Some new additive Runge Kutta methods and their applications

Size: px
Start display at page:

Download "Some new additive Runge Kutta methods and their applications"

Transcription

1 Journal of Computatonal and Appled Mathematcs 9 (6) Some new addtve Runge Kutta methods and ther applcatons Hongyu Lu,, Jun Zou Department of Mathematcs, The Chnese Unversty of Hong Kong, Shatn, N.T., Hong Kong Receved 4 November 4; receved n revsed form 4 January 5 Dedcated to Roderck S.C. Wong on the occason of hs 6th brthday Abstract We propose some new addtve Runge Kutta methods of orders rangng from to 4 that may be used for solvng some nonlnear system of ODEs, especally for the temporal dscretzaton of some nonlnear systems of PDEs wth constrants. Only lnear ODEs or PDEs need to be solved at each tme step wth these new methods. 5 Elsever B.V. All rghts reserved. Keywords: Addtve Runge Kutta method; A-stablty; Stffly accurate; Constraned PDEs. Prelmnares In ths paper we shall explore some new addtve Runge Kutta methods that may be used for solvng some nonlnear system of ODEs, especally for the temporal dscretzaton of some nonlnear systems of PDEs wth constrants: for example, the Naver Stokes equatons n ncompressble flow, where the velocty satsfes the dvergence-free constrant, and the mean-feld magnetc nducton system n geodynamo modelng, where the magnetc feld has to satsfy the dvergence-free constrant. Only lnear ODEs or PDEs need to be solved at each tme step wth these new addtve Runge Kutta methods. Correspondng author. E-mal addresses: hylu@math.cuhk.edu.hk (H. Lu), zou@math.cuhk.edu.hk (J. Zou). Department of Mathematcs, The Chnese Unversty of Hong Kong, Shatn, N.T., Hong Kong, supported by a Drect Grant of CUHK. Substantally supported by Hong Kong RGC Grants (Projects 4343 and 448/P) /$ - see front matter 5 Elsever B.V. All rghts reserved. do:.6/j.cam.5..

2 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) We start wth the ntroducton of some basc notatons and results on the exstng Runge Kutta (RK) methods. As usual, we shall represent a standard s-stage RK method (cf. [7]) by the tableau () where A = (a j ) s,j= s the coeffcent matrx, bt = (b ) s = the weght vector, and c = (c ) s = a vector used to specfy the dscrete tmes. We shall use h> to denote the tme stepsze, and t n = t + nh, n =,,,..., for the dscrete tme ponts. When appled to the followng system of frst-order ODEs, y (t) = f(t,y) for t>t ; y(t ) = y, () where y R m and f : (t, ) R m R m s a nonlnear vector-valued functon, scheme () can be wrtten as y (n) = y n + h a j f(t n + c j h, y (n) j ), =,,...,s, (3) y n = y n + h j= = b f(t n + c h, y (n) ), (4) where y n s an approxmaton of y(t n ) and y (n) y(t n + c h). Wthout loss of generalty, t suffces for us to consder only one tme step of scheme (3) (4). So we shall set n = and wrte the ntermedate stage vectors y (n) as Y y(t + c h) and y y(t ). Then scheme (3) (4) can be expressed as Y = y + h a j f(t + c j h, Y j ), =,,...,s, (5) y = y + h j= b f(t + c h, Y ). (6) = In general, the parameters c s n (5) (6) are requred to satsfy the condtons c = a j, =,...,s, j= n order to essentally smplfy order condtons, especally for hgh-order methods. These condtons ndcate that the local truncaton error of each approxmaton n (5) s at least frst-order accurate. Stablty wll be one of our central ssues to be consdered when we construct the new RK methods. Recall that the stablty functon R(z) wth z = λh of an RK method s the approxmate soluton generated by one step of the method for the Dahlqust test problem: y (t) = λy, y =, (7)

3 76 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) wth Re{λ} <. Then the stablty regon of the method s defned to be S ={z C; R(z) }. (8) We know that the stablty functon of the mplct RK method () s gven (cf. [7]) by R(z) = det(i za + zbt ). (9) det(i za) Recall that a method s sad to be A(α)-stable f the sector S α ={z; arg( z) α, z = } s contaned n the stablty regon. An A( π )-stable method s called A-stable. It can be seen that an RK method s A-stable f and only f R(y) for all real y, R(z) s analytc for Re{z} <. () Clearly, an A-stable RK method must be mplct. For very stff problems, one may need schemes wth L-stablty, that s, schemes whch are A-stable and lm R(z) =. () z In many stff stuatons, the L-stablty condton () may be too strong. In fact, t s often suffcent n those very stff cases to requre that lm z R(z) equals some constant less than, for example, α < ; ths wll be used n our subsequent constructon of the addtve methods.. Addtve Runge Kutta methods The focus of our study s on the numercal soluton of the followng ODEs of addtve form: y (t) = f(t,y)+ g(t, y), () where y R m, f : (t, ) R m R m s stff, but lnear wth respect to y, and g(t, y) s nonlnear but not stff. We remark that our subsequent constructon technques can be equally appled to construct smlar schemes for the case where f(t,y)s lnear but non-stff, whle g(t, y) s nonlnear but stff. As we wll show n Secton, system () can also be PDEs wth constrants, where f(t,y) and g(t, y) are both dfferental operators of space varables. In fact, the PDEs wth constrants are the major systems our new addtve RK methods ntend to solve. The addtve methods use the dea of the RK method but am to provde a more effectve way to deal wth the ODEs of the addtve form (). One can already fnd some addtve RK methods n the lterature, whch can acheve favorable results n numercal solutons of certan stff addtve ODEs lke () (see [,5,6,9,]). But all the exstng addtve schemes are not sutable to the PDEs wth constrants, or are applcable but very expensve. When appled to Eq. (), an s-stage addtve RK method s a scheme of the form: y (n) = y (n ) s + h j= a j f(t n + c j h, y (n) j ) + h j= b j g(t n + c j h, y (n) j ), (3)

4 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) where =,,...,s and n =,, 3,...The followng row condtons are always assumed c = a j = b j, (4) j= j= n order to smplfy the order condtons so that t s possble to fnd some reasonable hgh-order schemes. By puttng f(t,y)= org(t, y) = n (), we can get from (3) two standard s-stage RK methods of form () wth the weghted values b = a s or b s. Hence, they are stffly accurate n parlance of [] and one can convenently represent an s-stage addtve RK method by the trple (c, A, B) or the followng tableau (wth weghts as n () no longer necessary): Due to the couplng between blocks A and B n (5), the order condtons and the stablty analyss for the addtve RK methods are much more complcated than the standard (). (5) 3. Stablty analyss The general addtve RK methods were frst ntroduced by Cooper and Sayfy (cf. [5,6]) for solvng the stff problem y (t) = F(t,y) for t>t ; y(t ) = y. (6) They assocated the addtve methods wth a sequence of decompostons of F(t,y)of the form F(t,y)= J (n) y + g (n) (t, y), n =,, 3,..., (7) where {J (n) } s chosen to be ndependent of t and often as an approxmaton to the Jacoban of F evaluated at some sequence of computed values. To study the stablty of the addtve methods, let y be the partcular soluton of the system y (t) = Jy + g(t, y), t > t, (8) whch has the ntal value y(t ) = y. It s known that f the trval soluton of y = Jy s exponentally stable,.e., Re{λ} < for all λ σ[j ], and that g(t, y) =o( y ), then the trval soluton of (8) s also exponentally stable, whch mples that there s an ε > such that f y ε, then y(t) has lmt zero. Usng ths model problem, Cooper and Sayfy establshed the stablty for ther addtve schemes assocated wth system (8). Theorem. Suppose that J s a constant m m matrx. The trval soluton of y (t) = Jy s exponentally stable and g(t, y) =o( y ). Furthermore, we assume that the (c, A, B) RK method s lnearly

5 78 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) mplct,.e., y (n) = y (n ) s + h j= a j Jy (n) j + h j= b j g(t n + hc j,y (n) j ), (9) and the scheme (c, A) s A-stable. Then for any fxed postve h and arbtrary y s (), method (9) unquely defnes a sequence y s (n) for =,,...,sand n=,, 3..., and there exsts a δ > such that f y s () δ, the sequence {y s (n) } has lmt zero. We are nterested n the more general addtve system (), nstead of system (6) wth specal Jacobantype decompostons (7). Clearly the results n Theorem cannot be used for the stablty estmates of the addtve methods (5), especally the smallness condton,.e., g(t, y) =o( y ), whch plays a key role n the proof of the above theorem, s usually not true to system () of our nterest. For our purpose, we shall take a smlar approach to the one n [] to drectly apply the lnear stablty analyss for standard RK methods to the followng test problem assocated wth (): dy dt = λ f y + λ g y, y =, () where λ f and λ g represent the egenvalues of f/ y and g/ y n (). They are complex parameters satsfyng 3 Re{λ f }, Re{λ g } and Re{λ f }? Re{λ g }. We emphasze that the addtve RK schemes proposed n [] are dfferent from the RK schemes of form (3) to be studed n the current paper, and cannot be used for those PDE systems wth constrants, see Secton. By applyng the addtve method (5) to the test problem (), we get after one tme step that y = R(z f,z g ), z f = hλ f, z g = hλ g, () where R(z f,z g ) s the stablty functon. Then we ntroduce our new defnton of the A(α)- and L-stablty for the addtve methods. Defnton. An A(α)-stablty doman of an addtve method (5) n the complex plane of z g = hλ g s defned as the doman where Sα f g ={z g C; R(z f,z g ) for all z f Sα f }, () S f α ={z f C; arg( z f ) α,z f = } {}. (3) An addtve method (5) s sad to be L-stable f t s A-stable (.e. A( π )-stable) and lm R(z f,z g ) =. (4) z f 3 Our subsequent analyss can be extended equally to the case Re{λ f } > Re{λg}.

