super convergence results and better rates in te H norm use a duality argument, and space considerations do not permit us to develop te corresponding estimates in te current setting. 3

4 In te context of parabolic equations DG scemes were first analyzed for linear parabolic problems by Jamet in [13] were Ok q results were proved and by Eriksson, Jonson and Tomée in [11] were Ok 2q 1 results were proved for smoot initial data among oters. An excellent exposition of teir results and, more generally, te DG metod for parabolic equations, can be found in Tomée s book [24]. In [24] nodal and interior estimates are presented in various norms. One may also consult [18] for te analysis of a related formulation based on te backward Euler sceme. Te relation between te DG sceme and adaptive tecniques was studied in [9] and [1]. Finally, some results concerning te analysis of parabolic integro-differential equations by discontinuous Galerkin metod are presented in [16] see also references terein. In [8] DuPont and Liu introduce te concept of symmetric error estimates for parabolic problems. Tey define suc an error estimate to be one of te form u u C inf u w, w U were u and u are te exact and approximate solutions respectively,. is an appropriate norm, and U is te discrete subspace in wic approximation solutions are sougt. Wile estimates of tis form are standard for elliptic problems, tis is not te case for evolution problems. For example, error estimates for evolution problems approximated by te implicit Euler sceme frequently involve terms of te form u tt L 2 Ω. Symmetric error estimates are useful for problems were te solution u may not be very regular, suc as control problems, and are used to develop a-posteriori error estimates for adaptive scemes. Symmetric error estimates for moving mes finite element metods were studied in [8, 17] see also references witin. Mes modification tecniques for finite elements ave also been introduced in [19] and [2]. For some earlier work on convection-dominated problems based on te metods of caracteristics and mes modification one may consult [6] and [7] respectively. An alternative to te symmetric error estimates are estimates of te form u u C u P u, 1.3 were P : U U is a projection wic exibits optimal interpolation properties if u is sufficiently smoot. Estimates of tis form enjoy te same advantages of tose proposed by DuPont and Liu. Below we construct an estimate of te form 1.3 for parabolic equations of te form 1.1; were te projection P u is te numerical approximation of an ODE, so is not local. However, u P u u P loc u + P u P loc u, were P loc is a local projection, so te first term can be estimated using classical interpolation teory. Te second term P u P loc u vanises if te same subspace of U is used in eac partition, t n ; oterwise, it depends solely upon te jump in te interpolant of te exact solution at te partition points {t n } N n=. Te size of te constant C in 1.3, and its dependence on various constants, play an important role; below we are careful to state te dependence of te constant upon te various coercivity constants and bounds assumed for te operator A. 4

5 Error estimates for Lagrange-Galerkin approximations of convection dominated problems for divergence-free velocity fields vanising on te boundary are presented in [3]. Issues related to te stability of Lagrange-Galerkin approximations are also discussed in [21]. Recently tere as been a lot of work on te development and analysis of discontinuous Galerkin metods for elliptic problems. A compreensive survey and comparison of tis work can be found in [2] wic contains many references related to tis approac Outline. In Section 2, we formulate and analyze te DG sceme for te ordinary differential equation corresponding to A.. In tis section we focus on te difficulties tat arise wen different subspaces of U may be used at every time step. Error estimates are first derived at times corresponding to te partition point. Additional arguments using discrete caracteristic functions are developed to estimate te error and at times inbetween te partition points. Te arguments we use appear to be new. A priori estimates for te DG approximations of 1.1 are developed in Section 3. Estimates are derived in te natural norms associated wit te parabolic problem; by natural we mean norms tat arise in te natural energy estimates obtained by multiplying 1.1 by u. Te results of Section 2 are used in an essential fasion. Indeed, te difficulties associated wit different subspaces of U at eac time step are circumvented by comparing te discrete solution of te parabolic equation wit and appropriate solution of an ODE. By using te discrete caracteristic functions developed in Section 2 we can avoid te self adjoint assumptions typically imposed upon A.. Wen te same discrete subspace of U is used for eac time step our tecniques generalize, and to some extend simplify, te classical analysis. Te reader interested in tis case only needs to read Sections 2.3 and 2.5 on te construction of discrete caracteristic functions, and Definition 2.2 for P loc from Section 2 before proceeding to Section 3. Remark 4 in Section 3 amplifies upon tis Notaton. Trougtout we assume tat te evolution of te solution to 1.1 takes place in a Hilbert space H and te operators A. are defined on anoter Hilbert space U wit U H H U, were eac of te embeddings are dense and continuous. Te inner product on H is denoted by.,. and te induced duality pairing between U and U will be dentoted by.,.. Te norm on H is often denoted by.. H, and we assume tat te norm on U can be written as. 2 U =. 2 U +. 2 H were. U is a semi-norm on U te principle part and is often denoted by. ;. 2 U = Standard notation of te form L 2 [, T ; U], H 1 [, T ; U ] etc. is used to indicate te temporal regularity of functions wit values in U, U etc. Approximations of 1.1 will be constructed on a partition = t < t 1 <... < t N = T of [, T ]. On eac interval of te form, t n ] a subspace U n of U is specified, and te approximate solutions will lie in te space U = {u L 2 [, T ; U] u,t n ] P k, t n ; U n }. Here P k, t n ; U n is te space of polynomials of degree k or less aving values in U n. Notice tat, by convention, we ave cosen functions in U to be left continuous 5

6 wit rigt limits. We will write u n for u t n = u t n, and let u n + dentote ut n +. Tis notation will is also used wit functions like te error e = u u. We always assume te exact solution, u, is in C[, T ; H] so tat te jump in te error at t n, dentoted by [e n ] is equal to [u n ] = u n + u n. 2. DG sceme for an ODE Background. In tis section we address te issues tat arise wen different discrete subspaces are used for eac step of a time of te DG sceme. It suffices to consider suc DG scemes in te context of an ODE in a Hilbert space since te additional terms appearing in te error estimates are te same for bot te ODE and parabolic PDE case. Also, te solution of te ODE will be used to obtain error estimates for te parabolic PDE under minimal regularity assumptions. We consider te problem of recovering a function u C[, T ; H] H 1 [, T ; U ] given te initial value u and its derivative f = u t. Specifically, we consider te DG finite element approximations for te initial value problem u t = f, u = u. 2.1 Recall tat H and U are related troug a pivot space construction, U H H U. In tis situation tere exists a unique u C[, T ; H] H 1 [, T ; U ] wic is te solution of te weak problem, ut, vt T u, v t = u, v + v C[, T ; H] H 1 [, T ; U ]. T f, v 2.2 Recall tat we write for te norm H and write te inner product in H as.,.. We also write u 2 U = u 2 + u 2 were. is a semi-norm on U te principle part. Te U, U duality pairing is denoted by, To approximate te solution of 2.2 we introduce a partition = t < t 1 <... < t N = T of [, T ] and a collection {U n}n n= of subspaces of U. Te DG metod constructs an approximate solution suc tat u U {u L 2 [, T ; H] u,t n ] P k, t n ; U n } tn u n, v n u, v t u n 1, v+ n 1 = f, v, 2.3 for all v U and eac n = 1, 2,... N. Recall tat u n u t n = u t n and we employ te standard notation u n +, u n for te traces from above and below respectively. Integration by parts gives te following alternative form of 2.3 u t, v + u n 1 + un 1, v n 1 + = tn f, v

