Constrained optimal discrimination designs for Fourier regression models

Size: px
Start display at page:

Download "Constrained optimal discrimination designs for Fourier regression models"

Transcription

1 Ann Inst Stat Math (29) 6:43 57 DOI.7/s Constraine optimal iscrimination esigns for Fourier regression moels Stefanie Bieermann Holger Dette Philipp Hoffmann Receive: 26 June 26 / Revise: 9 March 27 / Publishe online: 7 July 27 The Institute of Statistical Mathematics, Tokyo 27 Abstract In this article, the problem of constructing efficient iscrimination esigns in a Fourier regression moel is consiere. We propose esigns which maximize the power of the F-test, which iscriminates between the two highest orer moels, subject to the constraints that the tests that iscriminate between lower orer moels have at least some given relative power. A complete solution is presente in terms of the canonical moments of the optimal esigns, an for the special case of equal constraints even more specific formulae are available. Keywors Constraine optimal esigns Trigonometric regression D -optimal esigns Chebyshev polynomials Canonical moments Introuction The Fourier regression or trigonometric regression moel g 2 (x) = a + g 2 (x) = a + a j sin( jx) + j= j= b j cos( jx), x [ π, π], () j= a j sin( jx) + b j cos( jx), x [ π, π], (2) j= S. Bieermann University of Southampton, School of Mathematics, Highfiel SO7 BJ, UK H. Dette (B) P. Hoffmann Ruhr-Universität Bochum, Fakultät für Mathematik, 4478 Bochum, Germany holger.ette@ruhr-uni-bochum.e; holger.ette@rub.e

2 44 S. Bieermann et al. where IN, is use to escribe perioic phenomena (see, e.g., Maria (972), Kitsos et al. (988) or the collection of research papers in biology eite by Lestrel (997)). Moreover, there are several applications of trigonometric regression moels in two-imensional shape analysis in biology. We refer to Younker an Ehrlich (977) an Currie et al. (2) for concrete examples. The value 2 in () or2 in(2) is usually enote as the egree of the Fourier regression moel. The coefficients a, a,...,a, b,...,b enote unknown parameters, which have to be estimate from the ata. The problem of esigning experiments for moels of the form () has been iscusse by several authors; see, e.g., Karlin an Stuen (966), p. 347, Feorov (972), p. 94, Hill (978), Lau an Stuen (985) for optimal esigns on the full circle, as well as Dette et al. (22a) an Dette et al. (22b) for optimal esigns on a partial circle. Most authors concentrate on the problem of etermining optimal esigns for the estimation of the full vector of unknown parameters, whereas the problem of constructing optimal esigns for moel iscrimination has been consiere by Dette an Haller (998),Dette an Melas (23) an Zen an Tsai (24). The present paper is evote to the problem of constructing optimal iscrimination esigns using constraine optimality criteria. Constraine optimal esigns have primarily been consiere by Stigler (97), Stuen (982b) an Lee (988a,b), whereas Cook an Wong (994), Dette (995) an Clye an Chaloner (996) investigate the relation between this approach an compoun optimality criteria. Although these results are interesting from a theoretical point of view constraine optimal esigns still ha to be foun numerically an explicit results coul only be inferre in rare cases. In particular, it turns out that there is a one-to-one corresponence between compoun an constraine optimal esigns, which, however, can only be exploite in rare cases to fin the constraine optimal esign from the corresponing compoun optimal esign, which is usually much simpler to calculate. Dette an Franke (2) characterize constraine optimal iscriminating esigns for polynomial regression moels utilizing the theory of canonical moments, which was introuce by Skibinsky (967) an applie by Stuen (98, 982a,b, 989) for etermining optimal esigns in polynomial regression moels. The problem of fining constraine optimal iscriminating esigns for Fourier regression moels, however, has not been consiere yet. The present paper is evote to this problem. For the construction of constraine optimal esigns we assume that the highest frequency of the moel has been fixe an etermine the esign such that the coefficient corresponing to this frequency is estimate with maximal efficiency subject to the constraints that the coefficients corresponing to the highest frequencies in the moels of lower egree can be estimate with some guarantee efficiency. The optimality criterion is carefully escribe in Sect. 2. In Sect. 3 we briefly review some facts from the theory of canonical moments (see Dette an Stuen 997), which is the basic tool for the construction of optimal iscrimination esigns. A complete characterization of the constraine optimal escriminating esigns is given in terms of their canonical moments, an in the special case of equal bouns we further specify the optimal esigns in terms of their supporting polynomials an explicit formulae for the weights. Finally, in Sect. 4 we give some concluing remarks iscussing our approach.

3 Discrimination esigns for Fourier regression moels 45 2 Constraine optimal esigns in Fourier regression moels We assume that we can make n 2 + inepenent observations Y,...,Y n where Y i N (g(x i ), σ 2 ), an the regression function g(x) belongs to the class of trigonometric moels {g, g,...,g 2 } where g 2l an g 2l are efine in () an (2), respectively. For k =,...,2 we efine { (, sin(x), cos(x),...,sin( jx), cos( jx)) f k (x)= T, if k = 2 j (, sin(x), cos(x),...,sin(( j )x), cos(( j )x), sin( jx)) T, if k =2 j an θ k = { (a, a, b,...,a j, b j ) T if k = 2 j (a, a, b,...,a j, b j, a j ) T if k = 2 j. Then the moels g k (x) where k =,...,2, can be written as g k (x) = f k (x) T θ k. An approximate esign is a probability measure σ with finite support on the interval [ π, π] with the interpretation that observations are taken at the support points in proportion to the corresponing masses. The information matrix in the Fourier regression moel g k (x) is given by M k (σ ) = π π f k (x) fk T (x)σ(x). (3) An optimal esign maximizes an appropriate information function of the informationmatrix(seepukelsheim 993, p. 3). There are numerous criteria which can be use for the characterization of efficient esigns. Most of these criteria focus on precise parameter estimation in a moel of given egree. In many practical situations, however, it is not known before the experiment, up to which egree a Fourier regression moel shoul be fitte. As sparse moeling is avisable we turn our attention to esigns that allow successive testing of the higher orer coefficients with high power, thus guaranteeing goo iscrimination properties of the testing proceure. Our optimality criterion for constructing iscrimination esigns is therefore base on a multiple F-test, where, starting with the given regression moel g 2 (x) in () one tests the hypotheses H (2) : b =, H (2 ) : a =, H (2 2) : b =, H (2 3) : a =,..., H () : a =, in the moels g 2, g 2,...,g, successively, an ecies for the moel g k where k is the first inex for which the hypothesis H (k ) : θ k = is rejecte. (Note that this sequence of tests can be stoppe earlier if the minimal egree of the Fourier regression moel is pre-specifie.) The statistical properties of this testing proceure are elaborately presente in Anerson (994). The quantities corresponing to the noncentrality parameter of the F-test for the hypothesis H (k) are given by δ k (σ ) = (e T k M k (σ )e k ) k =,...,2, (4)

4 46 S. Bieermann et al. where e k enotes the (k + )th unit vector in R k+ an the esign σ is assume to have at least (2 + ) support points (see Pukelsheim 993, p. 7). A esign σk is calle D -optimal for the moel g k or D k-optimal if it maximizes δ k.thisisin fact equivalent to maximizing the power of the corresponing test. Collecting ata following a D k -optimal sampling scheme therefore gives the best results for iscriminating between moels g k an g k. To iscriminate between more than two ifferent moels, one has to construct an optimality criterion base on several functions δ k (σ ), k = 2, 2,...,. As these quantities are of ifferent scalings we stanarize them by using the corresponing efficiencies to make them comparable in size. The expression eff k (σ ) := δ k(σ ) δ k (σk k =,...,2 (5) ), is calle the D k -efficiency of the esign σ in the Fourier regression moel g k(x). Dette an Haller (998) propose to maximize a weighte p-mean of the efficiencies eff,...,eff 2 for the construction of an optimal esign for iscriminating between the moels {g,...g 2 }. In the present paper, we consier an alternative optimality criterion to obtain efficient iscriminating esigns. This approach is attractive if the main interest of the experimenter is in iscriminating between the two moels of highest egree, while at the same time the optimal esign shoul allow for an efficient iscrimination between the moels of lower egrees. We consier two criteria for etermining a constraine optimal iscriminating esign for the Fourier regression moel. The first approach consiers the highest cosine frequency as most important an a constraine optimal iscriminating esign σ maximizes eff 2 (σ ) subject to eff k (σ ) c k, k = 2, 2 2,...,2 2 j (6) for some j {,..., }. The secon criterion, however, etermines the esign which maximizes eff 2 (σ ) subject to the constraints eff k (σ ) c k, k =2, 2 2,...,2 2 j, (7) an j {,..., }. For both criteria, the quantities c 2 2 j,...,c 2 (, ) are given by the experimenter reflecting the esire minimal relative power of testing H (k), k = 2 2 j,...,2. A necessary conition for the existence of optimal esigns is c 2 2 j + c 2 2 j. Unlike criterion (6), criterion (7) correspons to the situation where the highest sine frequency is regare as most important, i.e. the power of testing H (2 ) is maximize whereas the preceing test of H (2) (an all the other hypotheses H (k) ) have some prespecifie minimal relative power. The situation where the experimenter prefers to start with moel (2) for some practical reason, an therefore start the testing proceure with H (2 ), can be incorporate in criterion (7) by putting c 2 =. For the solution of the constraine optimization problems (6) an (7) we nee several tools, which will be explaine in what follows.

