L ubomír Baňas 1 and Robert Nürnberg Introduction A POSTERIORI ESTIMATES FOR THE CAHN HILLIARD EQUATION WITH OBSTACLE FREE ENERGY

Size: px
Start display at page:

Download "L ubomír Baňas 1 and Robert Nürnberg Introduction A POSTERIORI ESTIMATES FOR THE CAHN HILLIARD EQUATION WITH OBSTACLE FREE ENERGY"

Transcription

1 ESAIM: MAN 43 (009) DOI: /man/ ESAIM: Mathmatical Modlling and Numrical Analysis A POSERIORI ESIMAES FOR HE CAHN HILLIARD EQUAION WIH OBSACLE FREE ENERGY L ubomír Baňas 1 and Robrt Nürnbrg Abstract. W driv a postriori stimats for a discrtization in spac of th standard Cahn Hilliard quation with a doubl obstacl fr nrgy. h drivd stimats ar robust and fficint, and in practic ar combind with a huristic tim stp adaptation. W prsnt numrical xprimnts in two and thr spac dimnsions and compar our mthod with an xisting huristic spatial msh adaptation algorithm. Mathmatics Subjct Classification. 65M60, 65M15, 65M50, 35K55. Rcivd January 14, 008. Rvisd Novmbr 8, 008. Publishd onlin Jun 1, Introduction In this papr w driv spatial a postriori rrorstimats for a pic-wis linar finit lmnt approximation of th following Cahn Hilliard quation: γ u = Δw in Ω := Ω [0,], t w = γδu + 1 γ Ψ (u) in Ω, u ν = w ν =0 on Ω (0,], u(, 0) = u 0 in Ω, (1.1) whr Ω is a convx polyhdral domain in R d, d =, 3, and >0 is a fixd positiv tim. Morovr, Ψ is a givn nrgy potntial, and in this papr w will tak Ψ to b th so calld doubl obstacl potntial Ψ(s) := { 1 (1 s ) if s [ 1, 1], if s/ [ 1, 1]. (1.) Kywords and phrass. Cahn Hilliard quation, obstacl fr nrgy, linar finit lmnts, a postriori stimats, adaptiv numrical mthods. Supportd by th EPSRC grant EP/C548973/1. 1 Dpartmnt of Mathmatics and th Maxwll Institut for Mathmatical Scincs, Hriot-Watt Univrsity, Edinburgh, EH14 4AS, UK. L.Banas@hw.ac.uk Dpartmnt of Mathmatics, Imprial Collg London, London, SW7 AZ, UK. Articl publishd by EDP Scincs c EDP Scincs, SMAI 009

2 1004 L. BAŇAS AND R. NÜRNBERG W not that othr choics of Ψ ar also possibl, s.g. (1.4) blow. In addition, th paramtr γ>0isan intraction lngth, which is small compard to th dimnsions of Ω. Equation (1.1) was originally introducd by Cahn and Hilliard to modl spinodal dcomposition and coarsning phnomna in binary alloys, s [11,1]. Hr u is dfind to b th diffrnc of th local concntrations of th two componnts of an alloy and hnc u is rstrictd to li in th intrval [ 1, 1]. Mor rcntly, th Cahn Hilliard quation has bn usd.g. as a phas fild modl for sharp intrfac volutions and to study phas transitions and intrfac dynamics in multiphas fluids, s.g. [7,,3] and th rfrncs thrin. W not that in (1.1) w hav usd a tim scaling, so that in th limit γ 0, w rcovr th wll known sharp intrfac motions by Mullins Skrka. W rcall that this limit was first formally shown in [6], and latr provd rigorously in [1]. W not that as proprtis of commrcially producd matrials dpnd on microstructurs which ar gnratd using spcial procssing tchniqus, such as phas sparationand coarsningmchanisms, accurat prdictions of microstructur or th volution of pattrn formation during phas sparation and coarsning ar of considrabl intrst in matrials scinc. As it is difficult to obtain such information by ral-lif xprimnts, rliabl numrical computations ar vry important. It is th aim of this papr to prov suitabl a postriori stimats for th discrt approximation of th considrd problm that can b usd to construct robust and rliabl msh rfinmnt algorithms in two and thr spac dimnsions, which allow for fficint and rliabl numrical simulations. h thory of Cahn and Hilliard is basd on th following Ginzburg Landau fr nrgy ( E(u) := γ u + γ 1 Ψ(u) ) dx. (1.3) Ω h first trm in th fr nrgy pnalizs larg gradints and th scond trm is th homognous fr nrgy. hn (1.1) can b drivd from mass balanc considrations as a gradint flow for th fr nrgy E(u), with th chmical potntial w := δe δu bing th variational drivativ of th nrgy E with rspct to u. For notational convninc in (1.1) it was implicitly assumd that th fr nrgy Ψ is diffrntiabl. An xampl for such a potntial function is Ψ(s) = 1 4 (s 1), (1.4) which has th advantag of bing smooth but th disadvantag that physically non-admissibl valus with u > 1 can b attaind during th volution. Of cours, th obstacl fr nrgy (1.) forcs u to stay within th intrval [ 1, 1] of physically maningful valus. his is a clar advantag ovr a formulation involving (1.4). Hnc, in this papr w will from now on considr th obstacl fr nrgy (1.). hn th chmical potntial w nds to b computd with th hlp of a variational inquality, s (.1) blow. It is this variational inquality which rquirs spcial attntion in dvloping an a postriori rror stimat. ypical volutions of (1.1) starting from a wll mixd initial stat bgin with a rlativly short arly phas, calld spinodal dcomposition, in which th local concntrations u grow towards th minimizrs ± 1of(1.). his lads to a stup, whr larg parts of th domain ar occupid by rgions whr u = ± 1, which ar sparatd by intrfacial rgions whr u < 1, in which u smoothly varis from 1 to1. hnfollowsa much slowr volution phas, in which th total volum of ths intrfacial rgions is dcrasd. his phas is calld coarsning. h thicknss of th intrfacial rgions, i.., th rgion whr u < 1, is asymptotically of ordr O(γ). As mntiond arlir, it can b shown that in th sharp intrfac limit (i.., whnγ 0) th long tim dynamics of quations (1.1) corrspond to th Mullins Skrka quation. Finit lmnt mthods for quation (1.1) with(1.) hav bn proposd and analyzd in [9], s also [5,6]. In addition, xistnc and uniqunss of th solution u, w to (1.1), as wll as rgularity rsults, wr shown in [8]. In [7] a finit lmnt approximation for a rlatd, so calld dgnrat, Cahn Hilliard quation was considrd, and in addition a huristic adaptiv msh rfinmnt algorithm was usd for numrical simulations in two spac dimnsions, in ordr to incras th fficincy of th computations. his approximation and th corrsponding msh rfinmnt hav rcntly bn xtndd to thr spac dimnsions in [3], s also [4]. hr xist numrous works on finit lmnt approximations of (1.1) with smooth potntials such as (1.4).

3 A POSERIORI ESIMAES FOR HE CAHN HILLIARD EQUAION 1005 Hr w rfr to.g. [16 18] and th rfrncs thrin. A postriori stimats for th Cahn Hilliard quation with th smooth potntial (1.4) hav vry rcntly bn obtaind in [19], whr th stimats for a continuous in tim smi-discrt approximation only dpnd on polynomial powrs of γ 1, a rsult which crucially dpnds on th spctral stimat from [13]. o our knowldg, so far thr is no work on a postriori stimats for th Cahn Hilliard quation with th obstacl potntial (1.4), apart from th numrical rsults in [], which ar basd on rsults rlatd to th work in this papr. It is th aim of this papr to prov a postriori stimats and xamin adaptiv finit lmnt mthods for (1.1) in two and thr spac dimnsions. Sinc thr is no spctral stimat corrsponding to that from [13] availabl for th non-smooth modl, w only xamin th rror du to th spatial discrtization. hrfor w rstrict our analysis to spatial a postriori rror stimats for a discrt in tim analogu of (1.1). In particular, w will driv stimats for a coupld systm that consists of an lliptic variational inquality involving two constant obstacls, and a linar lliptic problm; s (.1) blow. By using th idas of [7], whr rror stimats for linar finit lmnt approximations of lliptic obstacl problms ar introducd, w ar abl to obtain an stimat with localizd intrior rsidual, which nabls ffctiv and rliabl rror control by rfinmnt that is mainly concntratd in th intrfacial rgion, whr u < 1. h a postriori analysis of lliptic obstacl problms is a rlativly nw fild. A rsidual a postriori stimat with non-localizd intrior rsidual was obtaind in [14]. A sharpr stimat with localizd intrior rsidual was constructd in [8] for constant obstacls and in [7] for gnral obstacls. A short rviw on a postriori stimats for lliptic obstacl problms is givn in [10]. A postriori stimats for parabolic variational inqualitis wr drivd in [4] by xtnding th idas of [7]. W also rfr to work in optimal control thory, whr vry rcntly an rror stimator for a control problm with sid constraints involving PDEs and inquality constraints has bn introducd in [0,1]. Howvr, w strss that a crucial diffrnc btwn work on optimal control thory and work on obstacl problms involving variational inqualitis is that th formr only applis th inquality constraints on th right hand sid of th control PDE, and that th localization of th intrior rsidual is not ssntial to obtain a lowr bound for th rror,.g. s [1]. h papr is organizd as follows. In Sction, w introduc th continuous in spac and discrt in tim Cahn Hilliard quation and its finit lmnt approximation by conforming pic-wis linar lmnts. In Sction 3, w stablish an a postriori stimat with non-localizd rsidual, which can potntially lad to xtnsiv msh rfinmnt outsid of th intrfacial rgion, i.. in th rgion whr th solution u is constant. In Sction 4, w construct uppr and lowr bounds for th rror with localizd intrior rsidual. In Sction 5, w discuss a numbr of adaptiv algorithms for numrical computations. Finally, Sction 6 is dvotd to numrical xprimnts, whr w xamin th prformanc of th adaptiv algorithms in two and thr spac dimnsions.. Finit lmnt approximation W considr th following continuous in spac smi-discrt countrpart of th Cahn Hilliard quation obtaind by a backward-eulr tim discrtization of (1.1): Find u K:= {v H 1 (Ω) : v 1} and w H 1 (Ω) such that (u, φ)+ τ γ ( w, φ) = (f,φ) φ H1 (Ω), (.1) γ ( u, (ψ u)) (w, ψ u) (g, ψ u) ψ K, whr (φ, ψ) = Ω φψ is th L -innr product ovr Ω. hroughout this papr, w dnot th L -norm ovr D Ωby D, and similarly us 1,D for th H 1 -norm. For notational convninc, w drop th subscript in th cas D = Ω. In addition, w dnot th norm in th dual spac (H 1 (Ω)) by 1 and us, for th duality pairing btwn H 1 (Ω) and its dual. On introducing th linar finit lmnt spac V h := {φ C(Ω) : φ is linar h } H 1 (Ω),

