Identification of the diffusion coefficient in a nonlinear parabolic PDE and numerical methods

Size: px
Start display at page:

Download "Identification of the diffusion coefficient in a nonlinear parabolic PDE and numerical methods"

Transcription

1 ANZIAM J. 57 (E) pp.e E38, 26 E Identification of the diffusion coefficient in a nonlinear parabolic PDE and numerical methods Sameh Abidi Jamil Satouri 2 (Received 26 September 24; revised 27 April 26) Abstract This paper presents an iterative method to identify the diffusion in a semi-linear parabolic problem. This method can be generalized to other kind of problems, elliptic, parabolic and hyperbolic in twodimensional and three-dimensional case. The diffusion is obtained by solving an optimal control problem. By imposing specific conditions to the data, we build a sequence of linear problems which converge to the exact solution. We discretize our problem by a finite element method in the first case and a spectral method in the second case, using the sensibility method for approximating the gradient of the functional. Some numerical experiments prove the efficiency of this method. Keywords: Parabolic equation, diffusion, optimization, finite-elements, spectral method gives this article, c Austral. Mathematical Soc. 26. Published May 6, 26. issn (Print two pages per sheet of paper.) Copies of this article must not be made otherwise available on the internet; instead link directly to this url for this article.

2 Contents E2 Contents Introduction E2 2 The linearized problems E4 2. The Assumptions E4 2.2 Linearized problems E5 2.3 The variational Formulation of (P n ) E6 3 Continuity of the map Φ E6 4 The convergence E 5 Optimal control E5 6 The discrete problem E6 6. Finite element method E6 6.. The variational Problem E Calculating the gradient of J n E9 6.2 Spectral discretization of the problem (PV n ) E Calculating the gradient of J n E26 7 Some numerical experiments E27 References E35 Introduction In the phenomena of exchange at the sediment-water interface of the seabed or lagoons, the concentration u(x, t) where x is the spatial coordinate system and t is the time, is governed by the one-dimensional diffusion equation u t v u x x ( D u x ) = f(x, t, u),

3 Introduction E3 where D(x, t) is the diffusion coefficient that accounts both molecular diffusion and dispersion kinetics [4] and f is the production-destruction term. Defining Q T = Ω ], T[, for open interval Ω =], [ and T is a positive real number, χ T is the concentration at the time T, and v is the concentration at time t =. Neglecting the convective term, the problem in this paper is to find the coefficient of diffusion D minimizing the functional J(D) = 2 [u(x, T; D) χ T (x)] 2 dx, where for a given D, u(x, t; D) is the solution of the problem Find u(x, t) : Q T R, such that u t ( D u ) = f(x, t, u), (x, t) Q T, x x u(x, ) = v (x), x Ω, (P D ) u(, t) = u(, t) = g(t), t ], T[, where f, g and v > are given. Thus the problem min J(D) such that D U ad, u(x, t; D) solves (P D ), (Q D ) is an optimization problem. A similar problem in linear case was first studied by Bouchiba and Abidi [3]. However, here and elsewhere the formulation that is considered in this work deals with the nonlinear source term f. There are Several applications in this field, discussing parabolic and hyperbolic, linear and nonlinear systems describing numerical simulation results. These include control of distributed parameter systems [5, 6, 7], regularization for highly ill-posed distributed parameter estimation problems [4]. This performed numerical analysis based on an iterative Gauss Newton method and solved by Levenberg Marquardttype method, which applied to the output least squares formulation.

4 2 The linearized problems E4 In section 2, we study after linearizing, the existence and the uniqueness of a weak solution of the linearized problem of (P n ) for a fixed D. In section 3, we prove the continuity of the map Φ : D u(x, t, D). In section 4, we establish the convergence of the sequence (u n ) n N where u n is solution of (P n ) to u solution of (P D ). In section 5, we describe the optimal control problem under constraints (P D ) []. Indeed we prove the existence of a global optimal solution for the optimal control problem (Q) [, 2]. In section 6, we give a numerical algorithm to solve (P n ) by computing the gradient of the functional J [4], [3, linear version]. The numerical solution of the pde system is generated by both the finite elements method [7] and the spectral method [8, 9]. Finally in Section 7, we present a comparative study between the two methods (finite elements method and spectral method) to see the best convergence. 2 The linearized problems 2. The Assumptions Let O be an open set of R d, let p [, + ] and let k be a natural number. Define the Sobolev space W k,p (O) = {u L p (O) : α u L p (O) for all α k}, where α is a multi-index, and α is a partial derivative of u in the weak sense. We suppose that the nonlinear source term f in the first equation of (P D ), satisfies the following assumptions: f C ( ) 2 Ω ], ( ) T[ R W 2, Ω ], T[ R, (H) f (x, t, s), for all s R and (x, t) Ω ], T[, (H2) s s f(,, s) is concave on R, (H3) f(x, t, ) >. (H4)

5 2 The linearized problems E5 2.2 Linearized problems Using the first equation of (P D ), we denote the residual R(ψ) = ψ t ( D ψ ) f(,, ψ), x x so that u is a solution of R(u) =. By applying Newton s method linearized which has the advantage to converge rapidly towards our solution, we build a recurrent sequence of functions (u n ) n with u n : Q T R, for n N, u n = u n R(u n) R (u n ). When u is the first term, then u n t ( D u ) n f x x u (,, u n )u n = f u (,, u n )u n +f(,, u n ). We obtain a sequence of linear problems: initially u t ( D u ) = k, in Q T x x u (x, ) = v, x Ω, (P ) u (, t) = u (, t) = g(t), t ], T[, where k = f. For n, u n t ( D u ) n f x x u (x, t, u n )u n = F(x, t, u n ), in Q T, u n (x, ) = v, x Ω, (P n ) where u n (, t) = u n (, t) = g(t), t ], T[, F(x, t, u n ) = f(x, t, u n ) f u (x, t, u n )u n.

6 3 Continuity of the map Φ E6 2.3 The variational Formulation of (P n ) Let O be an open set of R d and let k be a natural number. Define the Sobolev space H k (O) = { u L 2 (O) : α u L 2 (O) for all α N d, α k } For a fixed D, the variational formulation of problem (P n ) is Find u n (, t) H (Ω), with u n (, t) ḡ(, t) H (Ω), such that d dt (u n(, t), v) + a D (u n (, t), v) = (F(, t, u n ), v), for all v H (Ω), u n (, ) = v, (PV n ) where u n (, t) : Ω R and ḡ(, t) is the lifting of the boundary condition g(t) in H (Ω) and coincides with g(t) on {, }, the boundary condition is reflected by the fact that u n (, t) ḡ(, t) H (Ω) [9, Proposition.2]. The bilinear form, for all v H (Ω), a D (u n, v) = D u n(, t) x v x dx f u (, t, u n )u n (, t)v dx. So for F(,, u n ) L 2 (, T, L 2 (Ω)) and ḡ(, t) continuous in Ω, v L 2 (Ω) and D L (Q T ) such that D α >, problem (PV n ) admits a unique solution [, 2], u n L 2 (, T, H (Ω) ) C (, T, L 2 (Ω) ). 3 Continuity of the map Φ The following Lemma is needed to prove the continuity of the map Φ. Lemma. If function u is a solution of problem (P D ), then u is bounded in L 2 (, T, H (Ω)).

7 3 Continuity of the map Φ E7 Proof: The variational formulation of problem (P D ) is Find u(, t) H (Ω); u(, t) ḡ(, t) H (Ω) such that d dt (u(, t), v) + ã D(u(, t), v) = (f(, t, u), v), for all v H (Ω), (PV D ) u(, ) = v, where ã D (u(, t), v) = D u(,t) v dx. We denote ū(, t) = u(, t) ḡ(, t). x x Then u(, t) is solution of problem (PV D ) if and only if ū(, t) is solution of the problem Find ū(, t) H (Ω) such that, for all v H (Ω), d dt (ū(, t), v) + ã D(ū(, t), v) = (f(, t, ū + ḡ(t)), v) ã D (ḡ(, t), v), u(, ) = v. () Using () in the particular case for v = ū(, t) H (Ω), such that D α > and the Cauchy Schwarz inequality combined with Poincare inequality we prove that u is bounded in L 2 (, T, H (Ω)). Let the map Φ : L (Q) L 2 (, T, H (Ω) ) D u where u is the unique solution of (P D ). Let U = {ϑ L (Q T ) : there exists α\ϑ α > }. For D, D 2 U, D D 2, we denote u i = Φ(D i ) i=,2 and f i (x, t) = f(x, t, u i ) i=,2, for all (x, t) Q T. Proposition 2. The map Φ is continuous u u 2 L2 (,T,H (Ω)) C D D 2 L (Q T ), (2) where C is a constant independent of D and D 2.

