Presentation is loading. Please wait.

Presentation is loading. Please wait.

May 9, 2001Applied Symbolic Computation1 Applied Symbolic Computation (CS 680/480) Lecture 6: Multiplication, Interpolation, and the Chinese Remainder.

Similar presentations


Presentation on theme: "May 9, 2001Applied Symbolic Computation1 Applied Symbolic Computation (CS 680/480) Lecture 6: Multiplication, Interpolation, and the Chinese Remainder."— Presentation transcript:

1 May 9, 2001Applied Symbolic Computation1 Applied Symbolic Computation (CS 680/480) Lecture 6: Multiplication, Interpolation, and the Chinese Remainder Theorem Jeremy R. Johnson

2 May 9, 2001Applied Symbolic Computation2 Introduction Objective: To derive a family of asymptotically fast integer multiplication algorithms using polynomial interpolation –Karatsuba’s Algorithm –Polynomial algebra –Interpolation –Vandermonde Matrices –Toom-Cook algorithm –Polynomial multiplication using interpolation –Polynomial version of the Chinese Remainder Theorem References: Lipson, Cormen et al.

3 May 9, 2001Applied Symbolic Computation3 Karatsuba’s Algorithm Using the classical pen and paper algorithm two n digit integers can be multiplied in O(n 2 ) operations. Karatsuba came up with a faster algorithm. Let A and B be two integers with –A = A 1 10 k + A 0, A 0 < 10 k –B = B 1 10 k + B 0, B 0 < 10 k –C = A*B = (A 1 10 k + A 0 )(B 1 10 k + B 0 ) = A 1 B 1 10 2k + (A 1 B 0 + A 0 B 1 )10 k + A 0 B 0 Instead this can be computed with 3 multiplications T 0 = A 0 B 0 T 1 = (A 1 + A 0 )(B 1 + B 0 ) T 2 = A 1 B 1 C = T 2 10 2k + (T 1 - T 0 - T 2 )10 k + T 0

4 May 9, 2001Applied Symbolic Computation4 Complexity of Karatsuba’s Algorithm Let T(n) be the time to compute the product of two n-digit numbers using Karatsuba’s algorithm. Assume n = 2 k. T(n) =  (n lg(3) ), lg(3)  1.58 T(n)  3T(n/2) + cn  3(3T(n/4) + c(n/2)) + cn = 3 2 T(n/2 2 ) + cn(3/2 + 1)  3 2 (3T(n/2 3 ) + c(n/4)) + cn(3/2 + 1) = 3 3 T(n/2 3 ) + cn(3 2 /2 2 + 3/2 + 1) …  3 i T(n/2 i ) + cn(3 i-1 /2 i-1 + … + 3/2 + 1)...  cn[((3/2) k - 1)/(3/2 -1)] --- Assuming T(1)  c  2c(3 k - 2 k )  2c3 lg(n) = 2cn lg(3)

5 May 9, 2001Applied Symbolic Computation5 Divide & Conquer Recurrence Assume T(n) = aT(n/b) +  (n) T(n) =  (n) [a < b] T(n) =  (nlog(n)) [a = b] T(n) =  (n log b (a) ) [a > b]

6 May 9, 2001Applied Symbolic Computation6 Polynomial Algebra Let F[x] denote the set of polynomials in the variable x whose coefficients are in the field F. F[x] becomes an algebra where +, * are defined by polynomial addition and multiplication.

7 May 9, 2001Applied Symbolic Computation7 Interpolation A polynomial of degree n is uniquely determined by its value at (n+1) distinct points. Theorem: Let A(x) and B(x) be polynomials of degree m. If A(  i ) = B(  i ) for i = 0,…,m, then A(x) = B(x). Proof. Recall that a polynomial of degree m has m roots. A(x) = Q(x)(x-  ) + A(  ), if A(  ) = 0, A(x) = Q(x)(x-  ), and deg(Q) = m-1 Consider the polynomial C(x) = A(x) - B(x). Since C(  i ) = A(  i ) - B(  i ) = 0, for m+1 points, C(x) = 0, and A(x) must equal B(x).

8 May 9, 2001Applied Symbolic Computation8 Lagrange Interpolation Formula Find a polynomial of degree m given its value at (m+1) distinct points. Assume A(  i ) = y i Observe that

9 May 9, 2001Applied Symbolic Computation9 Matrix Version of Polynomial Evaluation Let A(x) = a 3 x 3 + a 2 x 2 + a 1 x + a 0 Evaluation at the points , , ,  is obtained from the following matrix-vector product

10 May 9, 2001Applied Symbolic Computation10 Matrix Interpretation of Interpolation Let A(x) = a n x n + … + a 1 x +a 0 be a polynomial of degree n. The problem of determining the (n+1) coefficients a n,…,a 1,a 0 from the (n+1) values A(  0 ),…,A(  n ) is equivalent to solving the linear system

11 May 9, 2001Applied Symbolic Computation11 Vandermonde Matrix V(  0,…,  n ) is non-singular when  0,…,  n are distinct.