4 1006 L. BAŇAS AND R. NÜRNBERG whr Ω= h, w considr th following finit lmnt approximation of (.1): Find u h K h and w h V h such that (u h,φ)+ τ γ ( w h, φ) = (f,φ) φ V h, (.) γ ( u h, (ψ u h )) (w h,ψ u h ) (g, ψ u h ) ψ K h, whr K h := K V h. Not that in viw of th drivation of (.1), w usually hav f = u old h, g = 1 γ uold h in (.), whr u old h is th solution from th prvious tim stp. hn (.) corrsponds to on tim stp of th ar picwis linar functions, whr Vold h is th finit lmnt spac corrsponding to th prvious tim stp. his cas will simplify som stps in th analysis blow, in particular whn Vold h V h. W dnot by u = u u h, (.3) w = w w h. unconditionally stabl, fully discrt approximation in [9]. Also, in that cas f, g V h old W rcall th following wll-known rsult concrning V h : (φ, χ) (φ, χ) h C h φ χ φ, χ V h ; (.4) whr (φ, χ) h = Ω Ih (φχ)forφ, χ C(Ω) is th usual mass lumpd innr product, and I h is th usual Lagrang intrpolation oprator onto V h. In addition to th triangulation h, w introduc th st of its nods P h and dgs E h. W dnot th nodal basis functions of V h as (χ p h ) p P h,whrχ p h (q) =1ifp = q and χp h (q) = 0 othrwis. Morovr, for ach h and E h w dnot thir diamtr by h and h, rspctivly. W also introduc th local msh siz function h :Ω R, which is picwis constant and such that h = h for all h. For any st D Ω, w dfin th discrt nighbourhood of D by D = { h ; D }. In addition, in a slight abus of notation, w also introduc th short hand notation h α [ u h ] := E h h α [ u h ],whrα R. 3. A POSERIORI stimat with positivity prsrving intrpolation In this sction w xtnd th idas of [14], in ordr to show how it is possibl to driv an uppr bound for th rror of th finit lmnt approximation in a rlativly simpl mannr. h obtaind stimat, howvr, dos not tak into account crtain spcial proprtis of th solution, and may lad to xcssiv msh rfinmnt in practic, in aras whr th solution u is constant. W rcall th dfinition of th positivity prsrving intrpolation oprator Π h 0 : L 1 (Ω) V h H0 1 (Ω) from [14], i.., whavthatu 0 Π h 0 u 0 for all u L1 (Ω). It is thn a straightforward mattr to xtnd this dfinition to th Numann boundary condition and doubl obstacl prsnt hr, to obtain an analogous oprator Π h : L 1 (Ω) V h such that u K Π h u K h. (3.1) In fact, w can choos Π h to b th oprator givn in [5], Exampl 1.1. W hav th following approximation proprtis of Π h for u H 1 (Ω) and u h V h : Π h u u, (3.a) u Π h u Ch u, (3.b) u Π h u Ch 1/ u ẽ, (3.c) u h Π h u h C h 3/ [ u h ], (3.d) u h Π h u h C h [ u h ] ẽ, (3.) cf. [14], whr w rcall that is th union of all th lmnts surrounding, and similarly for ẽ.

5 A POSERIORI ESIMAES FOR HE CAHN HILLIARD EQUAION 1007 Choosing φ = w in (.1) andφ =Π h w in (.), w obtain that ( u, w )+ τ γ w =(f, w Π h w ) (u h, w Π h w ) τ ( wh, ( w Π h w ) ). (3.3) γ Nxt, w tak ψ = u h in (.1), lading to and ψ =Π h u K h, rcall (3.1), in (.), which givs γ ( u, u ) (w, u ) (g, u ) 0, (3.4) γ ( u h, u ) (w h, u ) (g, u ) ( g, Π h u u ) γ ( u h, (Π h u u) ) + ( w h, Π h u u ). (3.5) W hav th simpl idntity Π h u u =(Π h u u )+(Π h u h u h ). hrfor, aftr w subtract (3.5) from(3.4), w obtain γ u ( w, u ) ( g, u Π h ) ( u γ uh, ( u Π h u ) ) + ( w h, u Π h ) u + ( g + w h, Π h ) ( u h u h γ uh, (Π h u h u h ) ) (3.6). Furthr, aftr mploying intgration by parts, sinc Δu h =0,wobsrvthat ( u h, ψ) = u h ψ = { } u h ν ψ Δu h ψ h h = [ u h ] ψ ψ H 1 (Ω), E h (3.7) whr ν dnots th outward unit vctor to h. Hnc it follows from (3.3), on applying th Cauchy Schwartz and Young inqualitis togthr with (3.b) and (3.7), that ( u, w )+ τ γ w C γ τ h(u h f) + τ 4γ w + C τ γ h1/ [ w h ] + τ 4γ w. (3.8) Similarly, it follows from (3.6) that γ u ( w, u ) C γ h(g + w h) + γ 4 u + γc h 1/ [ u h ] + γ 4 u + ( g + w h, Π h u h u h ) γ ( uh, (Π h u h u h ) ). (3.9) h last two trms on th right-had sid of (3.9) can b stimatd, on noting (3.7) and(3.d)-(3.), as ( g + wh, Π h ) ( u h u h γ uh, (Π h u h u h ) ) C (γ h 1/ [ u h ] + 1γ ) h(g + w h). By combining th prvious quation with (3.8), (3.9) w arriv at ( ) τ γ w + γ u γ C τ h(f u h) + τ γ h1/ [ w h ] + 1 γ h(g + w h) + γ h 1/ [ u h ]. (3.10) Upon subsquntly rscaling w obtain [ ] γ τ w + γ u C τ h(u h f) + τ h 1/ [ w h ] + h(g + w h ) + γ h 1/ [ u h ]. (3.11)

6 1008 L. BAŇAS AND R. NÜRNBERG Rmark 3.1. h disadvantag of th abov stimat is that th intrior rsidual h(g + w h ) corrsponding to th variational inquality in (.) is not localizd to th noncontact st (s dfinition in th nxt sction), which can caus xcssiv msh rfinmnt in th contact st, whr th solution u h is constant and whr w h usually attains larg valus. Howvr, as th variational inquality in (.) trivially holds in th contact st, idally thr should b no contribution from th intrior rsidual to th a postriori rror stimat. his problm will b addrssd in th nxt sction. 4. A POSERIORI stimat with localizd intrior rsidual In this sction w driv an a postriori stimat with an intrior rsidual localizd to th intrfac, i.. th intrior rsidual inducd by th variational inquality in (.) is zro in th rgion whr u h =1. his rsult givs ris to mor fficint a postriori rror basd msh rfinmnt stratgis, and it is furthrmor a thortical justification for th construction of huristical msh adaptiv algorithms, whr th msh rfinmnt is concntratd in th intrfacial ara, i.. whr u h < 1. W xtnd th idas of [7,8] to th smi-discrt formulation (.1) of th Cahn Hilliard quation. Hr w dfin th discrt functions f h := I h f, g h := I h g and not that by dfinition w hav (g h,φ) h =(g, φ) h,(f h,φ) h =(f,φ) h for all φ C(Ω). Instad of th discrt formulation (.), w considr th following discrt problm: Find u h K h and w h V h such that (u h,φ) h + τ γ ( w h, φ) = (f h,φ) h φ V h, γ ( u h, (ψ u h )) (w h,ψ u h ) h (g h,ψ u) h ψ K h. (4.1) h abov formulation only diffrs from (.) in th zro ordr trms, whr w us th rducd discrt innr product (, ) h. Givn th tru solution u, and following th tchniqu in [7] for a singl obstacl, w obtain th partition of th domain Ω=C(u) N(u) F(u), (4.) whr th contact st C(u) is th maximal opn st A Ω such that u 1onA; th noncontact st N (u) := ɛ>0 B ɛ ;whrb ɛ is th maximal opn st B Ω such that u < 1 ɛ; th fr boundary F(u) isthstω\ (C(u) N(u)). h contact st can b furthr dcomposd as C(u) =C + (u) C (u), whr u = ± 1onC ± (u). W dfin th continuous rsidual σ(u) (H 1 (Ω)) as σ(u),ψ =(g, ψ)+(w, ψ) γ( u, ψ) ψ H 1 (Ω). (4.3) h following proprtis can b obtaind from th dfinition of σ (not, u 1 in C(u)) h discrt rsidual is dfind as σ h V h such that σ 0 in C + (u), (4.4a) σ 0 in C (u), (4.4b) σ = g + w in C(u), (4.4c) σ =0 in N (u). (4.4d) (σ h,ψ) h =(g h,ψ) h +(w h,ψ) h γ( u h, ψ) ψ V h. (4.5)

7 A POSERIORI ESIMAES FOR HE CAHN HILLIARD EQUAION 1009 Altrnativly w can writ (σ h,ψ) h =(g h,ψ) h +(w h,ψ) h + γ(δ h u h,ψ) h, (4.6) whr Δ h : V h V h is th usual discrt Laplacian on V h. W dfin th jump across an innr lmnt dg/fac = 1 E h as [ u h ] = 1 ( u h 1 u h ) ν, whr ν is a unit normal vctor of pointing from 1 to. For a Numann boundary dg E h Ω w dfin [ u h ] = u h ν, whr ν is h outward unit vctor to th boundary Ω. Similarly to (4.), th domain Ω can b dcomposd into Ω=C h (u h ) F h (u h ) N h (u h ), (4.7) whr C h (u h ) := C h (u h ) + C h (u h ), C h (u h ) ± := { h ; u h = ± 1on }, N h (u h ) := { h ; u h < 1on }, F h (u h ) := Ω\ [C h (u h ) N h (u h )]. In our contxt, C h dnots th subdomains with pur matrials, N h dnots th diffus intrfac and F h is th so-calld discrt fr boundary btwn C h and N h. Similarly as in (4.4), on can stablish for all nods p P h that σ h (p) 0 if p C + h, (4.8a) σ h (p) 0 if p C h, (4.8b) σ h (p) =g h (p)+w h (p) if u h = 1 on supp χ p h, (4.8c) σ h (p) =0 if u h (p) < 1. (4.8d) Not that Δ h u h =0inC h (u h ). Following [7], w dfin th Galrkin functional G h (H 1 (Ω)) as G h,ψ = γ( (u h u), ψ) (w h,ψ)+(w, ψ)+(σ h σ, ψ) ψ H 1 (Ω). (4.9) W dirctly hav from (4.3) that G h,ψ = γ( u h, ψ) (w h + g, ψ)+(σ h,ψ) ψ H 1 (Ω). (4.10) Lmma 4.1 (prturbd Galrkin orthogonality). hr xists a constant C dpnding only on th msh rgularity, such that G h 1,h := sup G h,ψ h = C ( γ h ) Δ h u h + g h g 1,h. ψ h V h, ψ h =1 Proof. On rcalling th dfinitions of (, ) h, σ h and G h,whavforanyψ h V h that G h,ψ h = (g, ψ h ) h +[γ( u h, ψ h ) (g + w h,ψ h ) h ]+(w h,ψ h ) h (w h + g, ψ h )+(σ h,ψ h ) = (g, ψ h ) h (g, ψ h )+(w h,ψ h ) h (w h,ψ h )+(σ h,ψ h ) (σ h,ψ h ) h = (g + w h σ h,ψ h ) h (g h + w h σ h,ψ h )+(g h g, ψ h ) = γ(δ h u h,ψ h ) h + γ(δ h u h,ψ h )+(g h g, ψ h ). (4.11)

8 1010 L. BAŇAS AND R. NÜRNBERG Furthrmor, it follows from (.4) that γ(δ h u h,ψ h ) h + γ(δ h u h,ψ h ) γ h Δ h u h ψ h, which yilds th dsird rsult. Also th following is just a gnralisation of [7], Lmma 3.4, xcpt that th trm σ h σ 1 appar on th lft hand sid of (4.1). Lmma 4.. h following inquality holds dos not Proof. It follows from (4.9) that γ (u h u) (w h w, u h u) C 1 γ G h 1 C (σ h σ, u h u). (4.1) γ (u h u) (w h w, u h u) = G h,u h u (σ h σ, u h u) G h 1 (u h u) (σ h σ, u h u). Hnc by Young s inquality w gt γ (u h u) (w h w, u h u) 1 γ G h 1 + γ (u h u) (σ h σ, u h u). (4.13) h assrtion of th lmma thn asily follows from th last inquality Global uppr bound In th following lmma w stimat th Galrkin functional. Lmma 4.3. hr xists a constant C dpnding only on th msh rgularity, such that ( ) 1/ G h 1 C γ h 1/ [ u h ] + h(g + w h σ h ) + γ h Δ h u h + g gh 1,h. E h Proof. For ϕ H 1 (Ω) w writ G h,ϕ = G h,ϕ I h ϕ + G h,i h ϕ, whr I h ϕ dnots th Clémnt intrpolant for ϕ, s[15]. h scond trm in th abov quation can b stimatd using Lmma 4.1 (th prturbd Galrkin orthogonality) and th proprtis of I h as G h,i h ϕ C G h 1,h I h ϕ C(γ h Δ h u h + g g h 1,h ) ϕ. Similarly, on rcalling (4.10), w can stimat th first trm using standard argumnts of a postriori stimation as G h,ϕ I h ϕ = γ( u h, (ϕ I h ϕ)) (w h + g σ h,ϕ I h ϕ) = γ [ u h ] (ϕ I h ϕ) (w h + g σ h )(ϕ I h ϕ) E h ( ) h 1/ C γ h 1/ [ u h ] + h(g + w h σ h ) ϕ, E h L () which concluds th proof. h following lmma is an adaption of [7], Proposition 3.7.