6 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) If we denote by S g the stablty doman of the explct part (c, B) n (5) as a standard RK method, we have S f g α S g. (5) Ths can be seen by takng λ f = n the test (), and means that n the stablty doman Sα f g, the explct part (c, B) s stable. In the forthcomng dscussons, snce the term f(t,y)n () s stff, t s natural for us to requre that the sem-explct part (c, A) s to be A(α)-stable, and we should have Sα f g. In case the frst part (c, A) s A-stable as a standard RK method, then the L-stablty of the addtve method (5) mples that ts frst part (c, A) s stll L-stable and ths can be verfed drectly. 4. Constructon of the addtve RK methods Our constructon of addtve RK methods s based on the satsfacton of both stablty and accuracy condtons. Correspondng to (5), we ntroduce the followng notatons for =,,..., s,σ=,, 3,... a (σ) = c σ σ j= a j c σ j b (σ) = c σ σ j= b j c σ j. (6) Then the Taylor expanson gves the order condtons for the general method (5) up to fourth order as follows (see [5]). An addtve RK method (3) s of order p 4 f and only f the condtons a () = b () =, =,,...,s, (7) a s (σ) =, b s (σ) =, σ p, (8) = = a s c τ a (σ) =, b s c τ a (σ) =, = = and the followng extra condtons for p = 4, a s = j= a j a j () =, a s c τ b (σ) =, σ + τ p, (9) b s c τ b (σ) =, σ + τ p (3) a s = j= a j b j () =, (3) a s = j= b s = j= b j a j () =, a j a j () =, a s = j= b s = j= b j b j () =, (3) a j b j () =, (33)

7 8 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) b s = j= b j a j () =, b s = j= are satsfed, where σ and τ take all possble postve nteger values. b j b j () =, (34) 5. Sem-mplct addtve RK methods As we can easly see, the addtve RK schemes of the general form (5) are fully mplct, and wll be very expensve when appled for solvng ODEs or for the tme marchng schemes of PDEs. Recall that our target system s of form (), where the lnear part f(t,y) s stff but the nonlnear part g(t, y) s not stff. It wll be more practcal to have schemes that are only mplct n terms of f(t,y)and explct n terms of g(t, y) as ths needs to solve only lnear systems at each ntermedate tme step. To further reduce the computatonal complexty, we would lke to have schemes that are only sem-mplct n terms of f(t,y), that s, the frst coeffcent matrx A s lower trangular whle the second coeffcent matrx B s strctly lower trangular n (5). Besdes, n order to take nto account certan constrant equatons n the PDEs (see Secton ), we requre that the computed value at each ntermedate step,.e., y (n) n (3), s gven mplctly and ths can be satsfed f a = for =, 3,...,s. On the other hand, t s qute necessary for every ntermedate stage y (n) of the method to gve physcal meanngful computaton, that s, y (n) should be the approxmated value of y(t) at tme t = t n + c h. Naturally, we shall take c s = and y s (n) as the fnal computed value of each tme step. In summary, the new addtve RK methods we are lookng for can now be descrbed by the followng tableau: Before startng our constructon, we shall dscuss a bt about the dfference between our new schemes and the exstng ones. The most mportant exstng addtve schemes are the ones developed n [6], whch take a ss = and can therefore be vewed as the combnaton of an (s )-stage dagonally mplct RK (DIRK) method (wth weghts b =a s,=,,...,s ) and an (s )-stage explct RK method (wth weghts b = b s, =,,...,s ). For the mplct part of those schemes, the dagonal entres take as many zeros as possble n order to make the derved scheme more effcent and ease the constructon smultaneously. On the contrary, our new schemes are requred to satsfy the condtons that a = for =, 3,...,s due to the applcaton problems n our mnd (see Secton ). If ths s dffcult, at least we should have a ss =. Now, n the sem-mplct addtve RK method (35), the (c, A) method s n fact an s-stage stffly accurate DIRK method (wth weghts b = a s, =,,...,s) and the (c, B) method s reduced to an (s )-stage explct RK method (wth weghts b = b s,=,,...,s ). Thus, our new addtve RK schemes can be vewed as the combnaton of an s-stage stffly accurate DIRK method and an (s )-stage explct RK method. To our best knowledge, there are no such addtve methods n the lterature. In fact, the constructon of such schemes are not straghtforward at all and we cannot see (35)

8 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) any drect applcaton of the exstng technques for the constructon of explct RK and DIRK methods to acheve our purpose here. Theorem below wll play the key role n our subsequent constructon of the schemes whch meet our needs. We are now gong to construct the addtve schemes of form (35) wth a = for =, 3,...,s.In order to deal wth the stffness of the lnear part f(t,y), we shall requre that the part (c, A) n (5) as a standard RK method s A(α)-stable. The algebrac condtons for a sem-explct RK method (c, A), where the frst dagonal element of A s zero, to be A-stable can be gven n terms of two sets of parameters {α } = and {β } = (see [6,4]). {β } = are defned by s ( τa rr ) = β τβ + τ β. r= Ths gves β = and β s = β s+ = =. Let e, e,...,e s be the natural bass for R s and let e = e + e + +e s be the vector wth unt elements. The terms e T s Ar e, r =,, 3,..., are the sums of the elements n rows s of A, A, A 3,...and for a method of order p t has that e T s Ar e = r!, r =,,...,p. Defne α s = α s+ = = and α r = β r β r e T s Ae + +( )r β e T s Ar e, r =,,...,s, so that α =. Then a method of order p s A-stable ff a rr for r =,,...,s and s y r r=π r ( ) r+j (β r j β j α r j α j ) j= y, where π s the ntegral part of p/ + and the astersk denotes that the terms j = r are halved. Snce n the sem-mplct addtve RK method (35), the (c, B) method s n fact reduced to be an (s )-stage standard explct RK method, the order of such methods cannot exceed s ; f s>5, the order cannot exceed s, see [7]. That s, f we want to construct a fourth-order scheme of ths type, t s at least of 5-stage. By observng Eqs. (6) (34), one can readly see that the order condtons turn out to be too tedous to deal wth. Together wth the A-stable condtons, the constructon becomes extremely dffcult. However, t wll be much easer f we have a s = b s n (35) for =,,...,s, as mplemented n [6], snce the number of the order condtons s almost halved as one can see from (6) (34). But ths conflcts wth our requrement that a ss = and b ss =. Fortunately, we have the followng soluton to overcome ths dlemma. Theorem. Assume that the s-stage addtve RK method (5) s of order p, c s = c s =, and the (s )th ntermedate stage s an approxmaton to y(t) of order p,.e., y (n) s =y(t +nh)+o(h p ). Then the method obtaned by replacng b s,s and b ss wth (b s,s + b ss ) and, respectvely, s stll of order p.

9 8 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) For the proof of ths theorem, we need the followng lemma: Lemma. Suppose Y, Z R M are two vectors and satsfy Z Y h{ A[f(Z) f(y)] + B[g(Z) g(y)] }+O(h p ), (36) where A, B R M M are two constant matrces, p s a postve nteger and f, g : R M R M are often dfferentable and have bounded frst-order dervatves. Then we have Z Y O(h p ). (37) Proof. By Eq. (36), we know that there must exst some constant C such that Z Y Ch Z Y + O(h Z Y ) + O(hp ). (38) Agan, by Eq. (36), we easly see that Z Y O(h). Usng ths and Eq. (38), we have Z Y O(h ). Repeatng ths process, we can derve (37). Proof of Theorem. Settng y (n) f(t n + c h, y Y (n) =. (n), F(Y (n) ) g(t n + c h, y (n) ) =., G(Y (n) ) ) =., y s (n) f(t n + c s h, y s (n) ) g(t n + c s h, y s (n) ) we can rewrte (3) as Y (n) = Y (n ) + h(a I)F(Y (n) ) + h(b I)G(Y (n) ), (39) where Y (k) s a block column vector consstng of s, m-dmenson column vector y s (k), A and B are the coeffcent matrces of the frst and second part of method (5), respectvely, I s the m m dentty matrx, and denotes the Kronecker product. Next, replacng b s,s and b ss wth b s,s + b ss and, respectvely, the obtaned new method for Eq. () s Z (n) = Z (n ) + h(a I)F(Z (n) ) + h( B I)G(Z (n) ), (4) where Z (n), Z (n) have the same usages as Y (n), Y (n), and B s the same as B wth only ts (s, s )- and (s, s)-elements beng b s,s + b s,s and, nstead of b s,s and b s,s of B. By subtracton of Eq. (39) and (4), we obtan Z (n) Y (n) = (Z (n ) Y (n ) ) + h{(a I)[F(Z (n) ) F(Y (n) )] + ( B I)[G(Z (n) ) G(Y (n) )]} + h[( B B) I]G(Y (n) ). (4)