5 Discrimination esigns for Fourier regression moels 47 It follows by stanar arguments (see Pukelsheim 993, Chap. 4, 5) that eff k is a concave function on the set of esigns on the interval [ π, π] an invariant with respect to a reflection of the esign σ at the origin. Consequently, if there exists a constraine optimal iscriminating esign, then there also exists an optimal esign in the set of all symmetric esigns on the interval [ π, π]. We note that these symmetric esigns inuce esigns ξ σ on the interval [, ] by the projection ξ σ (cos x) = { 2σ(x) = 2σ( x) if < x π σ() if x = (8) for any symmetric esign σ. The corresponing set of the measures ξ σ on [, ] will be enote by [,].ItwasshowninDette an Haller (998) that for any σ δ k (σ ) = where B (ξ σ ) = A (ξ σ ) =, an 2 2( j ) A j (ξ σ ) A j (ξ σ ) if k = 2 j 2 2( j ) B j (ξ σ ) B j (ξ σ ) if k = 2 j (9) ( A k (ξ σ ) = ) k z i+ j ξ σ (z) i, j= () an ( ) k B k (ξ σ ) = ( z 2 )z i+ j ξ σ (z) i, j= () enote the information matrices of the esign ξ σ on the interval [, ] for a homosceastic an a heterosceastic polynomial regression moel with efficiency function λ(z) = ( z 2 ) (see Karlin an Stuen 966), respectively. Consequently, the problem of etermining constraine optimal iscriminating esigns for the Fourier regression moel can be solve by maximizing a certain function over the set of probability measures on the interval [, ] an transforming the maximizing measure back via (8). The problem of maximizing the right han sie of (9) over the set [,] if k = 2 j is in fact the D -optimal esign for the orinary polynomial regression moel, while for o values of k the right han sie of (9) correspons to the weighte polynomial regression with efficiency function σ 2 (x) = σ 2 /( x 2 ), x (, ). The solutions of these problems are well known (see Stuen 968, 982b an yiel δ k (σk ) = max δ k (σ ) = (k =,...,2), an therefore the efficiency of a symmetric esign σ efine in (5) can be rewritten σ as eff k (σ ) = 2 2( j ) A j (ξ σ ) A j (ξ σ ) if k = 2 j 2 2( j ) B j (ξ σ ) B j (ξ σ ) if k = 2 j. (2)

6 48 S. Bieermann et al. 3 The solution of the constraine optimal esign problem For the characterization of the measure ξ σ [,] corresponing to the constraine optimal iscriminating esign σ by the relation (8) we require some basic facts about the theory of canonical moments which has been introuce by Stuen (98, 982a,b) in the context of optimal esign. We will only give a very brief heuristical introuction an refer to the monograph of Dette an Stuen (997)formore etails. It is well known that a probability measure on the interval [, ], say ξ, is etermine by its sequence of moments (m, m 2,...).Skibinsky (967) efine a one to one mapping from the sequences of orinary moments onto sequences (p, p 2,...) whose elements vary inepenently in the interval [, ]. For a given probability measure on the interval [, ] the element p j of the corresponing sequence is calle the jth canonical moment of ξ. If j is the first inex for which p j {, } then the sequence of canonical moments terminates at p j, an the measure is supporte at a finite number of points. The support points an corresponing masses can be foun explicitly by evaluating certain orthogonal polynomials (see Dette an Stuen 997, Chap. 3). The set of probability measures on the interval [, ] with first k canonical moments equal to (p,...,p k ) (, ) k [, ] is a singleton if an only if p k {, }. Otherwise there exists an uncountable number of probability measures corresponing to (p,...,p k ) (see Skibinsky 986). It turns out that the eterminants in (2) can be escribe in terms of the canonical moments p, p 2,...of the measure ξ σ (see Stuen 982b), that is k A k (ξ σ ) =2 k(k+) (q 2l 2 p 2l q 2l p 2l ) k l+ (3) l= k B k (ξ σ ) =2 k(k+) (p 2l 2 q 2l p 2l q 2l ) k l+ (4) l= where p =, q = an q j = p j for j. Observing (2), (3) an (4), we fin that the efficiencies are increasing functions of the terms p 2l q 2l, an consequently the o canonical moments of the optimal projection esign ξ σ satisfy p 2l = 2 l =,...,. (5) Therefore we can restrict ourselves to esigns with this property, an (2) reuces to eff k (σ ) = 2 2 j 2 p 2 j j l= q 2l p 2l if k = 2 j 2 2 j 2 q 2 j j l= q 2l p 2l if k = 2 j where p 2, p 4,... enote the canonical moments of even orer of the esign ξ σ [,] satisfying (5) an corresponing to the measure σ via (8). Our main result gives a characterization of the canonical moments of ξ σ. (6)

7 Discrimination esigns for Fourier regression moels 49 Theorem (a) If there exists a constraine optimal iscriminating esign for the vector (c 2 2 j,...,c 2 ) in (6), then there also exists a symmetric optimal iscriminating esign σ. The canonical moments up to the orer 2 of the corresponing projection ξ σ are etermine by the system of equations p 2n = 2 n =,..., p 2n = 2 n =,..., j max 2, c 2 2 j+2n 2 2n j+n l= j p 2l q 2l, p 2 2 j+2n = max 2, c 2 2 j+2n 2 2n j+n p 2l q 2l, p 2 = c j l= j p 2lq 2l. l= j if c 2 2 j+2n >c 2 2 j+2n if c 2 2 j+2n c 2 2 j+2n n =,..., j (b) If there exists a constraine optimal iscriminating esign for the vector (c 2 2 j,..., c 2 2, c 2 ) in (7), then there also exists a symmetric constraine optimal iscriminating esign σ. The canonical moments up to the orer 2 ofthe corresponing projection ξ σ are etermine by the system of equations p 2n = 2 n =,..., p 2n = 2 n =,..., j max 2, c 2 2 j+2n 2 2n j+n l= j p 2l q 2l, p 2 2 j+2n = max 2, c 2 2 j+2n 2 2n j+n p 2l q 2l, p 2 = c j l= j p 2lq 2l. l= j if c 2 2 j+2n >c 2 2 j+2n if c 2 2 j+2n c 2 2 j+2n n =,..., j A necessary conition for the existence of optimal esigns with respect to either criterion ((6) or (7)) is given by c 2 2 j + c 2 2 j.

8 5 S. Bieermann et al. Proof Because both parts are proven similarly, we restrict ourselves to a proof of part (a). By the previous iscussion the canonical moments of o orer, 3,...,2 must be /2. Note that j+n eff 2 2 j+2n (σ ) = p 2 2 j+2n 2 2( j+n ) l= p 2l q 2l, n =,..., j. In orer to maximize these efficiencies we have to choose the canonical moments such that the proucts p 2l q 2l are as large as possible. This can be accomplishe by choosing p 2l as close as possible to the value /2 such that the constraints in (6) are satisfie. Since there are no restrictions on the efficiencies eff (σ ),..., eff 2 2 j 2 (σ ) we obtain p 2 = = p 2 2 j 2 = 2. Substituting this choice into the formulae for the higher orer efficiencies, (6) reuces to j+n eff 2 2 j+2n (σ ) = q 2 2 j+2n 2 2n l= j j+n eff 2 2 j+2n (σ ) = p 2 2 j+2n 2 2n l= j p 2l q 2l p 2l q 2l. We start with the case n =, for which the representations eff 2 2 j (σ ) = q 2 2 j, eff 2 2 j (σ ) = p 2 2 j yiel the constraints p 2 2 j c 2 2 j, q 2 2 j c 2 2 j. Consequently any esign ξ σ for which p 2 2 j [c 2 2 j, c 2 2 j ] satisfies the constraints of orer 2 2 j an 2 2 j. We therefore assume that c 2 2 j + c 2 2 j in what follows to ensure the existence of such a esign. If 2 [c 2 2 j, c 2 2 j ] one can choose p 2 2 j = 2 to maximize p 2 2 jq 2 2 j.else we have either c 2 2 j 2 or c 2 2 j 2, an we choose p 2 2 j = c 2 2 j or p 2 2 j = c 2 2 j, respectively. If n > we note that the constraints eff 2 2 j+2n (σ ) c 2 2 j+2n an eff 2 2 j+2n (σ ) c 2 2 j+2n reuce to p 2 2 j+2n q 2 2 j+2n c 2 2 j+2n 2 2n =: c j+n 2 2 j+2n p 2l q 2l l= j c 2 2 j+2n 2 2n j+n l= j p 2l q 2l =: c 2 2 j+2n Therefore the same arguments as presente for the case n = yiel the corresponing result for p 2 2 j+2n. Finally we consier the case n = j, where there is only one constraint eff 2 (σ ) c 2, which can be rewritten as p 2 c l= p 2lq 2l.