9 A POSERIORI ESIMAES FOR HE CAHN HILLIARD EQUAION 1011 Lmma 4.4. h following inquality holds for th solutions u and u h of (.1) and (4.1), rspctivly. whr (σ h σ, u h u) C γ h 4 Δ h u h + 1 h w h + g h + γ h [ u h ], γ h h E h h = { h ; p 1,p P h, u h (p 1 ) =1and u h (p ) < 1}, E h = { E h ; P h } with P h = {p P h; u h (p) =1and u h 1 on supp χ p h } Proof. W rwrit (σ h σ, u h u) =(σ h,u h u)+(σ, u u h ). Sinc u h K, w can stimat th scond trm using (.1) (σ, u u h )=γ( u, (u h u)) (g, u h u) (w, u h u) 0. Nxt, on noting that Ω = C h N h h, w rwrit th first trm on th right-hand sid as (σ h,u h u) = σ h ( 1 u)+ σ h (1 u)+ σ h (u h u)+ σ h (u h u). C h C + h N h h Using (4.8a)-(4.8b) w gt Rcalling (4.8d) w hav C h σ h ( 1 u) 0, C + h σ h (1 u) 0. σ h (u h u) =0. N h h rmaining trm is stimatd as follows. Considr h u h (p ) < 1andσ h 0wgt and p 1,p P h, with u h (p 1 ) = ± 1, σ h (u h u) = σ h (u h 1) + σ h (1 u) σ h (u h 1) σ h u h 1, if σ h 0wgt σ h (u h u) = From [7], Lmma 3.6, w obtain (E h (p) :={ E h ; p }) u h 1 Ch E h (p 1) σ h (u h +1)+ σ h ( 1 u) σ h (u h +1) σ h u h +1. h [u h 1] 1/ Ch E h (p 1) h [u h ] 1/. W hav from σ h = σ h w h g h + w h + g h σ h w h g h + w h + g h,

10 101 L. BAŇAS AND R. NÜRNBERG that σ h σ h w h g h + w h + g h γh Δ h u h + w h + g h. Finally w gt σ h (u h u) C γh 4 Δ hu h + 1 γ h w h + g h + γ h [u h ], E h (p 1) which, on noting that E h = p P h E h(p), concluds th proof. h following lmma is a simpl consqunc of (4.1) and Lmmas 4.3 and 4.4. Lmma 4.5. [ γ (u h u) (w h w, u h u) C 1 γ h 1/ [ u h ] γ + h(g + w h σ h ) E h + γ h Δ h u h + g gh 1,h + ] h (w h + g h ). h nxt lmma givs an stimat for th first quation in (4.1). Lmma 4.6. [ τ γ (w h w) +(u h u, w h w) C γ τ τ γ h 1/ [ w h ] + h(u h f) E h + h (u h f h ) ] + f f h 1,h. h Proof. W start with th idntity τ γ ( w, φ)+( u,φ)= τ γ ( w, (φ I h φ)) + ( u,φ I h φ)+ τ γ ( w, I h φ)+( u,i h φ) for any φ H 1 (Ω). Nxt, according to (.1), (4.1), w can rwrit th abov quation as τ γ ( w, φ)+( u,φ)= τ γ ( w h, (φ I h φ)) + (f u h,φ I h φ)+ τ γ ( w, I h φ)+( u,i h φ). (4.14) Similarly as in Lmma 4.1, onnoting(.1), (4.1) and(.4), w obtain that τ γ ( w, I h φ)+( u,i h φ) h (u h f h ) φ + f f h 1,h φ, (4.15) and similarly to Lmma 4.3, whavthat τ γ ( w h, (φ I h φ)) + (f u h, (φ I h φ)) C [ ( ) τ 1/ ] h 1/ [ w h ] + h(u h f) φ. (4.16) γ E h h proof can b concludd by combining (4.15), (4.16) and(4.14), and by subsquntly applying a Young s inquality for φ = w.

11 A POSERIORI ESIMAES FOR HE CAHN HILLIARD EQUAION 1013 h following corollary is a simpl consqunc of Lmmas 4.5 and 4.6. Corollary 4.1. h following stimat is valid for u h, w h : γ (u h u) + τ (w h w) [ C τ h 1/ [ w h ] + γ τ h(u h f) + γ τ E h h (u h f) + γ h 1/ [ u h ] + h(g + w h σ h ) + h (w h + g h ) E h h + γ ] h Δ h u h + g gh 1,h + γ τ f f h 1,h. (4.17) Rmark 4.1. h stimat (4.17) diffrsfrom(3.11) in th following: th intrior rsidual, which is now h(g + w h σ h ) + h (w h +g h ), is localizd to th discrt noncontact st Ω \C h, on rcalling (4.8c); for simplicity, w did not considr coarsning in th drivation of (3.11), i.. th trms g g h 1,h, γ τ f f h 1,h ar not includd in (3.11); th trms γ τ h (u h f), γ h Δ h u h in (4.17) ar du to th us of th discrt innr product (, ) h and ar thrfor not prsnt in (3.11). Finally, w not that th quantity h Δ h u h is 0 within th discrt contact st, cf. [7], Rmark 3.7, and so it will not contribut to th a postriori rror stimat in that rgion. 4.. Local lowr bounds h o ach function f L (Ω) w assign a picwis constant function f dfind as f = 1 f h. Furthr, th so-calld local data oscillation is dfind as osc h (f,) = h (f f). Lmma 4.7. h following local stimat holds for all h [ γ h 1/ [ u h ] + h (g + w h σ h ) + γ h Δ h u h ] 1/ { C γ (u h u) + h (g g h ) + h (σ h σ) + h (w h w) +osc h (g + w h σ h,) Proof. h proof is basd on th local argumnt of Vrfürth [9]. With vry h, E h w rspctivly associat th standard canonical bubbl functions ψ, ψ. For tchnical rasons, w introduc th auxiliary function z h := γu h. hn, following a similar argumnt in [14], }

12 1014 L. BAŇAS AND R. NÜRNBERG for any h, w can construct a function φ := α ψ + β ψ,whrα, β ar chosn such that ([ z h ],φ ) = h [ z h ], (g + w h σ h,φ ) = h g + w h σ h, and [ ] 1/ φ C h 1/ [ z h ] + h (g + w h σ h ), [ ] 1/ φ Ch h 1/ [ z h ] + h (g + w h σ h ). W hav, on rcalling (3.7) and(4.3), that γ h 1/ [ u h ] + h (g + w h σ h ) = = ([ z h ],φ ) +(g + w h σ h,φ ) h 1/ [ z h ] + h (g + w h σ h ) = γ ([ u h ],φ ) +(g + w h σ h,φ ) = γ( u h, φ ) +(g + w h σ h,φ ) +(g + w h σ h (g + w h σ h ),φ ) = γ( (u u h ), φ ) (w w h,φ ) +(σ σ h,φ ) +(g + w h σ h (g + w h σ h ),φ ) γ (u u h ) φ + w w h φ + σ σ h φ + g + w h σ h (g + w h σ h ) φ [ ] C γ (u u h ) + h w w h + h σ σ h +osc h (g + w h σ h,) [ γ ] 1/ h 1/ [ u h ] + h (g + w h σ h ). (4.18) Nxt, w hav from(4.6) and an invrs inquality that γ h Δ h u h = h (g h + w h σ h ) Ch g h + w h σ h C [h g + w h σ h + h g g h +osc h (g + w h σ h,)]. (4.19) Finally, th assrtion of th lmma follows on combining (4.18) and(4.19) and on noting that h (g + w h σ h ) h (g + w h σ h ) +osc h(g + w h σ h,).

13 A POSERIORI ESIMAES FOR HE CAHN HILLIARD EQUAION 1015 Lmma 4.8. h following stimat holds for all h [ τ γ [ w h ] + h (u h f) + h (u h f) ] 1/ h 1/ C ( ) τ γ (w h w) + h (u h u) + h (f f h ) +osc h (u h f,). (4.0) Proof. h proof is similar to th proof of th prvious lmma. Similarly to bfor, w can construct a function φ := α ψ + β ψ,whrα, β ar chosn such that ([ w h ],φ ) = τ γ h [ w h ], (f u h,φ ) = h u h f, and φ C [ τ γ φ Ch [ τ h 1/ [ w h ] + h (u h f) ] 1/, γ h 1/ [ w h ] + h (u h f) ] 1/. W can writ, on rcalling (.1), that τ γ h 1/ [ w h ] + h (u h f) = τ ([ w h ],φ ) (u h f,φ ) γ = τ γ ( w h, φ ) (u h f,φ ) (u h f (u h f),φ ) = τ γ ( (w w h), φ ) +(u u h,φ ) (u h f (u h f),φ ) τ γ (w w h) φ + u u h φ + u h f (u h f) φ τ ] C[ γ (w w h) + h u u h +osc h (u h f,) [ τ γ h 1/ [ w h ] + h (u h f) ] 1/. (4.1) Finally, similarly to (4.19), w hav from an invrs inquality that h (u h f) C h (u h f) C [ h (u h f) +osc h (u h f,) ]. (4.) Combining (4.1) and(4.) concluds th proof. Rmark 4.. h rror quantitis in Lmmas 4.7 and 4.8 contain additional trms,.g. h (w h w) and h (u h u), which ar not prsnt in th rror xprssion for th uppr bound, cf. Corollary 4.1. hrfor, w ar not abl to combin ths two lmmas in ordr to obtain a lowr bound that corrsponds prcisly to th uppr rror stimat in Corollary 4.1, i.. a lowr bound for th rror γ (u h u) + τ (w h w).

14 1016 L. BAŇAS AND R. NÜRNBERG Naturally, such a lowr bound would b dsirabl, as it would giv a thortical proof of th fficincy of th drivd a postriori stimator. 5. Adaptiv algorithms In this sction w introduc svral msh adaption stratgis, that ar basd on th a postriori rror stimator drivd in Sction 4. hroughout this sction, w assum that f = u old h, g = 1 γ uold h aris from a fully discrt approximation of (1.1), whr u old h is th discrt solution from th prvious tim lvl. Hnc f, g ar picwis linar functions on Vold h, th finit lmnt spac from th prvious tim lvl, and thy will only diffr from f h = I h f and g h = I h g, rspctivly, if Vold h V h, i.., whn msh coarsning is mployd. W dfin th following local rror indicators: η u, = 1 h 1/ [ u h ] + 1 γ h (g + w h σ h ) + 1 γ h (g + w h ) ( S S) ; h η w, = τ γ h 1/ [ w h ] + 1 τ h (u h f) ; h (u h f). η c, = h Δ h u h + 1 τ h global rror indicators ar thn dfind as a corrsponding sum of local rror indicators, i.., η u = η w = h η u,, η w,, η c = η c,. h h By using th abov dfinition of th rror indicators, Corollary 4.1 can b rformulatd as (u h u) + τ [ γ (w h w) C η u + η w + η c + 1 γ g g h 1,h + 1 ] τ f f h 1,h. (5.1) Furthr, in th numrical xprimnts w masurd th rlativ rror by th indicator dfind as: η rl = η u + η w + η c u h 1 Rmark 5.1. h rror contributions in (5.1) can b classifid as follows η u corrsponds to th discrtization rror of u; η w corrsponds to th discrtization rror of w; η c corrsponds to th consistncy rror causd by th us of th mass lumpd product (, ) h ; th trms 1 γ g g h 1,h, 1 τ f f h 1,h corrspond to th rror in th approximation of th solution from th prvious tim-lvl causd by msh coarsning, i.., thy ar zro if no lmnts ar coarsnd; w introduc a huristic indicator η τ for tim stp control as follows η τ = 1 γ u h g h 1. Rmark 5.. h discrt dual norm 1,h is difficult to comput in practic, cf. [4], Rmark 5.. Instad of using th dual norm w dfin a simpl coarsning indicator using th L norm as follows: η h, = 1 γ g Ih g 1 γ g Ih g 1, 1 γ g g h 1,h,. Also not that for our choic f = u old h, g = 1 γ uold h,whavthat 1 γ g g h 1,h = τ γ 4 ( 1 τ f f h 1,h ). Hnc th trm 1 τ f f h 1,h can b nglctd, whn τ = O(γ ), which is gnrally th cas in our xprimnts.

