Preconditioning of conjugate gradient iterations in interior point MPC method

Size: px
Start display at page:

Download "Preconditioning of conjugate gradient iterations in interior point MPC method"

Transcription

1 MISUBISHI ELECRIC RESEARCH LABORAORIES Peconditioning of conjugate gadient iteations in inteio point MPC method Malyshev, A; Quiynen, R; Knyazev, A R August 25, 2018 Abstact hee ae seveal efficient diect solves fo stuctued systems of linea equations defining seach diections in pimal-dual inteio point methods applied to constained model pedictive contol poblems We popose eusing matix decompositions of diect solves as peconditiones in Kylov-subspace methods applied to subsequent iteations of the inteio point method, which esults in at least halving its asymptotic computational complexity We also analyze sensitivity of diect solves to the egulaization paametes IFAC Confeence on Nonlinea Model Pedictive Contol his wo may not be copied o epoduced in whole o in pat fo any commecial pupose Pemission to copy in whole o in pat without payment of fee is ganted fo nonpofit educational and eseach puposes povided that all such whole o patial copies include the following: a notice that such copying is by pemission of Mitsubishi Electic Reseach Laboatoies, Inc; an acnowledgment of the authos and individual contibutions to the wo; and all applicable potions of the copyight notice Copying, epoduction, o epublishing fo any othe pupose shall equie a license with payment of fee to Mitsubishi Electic Reseach Laboatoies, Inc All ights eseved Copyight c Mitsubishi Electic Reseach Laboatoies, Inc, Boadway, Cambidge, Massachusetts 02139

2

3 Peconditioning of conjugate gadient iteations in inteio point MPC method Alexande Malyshev Rien Quiynen Andew Knyazev Univesity of Begen, Depatment of Mathematics, Postbox 8700, 5020 Begen, Noway ( Mitsubishi Electic Reseach Laboatoies, 201 Boadway, 8th floo, Cambidge, MA 02139, USA ( s: Abstact: hee ae seveal efficient diect solves fo stuctued systems of linea equations defining seach diections in pimal-dual inteio point methods applied to constained model pedictive contol poblems We popose eusing matix decompositions of diect solves as peconditiones in Kylov-subspace methods applied to subsequent iteations of the inteio point method, which esults in at least halving its asymptotic computational complexity We also analyze sensitivity of diect solves to the egulaization paametes Keywods: model pedictive contol, inteio point, peconditioning, conjugate gadient 1 INRODUCION In this pape, peconditioning techniques ae poposed that may be useful fo numeical solution of optimal contol poblems in model pedictive contol (MPC); see (Rawlings et al, 2017) We conside MPC ove a eceding hoizon in the fom of a quadatic pogam (QP), whee the cost function is given by a convex quadatic pefomance index, the equality constaints ae the linea model dynamics, the inequality constaints on the state and contol inputs ae also linea; simila fomulations ae used in (Wang and Boyd, 2010; Shahzad et al, 2012; Fison et al, 2014; Quiynen et al, 2018) Such stuctued QPs appea as subpoblems within a sequential QP fo nonlinea MPC; see, eg, (Diehl et al, 2009) he inequality constaints in the QP poblem ae usually teated by the inteio point (IP) method o by the activeset method; cf (Batlett et al, 2000) he IP method is chaacteized by fast convegence, which is almost independent of the poblem size; see, eg, (Mehota, 1992; Gondzio, 2012, 2013) he active-set solves fo MPC ae highly competitive owing to athe small changes of the active set duing continuation steps, educing aithmetic costs by exploiting low-an updates of a Cholesy factoization when changing the active set as shown in (Gill et al, 1974; Kiches et al, 2011) Matices in systems of linea equations defining seach diections in the Newton method ae spase, consisting of blocs with banded matices Seveal efficient diect methods have been poposed fo solving such systems; see, eg, (Rao et al, 1998; Wang and Boyd, 2010; Domahidi et al, 2012; Fison and Jøgensen, 2013) Let n x, n u, n g, and N denote the dimensions of the state vaiable and the contol input, the numbe of the inequality constaints, and the hoizon length, espectively Aithmetic complexity of bloc stuctued diect solves used in the IP method typically amounts to O(N(n 3 x + n 3 u + n 3 g)) flops Numeical solution of linea systems of the Newton method, eg, using diect solves, is the most timeconsuming online pat of MPC Computational complexity can be educed if saddle-point linea systems of KK conditions ae eplaced by linea systems having symmetic positive definite (SPD) bloc stuctued matices Diect solves fo SPD linea systems ae commonly based on the Cholesy factoization, often implemented to tae advantage of the BLAS3 matix opeations, which is highly efficient on contempoay compute achitectues with advanced memoy hieachies he active-set methods may achieve theoetical complexity O(N(n 2 x+n 2 u+n 2 g)), if based on cost-efficient low-an updates, eg, of the Cholesy factoization, although the low-an updates can only use BLAS2; see (Andeson et al, 1999) Well-nown fast convegence of the IP method, its esilience to inexact implementations, and high pefomance of BLAS3 implementations of the diect solves can mae the IP method competitive with the active-set methods We have consideed the IP method in (Malyshev et al, 2018) using each of the thee diect solves: Riccati ecusion, Schu complement, and augmented Lagangian, with egulaization by adding the identity matix multiplied by a small paamete to some blocs he factoized Riccati ecusion and the augmented Lagangian egulaization have been found the best in tems of aithmetic opeations he spase linea systems of the Newton methods can be solved by iteative methods; see, eg, (Golub and Van Loan, 2013, Ch 11) If an efficient peconditione is available, the theoetical complexity of an iteative method is educed to O(N(n 2 x + n 2 u + n 2 g)), and the peconditioned iteative method outpefoms diect solves fo elatively lage values of n x, n u, n g he est of this pape is oganized as follows Section 2 intoduces the notation and the MPC QP poblem Section 3 discusses eductions to SPD systems We evisit eg-

4 ulaization of the Schu complement in Section 4 and the augmented Lagangian in Section 5 We show theoetically in Section 6 and numeically in Section 7 that the fome is less sensitive to the choice of the egulaization paamete, compaed to the latte Section 7 also illustates behavio of the poposed peconditioning in the conjugate gadient (CG) method, whee Cholesy factoizations computed at odd IP iteations ae eused fo peconditioning at even IP iteations to halve the computational cost Ou main contibutions include (a) euse of the Cholesy factoization fo peconditioning subsequent IP iteations and (b) analysis of sensitivity of the egulaized diect solves fom Sections 4 and 5 to the choice of egulaization paamete values ɛ and τ 2 QUADRAIC PROGRAMMING FOR MODEL PREDICIVE CONROL We conside the following quadatic pogam (QP) with espect to a decision vaiable d R n h, min 2 d Hd + h d subject to F d = f, Gd g, (1) whee n h = (n x + n u )N + n x, n f = n x (N + 1), and the matices H R n h n h, F R n f n h, G R ng n h and the vectos h, f, and g have the fom Q 0 S S 0 R H = 0 0 Q N 1 SN 1, S N 1 R N Q N I A 0 B 0 I F =, A N 1 B N 1 I G x 0 G u G = 0 0 G x N 1 G u N 1 0, G x N h x 0 h u f 0 0 g 0 h = h u, f = f 1, g = g 1 N 1 f N g N d 1 h x N he blocs I in F denote identity matices of ode n x We assume that the matix H is symmetic positive semidefinite We note that the matix H as well as the blocs G x i and Gu i of the matix G ae often diagonal Optimization poblem (1) is solved by the pimal-dual IP method; see, eg, (Wight, 1997; Shahzad et al, 2012) he IP method uses the following Kaush-Kuhn-uce (KK) optimality conditions Hd + F y + G z + h = 0, (2) F d f = 0, (3) Gd g + s = 0, (4) Z1 µs 1 1 = 0, (5) whee 1 = [1 1] R ng, and the matices Z = diag (z) and S = diag (s) ae diagonal he vectos y R n f and z R ng denote the Lagange multiplies he slac vecto s R ng with positive components and the small paamete µ > 0 ae equied by the IP method he KK conditions consist of linea equations (2)-(4) and nonlinea equation (5) he latte is usually substituted by ZS1 µ1 = 0 to impove scaling the esiduals in the Newton method he seach diection at iteation of the Newton method fo solving (2)-(5) is detemined by the system of linea equations H F G 0 d F y H G 0 0 I 0 0 S Z z s = with the esiduals H = Hd + F y + G z + h, F = F d f, G = Gd g + s, F G S, (6) S = Z S 1 σµ 1, whee σ (0, 1) he scala σ in the last fomula is called a centeing paamete he value of µ = (z ) s /n g is elated to a duality gap; eg, (Wight, 1997; Gondzio, 2012) Linea system (6) must be solved online epeatedly fo = 1, 2,, max, whee max is detemined by a stopping citeion, eg, based on the 2-nom of the esidual vecto in the ight-hand side of (6) he matix of linea system (6) becomes inceasingly ill-conditioned when some inequality constaints become active Nevetheless, the Newton IP method pefoms well, if system (6) is solved by diect methods, as discussed and justified in (Wight, 1998) We note that the matix of system (6) depends on the diagonal matices S and Z in the last bloc ow heefoe, a diect solve has to factoize the matix of (6) at each iteation, which is the most time-consuming opeation of the IP method Reducing costs of solving (6) is cucial 3 PRELIMINARY REDUCIONS BY BLOCK GAUSSIAN ELIMINAIONS he slac vaiable s is usually eliminated fom (6), which leads to the system of linea equations H F G F 0 0 d y = H F (7) G 0 W z W with the diagonal matix W = (Z ) 1 S he esidual G is substituted with the esidual W = G (Z ) 1 S he eliminated vaiable s is ecoveed by the fomula s = (Z ) 1 ( S + S z ) H F In (Shahzad et al, 2010a,b, 2012), the bloc is F 0 assumed to be nonsingula and (7) is educed to { W H F 1 } G + [G 0] z = F 0 0 [ M = W H F [G 0] F 0 ] 1 H F