9 Discrimination esigns for Fourier regression moels 5 In orer to maximize this expression one has to choose p 2 such that there is equality. This proves the final assertion of part (a) in Theorem. Remark Note that Theorem characterizes the canonical moments up to the orer 2 of the projection ξ σ of the (symmetric) constraine optimal iscriminating esign σ. In general p 2 {, } an in these cases there exists an infinite number of probability measures on the interval [, ] with the canonical moments p,...p 2 (see Skibinsky 986). Each of these measures correspons to a constraine optimal iscriminating esign by the projection (8). One possible choice among these esigns with a reasonable small support will be illustrate later in the proof of Theorem 2, Eq. (24). In what follows, we present two further results, where a solution of the constraine optimal esign problem can be foun explicitly. For this purpose let T j (x) an U j (x) enote the jth Chebyshev polynomial of the first an secon kin, respectively (see Rivlin 974). Theorem 2 Consier the constraine optimal esign problem in (6) where c 2 2 j = =c 2 2 = c (, ),c l < c(l= 2 2 j,...,2 ). If there exists a constraine optimal iscriminating esign, then there also exists a symmetric constraine optimal iscriminating esign σ. (a) If c > /2, efine κ = κ(c 2, c) = 2 c 2 2c + 4c 2((2c ) j 2c), P+ (x) = (xu j(x) 2κU j (x))t j (x) 2c(xU j (x) 2κU j 2 (x))t j (x), (7) (x) = (xu j (x) 2κU j (x))u j (x) 2c(xU j (x) 2κU j 2 (x))u j 2 (x) (8) The polynomial P + (x) has + istinct roots x,...x in the interval (, ), an the esign ξ σ with masses λ k = (x k ) x P + (x), k =..., (9) x=x k at x,...,x correspons to a constraine optimal iscriminating esign by the projection (8) for the optimization problem (6). (b) If c < /2, efine P+ (x) = T +(x) ( 2c)T (x), (x) = U (x) ( 2c)U 2 (x). The polynomial P + (x) has + istinct roots x,...,x in the interval (, ), an the esign ξ σ with masses (9) at x,...,x correspons to a constraine optimal iscrimination esign by the projection (8) for the optimization problem (6)

10 52 S. Bieermann et al. Remark 2 A necessary conition for the existence of a symmetric constraine optimal iscrimination esign for the esign problem (6) is κ>, which ensures that the value of the canonical moment p 2 will be within the interval (, ).Ifκ it is therefore recommene to moify the choices of c an c 2 accoringly so that κ attains a positive value, before starting to calculate the optimal esign. Proof We only prove part (a) of the Theorem. Part (b) follows by similar (an even simpler) arguments. Note that the canonical moments of the constraine optimal iscriminating esign can be obtaine by Theorem. The canonical moments of o orer satisfy p 2n =, n =,...,, (2) 2 while the canonical moments of even orer less or equal than 2 2 j 2aregiven by p 2n =, n =,..., j. (2) 2 For the next canonical moment of even orer we have from Theorem { } p 2 2 j = max 2, c = c, an it can be shown by a straightforwar inuction that p 2 2 j+2t = (2c )t 2c, t =,..., j. (22) 2 (2c )(t + ) 2c Note that this representation implies 2 < c < j +, 2 j because the canonical moments vary in the interval (, ). The remaining canonical moment of orer 2 is obtaine by a irect calculation, that is p 2 = c 2 (2c ) j 2c 2 c (2c )( j + ) 2c. (23) It follows from a straightforwar but teious calculation that p 2 (, ) κ>, which proves Remark 2. Note that (2) (23) o not etermine a esign on the interval [, ] (except in the case c 2 =, which is exclue). In orer to obtain a esign with finite support we exten this sequence by p 2+ = 2, p 2+2 =. (24) The esign ξσ on the interval [, ] with canonical moments (2) (24) is uniquely etermine an has + support points not incluing or (seeskibinsky 986),

11 Discrimination esigns for Fourier regression moels 53 entailing that the esign σ will be supporte on points. For the calculation of the support points an corresponing weights we apply Theorem 3.6. in Dette an Stuen (997). By this result the esign ξ σ has weights λ k = (x k ) x P + (x) (25) x=x k at the roots x,...,x of the polynomial P+ (x), where P + (x) an P() (x) are obtaine from the recursion W k+ (x) = xw k (x) q 2k 2 p 2k W k (x) (26) (note that p 2 j = 2 for j =,..., + ) with ifferent initial conitions, that is P+ (x) = W +(x) for W (x), W (x) (27) (x) = W + (x) for W (x), W (x). (28) We now calculate these polynomials using (2) (24) an begin with P+ (x). From the initial conition in (27) an (2) we obtain by a straightforwar calculation W j (x) = Observing (22) an 2 j T j(x), W j (x) = 2 j 2 T j (x) (29) ( q 2l 2 p 2l = ) (2c )(l + j) 2c (2c )(l + j) 2c 2 (2c )(l + j) 2c 2 (2c )(l + j + ) 2c = (2c )(l + j + ) 2c 4 (2c )(l + j + ) 2c = 4 ( j < l ), we obtain the recursion W j+ = xw j (x) 2 cw j (x) W l+ (x) = xw l (x) 4 W l (x) if j < l. Now a straightforwar inuction yiels W j+l (x) = 2 l+ j (U l(x)t j (x) 2cU l (x)t j (x)), l =,..., j.

12 54 S. Bieermann et al. We finally note that by (22) an (23) wehaveq 2 2 p 2 = κ, from which it follows that P+ (x) = [ 2 (xu j (x) 2κU j (x))t j (x) 2c(xU j (x) ] 2κU j 2 (x))t j (x) using (26) an (27). Observing the initial conitions in (28) it follows that the polynomial (x) can be calculate analogously, where (29) is replace by W j (x) = 2 j U j (x), W j (x) = 2 j 2 U j 2(x). Consequently, by a straightforwar inuction we obtain (x) = [ 2 (xu j (x) 2κU j (x))u j (x) ] 2c(xU j (x) 2κU j 2 (x))u j 2 (x), an the assertion (a) of Theorem follows from Theorem 3.6. in Dette an Stuen (997). We conclue this section with an analogue for the optimization problem (7). The proof is similar an omitte for brevity. Theorem 3 Consier the constraine optimal esign problem in (7) where c 2 2 j = = c 2 2 = c 2 = c (, ), c l < c (l = 2 2 j,...,2 3). If there exists a constraine optimal iscriminating esign then there also exists a symmetric constraine optimal iscriminating esign σ. (a) If c > /2, efine κ = κ(c 2, c) = c 2 4c, an consier for this κ the polynomials P+ (x) an P() (x) efine by (7) an (8), respectively. The polynomial P+ (x) has + istinct roots x,...,x in the interval (, ), an the esign ξ σ which has masses (9) at the points x,...,x correspons to a constraine optimal iscriminating esign for the optimization problem (7) by the projection (8). (b) If c < /2, efine P+ (x) = T +(x) + ( 2c)T (x), (x) = U (x) + ( 2c)U 2 (x).

