UMCS. Fast multidimensional Bernstein-Lagrange algorithms. The John Paul II Catholic University of Lublin, ul. Konstantynow 1H, Lublin, Poland
|
|
- Frederick Mathews
- 6 years ago
- Views:
Transcription
1 Annales Informatica AI XII, ) 7 18 DOI: /v Fast multidimensional Bernstein-Lagrange algorithms Joanna Kapusta 1, Ryszard Smarzewski 1 1 Institute of Mathematics and Computer Science, The John Paul II Catholic University of Lublin, ul. Konstantynow 1H, Lublin, Poland Abstract In this paper we present two fast algorithms for the Bézier curves and surfaces of an arbitrary dimension. The first algorithm evaluates the Bernstein-Bézier curves and surfaces at a set of specific points by using the fast Bernstein-Lagrange transformation. The second algorithm is an inversion of the first one. Both algorithms reduce the initial problem to computation of some discrete Fourier transformations in the case of geometrical subdivisions of the d-dimensional cube. Their orders of computational complexity are proportional to those of corresponding d-dimensional FFT-algorithm, i.e. to ON logn)+odn), where N denotes the order of the Bernstein-Bézier curves. 1 Introduction Let n =n 1,n 2,...,n d ) be a d-tuple of positive integers and K be a field. Moreover, let Q n be a lattice of all N = n 1 n 2...n d multi-indices α =α 1,α 2,...,α d ) with the integer coordinates satisfying inequalities 0 α i <n i for i =1,2,...,d. 1) Using the multi-index notation, we write Bernstein-Bézier vector polynomials of the variable x =x 1,x 2,...,x d ) K d in the form p n x)= α Q n f α B α x), 2) jkapusta@kul.lublin.pl rsmax@kul.lublin.pl
2 8 Fast multidimensional Bernstein-Lagrange algorithms where f α K s are the control points, the summation extends over all n 1 n 2 n d multi-indices α from the lattice Q n, and ) n 1 d ) B α x)= x α 1 x) n α 1 nj 1 = x αj j 1 x j ) nj 1 αj, 3) α where n 1=n 1 1,n 2 1,...,n d 1). Note that p n x) is a Bézier curve or surface in the case when d =1or d =2, respectively. Additionally, suppose that j=1 x α =x 1,α1,x 2,α2,...,x d,αd ),α=α 1,α 2,...,α d ) Q n, 4) are the points in K d such that coordinates α j x i,0,x i,1,...,x i,ni 1 i =1,2,...,d) 5) are pairwise distinct, i.e. x i,j x i,k, whenever j k. Then the Bernstein-Bézier vector polynomial p n x) can be written in the Lagrange form p n x)= α Q n y α L α x), 6) where y α = p n x α ) K s and L α x)= d L αi x i ),L αi x i )= n i 1 j=0 j α i In this paper we present two fast algorithms of the order x i x i,j x i,αi x i,j. 7) ON logn)+odn),n= Q n = n 1 n 2...n d, 8) for the Bernstein-Lagrange transformation T :f β ) β Qn y β ) β Qn, and its inverse, which is defined by T : y β = f α B α x β ),β Q n, 9) α Q n where the coordinates of points x β are such that x i,j = λ i γ j i i =1,2,...,d, j=0,1,...,n i 1) 10) with the scalarsλ i 0, γ i 0andγ i 1i =1,2,...,d) fixed ink. For the simplicity, these algorithms will be established under the additional assumption that s =1,which does not restrict the generality of our considerations. Since the coordinates x i,j j =0,1,...,n i 1) form geometrical progression, it follows that the points x β can be used e.g. in extrapolation problems [1]. It is not clear if it is possible to extend our fast algorithms to the case of arithmetic progression, or more generally to the case when x i,j = λ i x i,j 1 +δ i i =1,2,...,d, j=1,2,...,n i 1), 11) where λ i 0, δ i and x i,0 = κ i belong to the field K [2]. Of course, in order to evaluate the transformation T for the last coordinates one can use multidimensional algorithms
3 Joanna Kapusta, Ryszard Smarzewski 9 based on the de Casteljau algorithm, which have the computational complexity of the order greater than our algorithms, cf. [3], [4], [5] and [6]. Following [7] and [8], our algorithms will use the discrete Fourier transformation and its inverse, which are defined by F m : K m a b K m 12) m 1 b i = a k ψm ik and a i = 1 m 1 b k ψm ik m i =0,1,...,m 1), where ψ m is supposed to be a primitive root of the unity of order m in the field K. It is well known that discrete Fourier transformations can be computed by the famous FFT-algorithm, which has a running time of order Omlogm) [9]. 2 Fast multidimensional convolutions and deconvolutions In order to present fast algorithms for computation of the Bernstein-Lagrange transformations T and T 1, we need fast algorithms for multidimensional convolutions and deconvolutions. For this purpose, suppose that a =a 0,a 1,...) and b =b 0,b 1,...) are two finite or infinite sequences. Then the wrapped convolution is defined by while its deconvolution c i = c =c 0,c 1,...,c m 1 )=a m b 13) i a k b i k i =0,1,...,m 1), 14) is supposed to be the solution a = c m b = c m b 1 b 0 0) a 0 = c 0 /b 0, i 1 a i = c i a k b i k )/b 0 i =1,2,...,m 1) of the lower triangular system of equations 14). Moreover, the convolutionary inverse is such that where bx)= 15) d =d 0,d 1,...d r 1 )=b 1 16) r 1 1/bx)= d k x k +Ox r ), 17) m 1 b k x k and d k = dk dx k 1. 18) bx)) x=0
4 10 Fast multidimensional Bernstein-Lagrange algorithms The wrapped convolution satisfies the formula a m b = { F 1 m [F m a) F m b)]+f 1 m [F m Ψ a) F m Ψ b)]/ψ } /2, 19) where a =a 0,a 1,...,a m 1 ), b =b 0,b 1,...,b m 1 ), Ψ=1,ψ 1 2m,...,ψ m 1 2m ), ψ 2m is the primitive root of order2m of the unity ink, and vector operations of multiplication and division are defined coordinatewise. Formula 19) gives an extremely effective and fast algorithm of the order Om logm) to evaluate wrapped convolutions, which is observed implicitly in [7], see also [8]. Note that it can be also applied to evaluate m 1 b j = a i γ ij j =1,2,...,m 1). 20) Indeed, we have m 1 j m 1 b j = a i γ ij = r j p i q j i + p i q j i+1) j =0,1,...,m 1), 21) where Hence with and r j = i=j+1 j j 1 γ k,p j = a j γ k,q j = 1 j =0,1,...,m 1). j γ k j b j m 1) = r j m 1) d i c j i j = m 1,m,...,2m 2) c i = { pi for i =0,1,...,m 1, d i = 0 for i = m,m+1,...,2m 2 { qm 2 i for i =0,1,...,m 2, q i m 1) for i = m 1,m,...,2m 2. Consequently, if d =d i ) 2m 2, c =c i ) 2m 2 and r =r i ) m 1, then we get where b = d m c ) r, 22) d m c = P m d 2m 1 c) and the projection P m : K 2m 1 K m is defined by P m e)=e m 1,e m,...,e 2m 2 ),e=e 0,e 1,...,e 2m 2 ). 23) It is clear that the order of algorithm 22) is equal to Omlogm). Note, that another algorithm for computing 20), which has the same order of complexity, was presented in [10].
