OBTAINING SQUARES FROM THE PRODUCTS OF NON-SQUARE INTEGERS

Similar documents
Discrete Structures Lecture Primes and Greatest Common Divisor

A NEW METHOD FOR FACTORING LARGE SEMI-PRIMES

2x 1 7. A linear congruence in modular arithmetic is an equation of the form. Why is the solution a set of integers rather than a unique integer?

Intermediate Math Circles March 6, 2013 Number Theory I

2x 1 7. A linear congruence in modular arithmetic is an equation of the form. Why is the solution a set of integers rather than a unique integer?

This is a recursive algorithm. The procedure is guaranteed to terminate, since the second argument decreases each time.

8 Primes and Modular Arithmetic

Dixon s Factorization method

Chapter 9 Mathematics of Cryptography Part III: Primes and Related Congruence Equations

Integer factorization, part 1: the Q sieve. D. J. Bernstein

Factoring. there exists some 1 i < j l such that x i x j (mod p). (1) p gcd(x i x j, n).

Lecture 7 Number Theory Euiseong Seo

[Part 2] Asymmetric-Key Encipherment. Chapter 9. Mathematics of Cryptography. Objectives. Contents. Objectives

Algorithms. Shanks square forms algorithm Williams p+1 Quadratic Sieve Dixon s Random Squares Algorithm

Arithmetic, Algebra, Number Theory

Math 131 notes. Jason Riedy. 6 October, Linear Diophantine equations : Likely delayed 6

Primes and Factorization

Factoring Algorithms Pollard s p 1 Method. This method discovers a prime factor p of an integer n whenever p 1 has only small prime factors.

Algorithms (II) Yu Yu. Shanghai Jiaotong University

a the relation arb is defined if and only if = 2 k, k

SQUARE PATTERNS AND INFINITUDE OF PRIMES

Number Theory. CSS322: Security and Cryptography. Sirindhorn International Institute of Technology Thammasat University CSS322. Number Theory.

5.1. EXAMPLE 1 Decide whether the first number listed is divisible by the second. (a) 45; 9

4 Number Theory and Cryptography

Elementary factoring algorithms


On the Composite Terms in Sequence Generated from Mersenne-type Recurrence Relations

Finding Divisors, Number of Divisors and Summation of Divisors. Jane Alam Jan

Frequency Domain Finite Field Arithmetic for Elliptic Curve Cryptography

Math 412: Number Theory Lecture 3: Prime Decomposition of

Numbers. 2.1 Integers. P(n) = n(n 4 5n 2 + 4) = n(n 2 1)(n 2 4) = (n 2)(n 1)n(n + 1)(n + 2); 120 =

Numbers and their divisors

An integer p is prime if p > 1 and p has exactly two positive divisors, 1 and p.

Intermediate Math Circles February 14, 2018 Contest Prep: Number Theory

CHAPTER 1 REAL NUMBERS KEY POINTS

Basic elements of number theory

Basic elements of number theory

Number Theory A focused introduction

Table of Contents. 2013, Pearson Education, Inc.

Polynomial Formula at the origin of the Prime Numbers Generator

Chuck Garner, Ph.D. May 25, 2009 / Georgia ARML Practice

10 Problem 1. The following assertions may be true or false, depending on the choice of the integers a, b 0. a "

Primes and Modular Arithmetic! CSCI 2824, Fall 2014!!

Wednesday, February 21. Today we will begin Course Notes Chapter 5 (Number Theory).

An Introduction to Proof-based Mathematics Harvard/MIT ESP: Summer HSSP Isabel Vogt

All variables a, b, n, etc are integers unless otherwise stated. Each part of a problem is worth 5 points.