15 A POSERIORI ESIMAES FOR HE CAHN HILLIARD EQUAION 1017 Blow w outlin th dtaild dfinitions of th adaptiv algorithms that w usd for our numrical xprimnts. h huristic adaptiv algorithm (VOL1) wasusdin[3,7] for computations for th dgnrat Cahn Hilliard quation. h ida of th algorithm (VOL1) is to locally rfin th msh in such a way, that on has uniformly small lmnts of a prscribd volum vol( ) vol f for h \C h (u h ). h lmnts C h (u h ) ar coarsnd if vol( ) vol c / and rfind if vol( ) >vol c. Not, that in our implmntation w st vol f = h min /, vol c = h max/ indandvol f = h 3 min /6, vol c = h 3 max/6 in3d,whrh min and h max ar givn dsird minimum and maximum msh sizs, rspctivly. h scond adaptiv algorithm (VOL) is basd on th obsrvation that th stimator attains maximum valus at th lmnts from th discrt boundary F h (u h ). h (VOL) algorithm is similar to th (VOL1) algorithm with th addition of an adaptiv control of th constants vol f <vol c to kp th valu of η u blow a prscribd tolranc. Algorithm (VOL) (1) comput u h ; () for all h ; if F h and vol( ) >vol f mark for rfinmnt; if N h and vol( ) > vol f mark for rfinmnt, ls if vol( ) vol f mark for coarsning; if C h and vol( ) vol c /mark for coarsning, ls if vol( ) >vol c mark for rfinmnt; (3) rfin/coarsn msh; if no lmnts wr rfind/coarsnd procd with stp 4 ls procd with stp ; (4) comput η u,ifη u >OLst vol f := vol f / and procd with stp 1, ls procd with stp 5; (5) if η τ >OL τ dcras tim stp τ := τ/; if η τ < 0.01 OL τ incras tim stp τ := min{ τ,τ max }; (6) procd to th nxt tim lvl. h adaptiv algorithm (MAX) is similar to th maximum rror adaptiv stratgy from [] and is dscribd blow. For givn tolrancs OL and OL τ, and coarsning/rfinmnt paramtrs ɛ c, ε c, ε r, vol f, vol c w start with th msh from th prvious tim stp, i.., h = old h, and improv th msh for th nxt tim lvl with th following stps, whr w us th notation η max := max h η u,. Algorithm (MAX) (1) comput u h and η u,, η h,, h ; () for all h,ifη u >OLand η u, >ε r η max mark for rfinmnt; if η u, + η h, <ε c η max mark for coarsning; (3) if η τ >OL τ dcras tim stp τ := τ/; if η τ < 0.01 OL τ incras tim stp τ := min{ τ,τ max }; (4) procd to th nxt tim lvl. h constants ε r, ε c wr chosn as 0.6 and0.05, rspctivly. Not, that th algorithm (MAX) rally only uss th indicator η u for th msh rfinmnt. As confirmd by th numrical xprimnts blow, this also guarants th control ovr th rmaining rror contributions in practic. Rmark 5.3. W not that th coarsning stimat η h was not mployd in th adaptiv stratgy in []. h coarsning stimat is critically important whn computing spinodal dcomposition, whr msh coarsning may lad to an xcssiv loss of information and an unphysical ris of th discrt analogu of th fr nrgy (1.3). W usd a Uzawa-multigrid algorithm for th solution of th discrt systm of nonlinar algbraic quations arising from (.). For mor dtails on this itrativ solvr s [3,4] Failur of th non-localizd stimator 6. Numrical rsults W dmonstrat that a localizd stimator is ssntial for fficint numrical computations. W comput an volution of a squar to a circl for γ = 1 8π on a tim intrval (0, 10 4 ). W mploy th adaptiv stratgy (VOL1) with h min =1/3, h max =1. InFigur1 w display for t =10 4 th computd solution u h,thmsh,

16 1018 L. BAŇAS AND R. NÜRNBERG Figur 1. Solution u h, th msh and th indicators η u, η u at t =10 4. Figur. Solution u h and adaptiv mshs VOL1, VOL, MAX. th localizd stimator η u and non-localizd stimator from Sction 3 dfind as: η u = h ( 1 h 1/ [ u h ] ) + 1 γ h (g + w h ). Clarly, th indicator ηu dos not rflct th charactr of th solution proprly and lads to a substantial ovrstimation of th rror in th aras whr th solution is constant. On th othr hand, th localizd indicator η u is non-zro only in th intrfacial rgion. 6.. Comparison of diffrnt adaptiv stratgis, discrt convrgnc W compar th adaptiv algorithm (VOL1) with th adaptiv algorithm with rfinmnt along th fr boundary (VOL), th maximum stratgy (MAX), and th uniform global msh rfinmnt. In ordr to highlight th diffrncs btwn th adaptiv stratgis (VOL1), (VOL) and (MAX), w display in Figur an xampl of mshs gnratd by th rspctiv adaptiv stratgis. W xamin th convrgnc of η u, η w, η c with rspct to th numbr of dgrs of frdom, with th hlp of an xampl computation for an stablishd intrfac in th form of an llips and γ = 1 8π. W computd with uniform tim stps τ = ( h min ) ,whrh min is th minimum msh siz in th rspctiv computations. h bhaviour was similar at all tim lvls, and w thrfor only prsnt th rsults at tim t = ˆt := h profil of u h at tims t =0andt = ˆt for a uniform msh computation can b sn in Figur 3. h graphs of th dpndnc of η u, η w, η c on th numbr of vrtics at tim ˆt ardpictdinfigurs4 6, rspctivly. A logarithmic scaling is usd in th figurs, which allows us to intrprt th slop α as an xprimntal convrgnc rat of α, sinch #N h in D. h abov rsults support th assumption that th control of η u in th prsntd adaptiv algorithms (or in othr words th rfinmnt in th intrfacial ara only) is sufficint to guarant th control ovr th rmaining indicators η w, η c. h only qualitativ diffrnc btwn th uniform msh rfinmnt and adaptiv stratgis is in th convrgnc rats of η w, which appars to b O(h ) for uniform msh rfinmnt and O(h) for adaptiv msh rfinmnt. h diffrnc can b

17 A POSERIORI ESIMAES FOR HE CAHN HILLIARD EQUAION 1019 Figur 3. Solution u h at t =0andt = ˆt on a uniform msh with h =1/ MAX VOL1 VOL uniform O(h) Figur 4. Convrgnc of η u at t =10 5. Plot of stimator against numbr of dgrs of frdom (dof). accountd to th fact that η w is not localizd to th intrfacial rgion, th rgion that is mainly rfind by th adaptiv mthods. Not, that in th prsnt cas th wors convrgnc rats do not influnc th ovrall convrgnc rat, which is O(h). W conclud, that apart from th abov disadvantag of th (VOL1) algorithm thr is no significant qualitativ diffrnc in th prformanc of th thr adaptiv algorithms. h algorithm (MAX) is prhaps th most flxibl and ffctiv of all thr algorithms; howvr, its prformanc dpnds on th choic of th rfinmnt/coarsning constants. h algorithm (VOL1) is th simplst to implmnt Dpndnc of th stimator on γ W study th fficincy of th adaptiv algorithms with rspct to th paramtr γ. In ordr to obtain rliabl rsults it is dsird that th adaptiv algorithm producs mshs for which th stimat η rl (γ) OL,

18 100 L. BAŇAS AND R. NÜRNBERG 1 MAX VOL1 VOL uniform O(h) O(h^) Figur 5. Convrgnc of η w at t =10 5. Plot of stimator against dof. 10 MAX VOL1 VOL uniform O(h) Figur 6. Convrgnc of η c at t =10 5. Plot of stimator against dof.

19 A POSERIORI ESIMAES FOR HE CAHN HILLIARD EQUAION gamma=1/3pi gamma=1/16pi gamma=1/8pi Figur 7. Evolution of η rl for γ = 1 8π, 1 16π, 1 3π for th (MAX) algorithm. whr OL is a tolranc indpndnt of γ. On th othr hand, in ordr to obtain an fficint adaptiv msh rfinmnt, th minimum msh siz h min (γ), ndd in ordr to kp th rror blow a givn tolranc, should hav a linar dpndnc on γ. W computd an volution of a squar using th adaptiv algorithms (VOL1) and (MAX). In Figurs 7 and 8 w display th tim volution of η rl for γ i = 1, i =0, 1, for th two algorithms. h paramtrs for i 8π th adaptiv msh rfinmnt in th (VOL1) algorithm wr chosn as h min = γi γ, h 018 max = γi γ 0 for i =0, 1,. h tolranc in th (MAX) algorithm was OL = 0.45, which rsultd in similar maximum and minimum msh sizs in both algorithms. In ordr to xclud th influncs of th adaptiv tim stpping on th rror w usd a fixd tim stp τ = (γi ) 10 6 (γ 0 ). h rsults show that th volution of η rl is similar for diffrnt valus of γ if th numbr of msh points in th intrfac is kpt constant (i.. for th abov choics of h min/max ). his is a natural rquirmnt, which undrlins th fficincy of th adaptiv msh rfinmnt. In Figur 9 w display th computd solution u h and th undrlying adaptiv msh obtaind by th (MAX) algorithm for γ = 1 8π Spinodal dcomposition In th nxt xprimnt w prform an xampl computation of spinodal dcomposition. h initial data is obtaind by dfining a coars solution ũ 0 as a random prturbation around 0 on a uniform msh with h =1/0. A smooth initial condition u 0 is thn obtaind as u 0 (x) = ũ 0 (y) 1000 y x dy. Ω Not, that th abov intgral is computd approximatly. W computd th xampl for γ = 1 8π using a uniform msh with h =1/3 and using th adaptiv msh rfinmnt stratgy (MAX). W usd adaptiv tim-stpping basd on th indicator η τ, giving a tim stp siz 10 1 τ h solution and adaptiv msh at diffrnt tim lvls is displayd in Figur 10.

20 10 L. BAŇAS AND R. NÜRNBERG gamma=1/3pi gamma=1/16pi gamma=1/8pi Figur 8. Evolution of η rl for γ = 1 8π, 1 16π, 1 3π for th (VOL1) algorithm. Figur 9. u h at tims t =0,10 5,10 4 for γ = 1 8π.

21 A POSERIORI ESIMAES FOR HE CAHN HILLIARD EQUAION 103 Figur 10. Spinodal dcomposition: u h at tims t =0,10 6,3 10 6,10 5. h volution of th indicator η u is displayd in Figur 11. Clarly, th rror on th uniform msh is almost two tims largr than th tolranc in th initial part of th computation, whil th rror on adaptiv mshs is always blow th tolranc. W found that too much coarsning in th computations of spinodal dcomposition could lad to an unphysical ris in th discrt nrgy, which undrlins th importanc of th coarsning stimat Coarsning in 3D h last xprimnt is to dmonstrat th prformanc of adaptiv msh rfinmnt in 3D computations. h zro lvl st of th initial condition consistd of two cubs of slightly diffrnt sizs. W computd th xampl using th adaptiv stratgy (VOL1) with fixd tim stp τ =10 6. h volution of th zro lvl st of th computd solution and a cut through th adaptiv msh at x 3 = 0 ar displayd in Figur 1. h volution of η rl in Figur 13 indicats a good control of th approximation rror.