5 Joanna Kapusta, Ryszard Smarzewski 11 The wrapped convolution can be also applied to evaluate the multidimensional convolution u = a b, 24) of a hypermatrix a =a α ) α Qn and vector b =b i ) d, with b i =b i,0,b i,1,...,b i,ni 1). Here coordinates of u are equal to u α = β Q α a β b α β,α=α 1,α 2,...,α d ) Q n. 25) Definition 1 [2]). A hypermatrix w =w α ) α Qn = a i) b i K n1 n2 n d, 1 i d, 26) is said to be the i-th partial hypermatrix convolution of a hypermatrix a =a α ) α Qn and a vector b i =b i,0,b i,1,...,b i,ni 1), whenever each column w β1,...,β i 1,,β i+1,...,β d = a β1,...,β i 1,,β i+1,...,β d ni b i, 0 β j <n j 1, j =1,2,...,i 1,i+1,...,d, of the hypermatrix w is equal to the wrapped convolution of the column and vector b i. a β1,...,β i 1,,β i+1,...,β d = a β1,...,β i 1,j,β i+1,...,β d ) ni 1 j=0. 27) The notation of the partial hypermatrix convolutions enables to rewrite the multidimensional convolution u =u α ) α Qn in the following hypermatrix form ) ) u = a b =... a 1) b 1 ) 2) b 2 3)... d) b d 28) with b i =b i,0,b i,1,...,b i,ni 1) and a =a α ) α Qn [2]. Hence it is clear that the fast algorithm for computing an i-th partial hypermatrix convolution should evaluate N/n i one-dimensional convolutions for vectors of size n i. Therefore, algorithm 28) for computing the multidimensional convolution is of the order N/n 1 )On 1 logn 1 )+...+N/n d )On d logn d )=ON logn),n= n 1 n 2...n d. 29) The same order is in the algorithm ) ) v = a b =... a 1) b 1 2) b 2...) 3) d) b d, 30) where the i-th extended convolution a i) b i i =1,2,...,dis defined as in Definition 1 with a β1,...,β i 1,,β i+1,...,β d ni b i 31) replaced by P ni aβ1,...,β i 1,,β i+1,...,β d 2ni 1 b i ) and a α =0for α/ Q n. In a similar way one can define the hypermatrix deconvolution 32) a = u b, 33)
6 12 Fast multidimensional Bernstein-Lagrange algorithms whenever b i,0 0for i =0,1,...,n i 1. The only difference consists in replacing the operator i) of the i -th partial hypermatrix convolution in Definition 1 by the corresponding operator i) of the i-th partial hypermatrix deconvolution. In other words, each column of the hypermatrix should be equal to a =a α ) α Qn = w i) b i K n1 n2 n d, 1 i d, 34) a β1,...,β i 1,,β i+1,...,β d = w β1,...,β i 1,,β i+1,...,β d ni b i = w β1,...,β i 1,,β i+1,...,β d ni b 1 where 0 β j <n j. Then we have ) ) a =... u d) b d... ) 2) b 2 1) b 1 ) ) ) 35) =... u d) b ) b 1 2 1) b 1 1. d One can prove that the last algorithm for hypermatrix deconvolution is of the order i, ON logn),n= n 1 n 2...n d. 36) For this purpose, it is sufficient to observe that the convolutionary inverse of a vector b =b 0,b 1,...,b m 1 ) K m with b 0 0can be computed by the Newton method of the order Omlogm) [11]. More precisely, let x i+1 =2x i x 2 ib, i =0,1,..., 37) be the Newton iterative formula for the function f x)=x 1 b x 0). Moreover, suppose that the coefficients d 0,d 1,...,d 2 i 1 i 1) 38) of the inverse b0 +b 1 x+...b m 1 x m 1) 1 = d0 +d 1 x+...d 2 i 1x 2i 1 +Ox 2i ) 39) are already computed and that d k =0for all k 2 i. Then the single Newton iteration d =2 d d 2 i+1 d 2 i+1 b. doubles the number of evaluated coefficients d k k =0,1,...,2 i+1 1 ) of the convolutionary inverse. Hence we finally conclude that the iterative Newton formula d =2 d d 2 i d 2 i b, i =2,3,..., log 2 m, 40) with the starting vector d of the form 1 d =, b ) 1 b 0 b 2,0,0..., 0 generates the required convolutionary inverse d =d 0,d 1,...,d m 1 ) 41)
7 Joanna Kapusta, Ryszard Smarzewski 13 ofb =b 0,b 1,...,b m 1 ), b 0 0. Since the computational complexity of the convolution 2 i is equal to Oi2 i ), it is clear that the computational complexity of algorithm 40) is of the order O mlog 2 m+ m 2 log m ) log 22 = Omlogm). 42) This completes the proof that the algorithm 35) is of the order ON logn). 3 Fast Bernstein-Lagrange transformation Now we establish a fast algorithm for evaluating the multivariate polynomial p n x)= ) n 1 x α 1 x) n α 1 43) α α Q n f α at the points x β =x 1,β1,x 2,β2,...,x d,βd ) with the coordinates of the form For this purpose, note that x i,j = λ i γ j i i =1,2,...,d, j=0,1,...,n i 1). 44) p n x)= f α α Q n β Q n α = β Q n x β = β Q n a β x β, α Q β+1 f α where the coefficients a β are given by the formula a β = n 1) n β 1) ) ) n 1 n α 1 x α+β 1) β α β ) ) n 1 n α 1 1) β α α β α β α f α 1) αβ α). α Q β+1 Hence one can use the multidimensional convolution in order to get the algorithm ) ) f f a = r p t =... r 1) p 1 ) 2) p 2 ) 3) d) p d t, 46) where t =t α ) α Qn,r=r α ) α Qn and p i =p i,0,p i,1,...,p i,ni 1) are defined by 45) and r α = α, t α = n 1) n α 1), α Q n, 47) p i,l = 1)l l) i =1,2,...,d, l =0,1,...,n i 1). 48) Therefore, it follows from 28) and 29) that the coefficients a β can be computed by the algorithm 46) of the order ON logn). Furthermore, by inserting formula 44)
8 14 Fast multidimensional Bernstein-Lagrange algorithms into 45), we get y α = p n x α )= β Q n a β λ β γ αβ. Hence, we obtain n d 1 n 2 1 n 1 1 y α =... a β b β w 1,α1 β 1 w 2,α2 β 2...w d,αd β d q α, 49) β d =0 whenever we set and q β = β 2=0 β 1=0 β d j γj, k b β = j=1 w j,l = 1 l γj k d β j 1 λ βj j j=1 Finally, formula 49) yields the following theorem. γ k j, β Q n, 50) j =1,2,...,d, l=0,1,...,n j 1). 