Homepage Solution manuals Ivan Niven An Introduction to the Theory of Numbers Exercise 2.3.41* ($f : n \mapsto \sum_{0<a<n, (a,n) = 1} a$ is one-to-one)

Exercise 2.3.41* ($f : n \mapsto \sum_{0<a<n, (a,n) = 1} a$ is one-to-one)

Define f ( n ) as the sum of the positive integers less than n and prime to n . Prove that f ( m ) = f ( n ) implies that m = n .

Answers

Proof. By Problem 40, f ( n ) , the sum of the positive integers less than n and prime to n , is given, if n 2 , by

f ( n ) = ( n ) 2 .

Suppose that f ( m ) = f ( n ) , where m 2 , n 2 . Then

( n ) = ( m ) . (1)

If n = 1 , then f ( 1 ) = 0 < a < 1 a = 0 since the sum is empty. Therefore f ( m ) = 0 , so m = 1 . We can assume now that m 2 , n 2 .

Then the decompositions of n , m in prime factors are

n = p 1 a 1 p k a k , a i > 0 , p 1 > p 2 > > p k , m = q 1 b 1 q l b l , b j > 0 , q 1 > q 2 > > q l .

(It is more convenient here to arrange the prime numbers in descending order.)

Then (1) gives

P : = p 1 a 1 p k a k p 1 a 1 1 ( p 1 1 ) p k a k 1 ( p k 1 ) = q 1 b 1 q l b l q 1 b 1 1 ( q 1 1 ) q l b l 1 ( q l 1 ) ,

that is

P = p 1 2 a 1 1 p k 2 a k 1 ( p 1 1 ) ( p k 1 ) = q 1 2 b 1 1 q l 2 b l 1 ( q 1 1 ) ( q l 1 ) . (2) 

All prime factors of p 1 1 , , p k 1 are less that p 1 . It is important to note that the exponents 2 a i 1 are not zero, so that p i divides P for every index i . Therefore, the greatest prime factor of P is p 1 . Similarly, considering the right member, the greatest prime factor of P is q 1 . The unicity of the decomposition in prime factors shows that p 1 = q 1 .

Since p 1 does not divide any other factor p 2 2 a 2 1 , , p k 2 a k 1 , p 1 1 , , p k 1 . The exponent of p 1 in the prime decomposition of P is 2 a 1 1 , and similarly the exponent of q 1 ( = p 1 ) in the prime decomposition of P is 2 b 1 1 . Therefore

ν p 1 ( P ) = 2 a 1 1 = 2 b 1 1 ,

so a 1 = b 1 and p 1 2 a 1 1 = q 1 2 b 1 1 . If we simplify both members of (2) by p 1 2 a 1 1 = q 1 2 b 1 1 , and p 1 1 = q 1 1 , we obtain

P = p 2 2 a 2 1 p k 2 a k 1 ( p 2 1 ) ( p k 1 ) = q 2 2 b 2 1 q l 2 b l 1 ( q 2 1 ) ( q l 1 ) .

So we are in the same situation, which implies p 2 = q 2 , and a 2 = b 2 , p 2 2 a 2 1 = q 2 2 b 2 1 , by the same reasoning.

Without loss of generality, we can assume k l . Reasoning by induction, we can simplify, until exhaustion of one of the left member. If k < l , we obtain after all simplifications

1 = q k + 1 2 b k + 1 1 q l 2 b l 1 ( q k + 1 1 ) ( q l 1 ) > 1 .

This is impossible, so k = l . In this case, p i a i = q i b i for all i , 1 i k = l , so n = m .

The function f is injective. □

User profile picture
2024-08-17 09:39
Comments