10 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) In order to prove that the new method s also of order p, t s suffcent to show that the local truncaton error of z s (n) s O(h p+ ). Thus, we take n = and z s () = y s () = y. By notng that y (n) s s an approxmaton of y(t) of order p and y s (n) s of order p,wesee [( B B) I]G(Y () ) = b s,s [G(y () s ) G(y() s )] C y () s y() s O(h p ), (4) where C = b s,s L and L s the bound for g, and we have made use of the fact that the local truncaton errors of y () s and y() s are, respectvely, O(h p ) and O(h p+ ), by the assumpton of the theorem,.e., y () s y(t + h) =O(h p ), y s () y(t + h) =O(h p+ ). Hence, we derve by (4) Z () Y () h{ (A I)[F(Z () ) F(Y () )] + (B I)[G(Z () ) G(Y () )] }+O(h p+ ). (43) Ths mples, by Lemma, Z () Y () O(h p+ ). (44) Then we can deduce z s () y(t + h) z s () y s () + y s () y(t + h) Z () Y () + y s () y(t + h) O(h p+ ). (45) The proof s completed. We know from Theorem that for those pth-order addtve RK methods of form (35), where the frst (s ) stages form an (s )-stage addtve RK method,.e., a s = b s = for =,...,s, Theorem says f the frst (s )-stage s of order p, then by replacng b s,s and b ss by (b s,s + b ss ) and, respectvely, the resultng method s stll of order p. Ths provdes an mportant prncple, whch we can use to construct the desred addtve RK methods: To construct an addtve method of form (35) wth a ss = butb ss =, we can frst construct a pth order method that satsfes a s = b s ( =,,...,s), a ss = b ss =, c s = c s = and the frst (s )-stage s a (p )th order lnearly mplct addtve RK method. Then, replacng b s,s and b ss wth (b s,s + b ss ) and, respectvely, we wll acheve a new method of order p, whch s lnearly mplct of form (35) and satsfes that a ss = and b ss =. All our subsequent constructons of the new addtve RK methods are based on the above prncple. Wth ths prncple and the order condtons n Secton 4, we can construct addtve RK methods wth orders rangng from to 4. We shall gnore all the tedous dervatons but present many dfferent examples of such schemes n the subsequent sectons. For most of the schemes, the A( π )-stablty regon of the addtve RK method wll be plotted aganst the stablty regon of the correspondng explct part (c, B) as a standard explct RK method of form () for comparsons. Some numercal experments wth these new RK methods wll be presented to demonstrate ther accuraces, stabltes and effcences. In the sequel, we shall use the descrptons lke addtve RK.m.A.n or RK.m.L.n for some postve ntegers m and n. addtve RK.m.A(L).n ndcates an addtve RK method, whch s mth-order accurate and A-stable (or L-stable). The number n means t s the nth method of ths class lsted n ths paper.

11 84 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) Examples of addtve RK methods wth nonzero dagonal entres Ths secton presents some new addtve RK methods wth nonzero dagonal entres a for =, 3,...,s. The followng tableau gves a new class of addtve RK methods of order whch s A-stable: where <c<, α, β are postve and satsfy α β [αβ (α + β) + ]. If we take α = β β, β > orβ <, then the resultng method s L-stable. We frst take c = /, α = β = and get the followng A-stable one: Its stablty functon R(z f,z g ) as defned n () s R(z f,z g ) = ( z f z f ) + ( z f )z g + z g z f + z f and ts A( π )-stablty doman s plotted n Fg. (left). One can observe that the A( π )-stablty doman of the addtve method,.e. S f g π/, s slghtly smaller than that of the explct part as a conventonal RK method,.e. S g. For addtve stff problem (), f we only use the explct part, whch s a second-order

12 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) Fg.. Stablty doman of addtve (sold lne) and explct RK (dot) methods Fg.. Stablty doman of addtve (sold lne) and explct RK (dot) method. standard RK method, the step length should be chosen to satsfy (λ f + λ g )h S g, where λ f σ{ f y } and λ g σ{ g y }, n order to meet the stablty requrement. But as dscussed earler, the addtve method needs only to satsfy that λ g h S f g π/. Snce Re{λ f } > Re{λ g }, eventhough the stablty doman of the addtve method s smaller than that of ts explct part as n the conventonal RK method, t allows a relatvely much larger range to choose the step length h. Moreover, consderng the effcency, we can see that such an addtve method s defntely superor to the pure mplct method as t does not need to solve any nonlnear equatons. So the addtve methods of ths new type present some advantages over the exstng RK methods for solvng the addtve system (). Below are two more examples wth A-stablty by takng α=/, β=/, c =/ and α=/, β=/, c = /4, respectvely: These two methods have the same stablty functon R(z f,z g ), R(z f,z g ) = ( 4 z f ) + z g + z g z f + 4 z f

13 86 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) and ther A( π )-stablty doman s shown n Fg. (rght). One can see that these two schemes have a relatvely larger stablty regon (n fact, they almost concde wth that of the explct RK method), and thus should be more preferable n solvng those not too stff problems. The followng are two L-stable schemes: The stablty functons of these two schemes are, respectvely, R(z f,z g ) = [ + ( )z f ]+[ + ( )z f ]z g + /zg ( )z f + (3/ )zf, R(z f,z g ) = ( + 7/4z f ) + ( + 7/4z f )z g + /z g 3/4z f + 3/4z f and ther A( π )-stablty domans are gven n Fg. (left for RK..L. and rght for RK..L.). Next, we provde some thrd-order methods that should be at least of 4-stage as we ponted out earler. But one can show that there exsts a 4-stage lnearly mplct RK method of order 3 f and only f a 44 = (the proof s omtted here). Ths does not meet our requrement that a ss = for an s-stage method. So we can only try to fnd some 5-stage thrd-order methods. Ths s possble, and some A-stable schemes are gven below:

14 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) wth ts A( π )-stablty doman as shown n Fg. 3 (left). wth ts stablty regon as shown n Fg. 3 (rght). wth ts A( π )-stablty doman gven n Fg. 4 (left). Below s one example of L-stable schemes, wth ts A( π )-stablty doman gven n Fg. 4 (rght). Fnally, we present some fourth-order methods.a few 6-stage schemes wth a 44 = are gven n Secton 7, whch are applcable n some specal cases. Here, we lst two 7-stage A-stable methods.

15 88 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) Fg. 3. Stablty doman of addtve (sold lne) and explct RK (dot) method Fg. 4. Stablty doman of addtve (sold lne) and explct RK (dot) method. wth ts A( π )-stablty doman plotted n Fg. 5 (left). Its A( π )-stablty doman s plotted n Fg. 5 (rght). 7. Numercal experments In ths secton we carry out some numercal experments to attest the stablty and accuracy of the addtve RK methods constructed n Secton 6. Some notatons are needed. We shall use E(h) to denote the dscrete L -norm error between the computed soluton y h (t) by an addtve RK method and the exact soluton y(t) to system (), whle r(h)

16 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) Fg. 5. Stablty doman of addtve (sold lne) and explct RK (dot) method. measures the asymptotc convergence rate. The two parameters are gven by (wth h >h ) E(h) = h y(t ) y h (t ), r(h) = ln E(h / ) ln h. E(h ) h In our tests of numercal schemes, we wll use the usual practce to compute the numercal solutons n the transent phase where the stff terms f(t,y)n () contrbute to the solutons of the consdered stff problems and outsde ths phase the stff terms de out. The numercal solutons are computed wth a relatvely hgh-order explct method usng much smaller tme steps for accuracy n the transent phase. In all the tests, we take the fourth-order explct RK methods for computatons n the transent phase. Example. The frst model problem s taken to be y = Ay + g(y), y() =[,, ] T, (46) where y =[y,y,y 3 ] T, and A and g(y) are gven by [ ] 9 a A = 9, g(y) = b y [ y ] y, y 3 where two parameters a and b are delberately ntroduced to test dfferent stuatons. Snce σ(a) = {, 4 + 4, 4 4}, the problem s of type (). We shall test the cases wth dfferent a s and b s. Frst, we take a =, b= and, respectvely, to check the stablty of both RK..A. and RK.3.A.3, wth the tme range taken to be [, 5]. () b =, a =. The exact soluton n ths case s y (t) = /e 5t (cos 4t + sn 4t) + /e t, y (t) = /e 5t (cos 4t + sn 4t) + /e t, y 3 (t) = e 5t (sn 4t cos 4t). Fg. 6 s the numercal result of RK..A. to ths case wth h =.3,.,..