51) Theorem 1. If T :f α ) α Qn y α ) α Qn denotes the d-dimensional Bernstein- Lagrange transformation with the points x α =x 1,α1,x 2,α2,...,x d,αd ) defined as in 44), then T can be evaluated by the algorithm ) ) ) a b) 1) w 1 2) w 2 3)... d) w d T : y = a =... f... r 1) p 1 q, 52) ) ) 2) p 2 ) 3) d) p d t, 53) where elements of b =b α ) α Qn,q=q α ) α Qn, w i =w i,ni 2,...,w i,0,w i,0,w i,1,..., w i,ni 1),r=r α ) α Qn,t =t α ) α Qn andp i =p i,ni 2,...,p i,0,p i,0,w i,1,..., p i,ni 1) are defined as in formulae 47), 48), 50) and 51). Moreover, this algorithm has the running time of ON logn))+odn), where N = n 1 n 2 n d. We note that the term OdN) in the running time is an estimate of all auxiliary computations 47), 48), 50) and 51), which do not use convolutions. For the completeness of consideration, we summarize the algorithm for computing the multidimensional Bernstein-Lagrange transformation in more detail. Algorithm 1. The d - dimensional Bernstein-Lagrange transformation T with respect to the points x α =x 1,α1,x 2,α2,...,x d,αd ), where α =α 1,α 2,...,α d ) Q n, n =n 1,n 2,..., n d ) and x i,j = λ i γ j i i =1,2,...,d, j=0,1,2,...,n i 1). Input: A hypermatrix f =f α ) α Qn, scalar vectors λ =λ 1,λ 2,...,λ d ) and γ = γ 1,γ 2,...,γ d ) in K d, and the vector n =n 1,n 2,...,n d ) of positive integers. Output: A hypermatrix y =y α ) α Qn of values y α = p n x α ). 1. Use 47) to evaluate r α,t α for each α Q n. 2. Use 48) to evaluate p j,l for j =1,2,...,d,l=0,1,...,n j 1.
9 Joanna Kapusta, Ryszard Smarzewski Perform the componentwise division v = f/r. 4. For i from 1 to d do the following: 4.1. Compute the partial hypermatrix convolution v = v i) p i. 5. Perform the componentwise multiplication a = v t. 6. Use 50) to evaluate b β,q β for each β Q n. 7. Use 51) to evaluate w j,l for j =1,2,...,d,l=0,1,...,n j Perform the componentwise multiplication u = a b. 9. For i from 1 to d do the following: 9.1. Compute the extended partial hypermatrix convolution u = u i) w i. 10. Perform the componentwise multiplication y = u q. 11. Return y). 4 Inverse multidimensional Bernstein-Lagrange transformation Now we consider the inversion of the multidimensional Bernstein-Lagrange transformation T 1 :y α ) α Qn f α ) α Qn. 54) If we know coefficients y α = px α ) of the Lagrange polynomial 6) at the knots of the form x α =x 1,α1,x 2,α2,...,x d,αd ) 55) x i,j = λ i γ j i,,2,..,d, j =0,1,...,n i 1, 56) then we can find the multivariate divided differences = p n [x 1,0,...,x 1,α1 ;...;x d,0,...,x d,αd ]= c α of the Newton polynomial p n x)= α 1=0α 2=0 d α i 1 c α α Q n j=0 α d =0 β Q α+1 y β 57) d α i x i,βi x i,j ) j=0,j β i x i x i,j ), 58) by using an algorithm of the order ON logn))+odn) presented in [12]. Moreover, by using equality 56) the formula 58) can be rewritten in the following form ) n 1 n 2 n d d α i 1 p n x)=... c α1,α 2,...,α d x αi i 1 λ iγ j i. 59) x i Since we have see [13]) n 1 1 xq k ) = n [ n m] m=0 q j=0 1) m x m q mm 1) 2 60)
10 16 Fast multidimensional Bernstein-Lagrange algorithms with [ n m] q = n ) 1 q i [n] q = m 1 q i ) n m 1 q i [n m] ) q [m] q, it follows from 59) and 60) that where a β = 1 [β] q Hence we get p n x)= n 1 β 1 1n 2 β 2 1 α 1=0 α 2=0 a α = 1 [β] q n 1 1n 2 1 β 1=0 β 2=0... n d 1 d... a β1,β 2,...,β d n d β d 1 α d =0 α Q n β c α β d =0 x βi i, d [α i +β i ] q α i α i 1) 2 c α1,α 2,...,α d γi λ i ) αi. [α i ] q [α+β] q [α] q γ αα 1) 2 λ) α,α Q n, or equivalently ) ) ) a =... c v) d) d 1) d 2) 1) z d zd 1 z1 g, 62) where the elements of v =v α ), g =g α ) and z i =z i,0,z i,1,...,z i,ni 1) are defined by v α = z i,l = d αiαi 1) 1 [α i ] q γ 2 i λ i ) αi, g α = 1 [α] q, α Q n, d [n i l] q i =1,2,...,d,l=0,1,...,n i 1), and the reversed i-th partial hypermatrix convolution 61) 63) w i) z i = ŵ i) z i 64) is defined as the i-th partial hypermatrix convolution with its elements written in the reverse order, where ŵ is the hypermatrix with i-th column written in the reverse order, too. Finally, one can apply 46) to get the following theorem. Theorem 2. Let T 1 :y α ) α Qn f α ) α Qn be the inverse multidimensional Bernstein-Lagrange transformation with respect to the points x α =x 1,α1,x 2,α2,...,x d,αd ) with the coordinates of the form x i,j = λ i γ j i,,2,..,d, j =0,1,2,...,n i 1, 65)
11 Joanna Kapusta, Ryszard Smarzewski 17 where λ i 0, γ i 1andγ i 0. Then it can be evaluated by the algorithm a ) ) T 1 : f =... t d) p d ) d 1) p d ) p 1 r, 66) ) ) ) a =... c v) d) d 1) d 2) 1) z d zd 1 z1 g, 67) where the elements of t =t α ) α Qn,r =r α ) α Qn,g=g α ) α Qn,c=c α ) α Qn, v =v α ) α Qn,z i =z i,0,z i,1,...,z i,ni 1) and p i =p i,0,p i,1,...,p i,ni 1) are defined as in formulae 47), 48), 57) and 63). Moreover, this algorithm has the running time of ON logn)) +OdN), where N = n 1 n 2 n d. Algorithm 2. The inverse d-dimensional Bernstein-Lagrange transformation T 1 with respect to the points x α =x 1,α1,x 2,α2,...,x d,αd ), where α =α 1,α 2,...,α d ) Q n,n=n 1,n 2,..., n d ) and x i,j = λ i γ j i i =1,2,...,d, j=0,1,2,...,n i 1). Input: A hypermatrix y =y α ) α Qn, scalar vectors λ =λ 1,λ 2,...,λ d ), γ =γ 1,γ 2,...,γ d ) in K d, and the vector n =n 1,n 2,...,n d ) of positive integers. Output: A hypermatrix f =f α ) α Qn. 1. Use Algorithm 8 from [12] to evaluate c α for each α Q n. 2. Use 63) to evaluate v α,g α for each α Q n. 3. Use 63) to evaluate z j,l for j =1,2,...,d,l=0,1,...,n j Perform the componentwise multiplication v = c v. 5. For i from d down to 1 do the following: 5.1. Compute the reversed partial hypermatrix convolution v = v i) z i. 6. Perform the componentwise multiplication a = v g. 7. Use 47) to evaluate t α,r α for each α Q n. 8. Use 48) to evaluate p j,l for j =1,2,...,d,l=0,1,...,n j Perform the componentwise division a = a/t. 10. For i from d down to 1 do the following: Compute the partial hypermatrix deconvolution a = a i) p i. 11. Perform the componentwise multiplication f = a r. 12. Return f). 5 Conclusions and remarks In this paper, we present two new algorithms for the d-dimensional Bernstein- Lagrange transformation and its inverse for the points with the coordinates defined by the formulae x α =x 1,α1,x 2,α2,...,x d,αd ),α Q n 68) x i,j = λ i γ j i,,2,...,d, j=0,1,...,n i 1, where γ i 0, γ i 1and λ i 0are fixed.