13 Discrimination esigns for Fourier regression moels 55 The polynomial P + (x) has + istinct roots x,...,x in the interval (, ), an the esign ξ σ with masses (9) correspons to a constraine optimal iscrimination esign for the optimization problem (7) by the projection (8). 4 Concluing remarks an iscussion We have applie the theory of canonical moments to erive constraine optimal esigns for iscriminating between Fourier moels of ifferent egree. For general constraints, we have foun explicit recursive relations (see Theorem ) for the first 2 canonical moments of the optimal esigns, from which optimal esigns can be compute applying Theorem 3.6. in Dette an Stuen (997). For some special cases (with respect to the constraints) we present explicit formulae for the supporting polynomials an the weights, which allow the irect computation of one of the optimal esigns (see Theorems 2 an 3). The choice of the lower bouns c l, l = 2 2 j,...,2, for the efficiencies is up to the experimenter accoring to his interest in a specific testing problem. There is, however, no guarantee that there exists an optimal esign satisfying these constraints. Apart from the necessary conitions that κ > (see Remark 2) an c 2 2 j+2n +c 2 2 j+2n, where n =,..., j an c k are efine in the proof of Theorem there seems to be no simple check if a particular choice of values c l will be amissible. Naturally, the experimenter woul like to have the bouns as large as possible, which might, however, contraict the existence of an optimal esign. Once the bouns have been chosen, we recommen to either compute the optimal canonical moments by the recursive relations given in Theorem or the optimal supporting polynomials P+ (x) via Theorems 2 or 3 (if appropriate). If any of the canonical moments is outsie the open interval (, ) or the roots of the polynomial P+ (x) are either outsie (, ) or not istinct then there exists no optimal esign with respect to these constraints. In this situation, the choice of the constraints has to be moifie by lowering the values of the less important bouns. If the experimenter oes not favour any testing problem over another, a natural approach to choose the bouns is c l = /2, l {2 2j,...,2} for some j. Applying Theorem yiels that in this situation all canonical moments up to the orer 2 equal /2. A somewhat relate approach woul be to maximize the minimal efficiency eff l (σ ) for any l within some subset L of {, 2,...,2}. This criterion has alreay been consiere in Dette an Haller (998), Sect. 4, Theorem 4.3. From part (b) of this theorem, it follows for example that for the choice L ={2 2j,...,2}, the optimal esigns also have canonical moments p k = /2, k =,...,2. These esigns therefore coincie with the constraine optimal esigns where all values c l, l L, are equal to /2. This correspons to intuition since neither approach gives preference to any testing problem within L over another. For ifferent choices of L the reaer is referre to Sect. 4 in Dette an Haller (998). In practice, however, there will be higher interest in testing the higher orer coefficients in most situations. A challenging problem for future consierations will be the generalization of our results an techniques to moels in more than one imension. In this situation, one approach may be to fin optimal esigns within the class of prouct esigns [see, e.g.,

14 56 S. Bieermann et al. Dette an Stuen (997), Sect. 5.8, for some results on multivariate polynomials]. Another metho worth trying will be a lattice esign approach as escribe in Riccomagno et al. (997), i.e. embe the higher imensional problem into a oneimensional structure an exploit results in one imension, incluing those in the present paper. Acknowlegments This work of the authors was supporte by the SFB 475 (Komplexitätsreuktion in multivariaten Datenstrukturen). The work of H. Dette was supporte in part by a NIH grant awar IRGM72876:A. The authors woul also like to thank I. Gottschlich, who type this paper with consierable technical expertise. We are also grateful to two referees for their constructive comments, which le to a substantial improvement of an earlier version of this paper. References Anerson, T. W. (994). The statistical analysis of time series. Classics Library, New York: Wiley. Currie, A. J., Ganeshananam, S., Noiton, D. A., Garrick, D., Shelbourne, C. J. A., Oraguzie, N. (2). Quantitative evaluation of apple (Malus omestica Borkh.) fruit shape by principle component analysis of Fourier escriptors. Euphytica,, 29E27. Clye, M., Chaloner, K. (996). The equivalence of constraine an weighte esigns in multiple objective esign problems. Journal of the American Statistical Association, 9, Cook, R. D., Wong, W. K. (994). On the equivalence of constraine an compoun optimal esigns. Journal of the American Statistical Association, 89, Dette, H. (995). Discussion of the paper Constraine optimization of experimental esign by V. Feorov an D. Cook. Statistics, 26, Dette, H., Franke, T. (2). Constraine D-anD -optimal esigns for polynomial regression. Annals of Statistics, 28, Dette, H., Haller, G. (998). Optimal esigns for the ientification of the orer of a Fourier regression. Annals of Statistics, 26, Dette, H., Melas, V. B. (23). Optimal esigns for estimating iniviual coefficients in Fourier regression moels. Annals of Statistics, 3, Dette, H., Melas, V. B., Bieermann, S. (22a). A functional-algebraic etermination of D-optimal esigns for trigonometric regression moels on a partial circle. Statistics & Probability Letters, 58(4), Dette, H., Melas, V. B., Pepelyshev, A. (22b). D-optimal esigns for trigonometric regression moels on a partial circle. Annals of the Institute of Statistical Mathematics, 54(4), Dette, H., Stuen, W. J. (997). The theory of canonical moments with applications in statistics, probability an analysis. New York: Wiley. Feorov, V. V. (972). Theory of optimal experiments. New York: Acaemic Press. Hill, P. D. H. (978). A note on the equivalence of D-optimal esign measures for three rival linear moels. Biometrika, 65, Karlin, S., Stuen, W. J. (966). Tchebycheff systems: with applications in analysis an statistics. New York: Interscience. Kitsos, C. P., Titterington, D. M., Torsney, B. (988). An optimal esign problem in rhythmometry. Biometrics, 44, Lau, T. S., Stuen, W. J. (985). Optimal esigns for trigonometric an polynomial regression. Annals of Statistics, 3, Lee, C. M. S. (988a). Constraine optimal esigns. Journal of Statistical Planninng an Inference, 8, Lee, C. M. S. (988b). D-optimal esigns for polynomial regression, when lower egree parameters are more important. Utilitas Mathematica, 34, Lestrel, P. E. (997). Fourier escriptors an their applications in biology. Cambrige: Cambrige University Press. Maria, K. (972). The statistics of irectional ata. New York: Acaemic Press. Pukelsheim, F. (993). Optimal esign of experiments. New York: Wiley. Riccomagno, E., Schwabe, R., Wynn, H.P. (997). Lattice-base D-optimum esigns for Fourier regression moels. Annals of Statistics, 25(6), Rivlin, T. J. (974). Chebyshev polynomials. New York: Wiley.

15 Discrimination esigns for Fourier regression moels 57 Skibinsky, M. (967). The range of the (n + )th moment for istributions on [, ]. Journal of Applie Probability, 4, Skibinsky, M. (986). Principal representations an canonical moment sequences for istributions on an interval. Journal of Mathematical Analysis an Applications, 2, Stigler, S. (97). Optimal experimental esigns for polynomial regression. Journal of the American Statistical Association, 66, Stuen, W. J. (968). Optimal esigns on Tchebycheff points. Annals of Mathematical Statistics, 39, Stuen, W. J. (98). D s -optimal esigns for polynomial regression using continue fractions. Annals of Statistics, 8, Stuen, W. J. (982a). Optimal esigns for weighte polynomial regression using canonical moments. In S. S. Gupta, J. O. Berger (Es.), Thir Purue symposium on ecision theory an relate topics (Vol. 2, pp ). Stuen, W. J. (982b). Some robust type D-optimal esigns in polynomial regression. Journal of the American Statistical Association, 8, Stuen, W. J. (989). Note on some φ p -optimal esigns for polynomial regression. Annals of Statistics, 7, Younker, J. L., Ehrlich, R. (977). Fourier biometrics: Harmonic amplitues as multivariate shape escriptors. Systematic Zoology, 26, 336E42. Zen, M.-M., Tsai, M.-H. (24). Criterion-robust optimal esigns for moel iscrimination an parameter estimation in Fourier regression moels. Journal of Statistical Planninng an Inference, 24,

LATTICE-BASED D-OPTIMUM DESIGN FOR FOURIER REGRESSION

LATTICE-BASED D-OPTIMUM DESIGN FOR FOURIER REGRESSION The Annals of Statistics 1997, Vol. 25, No. 6, 2313 2327 LATTICE-BASED D-OPTIMUM DESIGN FOR FOURIER REGRESSION By Eva Riccomagno, 1 Rainer Schwabe 2 an Henry P. Wynn 1 University of Warwick, Technische

More information

The derivative of a function f(x) is another function, defined in terms of a limiting expression: f(x + δx) f(x)

The derivative of a function f(x) is another function, defined in terms of a limiting expression: f(x + δx) f(x) Y. D. Chong (2016) MH2801: Complex Methos for the Sciences 1. Derivatives The erivative of a function f(x) is another function, efine in terms of a limiting expression: f (x) f (x) lim x δx 0 f(x + δx)

More information

6 General properties of an autonomous system of two first order ODE

6 General properties of an autonomous system of two first order ODE 6 General properties of an autonomous system of two first orer ODE Here we embark on stuying the autonomous system of two first orer ifferential equations of the form ẋ 1 = f 1 (, x 2 ), ẋ 2 = f 2 (, x

More information

Least-Squares Regression on Sparse Spaces

Least-Squares Regression on Sparse Spaces Least-Squares Regression on Sparse Spaces Yuri Grinberg, Mahi Milani Far, Joelle Pineau School of Computer Science McGill University Montreal, Canaa {ygrinb,mmilan1,jpineau}@cs.mcgill.ca 1 Introuction

More information

Computing Exact Confidence Coefficients of Simultaneous Confidence Intervals for Multinomial Proportions and their Functions

Computing Exact Confidence Coefficients of Simultaneous Confidence Intervals for Multinomial Proportions and their Functions Working Paper 2013:5 Department of Statistics Computing Exact Confience Coefficients of Simultaneous Confience Intervals for Multinomial Proportions an their Functions Shaobo Jin Working Paper 2013:5

More information

Quantum Mechanics in Three Dimensions

Quantum Mechanics in Three Dimensions Physics 342 Lecture 20 Quantum Mechanics in Three Dimensions Lecture 20 Physics 342 Quantum Mechanics I Monay, March 24th, 2008 We begin our spherical solutions with the simplest possible case zero potential.