17 9 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) x y 3 y y y y y y y y Fg. 6. Numercal results of RK..A. wth h =.3,.,. n turn..8.6 y(t) y(t) y3(t) Fg. 7. True soluton of Eq. (46) (a =,b= ) computed wth the fourth-order RK method. Table Egenvalues of g/ y at dfferent tme ponts n the case a = and b = t λ g We observe that f the step length h gets larger, the numercal solutons vbrate more strongly and even dverge. If only the step length s chosen to satsfy that ah (here a s the λ g n ()) falls nto the A( π )-stablty doman of the method used, t s safe to have the numercal soluton to be damped. () b =, a =, where the theoretcal soluton s computed usng the 4-stage and fourth-order explct RK method (see [8]), whch s plotted n Fg. 7. Below we lst the egenvalues of g/ y at dfferent tme ponts (Table ):

18 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) y.6 y y y y y y y y Fg. 8. Numercal results of RK.3.A.3 wth h =.3,.,. n turn. Table Egenvalues of g/ y at dfferent tme ponts n the case a = and b = t λ g Table 3 Convergence order of the method RK.3.A. h E(h).4e 5.49e 6.93e 7.459e 8 3.3e 9 r(h) Table 4 Convergence order of the method RK.3.A.3 h E(h).7e 5.884e e e e 9 r(h) Fg. 8 shows the numercal behavors of RK.3.A.3 wth h =.3,.,., respectvely. We can also see that f λ g h falls nto the A( π )-stablty doman of the method, t s safe to have the decay result. () b =,a=, the theoretcal soluton computed as n the prevous case. The egenvalues of g/ y at some tme ponts are lsted n Table. Now, wth ths model, the convergence order of some schemes can be computed as n the followng tables. Table 3 s for RK.3.A.. Tables 4 and 5 are for RK.3.A.3 and RK..A., respectvely. To accurately fnd the exact convergence order of the fourth-order scheme RK.4.A., we construct the followng model: y = λy + αy, y() = (47)

19 9 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) Table 5 Convergence order of the method RK..A. h E(h).48e e 6.494e e 7 9.3e 8 r(h).... Table 6 Convergence order of the method RK.4.A. h E(h) 3.693e e 9.478e.76e.96e 4.47e 5 r(h) Table 7 Convergence order of the method RK.4.A. h E(h).e 8.9e e 5.e 9.9e 5 5.8e 6 r(h) whch has the exact soluton λe λt y(t) = α( e λt ) + λ. (48) In ths, we take λ =, α = and snce the stffness s brought n the consderaton of the soluton at the transent phase (whch we may regard as [,]) s taken as the true soluton, whch s equvalent to takng t =, and the computaton s carred out as n [,6] and we arrve at results shown n Table 6. The convergence of RK.4.A. s lsted n Table 7. From all the prevous tables, one can see that the numercal experments have clearly verfed the actual orders of the correspondng addtve RK methods. 8. Methods wth zero dagonal elements In ths secton, we wll gve a few addtve methods of form (35), whch allow some dagonal elements a to be zero, but the last entry a ss =. The followng scheme s the frst one lsted n Secton 6 wth α =, β = c = :

20 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) Ths method can be vewed as a drect combnaton of the Crank Ncolson method and the modfed md-pont formula [], and ts stablty functon as defned n () s R(z f,z g ) = + /z f + ( + /z f )z g + /z g /z f. The A( π )-stablty doman of the addtve method and the stablty doman of ts explct part (c, B) as a standard RK method are shown n Fg. 9 (left). In the remanng part of ths secton, we present some more such addtve methods wth order 3 and 4. where a,b,c,d are 4 real parameters satsfyng a>, b = 3a 6a 3, b + 4a + 8ab 5 3 = ( 6b)c. Takng a =,b = /3,c = 3,d = and a = /3,b=,c= 5/3,d =, we get the followng two schemes: These two methods have the same stablty functon and ther A(π/)-stablty doman s shown n Fg. 9 (rght).

21 94 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) Fg. 9. Stablty doman of addtve (sold lne) and explct RK (dot) method Fg.. Stablty doman of addtve (sold lne) and explct RK (dot) method. and ts A( π )-stablty doman s plotted n Fg. (left). Its A( π )-stablty doman s shown n Fg. (rght).

22 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) Table 8 Convergence order of the method RK.4.nA.5 h E(h).8e 5.44e 6 7.5e e 9.67e 7.68e r(h) Table 9 Convergence order of the method RK.4.nA.6 h E(h) 7.43e e 6.63e 7.6e 8 4.e.56e r(h) Some not A-stable fourth-order methods In ths secton, we present some fourth-order methods that are not A-stable, whch may be used for the nonlnear ODEs or PDEs of addtve form () that are not stff. The followng tables verfy the fourth-order convergence of addtve RK.4.nA.5 (Table 8) and addtve RK.4.nA.6 (Table 9) when appled to the model problem (47) n the non-stff case wth λ =. The followng are two more such schemes of fourth-order that are not A-stable:

23 96 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) Some applcatons In ths secton we dscuss brefly how to apply our new addtve RK methods to solve two mportant PDE systems wth constrants... Knematc magnetc nducton system In the knematc geodynamo modelng, the followng mean-feld dynamo magnetc nducton system s wdely used, B = (β(x) B) + f(x, t; u, B) n Ω (, T), (49) t B = n Ω (,T), (5) wth approprate ntal and boundary condtons [,3], where B s the magnetc feld of the dynamo system, β(x) s the magnetc Reynolds number, and f(, x, t; u, B) s a nonlnear vector-valued functon of flow u and magnetc feld B, whch determnes the key dynamo system so that the magnetc feld can sustan n the correspondng physcal system. Now, as an example, we apply the addtve method RK..A. to the temporal sem-dscretzaton of ths system and obtan the followng scheme advancng from tme t n to t n+ : B n = B n h (β(x) B n ) + h f(x, t n; u n, B n ), (5) B n =, (5)

24 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) B n+ = B n h [β(x) (B n + B n+ )]+h f(x, t n+/ ; u n+/, B n ), (53) B n+ =, (54) where B n B(t n ), B n B(t n+/), B n+ B(t n+ ). Clearly, f explct schemes are used, then the dvergence constrants cannot be enforced for both stage values B n and B n+. When mplct schemes as above are used, systems (5) (5) and (53) (54) for the two stage values B n and B n+ are well-posed [,3]... Naver Stokes system The Naver Stokes system s a general model for ncompressble flud dynamcs: u = ν Δu p (u )u + f n Ω (,T), (55) t u = n Ω (,T), (56) where u s the flud velocty, p s the pressure and f s an external source, and ν s the vscosty of the flud. Applyng the addtve method RK..L. to ths system, we obtan the followng scheme: where u n = u n + h (νδu n p n ) + h 5 (νδu n p n ) + h 4 { (u n )u n + f n }, (57) un =, (58) u n+ = u n + h 8 (νδu n p n ) + h (νδu n p n ) + 3h 8 (νδu n+ p n+ ) h{ (u n )u n + f n }+h{ (u n )u n + f t n +/4}, (59) u n+ =, (6) (u n,p n ) (u(t n ), p(t n )), (u n,p n ) (u(t n+/4), p(t n+/4 ). Clearly, f explct schemes are used, then the dvergence constrants cannot be enforced for both stage values u n and u n+. When mplct schemes as above are used, systems (57) (58) and (59) (6) for the two stage values (u n,p n ) and (u n+,p n+ ) are well posed. Acknowledgements The authors wsh to thank Prof. Geng Sun for some helpful dscussons and two anonymous referees for ther constructve comments whch led to a great mprovement of the results and the presentaton of the paper.

25 98 H. Lu, J. Zou / Journal of Computatonal and Appled Mathematcs 9 (6) References [] A.L. Araújo, A. Murua, J.M. Sanz-Serna, Symplectc methods based on decompostons, SIAM. J. Numer. Anal. 34 (997) [] K. Chan, K. Zhang, J. Zou, G. Schubert, A nonlnear vacllatng dynamo nduced by an electrcally heterogeneous mantle, Geophys. Res. Lett. 8 () [3] K. Chan, K. Zhang, J. Zou, G. Schubert, A nonlnear 3-D, sphercal alpha-square dynamo usng a fnte element method, Phys. Earth Planet. Int. 8 () [4] G.J. Cooper, A. Sayfy, Semexplct A-stable Runge Kutta methods, Math. Comp. 33 (979) [5] G.J. Cooper, A. Sayfy, Addtve methods for the numercal soluton of ordnary dfferental equatons, Math. Comp. 35 (98) [6] G.J. Cooper, A. Sayfy, Addtve Runge Kutta methods for stff ordnary dfferental equatons, Math. Comp. 4 (983) 7 8. [7] E. Harer, G. Wanner, Solvng Ordnary Dfferental Equatons, II, Stff and Dfferental-Algebrac Problems, Sprnger, Berln, Hedelberg, 99. [8] J.D. Lambert, Computatonal Methods n Ordnary Dfferental Equatons, Wley, New York, 973. [9] O.J. Laurent, Structure preservaton for constraned dynamcs wth super addtve Runge Kutta methods, SIAM J. Sc. Comput. (998) [] H. Lu, G. Sun, Sympletc RK methods and symplectc PRK methods wth real egenvalues, J. Comput. Math. (5) (4) [] A. Prothero, A. Robnson, On the stablty and accuracy of one-step methods for solvng stff systems of ordnary dfferental equatons, Math. Comput. 8 (974) [] X. Zhong, Addtve Sem-mplct Runge Kutta methods for computng hgh-speed nonequlbrum reactve flows, J. Comp. Phys. 8 (996) 9 3.