8 3 Continuity of the map Φ E8 Proof: We denote ū = ū ū 2. Using (2), after subtracting for all v H (Ω), ( ū (, t) = t ), v + (D D 2 ) ū (, t) x v x dx + D 2 ū (, t) x v x dx ( f (, t) f 2 (, t) ) v dx. (3) In particular, equation (3) is true for v = ū (, t) H (Ω) and after integration between and t ( t ) 2 ū (, t) 2 L 2 (Ω) + ū 2 (, t) D 2 x dx dσ t [ = ( f (, t) f 2 (, t) ) ] ū (, t) dx dσ t [ + (D 2 D ) ū ] (, t) ū (, t) dx dσ. x x We denote A = 2 ū (, t) 2 L 2 (Ω), B = t ( D 2 ū (, t) x 2 dx ) dσ. Then, as D α >, ( t ) ū 2 (, t) x dx dσ α t + α [ t [ ( f (, t) f 2 (, t) ) ] ū (, t) dx dσ (D 2 D ) ū (, t) x ū (, t) x dx ] dσ. (4)

9 3 Continuity of the map Φ E9 And also, 2 ū (, t) 2 L 2 (Ω) t + [ t ( f (, t) f 2 (, t) ) ] ū (, t) dx dσ [ (D 2 D ) ū ] (, t) ū (, t) dx dσ. x x The inequality (4) is true for all t ], T[. Recall the assumption (H) and using Taylor s formula, we obtain the following inequality which is related to B defined above ū 2 (, t) x L 2 (,T,L 2 (Ω)) α D 2 D L (Q T ) ū (, t) x L 2 (,T,L 2 (Ω)) + c α D 2 D L (Q T ) ū L 2 (,T,L 2 (Ω)) By Lemma, ū (, t) x 2 We obtain ū (, t) x L 2 (,T,L 2 (Ω)) 2 L 2 (,T,L 2 (Ω)) ū (, t) x α c D 2 D ū (, t) L (Q T ) x L 2 (,T,L 2 (Ω)) L 2 (,T,L 2 (Ω)) + c α D 2 D L (Q T ) ū (, t) L 2 (,T,L 2 (Ω)). C D 2 D L (Q T ) ū (, t) L2 (,T,H (Ω)). Similarly to (4), the following inequality in terms of the defined variable A, 2 ū (, t) 2 L 2 (Ω) C D 2 D L (Q T ) ū (, t) L2 (,T,H (Ω)).

10 4 The convergence E After integration between and T, ū (, t) 2 L 2 (,T,L 2 (Ω)) 2TC D 2 D L (Q T ) ū (, t) L2 (,T,H (Ω)). We set C = 2 sup(c, 2TC ), and conclude that ū (, t) L2 (,T,H (Ω)) C D 2 D L (Q T ), and finally u u 2 L2 (,T,H (Ω)) C D D 2 L (Q T ). 4 The convergence Remark 3. The symbol X Y denotes the continuous and dense embedding of X into Y. The symbol X Y denotes the weak convergence of X to Y. The symbol X Y denotes the strong convergence of X to Y. Consider now the sequence (u n ) n N where u n is solution of (P n ). Proposition 4. The sequence {u n } n= satisfies < u n+ u n u.

11 4 The convergence E Proof:. We first prove that u n+ u n. Let ω (x, t) = u (x, t) u (x, t), for all (x, t) Q T, with u is the solution of (P ) and u is the solution of P, (P n ) for n =, then u t ( D u ) x x u t ( D u ) = k. x x f u (,, u )u = F(,, u ), After subtracting (P ) and P, (P n ) for n =, we get ω t ( D ω ) f x x u (,,, u )ω = f(,, u ) k. Using the assumption (H2), the difference field ω via the maximum principle []. Subtracting P n+ and (P n ), setting ω n = u n+ u n, with u n is the solution of (P n ) and u n+ is the solution of P n+, (P n ) for n n +, ω n t x ( D ω ) n f x u (,, u n)ω n = G n, where via (H3), G n = f(,, u n ) f(,, u n ) f u (,, u n )ω n. Then, using the assumption (H2) the maximum principle gives ω n, for all n. 2. We prove now that u n > for all n. In problem (P ), k = f > and v > by the maximum principle u >. In problem P n+, (P n ) for n n +, under assumption (H3) and (H4) F(,, u n ) >. Using the assumption (H2) and v >, then via the maximum principle u n > for all n.

12 4 The convergence E2 Theorem 5. The sequence (u n ) n N where u n is solution of (P n ) converges to u, where u is a solution of (P D ). Proof: By Lemma, u n is bounded in L 2 (, T, H (Ω)) then, we can extract a subsequence still denoted u n such that u n u in L 2 (, T, H (Ω) ) weakly. (5) Using (H4), we prove that u >. But it appears difficult because from (5), it is generally not guaranteed that ( D u ) n ( D u ). x x x x Then, we increase the regularity of u using the singular perturbation method. For that we introduce a new bilinear form, continuous on H 2 (Ω), such that (ϕ, ψ) b(ϕ, ψ), b(ψ, ψ) β ψ 2 H 2 (Ω), for all ψ H2 (Ω). Using Riez s Theorem, we represent the bilinear form b by an operator B such that b(ϕ, ψ) = Bϕ, ψ H 2 (Ω),H 2 (Ω). Similarly for the bilinear form a D : it is represented by the operator A(D) such that a D (ϕ, ψ) = A(D)ϕ, ψ H (Ω),H (Ω). For ε sufficiently small and positive we denote a ε,d (ϕ, ψ) = εb(ϕ, ψ) + a D (ϕ, ψ), for all ϕ, ψ H 2 (Ω). We verify that a ε,d (ψ, ψ) ε ψ 2 H 2 (Ω) + c ψ 2 H (Ω),

13 4 The convergence E3 where c is a positive constant. The singular perturbation problem is u n,ε + (εb + A(D))u n,ε = F(,, u n ) in Q T, t u n,ε (x, ) = v (x) in Ω, (P n,ε ) where u n,ε L 2 (, T, H 2 (Ω)). The operator + (εb + A(D)) is the singular t perturbation operator of + A(D). By applying the theorem of Lions [2] t such that u n,ε H 2 (Ω) and a D (u n, v) is replaced by a ε,d (u n, v) and using the assumption of uniform coercivity, a ε,d (ψ, ψ) ε ψ 2 H 2 (Ω) + c ψ 2 H (Ω), (6) problem (P n,ε ) admits a unique solution u n,ε. Proposition 6. As ε the following strong convergences hold: u n,ε u in L 2 (, T, H (Ω)); εu n,ε in L 2 (, T, H 2 (Ω)); t u n,ε t u in L2 (, T, H 2 (Ω) ). Since ε is fixed positive, we deduce from (6) that u n is bounded in L 2 (, T, H 2 (Ω) ). (7) From the first equation of (P n,ε ), we prove that t u n,εis bounded in L 2 (, T, H 2 (Ω) ). (8) From (7), we extract a sequence, still denoted by u n, such that u n converge in a weak sense to u in L 2 (, T, H 2 (Ω)). Since the injection of H 2 (Ω) in H (Ω) is compact, there exists a subsequence, still denoted u n, which converge strongly to u in L 2 (, T, H (Ω)).

14 4 The convergence E4 From (8), we conclude that u n converge weakly to u t t then D u n x D u in Q T, x in a weak sense. Indeed, for all ψ D(Q T ), X = (x, t) Q T, D u n Q T x ψ dx = D u ( Q T x ψ dx + un D Q T x u x We conclude that because Q T D Q T D u n x ψ dx Q T D u x ψ dx, in L2 (, T, H 2 (Ω) ), ) ψ dx. ( un x u ) ψ dx x C u n x u x ψ L 2 (Q T ), L 2 (Q T ) and Since f(q T ) is C 2, then u n x u x. L 2 (Q T ) f u (X, u n) f (X, u), u and f(x, u n ) f u (X, u n )u n f(x, u) f (X, u)u. u We conclude that t u n ( D ) x x u n + f u (X, u n)u n t u (D x ) x u + f (X, u)u, u weakly in Q T. Therefore u n satisfies the two equations of (P n,ε ).

15 5 Optimal control E5 5 Optimal control In this section we are interested by the optimal control problem under constraints (P D ) [] min J(D) = [u(x, T; D) χ T (x)] 2 dx, such that 2 u t ( D u ) = f(x, t, u), (x, t) Q T, x x u(x, ) = v (x), x Ω, (Q) u(, t) = u(, t) = g(t), t ], T[, u C. The set C is a closed convex subset of C (Ω) = {ω C( Ω) : ω = on {, }}, the space of continuous functions on Ω vanishing on {, }. Remark 7. Sometimes we adopt the notation u(x, t; D) as a reminder that u implicitly depends on D. The set of admissible solutions is defined as U ad = {D L (Q T ) : u(x, t; D) satisfies the state equation in (P D )}. The function u(x, T; D), for all (x, t) Q T, is the solution of problem (P D ) at t = T, χ T is given in L 2 (Ω). We now prove the main result of this section. Theorem 8. If U ad is non-empty, then there exists a global optimal solution for the optimal control problem (Q). Proof: Because U ad is not empty, we may take a minimizing sequence D n U ad. We obtain that D n < which implies that D n is uniformly bounded in L (Q T ). Then we may extract a weakly convergent subsequence, also denoted by D n, which converge in a weak-star sense to D L (Q T ). And we denote u n (x, t) = ū(x, t; D n ), for all (x, t) Q T, by () and using

16 6 The discrete problem E6 Lemma combined with (H), u n is uniformly bounded in H (Ω). Then we extract a weakly convergent subsequence, also denoted by u n, such that u n ū H (Ω). In order to see that (ū ; D) is a solution of the (P D ) equations, the only problem is to pass to the limit in the nonlinear form a n = f(, t, u n)v dx of (). Due to the compact embedding of H (Ω) into L 2 (Ω) [2] and the continuity of the a n, and without forgetting that a n is bounded for all x, t and u n, it follows that f(, t, u n )v dx f(, t, u)v dx. Consequently, taking into account the linearity and continuity of all terms involved, the limit (ū ; D) satisfies the state equations. Since C is convex and closed, it is weakly closed, so u n ū H (Ω) and the embedding H (Ω) C (Ω) imply that ū C. Taking into consideration that J(D) is weakly lower semi continuous, then there exists a global optimal solution for the optimal control problem (Q). 6 The discrete problem 6. Finite element method In this section, to simplify, we study the problem (PV n ) in the case of homogeneous boundary conditions (g = ). 6.. The variational Problem We put H = L 2 (Ω), V = H (Ω), Ω =], [. Then V H V. We consider the variational problem (PV n ) which admits a unique solution u n L 2 (, T, H (Ω) ) C ([, T], L 2 (Ω)).