More information

Table of Common Derivatives By David Abraham

Table of Common Derivatives By David Abraham Prouct an Quotient Rules: Table of Common Derivatives By Davi Abraham [ f ( g( ] = [ f ( ] g( + f ( [ g( ] f ( = g( [ f ( ] g( g( f ( [ g( ] Trigonometric Functions: sin( = cos( cos( = sin( tan( = sec

More information

Lagrangian and Hamiltonian Mechanics

Lagrangian and Hamiltonian Mechanics Lagrangian an Hamiltonian Mechanics.G. Simpson, Ph.. epartment of Physical Sciences an Engineering Prince George s Community College ecember 5, 007 Introuction In this course we have been stuying classical

More information

Linear First-Order Equations

Linear First-Order Equations 5 Linear First-Orer Equations Linear first-orer ifferential equations make up another important class of ifferential equations that commonly arise in applications an are relatively easy to solve (in theory)

More information

A Note on Exact Solutions to Linear Differential Equations by the Matrix Exponential

A Note on Exact Solutions to Linear Differential Equations by the Matrix Exponential Avances in Applie Mathematics an Mechanics Av. Appl. Math. Mech. Vol. 1 No. 4 pp. 573-580 DOI: 10.4208/aamm.09-m0946 August 2009 A Note on Exact Solutions to Linear Differential Equations by the Matrix

More information

Time-of-Arrival Estimation in Non-Line-Of-Sight Environments

Time-of-Arrival Estimation in Non-Line-Of-Sight Environments 2 Conference on Information Sciences an Systems, The Johns Hopkins University, March 2, 2 Time-of-Arrival Estimation in Non-Line-Of-Sight Environments Sinan Gezici, Hisashi Kobayashi an H. Vincent Poor

More information

Discrete Mathematics

Discrete Mathematics Discrete Mathematics 309 (009) 86 869 Contents lists available at ScienceDirect Discrete Mathematics journal homepage: wwwelseviercom/locate/isc Profile vectors in the lattice of subspaces Dániel Gerbner

More information

Hyperbolic Moment Equations Using Quadrature-Based Projection Methods

Hyperbolic Moment Equations Using Quadrature-Based Projection Methods Hyperbolic Moment Equations Using Quarature-Base Projection Methos J. Koellermeier an M. Torrilhon Department of Mathematics, RWTH Aachen University, Aachen, Germany Abstract. Kinetic equations like the

More information

Lecture XII. where Φ is called the potential function. Let us introduce spherical coordinates defined through the relations

Lecture XII. where Φ is called the potential function. Let us introduce spherical coordinates defined through the relations Lecture XII Abstract We introuce the Laplace equation in spherical coorinates an apply the metho of separation of variables to solve it. This will generate three linear orinary secon orer ifferential equations:

More information

θ x = f ( x,t) could be written as

θ x = f ( x,t) could be written as 9. Higher orer PDEs as systems of first-orer PDEs. Hyperbolic systems. For PDEs, as for ODEs, we may reuce the orer by efining new epenent variables. For example, in the case of the wave equation, (1)

More information

19 Eigenvalues, Eigenvectors, Ordinary Differential Equations, and Control

19 Eigenvalues, Eigenvectors, Ordinary Differential Equations, and Control 19 Eigenvalues, Eigenvectors, Orinary Differential Equations, an Control This section introuces eigenvalues an eigenvectors of a matrix, an iscusses the role of the eigenvalues in etermining the behavior

More information

Math 342 Partial Differential Equations «Viktor Grigoryan

Math 342 Partial Differential Equations «Viktor Grigoryan Math 342 Partial Differential Equations «Viktor Grigoryan 6 Wave equation: solution In this lecture we will solve the wave equation on the entire real line x R. This correspons to a string of infinite

More information

Agmon Kolmogorov Inequalities on l 2 (Z d )

Agmon Kolmogorov Inequalities on l 2 (Z d ) Journal of Mathematics Research; Vol. 6, No. ; 04 ISSN 96-9795 E-ISSN 96-9809 Publishe by Canaian Center of Science an Eucation Agmon Kolmogorov Inequalities on l (Z ) Arman Sahovic Mathematics Department,

More information

Proof of SPNs as Mixture of Trees

Proof of SPNs as Mixture of Trees A Proof of SPNs as Mixture of Trees Theorem 1. If T is an inuce SPN from a complete an ecomposable SPN S, then T is a tree that is complete an ecomposable. Proof. Argue by contraiction that T is not a

More information

Lecture 2 Lagrangian formulation of classical mechanics Mechanics

Lecture 2 Lagrangian formulation of classical mechanics Mechanics Lecture Lagrangian formulation of classical mechanics 70.00 Mechanics Principle of stationary action MATH-GA To specify a motion uniquely in classical mechanics, it suffices to give, at some time t 0,

More information

Lower Bounds for the Smoothed Number of Pareto optimal Solutions

Lower Bounds for the Smoothed Number of Pareto optimal Solutions Lower Bouns for the Smoothe Number of Pareto optimal Solutions Tobias Brunsch an Heiko Röglin Department of Computer Science, University of Bonn, Germany brunsch@cs.uni-bonn.e, heiko@roeglin.org Abstract.

More information

Math 1B, lecture 8: Integration by parts

Math 1B, lecture 8: Integration by parts Math B, lecture 8: Integration by parts Nathan Pflueger 23 September 2 Introuction Integration by parts, similarly to integration by substitution, reverses a well-known technique of ifferentiation an explores

More information

Acute sets in Euclidean spaces

Acute sets in Euclidean spaces Acute sets in Eucliean spaces Viktor Harangi April, 011 Abstract A finite set H in R is calle an acute set if any angle etermine by three points of H is acute. We examine the maximal carinality α() of

More information

Characterizing Real-Valued Multivariate Complex Polynomials and Their Symmetric Tensor Representations

Characterizing Real-Valued Multivariate Complex Polynomials and Their Symmetric Tensor Representations Characterizing Real-Value Multivariate Complex Polynomials an Their Symmetric Tensor Representations Bo JIANG Zhening LI Shuzhong ZHANG December 31, 2014 Abstract In this paper we stuy multivariate polynomial

More information

Differentiability, Computing Derivatives, Trig Review

Differentiability, Computing Derivatives, Trig Review Unit #3 : Differentiability, Computing Derivatives, Trig Review Goals: Determine when a function is ifferentiable at a point Relate the erivative graph to the the graph of an original function Compute

More information

Calculus of Variations

Calculus of Variations 16.323 Lecture 5 Calculus of Variations Calculus of Variations Most books cover this material well, but Kirk Chapter 4 oes a particularly nice job. x(t) x* x*+ αδx (1) x*- αδx (1) αδx (1) αδx (1) t f t

More information

Robustness and Perturbations of Minimal Bases

Robustness and Perturbations of Minimal Bases Robustness an Perturbations of Minimal Bases Paul Van Dooren an Froilán M Dopico December 9, 2016 Abstract Polynomial minimal bases of rational vector subspaces are a classical concept that plays an important

More information

NOTES ON EULER-BOOLE SUMMATION (1) f (l 1) (n) f (l 1) (m) + ( 1)k 1 k! B k (y) f (k) (y) dy,

NOTES ON EULER-BOOLE SUMMATION (1) f (l 1) (n) f (l 1) (m) + ( 1)k 1 k! B k (y) f (k) (y) dy, NOTES ON EULER-BOOLE SUMMATION JONATHAN M BORWEIN, NEIL J CALKIN, AND DANTE MANNA Abstract We stuy a connection between Euler-MacLaurin Summation an Boole Summation suggeste in an AMM note from 196, which