5 he main advantage of such a eduction is that the matix H F 1 G M = [G 0] is computed only once fo all F 0 0 IP iteations he SPD system (W + M) z = M is then solved by peconditioned CG Howeve, cheap peconditioning via the diagonal pat of W + M may be not sufficient to povide fast convegence, while peconditioning using the bloc diagonal o tidiagonal pats of W + M equie expensive computations of the Cholesy factoization at evey iteation In contast, we eliminate the vaiable z fom linea system (7) he esulting 2 2 bloc linea system H F d F 0 y = E F (8) has the esidual E = H + G (W ) 1 W and the bloc H = H + G (W ) 1 G, (9) which is symmetic positive semidefinite and bloc diagonal with the same spasity patten as the matix H he eliminated vaiable z is ecoveed by the fomula z = (W ) 1 ( G d + W ) (10) Systems in the fom (8) ae called saddle point poblems Now that we have educed (6) to (8), futhe eductions of the linea system (8) with espect to the vaiables d o y ae possible In the next two sections, we descibe two popula altenative eductions, Schu complement- and augmented Lagangian-based, both supplemented with egulaization by adding the identity matix multiplied by a small paamete to some blocs 4 INEXAC NEWON MEHOD BY HE SCHUR COMPLEMEN APPROACH he elimination of d is most popula in the pimal-dual IP method; see (Wight, 1997) If the matix H in (9) is singula, it is usually eplaced by a egulaized SPD bloc Ĥ = H + ɛi, whee ɛ is a small positive constant A justification of such a eplacement is due to the fact that the IP algoithm can use an inexact Newton method, ie, the system (6) may be solved only appoximately; see, eg, (Nocedal and Wight, 2006) In the following egulaized analog of system (8), [Ĥ ] F d F 0 y = E F, (11) the fist equation of (11) is equivalent to d + (Ĥ ) 1 F y = (Ĥ ) 1 E and hence [ F (Ĥ ) 1 F ] y = F F (Ĥ ) 1 E (12) Equation (12), often called a nomal equation, has the SPD matix F (Ĥ ) 1 F, which is the Schu complement fo linea system (11) SPD linea system (12) can be solved by peconditioned CG Fo nonlinea MPC, peconditioning is poposed, eg, in (Knyazev et al, 2015; Knyazev and Malyshev, 2015, 2016) fo the Newton method Peconditioning in the IP context is widely discussed; see, eg, the suvey in (Gondzio, 2012) Inteesting elations between peconditiones fo nomal (12) and indefinite (8) equations ae discoveed in (Oliveia and Soensen, 2005) We popose using the Cholesy factoization of the matix F (Ĥ 1 ) 1 F, ie based on the values fom the pevious IP iteation 1, as a peconditione fo CG in iteation he aithmetic complexity of the Cholesy factoization of the matix F (Ĥ 1 ) 1 F is ( 19 N 3 n3 x + 4n 2 xn u + 2n x n 2 u + 1 ) 3 n3 u, whee only the leading tems ae taen into account 5 INEXAC NEWON MEHOD BY AUGMENED LAGRANGIAN REGULARIZAION In this section, we investigate elimination of the vaiable y fom system (8) Since the IP iteations allow the inexact Newton method, we conside the following egulaized analog of system (8), H F d F τi y = E F, (13) whee the positive egulaization paamete τ is sufficiently small Such a egulaization is often used in iteative methods fo numeical solution of saddle point poblems, especially fo peconditioning, eg, (Rees and Geif, 2007; Benzi and Wathen, 2008), and is also connected to the augmented Lagangian method in numeical optimization; see, eg, (Nocedal and Wight, 2006) Linea system (13) can be equivalently tansfomed to the bloc tiangula system [ H + 1 ] [ d τ F ] [ F 0 F τi y = E + 1 ] τ F F (14) he solution of (14) is given by the fomulas d = ( H + 1 τ F F ) 1 ( E + 1 τ F F ), (15) y = 1 τ ( F + F d ) (16) he main computational buden in fomulas (15)-(16) is solving the system with the SPD matix H + 1 τ F F in (15), which depends on Aithmetic complexity of the Cholesy factoization fo the matix H + 1 τ F F equals ( 7 N 3 n3 x + 4n 2 xn u + 2n x n 2 u + 1 ) 3 n3 u By analogy with the pevious section, we popose evaluating (15) by peconditioned CG, whee the peconditione is based on the Cholesy factoization of the system matix H τ F F fom the pevious IP iteation 6 ANALYSIS OF REGULARIZAION IN INEXAC NEWON MEHODS H F Regulaization of the saddle point matices is F 0 a well-nown topic in optimization theoy Fo example, (Vandebei, 1995) investigates LDL factoizations of the following matices, with both diagonal blocs egulaized, H + ɛi F (17) F τi F

6 (Saundes and omlin, 1996) conside the IP iteations, whee the inexact Newton method uses egulaized linea systems with the matices of fom (17) hey give a fomulation of the initial QP poblem, which is equivalent to the egulaized IP iteation, and pesent numeous tests showing high efficiency of the egulaization appoach In this section, we contibute into theoetical analysis of the egulaizations given by equations (11) and (13) Since the matix H is symmetic positive semidefinite, it has its symmetic positive semidefinite squae oot H 1/2 he convex saddle point poblem can be analyzed by the aid of the genealized singula value decomposition (GSVD) of the matix pai H 1/2 and F ; see, eg, (Golub and Van Loan, 2013) In ou case, thee exist othogonal matices U 1 and U 2 and a nonsingula matix V such that H 1/2 C = U 1 V I, F = U 2 [S 0] V, whee the diagonal matices C = c 1 c2 c nf, S = s 1 s2 s nf have nonnegative diagonal elements c i and s i such that c 2 i + s2 i = 1 fo i = 1,, n f We conside below the egulaization H + ɛv V instead of ou actual egulaization Ĥ = H + ɛi fo simplicity of pesentation heoem 1 he esidual matices ( ) ( ) R S = H F H + ɛv V F 1 I F 0 F 0 and ( ) ( ) R L = H F H F 1 I F 0 F τi satisfy the bounds R S ɛ cond(v ) max ( 1, S 1 ), (18) R L τ cond(v ) max ( S(S 2 + τc 2 ) 1, C 2 (S 2 + τc 2 ) 1 ) (19) Poof Bounds (18) and (19) follow fom the identities ( ) 0 ɛs 1 V R S = U2 ɛ ( ɛ I V U2) 0 0 and ( V R L = U2) 0 0 ( ) 1 0 V τs(s 2 + τc 2 ) 1 τc 2 (S 2 + τc 2 ) 1 U2 heoem 1 implies the following asymptotic esults in the situation whee the smallest singula value σ min (S) vanishes he esiduals fo the Schu complement and augmented Lagangian egulaizations ae small if ɛ σ min (S) and τ σ min (S) 2 We thus conclude that the Schu based egulaization has much wide inteval of ɛ, whee the egulaization wos, than the woing inteval of τ in the augmented Lagangian egulaization 7 CASE SUDY: CHAIN OF OSCILLAING MASSES Let us pesent peliminay esults fom numeical expeiments fo the classical linea contol example of a chain of n/2 unit masses, connected by spings of unit stiffness without damping, whee the ends ae attached to fixed walls We analyze computational pefomance and numeical obustness solving one QP at a paticula sampling instant a epesentative case fo pefoming full closedloop MPC simulations with cold initializations 71 Poblem Fomulation and Algoithm Implementation he continuous-time state-space system is discetized using a sample time τ = 05 while eeping the m inputs constant between the sampling instants he discete-time dynamics ae x i+1 = A i x i + B i u i, whee A i = exp ( τa c ), B i = A 1 c (A i I n ) 0 n/2 I m, and 0 n/2 m 2 1 0n/2 I A c = n/ , = n/2 0 n/2 1 2 he contol inputs and states ae subject to the inequality constaints 05 u(i) 05, i = 0,, N 1, 35 x(i) 35, i = 1,, N, whee x( ) denotes the position of the masses he cost matices in the MPC fomulation ae R i = 10 6 I, S i = 0, and Q i = C C = [ I n/2 0 ] [ In/2 0 ] he initial values of the state coespond to f 0 = 35 [ ] Ou numeical expeiments in MALAB ae pefomed on a test poblem of the following size: we conside 6 masses such that the state dimension equals n = 12, the numbe of contol inputs is m = 3 and the contol hoizon length is N = 30 he numbe of inequality constaints at each contol inteval is n + 2m = 18 Ou MALAB code implements the vaiant of the IP method descibed in moe detail by (Shahzad et al, 2012) and uses the same algoithmic constants and stopping citeion such as σ = 01, γ = 10 3, β = 2, ɛ = 10 7 We epot the esults of solving QP fom a cold stat at the initial time of the MPC contolle, whee the initial values fo x, u, y, z, and s ae equal to 1 he IP method fo QP geneally does not benefit much fom wam stating, especially fo S and Z, (Batlett et al, 2000) As a efeence fo futhe compaisons, we initially un the IP method in which the linea system (8) is solved exactly by the bacslash opeato in MALAB Ou implementation etuns a solution to the quadatic pogam afte 20 IP iteations and with the 2-nom of the IP esidual equal to [ H, F, G, S ] 2 =