7 2.2. Error Estimate at Partition Points. In tis elementary context it is possible to explicitly write an expression for te error at eac nodal partition point t n, see equation 2.5 below. A simple consequence of tis formula is te following teorem wic provides a decomposition of te error into te errors due to te canging of te spaces and te initial projection error. Teorem 2.1. Let u U be te approximate solution of 2.1 computed using te discontinuous Galerkin sceme 2.3and let P n : H U n denote te projection operator in H, and write ê n = P n ut n u n. Ten ê n = n n P j I Pi 1 ut i 1 + P n P n 1... P ê j=i In particular, ê n n P i I P i 1 ut i 1 + ê. 2.6 Remark 1. 1 Note tat P i I P i 1 = wen U i U i 1, so ê n ê wen te same discrete subspace is used at eac time. 2 If e n = ut n u n is te total error at time t n ten e n I P n ut n + ê n. Wen te same space is used at eac step te first term becomes I P ut n wic is useful only if ut n can be well approximated in U at all times. Wen tis is not te case, an ideal strategy would cose te spaces {U n } to track te solution so tat bot I P n ut n and te jump terms in equation 2.6 are small. Proof. Let e = u u be te total error and note tat te Galerkin ortogonality gives, e n, v n e, v t e n 1, v+ n 1 =. 2.7 If v t v n is independent of time, ten te middle term vanises to give e n, v n = e n 1, v n, v n U n. It follows tat so tat ê n = P n e n ê n = P n e n 1 = P n u u n 1 = P n u ± P n 1 u u n 1 = P n I P n 1 u + P n ê n 1. Using te above relation inductively yields

8 2.3. Discrete Caracteristic Functions. To compute te error at arbitrary times t [, t n we would like to substitute v = χ [,tu into equation 2.4 were χ [,t is te caracteristic function on [, t. Clearly tis function is not in U so in tis section we construct discrete approximations of suc caracteristic functions to circumvent tis problem. Te construction of te discrete caracteristic functions is invariant under translation so it is convinient to work on te interval [, τ wit τ = t n. We begin by considering polynomials p P k, τ. A discrete approximation of χ [,t p is te polynomial p { p P q, τ p = p} satisfying pq = t pq q P k 1, τ. Te above construction is motivated by te fact tat we may put q = p to obtain τ p p = t pp = p 2 t p 2. We next extend tis elementary construction to approximate functions of te form χ [,t v for v P k, τ; U were U is any subspace of H. If v P k, τ; U we can write v = k i= p itv i were {p i } P k, τ and {v i } U. If we define ṽ = k i= p itv i it is clear tat ṽ P k, τ; U satisfies ṽ = v, and ṽ, w = t v, w w P k 1, τ; U. 2.9 In te ODE setting we could ave directly defined ṽ directly from equation 2.9 instead of te two stage construction given ere. However, for parabolic equations it is useful to observe tat v is independent of te coice of te space U Error Estimates at Arbitrary Times. To estimate te error at an arbitrary time t [, t n we use te projection operator P loc n introduced in [11]. Definition Te projection P loc n P loc n u n = P n ut n, and : C[, t n ; H] P k, t n ; U n satisfies u P loc n u, v =, v P k 1, t n ; U n. Here we ave used te convention P loc n u n P loc n ut n and P n : H U n projection operator onto U n H. 2 Te projection P loc P loc : C[, T ; H] U satisfies u U and P loc u,t n ] = P loc n u [,t n ]. is te Tis projection satisfies te te standard approximation properties, [24, Teorem 12.1], and can be tougt of as te one step DG approximation of u t = f on te interval, t n ] wit exact inital data u. To see tis, write u = P loc n u and write test 8

9 te function as te derivative v t of a function v P k, t n ; U n. Integration by parts ten gives u n, v n u, v t u, v n 1 + = tn u t, v wic is identical in form to 2.3. For te parabolic problem we will use te analogus global projection, P below wic is te DG approximation of u t = f on all of [, T ]. We are now ready to prove te main result of tis section wic sows tat te error estimate of Teorem 2.1, wic eld at te discrete times, olds for every time. Teorem 2.3. Let u U be te approximate solution of 2.1 computed using te discontinuous Galerkin sceme 2.3. Let ê = P loc u u were P loc is te projection defined in Definition 2.2. Ten and êt n P i I P i 1 ut i 1 + ê, t, t n ], [ê n 1 ] n P i I P i 1 ut i 1 + ê. were [ê n 1 ] = ê n 1 + ên 1 is te jump in ê at. More generally, if.,. V êt V ê n V = for t, t n ]. is a semi inner product on U n ten n n P j I Pi 1 ut i 1 + P n P n 1... P ê 1 V, j=i Remark 2. 1 In Teorem 2.1 we used te notation ê n to denote P n e n = P n et n. Tis is consistant wit te notation for êt above, since by construction P loc n e n = P n e n. 2 Notice tat discreteness plays an essential role in te last inequality. We ave not assumed, for example, tat u C[, T ; U], yet te last inequality gives an expression for estimate te error of P n u in tis norm. Proof. Recalling te Galerkin ortogonality relation 2.7 and applying te definition of te projection P loc n sows tat or equivalently, ê n, v n ê, v t e n 1, v n 1 + =, ê t, v + ê n 1 + en 1, v n 1 + =. Define z P, t n ; U n to be te discrete Laplacian of ê; tat is, for eac t [, t n ] z t, w = êt, w V, for all w U n. 9