More information

Implicit Differentiation

Implicit Differentiation Implicit Differentiation Thus far, the functions we have been concerne with have been efine explicitly. A function is efine explicitly if the output is given irectly in terms of the input. For instance,

More information

Sturm-Liouville Theory

Sturm-Liouville Theory LECTURE 5 Sturm-Liouville Theory In the three preceing lectures I emonstrate the utility of Fourier series in solving PDE/BVPs. As we ll now see, Fourier series are just the tip of the iceberg of the theory

More information

The Principle of Least Action

The Principle of Least Action Chapter 7. The Principle of Least Action 7.1 Force Methos vs. Energy Methos We have so far stuie two istinct ways of analyzing physics problems: force methos, basically consisting of the application of

More information

A Modification of the Jarque-Bera Test. for Normality

A Modification of the Jarque-Bera Test. for Normality Int. J. Contemp. Math. Sciences, Vol. 8, 01, no. 17, 84-85 HIKARI Lt, www.m-hikari.com http://x.oi.org/10.1988/ijcms.01.9106 A Moification of the Jarque-Bera Test for Normality Moawa El-Fallah Ab El-Salam

More information

APPROXIMATE SOLUTION FOR TRANSIENT HEAT TRANSFER IN STATIC TURBULENT HE II. B. Baudouy. CEA/Saclay, DSM/DAPNIA/STCM Gif-sur-Yvette Cedex, France

APPROXIMATE SOLUTION FOR TRANSIENT HEAT TRANSFER IN STATIC TURBULENT HE II. B. Baudouy. CEA/Saclay, DSM/DAPNIA/STCM Gif-sur-Yvette Cedex, France APPROXIMAE SOLUION FOR RANSIEN HEA RANSFER IN SAIC URBULEN HE II B. Bauouy CEA/Saclay, DSM/DAPNIA/SCM 91191 Gif-sur-Yvette Ceex, France ABSRAC Analytical solution in one imension of the heat iffusion equation

More information

Tractability results for weighted Banach spaces of smooth functions

Tractability results for weighted Banach spaces of smooth functions Tractability results for weighte Banach spaces of smooth functions Markus Weimar Mathematisches Institut, Universität Jena Ernst-Abbe-Platz 2, 07740 Jena, Germany email: markus.weimar@uni-jena.e March

More information

Differentiability, Computing Derivatives, Trig Review. Goals:

Differentiability, Computing Derivatives, Trig Review. Goals: Secants vs. Derivatives - Unit #3 : Goals: Differentiability, Computing Derivatives, Trig Review Determine when a function is ifferentiable at a point Relate the erivative graph to the the graph of an

More information

Euler equations for multiple integrals

Euler equations for multiple integrals Euler equations for multiple integrals January 22, 2013 Contents 1 Reminer of multivariable calculus 2 1.1 Vector ifferentiation......................... 2 1.2 Matrix ifferentiation........................

More information

Diagonalization of Matrices Dr. E. Jacobs

Diagonalization of Matrices Dr. E. Jacobs Diagonalization of Matrices Dr. E. Jacobs One of the very interesting lessons in this course is how certain algebraic techniques can be use to solve ifferential equations. The purpose of these notes is

More information

Schrödinger s equation.

Schrödinger s equation. Physics 342 Lecture 5 Schröinger s Equation Lecture 5 Physics 342 Quantum Mechanics I Wenesay, February 3r, 2010 Toay we iscuss Schröinger s equation an show that it supports the basic interpretation of

More information

Math Notes on differentials, the Chain Rule, gradients, directional derivative, and normal vectors

Math Notes on differentials, the Chain Rule, gradients, directional derivative, and normal vectors Math 18.02 Notes on ifferentials, the Chain Rule, graients, irectional erivative, an normal vectors Tangent plane an linear approximation We efine the partial erivatives of f( xy, ) as follows: f f( x+

More information

A. Incorrect! The letter t does not appear in the expression of the given integral

A. Incorrect! The letter t does not appear in the expression of the given integral AP Physics C - Problem Drill 1: The Funamental Theorem of Calculus Question No. 1 of 1 Instruction: (1) Rea the problem statement an answer choices carefully () Work the problems on paper as neee (3) Question

More information

Lectures - Week 10 Introduction to Ordinary Differential Equations (ODES) First Order Linear ODEs

Lectures - Week 10 Introduction to Ordinary Differential Equations (ODES) First Order Linear ODEs Lectures - Week 10 Introuction to Orinary Differential Equations (ODES) First Orer Linear ODEs When stuying ODEs we are consiering functions of one inepenent variable, e.g., f(x), where x is the inepenent

More information

Lie symmetry and Mei conservation law of continuum system

Lie symmetry and Mei conservation law of continuum system Chin. Phys. B Vol. 20, No. 2 20 020 Lie symmetry an Mei conservation law of continuum system Shi Shen-Yang an Fu Jing-Li Department of Physics, Zhejiang Sci-Tech University, Hangzhou 3008, China Receive

More information

4.2 First Differentiation Rules; Leibniz Notation

4.2 First Differentiation Rules; Leibniz Notation .. FIRST DIFFERENTIATION RULES; LEIBNIZ NOTATION 307. First Differentiation Rules; Leibniz Notation In this section we erive rules which let us quickly compute the erivative function f (x) for any polynomial

More information

3.7 Implicit Differentiation -- A Brief Introduction -- Student Notes

3.7 Implicit Differentiation -- A Brief Introduction -- Student Notes Fin these erivatives of these functions: y.7 Implicit Differentiation -- A Brief Introuction -- Stuent Notes tan y sin tan = sin y e = e = Write the inverses of these functions: y tan y sin How woul we

More information

Abstract A nonlinear partial differential equation of the following form is considered:

Abstract A nonlinear partial differential equation of the following form is considered: M P E J Mathematical Physics Electronic Journal ISSN 86-6655 Volume 2, 26 Paper 5 Receive: May 3, 25, Revise: Sep, 26, Accepte: Oct 6, 26 Eitor: C.E. Wayne A Nonlinear Heat Equation with Temperature-Depenent

More information

Physics 5153 Classical Mechanics. The Virial Theorem and The Poisson Bracket-1

Physics 5153 Classical Mechanics. The Virial Theorem and The Poisson Bracket-1 Physics 5153 Classical Mechanics The Virial Theorem an The Poisson Bracket 1 Introuction In this lecture we will consier two applications of the Hamiltonian. The first, the Virial Theorem, applies to systems

More information

Final Exam Study Guide and Practice Problems Solutions

Final Exam Study Guide and Practice Problems Solutions Final Exam Stuy Guie an Practice Problems Solutions Note: These problems are just some of the types of problems that might appear on the exam. However, to fully prepare for the exam, in aition to making

More information

ensembles When working with density operators, we can use this connection to define a generalized Bloch vector: v x Tr x, v y Tr y

ensembles When working with density operators, we can use this connection to define a generalized Bloch vector: v x Tr x, v y Tr y Ph195a lecture notes, 1/3/01 Density operators for spin- 1 ensembles So far in our iscussion of spin- 1 systems, we have restricte our attention to the case of pure states an Hamiltonian evolution. Toay

More information

Optimal discrimination designs

Optimal discrimination designs Optimal discrimination designs Holger Dette Ruhr-Universität Bochum Fakultät für Mathematik 44780 Bochum, Germany e-mail: holger.dette@ruhr-uni-bochum.de Stefanie Titoff Ruhr-Universität Bochum Fakultät

More information

Survey Sampling. 1 Design-based Inference. Kosuke Imai Department of Politics, Princeton University. February 19, 2013

Survey Sampling. 1 Design-based Inference. Kosuke Imai Department of Politics, Princeton University. February 19, 2013 Survey Sampling Kosuke Imai Department of Politics, Princeton University February 19, 2013 Survey sampling is one of the most commonly use ata collection methos for social scientists. We begin by escribing

More information

Optimal designs for estimating the slope of a regression

Optimal designs for estimating the slope of a regression Optimal designs for estimating the slope of a regression Holger Dette Ruhr-Universität Bochum Fakultät für Mathematik 4480 Bochum, Germany e-mail: holger.dette@rub.de Viatcheslav B. Melas St. Petersburg

More information

A Path Planning Method Using Cubic Spiral with Curvature Constraint

A Path Planning Method Using Cubic Spiral with Curvature Constraint A Path Planning Metho Using Cubic Spiral with Curvature Constraint Tzu-Chen Liang an Jing-Sin Liu Institute of Information Science 0, Acaemia Sinica, Nankang, Taipei 5, Taiwan, R.O.C., Email: hartree@iis.sinica.eu.tw