7 # iteations IP-AL # iteations IP-SC Residual IP-AL Residual IP-SC o # CG iteations IP-AL IP-SC Fig 1 he numbe of IP iteations and the 2-nom of the esidual fo the solution with espect to ɛ o τ Ou poposed implementation of the IP method uses diect solves based on the Schu complement (IP-SC) o the augmented Lagangian (IP-AL) egulaization appoach he IP-SC and IP-AL algoithms use egulaization of the poblem with a small value of the paamete ɛ and τ, espectively Smalle values of ɛ and τ give bette appoximations to the KK matix, but too small values may lead to pohibitively lage condition numbes, fo which the IP method beas down in ou implementation using double pecision aithmetic Choosing the egulaization paamete caefully becomes inceasingly impotant when pefoming computations with a lowe pecision, eg, using single pecision aithmetic 72 Numeical Stability and Paamete Sensitivity Let us investigate the dependence of IP-SC on ɛ and IP- AL on τ Fo this pupose, we choose 100 values of ɛ and τ unifomly distibuted between 10 2 and in the logaithmic scale Figue 1 displays the numbe of IP iteations and the 2-nom of the IP esidual with espect to the value of the paamete ɛ = τ Ou numeical esults fo ɛ o τ between appoximately and 10 3 almost coincide with those fo the bacslash opeato, ie, the numbe of IP iteations equals 20 and the 2-nom of the IP esidual is vey close to 10 7 Both algoithms have simila behavio in Figue 1 fo the ange of values of ɛ and τ he IP-SC method is clealy pefeential fo pactically impotant small values, below 10 11, consistent with ou analysis in Section 6 IP-SC and IP-AL ae damatically diffeent fo the lage values of ɛ and τ within the inteval [10 35, 10 2 ] he algoithm IP-SC continues executing well, while IP-AL has vey lage esiduals and beas down 73 Peconditioned CG within the IP Method We epot ou peliminay esults pefoming expeiments with peconditioned CG inside the IP method he peconditione is applied at even IP iteations and uses the Cholesy factoization fom the pevious odd IP iteation Odd IP iteations ae esolved exactly, which technically o IP iteation numbe Fig 2 he numbe of peconditioned CG iteations in the IP method fo one MPC time step coesponds to one CG iteation he dimension of the state is n x = 240, coesponding to 120 masses, and the numbe of the contol inputs is n u = 3 We incease the dimension of the state 20 times, compaed to the pevious expeiments, in ode to mae the test case epesentative fo using CG, since unning CG on small-size linea systems may esult in atypical convegence behavio Figue 2 shows that the numbe of CG iteations in IP-AL with τ = 10 8 sometimes is gowing with the numbe of IP iteations his behavio can be obseved also less stongly fo the esults of the IP-SC method in Figue 2 We anticipate that tuning of IP-AL and IP-SC paametes may futhe impove the pefomance of the peconditioned CG method in both cases Pactical embedded MPC implementations equie solving each QP within tight timing constaints heefoe, it is common to limit the maximum numbe of IP iteations to a pedetemined fixed value Similaly, one may limit the numbe of CG iteations, eg, up to 10 iteations in Figue 2, at the potential cost of an incease in the oveall numbe of IP iteations Finally, one can change the ule of when to constuct a new peconditione, ie to compute the new stuctued Cholesy factoization Instead of constucting the new peconditione at each odd IP iteation, one could, eg, apply the same peconditione until the numbe of CG iteations passes a theshold, futhe inceasing pefomance 8 CONCLUSION We conside peconditioned CG fo computation of the seach diection in the inexact Newton based IP method to solve each QP in MPC We popose peconditioning by eusing matix factos of diect solves fo the Schu complement of the nomal equations o fo the augmented Lagangian egulaization appoach We can apply peconditioned CG, eg, at evey even iteation of the IP method with the Cholesy factos computed in the pevious odd iteation Such a vaiant could allow computing optimal contol solutions twice as fast compaed to the taditional use of the diect solves at evey IP/Newton iteation

8 We have caied out numeical tests of the two poposed egulaizations fo saddle point linea systems On the one hand, smalle computational complexity of the peconditiones maes the augmented Lagangian egulaization based appoach somewhat pefeential On the othe hand, ou petubation analysis and numeical expeiments demonstate that peconditioning based on the moe computationally expensive Schu complement is geneally moe obust to ound-off eos, which is advantageous fo low-pecision embedded MPC implementations REFERENCES Andeson, E, Bai, Z, Bischof, C, Blacfod, LS, Demmel, J, Dongaa, J, Coz, JD, Geenbaum, A, Hammaling, S, McKenney, A, and Soensen, D (1999) LAPACK Uses Guide SIAM, Philadelphia Batlett, R, Wächte, A, and Biegle, L (2000) Active set vs inteio point stategies fo model pedictive contol In Poc Ameican Contol Conf, Chicago, IL Benzi, M and Wathen, AJ (2008) Some peconditioning techniques fo saddle point poblems In WHA Schildes et al (ed), Model Ode Reduction: heoy, Reseach Aspects and Applications, Mathematics in Industy, vol 13, Spinge, Belin, Heidelbeg, Gemany Diehl, M, Feeau, HJ, and Havebee, N (2009) Efficient numeical methods fo nonlinea MPC and moving hoizon estimation In L Magni et al (ed), Nonlinea Model Pedictive Contol, LNCIS 384, Spinge, Heidelbeg, Gemany Domahidi, A, Zgaggen, AU, Zeilinge, MN, Moai, M, and Jones, CN (2012) Efficient inteio point methods fo multistage poblems aising in eceding hoizon contol In Poc 51st IEEE Conf Decision and Contol, Maui, Hawaii, USA Fletche, R (1987) Pactical Methods of Optimization Wiley, 2nd edition Fison, G and Jøgensen, JB (2013) Efficient implementation of the Riccati ecusion fo solving lineaquadatic contol poblems In Poc IEEE Multiconfeence on Systems and Contol, India Fison, G, Soensen, HB, Dammann, B, and Jøgensen, JB (2014) High-pefomance small-scale solves fo linea model pedictive contol In Poc Euopean Contol Conf (ECC), Fance Gill, P, Golub, G, Muay, W, and Saundes, MA (1974) Methods fo modifying matix factoizations Math Comp, 28(126), Golub, G and Van Loan, CV (2013) Matix computations he Johns Hopins Univ Pess, Baltimoe, 4th edition Gondzio, J (2012) Inteio point methods 25 yeas late Euopean J Opeational Reseach, 218, Gondzio, J (2013) Convegence analysis of an inexact feasible inteio point method fo convex quadatic pogamming SIAM J Optim, 23(3), Kiches, C, Boc, HG, Schlöde, JP, and Sage, S (2011) A factoization with update pocedues fo a KK matix aising in diect optimal contol Math Pog Comp, 3(4), Knyazev, A, Fujii, Y, and Malyshev, A (2015) Peconditioned continuation model pedictive contol In Poc SIAM Conf Contol Appl, , Pais, Fance Knyazev, A and Malyshev, A (2015) Peconditioning fo continuation model pedictive contol IFAC- PapesOnLine, 48(23), Knyazev, A and Malyshev, A (2016) Spase peconditioning fo model pedictive contol In Poc Ameican Contol Conf (ACC), Boston, MA, USA Knyazev, AV (2007) Obsevations on degeneate saddle point poblems Compute Methods in Applied Mechanics and Engineeing, 196(37), Malyshev, A, Quiynen, R, Knyazev, A,, and Caiano, SD (2018) A egulaized newton solve fo linea model pedictive contol In Poc Euopean Contol Conf (ECC) Limassol, Cypus Mehota, S (1992) On the implementation of a pimaldual inteio point method SIAM J Optim, 2(4), Nocedal, J and Wight, SJ (2006) Numeical Optimization Spinge, second edition Oliveia, ARL and Soensen, DC (2005) A new class of peconditiones fo lage-scale linea systems fom inteio point methods fo linea pogamming Linea Algeba Appl, 394, 1 24 Quiynen, R, Knyazev, A, and Caiano, SD (2018) Bloc stuctued peconditioning within an active-set method fo eal-time optimal contol In Poc Euopean Contol Conf (ECC) Limassol, Cypus Rao, CV, Wight, SJ, and Rawlings, JB (1998) Application of inteio-point methods to model pedictive contol J Optimiz heoy Appl, 99, Rawlings, JB, Mayne, DQ, and Diehl, M (2017) Model Pedictive Contol: heoy, Computation, and Design Nob Hill Publishing, Madison, WI, second edition Rees, and Geif, C (2007) A peconditione fo linea systems aising fom inteio point optimization methods SIAM J Sci Comp, 29(5), Saundes, M and omlin, JA (1996) Solving egulaized linea pogams using baie methods and KK systems ech ep sol 96-4, Systems Optimization Laboatoy, Depatment of Opeations Reseach, Stanfod Univesity, Stanfod, CA 94305, USA Shahzad, A, Keigan, EC, and Constantinides, GA (2010a) A fast well-conditionend inteio point method fo pedictive contol In Poc IEEE Conf Decision Contol, Atlanta, GA, USA, Decembe Shahzad, A, Keigan, EC, and Constantinides, GA (2010b) Peconditiones fo inexact inteio point methods fo pedictive contol In Poc Ameican Contol Conf, Baltimoe, MD, USA Shahzad, A, Keigan, EC, and Constantinides, GA (2012) A stable and efficient method fo solving a convex quadatic pogam with application to optimal contol SIAM J Optim, 22(4), Vandebei, RJ (1995) Symmetic quasidefinite matices SIAM J Optimization, 5, Wang, Y and Boyd, S (2010) Fast model pedictive contol using online optimization IEEE ans Contol Syst echnology, 18(2), Wight, MH (1998) Ill-conditioning and computational eo in inteio point methods fo nonlinea pogamming SIAM J Optim, 9(1), Wight, SJ (1997) Pimal-Dual Inteio Point Methods SIAM, Philadelphia