12 May 9, 2001Applied Symbolic Computation12 Polynomial Multiplication using Interpolation Compute C(x) = A(x)B(x), where degree(A(x)) = m, and degree(B(x)) = n. Degree(C(x)) = m+n, and C(x) is uniquely determined by its value at m+n+1 distinct points. [Evaluation] Compute A(  i ) and B(  i ) for distinct  i, i=0,…,m+n. [Pointwise Product] Compute C(  i ) = A(  i ) * B(  i ) for i=0,…,m+n. [Interpolation] Compute the coefficients of C(x) = c n x m+n + … + c 1 x +c 0 from the points C(  i ) = A(  i ) * B(  i ) for i=0,…,m+n.

13 May 9, 2001Applied Symbolic Computation13 Interpolation and Karatsuba’s Algorithm Let A(x) = A 1 x + A 0, B(x) = B 1 x + B, C(x) = A(x)B(x) = C 2 x 2 + C 1 x + C 0 Then A(10 k ) = A, B(10 k ) = B, and C = C(10 k ) = A(10 k )B(10 k ) = AB Use interpolation based algorithm: –Evaluate A(  ), A(  ), A(  ) and B(  ), B(  ), B(  ) for  = 0,  = 1, and  = . –Compute C(  ) = A(  )B(  ), C(  ) = A(  ) B(  ), C(  ) = A(  )B(  ) –Interpolate the coefficients C 2, C 1, and C 0 –Compute C = C 2 10 2k + C 1 10 k + C 0

14 May 9, 2001Applied Symbolic Computation14 Matrix Equation for Karatsuba’s Algorithm Modified Vandermonde Matrix Interpolation

15 May 9, 2001Applied Symbolic Computation15 Integer Multiplication Splitting the Inputs into 3 Parts Instead of breaking up the inputs into 2 equal parts as is done for Karatsuba’s algorithm, we can split the inputs into three equal parts. This algorithm is based on an interpolation based polynomial product of two quadratic polynomials. Let A(x) = A 2 x 2 + A 1 x + A 0, B(x) = B 2 x 2 + B 1 x + B, C(x) = A(x)B(x) = C 4 x 4 + C 3 x 3 + C 2 x 2 + C 1 x + C 0 Thus there are 5 products. The divide and conquer part still takes time = O(n). Therefore the total computing time T(n) = 5T(n/3) + O(n) =  (n log 3 (5) ), log 3 (5)  1.46

16 May 9, 2001Applied Symbolic Computation16 Asymptotically Fast Integer Multiplication We can obtain a sequence of asymptotically faster multiplication algorithms by splitting the inputs into more and more pieces. If we split A and B into k equal parts, then the corresponding multiplication algorithm is obtained from an interpolation based polynomial multiplication algorithm of two degree (k-1) polynomials. Since the product polynomial is of degree 2(k-1), we need to evaluate at 2k-1 points. Thus there are (2k-1) products. The divide and conquer part still takes time = O(n). Therefore the total computing time T(n) = (2k-1)T(n/k) + O(n) =  (n log k (2k-1) ).

17 May 9, 2001Applied Symbolic Computation17 Asymptotically Fast Integer Multiplication Using the previous construction we can find an algorithm to multiply two n digit integers in time  (n 1+  ) for any positive . –log k (2k-1) = log k (k(2-1/k)) = 1 + log k (2-1/k) –log k (2-1/k)  log k (2) = ln(2)/ln(k)  0. Can we do better? The answer is yes. There is a faster algorithm, with computing time  (nlog(n)loglog(n)), based on the fast Fourier transform (FFT). This algorithm is also based on interpolation and the polynomial version of the CRT.