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems Numercal Analyss by Dr. Anta Pal Assstant Professor Department of Mathematcs Natonal Insttute of Technology Durgapur Durgapur-713209 emal: anta.bue@gmal.com 1 . Chapter 5 Soluton of System of Lnear Equatons

More information

Appendix B. The Finite Difference Scheme

Appendix B. The Finite Difference Scheme 140 APPENDIXES Appendx B. The Fnte Dfference Scheme In ths appendx we present numercal technques whch are used to approxmate solutons of system 3.1 3.3. A comprehensve treatment of theoretcal and mplementaton

More information

Numerical Heat and Mass Transfer

Numerical Heat and Mass Transfer Master degree n Mechancal Engneerng Numercal Heat and Mass Transfer 06-Fnte-Dfference Method (One-dmensonal, steady state heat conducton) Fausto Arpno f.arpno@uncas.t Introducton Why we use models and

More information

Lectures - Week 4 Matrix norms, Conditioning, Vector Spaces, Linear Independence, Spanning sets and Basis, Null space and Range of a Matrix

Lectures - Week 4 Matrix norms, Conditioning, Vector Spaces, Linear Independence, Spanning sets and Basis, Null space and Range of a Matrix Lectures - Week 4 Matrx norms, Condtonng, Vector Spaces, Lnear Independence, Spannng sets and Bass, Null space and Range of a Matrx Matrx Norms Now we turn to assocatng a number to each matrx. We could

More information

CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE

CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE Analytcal soluton s usually not possble when exctaton vares arbtrarly wth tme or f the system s nonlnear. Such problems can be solved by numercal tmesteppng

More information

Difference Equations

Difference Equations Dfference Equatons c Jan Vrbk 1 Bascs Suppose a sequence of numbers, say a 0,a 1,a,a 3,... s defned by a certan general relatonshp between, say, three consecutve values of the sequence, e.g. a + +3a +1

More information

Lecture 12: Discrete Laplacian

Lecture 12: Discrete Laplacian Lecture 12: Dscrete Laplacan Scrbe: Tanye Lu Our goal s to come up wth a dscrete verson of Laplacan operator for trangulated surfaces, so that we can use t n practce to solve related problems We are mostly

More information

One-sided finite-difference approximations suitable for use with Richardson extrapolation

One-sided finite-difference approximations suitable for use with Richardson extrapolation Journal of Computatonal Physcs 219 (2006) 13 20 Short note One-sded fnte-dfference approxmatons sutable for use wth Rchardson extrapolaton Kumar Rahul, S.N. Bhattacharyya * Department of Mechancal Engneerng,

More information

Lecture 21: Numerical methods for pricing American type derivatives

Lecture 21: Numerical methods for pricing American type derivatives Lecture 21: Numercal methods for prcng Amercan type dervatves Xaoguang Wang STAT 598W Aprl 10th, 2014 (STAT 598W) Lecture 21 1 / 26 Outlne 1 Fnte Dfference Method Explct Method Penalty Method (STAT 598W)

More information

NON-CENTRAL 7-POINT FORMULA IN THE METHOD OF LINES FOR PARABOLIC AND BURGERS' EQUATIONS

NON-CENTRAL 7-POINT FORMULA IN THE METHOD OF LINES FOR PARABOLIC AND BURGERS' EQUATIONS IJRRAS 8 (3 September 011 www.arpapress.com/volumes/vol8issue3/ijrras_8_3_08.pdf NON-CENTRAL 7-POINT FORMULA IN THE METHOD OF LINES FOR PARABOLIC AND BURGERS' EQUATIONS H.O. Bakodah Dept. of Mathematc

More information

n α j x j = 0 j=1 has a nontrivial solution. Here A is the n k matrix whose jth column is the vector for all t j=0

n α j x j = 0 j=1 has a nontrivial solution. Here A is the n k matrix whose jth column is the vector for all t j=0 MODULE 2 Topcs: Lnear ndependence, bass and dmenson We have seen that f n a set of vectors one vector s a lnear combnaton of the remanng vectors n the set then the span of the set s unchanged f that vector

More information

Transfer Functions. Convenient representation of a linear, dynamic model. A transfer function (TF) relates one input and one output: ( ) system

Transfer Functions. Convenient representation of a linear, dynamic model. A transfer function (TF) relates one input and one output: ( ) system Transfer Functons Convenent representaton of a lnear, dynamc model. A transfer functon (TF) relates one nput and one output: x t X s y t system Y s The followng termnology s used: x y nput output forcng

More information

A Hybrid Variational Iteration Method for Blasius Equation

A Hybrid Variational Iteration Method for Blasius Equation Avalable at http://pvamu.edu/aam Appl. Appl. Math. ISSN: 1932-9466 Vol. 10, Issue 1 (June 2015), pp. 223-229 Applcatons and Appled Mathematcs: An Internatonal Journal (AAM) A Hybrd Varatonal Iteraton Method

More information

NUMERICAL DIFFERENTIATION

NUMERICAL DIFFERENTIATION NUMERICAL DIFFERENTIATION 1 Introducton Dfferentaton s a method to compute the rate at whch a dependent output y changes wth respect to the change n the ndependent nput x. Ths rate of change s called the

More information

5 The Rational Canonical Form

5 The Rational Canonical Form 5 The Ratonal Canoncal Form Here p s a monc rreducble factor of the mnmum polynomal m T and s not necessarly of degree one Let F p denote the feld constructed earler n the course, consstng of all matrces

More information

2.3 Nilpotent endomorphisms

2.3 Nilpotent endomorphisms s a block dagonal matrx, wth A Mat dm U (C) In fact, we can assume that B = B 1 B k, wth B an ordered bass of U, and that A = [f U ] B, where f U : U U s the restrcton of f to U 40 23 Nlpotent endomorphsms

More information

APPENDIX A Some Linear Algebra

APPENDIX A Some Linear Algebra APPENDIX A Some Lnear Algebra The collecton of m, n matrces A.1 Matrces a 1,1,..., a 1,n A = a m,1,..., a m,n wth real elements a,j s denoted by R m,n. If n = 1 then A s called a column vector. Smlarly,

More information

Linear Approximation with Regularization and Moving Least Squares

Linear Approximation with Regularization and Moving Least Squares Lnear Approxmaton wth Regularzaton and Movng Least Squares Igor Grešovn May 007 Revson 4.6 (Revson : March 004). 5 4 3 0.5 3 3.5 4 Contents: Lnear Fttng...4. Weghted Least Squares n Functon Approxmaton...

More information

A new Approach for Solving Linear Ordinary Differential Equations

A new Approach for Solving Linear Ordinary Differential Equations , ISSN 974-57X (Onlne), ISSN 974-5718 (Prnt), Vol. ; Issue No. 1; Year 14, Copyrght 13-14 by CESER PUBLICATIONS A new Approach for Solvng Lnear Ordnary Dfferental Equatons Fawz Abdelwahd Department of

More information

Affine transformations and convexity

Affine transformations and convexity Affne transformatons and convexty The purpose of ths document s to prove some basc propertes of affne transformatons nvolvng convex sets. Here are a few onlne references for background nformaton: http://math.ucr.edu/

More information

The lower and upper bounds on Perron root of nonnegative irreducible matrices

The lower and upper bounds on Perron root of nonnegative irreducible matrices Journal of Computatonal Appled Mathematcs 217 (2008) 259 267 wwwelsevercom/locate/cam The lower upper bounds on Perron root of nonnegatve rreducble matrces Guang-Xn Huang a,, Feng Yn b,keguo a a College

More information

More metrics on cartesian products

More metrics on cartesian products More metrcs on cartesan products If (X, d ) are metrc spaces for 1 n, then n Secton II4 of the lecture notes we defned three metrcs on X whose underlyng topologes are the product topology The purpose of

More information

General viscosity iterative method for a sequence of quasi-nonexpansive mappings

General viscosity iterative method for a sequence of quasi-nonexpansive mappings Avalable onlne at www.tjnsa.com J. Nonlnear Sc. Appl. 9 (2016), 5672 5682 Research Artcle General vscosty teratve method for a sequence of quas-nonexpansve mappngs Cuje Zhang, Ynan Wang College of Scence,

More information

Salmon: Lectures on partial differential equations. Consider the general linear, second-order PDE in the form. ,x 2

Salmon: Lectures on partial differential equations. Consider the general linear, second-order PDE in the form. ,x 2 Salmon: Lectures on partal dfferental equatons 5. Classfcaton of second-order equatons There are general methods for classfyng hgher-order partal dfferental equatons. One s very general (applyng even to

More information

8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS

8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS SECTION 8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS 493 8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS All the vector spaces you have studed thus far n the text are real vector spaces because the scalars

More information

Errors for Linear Systems

Errors for Linear Systems Errors for Lnear Systems When we solve a lnear system Ax b we often do not know A and b exactly, but have only approxmatons  and ˆb avalable. Then the best thng we can do s to solve ˆx ˆb exactly whch

More information

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X Statstcs 1: Probablty Theory II 37 3 EPECTATION OF SEVERAL RANDOM VARIABLES As n Probablty Theory I, the nterest n most stuatons les not on the actual dstrbuton of a random vector, but rather on a number

More information

STAT 309: MATHEMATICAL COMPUTATIONS I FALL 2018 LECTURE 16

STAT 309: MATHEMATICAL COMPUTATIONS I FALL 2018 LECTURE 16 STAT 39: MATHEMATICAL COMPUTATIONS I FALL 218 LECTURE 16 1 why teratve methods f we have a lnear system Ax = b where A s very, very large but s ether sparse or structured (eg, banded, Toepltz, banded plus