Conjugate Gradient Methods. Michael Bader. Summer term 2012

Conjugate Gradient Methods. Michael Bader. Summer term 2012 Gadient Methods Outlines Pat I: Quadatic Foms and Steepest Descent Pat II: Gadients Pat III: Summe tem 2012 Pat I: Quadatic Foms and Steepest Descent Outlines Pat I: Quadatic Foms and Steepest Descent

More information

Scientific Computing II

Scientific Computing II Scientific Computing II Conjugate Gadient Methods Michael Bade Summe 2014 Conjugate Gadient Methods, Summe 2014 1 Families of Iteative Solves elaxation methods: Jacobi-, Gauss-Seidel-Relaxation,... Ove-Relaxation-Methods

More information

Pearson s Chi-Square Test Modifications for Comparison of Unweighted and Weighted Histograms and Two Weighted Histograms

Pearson s Chi-Square Test Modifications for Comparison of Unweighted and Weighted Histograms and Two Weighted Histograms Peason s Chi-Squae Test Modifications fo Compaison of Unweighted and Weighted Histogams and Two Weighted Histogams Univesity of Akueyi, Bogi, v/noduslód, IS-6 Akueyi, Iceland E-mail: nikolai@unak.is Two

More information

A Power Method for Computing Square Roots of Complex Matrices

A Power Method for Computing Square Roots of Complex Matrices JOURNAL OF MATHEMATICAL ANALYSIS AND APPLICATIONS 13, 39345 1997 ARTICLE NO. AY975517 A Powe Method fo Computing Squae Roots of Complex Matices Mohammed A. Hasan Depatment of Electical Engineeing, Coloado

More information

Gradient-based Neural Network for Online Solution of Lyapunov Matrix Equation with Li Activation Function

Gradient-based Neural Network for Online Solution of Lyapunov Matrix Equation with Li Activation Function Intenational Confeence on Infomation echnology and Management Innovation (ICIMI 05) Gadient-based Neual Netwok fo Online Solution of Lyapunov Matix Equation with Li Activation unction Shiheng Wang, Shidong

More information

State tracking control for Takagi-Sugeno models

State tracking control for Takagi-Sugeno models State tacing contol fo Taagi-Sugeno models Souad Bezzaoucha, Benoît Max,3,DidieMaquin,3 and José Ragot,3 Abstact This wo addesses the model efeence tacing contol poblem It aims to highlight the encouteed

More information

Absorption Rate into a Small Sphere for a Diffusing Particle Confined in a Large Sphere

Absorption Rate into a Small Sphere for a Diffusing Particle Confined in a Large Sphere Applied Mathematics, 06, 7, 709-70 Published Online Apil 06 in SciRes. http://www.scip.og/jounal/am http://dx.doi.og/0.46/am.06.77065 Absoption Rate into a Small Sphee fo a Diffusing Paticle Confined in

More information

EVALUATION OF PRECONDITIONED KRYLOV SUBSPACE METHODS FOR NAVIER-STOKES EQUATIONS

EVALUATION OF PRECONDITIONED KRYLOV SUBSPACE METHODS FOR NAVIER-STOKES EQUATIONS The Seventh Asia-Pacific Confeence on Wind Engineeing, Novembe 8-12, 2009, Taipei, Taiwan EVALUATION OF PRECONDITIONED KRYLOV SUBSPACE METHODS FOR NAVIER-STOKES EQUATIONS Aishe Zhang 1,2, Ming Gu 3, Cuilan

More information

Localization of Eigenvalues in Small Specified Regions of Complex Plane by State Feedback Matrix

Localization of Eigenvalues in Small Specified Regions of Complex Plane by State Feedback Matrix Jounal of Sciences, Islamic Republic of Ian (): - () Univesity of Tehan, ISSN - http://sciencesutaci Localization of Eigenvalues in Small Specified Regions of Complex Plane by State Feedback Matix H Ahsani

More information

Hua Xu 3 and Hiroaki Mukaidani 33. The University of Tsukuba, Otsuka. Hiroshima City University, 3-4-1, Ozuka-Higashi

Hua Xu 3 and Hiroaki Mukaidani 33. The University of Tsukuba, Otsuka. Hiroshima City University, 3-4-1, Ozuka-Higashi he inea Quadatic Dynamic Game fo Discete-ime Descipto Systems Hua Xu 3 and Hioai Muaidani 33 3 Gaduate School of Systems Management he Univesity of suuba, 3-9- Otsua Bunyo-u, oyo -0, Japan xuhua@gssm.otsua.tsuuba.ac.jp

More information

STATE VARIANCE CONSTRAINED FUZZY CONTROL VIA OBSERVER-BASED FUZZY CONTROLLERS

STATE VARIANCE CONSTRAINED FUZZY CONTROL VIA OBSERVER-BASED FUZZY CONTROLLERS Jounal of Maine Science and echnology, Vol. 4, No., pp. 49-57 (6) 49 SAE VARIANCE CONSRAINED FUZZY CONROL VIA OBSERVER-BASED FUZZY CONROLLERS Wen-Je Chang*, Yi-Lin Yeh**, and Yu-eh Meng*** Key wods: takagi-sugeno

More information

NOTE. Some New Bounds for Cover-Free Families

NOTE. Some New Bounds for Cover-Free Families Jounal of Combinatoial Theoy, Seies A 90, 224234 (2000) doi:10.1006jcta.1999.3036, available online at http:.idealibay.com on NOTE Some Ne Bounds fo Cove-Fee Families D. R. Stinson 1 and R. Wei Depatment

More information

MULTILAYER PERCEPTRONS

MULTILAYER PERCEPTRONS Last updated: Nov 26, 2012 MULTILAYER PERCEPTRONS Outline 2 Combining Linea Classifies Leaning Paametes Outline 3 Combining Linea Classifies Leaning Paametes Implementing Logical Relations 4 AND and OR

More information

Information Retrieval Advanced IR models. Luca Bondi

Information Retrieval Advanced IR models. Luca Bondi Advanced IR models Luca Bondi Advanced IR models 2 (LSI) Pobabilistic Latent Semantic Analysis (plsa) Vecto Space Model 3 Stating point: Vecto Space Model Documents and queies epesented as vectos in the

More information

A matrix method based on the Fibonacci polynomials to the generalized pantograph equations with functional arguments

A matrix method based on the Fibonacci polynomials to the generalized pantograph equations with functional arguments A mati method based on the Fibonacci polynomials to the genealized pantogaph equations with functional aguments Ayşe Betül Koç*,a, Musa Çama b, Aydın Kunaz a * Coespondence: aysebetuloc @ selcu.edu.t a