17 6 The discrete problem E7 To build an approximation u n,h of u n, we choose a subspace of V composed of affine functions over an equal number of intervals of equal size. More precisely, let N be a natural number and h =, i N +. We N+ denote x i = ih, i N +, that subdivides the interval Ω = [, ] to N + intervals K i = [x i, x i+ ], i N, each of length h and V h = {v C ( Ω) : v() = v() =, v Ki P, i N}, a finite dimensional subspace of V []. Setting F n (t) = F(, t, u n ), the approached problem is Find u n,h (, t) V h, such that for all t [, T], d dt u n,h(, t), v h + a D (u n,h (, t), v h ) = F n (t), v h for all v h V h, u n,h (, ) = v,h. (PV nh ) The problem (PV nh ) admits an unique solution u n,h (, t) V h : the proof is obtained by direct application of Lions s Theorem [5]. We introduce the basis (ϕ i ) i N of V h, for j N. Then there exists u D n,,j and ud n,j functions of [, T] in R, such that, for all (x, t) Q T, v,h (x, t) = u D n,,j(t)ϕ j (x), u n,h (x, t) = u D n,j(t)ϕ j (x). j= Then problem (PV nh ) is equivalent to the differential system (ϕ j, ϕ i ) d dt ud n,j(t) + a D (ϕ j, ϕ i )u D n,j(t) = F n,i (t), i N, j= j= u D n,i() = u D n,,i, i N, (S ) where F n,i (t) = (F n (t), ϕ i ). We denote j= R D (t) = (a D (ϕ j, ϕ i )) i,j N the stiffness matrix, M = (ϕ j, ϕ i ) i,j N the mass matrix, u D n (t) = (u D n,(t), u D n,2(t),..., u D n,n(t)) T, and F n (t) = (F n, (t), F n,2 (t),..., F n,n (t)) T.

18 6 The discrete problem E8 The system (S ) is written as M d dt u D n (t) + R D (t)u D (t) = F n (t), n u D n () = u D n,. (S 2 ) A simple calculation gives the elements of M as (ϕ i, ϕ i ) = 2h 3, i N, We put Setting K n (t) = K n,i+ 2 K + n,i 2 (ϕ i, ϕ i+ ) = h 6, (ϕ i+, ϕ i ) = h, i N, 6 (ϕ i, ϕ j ) =, i j >. D i+ 2 = xi+ x i D(t) dx. f (, t, u u n )u n (, t)v dx, then = = xi+ x i K n (t)(x i+ x) 2 dx, xi x i K n (t)(x x i ) 2 dx, xi+ K n,i+ 2 = K n (t)(x i+ x)(x x i ) dx. x i We obtain for the coefficients of R D : a D (ϕ i, ϕ i ) = h 2 (D i+ 2 + D i 2 + K n,i+ 2 + K + n,i 2 ), i N, a D (ϕ i+, ϕ i ) = h 2 (D i+ 2 + K n,i+ 2 ), a D (ϕ i, ϕ i+ ) = h (D 2 i+ + K 2 n,i+ ), i N, 2 a D (ϕ i, ϕ j ) =, i j >. (9)

19 6 The discrete problem E9 Using the modified Euler method of integration, K n (t)ϕ 2 i dx = ( ) K + + K h 2 n,i 2 n,i+ 2 = 2h f ( xi + h, t, u D f 3 u n,i) + ( u + h f 6 u ( + f u x i + h 2, t, ud n,i + ud n,i 2 x i+ + h 2, t, ud n,i+ + ud n,i+2 2 ( xi+ + h, t, u D n,i+2) ) ), () and ( K n (t)ϕ i ϕ i+ dx = h f x i + h 24 u 2, t, ud n,i + ) ud n,i+. () 2 The system (S 2 ) is solved using a fourth order Runge Kutta method Calculating the gradient of J n Consider the problem min J n (D), D U ad, u n (,.; D) solves (P n ), (Q n ) where we define the functional J n (D) = 2 [u n (x, T; D) χ T,n (x)] 2 dx. (2) As χ T,n = N i= χ n,t (x i )ϕ i, we write J n (D) = [ N ( u D 2 n,i (T) χ n,t (x i ) ) ] 2 ϕ i (x) dx. (3) i=

20 6 The discrete problem E2 Expand the expression of J, J n (D) = h ( u D 3 n,i (T) χ n,t (x i ) ) 2 i= + h 6 N ( u D n,i+ (T) χ n,t (x i+ ) ) ( u D n,i(t) χ n,t (x i ) ). i= To calculate an approximation of the gradient J, we consider a mesh of the rectangle Ω [, T] by triangles T ij with vertices and T ij with vertices where δt = (ih, jδt), ((i + )h, jδt), (ih, (j + )δt), (ih, (j + )δt), (ih, (j + )δt), ((i + )h, jδt), T is a time step. There D M+ i,j D(ih, jδt). And we denote λ i respectively (λ i ), i =, 2, 3, the barycentric coordinates of a point from the triangle T ij, respectively T A simple calculation gives ij, D ij Tij = λ D ij + λ 2 D i+,j + λ 3 D i,j+, D ij T ij = λ D i+,j + λ 2D i,j+ + λ 3D i+,j. λ (x, t) = x h + t t (i + j + ) = λ (x, t), λ 2(x, t) = x h + i + = (λ 2(x, t) ), λ 3(x, t) = t t + j + = (λ 3(x, t) ). The calculation of D i+ 2 is more delicate when x varies between x i and x i+ and t between jδt and (j + )δt the point (x, t) T ij T ij so if x [x i, x i + a(t)] then (x, t) T i, if x [x i + a(t), x i ] then (x, t) T i.

21 6 The discrete problem E2 Figure : Mesh where a(t) = h( s j) (see Figure ), with s = t/δt so D i+ 2 = h 2 [α j(t)d ij + β j (t)d i+,j + γ j (t)d i,j+ + δ j (t)d i+,j+ ], (4) with α j (t) = (+j s) 2, β j (t) = (+j s)( j+s), γ j (t) = (s j)(2+j s), and δ j (t) = (s j) 2. Thus, knowledge of D ij, i N +, j M +, approach D(x, t) on Q T and consequently reduce the problem of minimization in finite dimensions: D will equate to a vector R k where k = (N + 2)(M + 2). We denote D = (D,..., D (N+), D,..., D (N+),..., D (M+),..., D (N+)(M+) ) T. The problem (Q n ) becomes min J n (D), D R k, D α. (Q nh )

22 6 The discrete problem E22 Then we obtain J n (D) = ( Jn D, J n D,..., J n D (N+),..., J n D (N+)(M+) ) T. As and J n D ij = J n u D k= n,k u D n,k D ij (t), ( ) T u D n u D n, (t) =, ud n,2,..., ud n,n. D ij D ij D ij D ij The expression (3) gives J n = 2 u n,k 3 h [ u D k,n(t) χ T,n (x k ) ] + h [ u D 6 k,n (T) χ T,n (x k ) ] + h 6 ud k+,n(t) χ T,n (x k+ ). Knowing that χ T (x ) = χ T (x N+ ) = and u D J n D ij, we assume = u D N+ =, to calculate ( ) T v ij (t) = u D n u D n, (t) =, ud n,2,..., ud n,n. D ij D ij D ij D ij Then By using (S 2 ), it follows v ij t = D ij ( ) u D n. t v ij t = M R D u D n D ij (t) M R D (t) u D n D ij (t).

23 6 The discrete problem E23 To calculate (v ij ), it suffices to solve the following system, using the fourthorder Runge Kutta method, v ij t = M R D D ij u D n (t) M R D (t)v ij, v ij () =. The matrix coefficients R D (t) and R D(t) D ij are calculated using (). 6.2 Spectral discretization of the problem (PV n ) In this section we recall some formulas in the spectral method in a reference field Λ =], [, then we use these results to write the variational formulation in our field of study Ω. Let N be an integer 2. Denote the space P N (Λ) of polynomials with degree N and the space P N (Λ) of polynomials in P N(Λ) vanishing on the boundary of Λ [8, 9]. We introduce the space P N (, ) of restrictions to Λ of polynomials with degree N. Setting ξ = and ξ N =, we introduce the N nodes ξ j, j N, and the N + weights ρ j, j N, of the Gauss Lobatto quadrature formula, recall that φ(ζ)dζ = φ(ξ j )ρ j. (5) We also recall [6, (3.2)] the following property, which is useful in what follows: for all ϕ N P N (, ), j= N ϕ N 2 L 2 (Λ) ϕ 2 N(ζ)ρ j 3 ϕ N 2 L 2 (Λ). (6) Relying on this formula, we introduce the discrete product [8, 9] defined on continuous functions u and v by (u, v) = u(ξ j )v(ξ j )ρ j. (7) j= j=

24 6 The discrete problem E24 It follows from (6) that this discrete product is a scalar product on P N (Λ). Let finally I N denote the Lagrange interpolation operator at the nodes ξ i, i N, with values in P N (Λ). We now assume that the function F n (t) is continuous on Ω. Thus the discrete problem is constructed from (PV n ) using the Galerkin method combined with numerical integration. It reads Find u n,n (, t) P N (Ω), with u n,n (, t) I N ḡ(, t) P N(Ω), such that for all v N P N(Ω) d dt (u n,n(, t), v N ) N + a D,N (u n,n (, t), v N ) = (F n (t), v N ) N, u n,n (, ) = v,n, (8) where the bilinear form ( u n,n (, t) a D,N (u n,n, v N ) = D N, v ) N. (9) x x N It follows from (PV n ), combined with Cauchy Schwarz inequalities, that the form a D,N is continuous on P N (Ω) P N (Ω), with norm bounded independently of N. Then to investigate the well-posedness of the problem we use Lions s theorem [2, 5]. To simplify the problem, we suppose that we know the values of F n (t) at the nodes of Ω. Let l i be Lagrange polynomials at the nodes ξ i, i N [6, Remark.5]. Then we calculate the coordinates u n,i (t) of solution u n,n (, t) in the base l i, for each t in [, T]. So We note and u n,n (x, t) = u n,i (t)l i (x). i= v N = l r, r N, D N (x, t) = m=p= D mp l m (x) l p (t).