12 18 Fast multidimensional Bernstein-Lagrange algorithms Roughly speaking, the main feature of these algorithms consists in splitting the computations into two steps. In the first step we compute only quantities, which require to perform only OdN) operations. The second step includes computations of d-dimensional convolutions or deconvolutions of the order ON logn). Thus, the computational complexity of this algorithms takes only ON logn)+odn) 69) operations, where N = n 1 n 2...n d. Moreover, if we make natural assumption that n i 2 for i =1,2,...,d, then log 2 N d and the order of the algorithm can be reduced to ON logn)). It should be emphasized, that parts 53) and 66) of the algorithms presented in Theorems 1 and 2 are valid for arbitrary points x α,α Q n. However, we do not know if the remaining parts of these algorithms are true for the points defined in 11). References [1] Stoer J., Bulirsch R., Introduction to Numerical Analysis, Springer - Verlag, New York [2] Kapusta J., Smarzewski R., Fast algorithms for multivariate interpolation and evaluation at special points, Journal of Complexity ): 332. [3] Farouki R., Rajan V. T., Algorithms for polynomials in Bernstein form, Computer Aided Geometric Design ): 1. [4] Mainara E., Pe na J. M., Evaluation algorithms for multivariate polynomials in Bernstein-Bézier form, Journal of Approximation Theory 143 1) 2006): 44. [5] Peters J., Evaluation and approximate evaluation of the multivariate Bernstein-Bézier form on a regularly partitioned simplex, ACM Transactions on Mathematical Software 204) 1994): 460. [6] Phien H. N., Dejdumrong N., Efficient algorithms for Bé zier curves, Computer Aided Geometric Design ): 247. [7] Aho A., Hopcroft J., Ullman J., The design and analysis of computer algorithms, Addison-Wesley, London [8] Smarzewski R., Kapusta J., Fast Lagrange-Newton transformations, Journal of Complexity ): 336. [9] Bini D., Pan V. Y., Polynomial and matrix computations: fundamental algorithms, Birkhäuser Verlag, [10] Aho A., Steiglitz K., Ullman J., Evaluating polynomials at fixed sets of points, SIAM Journal Comput ): 533. [11] Borwein J. M., Borwein P. B., Pi and the AGM: A study in analytic number theory and computational complexity, Canadian Mathematical Society Series of Monographs and Advanced Texts, John Wiley and Sons, New York, Chichester, Brisbane, Toronto, Singapore, [12] Kapusta J., An efficient algorithm for multivariate Maclaurin-Newton transformation, Annales Informatica AI VIII 2) 2008): 5. [13] Andrews G. E., The theory of partitions Encyclopedia of mathematics and its applications), Addison-Wesley Publishing Company, Powered by TCPDF
An efficient algorithm for multivariate Maclaurin Newton transformation
Annales UMCS Informatca AI VIII, 2 2008) 5 14 DOI: 10.2478/v10065-008-0020-6 An effcent algorthm for multvarate Maclaurn Newton transformaton Joanna Kapusta Insttute of Mathematcs and Computer Scence,
More informationCPSC 518 Introduction to Computer Algebra Schönhage and Strassen s Algorithm for Integer Multiplication
CPSC 518 Introduction to Computer Algebra Schönhage and Strassen s Algorithm for Integer Multiplication March, 2006 1 Introduction We have now seen that the Fast Fourier Transform can be applied to perform
More informationCPSC 518 Introduction to Computer Algebra Asymptotically Fast Integer Multiplication
CPSC 518 Introduction to Computer Algebra Asymptotically Fast Integer Multiplication 1 Introduction We have now seen that the Fast Fourier Transform can be applied to perform polynomial multiplication
More informationPolynomial Interpolation
Polynomial Interpolation (Com S 477/577 Notes) Yan-Bin Jia Sep 1, 017 1 Interpolation Problem In practice, often we can measure a physical process or quantity (e.g., temperature) at a number of points
More informationFFT: Fast Polynomial Multiplications
FFT: Fast Polynomial Multiplications Jie Wang University of Massachusetts Lowell Department of Computer Science J. Wang (UMass Lowell) FFT: Fast Polynomial Multiplications 1 / 20 Overview So far we have
More informationNumerical Methods with MATLAB
Numerical Methods with MATLAB A Resource for Scientists and Engineers G. J. BÖRSE Lehigh University PWS Publishing Company I(T)P AN!NTERNATIONAL THOMSON PUBLISHING COMPANY Boston Albany Bonn Cincinnati
More informationNumerical Analysis: Solving Systems of Linear Equations
Numerical Analysis: Solving Systems of Linear Equations Mirko Navara http://cmpfelkcvutcz/ navara/ Center for Machine Perception, Department of Cybernetics, FEE, CTU Karlovo náměstí, building G, office
More informationarxiv:math/ v1 [math.co] 13 Jul 2005
A NEW PROOF OF THE REFINED ALTERNATING SIGN MATRIX THEOREM arxiv:math/0507270v1 [math.co] 13 Jul 2005 Ilse Fischer Fakultät für Mathematik, Universität Wien Nordbergstrasse 15, A-1090 Wien, Austria E-mail:
More informationBarycentric coordinates for Lagrange interpolation over lattices on a simplex
Barycentric coordinates for Lagrange interpolation over lattices on a simplex Gašper Jaklič gasper.jaklic@fmf.uni-lj.si, Jernej Kozak jernej.kozak@fmf.uni-lj.si, Marjeta Krajnc marjetka.krajnc@fmf.uni-lj.si,
More informationFast Polynomial Multiplication
Fast Polynomial Multiplication Marc Moreno Maza CS 9652, October 4, 2017 Plan Primitive roots of unity The discrete Fourier transform Convolution of polynomials The fast Fourier transform Fast convolution
More informationProblem Set 5 Solutions
Problem Set 5 Solutions Section 4.. Use mathematical induction to prove each of the following: a) For each natural number n with n, n > + n. Let P n) be the statement n > + n. The base case, P ), is true
More informationScientific Computing
2301678 Scientific Computing Chapter 2 Interpolation and Approximation Paisan Nakmahachalasint Paisan.N@chula.ac.th Chapter 2 Interpolation and Approximation p. 1/66 Contents 1. Polynomial interpolation
More informationFundamental Theorem of Algebra
EE 387, Notes 13, Handout #20 Fundamental Theorem of Algebra Lemma: If f(x) is a polynomial over GF(q) GF(Q), then β is a zero of f(x) if and only if x β is a divisor of f(x). Proof: By the division algorithm,