More information

Implicit Constraint Enforcement for Rigid Body Dynamic Simulation

Implicit Constraint Enforcement for Rigid Body Dynamic Simulation Implicit Constaint Enfocement fo Rigid Body Dynamic Simulation Min Hong 1, Samuel Welch, John app, and Min-Hyung Choi 3 1 Division of Compute Science and Engineeing, Soonchunhyang Univesity, 646 Eupnae-i

More information

Central Coverage Bayes Prediction Intervals for the Generalized Pareto Distribution

Central Coverage Bayes Prediction Intervals for the Generalized Pareto Distribution Statistics Reseach Lettes Vol. Iss., Novembe Cental Coveage Bayes Pediction Intevals fo the Genealized Paeto Distibution Gyan Pakash Depatment of Community Medicine S. N. Medical College, Aga, U. P., India

More information

Using Laplace Transform to Evaluate Improper Integrals Chii-Huei Yu

Using Laplace Transform to Evaluate Improper Integrals Chii-Huei Yu Available at https://edupediapublicationsog/jounals Volume 3 Issue 4 Febuay 216 Using Laplace Tansfom to Evaluate Impope Integals Chii-Huei Yu Depatment of Infomation Technology, Nan Jeon Univesity of

More information

DonnishJournals

DonnishJournals DonnishJounals 041-1189 Donnish Jounal of Educational Reseach and Reviews. Vol 1(1) pp. 01-017 Novembe, 014. http:///dje Copyight 014 Donnish Jounals Oiginal Reseach Pape Vecto Analysis Using MAXIMA Savaş

More information

Fractional Tikhonov regularization for linear discrete ill-posed problems

Fractional Tikhonov regularization for linear discrete ill-posed problems BIT manuscipt No. (will be inseted by the edito) Factional Tikhonov egulaization fo linea discete ill-posed poblems Michiel E. Hochstenbach Lotha Reichel Received: date / Accepted: date Abstact Tikhonov

More information

Surveillance Points in High Dimensional Spaces

Surveillance Points in High Dimensional Spaces Société de Calcul Mathématique SA Tools fo decision help since 995 Suveillance Points in High Dimensional Spaces by Benad Beauzamy Januay 06 Abstact Let us conside any compute softwae, elying upon a lage

More information

Relating Branching Program Size and. Formula Size over the Full Binary Basis. FB Informatik, LS II, Univ. Dortmund, Dortmund, Germany

Relating Branching Program Size and. Formula Size over the Full Binary Basis. FB Informatik, LS II, Univ. Dortmund, Dortmund, Germany Relating Banching Pogam Size and omula Size ove the ull Binay Basis Matin Saueho y Ingo Wegene y Ralph Wechne z y B Infomatik, LS II, Univ. Dotmund, 44 Dotmund, Gemany z ankfut, Gemany sauehof/wegene@ls.cs.uni-dotmund.de

More information

Model and Controller Order Reduction for Infinite Dimensional Systems

Model and Controller Order Reduction for Infinite Dimensional Systems IT J. Eng. Sci., Vol. 4, No.,, -6 Model and Contolle Ode Reduction fo Infinite Dimensional Systems Fatmawati,*, R. Saagih,. Riyanto 3 & Y. Soehayadi Industial and Financial Mathematics Goup email: fatma47@students.itb.ac.id;

More information

ELASTIC ANALYSIS OF CIRCULAR SANDWICH PLATES WITH FGM FACE-SHEETS

ELASTIC ANALYSIS OF CIRCULAR SANDWICH PLATES WITH FGM FACE-SHEETS THE 9 TH INTERNATIONAL CONFERENCE ON COMPOSITE MATERIALS ELASTIC ANALYSIS OF CIRCULAR SANDWICH PLATES WITH FGM FACE-SHEETS R. Sbulati *, S. R. Atashipou Depatment of Civil, Chemical and Envionmental Engineeing,

More information

Temporal-Difference Learning

Temporal-Difference Learning .997 Decision-Making in Lage-Scale Systems Mach 17 MIT, Sping 004 Handout #17 Lectue Note 13 1 Tempoal-Diffeence Leaning We now conside the poblem of computing an appopiate paamete, so that, given an appoximation

More information

Conservative Averaging Method and its Application for One Heat Conduction Problem

Conservative Averaging Method and its Application for One Heat Conduction Problem Poceedings of the 4th WSEAS Int. Conf. on HEAT TRANSFER THERMAL ENGINEERING and ENVIRONMENT Elounda Geece August - 6 (pp6-) Consevative Aveaging Method and its Application fo One Heat Conduction Poblem

More information

COLLAPSING WALLS THEOREM

COLLAPSING WALLS THEOREM COLLAPSING WALLS THEOREM IGOR PAK AND ROM PINCHASI Abstact. Let P R 3 be a pyamid with the base a convex polygon Q. We show that when othe faces ae collapsed (otated aound the edges onto the plane spanned

More information

Directed Regression. Benjamin Van Roy Stanford University Stanford, CA Abstract

Directed Regression. Benjamin Van Roy Stanford University Stanford, CA Abstract Diected Regession Yi-hao Kao Stanfod Univesity Stanfod, CA 94305 yihaoao@stanfod.edu Benjamin Van Roy Stanfod Univesity Stanfod, CA 94305 bv@stanfod.edu Xiang Yan Stanfod Univesity Stanfod, CA 94305 xyan@stanfod.edu

More information

Computers and Mathematics with Applications

Computers and Mathematics with Applications Computes and Mathematics with Applications 58 (009) 9 7 Contents lists available at ScienceDiect Computes and Mathematics with Applications jounal homepage: www.elsevie.com/locate/camwa Bi-citeia single

More information

Encapsulation theory: radial encapsulation. Edmund Kirwan *

Encapsulation theory: radial encapsulation. Edmund Kirwan * Encapsulation theoy: adial encapsulation. Edmund Kiwan * www.edmundkiwan.com Abstact This pape intoduces the concept of adial encapsulation, wheeby dependencies ae constained to act fom subsets towads

More information

Identification of the degradation of railway ballast under a concrete sleeper

Identification of the degradation of railway ballast under a concrete sleeper Identification of the degadation of ailway ballast unde a concete sleepe Qin Hu 1) and Heung Fai Lam ) 1), ) Depatment of Civil and Achitectual Engineeing, City Univesity of Hong Kong, Hong Kong SAR, China.

More information

STABILITY AND PARAMETER SENSITIVITY ANALYSES OF AN INDUCTION MOTOR

STABILITY AND PARAMETER SENSITIVITY ANALYSES OF AN INDUCTION MOTOR HUNGARIAN JOURNAL OF INDUSTRY AND CHEMISTRY VESZPRÉM Vol. 42(2) pp. 109 113 (2014) STABILITY AND PARAMETER SENSITIVITY ANALYSES OF AN INDUCTION MOTOR ATTILA FODOR 1, ROLAND BÁLINT 1, ATTILA MAGYAR 1, AND

More information

Application of homotopy perturbation method to the Navier-Stokes equations in cylindrical coordinates

Application of homotopy perturbation method to the Navier-Stokes equations in cylindrical coordinates Computational Ecology and Softwae 5 5(): 9-5 Aticle Application of homotopy petubation method to the Navie-Stokes equations in cylindical coodinates H. A. Wahab Anwa Jamal Saia Bhatti Muhammad Naeem Muhammad

More information

Do Managers Do Good With Other People s Money? Online Appendix

Do Managers Do Good With Other People s Money? Online Appendix Do Manages Do Good With Othe People s Money? Online Appendix Ing-Haw Cheng Haison Hong Kelly Shue Abstact This is the Online Appendix fo Cheng, Hong and Shue 2013) containing details of the model. Datmouth

More information

This is a very simple sampling mode, and this article propose an algorithm about how to recover x from y in this condition.