18 May 9, 2001Applied Symbolic Computation18 Polynomial Algebra mod a Polynomial A(x)  B(x) (mod f(x))  f(x)|(A(x) - B(x)) This equivalence relation partitions polynomials in F[x] into equivalence classes where the class [A(x)] consists of the set {A(x) + k(x)f(x), where k(x) and f(x) are in F[x]}. Choose a representative for [A(x)] with degree < deg(f(x)). Can choose rem(A(x),f(x)). Arithmetic –[A(x)] + [B(x)] = [A(x) + B(x)] –[A(x)] * [B(x)] = [A(x)B(x)] The set of equivalence classes with arithmetic defined like this is denoted by F[x]/(f(x))

19 May 9, 2001Applied Symbolic Computation19 Modular Inverses Definition: B(x) is the inverse of A(x) mod f(x), if A(x)B(x)  1 (mod f(x)) The equation A(x)B(x)  1 (mod f(x)) has a solution iff gcd(A(x),f(x)) = 1. In particular, if f(x) is irreducible, then F[x]/(f(x)) is a field. By the Extended Euclidean Algorithm, there exist u(x) and v(x) such that A(x)u(x) + B(x)v(x) = gcd(A(x),f(x)). When gcd(A(x),f(x)) = 1, we get A(x)v(x) + f(x)v(x) = 1. Taking this equation mod f(x), we see that A(x)v(x)  1 (mod f(x))

20 May 9, 2001Applied Symbolic Computation20 Polynomial Version of the Chinese Remainder Theorem Theorem: Let f(x) and g(x) be polynomials in F[x] (coefficients in a field). Assume that gcd(f(x),g(x)) = 1. For any A 1 (x) and A 2 (x) there exist a polynomial A(x) with A(x)  A 1 (x) (mod f(x)) and A(x)  A 2 (x) (mod g(x)). Theorem: F[x]/(f(x)g(x))  F[x]/(f(x))  F[x]/(g(x)). I.E. There is a 1-1 mapping from F[x]/(f(x)g(x)) onto F[x]/(f(x))  F[x]/(g(x)) that preserves arithmetic. A(x)  (A(x) mod f(x), A(x) mod g(x))

21 May 9, 2001Applied Symbolic Computation21 Constructive Chinese Remainder Theorem Theorem: If gcd(f(x),g(x)) = 1, then there exist E f (x) and E g (x) (orthogonal idempotents) –E f (x)  1 (mod f(x)) –E f (x)  0 (mod g(x)) –E g (x)  0 (mod f(x)) –E g (x)  1 (mod g(x)) It follows that A 1 (x) E f (x) + A 2 (x) E g (x)  A 1 (x) (mod f(x)) and  A 2 (x) (mod g(x)). Proof. Since gcd(f(x),g(x)) = 1, by the Extended Euclidean Algorithm, there exist u(x) and v(x) with f(x)u(x) + g(x)v(x) = 1. Set E f (x) = g(x)v(x) and E g (x) = f(x)u(x)

22 May 9, 2001Applied Symbolic Computation22 Polynomial Evaluation, Interpolation and the CRT Since A(x) = Q(x)(x-  ) + A(  ), A(x)  A(  ) (mod (x-  )) If   , then gcd((x-  ),(x-  )) = 1. Therefore, we can apply the CRT to find a quadratic polynomial A(x) such that A(  ) = y and A(  ) = z. A(x) = A(  ) E  (x) + A(  ) E  (x), where –E  (x) = (x -  )/(  -  ) –E  (x) = (x -  )/(  -  ) Observe that –E  (  ) = 1 and E  (  ) = 0, so that E  (x)  1 (mod (x-  )) and E  (x)  0 (mod (x-  )). The equivalent results hold for E  (x)

23 May 9, 2001Applied Symbolic Computation23 Multifactor CRT The CRT can be generalized to the case when we have n pairwise relatively prime polynomials. If f 1 (x),…,f n (x) are pairwise relatively prime, i.e. gcd(f i (x),f j (x)) = 1 for i  j, then given A 1 (x),…,A n (x) there exists a polynomial A(x) such that A  A i (x) (mod f i (x)). Moreover, there exist a system of orthogonal idempotents: E 1 (x),…,E n (x), such that E i (x)  1 (mod f i (x)) and E i (x)  0 (mod f j (x)) for i  j. A(x) = A 1 (x)E 1 (x) + … + A n (x)E n (x)

24 May 9, 2001Applied Symbolic Computation24 Lagrange Interpolation and the CRT Assume that  0,  1,…,  n are distinct and let f i (x) = (x-  i ). Then gcd(f i (x),f j (x)) = 1 for i  j. Let Then E i (x)  1 (mod f i (x)) and E i (x)  0 (mod f j (x)) for i  j, and Lagrange’s interpolation formula is A(x) = A (  0 )E 0 (x) + … + A(  n )E n (x)

25 May 9, 2001Applied Symbolic Computation25 Matrix Interpretation of Interpolation The previous system has a solution when the  0,…,  n are distinct. The solution can be obtained using Lagrange interpolation. In fact, the inverse of V(  0,…,  n ) is obtained from the idempotents


Download ppt "May 9, 2001Applied Symbolic Computation1 Applied Symbolic Computation (CS 680/480) Lecture 6: Multiplication, Interpolation, and the Chinese Remainder."

Similar presentations


Ads by Google