22 104 L. BAŇAS AND R. NÜRNBERG 0.7 MAX uniform Figur 11. Evolution of th stimat η u for a uniform msh of fixd siz h =1/3 and th algorithm (MAX). Figur 1. 3D coarsning: zro lvl st of u h and cut through th msh at x 3 =0attims t =0,10 5,10 4,

23 A POSERIORI ESIMAES FOR HE CAHN HILLIARD EQUAION Figur 13. 3D coarsning: volution of η rl. Rfrncs [1] N.D. Alikakos, P.W. Bats and X.F. Chn, h convrgnc of solutions of th Cahn Hilliard quation to th solution of th Hl Shaw modl. Arch. Rational Mch. Anal. 18 (1994) [] L. Baňas and R. Nürnbrg, Adaptiv finit lmnt mthods for Cahn Hilliard quations. J. Comput. Appl. Math. 18 (008) 11. [3] L. Baňas and R. Nürnbrg, Finit lmnt approximation of a thr dimnsional phas fild modl for void lctromigration. J. Sci. Comp. 37 (008) 0 3. [4] L. Baňas and R. Nürnbrg, Phas fild computations for surfac diffusion and void lctromigration in R 3. Comput. Vis. Sci. (008), doi: /s [5] J.W. Barrtt and J.F. Blowy, Finit lmnt approximation of a modl for phas sparation of a multi-componnt alloy with non-smooth fr nrgy. Numr. Math. 77 (1997) [6] J.W. Barrtt, J.F. Blowy and H. Garck, Finit lmnt approximation of th Cahn Hilliard quation with dgnrat mobility. SIAM J. Numr. Anal. 37 (1999) [7] J.W. Barrtt, R. Nürnbrg and V. Styls, Finit lmnt approximation of a phas fild modl for void lctromigration. SIAM J. Numr. Anal. 4 (004) [8] J.F. Blowy and C.M. Elliott, h Cahn Hilliard gradint thory for phas sparation with non-smooth fr nrgy. Part I: Mathmatical analysis. Europan J. Appl. Math. (1991) [9] J.F. Blowy and C.M. Elliott, h Cahn Hilliard gradint thory for phas sparation with non-smooth fr nrgy. Part II: Numrical analysis. Europan J. Appl. Math. 3 (199) [10] D. Brass, A postriori rror stimators for obstacl problms anothr look. Numr. Math. 101 (005) [11] J.W. Cahn, On spinodal dcomposition. Acta Mtall. 9 (1961) [1] J.W. Cahn and J.E. Hilliard, Fr nrgy of a non-uniform systm. I. Intrfacial fr nrgy. J. Chm. Phys. 8 (1958) [13] X. Chn, Spctrum for th Alln Cahn, Cahn Hilliard, and phas-fild quations for gnric intrfacs. Comm. Partial Diffr. Equ. 19 (1994) [14] Z. Chn and R.H. Nochtto, Rsidual typ a postriori rror stimats for lliptic obstacl problms. Numr. Math. 84 (000) [15] P. Clémnt, Approximation by finit lmnt functions using local rgularization. RAIRO Anal. Numér. 9 (1975) [16] C.M. Elliott and Z. Songmu, On th Cahn Hilliard quation. Arch. Rational Mch. Anal. 96 (1986)

24 106 L. BAŇAS AND R. NÜRNBERG [17] C.M. Elliott, D.A. Frnch and F.A. Milnr, A scond ordr splitting mthod for th Cahn Hilliard quation. Numr. Math. 54 (1989) [18] X. Fng and A. Prohl, Error analysis of a mixd finit lmnt mthod for th Cahn Hilliard quation. Numr. Math. 99 (004) [19] X. Fng and H. Wu, A postriori rror stimats for finit lmnt approximations of th Cahn Hilliard quation and th Hl Shaw flow. J. Comput. Math. 6 (008) [0] M. Hintrmüllr and R.H.W. Hopp, Goal-orintd adaptivity in control constraind optimal control of partial diffrntial quations. SIAM J. Control Optim. 47 (008) [1] M. Hintrmüllr, R.H.W. Hopp, Y. Iliash and M. Kiwg, An a postriori rror analysis of adaptiv finit lmnt mthods for distributd lliptic control problms with control constraints. ESAIM: COCV 14 (008) [] J. Kim, K. Kang and J. Lowngrub, Consrvativ multigrid mthods for Cahn Hilliard fluids. J. Comput. Phys. 193 (004) [3] L. Modica, Gradint thory of phas transitions with boundary contact nrgy. Ann. Inst. H. Poincaré Anal. NonLinéair 4 (1987) [4] K.-S. Moon, R.H. Nochtto,. von Ptrsdorff and C.-S. Zhang, A postriori rror analysis for parabolic variational inqualitis. ESAIM: MAN 41 (007) [5] R.H. Nochtto and L.B. Wahlbin, Positivity prsrving finit lmnt approximation. Math. Comp. 71 (00) [6] R.L. Pgo, Front migration in th nonlinar Cahn-Hilliard quation. Proc. Roy. Soc. London Sr. A 4 (1989) [7] A. Vsr, Efficint and rliabl a postriori rror stimators for lliptic obstacl problms. SIAM J. Numr. Anal. 39 (001) [8] A. Vsr, On a postriori rror stimation for constant obstacl problms, in Numrical mthods for viscosity solutions and applications (Hraklion, 1999), M. Falcon and C. Makridakis Eds., Sr. Adv. Math. Appl. Sci. 59, World Sci. Publ., Rivr Edg, USA (001) [9] R. Vrfürth, A Rviw of a Postriori Error Estimation and Adaptiv Msh-Rfinmnt chniqus. ubnr-wily, Nw York (1996).

Another view for a posteriori error estimates for variational inequalities of the second kind

Another view for a posteriori error estimates for variational inequalities of the second kind Accptd by Applid Numrical Mathmatics in July 2013 Anothr viw for a postriori rror stimats for variational inqualitis of th scond kind Fi Wang 1 and Wimin Han 2 Abstract. In this papr, w giv anothr viw

More information

Einstein Equations for Tetrad Fields

Einstein Equations for Tetrad Fields Apiron, Vol 13, No, Octobr 006 6 Einstin Equations for Ttrad Filds Ali Rıza ŞAHİN, R T L Istanbul (Turky) Evry mtric tnsor can b xprssd by th innr product of ttrad filds W prov that Einstin quations for

More information

Homotopy perturbation technique

Homotopy perturbation technique Comput. Mthods Appl. Mch. Engrg. 178 (1999) 257±262 www.lsvir.com/locat/cma Homotopy prturbation tchniqu Ji-Huan H 1 Shanghai Univrsity, Shanghai Institut of Applid Mathmatics and Mchanics, Shanghai 272,

More information

A Weakly Over-Penalized Non-Symmetric Interior Penalty Method

A Weakly Over-Penalized Non-Symmetric Interior Penalty Method Europan Socity of Computational Mthods in Scincs and Enginring (ESCMSE Journal of Numrical Analysis, Industrial and Applid Mathmatics (JNAIAM vol.?, no.?, 2007, pp.?-? ISSN 1790 8140 A Wakly Ovr-Pnalizd

More information

AS 5850 Finite Element Analysis

AS 5850 Finite Element Analysis AS 5850 Finit Elmnt Analysis Two-Dimnsional Linar Elasticity Instructor Prof. IIT Madras Equations of Plan Elasticity - 1 displacmnt fild strain- displacmnt rlations (infinitsimal strain) in matrix form

More information

Construction of asymmetric orthogonal arrays of strength three via a replacement method

Construction of asymmetric orthogonal arrays of strength three via a replacement method isid/ms/26/2 Fbruary, 26 http://www.isid.ac.in/ statmath/indx.php?modul=prprint Construction of asymmtric orthogonal arrays of strngth thr via a rplacmnt mthod Tian-fang Zhang, Qiaoling Dng and Alok Dy

More information

Introduction to Arithmetic Geometry Fall 2013 Lecture #20 11/14/2013

Introduction to Arithmetic Geometry Fall 2013 Lecture #20 11/14/2013 18.782 Introduction to Arithmtic Gomtry Fall 2013 Lctur #20 11/14/2013 20.1 Dgr thorm for morphisms of curvs Lt us rstat th thorm givn at th nd of th last lctur, which w will now prov. Thorm 20.1. Lt φ:

More information

The Matrix Exponential

The Matrix Exponential Th Matrix Exponntial (with xrciss) by D. Klain Vrsion 207.0.05 Corrctions and commnts ar wlcom. Th Matrix Exponntial For ach n n complx matrix A, dfin th xponntial of A to b th matrix A A k I + A + k!

More information

Addition of angular momentum

Addition of angular momentum Addition of angular momntum April, 0 Oftn w nd to combin diffrnt sourcs of angular momntum to charactriz th total angular momntum of a systm, or to divid th total angular momntum into parts to valuat th

More information

LINEAR DELAY DIFFERENTIAL EQUATION WITH A POSITIVE AND A NEGATIVE TERM

LINEAR DELAY DIFFERENTIAL EQUATION WITH A POSITIVE AND A NEGATIVE TERM Elctronic Journal of Diffrntial Equations, Vol. 2003(2003), No. 92, pp. 1 6. ISSN: 1072-6691. URL: http://jd.math.swt.du or http://jd.math.unt.du ftp jd.math.swt.du (login: ftp) LINEAR DELAY DIFFERENTIAL

More information

The Matrix Exponential

The Matrix Exponential Th Matrix Exponntial (with xrciss) by Dan Klain Vrsion 28928 Corrctions and commnts ar wlcom Th Matrix Exponntial For ach n n complx matrix A, dfin th xponntial of A to b th matrix () A A k I + A + k!

More information

arxiv: v1 [math.na] 3 Mar 2016

arxiv: v1 [math.na] 3 Mar 2016 MATHEMATICS OF COMPUTATION Volum 00, Numbr 0, Pags 000 000 S 0025-5718(XX)0000-0 arxiv:1603.01024v1 [math.na] 3 Mar 2016 RESIDUAL-BASED A POSTERIORI ERROR ESTIMATE FOR INTERFACE PROBLEMS: NONCONFORMING

More information

1 Isoparametric Concept

1 Isoparametric Concept UNIVERSITY OF CALIFORNIA BERKELEY Dpartmnt of Civil Enginring Spring 06 Structural Enginring, Mchanics and Matrials Profssor: S. Govindj Nots on D isoparamtric lmnts Isoparamtric Concpt Th isoparamtric

More information

Addition of angular momentum

Addition of angular momentum Addition of angular momntum April, 07 Oftn w nd to combin diffrnt sourcs of angular momntum to charactriz th total angular momntum of a systm, or to divid th total angular momntum into parts to valuat

More information

Title: Vibrational structure of electronic transition

Title: Vibrational structure of electronic transition Titl: Vibrational structur of lctronic transition Pag- Th band spctrum sn in th Ultra-Violt (UV) and visibl (VIS) rgions of th lctromagntic spctrum can not intrprtd as vibrational and rotational spctrum

More information

A LOCALLY CONSERVATIVE LDG METHOD FOR THE INCOMPRESSIBLE NAVIER-STOKES EQUATIONS

A LOCALLY CONSERVATIVE LDG METHOD FOR THE INCOMPRESSIBLE NAVIER-STOKES EQUATIONS A LOCALLY CONSERVATIVE LDG METHOD FOR THE INCOMPRESSIBLE NAVIER-STOES EQUATIONS BERNARDO COCBURN, GUIDO ANSCHAT, AND DOMINI SCHÖTZAU Abstract. In this papr, a nw local discontinuous Galrkin mthod for th