More information

The Feynman path integral

The Feynman path integral The Feynman path ntegral Aprl 3, 205 Hesenberg and Schrödnger pctures The Schrödnger wave functon places the tme dependence of a physcal system n the state, ψ, t, where the state s a vector n Hlbert space

More information

CSci 6974 and ECSE 6966 Math. Tech. for Vision, Graphics and Robotics Lecture 21, April 17, 2006 Estimating A Plane Homography

CSci 6974 and ECSE 6966 Math. Tech. for Vision, Graphics and Robotics Lecture 21, April 17, 2006 Estimating A Plane Homography CSc 6974 and ECSE 6966 Math. Tech. for Vson, Graphcs and Robotcs Lecture 21, Aprl 17, 2006 Estmatng A Plane Homography Overvew We contnue wth a dscusson of the major ssues, usng estmaton of plane projectve

More information

1 Matrix representations of canonical matrices

1 Matrix representations of canonical matrices 1 Matrx representatons of canoncal matrces 2-d rotaton around the orgn: ( ) cos θ sn θ R 0 = sn θ cos θ 3-d rotaton around the x-axs: R x = 1 0 0 0 cos θ sn θ 0 sn θ cos θ 3-d rotaton around the y-axs:

More information

9 Characteristic classes

9 Characteristic classes THEODORE VORONOV DIFFERENTIAL GEOMETRY. Sprng 2009 [under constructon] 9 Characterstc classes 9.1 The frst Chern class of a lne bundle Consder a complex vector bundle E B of rank p. We shall construct

More information

Convexity preserving interpolation by splines of arbitrary degree

Convexity preserving interpolation by splines of arbitrary degree Computer Scence Journal of Moldova, vol.18, no.1(52), 2010 Convexty preservng nterpolaton by splnes of arbtrary degree Igor Verlan Abstract In the present paper an algorthm of C 2 nterpolaton of dscrete

More information

MMA and GCMMA two methods for nonlinear optimization

MMA and GCMMA two methods for nonlinear optimization MMA and GCMMA two methods for nonlnear optmzaton Krster Svanberg Optmzaton and Systems Theory, KTH, Stockholm, Sweden. krlle@math.kth.se Ths note descrbes the algorthms used n the author s 2007 mplementatons

More information

1 GSW Iterative Techniques for y = Ax

1 GSW Iterative Techniques for y = Ax 1 for y = A I m gong to cheat here. here are a lot of teratve technques that can be used to solve the general case of a set of smultaneous equatons (wrtten n the matr form as y = A), but ths chapter sn

More information

PHYS 705: Classical Mechanics. Newtonian Mechanics

PHYS 705: Classical Mechanics. Newtonian Mechanics 1 PHYS 705: Classcal Mechancs Newtonan Mechancs Quck Revew of Newtonan Mechancs Basc Descrpton: -An dealzed pont partcle or a system of pont partcles n an nertal reference frame [Rgd bodes (ch. 5 later)]

More information

Supplement: Proofs and Technical Details for The Solution Path of the Generalized Lasso

Supplement: Proofs and Technical Details for The Solution Path of the Generalized Lasso Supplement: Proofs and Techncal Detals for The Soluton Path of the Generalzed Lasso Ryan J. Tbshran Jonathan Taylor In ths document we gve supplementary detals to the paper The Soluton Path of the Generalzed

More information

princeton univ. F 17 cos 521: Advanced Algorithm Design Lecture 7: LP Duality Lecturer: Matt Weinberg

princeton univ. F 17 cos 521: Advanced Algorithm Design Lecture 7: LP Duality Lecturer: Matt Weinberg prnceton unv. F 17 cos 521: Advanced Algorthm Desgn Lecture 7: LP Dualty Lecturer: Matt Wenberg Scrbe: LP Dualty s an extremely useful tool for analyzng structural propertes of lnear programs. Whle there

More information

The Order Relation and Trace Inequalities for. Hermitian Operators

The Order Relation and Trace Inequalities for. Hermitian Operators Internatonal Mathematcal Forum, Vol 3, 08, no, 507-57 HIKARI Ltd, wwwm-hkarcom https://doorg/0988/mf088055 The Order Relaton and Trace Inequaltes for Hermtan Operators Y Huang School of Informaton Scence

More information

Inexact Newton Methods for Inverse Eigenvalue Problems

Inexact Newton Methods for Inverse Eigenvalue Problems Inexact Newton Methods for Inverse Egenvalue Problems Zheng-jan Ba Abstract In ths paper, we survey some of the latest development n usng nexact Newton-lke methods for solvng nverse egenvalue problems.

More information

Deriving the X-Z Identity from Auxiliary Space Method

Deriving the X-Z Identity from Auxiliary Space Method Dervng the X-Z Identty from Auxlary Space Method Long Chen Department of Mathematcs, Unversty of Calforna at Irvne, Irvne, CA 92697 chenlong@math.uc.edu 1 Iteratve Methods In ths paper we dscuss teratve

More information

Consistency & Convergence

Consistency & Convergence /9/007 CHE 374 Computatonal Methods n Engneerng Ordnary Dfferental Equatons Consstency, Convergence, Stablty, Stffness and Adaptve and Implct Methods ODE s n MATLAB, etc Consstency & Convergence Consstency

More information

PHYS 705: Classical Mechanics. Calculus of Variations II

PHYS 705: Classical Mechanics. Calculus of Variations II 1 PHYS 705: Classcal Mechancs Calculus of Varatons II 2 Calculus of Varatons: Generalzaton (no constrant yet) Suppose now that F depends on several dependent varables : We need to fnd such that has a statonary

More information

CSCE 790S Background Results

CSCE 790S Background Results CSCE 790S Background Results Stephen A. Fenner September 8, 011 Abstract These results are background to the course CSCE 790S/CSCE 790B, Quantum Computaton and Informaton (Sprng 007 and Fall 011). Each

More information

6.3.4 Modified Euler s method of integration

6.3.4 Modified Euler s method of integration 6.3.4 Modfed Euler s method of ntegraton Before dscussng the applcaton of Euler s method for solvng the swng equatons, let us frst revew the basc Euler s method of numercal ntegraton. Let the general from

More information

Some modelling aspects for the Matlab implementation of MMA

Some modelling aspects for the Matlab implementation of MMA Some modellng aspects for the Matlab mplementaton of MMA Krster Svanberg krlle@math.kth.se Optmzaton and Systems Theory Department of Mathematcs KTH, SE 10044 Stockholm September 2004 1. Consdered optmzaton

More information

Lecture 3. Ax x i a i. i i

Lecture 3. Ax x i a i. i i 18.409 The Behavor of Algorthms n Practce 2/14/2 Lecturer: Dan Spelman Lecture 3 Scrbe: Arvnd Sankar 1 Largest sngular value In order to bound the condton number, we need an upper bound on the largest

More information

Random Walks on Digraphs

Random Walks on Digraphs Random Walks on Dgraphs J. J. P. Veerman October 23, 27 Introducton Let V = {, n} be a vertex set and S a non-negatve row-stochastc matrx (.e. rows sum to ). V and S defne a dgraph G = G(V, S) and a drected

More information

On a direct solver for linear least squares problems

On a direct solver for linear least squares problems ISSN 2066-6594 Ann. Acad. Rom. Sc. Ser. Math. Appl. Vol. 8, No. 2/2016 On a drect solver for lnear least squares problems Constantn Popa Abstract The Null Space (NS) algorthm s a drect solver for lnear

More information

Bézier curves. Michael S. Floater. September 10, These notes provide an introduction to Bézier curves. i=0

Bézier curves. Michael S. Floater. September 10, These notes provide an introduction to Bézier curves. i=0 Bézer curves Mchael S. Floater September 1, 215 These notes provde an ntroducton to Bézer curves. 1 Bernsten polynomals Recall that a real polynomal of a real varable x R, wth degree n, s a functon of

More information

ISSN: ISO 9001:2008 Certified International Journal of Engineering and Innovative Technology (IJEIT) Volume 3, Issue 1, July 2013

ISSN: ISO 9001:2008 Certified International Journal of Engineering and Innovative Technology (IJEIT) Volume 3, Issue 1, July 2013 ISSN: 2277-375 Constructon of Trend Free Run Orders for Orthogonal rrays Usng Codes bstract: Sometmes when the expermental runs are carred out n a tme order sequence, the response can depend on the run

More information

THE CHINESE REMAINDER THEOREM. We should thank the Chinese for their wonderful remainder theorem. Glenn Stevens

THE CHINESE REMAINDER THEOREM. We should thank the Chinese for their wonderful remainder theorem. Glenn Stevens THE CHINESE REMAINDER THEOREM KEITH CONRAD We should thank the Chnese for ther wonderful remander theorem. Glenn Stevens 1. Introducton The Chnese remander theorem says we can unquely solve any par of

More information

Bezier curves. Michael S. Floater. August 25, These notes provide an introduction to Bezier curves. i=0

Bezier curves. Michael S. Floater. August 25, These notes provide an introduction to Bezier curves. i=0 Bezer curves Mchael S. Floater August 25, 211 These notes provde an ntroducton to Bezer curves. 1 Bernsten polynomals Recall that a real polynomal of a real varable x R, wth degree n, s a functon of the