NUMBER THEORY AND CODES. Álvaro Pelayo WUSTL

NORTHWESTERN UNIVERSITY Tuesday, Oct 6th, 2015 ANSWERS FALL 2015 NU PUTNAM SELECTION TEST

Math 5330 Spring Notes Congruences

Algebra Introduction to Polynomials

PUTNAM TRAINING NUMBER THEORY. Exercises 1. Show that the sum of two consecutive primes is never twice a prime.

is square-free. Suppose it were not and had a square factor x 2 p r 1. This is a contradiction, as the only possible primes dividing p r 1

Introduction to Number Theory

Number Theory. Modular Arithmetic

Math 3 Variable Manipulation Part 3 Polynomials A

For your quiz in recitation this week, refer to these exercise generators:

Clock Arithmetic. 1. If it is 9 o clock and you get out of school in 4 hours, when do you get out of school?

Mathematics of Public Key Cryptography

Numbers, Groups and Cryptography. Gordan Savin

EXAMPLES OF MORDELL S EQUATION

Winter Camp 2009 Number Theory Tips and Tricks

Number Theory Solutions Packet

Exercises Exercises. 2. Determine whether each of these integers is prime. a) 21. b) 29. c) 71. d) 97. e) 111. f) 143. a) 19. b) 27. c) 93.

Integers and Division

p = This is small enough that its primality is easily verified by trial division. A candidate prime above 1000 p of the form p U + 1 is

22. The Quadratic Sieve and Elliptic Curves. 22.a The Quadratic Sieve

3 The fundamentals: Algorithms, the integers, and matrices

MATH 4400 SOLUTIONS TO SOME EXERCISES. 1. Chapter 1

download instant at

MATH 2112/CSCI 2112, Discrete Structures I Winter 2007 Toby Kenney Homework Sheet 5 Hints & Model Solutions

Mathematics of Cryptography

21. Factoring Numbers I

Cryptography. Number Theory with AN INTRODUCTION TO. James S. Kraft. Lawrence C. Washington. CRC Press

cse 311: foundations of computing Fall 2015 Lecture 12: Primes, GCD, applications

Finding Prime Factors

Chapter 2. Divisibility. 2.1 Common Divisors

Number Tree LCM HCF Divisibility Rules Power cycle Remainder Theorem Remainder of powers a n b n Last and Second last digit Power of Exponents Euler s

Before we talk about prime numbers, we will spend some time with divisibility because there is

Solution Sheet (i) q = 5, r = 15 (ii) q = 58, r = 15 (iii) q = 3, r = 7 (iv) q = 6, r = (i) gcd (97, 157) = 1 = ,

Math 319 Problem Set #2 Solution 14 February 2002

MATHEMATICS IN EVERYDAY LIFE 8

cse 311: foundations of computing Spring 2015 Lecture 12: Primes, GCD, applications

On the Cardinality of Mersenne Primes

CS 5319 Advanced Discrete Structure. Lecture 9: Introduction to Number Theory II

Introduction to Cryptology. Lecture 20

KTH, NADA , and D1449 Kryptografins grunder. Lecture 6: RSA. Johan Håstad, transcribed by Martin Lindkvist

If something is repeated, look for what stays unchanged!

Massachusetts Tests for Educator Licensure (MTEL )

What is The Continued Fraction Factoring Method

Elliptic Curves Spring 2013 Lecture #12 03/19/2013

God may not play dice with the universe, but something strange is going on with the prime numbers.

Senior Math Circles Cryptography and Number Theory Week 2

Homework #2 solutions Due: June 15, 2012

Factorization & Primality Testing

Outline. Number Theory and Modular Arithmetic. p-1. Definition: Modular equivalence a b [mod n] (a mod n) = (b mod n) n (a-b)

Part 2 - Beginning Algebra Summary

MODULAR ARITHMETIC I

Congruence Classes. Number Theory Essentials. Modular Arithmetic Systems

part 2: detecting smoothness part 3: the number-field sieve

MTH 346: The Chinese Remainder Theorem

Transcription:

OBTAINING SQUARES FROM THE PRODUCTS OF NON-SQUARE INTEGERS The difference between two neighboring squares n 2 and (n+1) 2 is equal to 2n+1 for any integer n=1,2,3,. Thus the numbers generated by n 2 -A can never be perfect squares if A<2n. The question which now arises is Is it possible to multiply some of these non-square numbers together to form a perfect square?. The answer is yes and it forms the basis of most sieve factorization methods for large semi-primes. Let us look at the problem a bit more in detail. First we construct the following table in which we have calculated n 2-1 for the first ten integers starting with n= 2. We also have written down the exponent vector produced by a prime number factorization of these numbers- n n 2-1 Exponent Vector (n 2-1) 2 3 [01000] 3 8 [30000] 4 15 [01100] 5 24 [31000] 6 35 [00110] 7 48 [41000] 8 63 [02010] 9 80 [40100] 10 99 [02001] 11 120 [31100] Note the exponent vectors represent the powers of the prime products representing a number. Thus [0 2 0 1 0 ] represents the number 2 0 3 2 5 0 7 1 11 0 =63. What is noticed is that none of the numbers n 2-1 have all even elements in their exponent vectors and hence cannot represent perfect squares by themselves. However the product of several of these n 2-1 numbers can lead to perfect squares if the sum of the exponent elements are all even numbers. A brief inspection of the above table shows that the product of 24 35 63 120=6350400 produces the exponent vector [(3+3+0+0),(1+1+0+2),(1+1+0+0),(1+0+1+0),(0+0+1+1)]=[6 4 2 2]. Halving the elements in this last vector produces the square root of 6350400, namely, 2520. There are also other combinations in the table which produce perfect squares. The product 8 15 120 =14400 has the exponent vector [6 2 2 0] so that [3 1 1 0] produces the square root 8 3 5=120. In matrix form we can represent the above collection of exponent vectors as-

01000 30000 01100 31000 M= 00110for n 2-1= 41000 02010 40100 02001 31100 3 8 15 24 35 48 63 80 99 120 2 3 4 5 6 7 8 9 and n= 10 11 Each row represents the exponent vector of one of the n 2-1 numbers starting with n=2 and going through n=11. Thus the 8th row in the matrix represents the number 80=2 4 5 1. With an appropriate selection of different numbers n 2-1 one can form exponent vectors consisting of all even elements and hence will produce a perfect square. For the above matrix let us look at the additional exponent vector produced by adding rows 2 to row 4 to row 6.This produces the vector- V = 10200 so that we have the perfect square K=2 10 3 2 5 0 7 0 =9216=(96) 2 In finding perfect squares one wants to generally deal only with smooth values of n 2-1. That is, numbers where the highest prime power of each n 2-1 is kept as low as possible in order to minimize matrix size. This can be accomplished by just neglecting those generated numbers where we have vectors with elements beyond a fixed ithprime. The matrix M above is an example of where no row exceeds five elements. Hence we have a 5 th prime cut-off. Should a number produced by the generating formula violate a given smoothness condition it is removed from the list. Other exponent vectors with even elements generated from the above matrix are- V=[6,2,0], V=[8,2,2], and V=[14,2,2] These vectors produce the perfect squares K=576, 57600, and 3686400, respectively. Their square roots are 24, 240, and 1920. The number generator used above is just one of an infinite number of other generators one may use to generate exponent vectors. Perhaps one of the most interesting is the quadratic generator- =