More information

Symmetric centrosymmetric matrix vector multiplication

Symmetric centrosymmetric matrix vector multiplication Linar Algbra and its Applications 320 (2000) 193 198 www.lsvir.com/locat/laa Symmtric cntrosymmtric matrix vctor multiplication A. Mlman 1 Dpartmnt of Mathmatics, Univrsity of San Francisco, San Francisco,

More information

(Upside-Down o Direct Rotation) β - Numbers

(Upside-Down o Direct Rotation) β - Numbers Amrican Journal of Mathmatics and Statistics 014, 4(): 58-64 DOI: 10593/jajms0140400 (Upsid-Down o Dirct Rotation) β - Numbrs Ammar Sddiq Mahmood 1, Shukriyah Sabir Ali,* 1 Dpartmnt of Mathmatics, Collg

More information

Cramér-Rao Inequality: Let f(x; θ) be a probability density function with continuous parameter

Cramér-Rao Inequality: Let f(x; θ) be a probability density function with continuous parameter WHEN THE CRAMÉR-RAO INEQUALITY PROVIDES NO INFORMATION STEVEN J. MILLER Abstract. W invstigat a on-paramtr family of probability dnsitis (rlatd to th Parto distribution, which dscribs many natural phnomna)

More information

3 Finite Element Parametric Geometry

3 Finite Element Parametric Geometry 3 Finit Elmnt Paramtric Gomtry 3. Introduction Th intgral of a matrix is th matrix containing th intgral of ach and vry on of its original componnts. Practical finit lmnt analysis rquirs intgrating matrics,

More information

Derangements and Applications

Derangements and Applications 2 3 47 6 23 Journal of Intgr Squncs, Vol. 6 (2003), Articl 03..2 Drangmnts and Applications Mhdi Hassani Dpartmnt of Mathmatics Institut for Advancd Studis in Basic Scincs Zanjan, Iran mhassani@iasbs.ac.ir

More information

A Propagating Wave Packet Group Velocity Dispersion

A Propagating Wave Packet Group Velocity Dispersion Lctur 8 Phys 375 A Propagating Wav Packt Group Vlocity Disprsion Ovrviw and Motivation: In th last lctur w lookd at a localizd solution t) to th 1D fr-particl Schrödingr quation (SE) that corrsponds to

More information

22/ Breakdown of the Born-Oppenheimer approximation. Selection rules for rotational-vibrational transitions. P, R branches.

22/ Breakdown of the Born-Oppenheimer approximation. Selection rules for rotational-vibrational transitions. P, R branches. Subjct Chmistry Papr No and Titl Modul No and Titl Modul Tag 8/ Physical Spctroscopy / Brakdown of th Born-Oppnhimr approximation. Slction ruls for rotational-vibrational transitions. P, R branchs. CHE_P8_M

More information

Lecture 37 (Schrödinger Equation) Physics Spring 2018 Douglas Fields

Lecture 37 (Schrödinger Equation) Physics Spring 2018 Douglas Fields Lctur 37 (Schrödingr Equation) Physics 6-01 Spring 018 Douglas Filds Rducd Mass OK, so th Bohr modl of th atom givs nrgy lvls: E n 1 k m n 4 But, this has on problm it was dvlopd assuming th acclration

More information

Discontinuous Galerkin and Mimetic Finite Difference Methods for Coupled Stokes-Darcy Flows on Polygonal and Polyhedral Grids

Discontinuous Galerkin and Mimetic Finite Difference Methods for Coupled Stokes-Darcy Flows on Polygonal and Polyhedral Grids Discontinuous Galrkin and Mimtic Finit Diffrnc Mthods for Coupld Stoks-Darcy Flows on Polygonal and Polyhdral Grids Konstantin Lipnikov Danail Vassilv Ivan Yotov Abstract W study locally mass consrvativ

More information

On the irreducibility of some polynomials in two variables

On the irreducibility of some polynomials in two variables ACTA ARITHMETICA LXXXII.3 (1997) On th irrducibility of som polynomials in two variabls by B. Brindza and Á. Pintér (Dbrcn) To th mmory of Paul Erdős Lt f(x) and g(y ) b polynomials with intgral cofficints

More information

Difference -Analytical Method of The One-Dimensional Convection-Diffusion Equation

Difference -Analytical Method of The One-Dimensional Convection-Diffusion Equation Diffrnc -Analytical Mthod of Th On-Dimnsional Convction-Diffusion Equation Dalabav Umurdin Dpartmnt mathmatic modlling, Univrsity of orld Economy and Diplomacy, Uzbistan Abstract. An analytical diffrncing

More information

Higher order derivatives

Higher order derivatives Robrto s Nots on Diffrntial Calculus Chaptr 4: Basic diffrntiation ruls Sction 7 Highr ordr drivativs What you nd to know alrady: Basic diffrntiation ruls. What you can larn hr: How to rpat th procss of

More information

Dynamic Modelling of Hoisting Steel Wire Rope. Da-zhi CAO, Wen-zheng DU, Bao-zhu MA *

Dynamic Modelling of Hoisting Steel Wire Rope. Da-zhi CAO, Wen-zheng DU, Bao-zhu MA * 17 nd Intrnational Confrnc on Mchanical Control and Automation (ICMCA 17) ISBN: 978-1-6595-46-8 Dynamic Modlling of Hoisting Stl Wir Rop Da-zhi CAO, Wn-zhng DU, Bao-zhu MA * and Su-bing LIU Xi an High

More information

Data Assimilation 1. Alan O Neill National Centre for Earth Observation UK

Data Assimilation 1. Alan O Neill National Centre for Earth Observation UK Data Assimilation 1 Alan O Nill National Cntr for Earth Obsrvation UK Plan Motivation & basic idas Univariat (scalar) data assimilation Multivariat (vctor) data assimilation 3d-Variational Mthod (& optimal

More information

2008 AP Calculus BC Multiple Choice Exam

2008 AP Calculus BC Multiple Choice Exam 008 AP Multipl Choic Eam Nam 008 AP Calculus BC Multipl Choic Eam Sction No Calculator Activ AP Calculus 008 BC Multipl Choic. At tim t 0, a particl moving in th -plan is th acclration vctor of th particl

More information

Chapter 10. The singular integral Introducing S(n) and J(n)

Chapter 10. The singular integral Introducing S(n) and J(n) Chaptr Th singular intgral Our aim in this chaptr is to rplac th functions S (n) and J (n) by mor convnint xprssions; ths will b calld th singular sris S(n) and th singular intgral J(n). This will b don

More information

Finite element discretization of Laplace and Poisson equations

Finite element discretization of Laplace and Poisson equations Finit lmnt discrtization of Laplac and Poisson quations Yashwanth Tummala Tutor: Prof S.Mittal 1 Outlin Finit Elmnt Mthod for 1D Introduction to Poisson s and Laplac s Equations Finit Elmnt Mthod for 2D-Discrtization

More information

That is, we start with a general matrix: And end with a simpler matrix:

That is, we start with a general matrix: And end with a simpler matrix: DIAGON ALIZATION OF THE STR ESS TEN SOR INTRO DUCTIO N By th us of Cauchy s thorm w ar abl to rduc th numbr of strss componnts in th strss tnsor to only nin valus. An additional simplification of th strss

More information

Recall that by Theorems 10.3 and 10.4 together provide us the estimate o(n2 ), S(q) q 9, q=1

Recall that by Theorems 10.3 and 10.4 together provide us the estimate o(n2 ), S(q) q 9, q=1 Chaptr 11 Th singular sris Rcall that by Thorms 10 and 104 togthr provid us th stimat 9 4 n 2 111 Rn = SnΓ 2 + on2, whr th singular sris Sn was dfind in Chaptr 10 as Sn = q=1 Sq q 9, with Sq = 1 a q gcda,q=1

More information

Symmetric Interior penalty discontinuous Galerkin methods for elliptic problems in polygons

Symmetric Interior penalty discontinuous Galerkin methods for elliptic problems in polygons Symmtric Intrior pnalty discontinuous Galrkin mthods for lliptic problms in polygons F. Müllr and D. Schötzau and Ch. Schwab Rsarch Rport No. 2017-15 March 2017 Sminar für Angwandt Mathmatik Eidgnössisch

More information

c 2009 Society for Industrial and Applied Mathematics

c 2009 Society for Industrial and Applied Mathematics SIAM J. NUMER. ANAL. Vol. 47 No. 3 pp. 2132 2156 c 2009 Socity for Industrial and Applid Mathmatics RECOVERY-BASED ERROR ESTIMATOR FOR INTERFACE PROBLEMS: CONFORMING LINEAR ELEMENTS ZHIQIANG CAI AND SHUN

More information

Brief Introduction to Statistical Mechanics

Brief Introduction to Statistical Mechanics Brif Introduction to Statistical Mchanics. Purpos: Ths nots ar intndd to provid a vry quick introduction to Statistical Mchanics. Th fild is of cours far mor vast than could b containd in ths fw pags.

More information

CE 530 Molecular Simulation

CE 530 Molecular Simulation CE 53 Molcular Simulation Lctur 8 Fr-nrgy calculations David A. Kofk Dpartmnt of Chmical Enginring SUNY Buffalo kofk@ng.buffalo.du 2 Fr-Enrgy Calculations Uss of fr nrgy Phas quilibria Raction quilibria

More information

SECTION where P (cos θ, sin θ) and Q(cos θ, sin θ) are polynomials in cos θ and sin θ, provided Q is never equal to zero.

SECTION where P (cos θ, sin θ) and Q(cos θ, sin θ) are polynomials in cos θ and sin θ, provided Q is never equal to zero. SETION 6. 57 6. Evaluation of Dfinit Intgrals Exampl 6.6 W hav usd dfinit intgrals to valuat contour intgrals. It may com as a surpris to larn that contour intgrals and rsidus can b usd to valuat crtain

More information

Deift/Zhou Steepest descent, Part I

Deift/Zhou Steepest descent, Part I Lctur 9 Dift/Zhou Stpst dscnt, Part I W now focus on th cas of orthogonal polynomials for th wight w(x) = NV (x), V (x) = t x2 2 + x4 4. Sinc th wight dpnds on th paramtr N N w will writ π n,n, a n,n,

More information

Rational Approximation for the one-dimensional Bratu Equation

Rational Approximation for the one-dimensional Bratu Equation Intrnational Journal of Enginring & Tchnology IJET-IJES Vol:3 o:05 5 Rational Approximation for th on-dimnsional Bratu Equation Moustafa Aly Soliman Chmical Enginring Dpartmnt, Th British Univrsity in

More information

Symmetric Interior Penalty Galerkin Method for Elliptic Problems

Symmetric Interior Penalty Galerkin Method for Elliptic Problems Symmtric Intrior Pnalty Galrkin Mthod for Elliptic Problms Ykatrina Epshtyn and Béatric Rivièr Dpartmnt of Mathmatics, Univrsity of Pittsburgh, 3 Thackray, Pittsburgh, PA 56, U.S.A. Abstract This papr

More information

2.3 Matrix Formulation

2.3 Matrix Formulation 23 Matrix Formulation 43 A mor complicatd xampl ariss for a nonlinar systm of diffrntial quations Considr th following xampl Exampl 23 x y + x( x 2 y 2 y x + y( x 2 y 2 (233 Transforming to polar coordinats,

More information

ON A SECOND ORDER RATIONAL DIFFERENCE EQUATION

ON A SECOND ORDER RATIONAL DIFFERENCE EQUATION Hacttp Journal of Mathmatics and Statistics Volum 41(6) (2012), 867 874 ON A SECOND ORDER RATIONAL DIFFERENCE EQUATION Nourssadat Touafk Rcivd 06:07:2011 : Accptd 26:12:2011 Abstract In this papr, w invstigat

More information

u 3 = u 3 (x 1, x 2, x 3 )

u 3 = u 3 (x 1, x 2, x 3 ) Lctur 23: Curvilinar Coordinats (RHB 8.0 It is oftn convnint to work with variabls othr than th Cartsian coordinats x i ( = x, y, z. For xampl in Lctur 5 w mt sphrical polar and cylindrical polar coordinats.