25 6 The discrete problem E25 The variational problem (8) (l i, l r ) dt u n,i(t) + i= u n,i () = v,i, a D,N (l i, l r )u n,i (t) = F n (t)( ξ r )ρ r, (2) i= (2) where u n,i (t) are the values of u n,n (t) at the nodes ξ i of the Gauss Lobatto in Ω. We obtain N equations with N unknowns. So (9) is equivalent to the linear system M d dt u D n (t) + R D (t)u D (t) = F n (t), n u D n () = ṽ, (22) where u D n (t) is the vector of coordinates u n,i (t), i N. M is the mass matrix, R D (t) is the stiffness matrix. F n (t) is the vector of coordinates 2 ( f( ξ r, t, u n ( ξ r, t)) f s ( ξ r, t, u n ( ξ r, t))u n ( ξ r ) a D,n (l, l r )u n ( ξ, t) + a D,n (l N, l r )u n ( ξ N, t), r N. ṽ is constituted with the values v,i, i N, at the nodes ξ i of the Gauss Lobatto in Ω. The coefficients of M are written as [3, Appendix] { if i < r N, (l i, l r ) = if i = r N. 2 N(N+)L 2 N ( ξ i ) And the coefficients of R D (t) are a D,N (l i, l r ) = 2 lp (t) p= ) D kp l i( ξ k )l r( ξ k )ρ k k= f 2 s ( ξ r, t, u n ( ξ r, t))δ ir ρ r. ρ r

26 6 The discrete problem E Calculating the gradient of J n Consider the problem (Q n ) as u n,n (t) (x) = u D n,jl j (x), χ T (x) = j= χ T,i l i (x), i= and setting u D n ( ξk ) (T) = u D n,k (T), it follows J n (D) = k= [ u D n,k(t) χ T ( ξ k )] 2 ρk. (23) To calculate an approximation of the gradient J n (2), we consider a mesh of rectangle Ω [, T]. Then J n (D) D mp = We then observe that, via (23), k= J n (D) u D n,k u D n,k D mp. J n (D) u n,k = 2(u D n,k(t) χ T,k )ρ k. And then find ( u D n (t) u D n, (t) =,..., ud n,n (t) ), D mp D mp D mp we assume v m,p (t) = u D n (t) D mp, and using (22) gives v m,p t = M (R Du D n (t)) D mp, v m,p () =. (24)

27 7 Some numerical experiments E27 To solve (24), R D u D n = 2 m= p= ( N D mp and finally ( R D u D n (t) N = 2 D mp i= j= i= j= ) u ij l i( ξ m ) l j (t) l r( ξ m ) l p (t) ρ m, ) u ij l i( ξ m ) l j (t) l r( ξ m ) l p (t) ρ m. 7 Some numerical experiments In this section we present some numerical experiments using matlab as a computational tool. The algorithm of resolution of the problem (Q), is divided into two parts. Firstly we solve the problem (PV n ), in the matrix form (S 2 ) and (22) in both methods. The resolution of U is done knowing D = D(, ). The method used is a fourth order Runge Kutta. Secondly, once U is calculated, it is then introduced in the minimization problem (Q D ) to recover D. The best method used in our case is a Levenberg Marquardt method based on the calculation of gradient of J(D) detailed in subsections 6..2 and We continue this process until we get the desired convergence. The algorithm requires about 7% of total time for calculating the solution u and 93% to calculate D. Runge Kutta is used three times in the program. In the Levenberg Marquardt algorithm, the iteration number to achieve maximum convergence of D varies somewhat as function of N, and it does not exceed five iterations. We need also just a few iterations to assure the convergence of the global system. Table reports the average cpu time (in

28 7 Some numerical experiments E28 Table : The cpu time in seconds required for one iteration of Runge Kutta. N fem fem2 sp.m seconds) required for one iteration of Runge Kutta using the three methods with N varying between 8 and 36. We use the bicgstabl method (gmres method is also a suitable alternative). Several preconditioners can be found in literature. In our case a preconditioner performing an incomplete lu factorization is the best appropriate to be used for all three methods. The convergence is assured in four iterations at most. Without using any preconditioner the number of iterations increases as a function of N. For example by fixing a tolerance 6 the iteration number exceeds ten for N greater than 3. Without loss of generality, we consider the domain of study Q T. Numerically we prove the convergence of the same problem in a more general case taking u(, t) = g (t) and u(, t) = g 2 (t) where g and g 2 are given. We present a complete study of the problem, using finite element method (explained in subsection 6.). We give in this method two different approaches. In the first approach (fem) we use the trapezoidal formula for approximating integrals. Regarding the second approach (fem2) we employ the Gauss Lobatto formula. Finally, we present a spectral method (sp.m), with a comparative study for each test case.

29 7 Some numerical experiments E29. We consider the following analytical functions: u(x, t) = (x 2 + 2)(t + ), D(x, t) = (x + )(t + ), (25) 24 f(x, t) = s(x, t)u 2 + h(x, t), with s(x, t) = (t 2 + )e x, and h(x, t) verifies the equation (P D ). Figures 2 and 3 plot the two quantities log ( u u N L 2 (Ω) ) and log ( D D N L 2 (Ω) ) as functions of log (δt). We apply this for both the methods. In all these calculations we take T =. and N = We consider the following functions: u(x, t) = e t (x + ), D(x, t) = 2 e t (x 2 + ), (26) f(x, t) = s(x, t)u 2 + h(x, t), with s(x, t) = tx 2, and h(x, t) verifies the equation (P D ). Figures 4 and 5 plot the two quantities log ( u u N L 2 (Ω) ) and log ( D D N L 2 (Ω) ) as functions of log (δt). We apply this for both the methods. All these calculations are made by fixing T =. and N = 35. In both the cases (first test and second test) presented, we find that the spectral method is more precise than finite elements method. The slope of the curve estimations in approach 2 (finite element method) and the spectral method proves the good convergence in time of the solutions u and D. While in approach (finite element method) the slopes of the curve estimations decreases very slowly with time, this result is due to the poor convergence space stops the evolution in time of errors. 3. For the same example as the first case, Table 2 presents the error u u N and Figure 6 the approached solution u N for times t varying between and in the nodes ξ i, i =,..., 8. The results are for T =, δt = 3 and N = 8.

30 7 Some numerical experiments E3 Figure 2: L 2 error between the exact solution u given in the first test and the experimental solution u N using the three approaches. 2 4 log u u N L 2 (Ω) f.e.m. f.e.m.2 sp.m log (δt) Figure 3: L 2 error between the exact solution D given in the first test and the experimental solution D N. log D D N L 2 (Ω) 5 f.e.m. f.e.m.2 sp.m log (δt)

31 7 Some numerical experiments E3 Figure 4: L 2 error between the exact solution u given in the second test and the experimental solution u N. 5 log u u N L 2 (Ω) f.e.m. f.e.m.2 sp.m log (δt) Figure 5: L 2 error between the exact solution D given in the second test and the experimental solution D N. log D D N L 2 (Ω) 5 f.e.m. f.e.m.2 sp.m log (δt)

32 7 Some numerical experiments E32 Table 2: The values u(ξ i,.) u N (ξ i,.) in the third test calculated at the nodes ξ i for different values of t, with N = 8 and T =. t ξ ξ 2 ξ 3 ξ 4 ξ 5 ξ 6 ξ 7 ξ 8. e 3 e 3 e 2 7e 2 5e 2 e 2e 3e.2 e 3 e 2 2e 2 7e 2 2e 5e e e.3 e 2 2e 2 6e 2 e 5e e 2e 3e.4 e 2 4e 2 e 3e e 2e 5e 6e.5 2e 2 7e 2 2e 5e e 4e 9e e 9.6 4e 2 e 3e e 2e 6e e 9 e 9.7 6e 2 e 5e e 4e e 9 2e 9 2e 9.8 9e 2 2e 8e 2e 5e e 9 2e 9 3e 9.9 e 4e e 3e 8e e 9 3e 9 4e 9 2e 6e e 4e e 9 2e 9 4e 9 5e 9 4. We study the convergence of solutions depending on the parameter n. Figures 7 and 8 show the evolution of curves as a function of N; N varying between 6 and 24. We take T =. and δt = 2, and the exact solutions are u(x, t) = (t + )(x 2 + ) /2 D(x, t) =.25 + t (x2 + ) 3/2, (27) f(x, t) = s(x, t)u 4 + h(x, t), with s(x, t) = e t2 + (x 2 + ), and h(x, t) verifies the equation (P D ). In this case u is a singular function. The maximum error of the solutions is reached at N = 24 in approach and 2 of finite element methods, and at N = 4 in the spectral method. Knowing that, if we take u as a polynomial function, then the maximum is reached from N = 4 for the two methods.