More information

Qubit channels that achieve capacity with two states

Qubit channels that achieve capacity with two states Qubit channels that achieve capacity with two states Dominic W. Berry Department of Physics, The University of Queenslan, Brisbane, Queenslan 4072, Australia Receive 22 December 2004; publishe 22 March

More information

Quantum mechanical approaches to the virial

Quantum mechanical approaches to the virial Quantum mechanical approaches to the virial S.LeBohec Department of Physics an Astronomy, University of Utah, Salt Lae City, UT 84112, USA Date: June 30 th 2015 In this note, we approach the virial from

More information

Introduction to the Vlasov-Poisson system

Introduction to the Vlasov-Poisson system Introuction to the Vlasov-Poisson system Simone Calogero 1 The Vlasov equation Consier a particle with mass m > 0. Let x(t) R 3 enote the position of the particle at time t R an v(t) = ẋ(t) = x(t)/t its

More information

Entanglement is not very useful for estimating multiple phases

Entanglement is not very useful for estimating multiple phases PHYSICAL REVIEW A 70, 032310 (2004) Entanglement is not very useful for estimating multiple phases Manuel A. Ballester* Department of Mathematics, University of Utrecht, Box 80010, 3508 TA Utrecht, The

More information

Separation of Variables

Separation of Variables Physics 342 Lecture 1 Separation of Variables Lecture 1 Physics 342 Quantum Mechanics I Monay, January 25th, 2010 There are three basic mathematical tools we nee, an then we can begin working on the physical

More information

Laplace s Equation in Cylindrical Coordinates and Bessel s Equation (II)

Laplace s Equation in Cylindrical Coordinates and Bessel s Equation (II) Laplace s Equation in Cylinrical Coorinates an Bessel s Equation (II Qualitative properties of Bessel functions of first an secon kin In the last lecture we foun the expression for the general solution

More information

The Exact Form and General Integrating Factors

The Exact Form and General Integrating Factors 7 The Exact Form an General Integrating Factors In the previous chapters, we ve seen how separable an linear ifferential equations can be solve using methos for converting them to forms that can be easily

More information

arxiv: v4 [math.pr] 27 Jul 2016

arxiv: v4 [math.pr] 27 Jul 2016 The Asymptotic Distribution of the Determinant of a Ranom Correlation Matrix arxiv:309768v4 mathpr] 7 Jul 06 AM Hanea a, & GF Nane b a Centre of xcellence for Biosecurity Risk Analysis, University of Melbourne,

More information

SYNCHRONOUS SEQUENTIAL CIRCUITS

SYNCHRONOUS SEQUENTIAL CIRCUITS CHAPTER SYNCHRONOUS SEUENTIAL CIRCUITS Registers an counters, two very common synchronous sequential circuits, are introuce in this chapter. Register is a igital circuit for storing information. Contents

More information

Balancing Expected and Worst-Case Utility in Contracting Models with Asymmetric Information and Pooling

Balancing Expected and Worst-Case Utility in Contracting Models with Asymmetric Information and Pooling Balancing Expecte an Worst-Case Utility in Contracting Moels with Asymmetric Information an Pooling R.B.O. erkkamp & W. van en Heuvel & A.P.M. Wagelmans Econometric Institute Report EI2018-01 9th January

More information

The canonical controllers and regular interconnection

The canonical controllers and regular interconnection Systems & Control Letters ( www.elsevier.com/locate/sysconle The canonical controllers an regular interconnection A.A. Julius a,, J.C. Willems b, M.N. Belur c, H.L. Trentelman a Department of Applie Mathematics,

More information

The chromatic number of graph powers

The chromatic number of graph powers Combinatorics, Probability an Computing (19XX) 00, 000 000. c 19XX Cambrige University Press Printe in the Unite Kingom The chromatic number of graph powers N O G A A L O N 1 an B O J A N M O H A R 1 Department

More information

Thermal conductivity of graded composites: Numerical simulations and an effective medium approximation

Thermal conductivity of graded composites: Numerical simulations and an effective medium approximation JOURNAL OF MATERIALS SCIENCE 34 (999)5497 5503 Thermal conuctivity of grae composites: Numerical simulations an an effective meium approximation P. M. HUI Department of Physics, The Chinese University

More information

Robust Forward Algorithms via PAC-Bayes and Laplace Distributions. ω Q. Pr (y(ω x) < 0) = Pr A k

Robust Forward Algorithms via PAC-Bayes and Laplace Distributions. ω Q. Pr (y(ω x) < 0) = Pr A k A Proof of Lemma 2 B Proof of Lemma 3 Proof: Since the support of LL istributions is R, two such istributions are equivalent absolutely continuous with respect to each other an the ivergence is well-efine

More information

ALGEBRAIC AND ANALYTIC PROPERTIES OF ARITHMETIC FUNCTIONS

ALGEBRAIC AND ANALYTIC PROPERTIES OF ARITHMETIC FUNCTIONS ALGEBRAIC AND ANALYTIC PROPERTIES OF ARITHMETIC FUNCTIONS MARK SCHACHNER Abstract. When consiere as an algebraic space, the set of arithmetic functions equippe with the operations of pointwise aition an

More information

Lower bounds on Locality Sensitive Hashing

Lower bounds on Locality Sensitive Hashing Lower bouns on Locality Sensitive Hashing Rajeev Motwani Assaf Naor Rina Panigrahy Abstract Given a metric space (X, X ), c 1, r > 0, an p, q [0, 1], a istribution over mappings H : X N is calle a (r,

More information

Improving Estimation Accuracy in Nonrandomized Response Questioning Methods by Multiple Answers

Improving Estimation Accuracy in Nonrandomized Response Questioning Methods by Multiple Answers International Journal of Statistics an Probability; Vol 6, No 5; September 207 ISSN 927-7032 E-ISSN 927-7040 Publishe by Canaian Center of Science an Eucation Improving Estimation Accuracy in Nonranomize

More information

Assignment 1. g i (x 1,..., x n ) dx i = 0. i=1

Assignment 1. g i (x 1,..., x n ) dx i = 0. i=1 Assignment 1 Golstein 1.4 The equations of motion for the rolling isk are special cases of general linear ifferential equations of constraint of the form g i (x 1,..., x n x i = 0. i=1 A constraint conition

More information

ON THE OPTIMALITY SYSTEM FOR A 1 D EULER FLOW PROBLEM

ON THE OPTIMALITY SYSTEM FOR A 1 D EULER FLOW PROBLEM ON THE OPTIMALITY SYSTEM FOR A D EULER FLOW PROBLEM Eugene M. Cliff Matthias Heinkenschloss y Ajit R. Shenoy z Interisciplinary Center for Applie Mathematics Virginia Tech Blacksburg, Virginia 46 Abstract

More information

7.1 Support Vector Machine

7.1 Support Vector Machine 67577 Intro. to Machine Learning Fall semester, 006/7 Lecture 7: Support Vector Machines an Kernel Functions II Lecturer: Amnon Shashua Scribe: Amnon Shashua 7. Support Vector Machine We return now to

More information

Perfect Matchings in Õ(n1.5 ) Time in Regular Bipartite Graphs

Perfect Matchings in Õ(n1.5 ) Time in Regular Bipartite Graphs Perfect Matchings in Õ(n1.5 ) Time in Regular Bipartite Graphs Ashish Goel Michael Kapralov Sanjeev Khanna Abstract We consier the well-stuie problem of fining a perfect matching in -regular bipartite

More information

QF101: Quantitative Finance September 5, Week 3: Derivatives. Facilitator: Christopher Ting AY 2017/2018. f ( x + ) f(x) f(x) = lim

QF101: Quantitative Finance September 5, Week 3: Derivatives. Facilitator: Christopher Ting AY 2017/2018. f ( x + ) f(x) f(x) = lim QF101: Quantitative Finance September 5, 2017 Week 3: Derivatives Facilitator: Christopher Ting AY 2017/2018 I recoil with ismay an horror at this lamentable plague of functions which o not have erivatives.