10 Ten set v = z χ [,tz to be te approximate caracteristic function of z constructed in Section 2.3. Tis coice of test function gives 1/2 êt 2 V ê n V + ê n 1 + P en 1, ê n 1 + V =, so tat 1/2 êt 2 V + 1/2 ê n V P n e n 1, ê n 1 + V =. 2.1 From equation 2.8 we find ê n = P n e n 1, so an application of te Caucy Scwarz inequality and ab 1/2a 2 + b 2 sows êt V ê n V as claimed. To establis te bound on te jump term we set V = H in 2.1 and rearrange te terms to get 1/2 êt 2 + 1/2 ê n ê n 1, ê n 1 + = I P n 1u, ê n 1 +. Next let t + and complete te square on te left to obtain 1/2 ê n /2 ê n 1 + ên 1 2 = 1/2 ê n I P n 1 u, ê n 1 +. It follows tat ê n 1 + ên 1 2 ê n P n I P n 1 u 2 ê n 1 + P n I P n 1 u 2, and again we use te estimate establised in Teorem 2.1 to complete te proof. Te following definition and corollary provide a concise synopsis of te results of tis section in a form useful for te analysis of te parabolic problem in te next section. Definition 2.4. Te projection P : C[, T ; H] H 1 [, T ; U ] U is te discontinuous Galerkin approximation of te function reconstructed from it s derivative and initial data. Tat is, if u = P u, ten u is te solution of 2.3 were f = u. Te previous teorem can ten be interprated as an estimate of te difference between te global projection P u and te local projection P loc u. Corollary 2.5. Let u C[, T ; H] H 1 [, T ; U ], ten and n P u P loc ut P i I P i 1 ut i 1 + ê, t, t n ], [P u P loc un 1 ] n P i I P i 1 ut i 1 + ê, were ê = P u u. More generally, if.,. V P u P loc ut V is a semi inner product on U n ten n n P j I Pi 1 ut i 1 + P n P n 1... P ê 1 V, j=i 1

11 for t, t n ]. Remark 3. An alternative to using te last expression to estimate te error in oter norms is to postulate an inverse inequality. For example, if v C inv v for all v n U n, ten for t tn 1, t n ] P u P loc ut C inv n P i I P i 1 ut i 1 + ê. If te solution as sufficent regularity projection errors in te different norms typically satisfy C inv e e Estimates for Discrete Caracteristic Functions. Our construction of te discrete caracteristic functions in Section 2.3 was purely algebraic and, for te ODE, no bounds were required for te mapping e ẽ. Te analysis of te parabolic problem requires te bounds developed next. In te next two lemmas, te function p p will refer to te map constructed in Secton 2.3. Lemma 2.6. Te mapping p p in P k, τ is linear, continuous and tere exists a constant Ĉk depending only upon k suc tat p p L 2,τ Ĉk p 2 L 2 t,τ. Moreover, p χ [,t p L 2,τ p p L 2,τ + p χ [,t p L 2,τ 1 + Ĉk p L 2 t,τ and p L 2,τ 1 + Ĉk p L 2,τ. Proof. Since p = p we can write p p = t p wit p P k 1, τ. Ten using te definition of p, t pq = p pq = pq q P k 1, τ. t Setting q = p we obtain, c k τ p 2 t p 2 = p p, t were we use te equivalence of norms on P k, and scale to make c k independent of τ. It follows tat te matrix corresponding to te bilinear form is non-singular, so solutions exist. Te Caucy-Scwarz inequality gives wic implies, c 2 k p p 2 = c 2 k c k τ 2 p 2 t p 2, t 2 p 2 c 2 τ 2 p 2 t p 2. 11

12 We next consider te induced mapping v ṽ on P k, τ; V were V is any semi inner product space. Recall tat if v = k i= p itv i ten ṽ = k i= p itv i. Since tis construction is purely algebraic it follows tat ṽ P k, τ; V satisfies ṽ = v and ṽ, w V = t v, w V w P k 1, τ; V Lemma 2.7. Let V be a semi-inner product space, ten te mapping k i= p itv i k i= p itv i on P k, τ; V is continuous in L 2 [,τ;v ]. In particular, ṽ L 2 [,τ;v ] C k v L 2 [,τ;v ], and ṽ χ [,t v L 2 [,τ;v ] C k v L 2 [,τ;v ], were C k = k + 1 1/2 1 + Ĉk. Proof. Witout loss of generality write v = k i= p itv i were {p i } form an ortonormal basis of P k, τ in L 2, τ, so tat v 2 L 2 [,τ;v ] = k i= v i 2 V. Te lemma ten follows by direct computation: ṽ 2 V = k i,j= p i t p j tv i, v j V k p i L 2,τ p j L 2,τ v i V v j V i,j= 1 + Ĉk 2 Te second estimate follows similarly. k v i V v j V i,j= k 1 + Ĉk 2 k + 1 v i 2 V 1 + Ĉk 2 k DG Sceme for Parabolic PDE s. i= v 2 V Formulation of te DG sceme.. We turn our attention on te approximation of 1.1 using te discontinuous Galerkin sceme. In order to derive optimal error estimates we extend te ideas introduced in Section 2. We denote by a, te natural bilinear form associated wit A. We assume a.,. satisfies te following continuity and coercivity conditions. Assumption 1. Tere exist non-negative constants C a, C α, c a, c α suc tat 12

13 1. Continuity of te bilinear form and data: at; u, v c a u 2 + C a u c a v 2 + C a v and f, v f ca u 2 + C a u Corecivity of te bilinear form: at; u, u c α u 2 C α u 2 In tis context te natural weak formulation of 1.1 is to find u U L 2 [, T ; U] H 1 [, T ; U ] suc tat ut, vt + T u, v t + at, u, v = u, v + T F, v, 3.1 for all v U. We empasize tat U is te natural space to seek convergence, so ideally we would like to derive estimates in a related norm. To approximate te solution of te above weak formulation we introduce a partition = t < t 1 <... < t N = T of [, T ] and on eac partition we construct a closed subspace U n U. Te discontinuous Galerkin metod constructs an approximate solution u,t n ] P k, t n ; U n satisfying u n, v n + u, v t + a ; u, v = u n 1, v n tn F, v v P k, t n ; U n. Integration of te temporal term by parts yields te representation: 3.2 u t, v + a ; u, v + u n 1 + un 1, v+ n = F, v v P k, t n ; U n Preliminary Estimates. Classical bounds for te parabolic equation 3.2 are obtained upon selecting u = v in equation 3.1; te discrete analogue would be to set v = u in 3.3. Upon observing tat u t, u + u n 1 + un 1, u n 1 + = 1 2 un un un 1 + un 1 2, standard enery arguments use te continuity and coercivity assumptions 1 lead to te inequality u n 2 + c α u n u 2 + u n 1 u n ca /c α F 2 + 2C α + C a u 2. 13

14 Wen u P k, t n ; U n wit k 2 tis inequality does not control u s for s, t n. Wen u is piecewise constant or linear in time k = or 1 te terms on te left and side will dominate te last term on te rigt for sufficiently small time steps [24, Teorem 12.4]. Te case k > 2 as not been completly addressed previouly; typically strict coercivity is assumed so tat C α. In tis situation it is possible to write t m 1 F, u t m 1 1/2ɛ F 2 + ɛ/2 u 2 + u 2 and using te Poincare inequality to control te last term. Similar difficulties are encountered wit error estimates wen k 2. Letting e = u u te ortogonality condition becomes e n, v n + e, v t + a ; e, v = e n 1, v+ n 1 for all v P k, t n ; U n. Writing e = u u = u P u + P u u e + e p, were P is te projection defined in Definition 2.4 we compute e n, vn + e, v t + a ; e, v e n 1, v n 1 = e n p, v n + e p, v t + e n 1 p + tn, v+ n 1 a ; e p, v Since P u is te discontinuous Galerkin solution of an ordinary differential equation it follows tat e p = u P u satisfies te orgogonality condition 2.7. We ten conclude tat all but te last term on te rigt and side of te above expression vanises so tat e n, vn + e, v t + a ; e, v tn = e n 1, v n 1 + a ; e p, v. 3.5 Tis expresion is identical in form to te original sceme 3.2 for u wit F. = ae p,.. Putting v = e we obtain te analogue of equation 3.4, e n 2 + c α e n e 2 + e n 1 e n c a /c α c a e p 2 + C a e p 2 + 2C α + C a e 2. Again te natural energy arguments for te DG sceme fail to control e t for t, t n. 14