This is a very simple sampling mode, and this article propose an algorithm about how to recover x from y in this condition. 3d Intenational Confeence on Multimedia echnology(icm 03) A Simple Compessive Sampling Mode and the Recovey of Natue Images Based on Pixel Value Substitution Wenping Shao, Lin Ni Abstact: Compessive Sampling

More information

J. Electrical Systems 1-3 (2005): Regular paper

J. Electrical Systems 1-3 (2005): Regular paper K. Saii D. Rahem S. Saii A Miaoui Regula pape Coupled Analytical-Finite Element Methods fo Linea Electomagnetic Actuato Analysis JES Jounal of Electical Systems In this pape, a linea electomagnetic actuato

More information

International Journal of Mathematical Archive-3(12), 2012, Available online through ISSN

International Journal of Mathematical Archive-3(12), 2012, Available online through  ISSN Intenational Jounal of Mathematical Achive-3(), 0, 480-4805 Available online though www.ijma.info ISSN 9 504 STATISTICAL QUALITY CONTROL OF MULTI-ITEM EOQ MOEL WITH VARYING LEAING TIME VIA LAGRANGE METHO

More information

The Substring Search Problem

The Substring Search Problem The Substing Seach Poblem One algoithm which is used in a vaiety of applications is the family of substing seach algoithms. These algoithms allow a use to detemine if, given two chaacte stings, one is

More information

ASTR415: Problem Set #6

ASTR415: Problem Set #6 ASTR45: Poblem Set #6 Cuan D. Muhlbege Univesity of Mayland (Dated: May 7, 27) Using existing implementations of the leapfog and Runge-Kutta methods fo solving coupled odinay diffeential equations, seveal

More information

Chapter 5 Linear Equations: Basic Theory and Practice

Chapter 5 Linear Equations: Basic Theory and Practice Chapte 5 inea Equations: Basic Theoy and actice In this chapte and the next, we ae inteested in the linea algebaic equation AX = b, (5-1) whee A is an m n matix, X is an n 1 vecto to be solved fo, and

More information

15.081J/6.251J Introduction to Mathematical Programming. Lecture 6: The Simplex Method II

15.081J/6.251J Introduction to Mathematical Programming. Lecture 6: The Simplex Method II 15081J/6251J Intoduction to Mathematical Pogamming ectue 6: The Simplex Method II 1 Outline Revised Simplex method Slide 1 The full tableau implementation Anticycling 2 Revised Simplex Initial data: A,

More information

arxiv: v1 [math.na] 8 Feb 2013

arxiv: v1 [math.na] 8 Feb 2013 A mixed method fo Diichlet poblems with adial basis functions axiv:1302.2079v1 [math.na] 8 Feb 2013 Nobet Heue Thanh Tan Abstact We pesent a simple discetization by adial basis functions fo the Poisson

More information

EM Boundary Value Problems

EM Boundary Value Problems EM Bounday Value Poblems 10/ 9 11/ By Ilekta chistidi & Lee, Seung-Hyun A. Geneal Desciption : Maxwell Equations & Loentz Foce We want to find the equations of motion of chaged paticles. The way to do

More information

Likelihood vs. Information in Aligning Biopolymer Sequences. UCSD Technical Report CS Timothy L. Bailey

Likelihood vs. Information in Aligning Biopolymer Sequences. UCSD Technical Report CS Timothy L. Bailey Likelihood vs. Infomation in Aligning Biopolyme Sequences UCSD Technical Repot CS93-318 Timothy L. Bailey Depatment of Compute Science and Engineeing Univesity of Califonia, San Diego 1 Febuay, 1993 ABSTRACT:

More information

Vector d is a linear vector function of vector d when the following relationships hold:

Vector d is a linear vector function of vector d when the following relationships hold: Appendix 4 Dyadic Analysis DEFINITION ecto d is a linea vecto function of vecto d when the following elationships hold: d x = a xxd x + a xy d y + a xz d z d y = a yxd x + a yy d y + a yz d z d z = a zxd

More information

Hammerstein Model Identification Based On Instrumental Variable and Least Square Methods

Hammerstein Model Identification Based On Instrumental Variable and Least Square Methods Intenational Jounal of Emeging Tends & Technology in Compute Science (IJETTCS) Volume 2, Issue, Januay Febuay 23 ISSN 2278-6856 Hammestein Model Identification Based On Instumental Vaiable and Least Squae

More information

CASCADE OPTIMIZATION AND CONTROL OF BATCH REACTORS

CASCADE OPTIMIZATION AND CONTROL OF BATCH REACTORS CASCADE OPIMIZAION AND CONROL OF BACH REACORS Xiangming Hua, Sohab Rohani and Athu Jutan* Depatment of Chemical and Biochemical Engineeing Univesity of Westen Ontaio, London, Canada N6A B9 * ajutan@uwo.ca

More information

The Congestion of n-cube Layout on a Rectangular Grid S.L. Bezrukov J.D. Chavez y L.H. Harper z M. Rottger U.-P. Schroeder Abstract We consider the pr

The Congestion of n-cube Layout on a Rectangular Grid S.L. Bezrukov J.D. Chavez y L.H. Harper z M. Rottger U.-P. Schroeder Abstract We consider the pr The Congestion of n-cube Layout on a Rectangula Gid S.L. Bezukov J.D. Chavez y L.H. Hape z M. Rottge U.-P. Schoede Abstact We conside the poblem of embedding the n-dimensional cube into a ectangula gid

More information

Geometry of the homogeneous and isotropic spaces

Geometry of the homogeneous and isotropic spaces Geomety of the homogeneous and isotopic spaces H. Sonoda Septembe 2000; last evised Octobe 2009 Abstact We summaize the aspects of the geomety of the homogeneous and isotopic spaces which ae most elevant

More information

Aalborg Universitet. Load Estimation from Natural input Modal Analysis Aenlle, Manuel López; Brincker, Rune; Canteli, Alfonso Fernández

Aalborg Universitet. Load Estimation from Natural input Modal Analysis Aenlle, Manuel López; Brincker, Rune; Canteli, Alfonso Fernández Aalbog Univesitet Load Estimation fom atual input Modal Analysis Aenlle, Manuel López; Bincke, Rune; Canteli, Alfonso Fenández Published in: Confeence Poceedings Publication date: 005 Document Vesion Publishe's

More information

Fall 2014 Randomized Algorithms Oct 8, Lecture 3

Fall 2014 Randomized Algorithms Oct 8, Lecture 3 Fall 204 Randomized Algoithms Oct 8, 204 Lectue 3 Pof. Fiedich Eisenband Scibes: Floian Tamè In this lectue we will be concened with linea pogamming, in paticula Clakson s Las Vegas algoithm []. The main

More information

An Exact Solution of Navier Stokes Equation

An Exact Solution of Navier Stokes Equation An Exact Solution of Navie Stokes Equation A. Salih Depatment of Aeospace Engineeing Indian Institute of Space Science and Technology, Thiuvananthapuam, Keala, India. July 20 The pincipal difficulty in

More information

How to Obtain Desirable Transfer Functions in MIMO Systems Under Internal Stability Using Open and Closed Loop Control

How to Obtain Desirable Transfer Functions in MIMO Systems Under Internal Stability Using Open and Closed Loop Control How to Obtain Desiable ansfe Functions in MIMO Sstems Unde Intenal Stabilit Using Open and losed Loop ontol echnical Repot of the ISIS Goup at the Univesit of Note Dame ISIS-03-006 June, 03 Panos J. Antsaklis

More information

arxiv: v2 [physics.data-an] 15 Jul 2015

arxiv: v2 [physics.data-an] 15 Jul 2015 Limitation of the Least Squae Method in the Evaluation of Dimension of Factal Bownian Motions BINGQIANG QIAO,, SIMING LIU, OUDUN ZENG, XIANG LI, and BENZONG DAI Depatment of Physics, Yunnan Univesity,

More information

MATH 220: SECOND ORDER CONSTANT COEFFICIENT PDE. We consider second order constant coefficient scalar linear PDEs on R n. These have the form

MATH 220: SECOND ORDER CONSTANT COEFFICIENT PDE. We consider second order constant coefficient scalar linear PDEs on R n. These have the form MATH 220: SECOND ORDER CONSTANT COEFFICIENT PDE ANDRAS VASY We conside second ode constant coefficient scala linea PDEs on R n. These have the fom Lu = f L = a ij xi xj + b i xi + c i whee a ij b i and

More information

FE FORMULATIONS FOR PLASTICITY

FE FORMULATIONS FOR PLASTICITY G These slides ae designed based on the book: Finite Elements in Plasticity Theoy and Pactice, D.R.J. Owen and E. Hinton, 970, Pineidge Pess Ltd., Swansea, UK. Couse Content: A INTRODUCTION AND OVERVIEW

More information

A Converse to Low-Rank Matrix Completion

A Converse to Low-Rank Matrix Completion A Convese to Low-Rank Matix Completion Daniel L. Pimentel-Alacón, Robet D. Nowak Univesity of Wisconsin-Madison Abstact In many pactical applications, one is given a subset Ω of the enties in a d N data

More information

arxiv: v1 [math.co] 1 Apr 2011

arxiv: v1 [math.co] 1 Apr 2011 Weight enumeation of codes fom finite spaces Relinde Juius Octobe 23, 2018 axiv:1104.0172v1 [math.co] 1 Ap 2011 Abstact We study the genealized and extended weight enumeato of the - ay Simplex code and

More information

by numerous studies); ii) MSWT spectrum is symmetric with respect to point while numerical methods give asymmetrical spectrum with gap = = 2.

by numerous studies); ii) MSWT spectrum is symmetric with respect to point while numerical methods give asymmetrical spectrum with gap = = 2. SPIN-WAVE THEORY FOR S= ANTIFERROMAGNETIC ISOTROPIC CHAIN D. V. Spiin V.I. Venadsii Tauida National Univesity, Yaltinsaya st. 4, Simfeopol, 957, Cimea, Uaine E-mail: spiin@cimea.edu, spiin@tnu.cimea.ua