More informationSolutions of the Harmonic Oscillator Equation in a B-Polynomial Basis
Adv. Studies Theor. Phys., Vol. 3, 2009, no. 12, 451-460 Solutions of the Harmonic Oscillator Equation in a B-Polynomial Basis Muhammad I. Bhatti Department of Physics and Geology University of Texas-Pan
More informationInterpolation and Deformations A short cookbook
Interpolation and Deformations A short cookbook 600.445 Fall 2000; Updated: 29 September 205 Linear Interpolation p ρ 2 2 = [ 40 30 20] = 20 T p ρ = [ 0 5 20] = 5 p 3 = ρ3 =? 0?? T [ 20 20 20] T 2 600.445
More informationDifferentiation matrices in polynomial bases
Math Sci () 5 DOI /s9---x ORIGINAL RESEARCH Differentiation matrices in polynomial bases A Amiraslani Received January 5 / Accepted April / Published online April The Author(s) This article is published
More informationOptimal global rates of convergence for interpolation problems with random design
Optimal global rates of convergence for interpolation problems with random design Michael Kohler 1 and Adam Krzyżak 2, 1 Fachbereich Mathematik, Technische Universität Darmstadt, Schlossgartenstr. 7, 64289
More informationInterpolation and polynomial approximation Interpolation
Outline Interpolation and polynomial approximation Interpolation Lagrange Cubic Splines Approximation B-Splines 1 Outline Approximation B-Splines We still focus on curves for the moment. 2 3 Pierre Bézier
More informationDivide and Conquer. Maximum/minimum. Median finding. CS125 Lecture 4 Fall 2016
CS125 Lecture 4 Fall 2016 Divide and Conquer We have seen one general paradigm for finding algorithms: the greedy approach. We now consider another general paradigm, known as divide and conquer. We have
More informationLinear and Integer Programming - ideas
Linear and Integer Programming - ideas Paweł Zieliński Institute of Mathematics and Computer Science, Wrocław University of Technology, Poland http://www.im.pwr.wroc.pl/ pziel/ Toulouse, France 2012 Literature
More informationApplication of the Bernstein Polynomials for Solving Volterra Integral Equations with Convolution Kernels
Filomat 3:4 (216), 145 152 DOI 1.2298/FIL16445A Published by Faculty of Sciences and Mathematics, University of Niš, Serbia Available at: http://www.pmf.ni.ac.rs/filomat Application of the Bernstein Polynomials
More informationTwo hours. To be provided by Examinations Office: Mathematical Formula Tables. THE UNIVERSITY OF MANCHESTER. 29 May :45 11:45
Two hours MATH20602 To be provided by Examinations Office: Mathematical Formula Tables. THE UNIVERSITY OF MANCHESTER NUMERICAL ANALYSIS 1 29 May 2015 9:45 11:45 Answer THREE of the FOUR questions. If more
More informationRecurrence Relations between Symmetric Polynomials of n-th Order
Applied Mathematical Sciences, Vol. 8, 214, no. 15, 5195-522 HIKARI Ltd, www.m-hikari.com http://dx.doi.org/1.12988/ams.214.47525 Recurrence Relations between Symmetric Polynomials of n-th Order Yuriy
More informationON A WEIGHTED INTERPOLATION OF FUNCTIONS WITH CIRCULAR MAJORANT
ON A WEIGHTED INTERPOLATION OF FUNCTIONS WITH CIRCULAR MAJORANT Received: 31 July, 2008 Accepted: 06 February, 2009 Communicated by: SIMON J SMITH Department of Mathematics and Statistics La Trobe University,
More informationThe discrete and fast Fourier transforms
The discrete and fast Fourier transforms Marcel Oliver Revised April 7, 1 1 Fourier series We begin by recalling the familiar definition of the Fourier series. For a periodic function u: [, π] C, we define
More information14 Lecture 14 Local Extrema of Function
14 Lecture 14 Local Extrema of Function 14.1 Taylor s Formula with Lagrangian Remainder Term Theorem 14.1. Let n N {0} and f : (a,b) R. We assume that there exists f (n+1) (x) for all x (a,b). Then for
More informationEquations with regular-singular points (Sect. 5.5).
Equations with regular-singular points (Sect. 5.5). Equations with regular-singular points. s: Equations with regular-singular points. Method to find solutions. : Method to find solutions. Recall: The
More informationApplication of Polynomial Interpolation in the Chinese Remainder Problem
Journal of Mathematical Research with Applications Jan., 017, Vol. 37, No. 1, pp. 90 96 DOI:10.3770/j.issn:095-651.017.01.008 Http://jmre.dlut.edu.cn Application of Polynomial Interpolation in the Chinese
More informationE.J. Barbeau. Polynomials. With 36 Illustrations. Springer
E.J. Barbeau Polynomials With 36 Illustrations Springer Contents Preface Acknowledgment of Problem Sources vii xiii 1 Fundamentals 1 /l.l The Anatomy of a Polynomial of a Single Variable 1 1.1.5 Multiplication
More informationON POSITIVE, LINEAR AND QUADRATIC BOOLEAN FUNCTIONS
Kragujevac Journal of Mathematics Volume 36 Number 2 (2012), Pages 177 187. ON POSITIVE, LINEAR AND QUADRATIC BOOLEAN FUNCTIONS SERGIU RUDEANU Abstract. In [1], [2] it was proved that a function f : {0,
More informationMatrix Methods for the Bernstein Form and Their Application in Global Optimization
Matrix Methods for the Bernstein Form and Their Application in Global Optimization June, 10 Jihad Titi Jürgen Garloff University of Konstanz Department of Mathematics and Statistics and University of Applied
More informationSymmetric functions and the Vandermonde matrix
J ournal of Computational and Applied Mathematics 172 (2004) 49-64 Symmetric functions and the Vandermonde matrix Halil Oruç, Hakan K. Akmaz Department of Mathematics, Dokuz Eylül University Fen Edebiyat
More informationInterpolation & Polynomial Approximation. Hermite Interpolation I
Interpolation & Polynomial Approximation Hermite Interpolation I Numerical Analysis (th Edition) R L Burden & J D Faires Beamer Presentation Slides prepared by John Carroll Dublin City University c 2011
More informationNodal bases for the serendipity family of finite elements
Foundations of Computational Mathematics manuscript No. (will be inserted by the editor) Nodal bases for the serendipity family of finite elements Michael S. Floater Andrew Gillette Received: date / Accepted:
More informationLinear Algebra 1. M.T.Nair Department of Mathematics, IIT Madras. and in that case x is called an eigenvector of T corresponding to the eigenvalue λ.