15 Remark 4. Te projection constructed using te discrete solution of te corresponding ODE is necessary wen different subspaces are used in eac time step, i.e., U n U n 1. Using te standard projection, P loc u in place of P as in [24] gives e n, vn + e, v t + a ; e, v tn = e n 1, v n 1 + a ; e n 1 p, v + e p, v n 1 +. wit e p = u P loc u. Note tat te last term is equal to en 1 p, v n 1 + w for every w U n 1, so wen U n = U n 1 we may pick w = v n 1 + to obtain Stability and Error Estimates. In tis section we sow ow te estimates in 3.4 and 3.6 can be augmented to provide bounds on te solution and te error for all times; in particular, for te intermediate times t, t n. To do tis we use te discrete caracteristic functions developed in Section 2.3. Since equations 3.2 for u and 3.5 for e are identical in form, te same line of argument can be applied to eiter of tem. For tis reason we will only prove te teorem for te error estimate and will simply state te analogous bound for u. Teorem 3.1. Let U H U be a dense embedding of Hilbert spaces and U be te subspace of L 2 [, T ; U] defined in Section 2.1 and let te bilinear form a : U U R and linear form F : U R satisfy Assumptions 1. Let u L 2 [, T ; U] H 1 [, T ; U ] be te solution of 3.1 and u U be te approximate solution computed using te discontinuous Galerkin sceme 3.2 on te partition = t < t 1 <... < t N = T and set τ max n t n. Ten tere exists a constant C > depending only on k troug te constant C k of Lemma 2.7, te constants C a, C α, and te ratio c a /c α, suc tat and n 1 1 λ u n 2 + λ sup us 2 + e Ctn 1 t i u i u i + 2 s t n +1 λ c α e Ctn s u s 2 ds T Oτ e Ctn u 2 + Cλ e Ctn s F s 2 ds, n 1 1 λ e n 2 + λ sup e s 2 + e Ctn 1 t i e i ei + 2 s t n i= +1 λ c α e Ctn s e s 2 ds T Oτ e Ctn e 2 + Cλ e Ctn s c a e p s 2 + C a e p s 2 ds, 15 i=

16 provided Cτ < 1. Here λ = 1/2C k + 4C k c a /c α + 1, 1, and e p = u P u and e = P u u were P : C[, T ; H] H 1 [, T ; U ] U is te projection defined in Definition 2.4. Proof. Since te line of argument to prove eac inequality is identical we only prove te second. Let ẽ P k, t n ; U n be te discrete approximation of χ [t n,te t constructed in Lemma 2.7. Setting v = ẽ in equation 3.5 and moving te term a.; e, ẽ to te rigt and side gives 1 2 e t en 1 e n = 1 2 en 1 2 a.; e p, ẽ + a ; e, ẽ. We estimate te eac of te last two terms seperatly. a.; e, ẽ c a e 2 + C a e 2 1/2 c a ẽ 2 + C a ẽ 2 1/2 t n 1 t n 1/2 c a e 2 + C a e 2 c a ẽ 2 + C a ẽ 2 t n 1 t n C k c a e 2 + C a e 2. Lemma 2.7 was used to bound ẽ in terms of e in te last line. A similar computation sows a.; e p, ẽ C k 2 Combining te above gives 1/2 c a /c α + 1c a e p 2 + C a e p 2 + c α e 2 + C a e 2. e t 2 + e n 1 e n e n C k c a /c α + 1c a e p 2 + C a e p 2 + c α 1 + 2c a /c α e 2 + 3C a e 2 We now consider te convex combination of 1 λ equation 3.4 and λ equation 3.7. We coose te coefficient, λ, so tat te term involving e 2 on te rigt of 3.7 is dominated by te corresponding term on te left of 3.4. Speciffically, let λc k 1 + 2c a /c α = 1/21 λ, or λ = Tis gives an estimate of te form 1 λ e n 2 + λ e t λ c α 2 e n Cλ 16 e 2 + e n 1 1 2C k + 4C k c a /c α + 1. e n c a e p 2 + C a e p 2 + e 2

17 were C depends only upon te coercivity constant c α and c a troug te ratio c a /c α. Bound te first and last terms on te rigt by and e n λ e n λ sup e s 2 t n 2 <s e 2 τ n sup e s 2, τ n t n, <s t n respectivly, and select te time t on te left so tat e t = sup <s t n e s to get 1 λ e n 2 + λ1 Cτ n sup e t λ c α e 2 + e n 1 <s t n 2 1 λ e n λ sup t n 2 <s e s 2 + Cλ e n c a e p 2 + C a e p 2. Upon introducing a factor 1 Cτ n in front of te first term, tis inequality takes te form 1 Cτ n α n + β n α n 1 + f n, and te teorem follows from te discrete Gronwall inequality. Remark 5. Note tat te above estimate consists of norms tat measure te beaviour of te solution at te nodal, interior and jump points. All norms included are te natural ones and no additional regularity is required on te rigt and side. Teorem 3.1 essentially sows tat te error for te parabolic equation can be bounded by te error of te ODE. If te norms.,. 2 and jump term J N e are defined by and v 2 = v 2 2 = T T sup vs 2 + c α e CT s vs 2 ds s T T e CT s vs 2 ds + c a e CT s vs 2 ds, J 2 Nv = N 1 i= e CT ti v i v i + 2, ten Teorem 3.1 states P u u 2 +JNP 2 u u CT P u u 2 + u P u 2 2 Since T ect s e p s 2 e CT sup s T e p s 2 various symetric error estimates follow. For example, setting u = P u and using te triangle inequlality gives u u 2 CT u P u 2, and u u CT u P u. Since P u is not a local projection, classical interpolation teory does not immediatly yield rates of convergence for a specific problem. However, we can use te results of. 17