More information

Appendix for Causal Interaction in Factorial Experiments: Application to Conjoint Analysis

Appendix for Causal Interaction in Factorial Experiments: Application to Conjoint Analysis A Appendx for Causal Interacton n Factoral Experments: Applcaton to Conjont Analyss Mathematcal Appendx: Proofs of Theorems A. Lemmas Below, we descrbe all the lemmas, whch are used to prove the man theorems

More information

Inner Product. Euclidean Space. Orthonormal Basis. Orthogonal

Inner Product. Euclidean Space. Orthonormal Basis. Orthogonal Inner Product Defnton 1 () A Eucldean space s a fnte-dmensonal vector space over the reals R, wth an nner product,. Defnton 2 (Inner Product) An nner product, on a real vector space X s a symmetrc, blnear,

More information

LOW BIAS INTEGRATED PATH ESTIMATORS. James M. Calvin

LOW BIAS INTEGRATED PATH ESTIMATORS. James M. Calvin Proceedngs of the 007 Wnter Smulaton Conference S G Henderson, B Bller, M-H Hseh, J Shortle, J D Tew, and R R Barton, eds LOW BIAS INTEGRATED PATH ESTIMATORS James M Calvn Department of Computer Scence

More information

Uniqueness of Weak Solutions to the 3D Ginzburg- Landau Model for Superconductivity

Uniqueness of Weak Solutions to the 3D Ginzburg- Landau Model for Superconductivity Int. Journal of Math. Analyss, Vol. 6, 212, no. 22, 195-114 Unqueness of Weak Solutons to the 3D Gnzburg- Landau Model for Superconductvty Jshan Fan Department of Appled Mathematcs Nanjng Forestry Unversty

More information

Finite Element Modelling of truss/cable structures

Finite Element Modelling of truss/cable structures Pet Schreurs Endhoven Unversty of echnology Department of Mechancal Engneerng Materals echnology November 3, 214 Fnte Element Modellng of truss/cable structures 1 Fnte Element Analyss of prestressed structures

More information

Supplementary Notes for Chapter 9 Mixture Thermodynamics

Supplementary Notes for Chapter 9 Mixture Thermodynamics Supplementary Notes for Chapter 9 Mxture Thermodynamcs Key ponts Nne major topcs of Chapter 9 are revewed below: 1. Notaton and operatonal equatons for mxtures 2. PVTN EOSs for mxtures 3. General effects

More information

The Finite Element Method

The Finite Element Method The Fnte Element Method GENERAL INTRODUCTION Read: Chapters 1 and 2 CONTENTS Engneerng and analyss Smulaton of a physcal process Examples mathematcal model development Approxmate solutons and methods of

More information

Problem Set 9 Solutions

Problem Set 9 Solutions Desgn and Analyss of Algorthms May 4, 2015 Massachusetts Insttute of Technology 6.046J/18.410J Profs. Erk Demane, Srn Devadas, and Nancy Lynch Problem Set 9 Solutons Problem Set 9 Solutons Ths problem

More information

High resolution entropy stable scheme for shallow water equations

High resolution entropy stable scheme for shallow water equations Internatonal Symposum on Computers & Informatcs (ISCI 05) Hgh resoluton entropy stable scheme for shallow water equatons Xaohan Cheng,a, Yufeng Ne,b, Department of Appled Mathematcs, Northwestern Polytechncal

More information

ON A DETERMINATION OF THE INITIAL FUNCTIONS FROM THE OBSERVED VALUES OF THE BOUNDARY FUNCTIONS FOR THE SECOND-ORDER HYPERBOLIC EQUATION

ON A DETERMINATION OF THE INITIAL FUNCTIONS FROM THE OBSERVED VALUES OF THE BOUNDARY FUNCTIONS FOR THE SECOND-ORDER HYPERBOLIC EQUATION Advanced Mathematcal Models & Applcatons Vol.3, No.3, 2018, pp.215-222 ON A DETERMINATION OF THE INITIAL FUNCTIONS FROM THE OBSERVED VALUES OF THE BOUNDARY FUNCTIONS FOR THE SECOND-ORDER HYPERBOLIC EUATION

More information

Lecture 13 APPROXIMATION OF SECOMD ORDER DERIVATIVES

Lecture 13 APPROXIMATION OF SECOMD ORDER DERIVATIVES COMPUTATIONAL FLUID DYNAMICS: FDM: Appromaton of Second Order Dervatves Lecture APPROXIMATION OF SECOMD ORDER DERIVATIVES. APPROXIMATION OF SECOND ORDER DERIVATIVES Second order dervatves appear n dffusve

More information

Workshop: Approximating energies and wave functions Quantum aspects of physical chemistry

Workshop: Approximating energies and wave functions Quantum aspects of physical chemistry Workshop: Approxmatng energes and wave functons Quantum aspects of physcal chemstry http://quantum.bu.edu/pltl/6/6.pdf Last updated Thursday, November 7, 25 7:9:5-5: Copyrght 25 Dan Dll (dan@bu.edu) Department

More information

Linear, affine, and convex sets and hulls In the sequel, unless otherwise specified, X will denote a real vector space.

Linear, affine, and convex sets and hulls In the sequel, unless otherwise specified, X will denote a real vector space. Lnear, affne, and convex sets and hulls In the sequel, unless otherwse specfed, X wll denote a real vector space. Lnes and segments. Gven two ponts x, y X, we defne xy = {x + t(y x) : t R} = {(1 t)x +

More information

Prof. Dr. I. Nasser Phys 630, T Aug-15 One_dimensional_Ising_Model

Prof. Dr. I. Nasser Phys 630, T Aug-15 One_dimensional_Ising_Model EXACT OE-DIMESIOAL ISIG MODEL The one-dmensonal Isng model conssts of a chan of spns, each spn nteractng only wth ts two nearest neghbors. The smple Isng problem n one dmenson can be solved drectly n several

More information

EEE 241: Linear Systems

EEE 241: Linear Systems EEE : Lnear Systems Summary #: Backpropagaton BACKPROPAGATION The perceptron rule as well as the Wdrow Hoff learnng were desgned to tran sngle layer networks. They suffer from the same dsadvantage: they

More information

Physics 5153 Classical Mechanics. Principle of Virtual Work-1

Physics 5153 Classical Mechanics. Principle of Virtual Work-1 P. Guterrez 1 Introducton Physcs 5153 Classcal Mechancs Prncple of Vrtual Work The frst varatonal prncple we encounter n mechancs s the prncple of vrtual work. It establshes the equlbrum condton of a mechancal

More information

Some Comments on Accelerating Convergence of Iterative Sequences Using Direct Inversion of the Iterative Subspace (DIIS)

Some Comments on Accelerating Convergence of Iterative Sequences Using Direct Inversion of the Iterative Subspace (DIIS) Some Comments on Acceleratng Convergence of Iteratve Sequences Usng Drect Inverson of the Iteratve Subspace (DIIS) C. Davd Sherrll School of Chemstry and Bochemstry Georga Insttute of Technology May 1998

More information

Modelli Clamfim Equazione del Calore Lezione ottobre 2014

Modelli Clamfim Equazione del Calore Lezione ottobre 2014 CLAMFIM Bologna Modell 1 @ Clamfm Equazone del Calore Lezone 17 15 ottobre 2014 professor Danele Rtell danele.rtell@unbo.t 1/24? Convoluton The convoluton of two functons g(t) and f(t) s the functon (g

More information

Singular Value Decomposition: Theory and Applications

Singular Value Decomposition: Theory and Applications Sngular Value Decomposton: Theory and Applcatons Danel Khashab Sprng 2015 Last Update: March 2, 2015 1 Introducton A = UDV where columns of U and V are orthonormal and matrx D s dagonal wth postve real

More information

A Functionally Fitted 3-stage ESDIRK Method Kazufumi Ozawa Akita Prefectural University Honjo Akita , Japan

A Functionally Fitted 3-stage ESDIRK Method Kazufumi Ozawa Akita Prefectural University Honjo Akita , Japan A Functonally Ftted 3-stage ESDIRK Method Kazufum Ozawa Akta Prefectural Unversty Hono Akta 05-0055, Japan ozawa@akta-pu.ac.p Abstract A specal class of Runge-Kutta (-Nyström) methods called functonally

More information

Introduction to Vapor/Liquid Equilibrium, part 2. Raoult s Law:

Introduction to Vapor/Liquid Equilibrium, part 2. Raoult s Law: CE304, Sprng 2004 Lecture 4 Introducton to Vapor/Lqud Equlbrum, part 2 Raoult s Law: The smplest model that allows us do VLE calculatons s obtaned when we assume that the vapor phase s an deal gas, and

More information

4DVAR, according to the name, is a four-dimensional variational method.

4DVAR, according to the name, is a four-dimensional variational method. 4D-Varatonal Data Assmlaton (4D-Var) 4DVAR, accordng to the name, s a four-dmensonal varatonal method. 4D-Var s actually a drect generalzaton of 3D-Var to handle observatons that are dstrbuted n tme. The

More information

Structure and Drive Paul A. Jensen Copyright July 20, 2003

Structure and Drive Paul A. Jensen Copyright July 20, 2003 Structure and Drve Paul A. Jensen Copyrght July 20, 2003 A system s made up of several operatons wth flow passng between them. The structure of the system descrbes the flow paths from nputs to outputs.

