Vous êtes sur la page 1sur 12

Basic Number Theory By Dharmendra Kumar

Introduction

We denote the set of natural numbers or positive integers by N and is dened as follows N = {1, 2, 3, 4, . . .} and the set of integers Z = {. . . , 4, 3, 2, 1, 0, 1, 2, 3, 4, . . .} nally Z0 denote the set of non negative integers {0,1,2,3,4,. . . } Denition Let a, c N.then a divides c if b N such that c = ab more generally, for a, c Z with a = 0, we say a | c if b Z such that c = ab The greatest common divisor of integers a and c with (a, b) = (0, 0) is a positive integer satisfying (i) d | a and d | b (ii) if e | a and e | b then e | d where e N, we denote d by d = (a, b) Example 1.1 we observe that (0, 1) = 1 Every natural number has at least two divisors 1 and . These are called improper divisors. A dividor of which is not improper is called proper divisor. for example 1 and 6 are improper divisors of 6 and 2 and 3 are proper divisors of 6. Denition Let n N. then n is called prime if (i) n > 1 (ii) n has no proper divisor, if n is not prime then n is called composite number Theorem 1.1 Every composite number n > 1 can be written as a product of primes 1

Proof The proof is by induction on n. The assertion is valid if n = 2. Let n 3. If n is prime, we are through. Assume that n is composite. We see that n has a proper divisor n1 with 1 < n1 < n. If n1 is not prime, then n1 has a proper divisor n2 wth 1 < n2 < n1 < n further n2 is a proper divisor of n. Continuing as above, we have 1 < nk < nk1 < ........ < n1 < n and nk is a divisor of n. This chain must be terminate. Consequently there exists a prime number p such that p | n. Now 1 < n < n. By hypothesis, we derive p that n is a product of primes and hence n is the product of primes p We arrange the set of prime numbers in the increasing order 2, 3, 5, 7, 11, . . .. We denote pi with i 1 the ith prime. Then p1 = 2, p2 = 3,... and p1 < p2 < p3 < p4 < . . . As we move along the seq of natural numbers, we see that the prime numbers occur more and more rare. A natural number n is composite if it is divisible by a prime number p < n. The number of primes < n increases with n. Therefore the possibility of n being composite increases with n. Infact there are arbitrary long chains. n! + 2, n! + 2, ....., n! + n where n 2 of composite numbers. But it is not true that all natural numbers are composites after a certain stage. This was proved by Euclid around 300 BC. Theorem 1.2 There are innitely many primes. Proof Let N 1 and we consider

a = p1 p2 ....pN + 1 By theorem 1, there exists a prime p dividing a. We see that p is a dierent from p1 , p2 ..., pN otherwise p divides 1 which is a contradiction. Thus for every N 1, we can nd a prime larger than pN . Hence the assertion follows. For x > 0, we denote

(x) =
px

where the sum is taken over all primes p is not exceeding x. Here the empty sum is interpreted as zero. Thus theorem 2 can be restated as. Theorem 1.3 limx (x) = . Primes have been studied extensively. There are problems on primes which are inreactable. On the other hand there are problems which can be either solved or solved practically. We mention few of them. Let a, c, m Z with m = 0. Then if a b(mod m) m | (a b) (i) If p is prime, then (p 1)! + 1 0 (mod p) Example 1.2 (i) 4! + 1 = 25 0(mod 5) (ii) 6! + 1 = 721 0(mod 7) The above theorem is attributed to Wilson. But it was stated by Waring in 1770 and in little later a proof was given by Lagrange. (ii) Let p be a prime and a N such that (a, p) = 1 then ap1 1(mod p) This result was proved by Picrre de Fermat. (iii) Every prime p 1(mod 4) can be written uniquely as a sum of of two squares. Thus 5 = 12 + 22 , 13 = 22 + 32 . This was again proved by Fermat. The sum of two squares can be congruent to 0,1 or 2 (mod 4). Therefore p 3(mod 4) can not be written as sum of two squares. (iv) Carl Friedrich Guuss conjectured at the age of 15 that