Linear Algebra 1 M.T.Nair Department of Mathematics, IIT Madras 1 Eigenvalues and Eigenvectors 1.1 Definition and Examples Definition 1.1. Let V be a vector space (over a field F) and T : V V be a linear
More informationTetrahedral C m Interpolation by Rational Functions
Tetrahedral C m Interpolation by Rational Functions Guoliang Xu State Key Laboratory of Scientific and Engineering Computing, ICMSEC, Chinese Academy of Sciences Chuan I Chu Weimin Xue Department of Mathematics,
More informationCSE 421 Algorithms. T(n) = at(n/b) + n c. Closest Pair Problem. Divide and Conquer Algorithms. What you really need to know about recurrences
CSE 421 Algorithms Richard Anderson Lecture 13 Divide and Conquer What you really need to know about recurrences Work per level changes geometrically with the level Geometrically increasing (x > 1) The
More informationFinite Elements. Colin Cotter. January 18, Colin Cotter FEM
Finite Elements January 18, 2019 The finite element Given a triangulation T of a domain Ω, finite element spaces are defined according to 1. the form the functions take (usually polynomial) when restricted
More informationMath Introduction to Numerical Analysis - Class Notes. Fernando Guevara Vasquez. Version Date: January 17, 2012.
Math 5620 - Introduction to Numerical Analysis - Class Notes Fernando Guevara Vasquez Version 1990. Date: January 17, 2012. 3 Contents 1. Disclaimer 4 Chapter 1. Iterative methods for solving linear systems
More informationPolynomial evaluation and interpolation on special sets of points
Polynomial evaluation and interpolation on special sets of points Alin Bostan and Éric Schost Laboratoire STIX, École polytechnique, 91128 Palaiseau, France Abstract We give complexity estimates for the
More informationSupermodular ordering of Poisson arrays
Supermodular ordering of Poisson arrays Bünyamin Kızıldemir Nicolas Privault Division of Mathematical Sciences School of Physical and Mathematical Sciences Nanyang Technological University 637371 Singapore
More informationNumerical Method for Solving Fuzzy Nonlinear Equations
Applied Mathematical Sciences, Vol. 2, 2008, no. 24, 1191-1203 Numerical Method for Solving Fuzzy Nonlinear Equations Javad Shokri Department of Mathematics, Urmia University P.O.Box 165, Urmia, Iran j.shokri@mail.urmia.ac.ir
More informationInverses. Stephen Boyd. EE103 Stanford University. October 28, 2017
Inverses Stephen Boyd EE103 Stanford University October 28, 2017 Outline Left and right inverses Inverse Solving linear equations Examples Pseudo-inverse Left and right inverses 2 Left inverses a number
More informationAppendix A Taylor Approximations and Definite Matrices
Appendix A Taylor Approximations and Definite Matrices Taylor approximations provide an easy way to approximate a function as a polynomial, using the derivatives of the function. We know, from elementary
More informationMATRICES. a m,1 a m,n A =
MATRICES Matrices are rectangular arrays of real or complex numbers With them, we define arithmetic operations that are generalizations of those for real and complex numbers The general form a matrix of
More informationBASIC MATRIX ALGEBRA WITH ALGORITHMS AND APPLICATIONS ROBERT A. LIEBLER CHAPMAN & HALL/CRC
BASIC MATRIX ALGEBRA WITH ALGORITHMS AND APPLICATIONS ROBERT A. LIEBLER CHAPMAN & HALL/CRC A CRC Press Company Boca Raton London New York Washington, D.C. Contents Preface Examples Major results/proofs
More information1 + lim. n n+1. f(x) = x + 1, x 1. and we check that f is increasing, instead. Using the quotient rule, we easily find that. 1 (x + 1) 1 x (x + 1) 2 =
Chapter 5 Sequences and series 5. Sequences Definition 5. (Sequence). A sequence is a function which is defined on the set N of natural numbers. Since such a function is uniquely determined by its values
More informationLecture Note 3: Interpolation and Polynomial Approximation. Xiaoqun Zhang Shanghai Jiao Tong University
Lecture Note 3: Interpolation and Polynomial Approximation Xiaoqun Zhang Shanghai Jiao Tong University Last updated: October 10, 2015 2 Contents 1.1 Introduction................................ 3 1.1.1
More information(0, 0), (1, ), (2, ), (3, ), (4, ), (5, ), (6, ).
1 Interpolation: The method of constructing new data points within the range of a finite set of known data points That is if (x i, y i ), i = 1, N are known, with y i the dependent variable and x i [x
More informationPolynomial approximation via de la Vallée Poussin means
Summer School on Applied Analysis 2 TU Chemnitz, September 26-3, 2 Polynomial approximation via de la Vallée Poussin means Lecture 2: Discrete operators Woula Themistoclakis CNR - National Research Council
More informationNumerical Analysis. A Comprehensive Introduction. H. R. Schwarz University of Zürich Switzerland. with a contribution by
Numerical Analysis A Comprehensive Introduction H. R. Schwarz University of Zürich Switzerland with a contribution by J. Waldvogel Swiss Federal Institute of Technology, Zürich JOHN WILEY & SONS Chichester
More informationChapter 1: Systems of linear equations and matrices. Section 1.1: Introduction to systems of linear equations
Chapter 1: Systems of linear equations and matrices Section 1.1: Introduction to systems of linear equations Definition: A linear equation in n variables can be expressed in the form a 1 x 1 + a 2 x 2
More informationGraeffe s, Chebyshev-like, and Cardinal s Processes for Splitting a Polynomial into Factors*
JOURNAL OF COMPLEXITY 12, 492 511 (1996) ARTICLE NO 0030 Graeffe s, Chebyshev-like, and Cardinal s Processes for Splitting a Polynomial into Factors* DARIO BINI Dipartimento di Matematica, Università di
More informationLectures notes. Rheology and Fluid Dynamics
ÉC O L E P O L Y T E C H N IQ U E FÉ DÉR A L E D E L A U S A N N E Christophe Ancey Laboratoire hydraulique environnementale (LHE) École Polytechnique Fédérale de Lausanne Écublens CH-05 Lausanne Lectures
More informationConvergence under Subdivision and Complexity of Polynomial Minimization in the Simplicial Bernstein Basis
Convergence under Subdivision and Complexity of Polynomial Minimization in the Simplicial Bernstein Basis Richard Leroy IRMAR, University of Rennes 1 richard.leroy@ens-cachan.org Abstract In this article,
More informationGiven a sequence a 1, a 2,...of numbers, the finite sum a 1 + a 2 + +a n,wheren is an nonnegative integer, can be written
A Summations When an algorithm contains an iterative control construct such as a while or for loop, its running time can be expressed as the sum of the times spent on each execution of the body of the
More informationPositive definite solutions of some matrix equations
Available online at www.sciencedirect.com Linear Algebra and its Applications 49 (008) 40 44 www.elsevier.com/locate/laa Positive definite solutions of some matrix equations Aleksandar S. Cvetković, Gradimir
More informationCover Page. The handle holds various files of this Leiden University dissertation
Cover Page The handle http://hdl.handle.net/1887/32076 holds various files of this Leiden University dissertation Author: Junjiang Liu Title: On p-adic decomposable form inequalities Issue Date: 2015-03-05
More informationNotes on Mathematics
Notes on Mathematics - 12 1 Peeyush Chandra, A. K. Lal, V. Raghavendra, G. Santhanam 1 Supported by a grant from MHRD 2 Contents I Linear Algebra 7 1 Matrices 9 1.1 Definition of a Matrix......................................