33 7 Some numerical experiments E33 Figure 6: The curves of u N (ξ i, t) as a function of t, calculated in the third test at the nodes ξ i with N = 8, T = and δt = 3..8 Estimations by Finite elements method.6 u(.,t))in the values ξ i,i=: u(ξ 8,t) u(ξ,t) t 5. In this test, we take u f(x, t) = ( (x + )(t + ) )2 + x + 5, u(x, ) = x +, (28) { (t + ) on {} g(t) = Ω, 2(t + ) on {} Ω (29), D(x, ) = x +. The solution in this case is unknown. The calculations done with the second approach (finite element methods) and spectral method are presented in Figures 9 and. Figures 9 and are almost identical, this proves the good convergence of the two algorithms.

34 7 Some numerical experiments E34 Figure 7: L 2 error between the exact solution u given in the fourth test and the experimental solution u N. 2 f.e.m. f.e.m.2 sp.m. log u u N L 2 (Ω) N Figure 8: L 2 error between the exact solution D given in the fourth test and the experimental solution D N. 2 f.e.m. f.e.m.2 sp.m. 3 log D D N L 2 (Ω) N

35 t t References E35 Figure 9: The discrete solution u N and D N using spectral method with data f and g given in the fifth test. Sp.m isovalues of u u t.5 x x.5 isovalues of D.8.8 D t.5 x x.2 Conclusion The selection of the two algorithms introduced in the paper gives a good and stable numerical convergence of the two methods. The results of the numerical simulations that we performed confirm all theoretical results found. References [] P. A. Raviart, D. J. M. Thomas, Introduction à l analyse numérique des équations aux dérivées partielles, Masson (983). E6, E, E7 [2] J.L. Lions, E. Magenes, Problèmes aux limites non homogènes et applications, Vol., Dunod (968). E6, E3, E6, E24

36 t t References E36 Figure : The discrete solution u N and D N using finite elements method with data f and g given in the fifth test. F.e.m.2 isovalues of u u t.5 x x.5 isovalues of D.8.8 D t.5 x x.2 [3] M. Bouchiba, S. Abidi, Identification of the Diffusion in a lineair Parabolic problem, International Journal of Applied Mathematics, Vol. 23, No. 3, 2, E3, E4 [4] M. Kern, Problèmes Inverses Aspects Numériques, INRIA (23). E3, E4 [5] J. E. Rakotoson, J. M. Rakotoson, Analyse Fonctionnelle appliquée aux équations aux dérivées partielles, Presse Universitaires de France, (999). E7, E24 [6] C. Bernardi, Y. Maday, Spectral Methods, in the Handbook of Numerical Analysis V, P. G. Ciarlet & J. L. Lions eds., North-Holland, (997), E23, E24

37 References E37 [7] V. Girault, P.-A. Raviart, Finite Element Methods for Navier Stokes Equations, Theory and Algorithms, Springer-Verlag (986). E4 [8] C. Bernardi, Y. Maday, F. Rapetti, Discrétisation Variationnelles de problèmes aux limites elliptiques, Springer-Verlag (24). E4, E23 [9] C. Bernardi, Y. Maday, Approximations spectrales de problèmes aux limites elliptiques, Springer-Verlag (992). E4, E6, E23 [] I. Ekeland, R. Teman, Analyse Convexe et problèmes Variationnels, Bordas (9674). E4, E5 [] J. L. Lions, Controle Optimal de systèmes gouvernés par des équations aux dérivées partielles, Dunod, Paris, (968). E4 [2] J. Henry, Étude de la contrôlabilité de certaines équations paraboliques, Thèse d état, Université Paris VI, (978). E4 [3] J. Satouri, Méthodes déléments spectraux avec joints pour des géométries axisymétriques, Thèse de l Université Pierre et Marie Curie, Paris VI (2). E25 [4] K. Van de Doel, U. M. Ascher, On level set regularization for highly ill-posed distributed parameter systems, Journal of Computational Physics, 26, 26: E3 [5] M. Böhm, M. A. Demetriou, S. Reich, and I. G. Rosen, Model Reference Adaptive Control of Distributed Parameter Systems, SIAM J. Control Optim., 36(), 33 8; doi:.2/rnc.98 E3 [6] C. Jia, A note on the model reference adaptive control of linear parabolic systems with constant coefficients. Journal of Systems Science and Complexity 24, 7; doi:.7/s E3 [7] D. B. Pietria, M. Krstic, Output-feedback adaptive control of a wave PDE with boundary anti-damping. Automatica 5, 47 45; doi:.6/j.automatica E3

38 References E38 Author addresses. Sameh Abidi, University of Tunis El Manar, Faculty of Sciences of Tunis, Department of Mathematics, 6 Tunis, Tunisia. mailto:labidisameh@yahoo.fr. 2. Jamil Satouri, IPEIT-University of Tunis, 2 street Jawaher Lel Nahra-89, Monfleury, Tunisia. mailto:jamil.satouri@yahoo.fr

Mixed exterior Laplace s problem

Mixed exterior Laplace s problem Mixed exterior Laplace s problem Chérif Amrouche, Florian Bonzom Laboratoire de mathématiques appliquées, CNRS UMR 5142, Université de Pau et des Pays de l Adour, IPRA, Avenue de l Université, 64000 Pau

More information

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

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

More information

OPTIMAL CONTROL AND STRANGE TERM FOR A STOKES PROBLEM IN PERFORATED DOMAINS

OPTIMAL CONTROL AND STRANGE TERM FOR A STOKES PROBLEM IN PERFORATED DOMAINS PORTUGALIAE MATHEMATICA Vol. 59 Fasc. 2 2002 Nova Série OPTIMAL CONTROL AND STRANGE TERM FOR A STOKES PROBLEM IN PERFORATED DOMAINS J. Saint Jean Paulin and H. Zoubairi Abstract: We study a problem of

More information

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

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

More information

Finite Differences for Differential Equations 28 PART II. Finite Difference Methods for Differential Equations

Finite Differences for Differential Equations 28 PART II. Finite Difference Methods for Differential Equations Finite Differences for Differential Equations 28 PART II Finite Difference Methods for Differential Equations Finite Differences for Differential Equations 29 BOUNDARY VALUE PROBLEMS (I) Solving a TWO

More information

EXISTENCE OF SOLUTIONS FOR A CLASS OF VARIATIONAL INEQUALITIES

EXISTENCE OF SOLUTIONS FOR A CLASS OF VARIATIONAL INEQUALITIES Journal of Matematics and Statistics 9 (4: 35-39, 3 ISSN: 549-3644 3 doi:.3844/jmssp.3.35.39 Publised Online 9 (4 3 (ttp://www.tescipub.com/jmss.toc EXISTENCE OF SOLUTIONS FOR A CLASS OF ARIATIONAL INEQUALITIES

More information

Overlapping Schwarz preconditioners for Fekete spectral elements

Overlapping Schwarz preconditioners for Fekete spectral elements Overlapping Schwarz preconditioners for Fekete spectral elements R. Pasquetti 1, L. F. Pavarino 2, F. Rapetti 1, and E. Zampieri 2 1 Laboratoire J.-A. Dieudonné, CNRS & Université de Nice et Sophia-Antipolis,

More information

Lecture 4: Numerical solution of ordinary differential equations

Lecture 4: Numerical solution of ordinary differential equations Lecture 4: Numerical solution of ordinary differential equations Department of Mathematics, ETH Zürich General explicit one-step method: Consistency; Stability; Convergence. High-order methods: Taylor

More information

The Navier-Stokes problem in velocity-pressure formulation :convergence and Optimal Control

The Navier-Stokes problem in velocity-pressure formulation :convergence and Optimal Control The Navier-Stokes problem in velocity-pressure formulation :convergence and Optimal Control A.Younes 1 A. Jarray 2 1 Faculté des Sciences de Tunis, Tunisie. e-mail :younesanis@yahoo.fr 2 Faculté des Sciences

More information

The Mortar Finite Element Method for Contact Problems

The Mortar Finite Element Method for Contact Problems The Mortar Finite Element Method for Contact Problems F. Ben Belgacem, P. Hild, P. Laborde Mathématiques pour l Industrie et la Physique, Unité Mixte de Recherche CNRS UPS INSAT (UMR 5640), Université

More information

Simple Examples on Rectangular Domains

Simple Examples on Rectangular Domains 84 Chapter 5 Simple Examples on Rectangular Domains In this chapter we consider simple elliptic boundary value problems in rectangular domains in R 2 or R 3 ; our prototype example is the Poisson equation

More information

Error estimates for the Raviart-Thomas interpolation under the maximum angle condition

Error estimates for the Raviart-Thomas interpolation under the maximum angle condition Error estimates for the Raviart-Thomas interpolation under the maximum angle condition Ricardo G. Durán and Ariel L. Lombardi Abstract. The classical error analysis for the Raviart-Thomas interpolation

More information

A penalty algorithm for the spectral element. discretization of the Stokes problem

A penalty algorithm for the spectral element. discretization of the Stokes problem A penalty algorithm for the spectral element discretization of the Stokes problem Christine Bernardi, Adel Blouza, Nejmeddine Chorfi 3, and Nizar Kharrat 4 Abstract: The penalty method when applied to

More information

Optimal Control Approaches for Some Geometric Optimization Problems

Optimal Control Approaches for Some Geometric Optimization Problems Optimal Control Approaches for Some Geometric Optimization Problems Dan Tiba Abstract This work is a survey on optimal control methods applied to shape optimization problems. The unknown character of the

More information

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

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

More information

Juan Vicente Gutiérrez Santacreu Rafael Rodríguez Galván. Departamento de Matemática Aplicada I Universidad de Sevilla

Juan Vicente Gutiérrez Santacreu Rafael Rodríguez Galván. Departamento de Matemática Aplicada I Universidad de Sevilla Doc-Course: Partial Differential Equations: Analysis, Numerics and Control Research Unit 3: Numerical Methods for PDEs Part I: Finite Element Method: Elliptic and Parabolic Equations Juan Vicente Gutiérrez

More information

arxiv: v1 [math.na] 27 Jan 2016

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

More information

256 Summary. D n f(x j ) = f j+n f j n 2n x. j n=1. α m n = 2( 1) n (m!) 2 (m n)!(m + n)!. PPW = 2π k x 2 N + 1. i=0?d i,j. N/2} N + 1-dim.