18 Section 2.4 to estimate te rigt and sides of te above in terms of te local projection P loc u. Teorem 3.2. Under te assumptions in teorem 3.1 tere exists a positive constant CT depending only on k troug te constant C k of Lemma 2.7, te constants C, λ, C a, C α, and te ratio c a /c α, and te final time T suc tat te following estimate olds u u CT e + u Ploc u 2 + P u P loc u N CT e + u Ploc u 2 + P i I P i 1 ut i 1 + n c a max n P j I Pi 1 ut i 1 + P n... P 1 e, 1 n N j=i were e = P u u, and P loc u is te local projection defined in Definition 2.2, and P n : H U n is te ortogonal projection. Remark 6. 1 Upon assuming an inverse inequality of te form v C inv v for all v U n, n =, 1,..., and >, te above estimate simplifies to u u CT u P loc u c a C inv N P i I P i 1 ut i 1 + e. 2 Estimates for te jump terms J N u u can also be obtained; owever, tey will converge at a reduced rate; specifically, J N u u CT u P loc u 2 +J N u P loc u + N N P i I P i 1 ut i 1 + e + n c a max n P j I Pi 1 ut i 1 + P n... P 1 e, 1 n N j=i Appendix A. Discrete Gronwall Inequality. If 1 Cτ n a n + b n a n 1 + f n te discrete Gronwall inequality states tat if max n Cτ n < 1 ten a N + N n=1 b n N i=n 1 Cτ i a N N 1 Cτ i + f n N i=n 1 Cτ i. n=1 Since N exp Cτ i i=n 1 N i=n 1 Cτ i 1 1 N N Cτ i 2 exp Cτ i i=n i=n 18

19 and 1 N i=n Cτ i 2 1 C 2 T τ, were τ = max n τ n we may write a N + N e CtN t n b n 1 + T Oτ e CtN a + n=1 were t n = n τ n. N e CtN t n f n, n=1 REFERENCES [1] R. ADAMS, Sobolev Spaces, Academic Press, New York, [2] D.N. ARNOLD, F. BREZZI, B. COCKBURN AND D.L. MARINI, Unified analysis of discontinuous Galerkin metods for elliptic problems, SIAM J. Num. Anal. 39, 21-2, pp [3] M. BAUSE AND P. KNABNER, Uniform error analysis for Lagrange-Galerkin approximations of convection-dominated problems, SIAM J. Num. Anal. 39, pp [4] B. COCKBURN, G. E. KARNADIAKIS, AND C. W. SHU, Te development of discontinuous Galerkin metods, in Discontinuous Galerkin metods Newport, RI, 1999, Springer, Berlin, 2, pp 3-5. [5] M. DELFOUR, W. HAGER AND F. TROCHU, Discontinuous Galerkin Metods for Ordinary Differential Equations, Mat. Comp, 36, 1981, pp [6] J. DOUGLAS Jr. AND T. F. RUSSEL, Numerical metods for convection dominated difussion problems based on combining te metod of caracteristics wit finite element or finite difference procedures, SIAM J. Numer. Anal., 19, 1982, pp [7] T. DUPONT, Mes modification for evolutionary equations, Mat. Comp., 39, 1982, pp [8] T. F. DUPONT AND YINGJIE LIU, Symmetric error estimates for moving mes Galerkin metods for advection-diffusion equations, SIAM J. Numer. Anal. 4, 22, pp [9] K. ERIKSSON AND C. JOHNSON, Adaptive finite element metods for parabolic problems. I. A linear model problem, SIAM J. Numer. Anal., 28, 1991, pp [1] K. ERIKSSON AND C. JOHNSON, Adaptive finite element metods for parabolic problems. II. Optimal error estimates in L L 2 and L L, SIAM J. Numer. Anal., 32, 1995, pp [11] K. ERIKSSON, C. JOHNSON AND V. THOMÉE, Time discretization of parabolic problems by te discontinuous Galerkin metod, RAIRO Modél. Mat. Anal. Numér., 29, 1985, pp [12] D. ESTEP AND S. LARSSON, Te discontinuous Galerkin metod for semilinear parabolic equations, RAIRO Modél. Mat. Anal. Numér., 27, 1993, pp [13] P. JAMET, Galerkin-type approximations wic are discontinuous in time for parabolic equations in a varible domain, SIAM J. Numer. Anal. 15, 1978, pp [14] C. JOHNSON, Numerical Solution of Partial Differential Equations by te Finite Element Metod, Cambridge, [15] P. LASAINT AND P.-A. RAVIART, On a finite element metod for solving te neutron transport equation, in Matematical aspects of finite elements in partial differential equations, C. de Boor, ed., Academic Press, New York, 1974, pp [16] S. LARSSON, V. THOMÉE AND L.B. WALHBIN, Numerical solution of parabolic intregrodifferential equations by te discontinuous Galerkin metod, Mat. Comp. 67, 1998, pp

20 [17] YINGJIE LIU, R.E. BANK, T.F DUPONT, S. GARCIA AND R.P. SANTOS, Symmetric error estimates for moving mes mixed metods for advection-diffusion equations, SIAM J. Numer. Anal. 4, 23, pp [18] M. LUSKIN AND R. RANNACHER, On te smooting property of te Galerkin metod for parabolic equations, SIAM J. Numer. Anal. 19, 1982, pp [19] K. MILLER, Moving finite elements II, SIAM J. Numer. Anal. 18, 1981, pp [2] K. MILLER AND R.N. MILLER, Moving finite elements I, SIAM J. Numer. Anal. 18, 1981, pp [21] K.W. MORTON, A. PRIESTLEY AND E. SÜLI, Stability of te Lagrange-Galerkin metod wit nonexact integration, RAIRO Modél. Mat. Anal. Numér., 22, 1988, pp [22] SHOWALTER, R. E., Hilbert Space Metods for Partial Differential Equations, Pitman, [23] SHOWALTER, R. E., Monotone operators in Banac space and nonlinear partial differential equations, American Matematical Society, Providence, RI, [24] V. THOMÉE, Galerkin finite element metods for parabolic problems, Spinger-Verlag, Berlin, 1997 [25] N. J. WALKINGTON, Convergence of te discontinuous Galerkin metod for discontinuous solutions, preprint. 2

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

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

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

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

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

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

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

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

arxiv: v1 [math.na] 12 Mar 2018

arxiv: v1 [math.na] 12 Mar 2018 ON PRESSURE ESTIMATES FOR THE NAVIER-STOKES EQUATIONS J A FIORDILINO arxiv:180304366v1 [matna 12 Mar 2018 Abstract Tis paper presents a simple, general tecnique to prove finite element metod (FEM) pressure

More information

FINITE ELEMENT EXTERIOR CALCULUS FOR PARABOLIC EVOLUTION PROBLEMS ON RIEMANNIAN HYPERSURFACES

FINITE ELEMENT EXTERIOR CALCULUS FOR PARABOLIC EVOLUTION PROBLEMS ON RIEMANNIAN HYPERSURFACES FINITE ELEMENT EXTERIOR CALCULUS FOR PARABOLIC EVOLUTION PROBLEMS ON RIEMANNIAN HYPERSURFACES MICHAEL HOLST AND CHRIS TIEE ABSTRACT. Over te last ten years, te Finite Element Exterior Calculus (FEEC) as