More informationMTS 105 LECTURE 5: SEQUENCE AND SERIES 1.0 SEQUENCE
MTS 105 LECTURE 5: SEQUENCE AND SERIES 1.0 SEQUENCE A sequence is an endless succession of numbers placed in a certain order so that there is a first number, a second and so on. Consider, for example,
More informationA chain rule for multivariate divided differences
A chain rule for multivariate divided differences Michael S. Floater Abstract In this paper we derive a formula for divided differences of composite functions of several variables with respect to rectangular
More informationLECTURE NOTES ELEMENTARY NUMERICAL METHODS. Eusebius Doedel
LECTURE NOTES on ELEMENTARY NUMERICAL METHODS Eusebius Doedel TABLE OF CONTENTS Vector and Matrix Norms 1 Banach Lemma 20 The Numerical Solution of Linear Systems 25 Gauss Elimination 25 Operation Count
More informationCoding Theory and Applications. Solved Exercises and Problems of Cyclic Codes. Enes Pasalic University of Primorska Koper, 2013
Coding Theory and Applications Solved Exercises and Problems of Cyclic Codes Enes Pasalic University of Primorska Koper, 2013 Contents 1 Preface 3 2 Problems 4 2 1 Preface This is a collection of solved
More informationHermite Interpolation and Sobolev Orthogonality
Acta Applicandae Mathematicae 61: 87 99, 2000 2000 Kluwer Academic Publishers Printed in the Netherlands 87 Hermite Interpolation and Sobolev Orthogonality ESTHER M GARCÍA-CABALLERO 1,, TERESA E PÉREZ
More informationChapter 3 Vector Spaces and Linear Transformations
Chapter 3 Vector Spaces and Linear Transformations 3. Vector Spaces Recall that a vector in the xy plane is a line segment directed from the origin to a point in the plane as shown in Figure 3.. Recall
More informationJournal of Computational and Applied Mathematics
Journal of Computational and Applied Mathematics 233 2010) 1554 1561 Contents lists available at ScienceDirect Journal of Computational and Applied Mathematics journal homepage: wwwelseviercom/locate/cam
More informationConvex Quadratic Approximation
Convex Quadratic Approximation J. Ben Rosen 1 and Roummel F. Marcia 2 Abstract. For some applications it is desired to approximate a set of m data points in IR n with a convex quadratic function. Furthermore,
More informationNumerical Solution of Nonlocal Parabolic Partial Differential Equation via Bernstein Polynomial Method
Punjab University Journal of Mathematics (ISSN 116-2526) Vol.48(1)(216) pp. 47-53 Numerical Solution of Nonlocal Parabolic Partial Differential Equation via Bernstein Polynomial Method Kobra Karimi Department
More informationReliability of Coherent Systems with Dependent Component Lifetimes
Reliability of Coherent Systems with Dependent Component Lifetimes M. Burkschat Abstract In reliability theory, coherent systems represent a classical framework for describing the structure of technical
More informationABSTRACT. Department of Mathematics. interesting results. A graph on n vertices is represented by a polynomial in n
ABSTRACT Title of Thesis: GRÖBNER BASES WITH APPLICATIONS IN GRAPH THEORY Degree candidate: Angela M. Hennessy Degree and year: Master of Arts, 2006 Thesis directed by: Professor Lawrence C. Washington
More informationarxiv:math/ v1 [math.na] 12 Jul 2004
arxiv:math/0407177v1 [math.na] 12 Jul 2004 On improving the accuracy of Horner s and Goertzel s algorithms Alica Smoktunowicz and Iwona Wróbel Faculty of Mathematics and Information Science, Warsaw University
More informationA New Algorithm to Compute Terms in Special Types of Characteristic Sequences
A New Algorithm to Compute Terms in Special Types of Characteristic Sequences Kenneth J. Giuliani 1 and Guang Gong 2 1 Dept. of Mathematical and Computational Sciences University of Toronto at Mississauga
More informationSUMS OF ENTIRE FUNCTIONS HAVING ONLY REAL ZEROS
PROCEEDINGS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 00, Number 0, Pages 000 000 S 0002-9939(XX)0000-0 SUMS OF ENTIRE FUNCTIONS HAVING ONLY REAL ZEROS STEVEN R. ADAMS AND DAVID A. CARDON (Communicated
More informationPublic-key Cryptography: Theory and Practice
Public-key Cryptography Theory and Practice Department of Computer Science and Engineering Indian Institute of Technology Kharagpur Chapter 2: Mathematical Concepts Divisibility Congruence Quadratic Residues
More informationSOLUTION OF GENERALIZED LINEAR VECTOR EQUATIONS IN IDEMPOTENT ALGEBRA
, pp. 23 36, 2006 Vestnik S.-Peterburgskogo Universiteta. Matematika UDC 519.63 SOLUTION OF GENERALIZED LINEAR VECTOR EQUATIONS IN IDEMPOTENT ALGEBRA N. K. Krivulin The problem on the solutions of homogeneous
More informationIntroduction to Number Theory
INTRODUCTION Definition: Natural Numbers, Integers Natural numbers: N={0,1,, }. Integers: Z={0,±1,±, }. Definition: Divisor If a Z can be writeen as a=bc where b, c Z, then we say a is divisible by b or,
More informationParameter inequalities for orthogonal arrays with mixed levels
Parameter inequalities for orthogonal arrays with mixed levels Wiebke S. Diestelkamp Department of Mathematics University of Dayton Dayton, OH 45469-2316 wiebke@udayton.edu Designs, Codes and Cryptography,
More informationarxiv: v2 [math.co] 14 May 2017
arxiv:1606.07823v2 [math.co] 14 May 2017 A Note on the Alon-Kleitman Argument for Sum-free Subset Theorem Zhengjun Cao 1, Lihua Liu 2, Abstract. In 1990, Alon and Kleitman proposed an argument for the
More information4 Unit Math Homework for Year 12
Yimin Math Centre 4 Unit Math Homework for Year 12 Student Name: Grade: Date: Score: Table of contents 3 Topic 3 Polynomials Part 2 1 3.2 Factorisation of polynomials and fundamental theorem of algebra...........