lim

(x)
x log x

=1

This was proved independently by Hadamard and Charles de Vallee Poussin in 1896. This is known as the Prime Number Theorem. (v) The dierence between any two consecutive primes is at least two other than 2 and 3. A well known conjecture states whether there are innitely many primes p such that p+2 is also prime. Such primes are so called twin primes and the above conjecture is known as Twin Prime Conjecture. So far this remains unproved. This has been modied to make it more accessible by Sievs methods due to Brun and Selberg. Selberg proved that there are innitely many positive integers n such that n is composed of at most three prime factor and n+2 is composed three prime factors counted with their multiplicity. Theorem 1.4 There are innitely many primes p 3(mod 4). Proof Let 3, 7, ...., Ps be the rst s primes 3(mod 4). Put Ps = p. Now we consider M = (4.3.7...p) 1 then M 3(mod 4) A product of positive integers each 1(mod 4) is again 1(mod 4). Therefore M must have a prime factor such that 1(mod 4). therefore it has to be 3(mod 3) and it is dierent from 3, 7, ...., p otherwise it divides 1 which is not the case. Dirichlet proved in 1839 the following: Let a and m > 0 be integers with (a, m) = 1. Then there are innitely many primes in the arithmatic progression. a, a + m, a + 2m, . . . Exercise 1.1 There are inntely many primes 5(mod6). We have proved that every integer n > 1 can be written as a product of primes. Infact Euclid showed that this representation of n as product if primes is unique up to order. Thus 4

300 = 2.2.3.5.5 = 2.3.2.5.5 This is a special feature of positive integers. We can explain it by the following example: Example 1.3 Let us consider positive integers 1(mod3) greater than 4, 7, 10 . . . the product of any two each 1(mod3) is again 1(mod3). An element form this set is called primitive if it has no divisor from this set other than itself. for example 4, 7, 10, 13 are primitive. But 16 is not since 16 = 4.4. It is clear that every element can be written as a product of primitive elements. But this representstion is need not be unique which is clear from following 100 = 4.25 = 10.10 Theorem 1.5 There are innitely many primes 1(mod4) The proof depends on the following theorem Theorem 1.6 Let n 2 and A be positive integer such that n | (A2 + 1) then n is sum two squares Let us assume the result of theorem 1.6 to prove theorem 1.5. We will prove this theorem in the next chapter. Proof Suppose that there are only primes 5, 13, ...p which are 1(mod4). We put M = A2 + 1 where A = 2.5.13...P . But by theorem 4 there exists prime p > 2 such that p | A2 + 1. Now applying theorem 6 to derive p = u2 + v 2 . Therefore p 1(mod4). Since the square of an integer is either 0(mod4) or 1(mod4). Now the assertion follows from the arguement of Euclid that p is dierent from 5, 13, ..P otherwise p divides 1, wihch is contradiction. We conclude this chapter by introducing Dirichlets Fundamental Principle known as Box Principle. Denition Let N 1. If there are N+1 objects in boxes, then there is at least one box containg two or more objects. This is known as Box Principle Theorem 1.7 (Dirichlet) Let be real irrational and m 1 be an integer. Then there exist integers h and k with 0 < k m such that 5

0 < | h | < k

1 . mk

Proof We consider the set {0.}, {1.},...,{m.}. The elements of the above set are distinct since is irrational and they are m + 1 in number. Further they lie in the element [0, 1). We partition [0, 1) into m subintervals.
1 2 1 [0, m ),[ m , m ),...[ m1 , m ) m m

By Box Principle, there exist integers 0 k1 m, 0 k2 m with k1 > k2 such that 0 < |{k1 } {k2 }| < We have 0 < |k1 [k1 ] k2 + [k2 ]| <
1 m 1 m 1 m

We put k = k1 k2 , h = [k1 ] [k2 ] then 0 k m and 0 < |k h| < 1 that is o < | h | < mk k

Corollary 1.8 Let be a real irrational. There are innitely many pairs (h,k) of relatively prime integers with k > 0 such that 0 < | h | < k
1 k2

Proof For m 1, we nd a pair (hm , km ) with 0 < km m such that 0 < |


hm | km

<

1 km

1 2 km

Further we may assume that (hm , km ) = 1 and the pairs (hm , km ) are innitely many in number since
1 mkm

0 as m

Exercise 1.2 Let m 1 and be irrational real. Then there exist integers h and k with 0 < k m such that | h | < k
1 (m+1)k

Chapter

In this chapter we show that the representation of n > 1 as a product of primes is unique(up to order). For this , we consider fractions of non negative c integers. Let a, b, c, d Z with bd = 0. Then the fractions a and d satisfy b
a b c = d, a b c > d, a b

<

c d

if ad = bc, ad > bc, ad < bc respectively. Further a is reduced fractions which are not identical can not b be equal. For this we introduce Farey fractions introduced by Farey in 1816. For N 1, we write FN the sequence order in non decreasing order of reduced fractions a [0, 1] with a Z0 , b N and b N . If two elements b in FN are equal, then their numerartors are not equal and we order them according to the increasing order of their sizes.we have F1 : 0 , 1 1 1
1 F2 : 0 , 1 , 1 1 2 2 F3 : 0 , 1 , 1 , 3 , 1 1 3 2 1 2 1 F4 : 0 , 1 , 1 , 1 , 3 , 3 , 1 1 4 3 2 4