256 Summary. D n f(x j ) = f j+n f j n 2n x. j n=1. α m n = 2( 1) n (m!) 2 (m n)!(m + n)!. PPW = 2π k x 2 N + 1. i=0?d i,j. N/2} N + 1-dim. 56 Summary High order FD Finite-order finite differences: Points per Wavelength: Number of passes: D n f(x j ) = f j+n f j n n x df xj = m α m dx n D n f j j n= α m n = ( ) n (m!) (m n)!(m + n)!. PPW =

More information

A mixed finite element approximation of the Stokes equations with the boundary condition of type (D+N)

A mixed finite element approximation of the Stokes equations with the boundary condition of type (D+N) wwwijmercom Vol2, Issue1, Jan-Feb 2012 pp-464-472 ISSN: 2249-6645 A mixed finite element approximation of the Stokes equations with the boundary condition of type (D+N) Jaouad El-Mekkaoui 1, Abdeslam Elakkad

More information

Research Article A Two-Grid Method for Finite Element Solutions of Nonlinear Parabolic Equations

Research Article A Two-Grid Method for Finite Element Solutions of Nonlinear Parabolic Equations Abstract and Applied Analysis Volume 212, Article ID 391918, 11 pages doi:1.1155/212/391918 Research Article A Two-Grid Method for Finite Element Solutions of Nonlinear Parabolic Equations Chuanjun Chen

More information

Spectral discretization of Darcy s equations coupled with the heat equation

Spectral discretization of Darcy s equations coupled with the heat equation Spectral discretization of Darcy s equations coupled with the heat equation Christine Bernardi, Sarra Maarouf, Driss Yakoubi To cite this version: Christine Bernardi, Sarra Maarouf, Driss Yakoubi. Spectral

More information

Block-Structured Adaptive Mesh Refinement

Block-Structured Adaptive Mesh Refinement Block-Structured Adaptive Mesh Refinement Lecture 2 Incompressible Navier-Stokes Equations Fractional Step Scheme 1-D AMR for classical PDE s hyperbolic elliptic parabolic Accuracy considerations Bell

More information

Unconditionally stable scheme for Riccati equation

Unconditionally stable scheme for Riccati equation ESAIM: Proceedings, Vol. 8, 2, 39-52 Contrôle des systèmes gouvernés par des équations aux dérivées partielles http://www.emath.fr/proc/vol.8/ Unconditionally stable scheme for Riccati equation François

More information

STOKES PROBLEM WITH SEVERAL TYPES OF BOUNDARY CONDITIONS IN AN EXTERIOR DOMAIN

STOKES PROBLEM WITH SEVERAL TYPES OF BOUNDARY CONDITIONS IN AN EXTERIOR DOMAIN Electronic Journal of Differential Equations, Vol. 2013 2013, No. 196, pp. 1 28. ISSN: 1072-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu ftp ejde.math.txstate.edu STOKES PROBLEM

More information

Weak Solutions to Nonlinear Parabolic Problems with Variable Exponent

Weak Solutions to Nonlinear Parabolic Problems with Variable Exponent International Journal of Mathematical Analysis Vol. 1, 216, no. 12, 553-564 HIKARI Ltd, www.m-hikari.com http://dx.doi.org/1.12988/ijma.216.6223 Weak Solutions to Nonlinear Parabolic Problems with Variable

More information

Numerical Analysis: Interpolation Part 1

Numerical Analysis: Interpolation Part 1 Numerical Analysis: Interpolation Part 1 Computer Science, Ben-Gurion University (slides based mostly on Prof. Ben-Shahar s notes) 2018/2019, Fall Semester BGU CS Interpolation (ver. 1.00) AY 2018/2019,

More information

Existence of minimizers for the pure displacement problem in nonlinear elasticity

Existence of minimizers for the pure displacement problem in nonlinear elasticity Existence of minimizers for the pure displacement problem in nonlinear elasticity Cristinel Mardare Université Pierre et Marie Curie - Paris 6, Laboratoire Jacques-Louis Lions, Paris, F-75005 France Abstract

More information

On an Approximation Result for Piecewise Polynomial Functions. O. Karakashian

On an Approximation Result for Piecewise Polynomial Functions. O. Karakashian BULLETIN OF THE GREE MATHEMATICAL SOCIETY Volume 57, 010 (1 7) On an Approximation Result for Piecewise Polynomial Functions O. arakashian Abstract We provide a new approach for proving approximation results

More information

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

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

More information

Spectral element schemes for the. Korteweg-de Vries and Saint-Venant equations

Spectral element schemes for the. Korteweg-de Vries and Saint-Venant equations Spectral element schemes for the Korteweg-de Vries and Saint-Venant equations R. PASQUETTI a a. Université Côte d Azur, CNRS, Inria, LJAD, France, richard.pasquetti@unice.fr... Résumé : Les sytèmes hyperboliques

More information

2.29 Numerical Fluid Mechanics Spring 2015 Lecture 9

2.29 Numerical Fluid Mechanics Spring 2015 Lecture 9 Spring 2015 Lecture 9 REVIEW Lecture 8: Direct Methods for solving (linear) algebraic equations Gauss Elimination LU decomposition/factorization Error Analysis for Linear Systems and Condition Numbers

More information

INTRODUCTION TO FINITE ELEMENT METHODS

INTRODUCTION TO FINITE ELEMENT METHODS INTRODUCTION TO FINITE ELEMENT METHODS LONG CHEN Finite element methods are based on the variational formulation of partial differential equations which only need to compute the gradient of a function.

More information

LECTURE 3: DISCRETE GRADIENT FLOWS

LECTURE 3: DISCRETE GRADIENT FLOWS LECTURE 3: DISCRETE GRADIENT FLOWS Department of Mathematics and Institute for Physical Science and Technology University of Maryland, USA Tutorial: Numerical Methods for FBPs Free Boundary Problems and

More information

Glowinski Pironneau method for the 3D ω-ψ equations

Glowinski Pironneau method for the 3D ω-ψ equations 280 GUERMOND AND QUARTAPELLE Glowinski Pironneau method for the 3D ω-ψ equations Jean-Luc Guermond and Luigi Quartapelle 1 LIMSI CNRS, Orsay, France, and Dipartimento di Fisica, Politecnico di Milano,

More information

i=1 α i. Given an m-times continuously

i=1 α i. Given an m-times continuously 1 Fundamentals 1.1 Classification and characteristics Let Ω R d, d N, d 2, be an open set and α = (α 1,, α d ) T N d 0, N 0 := N {0}, a multiindex with α := d i=1 α i. Given an m-times continuously differentiable

More information

Une méthode de pénalisation par face pour l approximation des équations de Navier-Stokes à nombre de Reynolds élevé

Une méthode de pénalisation par face pour l approximation des équations de Navier-Stokes à nombre de Reynolds élevé Une méthode de pénalisation par face pour l approximation des équations de Navier-Stokes à nombre de Reynolds élevé CMCS/IACS Ecole Polytechnique Federale de Lausanne Erik.Burman@epfl.ch Méthodes Numériques

More information

Basics and some applications of the mortar element method

Basics and some applications of the mortar element method GAMM-Mitt. 28, No. 2, 97 123 (2005) Basics and some applications of the mortar element method Christine Bernardi 1, Yvon Maday 1, and Francesca Rapetti 2 1 Laboratoire Jacques-Louis Lions, C.N.R.S. & université

More information

Weak Controllability and the New Choice of Actuators

Weak Controllability and the New Choice of Actuators Global Journal of Pure and Applied Mathematics ISSN 973-1768 Volume 14, Number 2 (218), pp 325 33 c Research India Publications http://wwwripublicationcom/gjpamhtm Weak Controllability and the New Choice

More information

Finite difference method for elliptic problems: I

Finite difference method for elliptic problems: I Finite difference method for elliptic problems: I Praveen. C praveen@math.tifrbng.res.in Tata Institute of Fundamental Research Center for Applicable Mathematics Bangalore 560065 http://math.tifrbng.res.in/~praveen

More information

AMS subject classifications. Primary, 65N15, 65N30, 76D07; Secondary, 35B45, 35J50

AMS subject classifications. Primary, 65N15, 65N30, 76D07; Secondary, 35B45, 35J50 A SIMPLE FINITE ELEMENT METHOD FOR THE STOKES EQUATIONS LIN MU AND XIU YE Abstract. The goal of this paper is to introduce a simple finite element method to solve the Stokes equations. This method is in

More information

Chapter 1: The Finite Element Method

Chapter 1: The Finite Element Method Chapter 1: The Finite Element Method Michael Hanke Read: Strang, p 428 436 A Model Problem Mathematical Models, Analysis and Simulation, Part Applications: u = fx), < x < 1 u) = u1) = D) axial deformation

More information

Lecture Note III: Least-Squares Method

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

More information

Invertibility of discrete distributed systems: A state space approach

Invertibility of discrete distributed systems: A state space approach Invertibility of discrete distributed systems: A state space approach J Karrakchou Laboratoire d étude et de Recherche en Mathématiques Appliquées Ecole Mohammadia d Ingénieurs, BP: 765, Rabat-Agdal, Maroc

More information

Solution Sheet 3. Solution Consider. with the metric. We also define a subset. and thus for any x, y X 0

Solution Sheet 3. Solution Consider. with the metric. We also define a subset. and thus for any x, y X 0 Solution Sheet Throughout this sheet denotes a domain of R n with sufficiently smooth boundary. 1. Let 1 p

More information

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

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

More information

Theory of PDE Homework 2

Theory of PDE Homework 2 Theory of PDE Homework 2 Adrienne Sands April 18, 2017 In the following exercises we assume the coefficients of the various PDE are smooth and satisfy the uniform ellipticity condition. R n is always an

More information

An optimal control problem governed by implicit evolution quasi-variational inequalities

An optimal control problem governed by implicit evolution quasi-variational inequalities Annals of the University of Bucharest (mathematical series) 4 (LXII) (213), 157 166 An optimal control problem governed by implicit evolution quasi-variational inequalities Anca Capatina and Claudia Timofte

More information

PART IV Spectral Methods

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

More information

Numerical Methods for the Navier-Stokes equations

Numerical Methods for the Navier-Stokes equations Arnold Reusken Numerical Methods for the Navier-Stokes equations January 6, 212 Chair for Numerical Mathematics RWTH Aachen Contents 1 The Navier-Stokes equations.............................................

