Scribd
Upload
41 views9 pages

3 D and C - Karatsuba

This document discusses the Karatsuba algorithm for fast multiplication of large integers. It explains the naive multiplication approach, then describes how Karatsuba improves on this by dividing the integers into parts and combining the results, achieving a running time of θ(n1.58) rather than θ(n2). Examples are provided to demonstrate the algorithm.

Uploaded by

ctxxfx5ykz
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
41 views9 pages

3 D and C - Karatsuba

This document discusses the Karatsuba algorithm for fast multiplication of large integers. It explains the naive multiplication approach, then describes how Karatsuba improves on this by dividing the integers into parts and combining the results, achieving a running time of θ(n1.58) rather than θ(n2). Examples are provided to demonstrate the algorithm.

Uploaded by

ctxxfx5ykz
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

Design and Analysis of Algorithms

Divide and Conquer


(Karatsuba’s Fast Multiplication)

Dr. D. P. Acharjya
Professor, SCOPE
Office: SJT Annex 201E
Email: [email protected]

1/9/2024 Dr. D. P. Acharjya 1


Naïve Approach of Multiplication
 Consider two integers a and b
 Convert the integers into binary strings of same
length.
 Let two integers be a = 12 and b = 10
 The binary representation is given as:
 a = 12 = 1100 and b = 1010
 a  b = 1100  1010
 The output = 120 = 1111000
9 January 2024 Dr. D. P. Acharjya 2
Continued …

 Number of multiplications = 16 = 42
 If there are n bits, The computing time = (n2)
9 January 2024 Dr. D. P. Acharjya 3
Karatsuba’s Fast Multiplication
 Consider two integers a and b
 Divide both the integers into two equal parts such as
a = a1a2 and b = b1b2
 Compute A = a1b1
 Compute B = a2b2
 Compute C = (a1 + a2) (b1 + b2)
 Compute D = C – A – B
 The product ab = bn  A + bn/2  D + B
 Where b is the base of the integer and n is the length
of the integers.
9 January 2024 Dr. D. P. Acharjya 4
Computing Time
 Let the length of two integers a and b be n.
 Length of a1, a2, b1, and b2 be (n/2)
A = a1b1  Computing Time = T(n/2)
B = a2b2  Computing Time = T(n/2)
C = (a1 + a2) (b1 + b2)  Computing Time = T(n/2)
D = C – A – B  Computing Time = constant
ab =bn  A + bn/2  D + B  Computing Time = Constant
 The recurrence relation is
T(n) = 3T(n/2) + k
9 January 2024 Dr. D. P. Acharjya 5
Continued …

9 January 2024 Dr. D. P. Acharjya 6


Continued …

 The general multiplication takes (n2) whereas


Karatsuba’s multiplication takes time (n1.58) and hence
runs faster.
 The multiplication algorithm also works for binary
numbers.
1/9/2024 Dr. D. P. Acharjya 7
Numerical Illustration
 Consider two integers a = 41 and b = 42
 Therefore, 41 = a1a2 and 42 = b1b2
a1= 4, a2=1 and b1= 4, b2=2
 A = a1b1 = 4  4=16
 B = a2b2 = 1  2 =2
 C = (a1 + a2) (b1 + b2) = (4+1) (4+2)=30
 D = C – A – B = 30 – 16 – 2 =12
 ab = bn  A + bn/2  D + B = 102  16 + 101  12 + 2
= 1600 + 120 + 2 = 1722
Here, n = 2 (length of a and b) and b = base of decimals =10
1/9/2024 Dr. D. P. Acharjya 8
Note
 If the length of numbers is not even, the add 0
(zeros) at the left side of the numbers to make
even length.
 For numbers with length > 2, the algorithm is to
be applied recursively.
 For example: let a =312 and b =408
Thus a = a1a2= 312 = 0312; a1= 03 and a2=12
b = b1b2= 408 = 0408; b1= 04 and b2=08
 Solve recursively.
1/9/2024 Dr. D. P. Acharjya 9

You might also like