FN is called Farey sequence of order N. We observe that all the fractions appearing in F1 ,F2 ,F3 are distinct. We shall prove that this is the case in general. This is an immediate consequence of following theorem.
l Theorem 2.1 Let h and m be two consecutive fractions in the Farey sek quence of order N. Then hm = kl 1 i.e.

h l = 1 k m Proof The proof is by induction on N. By direct checking, the asserstion is valid for 1 N 5. We assume that the result for N prove it for N + 1. / Let a be a reduced fraction in [0, 1] such that a FN then b N + 1. b b h l l Fuether we nd consecutive fractions k and m in FN such that h a m k b and the equality does not hold at both the fractions. We write 7

= ak bh 0 = am + bl 0 solving for a and b in ak bh = 0 am + bl = 0, we have


a h+l

b k+m

1 klhm

1 1

by induction hypothesis. Thus a = h + l and b = k + m thus


a b

h+l k+m

where 0, 0, + > 0 and (, ) = 1 since (a, b) = 1. Let = 0 then


a b

h k

which implies that = 1 since

a b

is reduced. Then b=kN

which contradicts b N + 1. Thus = 0. similarly = 0 thus


a b

h+k k+m

with 0, 0 and (, ) = 1

On the other hand


h k

h+l k+m

l m

for 1, 1 and (, ) = 1. Thus k + m N + 1 for , 1 Let


a b

FN +1 and

a b

FN . Then this implies that = = 1 /


a b

h+l k+m

Now h h+l h l = = 1 k k+m k m and h+l l h l = = 1 k+m m k m by induction hypothesis. Hence the assertion is valid for N + 1. This completes the proof. Corollary 2.2 Let a and b Then they are not equal. Proof Assume that
a b c d

be reduced fractions which are not identical.

c d

i.e. ad = bc By considering their fractional parts, we may assume that they belong to c [0, 1]. Further b = 0 since the fractions a and d are not identical. We may b a c suppose b < d. Then b , d Fd Denition For fractions l . m
h k

and

l , m

we call

h+l k+m

the mediant between

h k

and

Infact we have proved the following while proving Theorem 2.1 Theorem 2.3 The fractions in FN +1 but not in FN are mediants of fractions in FN Exercise 2.1 List F6 and F7 . 9

Theorem 2.4 The denominators of two consecutive fractions of FN add upto at least N + 1. Proof Let
h k

and

l m

be consecutive fractions in FN . We observe that


h k

<
l m

h+l k+m

<

l m

Therefore

h+l k+m

FN since /

h k

and

are consecutive hence

k + m N + 1. Exercise 2.2 1. If N > 1, no two consecutive fractions in FN have the same denominator. 2. If h , h and h are consecutive fractions belonging to the same Farey k k k sequence, then
h k

h+h k+k

We drive from theorem 2.1 the following result. Theorem 2.5 Let a 1, b 1 with (a, b) = 1 be integers. Then there exist integers x and y such that

ax+by=1. (2.1) Proof If a = b, then a = b = 1 and the assertion is follows by taking x = 1,y = 0. Further we may assume that 0 < a < b. Then a Fb . Let h b k be adjacent fraction in Fb
h k

<

a b

Then by theorem 1.1 h a = 1 k b 10

i.e. ak bh = 1 Now the assertion follows by taking x = k and y = h Exercise 2.3 1. Let (a, b) = 1 and c Z. Then the equation ax + by = c where x, y Z has a soluion. 2. Let (a, b) = d. Then the eqution ax + by = d where x, y Z has a soltion. Theorem 2.5 implies immediately Euclids lemma. Theorem 2.6 Let a,b and c be positive integers with (a, b) = 1 and a | bc. Then a | c Proof By Theorem 2.5, we nd x, y Z such that ax + by = 1 then acx + bcy = c. Since a | acx and a | bcy, we see that a | c Corollary 2.7 Let a 1, b 1 with (a, b) = 1 be integers. Then all the solution of eq. (2.1) in integers x and y are given by x = x0 bt, y = y0 at where x0 , y0 is a solution of (2.1) and t an integer. Proof First observe that x = x0 bt, y = y0 at is a solution of the (2.1) since

11

Vous aimerez peut-être aussi