More information

SOME GENERAL NUMERICAL RADIUS INEQUALITIES FOR THE OFF-DIAGONAL PARTS OF 2 2 OPERATOR MATRICES

SOME GENERAL NUMERICAL RADIUS INEQUALITIES FOR THE OFF-DIAGONAL PARTS OF 2 2 OPERATOR MATRICES italian jounal of pue and applied mathematics n. 35 015 (433 44) 433 SOME GENERAL NUMERICAL RADIUS INEQUALITIES FOR THE OFF-DIAGONAL PARTS OF OPERATOR MATRICES Watheq Bani-Domi Depatment of Mathematics

More information

A NEW VARIABLE STIFFNESS SPRING USING A PRESTRESSED MECHANISM

A NEW VARIABLE STIFFNESS SPRING USING A PRESTRESSED MECHANISM Poceedings of the ASME 2010 Intenational Design Engineeing Technical Confeences & Computes and Infomation in Engineeing Confeence IDETC/CIE 2010 August 15-18, 2010, Monteal, Quebec, Canada DETC2010-28496

More information

On Polynomials Construction

On Polynomials Construction Intenational Jounal of Mathematical Analysis Vol., 08, no. 6, 5-57 HIKARI Ltd, www.m-hikai.com https://doi.og/0.988/ima.08.843 On Polynomials Constuction E. O. Adeyefa Depatment of Mathematics, Fedeal

More information

LET a random variable x follows the two - parameter

LET a random variable x follows the two - parameter INTERNATIONAL JOURNAL OF MATHEMATICS AND SCIENTIFIC COMPUTING ISSN: 2231-5330, VOL. 5, NO. 1, 2015 19 Shinkage Bayesian Appoach in Item - Failue Gamma Data In Pesence of Pio Point Guess Value Gyan Pakash

More information

A Comparison and Contrast of Some Methods for Sample Quartiles

A Comparison and Contrast of Some Methods for Sample Quartiles A Compaison and Contast of Some Methods fo Sample Quatiles Anwa H. Joade and aja M. Latif King Fahd Univesity of Petoleum & Mineals ABSTACT A emainde epesentation of the sample size n = 4m ( =, 1, 2, 3)

More information

Hydroelastic Analysis of a 1900 TEU Container Ship Using Finite Element and Boundary Element Methods

Hydroelastic Analysis of a 1900 TEU Container Ship Using Finite Element and Boundary Element Methods TEAM 2007, Sept. 10-13, 2007,Yokohama, Japan Hydoelastic Analysis of a 1900 TEU Containe Ship Using Finite Element and Bounday Element Methods Ahmet Egin 1)*, Levent Kaydıhan 2) and Bahadı Uğulu 3) 1)

More information

F-IF Logistic Growth Model, Abstract Version

F-IF Logistic Growth Model, Abstract Version F-IF Logistic Gowth Model, Abstact Vesion Alignments to Content Standads: F-IFB4 Task An impotant example of a model often used in biology o ecology to model population gowth is called the logistic gowth

More information

A dual-reciprocity boundary element method for axisymmetric thermoelastodynamic deformations in functionally graded solids

A dual-reciprocity boundary element method for axisymmetric thermoelastodynamic deformations in functionally graded solids APCOM & ISCM 11-14 th Decembe, 013, Singapoe A dual-ecipocity bounday element method fo axisymmetic themoelastodynamic defomations in functionally gaded solids *W. T. Ang and B. I. Yun Division of Engineeing

More information

Stanford University CS259Q: Quantum Computing Handout 8 Luca Trevisan October 18, 2012

Stanford University CS259Q: Quantum Computing Handout 8 Luca Trevisan October 18, 2012 Stanfod Univesity CS59Q: Quantum Computing Handout 8 Luca Tevisan Octobe 8, 0 Lectue 8 In which we use the quantum Fouie tansfom to solve the peiod-finding poblem. The Peiod Finding Poblem Let f : {0,...,

More information

Research Design - - Topic 17 Multiple Regression & Multiple Correlation: Two Predictors 2009 R.C. Gardner, Ph.D.

Research Design - - Topic 17 Multiple Regression & Multiple Correlation: Two Predictors 2009 R.C. Gardner, Ph.D. Reseach Design - - Topic 7 Multiple Regession & Multiple Coelation: Two Pedictos 009 R.C. Gadne, Ph.D. Geneal Rationale and Basic Aithmetic fo two pedictos Patial and semipatial coelation Regession coefficients

More information

On the Quasi-inverse of a Non-square Matrix: An Infinite Solution

On the Quasi-inverse of a Non-square Matrix: An Infinite Solution Applied Mathematical Sciences, Vol 11, 2017, no 27, 1337-1351 HIKARI Ltd, wwwm-hikaicom https://doiog/1012988/ams20177273 On the Quasi-invese of a Non-squae Matix: An Infinite Solution Ruben D Codeo J

More information

ON INDEPENDENT SETS IN PURELY ATOMIC PROBABILITY SPACES WITH GEOMETRIC DISTRIBUTION. 1. Introduction. 1 r r. r k for every set E A, E \ {0},

ON INDEPENDENT SETS IN PURELY ATOMIC PROBABILITY SPACES WITH GEOMETRIC DISTRIBUTION. 1. Introduction. 1 r r. r k for every set E A, E \ {0}, ON INDEPENDENT SETS IN PURELY ATOMIC PROBABILITY SPACES WITH GEOMETRIC DISTRIBUTION E. J. IONASCU and A. A. STANCU Abstact. We ae inteested in constucting concete independent events in puely atomic pobability

More information

ANA BERRIZBEITIA, LUIS A. MEDINA, ALEXANDER C. MOLL, VICTOR H. MOLL, AND LAINE NOBLE

ANA BERRIZBEITIA, LUIS A. MEDINA, ALEXANDER C. MOLL, VICTOR H. MOLL, AND LAINE NOBLE THE p-adic VALUATION OF STIRLING NUMBERS ANA BERRIZBEITIA, LUIS A. MEDINA, ALEXANDER C. MOLL, VICTOR H. MOLL, AND LAINE NOBLE Abstact. Let p > 2 be a pime. The p-adic valuation of Stiling numbes of the

More information

Research Article On Alzer and Qiu s Conjecture for Complete Elliptic Integral and Inverse Hyperbolic Tangent Function

Research Article On Alzer and Qiu s Conjecture for Complete Elliptic Integral and Inverse Hyperbolic Tangent Function Abstact and Applied Analysis Volume 011, Aticle ID 697547, 7 pages doi:10.1155/011/697547 Reseach Aticle On Alze and Qiu s Conjectue fo Complete Elliptic Integal and Invese Hypebolic Tangent Function Yu-Ming

More information

APPLICATION OF MAC IN THE FREQUENCY DOMAIN

APPLICATION OF MAC IN THE FREQUENCY DOMAIN PPLICION OF MC IN HE FREQUENCY DOMIN D. Fotsch and D. J. Ewins Dynamics Section, Mechanical Engineeing Depatment Impeial College of Science, echnology and Medicine London SW7 2B, United Kingdom BSRC he

More information

Numerical Solution of Boundary Value Problems for the Laplacian in R 3 in the Case of Complex Boundary Surface

Numerical Solution of Boundary Value Problems for the Laplacian in R 3 in the Case of Complex Boundary Surface Computational Applied Mathematics Jounal 5; (: 9-5 Published online Febuay 5 (http://www.aascit.og/jounal/camj Numeical Solution of Bounday Value Poblems fo the Laplacian in R in the Case of Complex Bounday

More information

4/18/2005. Statistical Learning Theory

4/18/2005. Statistical Learning Theory Statistical Leaning Theoy Statistical Leaning Theoy A model of supevised leaning consists of: a Envionment - Supplying a vecto x with a fixed but unknown pdf F x (x b Teache. It povides a desied esponse

More information

Discrete LQ optimal control with integral action: A simple controller on incremental form for MIMO systems

Discrete LQ optimal control with integral action: A simple controller on incremental form for MIMO systems Modeling, Identification and Contol, Vol., No., 1, pp. 5, ISSN 189 18 Discete LQ optimal contol with integal action: A simple contolle on incemental fom fo MIMO systems David Di Ruscio Telemak Univesity

More information

A Bijective Approach to the Permutational Power of a Priority Queue

A Bijective Approach to the Permutational Power of a Priority Queue A Bijective Appoach to the Pemutational Powe of a Pioity Queue Ia M. Gessel Kuang-Yeh Wang Depatment of Mathematics Bandeis Univesity Waltham, MA 02254-9110 Abstact A pioity queue tansfoms an input pemutation

More information

Functions Defined on Fuzzy Real Numbers According to Zadeh s Extension

Functions Defined on Fuzzy Real Numbers According to Zadeh s Extension Intenational Mathematical Foum, 3, 2008, no. 16, 763-776 Functions Defined on Fuzzy Real Numbes Accoding to Zadeh s Extension Oma A. AbuAaqob, Nabil T. Shawagfeh and Oma A. AbuGhneim 1 Mathematics Depatment,

More information

Asymptotically Lacunary Statistical Equivalent Sequence Spaces Defined by Ideal Convergence and an Orlicz Function

Asymptotically Lacunary Statistical Equivalent Sequence Spaces Defined by Ideal Convergence and an Orlicz Function "Science Stays Tue Hee" Jounal of Mathematics and Statistical Science, 335-35 Science Signpost Publishing Asymptotically Lacunay Statistical Equivalent Sequence Spaces Defined by Ideal Convegence and an

More information

( ) [ ] [ ] [ ] δf φ = F φ+δφ F. xdx.