More information

College of Computer & Information Science Fall 2009 Northeastern University 20 October 2009

College of Computer & Information Science Fall 2009 Northeastern University 20 October 2009 College of Computer & Informaton Scence Fall 2009 Northeastern Unversty 20 October 2009 CS7880: Algorthmc Power Tools Scrbe: Jan Wen and Laura Poplawsk Lecture Outlne: Prmal-dual schema Network Desgn:

More information

For now, let us focus on a specific model of neurons. These are simplified from reality but can achieve remarkable results.

For now, let us focus on a specific model of neurons. These are simplified from reality but can achieve remarkable results. Neural Networks : Dervaton compled by Alvn Wan from Professor Jtendra Malk s lecture Ths type of computaton s called deep learnng and s the most popular method for many problems, such as computer vson

More information

Randić Energy and Randić Estrada Index of a Graph

Randić Energy and Randić Estrada Index of a Graph EUROPEAN JOURNAL OF PURE AND APPLIED MATHEMATICS Vol. 5, No., 202, 88-96 ISSN 307-5543 www.ejpam.com SPECIAL ISSUE FOR THE INTERNATIONAL CONFERENCE ON APPLIED ANALYSIS AND ALGEBRA 29 JUNE -02JULY 20, ISTANBUL

More information

MATH 241B FUNCTIONAL ANALYSIS - NOTES EXAMPLES OF C ALGEBRAS

MATH 241B FUNCTIONAL ANALYSIS - NOTES EXAMPLES OF C ALGEBRAS MATH 241B FUNCTIONAL ANALYSIS - NOTES EXAMPLES OF C ALGEBRAS These are nformal notes whch cover some of the materal whch s not n the course book. The man purpose s to gve a number of nontrval examples

More information

SIO 224. m(r) =(ρ(r),k s (r),µ(r))

SIO 224. m(r) =(ρ(r),k s (r),µ(r)) SIO 224 1. A bref look at resoluton analyss Here s some background for the Masters and Gubbns resoluton paper. Global Earth models are usually found teratvely by assumng a startng model and fndng small

More information

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity LINEAR REGRESSION ANALYSIS MODULE IX Lecture - 30 Multcollnearty Dr. Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur 2 Remedes for multcollnearty Varous technques have

More information

The Geometry of Logit and Probit

The Geometry of Logit and Probit The Geometry of Logt and Probt Ths short note s meant as a supplement to Chapters and 3 of Spatal Models of Parlamentary Votng and the notaton and reference to fgures n the text below s to those two chapters.

More information

Module 9. Lecture 6. Duality in Assignment Problems

Module 9. Lecture 6. Duality in Assignment Problems Module 9 1 Lecture 6 Dualty n Assgnment Problems In ths lecture we attempt to answer few other mportant questons posed n earler lecture for (AP) and see how some of them can be explaned through the concept

More information

A MODIFIED METHOD FOR SOLVING SYSTEM OF NONLINEAR EQUATIONS

A MODIFIED METHOD FOR SOLVING SYSTEM OF NONLINEAR EQUATIONS Journal of Mathematcs and Statstcs 9 (1): 4-8, 1 ISSN 1549-644 1 Scence Publcatons do:1.844/jmssp.1.4.8 Publshed Onlne 9 (1) 1 (http://www.thescpub.com/jmss.toc) A MODIFIED METHOD FOR SOLVING SYSTEM OF

More information

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS Avalable onlne at http://sck.org J. Math. Comput. Sc. 3 (3), No., 6-3 ISSN: 97-537 COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

More information

A New Refinement of Jacobi Method for Solution of Linear System Equations AX=b

A New Refinement of Jacobi Method for Solution of Linear System Equations AX=b Int J Contemp Math Scences, Vol 3, 28, no 17, 819-827 A New Refnement of Jacob Method for Soluton of Lnear System Equatons AX=b F Naem Dafchah Department of Mathematcs, Faculty of Scences Unversty of Gulan,

More information

Kernel Methods and SVMs Extension

Kernel Methods and SVMs Extension Kernel Methods and SVMs Extenson The purpose of ths document s to revew materal covered n Machne Learnng 1 Supervsed Learnng regardng support vector machnes (SVMs). Ths document also provdes a general

More information

PHYS 705: Classical Mechanics. Canonical Transformation II

PHYS 705: Classical Mechanics. Canonical Transformation II 1 PHYS 705: Classcal Mechancs Canoncal Transformaton II Example: Harmonc Oscllator f ( x) x m 0 x U( x) x mx x LT U m Defne or L p p mx x x m mx x H px L px p m p x m m H p 1 x m p m 1 m H x p m x m m

More information

ELASTIC WAVE PROPAGATION IN A CONTINUOUS MEDIUM

ELASTIC WAVE PROPAGATION IN A CONTINUOUS MEDIUM ELASTIC WAVE PROPAGATION IN A CONTINUOUS MEDIUM An elastc wave s a deformaton of the body that travels throughout the body n all drectons. We can examne the deformaton over a perod of tme by fxng our look

More information

The Two-scale Finite Element Errors Analysis for One Class of Thermoelastic Problem in Periodic Composites

The Two-scale Finite Element Errors Analysis for One Class of Thermoelastic Problem in Periodic Composites 7 Asa-Pacfc Engneerng Technology Conference (APETC 7) ISBN: 978--6595-443- The Two-scale Fnte Element Errors Analyss for One Class of Thermoelastc Problem n Perodc Compostes Xaoun Deng Mngxang Deng ABSTRACT

More information

Nonlinear Overlapping Domain Decomposition Methods

Nonlinear Overlapping Domain Decomposition Methods Nonlnear Overlappng Doman Decomposton Methods Xao-Chuan Ca 1 Department of Computer Scence, Unversty of Colorado at Boulder, Boulder, CO 80309, ca@cs.colorado.edu Summary. We dscuss some overlappng doman

More information

Perron Vectors of an Irreducible Nonnegative Interval Matrix

Perron Vectors of an Irreducible Nonnegative Interval Matrix Perron Vectors of an Irreducble Nonnegatve Interval Matrx Jr Rohn August 4 2005 Abstract As s well known an rreducble nonnegatve matrx possesses a unquely determned Perron vector. As the man result of

More information

for Linear Systems With Strictly Diagonally Dominant Matrix

for Linear Systems With Strictly Diagonally Dominant Matrix MATHEMATICS OF COMPUTATION, VOLUME 35, NUMBER 152 OCTOBER 1980, PAGES 1269-1273 On an Accelerated Overrelaxaton Iteratve Method for Lnear Systems Wth Strctly Dagonally Domnant Matrx By M. Madalena Martns*

More information

Time-Varying Systems and Computations Lecture 6

Time-Varying Systems and Computations Lecture 6 Tme-Varyng Systems and Computatons Lecture 6 Klaus Depold 14. Januar 2014 The Kalman Flter The Kalman estmaton flter attempts to estmate the actual state of an unknown dscrete dynamcal system, gven nosy

More information

arxiv: v1 [math.ho] 18 May 2008

arxiv: v1 [math.ho] 18 May 2008 Recurrence Formulas for Fbonacc Sums Adlson J. V. Brandão, João L. Martns 2 arxv:0805.2707v [math.ho] 8 May 2008 Abstract. In ths artcle we present a new recurrence formula for a fnte sum nvolvng the Fbonacc

More information

Math 217 Fall 2013 Homework 2 Solutions

Math 217 Fall 2013 Homework 2 Solutions Math 17 Fall 013 Homework Solutons Due Thursday Sept. 6, 013 5pm Ths homework conssts of 6 problems of 5 ponts each. The total s 30. You need to fully justfy your answer prove that your functon ndeed has

More information

Yong Joon Ryang. 1. Introduction Consider the multicommodity transportation problem with convex quadratic cost function. 1 2 (x x0 ) T Q(x x 0 )

Yong Joon Ryang. 1. Introduction Consider the multicommodity transportation problem with convex quadratic cost function. 1 2 (x x0 ) T Q(x x 0 ) Kangweon-Kyungk Math. Jour. 4 1996), No. 1, pp. 7 16 AN ITERATIVE ROW-ACTION METHOD FOR MULTICOMMODITY TRANSPORTATION PROBLEMS Yong Joon Ryang Abstract. The optmzaton problems wth quadratc constrants often

More information

The Exact Formulation of the Inverse of the Tridiagonal Matrix for Solving the 1D Poisson Equation with the Finite Difference Method

The Exact Formulation of the Inverse of the Tridiagonal Matrix for Solving the 1D Poisson Equation with the Finite Difference Method Journal of Electromagnetc Analyss and Applcatons, 04, 6, 0-08 Publshed Onlne September 04 n ScRes. http://www.scrp.org/journal/jemaa http://dx.do.org/0.46/jemaa.04.6000 The Exact Formulaton of the Inverse

More information

Group Analysis of Ordinary Differential Equations of the Order n>2

Group Analysis of Ordinary Differential Equations of the Order n>2 Symmetry n Nonlnear Mathematcal Physcs 997, V., 64 7. Group Analyss of Ordnary Dfferental Equatons of the Order n> L.M. BERKOVICH and S.Y. POPOV Samara State Unversty, 4430, Samara, Russa E-mal: berk@nfo.ssu.samara.ru

More information