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

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

arxiv: v1 [math.na] 18 Jul 2015

arxiv: v1 [math.na] 18 Jul 2015 ENERGY NORM ERROR ESTIMATES FOR FINITE ELEMENT DISCRETIZATION OF PARABOLIC PROBLEMS HERBERT EGGER arxiv:150705183v1 [matna] 18 Jul 2015 Department of Matematics, TU Darmstadt, Germany Abstract We consider

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

Blanca Bujanda, Juan Carlos Jorge NEW EFFICIENT TIME INTEGRATORS FOR NON-LINEAR PARABOLIC PROBLEMS

Blanca Bujanda, Juan Carlos Jorge NEW EFFICIENT TIME INTEGRATORS FOR NON-LINEAR PARABOLIC PROBLEMS Opuscula Matematica Vol. 26 No. 3 26 Blanca Bujanda, Juan Carlos Jorge NEW EFFICIENT TIME INTEGRATORS FOR NON-LINEAR PARABOLIC PROBLEMS Abstract. In tis work a new numerical metod is constructed for time-integrating

More information

Solving Continuous Linear Least-Squares Problems by Iterated Projection

Solving Continuous Linear Least-Squares Problems by Iterated Projection Solving Continuous Linear Least-Squares Problems by Iterated Projection by Ral Juengling Department o Computer Science, Portland State University PO Box 75 Portland, OR 977 USA Email: juenglin@cs.pdx.edu

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

COMPACTNESS PROPERTIES OF THE DG AND CG TIME STEPPING SCHEMES FOR PARABOLIC EQUATIONS. u t + A(u) = f(u).

COMPACTNESS PROPERTIES OF THE DG AND CG TIME STEPPING SCHEMES FOR PARABOLIC EQUATIONS. u t + A(u) = f(u). COMPACTNESS PROPERTIES OF THE DG AND CG TIME STEPPING SCHEMES FOR PARABOLIC EQUATIONS NOEL J. WALKINGTON Abstract. It is sown tat for a broad class of equations tat numerical solutions computed using te

More information

Some Error Estimates for the Finite Volume Element Method for a Parabolic Problem

Some Error Estimates for the Finite Volume Element Method for a Parabolic Problem Computational Metods in Applied Matematics Vol. 13 (213), No. 3, pp. 251 279 c 213 Institute of Matematics, NAS of Belarus Doi: 1.1515/cmam-212-6 Some Error Estimates for te Finite Volume Element Metod

More information

A SPLITTING LEAST-SQUARES MIXED FINITE ELEMENT METHOD FOR ELLIPTIC OPTIMAL CONTROL PROBLEMS

A SPLITTING LEAST-SQUARES MIXED FINITE ELEMENT METHOD FOR ELLIPTIC OPTIMAL CONTROL PROBLEMS INTERNATIONAL JOURNAL OF NUMERICAL ANALSIS AND MODELING Volume 3, Number 4, Pages 6 626 c 26 Institute for Scientific Computing and Information A SPLITTING LEAST-SQUARES MIED FINITE ELEMENT METHOD FOR

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

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

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

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

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

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

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

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

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

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

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

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

Key words. Finite element method; convection-diffusion-reaction; nonnegativity; boundedness

Key words. Finite element method; convection-diffusion-reaction; nonnegativity; boundedness PRESERVING NONNEGATIVITY OF AN AFFINE FINITE ELEMENT APPROXIMATION FOR A CONVECTION-DIFFUSION-REACTION PROBLEM JAVIER RUIZ-RAMÍREZ Abstract. An affine finite element sceme approximation of a time dependent

More information

arxiv: v1 [math.na] 19 Mar 2018

arxiv: v1 [math.na] 19 Mar 2018 A primal discontinuous Galerkin metod wit static condensation on very general meses arxiv:1803.06846v1 [mat.na] 19 Mar 018 Alexei Lozinski Laboratoire de Matématiques de Besançon, CNRS UMR 663, Univ. Bourgogne

More information

Finite Difference Method

Finite Difference Method Capter 8 Finite Difference Metod 81 2nd order linear pde in two variables General 2nd order linear pde in two variables is given in te following form: L[u] = Au xx +2Bu xy +Cu yy +Du x +Eu y +Fu = G According

More information

Part VIII, Chapter 39. Fluctuation-based stabilization Model problem

Part VIII, Chapter 39. Fluctuation-based stabilization Model problem Part VIII, Capter 39 Fluctuation-based stabilization Tis capter presents a unified analysis of recent stabilization tecniques for te standard Galerkin approximation of first-order PDEs using H 1 - conforming

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

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

Chapter 10. Function approximation Function approximation. The Lebesgue space L 2 (I)

Chapter 10. Function approximation Function approximation. The Lebesgue space L 2 (I) Capter 1 Function approximation We ave studied metods for computing solutions to algebraic equations in te form of real numbers or finite dimensional vectors of real numbers. In contrast, solutions to

More information

1. Introduction. Consider a semilinear parabolic equation in the form

1. Introduction. Consider a semilinear parabolic equation in the form A POSTERIORI ERROR ESTIMATION FOR PARABOLIC PROBLEMS USING ELLIPTIC RECONSTRUCTIONS. I: BACKWARD-EULER AND CRANK-NICOLSON METHODS NATALIA KOPTEVA AND TORSTEN LINSS Abstract. A semilinear second-order parabolic

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

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

Arbitrary order exactly divergence-free central discontinuous Galerkin methods for ideal MHD equations

Arbitrary order exactly divergence-free central discontinuous Galerkin methods for ideal MHD equations Arbitrary order exactly divergence-free central discontinuous Galerkin metods for ideal MHD equations Fengyan Li, Liwei Xu Department of Matematical Sciences, Rensselaer Polytecnic Institute, Troy, NY

More information

Parabolic PDEs: time approximation Implicit Euler

Parabolic PDEs: time approximation Implicit Euler Part IX, Capter 53 Parabolic PDEs: time approximation We are concerned in tis capter wit bot te time and te space approximation of te model problem (52.4). We adopt te metod of line introduced in 52.2.