( ) [ ] [ ] [ ] δf φ = F φ+δφ F. xdx. 9. LAGRANGIAN OF THE ELECTROMAGNETIC FIELD In the pevious section the Lagangian and Hamiltonian of an ensemble of point paticles was developed. This appoach is based on a qt. This discete fomulation can

More information

An Adaptive Neural-Network Model-Following Speed Control of PMSM Drives for Electric Vehicle Applications

An Adaptive Neural-Network Model-Following Speed Control of PMSM Drives for Electric Vehicle Applications Poceedings of the 9th WSEAS Intenational Confeence on Applied Mathematics, Istanbul, Tuey, May 27-29, 2006 (pp412-417) An Adaptive Neual-Netwo Model-Following Speed Contol of PMSM Dives fo Electic Vehicle

More information

A FURTHER SUBSPACE METHOD OPTIMIZATION FOR TRACKING REAL VALUED SINUSOIDS IN NOISE

A FURTHER SUBSPACE METHOD OPTIMIZATION FOR TRACKING REAL VALUED SINUSOIDS IN NOISE Électonique et tansmission de l infomation A FURHER SUBSPACE MEHOD OPIMIZAION FOR RACKING REAL VALUED SINUSOIDS IN NOISE ŞEFAN SLAVNICU, SILVIU CIOCHINĂ Key wods: Subspace tacking, Fequency estimation,

More information

Duality between Statical and Kinematical Engineering Systems

Duality between Statical and Kinematical Engineering Systems Pape 00, Civil-Comp Ltd., Stiling, Scotland Poceedings of the Sixth Intenational Confeence on Computational Stuctues Technology, B.H.V. Topping and Z. Bittna (Editos), Civil-Comp Pess, Stiling, Scotland.

More information

CALCULATING THE NUMBER OF TWIN PRIMES WITH SPECIFIED DISTANCE BETWEEN THEM BASED ON THE SIMPLEST PROBABILISTIC MODEL

CALCULATING THE NUMBER OF TWIN PRIMES WITH SPECIFIED DISTANCE BETWEEN THEM BASED ON THE SIMPLEST PROBABILISTIC MODEL U.P.B. Sci. Bull. Seies A, Vol. 80, Iss.3, 018 ISSN 13-707 CALCULATING THE NUMBER OF TWIN PRIMES WITH SPECIFIED DISTANCE BETWEEN THEM BASED ON THE SIMPLEST PROBABILISTIC MODEL Sasengali ABDYMANAPOV 1,

More information

A Multivariate Normal Law for Turing s Formulae

A Multivariate Normal Law for Turing s Formulae A Multivaiate Nomal Law fo Tuing s Fomulae Zhiyi Zhang Depatment of Mathematics and Statistics Univesity of Noth Caolina at Chalotte Chalotte, NC 28223 Abstact This pape establishes a sufficient condition

More information

Multiple Criteria Secretary Problem: A New Approach

Multiple Criteria Secretary Problem: A New Approach J. Stat. Appl. Po. 3, o., 9-38 (04 9 Jounal of Statistics Applications & Pobability An Intenational Jounal http://dx.doi.og/0.785/jsap/0303 Multiple Citeia Secetay Poblem: A ew Appoach Alaka Padhye, and

More information

Chapter 3 Optical Systems with Annular Pupils

Chapter 3 Optical Systems with Annular Pupils Chapte 3 Optical Systems with Annula Pupils 3 INTRODUCTION In this chapte, we discuss the imaging popeties of a system with an annula pupil in a manne simila to those fo a system with a cicula pupil The

More information

Nuclear size corrections to the energy levels of single-electron atoms

Nuclear size corrections to the energy levels of single-electron atoms Nuclea size coections to the enegy levels of single-electon atoms Babak Nadii Nii a eseach Institute fo Astonomy and Astophysics of Maagha (IAAM IAN P. O. Box: 554-44. Abstact A study is made of nuclea

More information

COMP Parallel Computing SMM (3) OpenMP Case Study: The Barnes-Hut N-body Algorithm

COMP Parallel Computing SMM (3) OpenMP Case Study: The Barnes-Hut N-body Algorithm COMP 633 - Paallel Computing Lectue 8 Septembe 14, 2017 SMM (3) OpenMP Case Study: The Banes-Hut N-body Algoithm Topics Case study: the Banes-Hut algoithm Study an impotant algoithm in scientific computing»

More information

Moment-free numerical approximation of highly oscillatory integrals with stationary points

Moment-free numerical approximation of highly oscillatory integrals with stationary points Moment-fee numeical appoximation of highly oscillatoy integals with stationay points Sheehan Olve Abstact We pesent a method fo the numeical quadatue of highly oscillatoy integals with stationay points.

More information

EVOLUTIONARY COMPUTING FOR METALS PROPERTIES MODELLING

EVOLUTIONARY COMPUTING FOR METALS PROPERTIES MODELLING EVOLUTIONARY COMPUTING FOR METALS PROPERTIES MODELLING M.F. Abbod*, M. Mahfouf, D.A. Linkens and Sellas, C.M. IMMPETUS Institute fo Micostuctue and Mechanical Popeties Engineeing, The Univesity of Sheffield

More information

A Backward Identification Problem for an Axis-Symmetric Fractional Diffusion Equation

A Backward Identification Problem for an Axis-Symmetric Fractional Diffusion Equation Mathematical Modelling and Analysis Publishe: Taylo&Fancis and VGTU Volume 22 Numbe 3, May 27, 3 32 http://www.tandfonline.com/tmma https://doi.og/.3846/3926292.27.39329 ISSN: 392-6292 c Vilnius Gediminas

More information

PROBLEM SET #1 SOLUTIONS by Robert A. DiStasio Jr.

PROBLEM SET #1 SOLUTIONS by Robert A. DiStasio Jr. POBLM S # SOLUIONS by obet A. DiStasio J. Q. he Bon-Oppenheime appoximation is the standad way of appoximating the gound state of a molecula system. Wite down the conditions that detemine the tonic and

More information

MAGNETIC FIELD AROUND TWO SEPARATED MAGNETIZING COILS

MAGNETIC FIELD AROUND TWO SEPARATED MAGNETIZING COILS The 8 th Intenational Confeence of the Slovenian Society fo Non-Destuctive Testing»pplication of Contempoay Non-Destuctive Testing in Engineeing«Septembe 1-3, 5, Potoož, Slovenia, pp. 17-1 MGNETIC FIELD

More information

A Relativistic Electron in a Coulomb Potential

A Relativistic Electron in a Coulomb Potential A Relativistic Electon in a Coulomb Potential Alfed Whitehead Physics 518, Fall 009 The Poblem Solve the Diac Equation fo an electon in a Coulomb potential. Identify the conseved quantum numbes. Specify

More information

C e f paamete adaptation f (' x) ' ' d _ d ; ; e _e K p K v u ^M() RBF NN ^h( ) _ obot s _ s n W ' f x x xm xm f x xm d Figue : Block diagam of comput

C e f paamete adaptation f (' x) ' ' d _ d ; ; e _e K p K v u ^M() RBF NN ^h( ) _ obot s _ s n W ' f x x xm xm f x xm d Figue : Block diagam of comput A Neual-Netwok Compensato with Fuzzy Robustication Tems fo Impoved Design of Adaptive Contol of Robot Manipulatos Y.H. FUNG and S.K. TSO Cente fo Intelligent Design, Automation and Manufactuing City Univesity

More information

A Deep Convolutional Neural Network Based on Nested Residue Number System

A Deep Convolutional Neural Network Based on Nested Residue Number System A Deep Convolutional Neual Netwok Based on Nested Residue Numbe System Hioki Nakahaa Ehime Univesity, Japan Tsutomu Sasao Meiji Univesity, Japan Abstact A pe-tained deep convolutional neual netwok (DCNN)

More information

This aticle was oiginally published in a jounal published by Elsevie, the attached copy is povided by Elsevie fo the autho s benefit fo the benefit of the autho s institution, fo non-commecial eseach educational

More information

Numerical Integration

Numerical Integration MCEN 473/573 Chapte 0 Numeical Integation Fall, 2006 Textbook, 0.4 and 0.5 Isopaametic Fomula Numeical Integation [] e [ ] T k = h B [ D][ B] e B Jdsdt In pactice, the element stiffness is calculated numeically.

More information