More information

Differentiability with respect to initial data for a scalar conservation law

Differentiability with respect to initial data for a scalar conservation law Differentiability with respect to initial data for a scalar conservation law François BOUCHUT François JAMES Abstract We linearize a scalar conservation law around an entropy initial datum. The resulting

More information

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

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

More information

EXISTENCE AND REGULARITY OF SOLUTIONS FOR STOKES SYSTEMS WITH NON-SMOOTH BOUNDARY DATA IN A POLYHEDRON

EXISTENCE AND REGULARITY OF SOLUTIONS FOR STOKES SYSTEMS WITH NON-SMOOTH BOUNDARY DATA IN A POLYHEDRON Electronic Journal of Differential Equations, Vol. 2017 (2017), No. 147, pp. 1 10. ISSN: 1072-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu EXISTENCE AND REGULARITY OF SOLUTIONS FOR

More information

A note on accurate and efficient higher order Galerkin time stepping schemes for the nonstationary Stokes equations

A note on accurate and efficient higher order Galerkin time stepping schemes for the nonstationary Stokes equations A note on accurate and efficient higher order Galerkin time stepping schemes for the nonstationary Stokes equations S. Hussain, F. Schieweck, S. Turek Abstract In this note, we extend our recent work for

More information

Calculus of Variations. Final Examination

Calculus of Variations. Final Examination Université Paris-Saclay M AMS and Optimization January 18th, 018 Calculus of Variations Final Examination Duration : 3h ; all kind of paper documents (notes, books...) are authorized. The total score 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

1 Directional Derivatives and Differentiability

1 Directional Derivatives and Differentiability Wednesday, January 18, 2012 1 Directional Derivatives and Differentiability Let E R N, let f : E R and let x 0 E. Given a direction v R N, let L be the line through x 0 in the direction v, that is, L :=

More information

Algebraic flux correction and its application to convection-dominated flow. Matthias Möller

Algebraic flux correction and its application to convection-dominated flow. Matthias Möller Algebraic flux correction and its application to convection-dominated flow Matthias Möller matthias.moeller@math.uni-dortmund.de Institute of Applied Mathematics (LS III) University of Dortmund, Germany

More information

PDEs in Image Processing, Tutorials

PDEs in Image Processing, Tutorials PDEs in Image Processing, Tutorials Markus Grasmair Vienna, Winter Term 2010 2011 Direct Methods Let X be a topological space and R: X R {+ } some functional. following definitions: The mapping R is lower

More information

On the uniform Poincaré inequality

On the uniform Poincaré inequality On the uniform Poincaré inequality Abdesslam oulkhemair, Abdelkrim Chakib To cite this version: Abdesslam oulkhemair, Abdelkrim Chakib. On the uniform Poincaré inequality. Communications in Partial Differential

More information

b i (x) u + c(x)u = f in Ω,

b i (x) u + c(x)u = f in Ω, SIAM J. NUMER. ANAL. Vol. 39, No. 6, pp. 1938 1953 c 2002 Society for Industrial and Applied Mathematics SUBOPTIMAL AND OPTIMAL CONVERGENCE IN MIXED FINITE ELEMENT METHODS ALAN DEMLOW Abstract. An elliptic

More information

Well-balanced DG scheme for Euler equations with gravity

Well-balanced DG scheme for Euler equations with gravity Well-balanced DG scheme for Euler equations with gravity Praveen Chandrashekar praveen@tifrbng.res.in Center for Applicable Mathematics Tata Institute of Fundamental Research Bangalore 560065 Higher Order

More information

Abstract. 1. Introduction

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

More information

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

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

More information

BLOWUP THEORY FOR THE CRITICAL NONLINEAR SCHRÖDINGER EQUATIONS REVISITED

BLOWUP THEORY FOR THE CRITICAL NONLINEAR SCHRÖDINGER EQUATIONS REVISITED BLOWUP THEORY FOR THE CRITICAL NONLINEAR SCHRÖDINGER EQUATIONS REVISITED TAOUFIK HMIDI AND SAHBI KERAANI Abstract. In this note we prove a refined version of compactness lemma adapted to the blowup analysis

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

PSEUDO-COMPRESSIBILITY METHODS FOR THE UNSTEADY INCOMPRESSIBLE NAVIER-STOKES EQUATIONS

PSEUDO-COMPRESSIBILITY METHODS FOR THE UNSTEADY INCOMPRESSIBLE NAVIER-STOKES EQUATIONS PSEUDO-COMPRESSIBILITY METHODS FOR THE UNSTEADY INCOMPRESSIBLE NAVIER-STOKES EQUATIONS Jie Shen Department of Mathematics, Penn State University University Par, PA 1680, USA Abstract. We present in this

More information

Pseudo-monotonicity and degenerate elliptic operators of second order

Pseudo-monotonicity and degenerate elliptic operators of second order 2002-Fez conference on Partial Differential Equations, Electronic Journal of Differential Equations, Conference 09, 2002, pp 9 24. http://ejde.math.swt.edu or http://ejde.math.unt.edu ftp ejde.math.swt.edu

More information

Greedy control. Martin Lazar University of Dubrovnik. Opatija, th Najman Conference. Joint work with E: Zuazua, UAM, Madrid

Greedy control. Martin Lazar University of Dubrovnik. Opatija, th Najman Conference. Joint work with E: Zuazua, UAM, Madrid Greedy control Martin Lazar University of Dubrovnik Opatija, 2015 4th Najman Conference Joint work with E: Zuazua, UAM, Madrid Outline Parametric dependent systems Reduced basis methods Greedy control

More information

Spectral element discretization of the vorticity, velocity and pressure formulation of the Navier Stokes problem

Spectral element discretization of the vorticity, velocity and pressure formulation of the Navier Stokes problem Spectral element discretization of the vorticity, velocity and pressure formulation of the Navier Stokes problem by K. Amoura, M. Azaïez 2, C. Bernardi 3, N. Chorfi 4 and S. Saadi Abstract: The two-dimensional

More information

Adaptive Finite Element Methods Lecture Notes Winter Term 2017/18. R. Verfürth. Fakultät für Mathematik, Ruhr-Universität Bochum

Adaptive Finite Element Methods Lecture Notes Winter Term 2017/18. R. Verfürth. Fakultät für Mathematik, Ruhr-Universität Bochum Adaptive Finite Element Methods Lecture Notes Winter Term 2017/18 R. Verfürth Fakultät für Mathematik, Ruhr-Universität Bochum Contents Chapter I. Introduction 7 I.1. Motivation 7 I.2. Sobolev and finite

More information

A Bound-Preserving Fourth Order Compact Finite Difference Scheme for Scalar Convection Diffusion Equations

A Bound-Preserving Fourth Order Compact Finite Difference Scheme for Scalar Convection Diffusion Equations A Bound-Preserving Fourth Order Compact Finite Difference Scheme for Scalar Convection Diffusion Equations Hao Li Math Dept, Purdue Univeristy Ocean University of China, December, 2017 Joint work with

More information

Analysis of an Adaptive Finite Element Method for Recovering the Robin Coefficient

Analysis of an Adaptive Finite Element Method for Recovering the Robin Coefficient Analysis of an Adaptive Finite Element Method for Recovering the Robin Coefficient Yifeng Xu 1 Jun Zou 2 Abstract Based on a new a posteriori error estimator, an adaptive finite element method is proposed

More information

10 The Finite Element Method for a Parabolic Problem

10 The Finite Element Method for a Parabolic Problem 1 The Finite Element Method for a Parabolic Problem In this chapter we consider the approximation of solutions of the model heat equation in two space dimensions by means of Galerkin s method, using piecewise

More information

Well-balanced DG scheme for Euler equations with gravity

Well-balanced DG scheme for Euler equations with gravity Well-balanced DG scheme for Euler equations with gravity Praveen Chandrashekar praveen@tifrbng.res.in Center for Applicable Mathematics Tata Institute of Fundamental Research Bangalore 560065 Dept. of

More information

[2] (a) Develop and describe the piecewise linear Galerkin finite element approximation of,

[2] (a) Develop and describe the piecewise linear Galerkin finite element approximation of, 269 C, Vese Practice problems [1] Write the differential equation u + u = f(x, y), (x, y) Ω u = 1 (x, y) Ω 1 n + u = x (x, y) Ω 2, Ω = {(x, y) x 2 + y 2 < 1}, Ω 1 = {(x, y) x 2 + y 2 = 1, x 0}, Ω 2 = {(x,

More information

Numerical Solution of Heat Equation by Spectral Method

Numerical Solution of Heat Equation by Spectral Method Applied Mathematical Sciences, Vol 8, 2014, no 8, 397-404 HIKARI Ltd, wwwm-hikaricom http://dxdoiorg/1012988/ams201439502 Numerical Solution of Heat Equation by Spectral Method Narayan Thapa Department