More information

Compound Optimal Designs for Percentile Estimation in Dose-Response Models with Restricted Design Intervals

Compound Optimal Designs for Percentile Estimation in Dose-Response Models with Restricted Design Intervals Compound Optimal Designs for Percentile Estimation in Dose-Response Models with Restricted Design Intervals Stefanie Biedermann 1, Holger Dette 1, Wei Zhu 2 Abstract In dose-response studies, the dose

More information

On colour-blind distinguishing colour pallets in regular graphs

On colour-blind distinguishing colour pallets in regular graphs J Comb Optim (2014 28:348 357 DOI 10.1007/s10878-012-9556-x On colour-blin istinguishing colour pallets in regular graphs Jakub Przybyło Publishe online: 25 October 2012 The Author(s 2012. This article

More information

Introduction to variational calculus: Lecture notes 1

Introduction to variational calculus: Lecture notes 1 October 10, 2006 Introuction to variational calculus: Lecture notes 1 Ewin Langmann Mathematical Physics, KTH Physics, AlbaNova, SE-106 91 Stockholm, Sween Abstract I give an informal summary of variational

More information

arxiv: v1 [math.co] 29 May 2009

arxiv: v1 [math.co] 29 May 2009 arxiv:0905.4913v1 [math.co] 29 May 2009 simple Havel-Hakimi type algorithm to realize graphical egree sequences of irecte graphs Péter L. Erős an István Miklós. Rényi Institute of Mathematics, Hungarian

More information

11.7. Implicit Differentiation. Introduction. Prerequisites. Learning Outcomes

11.7. Implicit Differentiation. Introduction. Prerequisites. Learning Outcomes Implicit Differentiation 11.7 Introuction This Section introuces implicit ifferentiation which is use to ifferentiate functions expresse in implicit form (where the variables are foun together). Examples

More information

Research Article When Inflation Causes No Increase in Claim Amounts

Research Article When Inflation Causes No Increase in Claim Amounts Probability an Statistics Volume 2009, Article ID 943926, 10 pages oi:10.1155/2009/943926 Research Article When Inflation Causes No Increase in Claim Amounts Vytaras Brazauskas, 1 Bruce L. Jones, 2 an

More information

u t v t v t c a u t b a v t u t v t b a

u t v t v t c a u t b a v t u t v t b a Nonlinear Dynamical Systems In orer to iscuss nonlinear ynamical systems, we must first consier linear ynamical systems. Linear ynamical systems are just systems of linear equations like we have been stuying

More information

Designing Information Devices and Systems II Fall 2017 Note Theorem: Existence and Uniqueness of Solutions to Differential Equations

Designing Information Devices and Systems II Fall 2017 Note Theorem: Existence and Uniqueness of Solutions to Differential Equations EECS 6B Designing Information Devices an Systems II Fall 07 Note 3 Secon Orer Differential Equations Secon orer ifferential equations appear everywhere in the real worl. In this note, we will walk through

More information

EVALUATING HIGHER DERIVATIVE TENSORS BY FORWARD PROPAGATION OF UNIVARIATE TAYLOR SERIES

EVALUATING HIGHER DERIVATIVE TENSORS BY FORWARD PROPAGATION OF UNIVARIATE TAYLOR SERIES MATHEMATICS OF COMPUTATION Volume 69, Number 231, Pages 1117 1130 S 0025-5718(00)01120-0 Article electronically publishe on February 17, 2000 EVALUATING HIGHER DERIVATIVE TENSORS BY FORWARD PROPAGATION

More information

12.11 Laplace s Equation in Cylindrical and

12.11 Laplace s Equation in Cylindrical and SEC. 2. Laplace s Equation in Cylinrical an Spherical Coorinates. Potential 593 2. Laplace s Equation in Cylinrical an Spherical Coorinates. Potential One of the most important PDEs in physics an engineering

More information

Topic 7: Convergence of Random Variables

Topic 7: Convergence of Random Variables Topic 7: Convergence of Ranom Variables Course 003, 2016 Page 0 The Inference Problem So far, our starting point has been a given probability space (S, F, P). We now look at how to generate information

More information

Three-Variable Bracket Polynomial for Two-Bridge Knots

Three-Variable Bracket Polynomial for Two-Bridge Knots Three-Variable Bracket Polynomial for Two-Brige Knots Matthew Overuin Research Experience for Unergrauates California State University, San Bernarino San Bernarino, CA 92407 August 23, 2013 Abstract In

More information

Calculus and optimization

Calculus and optimization Calculus an optimization These notes essentially correspon to mathematical appenix 2 in the text. 1 Functions of a single variable Now that we have e ne functions we turn our attention to calculus. A function

More information

Review of Differentiation and Integration for Ordinary Differential Equations

Review of Differentiation and Integration for Ordinary Differential Equations Schreyer Fall 208 Review of Differentiation an Integration for Orinary Differential Equations In this course you will be expecte to be able to ifferentiate an integrate quickly an accurately. Many stuents

More information

2 Viktor G. Kurotschka, Rainer Schwabe. 1) In the case of a small experimental region, mathematically described by

2 Viktor G. Kurotschka, Rainer Schwabe. 1) In the case of a small experimental region, mathematically described by HE REDUION OF DESIGN PROLEMS FOR MULIVARIAE EXPERIMENS O UNIVARIAE POSSIILIIES AND HEIR LIMIAIONS Viktor G Kurotschka, Rainer Schwabe Freie Universitat erlin, Mathematisches Institut, Arnimallee 2{6, D-4

More information

PDE Notes, Lecture #11

PDE Notes, Lecture #11 PDE Notes, Lecture # from Professor Jalal Shatah s Lectures Febuary 9th, 2009 Sobolev Spaces Recall that for u L loc we can efine the weak erivative Du by Du, φ := udφ φ C0 If v L loc such that Du, φ =

More information

23 Implicit differentiation

23 Implicit differentiation 23 Implicit ifferentiation 23.1 Statement The equation y = x 2 + 3x + 1 expresses a relationship between the quantities x an y. If a value of x is given, then a corresponing value of y is etermine. For

More information

Similar Operators and a Functional Calculus for the First-Order Linear Differential Operator

Similar Operators and a Functional Calculus for the First-Order Linear Differential Operator Avances in Applie Mathematics, 9 47 999 Article ID aama.998.067, available online at http: www.iealibrary.com on Similar Operators an a Functional Calculus for the First-Orer Linear Differential Operator

More information

arxiv: v1 [hep-lat] 19 Nov 2013

arxiv: v1 [hep-lat] 19 Nov 2013 HU-EP-13/69 SFB/CPP-13-98 DESY 13-225 Applicability of Quasi-Monte Carlo for lattice systems arxiv:1311.4726v1 [hep-lat] 19 ov 2013, a,b Tobias Hartung, c Karl Jansen, b Hernan Leovey, Anreas Griewank

More information

Generalizing Kronecker Graphs in order to Model Searchable Networks

Generalizing Kronecker Graphs in order to Model Searchable Networks Generalizing Kronecker Graphs in orer to Moel Searchable Networks Elizabeth Boine, Babak Hassibi, Aam Wierman California Institute of Technology Pasaena, CA 925 Email: {eaboine, hassibi, aamw}@caltecheu

More information

The Three-dimensional Schödinger Equation

The Three-dimensional Schödinger Equation The Three-imensional Schöinger Equation R. L. Herman November 7, 016 Schröinger Equation in Spherical Coorinates We seek to solve the Schröinger equation with spherical symmetry using the metho of separation

More information

Optimal CDMA Signatures: A Finite-Step Approach

Optimal CDMA Signatures: A Finite-Step Approach Optimal CDMA Signatures: A Finite-Step Approach Joel A. Tropp Inst. for Comp. Engr. an Sci. (ICES) 1 University Station C000 Austin, TX 7871 jtropp@ices.utexas.eu Inerjit. S. Dhillon Dept. of Comp. Sci.

More information

Generalization of the persistent random walk to dimensions greater than 1

Generalization of the persistent random walk to dimensions greater than 1 PHYSICAL REVIEW E VOLUME 58, NUMBER 6 DECEMBER 1998 Generalization of the persistent ranom walk to imensions greater than 1 Marián Boguñá, Josep M. Porrà, an Jaume Masoliver Departament e Física Fonamental,

More information

Partial Differential Equations

Partial Differential Equations Chapter Partial Differential Equations. Introuction Have solve orinary ifferential equations, i.e. ones where there is one inepenent an one epenent variable. Only orinary ifferentiation is therefore involve.

More information

Mathematical Review Problems

Mathematical Review Problems Fall 6 Louis Scuiero Mathematical Review Problems I. Polynomial Equations an Graphs (Barrante--Chap. ). First egree equation an graph y f() x mx b where m is the slope of the line an b is the line's intercept

More information

THE VAN KAMPEN EXPANSION FOR LINKED DUFFING LINEAR OSCILLATORS EXCITED BY COLORED NOISE

THE VAN KAMPEN EXPANSION FOR LINKED DUFFING LINEAR OSCILLATORS EXCITED BY COLORED NOISE Journal of Soun an Vibration (1996) 191(3), 397 414 THE VAN KAMPEN EXPANSION FOR LINKED DUFFING LINEAR OSCILLATORS EXCITED BY COLORED NOISE E. M. WEINSTEIN Galaxy Scientific Corporation, 2500 English Creek

More information