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

ERROR ESTIMATES FOR A FULLY DISCRETIZED SCHEME TO A CAHN-HILLIARD PHASE-FIELD MODEL FOR TWO-PHASE INCOMPRESSIBLE FLOWS

ERROR ESTIMATES FOR A FULLY DISCRETIZED SCHEME TO A CAHN-HILLIARD PHASE-FIELD MODEL FOR TWO-PHASE INCOMPRESSIBLE FLOWS ERROR ESTIMATES FOR A FULLY DISCRETIZED SCHEME TO A CAHN-HILLIARD PHASE-FIELD MODEL FOR TWO-PHASE INCOMPRESSIBLE FLOWS YONGYONG CAI, AND JIE SHEN Abstract. We carry out in tis paper a rigorous error analysis

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

Priority Program 1253

Priority Program 1253 Deutsce Forscungsgemeinscaft Priority Program 53 Optimization wit Partial Differential Equations K. Deckelnick, A. Günter and M. Hinze Finite element approximation of elliptic control problems wit constraints

More information

Inf sup testing of upwind methods

Inf sup testing of upwind methods INTERNATIONAL JOURNAL FOR NUMERICAL METHODS IN ENGINEERING Int. J. Numer. Met. Engng 000; 48:745 760 Inf sup testing of upwind metods Klaus-Jurgen Bate 1; ;, Dena Hendriana 1, Franco Brezzi and Giancarlo

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

Parallel algorithm for convection diffusion system based on least squares procedure

Parallel algorithm for convection diffusion system based on least squares procedure Zang et al. SpringerPlus (26 5:69 DOI.86/s464-6-3333-8 RESEARCH Parallel algoritm for convection diffusion system based on least squares procedure Open Access Jiansong Zang *, Hui Guo 2, Hongfei Fu 3 and

More information

ELLIPTIC RECONSTRUCTION AND A POSTERIORI ERROR ESTIMATES FOR PARABOLIC PROBLEMS

ELLIPTIC RECONSTRUCTION AND A POSTERIORI ERROR ESTIMATES FOR PARABOLIC PROBLEMS ELLIPTIC RECONSTRUCTION AND A POSTERIORI ERROR ESTIMATES FOR PARABOLIC PROBLEMS CHARALAMBOS MAKRIDAKIS AND RICARDO H. NOCHETTO Abstract. It is known that the energy technique for a posteriori error analysis

More information

A New Class of Zienkiewicz-Type Nonconforming Element in Any Dimensions

A New Class of Zienkiewicz-Type Nonconforming Element in Any Dimensions Numerisce Matematik manuscript No. will be inserted by te editor A New Class of Zienkiewicz-Type Nonconforming Element in Any Dimensions Wang Ming 1, Zong-ci Si 2, Jincao Xu1,3 1 LMAM, Scool of Matematical

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

OSCILLATION OF SOLUTIONS TO NON-LINEAR DIFFERENCE EQUATIONS WITH SEVERAL ADVANCED ARGUMENTS. Sandra Pinelas and Julio G. Dix

OSCILLATION OF SOLUTIONS TO NON-LINEAR DIFFERENCE EQUATIONS WITH SEVERAL ADVANCED ARGUMENTS. Sandra Pinelas and Julio G. Dix Opuscula Mat. 37, no. 6 (2017), 887 898 ttp://dx.doi.org/10.7494/opmat.2017.37.6.887 Opuscula Matematica OSCILLATION OF SOLUTIONS TO NON-LINEAR DIFFERENCE EQUATIONS WITH SEVERAL ADVANCED ARGUMENTS Sandra

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

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

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

3.4 Worksheet: Proof of the Chain Rule NAME

3.4 Worksheet: Proof of the Chain Rule NAME Mat 1170 3.4 Workseet: Proof of te Cain Rule NAME Te Cain Rule So far we are able to differentiate all types of functions. For example: polynomials, rational, root, and trigonometric functions. We are

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

Decay of solutions of wave equations with memory

Decay of solutions of wave equations with memory Proceedings of te 14t International Conference on Computational and Matematical Metods in Science and Engineering, CMMSE 14 3 7July, 14. Decay of solutions of wave equations wit memory J. A. Ferreira 1,

More information

Stability properties of a family of chock capturing methods for hyperbolic conservation laws

Stability properties of a family of chock capturing methods for hyperbolic conservation laws Proceedings of te 3rd IASME/WSEAS Int. Conf. on FLUID DYNAMICS & AERODYNAMICS, Corfu, Greece, August 0-, 005 (pp48-5) Stability properties of a family of cock capturing metods for yperbolic conservation

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

FEM solution of the ψ-ω equations with explicit viscous diffusion 1

FEM solution of the ψ-ω equations with explicit viscous diffusion 1 FEM solution of te ψ-ω equations wit explicit viscous diffusion J.-L. Guermond and L. Quartapelle 3 Abstract. Tis paper describes a variational formulation for solving te D time-dependent incompressible

More information

Global Existence of Classical Solutions for a Class Nonlinear Parabolic Equations

Global Existence of Classical Solutions for a Class Nonlinear Parabolic Equations Global Journal of Science Frontier Researc Matematics and Decision Sciences Volume 12 Issue 8 Version 1.0 Type : Double Blind Peer Reviewed International Researc Journal Publiser: Global Journals Inc.

More information

Efficient, unconditionally stable, and optimally accurate FE algorithms for approximate deconvolution models of fluid flow

Efficient, unconditionally stable, and optimally accurate FE algorithms for approximate deconvolution models of fluid flow Efficient, unconditionally stable, and optimally accurate FE algoritms for approximate deconvolution models of fluid flow Leo G. Rebolz Abstract Tis paper addresses an open question of ow to devise numerical

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

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

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

Department of Mathematical Sciences University of South Carolina Aiken Aiken, SC 29801

Department of Mathematical Sciences University of South Carolina Aiken Aiken, SC 29801 RESEARCH SUMMARY AND PERSPECTIVES KOFFI B. FADIMBA Department of Matematical Sciences University of Sout Carolina Aiken Aiken, SC 29801 Email: KoffiF@usca.edu 1. Introduction My researc program as focused

More information

CLEMSON U N I V E R S I T Y

CLEMSON U N I V E R S I T Y A Fractional Step θ-metod for Convection-Diffusion Equations Jon Crispell December, 006 Advisors: Dr. Lea Jenkins and Dr. Vincent Ervin Fractional Step θ-metod Outline Crispell,Ervin,Jenkins Motivation

More information

SECTION 3.2: DERIVATIVE FUNCTIONS and DIFFERENTIABILITY

SECTION 3.2: DERIVATIVE FUNCTIONS and DIFFERENTIABILITY (Section 3.2: Derivative Functions and Differentiability) 3.2.1 SECTION 3.2: DERIVATIVE FUNCTIONS and DIFFERENTIABILITY LEARNING OBJECTIVES Know, understand, and apply te Limit Definition of te Derivative

More information

H(div) conforming and DG methods for incompressible Euler s equations

H(div) conforming and DG methods for incompressible Euler s equations H(div) conforming and DG metods for incompressible Euler s equations Jonny Guzmán Filánder A. Sequeira Ci-Wang Su Abstract H(div) conforming and discontinuous Galerkin (DG) metods are designed for incompressible

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

Consider a function f we ll specify which assumptions we need to make about it in a minute. Let us reformulate the integral. 1 f(x) dx.