More information

EXISTENCE OF SOLUTIONS TO THE CAHN-HILLIARD/ALLEN-CAHN EQUATION WITH DEGENERATE MOBILITY

EXISTENCE OF SOLUTIONS TO THE CAHN-HILLIARD/ALLEN-CAHN EQUATION WITH DEGENERATE MOBILITY Electronic Journal of Differential Equations, Vol. 216 216), No. 329, pp. 1 22. ISSN: 172-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu EXISTENCE OF SOLUTIONS TO THE CAHN-HILLIARD/ALLEN-CAHN

More information

A Stable Spectral Difference Method for Triangles

A Stable Spectral Difference Method for Triangles A Stable Spectral Difference Method for Triangles Aravind Balan 1, Georg May 1, and Joachim Schöberl 2 1 AICES Graduate School, RWTH Aachen, Germany 2 Institute for Analysis and Scientific Computing, Vienna

More information

SHARP BOUNDARY TRACE INEQUALITIES. 1. Introduction

SHARP BOUNDARY TRACE INEQUALITIES. 1. Introduction SHARP BOUNDARY TRACE INEQUALITIES GILES AUCHMUTY Abstract. This paper describes sharp inequalities for the trace of Sobolev functions on the boundary of a bounded region R N. The inequalities bound (semi-)norms

More information

CONVERGENCE THEORY. G. ALLAIRE CMAP, Ecole Polytechnique. 1. Maximum principle. 2. Oscillating test function. 3. Two-scale convergence

CONVERGENCE THEORY. G. ALLAIRE CMAP, Ecole Polytechnique. 1. Maximum principle. 2. Oscillating test function. 3. Two-scale convergence 1 CONVERGENCE THEOR G. ALLAIRE CMAP, Ecole Polytechnique 1. Maximum principle 2. Oscillating test function 3. Two-scale convergence 4. Application to homogenization 5. General theory H-convergence) 6.

More information

ITERATIVE METHODS FOR NONLINEAR ELLIPTIC EQUATIONS

ITERATIVE METHODS FOR NONLINEAR ELLIPTIC EQUATIONS ITERATIVE METHODS FOR NONLINEAR ELLIPTIC EQUATIONS LONG CHEN In this chapter we discuss iterative methods for solving the finite element discretization of semi-linear elliptic equations of the form: find

More information

ON THE UNIQUENESS IN THE 3D NAVIER-STOKES EQUATIONS

ON THE UNIQUENESS IN THE 3D NAVIER-STOKES EQUATIONS ON THE UNIQUENESS IN THE 3D NAVIER-STOKES EQUATIONS Abdelhafid Younsi To cite this version: Abdelhafid Younsi. ON THE UNIQUENESS IN THE 3D NAVIER-STOKES EQUATIONS. 4 pages. 212. HAL Id:

More information

Optimal stopping time formulation of adaptive image filtering

Optimal stopping time formulation of adaptive image filtering Optimal stopping time formulation of adaptive image filtering I. Capuzzo Dolcetta, R. Ferretti 19.04.2000 Abstract This paper presents an approach to image filtering based on an optimal stopping time problem

More information

Relaxation methods and finite element schemes for the equations of visco-elastodynamics. Chiara Simeoni

Relaxation methods and finite element schemes for the equations of visco-elastodynamics. Chiara Simeoni Relaxation methods and finite element schemes for the equations of visco-elastodynamics Chiara Simeoni Department of Information Engineering, Computer Science and Mathematics University of L Aquila (Italy)

More information

TRANSPORT IN POROUS MEDIA

TRANSPORT IN POROUS MEDIA 1 TRANSPORT IN POROUS MEDIA G. ALLAIRE CMAP, Ecole Polytechnique 1. Introduction 2. Main result in an unbounded domain 3. Asymptotic expansions with drift 4. Two-scale convergence with drift 5. The case

More information

Analysis of Two-Grid Methods for Nonlinear Parabolic Equations by Expanded Mixed Finite Element Methods

Analysis of Two-Grid Methods for Nonlinear Parabolic Equations by Expanded Mixed Finite Element Methods Advances in Applied athematics and echanics Adv. Appl. ath. ech., Vol. 1, No. 6, pp. 830-844 DOI: 10.408/aamm.09-m09S09 December 009 Analysis of Two-Grid ethods for Nonlinear Parabolic Equations by Expanded

More information

University of Houston, Department of Mathematics Numerical Analysis, Fall 2005

University of Houston, Department of Mathematics Numerical Analysis, Fall 2005 3 Numerical Solution of Nonlinear Equations and Systems 3.1 Fixed point iteration Reamrk 3.1 Problem Given a function F : lr n lr n, compute x lr n such that ( ) F(x ) = 0. In this chapter, we consider

More information

Finite Volume Discretization of a Class of Parabolic Equations with Discontinuous and Oscillating Coefficients

Finite Volume Discretization of a Class of Parabolic Equations with Discontinuous and Oscillating Coefficients Applied Mathematical Sciences, Vol, 8, no 3, - 35 HIKARI Ltd, wwwm-hikaricom https://doiorg/988/ams876 Finite Volume Discretization of a Class of Parabolic Equations with Discontinuous and Oscillating

More information

COMBINING MAXIMAL REGULARITY AND ENERGY ESTIMATES FOR TIME DISCRETIZATIONS OF QUASILINEAR PARABOLIC EQUATIONS

COMBINING MAXIMAL REGULARITY AND ENERGY ESTIMATES FOR TIME DISCRETIZATIONS OF QUASILINEAR PARABOLIC EQUATIONS COMBINING MAXIMAL REGULARITY AND ENERGY ESTIMATES FOR TIME DISCRETIZATIONS OF QUASILINEAR PARABOLIC EQUATIONS GEORGIOS AKRIVIS, BUYANG LI, AND CHRISTIAN LUBICH Abstract. We analyze fully implicit and linearly

More information

The mortar element method for quasilinear elliptic boundary value problems

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

More information

A Least-Squares Finite Element Approximation for the Compressible Stokes Equations

A Least-Squares Finite Element Approximation for the Compressible Stokes Equations A Least-Squares Finite Element Approximation for the Compressible Stokes Equations Zhiqiang Cai, 1 Xiu Ye 1 Department of Mathematics, Purdue University, 1395 Mathematical Science Building, West Lafayette,

More information

Weak Convergence Methods for Energy Minimization

Weak Convergence Methods for Energy Minimization Weak Convergence Methods for Energy Minimization Bo Li Department of Mathematics University of California, San Diego E-mail: bli@math.ucsd.edu June 3, 2007 Introduction This compact set of notes present

More information

A posteriori analysis of a discontinuous Galerkin scheme for a diffuse interface model

A posteriori analysis of a discontinuous Galerkin scheme for a diffuse interface model A posteriori analysis of a discontinuous Galerkin scheme for a diffuse interface model Jan Giesselmann joint work with Ch. Makridakis (Univ. of Sussex), T. Pryer (Univ. of Reading) 9th DFG-CNRS WORKSHOP

More information

Local null controllability of the N-dimensional Navier-Stokes system with N-1 scalar controls in an arbitrary control domain

Local null controllability of the N-dimensional Navier-Stokes system with N-1 scalar controls in an arbitrary control domain Local null controllability of the N-dimensional Navier-Stokes system with N-1 scalar controls in an arbitrary control domain Nicolás Carreño Université Pierre et Marie Curie-Paris 6 UMR 7598 Laboratoire

More information

On Friedrichs inequality, Helmholtz decomposition, vector potentials, and the div-curl lemma. Ben Schweizer 1

On Friedrichs inequality, Helmholtz decomposition, vector potentials, and the div-curl lemma. Ben Schweizer 1 On Friedrichs inequality, Helmholtz decomposition, vector potentials, and the div-curl lemma Ben Schweizer 1 January 16, 2017 Abstract: We study connections between four different types of results that

More information

EXISTENCE AND REGULARITY RESULTS FOR SOME NONLINEAR PARABOLIC EQUATIONS

EXISTENCE AND REGULARITY RESULTS FOR SOME NONLINEAR PARABOLIC EQUATIONS EXISTECE AD REGULARITY RESULTS FOR SOME OLIEAR PARABOLIC EUATIOS Lucio BOCCARDO 1 Andrea DALL AGLIO 2 Thierry GALLOUËT3 Luigi ORSIA 1 Abstract We prove summability results for the solutions of nonlinear

More information

Solving PDEs with freefem++

Solving PDEs with freefem++ Solving PDEs with freefem++ Tutorials at Basque Center BCA Olivier Pironneau 1 with Frederic Hecht, LJLL-University of Paris VI 1 March 13, 2011 Do not forget That everything about freefem++ is at www.freefem.org

More information

Numerical Analysis Comprehensive Exam Questions

Numerical Analysis Comprehensive Exam Questions Numerical Analysis Comprehensive Exam Questions 1. Let f(x) = (x α) m g(x) where m is an integer and g(x) C (R), g(α). Write down the Newton s method for finding the root α of f(x), and study the order

More information

Title of your thesis here

Title of your thesis here Title of your thesis here Your name here A Thesis presented for the degree of Doctor of Philosophy Your Research Group Here Department of Mathematical Sciences University of Durham England Month and Year

More information

Time-dependent Dirichlet Boundary Conditions in Finite Element Discretizations

Time-dependent Dirichlet Boundary Conditions in Finite Element Discretizations Time-dependent Dirichlet Boundary Conditions in Finite Element Discretizations Peter Benner and Jan Heiland November 5, 2015 Seminar Talk at Uni Konstanz Introduction Motivation A controlled physical processes

More information