More informationNUMERICAL METHODS FOR ENGINEERING APPLICATION
NUMERICAL METHODS FOR ENGINEERING APPLICATION Second Edition JOEL H. FERZIGER A Wiley-Interscience Publication JOHN WILEY & SONS, INC. New York / Chichester / Weinheim / Brisbane / Singapore / Toronto
More informationLinear Algebra and Matrix Inversion
Jim Lambers MAT 46/56 Spring Semester 29- Lecture 2 Notes These notes correspond to Section 63 in the text Linear Algebra and Matrix Inversion Vector Spaces and Linear Transformations Matrices are much
More informationTwo chain rules for divided differences
Two chain rules for divided differences and Faà di Bruno s formula Michael S Floater and Tom Lyche Abstract In this paper we derive two formulas for divided differences of a function of a function Both
More informationThe Essentials of CAGD
The Essentials of CAGD Chapter 4: Bézier Curves: Cubic and Beyond Gerald Farin & Dianne Hansford CRC Press, Taylor & Francis Group, An A K Peters Book www.farinhansford.com/books/essentials-cagd c 2000
More informationPolynomial Multiplication over Finite Fields using Field Extensions and Interpolation
009 19th IEEE International Symposium on Computer Arithmetic Polynomial Multiplication over Finite Fields using Field Extensions and Interpolation Murat Cenk Department of Mathematics and Computer Science
More informationAN ALGORITHM FOR COMPUTING FUNDAMENTAL SOLUTIONS OF DIFFERENCE OPERATORS
AN ALGORITHM FOR COMPUTING FUNDAMENTAL SOLUTIONS OF DIFFERENCE OPERATORS HENRIK BRANDÉN, AND PER SUNDQVIST Abstract We propose an FFT-based algorithm for computing fundamental solutions of difference operators
More information0 Introduction Let the matrix A C nn be given by all its n entries The problem is to compute the product Ab of a matrix A by input vector b C n Using
Fast algorithms with preprocessing for matrix{vector multiplication problems IGohberg and VOlshevsky School of Mathematical Sciences Raymond and Beverly Sackler Faculty of Exact Sciences Tel Aviv University,
More informationInterpolation and polynomial approximation Interpolation
Outline Interpolation and polynomial approximation Interpolation Lagrange Cubic Approximation Bézier curves B- 1 Some vocabulary (again ;) Control point : Geometric point that serves as support to the
More informationOn the BMS Algorithm
On the BMS Algorithm Shojiro Sakata The University of Electro-Communications Department of Information and Communication Engineering Chofu-shi, Tokyo 182-8585, JAPAN Abstract I will present a sketch of
More informationSylvester-Gallai Theorems for Complex Numbers and Quaternions
Sylvester-Gallai Theorems for Complex Numbers and Quaternions The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters Citation Elkies,
More informationChapter 5 Divide and Conquer
CMPT 705: Design and Analysis of Algorithms Spring 008 Chapter 5 Divide and Conquer Lecturer: Binay Bhattacharya Scribe: Chris Nell 5.1 Introduction Given a problem P with input size n, P (n), we define
More informationMethod of Multivariate Lagrange Interpolation for Generating Bivariate Bonferroni-Type Inequalities
R u t c o r Research R e p o r t Method of Multivariate Lagrange Interpolation for Generating Bivariate Bonferroni-Type Inequalities Gergely Mádi-Nagy a András Prékopa b RRR 10-2009, June 2009 RUTCOR Rutgers
More informationGomory Cuts. Chapter 5. Linear Arithmetic. Decision Procedures. An Algorithmic Point of View. Revision 1.0
Chapter 5 Linear Arithmetic Decision Procedures An Algorithmic Point of View D.Kroening O.Strichman Revision 1.0 Cutting planes Recall that in Branch & Bound we first solve a relaxed problem (i.e., no
More informationInterpolation and Approximation
Interpolation and Approximation The Basic Problem: Approximate a continuous function f(x), by a polynomial p(x), over [a, b]. f(x) may only be known in tabular form. f(x) may be expensive to compute. Definition:
More informationFast evaluation of iterated multiplication of very large polynomials: An application to chinese remainder theory
ANZIAM J. 48 (CTAC2006) pp.c709 C724, 2007 C709 Fast evaluation of iterated multiplication of very large polynomials: An application to chinese remainder theory D. Laing 1 B. Litow 2 (Received 30 August
More information15.083J/6.859J Integer Optimization. Lecture 10: Solving Relaxations
15.083J/6.859J Integer Optimization Lecture 10: Solving Relaxations 1 Outline The key geometric result behind the ellipsoid method Slide 1 The ellipsoid method for the feasibility problem The ellipsoid
More information2.1 Sets. Definition 1 A set is an unordered collection of objects. Important sets: N, Z, Z +, Q, R.
2. Basic Structures 2.1 Sets Definition 1 A set is an unordered collection of objects. Important sets: N, Z, Z +, Q, R. Definition 2 Objects in a set are called elements or members of the set. A set is
More informationarxiv: v2 [math.nt] 4 Nov 2012
INVERSE ERDŐS-FUCHS THEOREM FOR k-fold SUMSETS arxiv:12093233v2 [mathnt] 4 Nov 2012 LI-XIA DAI AND HAO PAN Abstract We generalize a result of Ruzsa on the inverse Erdős-Fuchs theorem for k-fold sumsets
More informationA Saturation Algorithm for Homogeneous Binomial Ideals
A Saturation Algorithm for Homogeneous Binomial Ideals Deepanjan Kesh and Shashank K Mehta Indian Institute of Technology, Kanpur - 208016, India, {deepkesh,skmehta}@cse.iitk.ac.in Abstract. Let k[x 1,...,
More informationNovel Approach to Analysis of Nonlinear Recursions. 1 Department of Physics, Bar-Ilan University, Ramat-Gan, ISRAEL
Novel Approach to Analysis of Nonlinear Recursions G.Berkolaiko 1 2, S. Rabinovich 1,S.Havlin 1 1 Department of Physics, Bar-Ilan University, 529 Ramat-Gan, ISRAEL 2 Department of Mathematics, Voronezh
More informationLecture 9: Submatrices and Some Special Matrices
Lecture 9: Submatrices and Some Special Matrices Winfried Just Department of Mathematics, Ohio University February 2, 2018 Submatrices A submatrix B of a given matrix A is any matrix that can be obtained
More information