This is recognized as a Diophantine Equation where x=n, y=sqrt(n 2 -N) and instead of 1 we replace it by the larger integer N. Using this formula as our number generator, we get a new table - x=n y 2 =x 2 -N ExponentVector x a >sqrt(n) x 2 a -N [x 2a x 3a x 5a x 7a ] x b x 2 b -N [x b2 x b3 x b5 x b7 ] x c x 2 c -N [x c2 x c3 x c5 x c7 ]... For N=15 and x a =4, x b =5, etc we find the matrix composed of y 2 =K=x 2 -N numbers with no more than 4 elements per row to be- 0000 1010 1 10 4 5 M= 0101 0002 1111 for x 2-15= 21 49 210 = 6 8 15 In this case the first and fourth row indicate the perfect squares K=1 and K=49. Adding together the second, third and fifths row produces another perfect square defined by V=[2 2 2 2]. It reads K=10 21 210=44100 and sqrt(k)=210. You will notice in these last three result that, if we define X=5 6 15=450, we can write things in the Diophantine form- That is- (X-sqrt(K))(X+sqrt(K))=const.N (450-210)(450+210)=240 660=4080 15 In terms of modular arithmetic these results can be expressed as- (X 2 -K) mod N =0 Thus (240) mod 15 and (660) mod 15 are both equal to zero. The vector V can be used to represent any positive integer including large ones such as the composite Mersenne Number- N=2 11-1=2047=23 89 Its exponent vector has all zero elements except for a one at position 9 and a one at position 24. That is-

N=[0000000010000000000000010] An even more compact form occurs for N=10 m where V=[m 0 m ]. Thus the googol which is defined as G=10 100 can be represented by the three element exponent vector V=[100 0 100 ]. Instead of using a generating formula for finding the M matrix, we can also construct matrixes based directly on all the positive integers 1, 2, 3, etc. but keeping only those values which are sufficiently smooth. Allowing no elements beyond a finite power for 5 we get the following 22 x 3 matrix for all allowed exponent vectors generated between n=2 and 50-100 010 200 001 110 300 020 101 210 011 400 M= 120 201 310 002 030 500 220 301 021 410 102 2 3 4 5 6 8 9 10 12 15 16 for n =18 20 24 25 27 32 36 40 45 48 50 We can use this matrix M to generate a great number of perfect squares although only a very few of the rows represent perfect squares to begin with. As an example consider the product- K=3 4 6 8 18 24 48=11943936 with the corresponding exponent vector V=[14 6 0] Its root is sqrt(k)=2 7 3 3 =128 27=3456. Another possibility arises from K=27 36 6 48 25 3 whose exponent vector reads V=[16 16 6 ]. Here K= 44079842304000000 and sqrt(k) = 2 8 3 8 5 3 = 209952000

Finally let us demonstrate how one goes about factoring larger semi-primes. Start with the number N=24139 whose sqrt(n)=155.3673... First we construct the M matrix using the numbers generator n 2 -N and restricting ourselves to vectors no longer than five elements. This produces- 01201 14001 825 1782 158 161 M= 15100 11400 02102 for n 2 -N= 2430 3750 5790 and n= 163 167 172 Adding together the elements in rows 3, 4 and 5 produces the exponent vector V=[2 8 6 0 2] which produces the perfect square K=2 2 3 8 5 6 7 0 11 2 =49617562500 with sqrt(k)=222750. The corresponding X=163 167 172=4682012. We thus have- (4682012-222750)(4682012+222750)=4459262 4904762=const(24139) which produces [4682012^2-49617562500] mod 24139=0. Also we can pick out the components p and q by performing the following operation involving the greatest common divisor- gcd(24139,4459262)=q=239 and gcd(24139,4904762)=p=101 That is we have factored the semi-prime into 24139=101 x 239. This procedure demonstrates essentially the route taken by all sieve factorization techniques for breaking large semi-primes N. The present factoring record lies somewhere around 500 digit length for N and requires months of computer runs with the fastest existing computers. It should be pointed out that one can factor the present number N=24139 even faster by using the variation of the Fermat technique which we discussed in a previous article. One needs to simply run the one line program- for n from 156 to 180 do {n, n^2-24139,ifactor(n^2-24139)}od; until a value of n is reached where the exponents of the ifactor terms produce an exponent vector whose elements are all even. This occurs at n=170 and yields the vector- V=[ 0200000020] Thus K=3 2 23 2 =4761 and sqrt(k)=69. So (170-69)(170+69)=N or p=101 and q=239.

U.H.Kurzweg April 2012