More information

Reparameterization and Adaptive Quadrature for the Isogeometric Discontinuous Galerkin Method

Reparameterization and Adaptive Quadrature for the Isogeometric Discontinuous Galerkin Method Rparamtrization and Adaptiv Quadratur for th Isogomtric Discontinuous Galrkin Mthod Agns Silr, Brt Jüttlr 2 Doctoral Program Computational Mathmatics 2 Institut of Applid Gomtry Johanns Kplr Univrsity

More information

Bifurcation Theory. , a stationary point, depends on the value of α. At certain values

Bifurcation Theory. , a stationary point, depends on the value of α. At certain values Dnamic Macroconomic Thor Prof. Thomas Lux Bifurcation Thor Bifurcation: qualitativ chang in th natur of th solution occurs if a paramtr passs through a critical point bifurcation or branch valu. Local

More information

MATH 319, WEEK 15: The Fundamental Matrix, Non-Homogeneous Systems of Differential Equations

MATH 319, WEEK 15: The Fundamental Matrix, Non-Homogeneous Systems of Differential Equations MATH 39, WEEK 5: Th Fundamntal Matrix, Non-Homognous Systms of Diffrntial Equations Fundamntal Matrics Considr th problm of dtrmining th particular solution for an nsmbl of initial conditions For instanc,

More information

cycle that does not cross any edges (including its own), then it has at least

cycle that does not cross any edges (including its own), then it has at least W prov th following thorm: Thorm If a K n is drawn in th plan in such a way that it has a hamiltonian cycl that dos not cross any dgs (including its own, thn it has at last n ( 4 48 π + O(n crossings Th

More information

Inference Methods for Stochastic Volatility Models

Inference Methods for Stochastic Volatility Models Intrnational Mathmatical Forum, Vol 8, 03, no 8, 369-375 Infrnc Mthods for Stochastic Volatility Modls Maddalna Cavicchioli Cá Foscari Univrsity of Vnic Advancd School of Economics Cannargio 3, Vnic, Italy

More information

The Equitable Dominating Graph

The Equitable Dominating Graph Intrnational Journal of Enginring Rsarch and Tchnology. ISSN 0974-3154 Volum 8, Numbr 1 (015), pp. 35-4 Intrnational Rsarch Publication Hous http://www.irphous.com Th Equitabl Dominating Graph P.N. Vinay

More information

1 Minimum Cut Problem

1 Minimum Cut Problem CS 6 Lctur 6 Min Cut and argr s Algorithm Scribs: Png Hui How (05), Virginia Dat: May 4, 06 Minimum Cut Problm Today, w introduc th minimum cut problm. This problm has many motivations, on of which coms

More information

BINOMIAL COEFFICIENTS INVOLVING INFINITE POWERS OF PRIMES. 1. Statement of results

BINOMIAL COEFFICIENTS INVOLVING INFINITE POWERS OF PRIMES. 1. Statement of results BINOMIAL COEFFICIENTS INVOLVING INFINITE POWERS OF PRIMES DONALD M. DAVIS Abstract. If p is a prim and n a positiv intgr, lt ν p (n dnot th xponnt of p in n, and u p (n n/p νp(n th unit part of n. If α

More information

A ROBUST NONCONFORMING H 2 -ELEMENT

A ROBUST NONCONFORMING H 2 -ELEMENT MAHEMAICS OF COMPUAION Volum 70, Numbr 234, Pags 489 505 S 0025-5718(00)01230-8 Articl lctronically publishd on Fbruary 23, 2000 A ROBUS NONCONFORMING H 2 -ELEMEN RYGVE K. NILSSEN, XUE-CHENG AI, AND RAGNAR

More information

Supplementary Materials

Supplementary Materials 6 Supplmntary Matrials APPENDIX A PHYSICAL INTERPRETATION OF FUEL-RATE-SPEED FUNCTION A truck running on a road with grad/slop θ positiv if moving up and ngativ if moving down facs thr rsistancs: arodynamic

More information

ON THE DISTRIBUTION OF THE ELLIPTIC SUBSET SUM GENERATOR OF PSEUDORANDOM NUMBERS

ON THE DISTRIBUTION OF THE ELLIPTIC SUBSET SUM GENERATOR OF PSEUDORANDOM NUMBERS INTEGERS: ELECTRONIC JOURNAL OF COMBINATORIAL NUMBER THEORY 8 2008, #A3 ON THE DISTRIBUTION OF THE ELLIPTIC SUBSET SUM GENERATOR OF PSEUDORANDOM NUMBERS Edwin D. El-Mahassni Dpartmnt of Computing, Macquari

More information

The van der Waals interaction 1 D. E. Soper 2 University of Oregon 20 April 2012

The van der Waals interaction 1 D. E. Soper 2 University of Oregon 20 April 2012 Th van dr Waals intraction D. E. Sopr 2 Univrsity of Orgon 20 pril 202 Th van dr Waals intraction is discussd in Chaptr 5 of J. J. Sakurai, Modrn Quantum Mchanics. Hr I tak a look at it in a littl mor

More information

UNIFIED A POSTERIORI ERROR ESTIMATOR FOR FINITE ELEMENT METHODS FOR THE STOKES EQUATIONS

UNIFIED A POSTERIORI ERROR ESTIMATOR FOR FINITE ELEMENT METHODS FOR THE STOKES EQUATIONS UNIFIED A POSTERIORI ERROR ESTIMATOR FOR FINITE ELEMENT METHODS FOR THE STOKES EQUATIONS JUNPING WANG, YANQIU WANG, AND XIU YE Abstract. This papr is concrnd with rsidual typ a postriori rror stimators

More information

Application of Vague Soft Sets in students evaluation

Application of Vague Soft Sets in students evaluation Availabl onlin at www.plagiarsarchlibrary.com Advancs in Applid Scinc Rsarch, 0, (6):48-43 ISSN: 0976-860 CODEN (USA): AASRFC Application of Vagu Soft Sts in studnts valuation B. Chtia*and P. K. Das Dpartmnt

More information

Quasi-Classical States of the Simple Harmonic Oscillator

Quasi-Classical States of the Simple Harmonic Oscillator Quasi-Classical Stats of th Simpl Harmonic Oscillator (Draft Vrsion) Introduction: Why Look for Eignstats of th Annihilation Oprator? Excpt for th ground stat, th corrspondnc btwn th quantum nrgy ignstats

More information

COMPUTER GENERATED HOLOGRAMS Optical Sciences 627 W.J. Dallas (Monday, April 04, 2005, 8:35 AM) PART I: CHAPTER TWO COMB MATH.

COMPUTER GENERATED HOLOGRAMS Optical Sciences 627 W.J. Dallas (Monday, April 04, 2005, 8:35 AM) PART I: CHAPTER TWO COMB MATH. C:\Dallas\0_Courss\03A_OpSci_67\0 Cgh_Book\0_athmaticalPrliminaris\0_0 Combath.doc of 8 COPUTER GENERATED HOLOGRAS Optical Scincs 67 W.J. Dallas (onday, April 04, 005, 8:35 A) PART I: CHAPTER TWO COB ATH

More information

Hydrogen Atom and One Electron Ions

Hydrogen Atom and One Electron Ions Hydrogn Atom and On Elctron Ions Th Schrödingr quation for this two-body problm starts out th sam as th gnral two-body Schrödingr quation. First w sparat out th motion of th cntr of mass. Th intrnal potntial

More information

5.80 Small-Molecule Spectroscopy and Dynamics

5.80 Small-Molecule Spectroscopy and Dynamics MIT OpnCoursWar http://ocw.mit.du 5.80 Small-Molcul Spctroscopy and Dynamics Fall 008 For information about citing ths matrials or our Trms of Us, visit: http://ocw.mit.du/trms. Lctur # 3 Supplmnt Contnts

More information

u x v x dx u x v x v x u x dx d u x v x u x v x dx u x v x dx Integration by Parts Formula

u x v x dx u x v x v x u x dx d u x v x u x v x dx u x v x dx Integration by Parts Formula 7. Intgration by Parts Each drivativ formula givs ris to a corrsponding intgral formula, as w v sn many tims. Th drivativ product rul yilds a vry usful intgration tchniqu calld intgration by parts. Starting

More information

RELIABLE AND EFFICIENT A POSTERIORI ERROR ESTIMATES OF DG METHODS FOR A SIMPLIFIED FRICTIONAL CONTACT PROBLEM

RELIABLE AND EFFICIENT A POSTERIORI ERROR ESTIMATES OF DG METHODS FOR A SIMPLIFIED FRICTIONAL CONTACT PROBLEM INTERNATIONAL JOURNAL OF NUMERICAL ANALYSIS AND MODELING Volum 16, Numbr 1, Pags 49 62 c 2019 Institut for Scintific Computing and Information RELIABLE AND EFFICIENT A POSTERIORI ERROR ESTIMATES OF DG

More information

BINOMIAL COEFFICIENTS INVOLVING INFINITE POWERS OF PRIMES

BINOMIAL COEFFICIENTS INVOLVING INFINITE POWERS OF PRIMES BINOMIAL COEFFICIENTS INVOLVING INFINITE POWERS OF PRIMES DONALD M. DAVIS Abstract. If p is a prim (implicit in notation and n a positiv intgr, lt ν(n dnot th xponnt of p in n, and U(n n/p ν(n, th unit

More information

CS 361 Meeting 12 10/3/18

CS 361 Meeting 12 10/3/18 CS 36 Mting 2 /3/8 Announcmnts. Homwork 4 is du Friday. If Friday is Mountain Day, homwork should b turnd in at my offic or th dpartmnt offic bfor 4. 2. Homwork 5 will b availabl ovr th wknd. 3. Our midtrm

More information

CPSC 665 : An Algorithmist s Toolkit Lecture 4 : 21 Jan Linear Programming

CPSC 665 : An Algorithmist s Toolkit Lecture 4 : 21 Jan Linear Programming CPSC 665 : An Algorithmist s Toolkit Lctur 4 : 21 Jan 2015 Lcturr: Sushant Sachdva Linar Programming Scrib: Rasmus Kyng 1. Introduction An optimization problm rquirs us to find th minimum or maximum) of

More information

Fourier Transforms and the Wave Equation. Key Mathematics: More Fourier transform theory, especially as applied to solving the wave equation.