Consider a function f we ll specify which assumptions we need to make about it in a minute. Let us reformulate the integral. 1 f(x) dx. Capter 2 Integrals as sums and derivatives as differences We now switc to te simplest metods for integrating or differentiating a function from its function samples. A careful study of Taylor expansions

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

Grad-div stabilization for the evolutionary Oseen problem with inf-sup stable finite elements

Grad-div stabilization for the evolutionary Oseen problem with inf-sup stable finite elements Noname manuscript No. will be inserted by te editor Grad-div stabilization for te evolutionary Oseen problem wit inf-sup stable finite elements Javier de Frutos Bosco García-Arcilla Volker Jon Julia Novo

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

Semigroups of Operators

Semigroups of Operators Lecture 11 Semigroups of Operators In tis Lecture we gater a few notions on one-parameter semigroups of linear operators, confining to te essential tools tat are needed in te sequel. As usual, X is a real

More information

AMS 147 Computational Methods and Applications Lecture 09 Copyright by Hongyun Wang, UCSC. Exact value. Effect of round-off error.

AMS 147 Computational Methods and Applications Lecture 09 Copyright by Hongyun Wang, UCSC. Exact value. Effect of round-off error. Lecture 09 Copyrigt by Hongyun Wang, UCSC Recap: Te total error in numerical differentiation fl( f ( x + fl( f ( x E T ( = f ( x Numerical result from a computer Exact value = e + f x+ Discretization error

More information

ch (for some fixed positive number c) reaching c

ch (for some fixed positive number c) reaching c GSTF Journal of Matematics Statistics and Operations Researc (JMSOR) Vol. No. September 05 DOI 0.60/s4086-05-000-z Nonlinear Piecewise-defined Difference Equations wit Reciprocal and Cubic Terms Ramadan

More information

A Demonstration of the Advantage of Asymptotic Preserving Schemes over Standard Finite Volume Schemes

A Demonstration of the Advantage of Asymptotic Preserving Schemes over Standard Finite Volume Schemes A Demonstration of te Advantage of Asymptotic Preserving Scemes over Standard Finite Volume Scemes Jocen Scütz Berict Nr. 366 Juni 213 Key words: conservation laws, asymptotic metods, finite volume metods,

More information

HOMEWORK HELP 2 FOR MATH 151

HOMEWORK HELP 2 FOR MATH 151 HOMEWORK HELP 2 FOR MATH 151 Here we go; te second round of omework elp. If tere are oters you would like to see, let me know! 2.4, 43 and 44 At wat points are te functions f(x) and g(x) = xf(x)continuous,

More information

Chapter 4: Numerical Methods for Common Mathematical Problems

Chapter 4: Numerical Methods for Common Mathematical Problems 1 Capter 4: Numerical Metods for Common Matematical Problems Interpolation Problem: Suppose we ave data defined at a discrete set of points (x i, y i ), i = 0, 1,..., N. Often it is useful to ave a smoot

More information

WEIGHTED ERROR ESTIMATES OF THE CONTINUOUS INTERIOR PENALTY METHOD FOR SINGULARLY PERTURBED PROBLEMS

WEIGHTED ERROR ESTIMATES OF THE CONTINUOUS INTERIOR PENALTY METHOD FOR SINGULARLY PERTURBED PROBLEMS WEIGHTED ERROR ESTIMATES OF THE CONTINUOUS INTERIOR PENALTY METHOD FOR SINGULARLY PERTURBED PROBLEMS ERIK BURMAN, JOHNNY GUZMÁN, AND DMITRIY LEYKEKHMAN Abstract. In tis paper we analyze local properties

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

Analysis of the grad-div stabilization for the time-dependent Navier Stokes equations with inf-sup stable finite elements

Analysis of the grad-div stabilization for the time-dependent Navier Stokes equations with inf-sup stable finite elements arxiv:161.517v3 [mat.na] 2 May 217 Analysis of te grad-div stabilization for te time-dependent Navier Stokes equations wit inf-sup stable finite elements Javier de Frutos Bosco García-Arcilla Volker Jon

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

MA455 Manifolds Solutions 1 May 2008

MA455 Manifolds Solutions 1 May 2008 MA455 Manifolds Solutions 1 May 2008 1. (i) Given real numbers a < b, find a diffeomorpism (a, b) R. Solution: For example first map (a, b) to (0, π/2) and ten map (0, π/2) diffeomorpically to R using

More information

ETNA Kent State University

ETNA Kent State University Electronic Transactions on Numerical Analysis. Volume 34, pp. 14-19, 2008. Copyrigt 2008,. ISSN 1068-9613. ETNA A NOTE ON NUMERICALLY CONSISTENT INITIAL VALUES FOR HIGH INDEX DIFFERENTIAL-ALGEBRAIC EQUATIONS

More information

Smoothed projections in finite element exterior calculus

Smoothed projections in finite element exterior calculus Smooted projections in finite element exterior calculus Ragnar Winter CMA, University of Oslo Norway based on joint work wit: Douglas N. Arnold, Minnesota, Ricard S. Falk, Rutgers, and Snorre H. Cristiansen,

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

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

A SYMMETRIC NODAL CONSERVATIVE FINITE ELEMENT METHOD FOR THE DARCY EQUATION

A SYMMETRIC NODAL CONSERVATIVE FINITE ELEMENT METHOD FOR THE DARCY EQUATION A SYMMETRIC NODAL CONSERVATIVE FINITE ELEMENT METHOD FOR THE DARCY EQUATION GABRIEL R. BARRENECHEA, LEOPOLDO P. FRANCA 1 2, AND FRÉDÉRIC VALENTIN Abstract. Tis work introduces and analyzes novel stable

More information

A posteriori error estimates for non-linear parabolic equations

A posteriori error estimates for non-linear parabolic equations A posteriori error estimates for non-linear parabolic equations R Verfürt Fakultät für Matematik, Rur-Universität Bocum, D-4478 Bocum, Germany E-mail address: rv@num1rur-uni-bocumde Date: December 24 Summary:

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

Finite Difference Methods Assignments

Finite Difference Methods Assignments Finite Difference Metods Assignments Anders Söberg and Aay Saxena, Micael Tuné, and Maria Westermarck Revised: Jarmo Rantakokko June 6, 1999 Teknisk databeandling Assignment 1: A one-dimensional eat equation

More information

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

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

More information

SMAI-JCM SMAI Journal of Computational Mathematics

SMAI-JCM SMAI Journal of Computational Mathematics SMAI-JCM SMAI Journal of Computational Matematics Compatible Maxwell solvers wit particles II: conforming and non-conforming 2D scemes wit a strong Faraday law Martin Campos Pinto & Eric Sonnendrücker

More information

CHAPTER 4. Elliptic PDEs

CHAPTER 4. Elliptic PDEs CHAPTER 4 Elliptic PDEs One of te main advantages of extending te class of solutions of a PDE from classical solutions wit continuous derivatives to weak solutions wit weak derivatives is tat it is easier

More information