Fourier Transforms and the Wave Equation. Key Mathematics: More Fourier transform theory, especially as applied to solving the wave equation. Lur 7 Fourir Transforms and th Wav Euation Ovrviw and Motivation: W first discuss a fw faturs of th Fourir transform (FT), and thn w solv th initial-valu problm for th wav uation using th Fourir transform

More information

An Extensive Study of Approximating the Periodic. Solutions of the Prey Predator System

An Extensive Study of Approximating the Periodic. Solutions of the Prey Predator System pplid athmatical Scincs Vol. 00 no. 5 5 - n xtnsiv Study of pproximating th Priodic Solutions of th Pry Prdator Systm D. Vnu Gopala Rao * ailing addrss: Plot No.59 Sctor-.V.P.Colony Visahapatnam 50 07

More information

Forces. Quantum ElectroDynamics. α = = We have now:

Forces. Quantum ElectroDynamics. α = = We have now: W hav now: Forcs Considrd th gnral proprtis of forcs mdiatd by xchang (Yukawa potntial); Examind consrvation laws which ar obyd by (som) forcs. W will nxt look at thr forcs in mor dtail: Elctromagntic

More information

surface of a dielectric-metal interface. It is commonly used today for discovering the ways in

surface of a dielectric-metal interface. It is commonly used today for discovering the ways in Surfac plasmon rsonanc is snsitiv mchanism for obsrving slight changs nar th surfac of a dilctric-mtal intrfac. It is commonl usd toda for discovring th was in which protins intract with thir nvironmnt,

More information

Elements of Statistical Thermodynamics

Elements of Statistical Thermodynamics 24 Elmnts of Statistical Thrmodynamics Statistical thrmodynamics is a branch of knowldg that has its own postulats and tchniqus. W do not attmpt to giv hr vn an introduction to th fild. In this chaptr,

More information

An interior penalty method for a two dimensional curl-curl and grad-div problem

An interior penalty method for a two dimensional curl-curl and grad-div problem ANZIAM J. 50 (CTAC2008) pp.c947 C975, 2009 C947 An intrior pnalty mthod for a two dimnsional curl-curl and grad-div problm S. C. Brnnr 1 J. Cui 2 L.-Y. Sung 3 (Rcivd 30 Octobr 2008; rvisd 30 April 2009)

More information

MORTAR ADAPTIVITY IN MIXED METHODS FOR FLOW IN POROUS MEDIA

MORTAR ADAPTIVITY IN MIXED METHODS FOR FLOW IN POROUS MEDIA INTERNATIONAL JOURNAL OF NUMERICAL ANALYSIS AND MODELING Volum 2, Numbr 3, Pags 241 282 c 25 Institut for Scintific Computing and Information MORTAR ADAPTIVITY IN MIXED METHODS FOR FLOW IN POROUS MEDIA

More information

NONCONFORMING FINITE ELEMENTS FOR REISSNER-MINDLIN PLATES

NONCONFORMING FINITE ELEMENTS FOR REISSNER-MINDLIN PLATES NONCONFORMING FINITE ELEMENTS FOR REISSNER-MINDLIN PLATES C. CHINOSI Dipartimnto di Scinz Tcnologi Avanzat, Univrsità dl Pimont Orintal, Via Bllini 5/G, 5 Alssandria, Italy E-mail: chinosi@mfn.unipmn.it

More information

Mathematics. Complex Number rectangular form. Quadratic equation. Quadratic equation. Complex number Functions: sinusoids. Differentiation Integration

Mathematics. Complex Number rectangular form. Quadratic equation. Quadratic equation. Complex number Functions: sinusoids. Differentiation Integration Mathmatics Compl numbr Functions: sinusoids Sin function, cosin function Diffrntiation Intgration Quadratic quation Quadratic quations: a b c 0 Solution: b b 4ac a Eampl: 1 0 a= b=- c=1 4 1 1or 1 1 Quadratic

More information

NEW APPLICATIONS OF THE ABEL-LIOUVILLE FORMULA

NEW APPLICATIONS OF THE ABEL-LIOUVILLE FORMULA NE APPLICATIONS OF THE ABEL-LIOUVILLE FORMULA Mirca I CÎRNU Ph Dp o Mathmatics III Faculty o Applid Scincs Univrsity Polithnica o Bucharst Cirnumirca @yahoocom Abstract In a rcnt papr [] 5 th indinit intgrals

More information

Ch. 24 Molecular Reaction Dynamics 1. Collision Theory

Ch. 24 Molecular Reaction Dynamics 1. Collision Theory Ch. 4 Molcular Raction Dynamics 1. Collision Thory Lctur 16. Diffusion-Controlld Raction 3. Th Matrial Balanc Equation 4. Transition Stat Thory: Th Eyring Equation 5. Transition Stat Thory: Thrmodynamic

More information

Introduction to Condensed Matter Physics

Introduction to Condensed Matter Physics Introduction to Condnsd Mattr Physics pcific hat M.P. Vaughan Ovrviw Ovrviw of spcific hat Hat capacity Dulong-Ptit Law Einstin modl Dby modl h Hat Capacity Hat capacity h hat capacity of a systm hld at

More information

A Prey-Predator Model with an Alternative Food for the Predator, Harvesting of Both the Species and with A Gestation Period for Interaction

A Prey-Predator Model with an Alternative Food for the Predator, Harvesting of Both the Species and with A Gestation Period for Interaction Int. J. Opn Problms Compt. Math., Vol., o., Jun 008 A Pry-Prdator Modl with an Altrnativ Food for th Prdator, Harvsting of Both th Spcis and with A Gstation Priod for Intraction K. L. arayan and. CH. P.

More information

ME469A Numerical Methods for Fluid Mechanics

ME469A Numerical Methods for Fluid Mechanics ME469A Numrical Mthods for Fluid Mchanics Handout #5 Gianluca Iaccarino Finit Volum Mthods Last tim w introducd th FV mthod as a discrtization tchniqu applid to th intgral form of th govrning quations

More information

Full Waveform Inversion Using an Energy-Based Objective Function with Efficient Calculation of the Gradient

Full Waveform Inversion Using an Energy-Based Objective Function with Efficient Calculation of the Gradient Full Wavform Invrsion Using an Enrgy-Basd Objctiv Function with Efficint Calculation of th Gradint Itm yp Confrnc Papr Authors Choi, Yun Sok; Alkhalifah, ariq Ali Citation Choi Y, Alkhalifah (217) Full

More information

APPROXIMATION THEORY, II ACADEMIC PRfSS. INC. New York San Francioco London. J. T. aden

APPROXIMATION THEORY, II ACADEMIC PRfSS. INC. New York San Francioco London. J. T. aden Rprintd trom; APPROXIMATION THORY, II 1976 ACADMIC PRfSS. INC. Nw York San Francioco London \st. I IITBRID FINIT L}ffiNT }ffithods J. T. adn Som nw tchniqus for dtrmining rror stimats for so-calld hybrid

More information

A Family of Discontinuous Galerkin Finite Elements for the Reissner Mindlin plate

A Family of Discontinuous Galerkin Finite Elements for the Reissner Mindlin plate A Family of Discontinuous Galrkin Finit Elmnts for th Rissnr Mindlin plat Douglas N. Arnold Franco Brzzi L. Donatlla Marini Sptmbr 28, 2003 Abstract W dvlop a family of locking-fr lmnts for th Rissnr Mindlin

More information

The pn junction: 2 Current vs Voltage (IV) characteristics

The pn junction: 2 Current vs Voltage (IV) characteristics Th pn junction: Currnt vs Voltag (V) charactristics Considr a pn junction in quilibrium with no applid xtrnal voltag: o th V E F E F V p-typ Dpltion rgion n-typ Elctron movmnt across th junction: 1. n

More information

Sliding Mode Flow Rate Observer Design

Sliding Mode Flow Rate Observer Design Sliding Mod Flow Rat Obsrvr Dsign Song Liu and Bin Yao School of Mchanical Enginring, Purdu Univrsity, Wst Lafaytt, IN797, USA liu(byao)@purdudu Abstract Dynamic flow rat information is ndd in a lot of

More information

Basic Polyhedral theory

Basic Polyhedral theory Basic Polyhdral thory Th st P = { A b} is calld a polyhdron. Lmma 1. Eithr th systm A = b, b 0, 0 has a solution or thr is a vctorπ such that π A 0, πb < 0 Thr cass, if solution in top row dos not ist

More information

Finding low cost TSP and 2-matching solutions using certain half integer subtour vertices

Finding low cost TSP and 2-matching solutions using certain half integer subtour vertices Finding low cost TSP and 2-matching solutions using crtain half intgr subtour vrtics Sylvia Boyd and Robrt Carr Novmbr 996 Introduction Givn th complt graph K n = (V, E) on n nods with dg costs c R E,

More information

Exam 1. It is important that you clearly show your work and mark the final answer clearly, closed book, closed notes, no calculator.

Exam 1. It is important that you clearly show your work and mark the final answer clearly, closed book, closed notes, no calculator. Exam N a m : _ S O L U T I O N P U I D : I n s t r u c t i o n s : It is important that you clarly show your work and mark th final answr clarly, closd book, closd nots, no calculator. T i m : h o u r

More information

ABEL TYPE THEOREMS FOR THE WAVELET TRANSFORM THROUGH THE QUASIASYMPTOTIC BOUNDEDNESS

ABEL TYPE THEOREMS FOR THE WAVELET TRANSFORM THROUGH THE QUASIASYMPTOTIC BOUNDEDNESS Novi Sad J. Math. Vol. 45, No. 1, 2015, 201-206 ABEL TYPE THEOREMS FOR THE WAVELET TRANSFORM THROUGH THE QUASIASYMPTOTIC BOUNDEDNESS Mirjana Vuković 1 and Ivana Zubac 2 Ddicatd to Acadmician Bogoljub Stanković

More information

Estimation of apparent fraction defective: A mathematical approach

Estimation of apparent fraction defective: A mathematical approach Availabl onlin at www.plagiarsarchlibrary.com Plagia Rsarch Library Advancs in Applid Scinc Rsarch, 011, (): 84-89 ISSN: 0976-8610 CODEN (USA): AASRFC Estimation of apparnt fraction dfctiv: A mathmatical

More information

On spanning trees and cycles of multicolored point sets with few intersections

On spanning trees and cycles of multicolored point sets with few intersections On spanning trs and cycls of multicolord point sts with fw intrsctions M. Kano, C. Mrino, and J. Urrutia April, 00 Abstract Lt P 1,..., P k b a collction of disjoint point sts in R in gnral position. W

More information

Principles of Humidity Dalton s law

Principles of Humidity Dalton s law Principls of Humidity Dalton s law Air is a mixtur of diffrnt gass. Th main gas componnts ar: Gas componnt volum [%] wight [%] Nitrogn N 2 78,03 75,47 Oxygn O 2 20,99 23,20 Argon Ar 0,93 1,28 Carbon dioxid

More information

A POSTERIORI ERROR ESTIMATION FOR AN INTERIOR PENALTY TYPE METHOD EMPLOYING H(DIV) ELEMENTS FOR THE STOKES EQUATIONS

A POSTERIORI ERROR ESTIMATION FOR AN INTERIOR PENALTY TYPE METHOD EMPLOYING H(DIV) ELEMENTS FOR THE STOKES EQUATIONS A POSTERIORI ERROR ESTIMATION FOR AN INTERIOR PENALTY TYPE METHOD EMPLOYING H(DIV) ELEMENTS FOR THE STOKES EQUATIONS JUNPING WANG, YANQIU WANG, AND XIU YE Abstract. This papr stablishs a postriori rror

More information

PHYSICS 489/1489 LECTURE 7: QUANTUM ELECTRODYNAMICS

PHYSICS 489/1489 LECTURE 7: QUANTUM ELECTRODYNAMICS PHYSICS 489/489 LECTURE 7: QUANTUM ELECTRODYNAMICS REMINDER Problm st du today 700 in Box F TODAY: W invstigatd th Dirac quation it dscribs a rlativistic spin /2 particl implis th xistnc of antiparticl

More information

A LOCALLY DIVERGENCE-FREE INTERIOR PENALTY METHOD FOR TWO-DIMENSIONAL CURL-CURL PROBLEMS

A LOCALLY DIVERGENCE-FREE INTERIOR PENALTY METHOD FOR TWO-DIMENSIONAL CURL-CURL PROBLEMS A LOCALLY DIVERGENCE-FREE INTERIOR PENALTY METHOD FOR TWO-DIMENSIONAL CURL-CURL PROBLEMS SUSANNE C. BRENNER, FENGYAN LI, AND LI-YENG SUNG Abstract. An intrior pnalty mthod for crtain two-dimnsional curl-curl

More information

Analysis of a Discontinuous Finite Element Method for the Coupled Stokes and Darcy Problems

Analysis of a Discontinuous Finite Element Method for the Coupled Stokes and Darcy Problems Journal of Scintific Computing, Volums and 3, Jun 005 ( 005) DOI: 0.007/s095-004-447-3 Analysis of a Discontinuous Finit Elmnt Mthod for th Coupld Stoks and Darcy Problms Béatric Rivièr Rcivd July 8, 003;

More information

16. Electromagnetics and vector elements (draft, under construction)

16. Electromagnetics and vector elements (draft, under construction) 16. Elctromagntics (draft)... 1 16.1 Introduction... 1 16.2 Paramtric coordinats... 2 16.3 Edg Basd (Vctor) Finit Elmnts... 4 16.4 Whitny vctor lmnts... 5 16.5 Wak Form... 8 16.6 Vctor lmnt matrics...

More information