On some interesting tasks fron classical theory of numbers

Zurab Aghdgomelashvili
Doctor of Mathematics

Department of Mathematics of Faculty of Informatics and control

Systems of Georgian Technical University, 77, M. Kostava
Street, 0175, Tbilisi, Georgia

E-mail: [email protected]

Keywords: Prime number; Some Specific Diophantine equations; Pseudo-Fibonacci number.

Abstract: Expressions of type (𝑝𝑞 − 1)/(𝑝 − 1) and (a2 + ab + b2 ) , where a, b are natural and

p, q are prime numbers, are studied.

Introduction
Solving equations in integers is one of the most beautiful parts of mathematics. Over time,
many methods have been collected for solving specific Diophantine equations, but general methods
for investigating them were developed only in the XX century. In 1900, at the X International
Congress of Mathematicians, held in Paris, the famous German mathematician D. Hilbert formulated
23 key mathematical problems in his report. Hilbert's tenth problem was called as "The Task of
solving Diophantine equations". Its essence was as follows: is given a Diophantine equation with
integer coefficients and several unknowns. Does exist or not an algorithm with a finite number of
operations for solving this equation in integers.

We have shown that the entire class of tasks would be easily solved using one simple lemma;
pq − 1
in addition, is stated the task to study type expressions, where p and q are prime numbers,
p −1

with properties; some properties of the (a2 + ab + b2 ) type expression were studied, and so on.
The following issues are considered in the work:
– Obviously, for prime numbers p and q, it is of great interest to determine the quantity of those
pq − 1
prime divisors of the A = number that are less than p. To do this, we have considered:
p −1
Theorem 1. Let's say that p and q are odd prime numbers and p = 2q + 1. Then from the various
 2

pq − 1
prime divisors of number A = , taken separately, only one is less than p. A has at least two
p −1
different prime divisors.
Theorem 2. Let's say that p and q are odd prime numbers and p < 2q + 1. Then all prime divisors
pq − 1
of the number A = are greater than p;
p −1
Theorem 3. Let's say that q is an odd prime number and, p  N \ {1} p ]1; q] [q + 2;2q] , then

pq − 1
each of the various prime divisors of the number A = , taken separately, is greater than p;
p −1
Theorem 4. Let's say that q is a prime odd number and p  {q + 1; 2q + 1}, then from different
pq − 1
prime divisors of number A = , taken separately, only one of them is less than p. A has at least
p −1
two different prime divisors.
yz −1
Task 1. Let’s solve the equation 2 x = in natural x, y, z numbers. At the same time y
y −1
must be a prime number.
yz −1
Task 2. Let’s solve the equation 3x = in natural x, y, z numbers. At the same time y must
y −1
be a prime number.
yz −1
Task 3. Let’s solve the equation p x = , where p{5; 7; 11; 13; ...} prime number x,
y −1

y  N and y – is prime number.

 – is stated the Lemma, by that easily would be solved the class of tasks:
Lemma 1. Let’s say that a, b, n  N and (a, b) = 1. Let’s prove that if a n  0 (mod | a − b |) ,

or bn  0 (mod | a − b |) , then |a – b| = 1.

Let’s solve the equations ( – ) in natural x, y numbers:

z
x+ y x− y
I.   =x −y ;
z z
VI. ( x + y) = xy ;
 2 
II. ( x + y) z = (2 x) z + y z ; VII. ( x + y) x − y = y x ;

III. ( x + y) = (3x) + y ; VIII. ( x + y) = ( x − y) , ( x  y) ;

z z z y x

IV. ( y − x) x + y = x y , ( y  x) ; IX. ( x − y ) x + y = x x − y ;

V. ( y − x) x + y = y x , ( y  x) ; X. ( x + y) x − y = ( x − y) x , ( y  x) .
 3

theorem 5. If a, b  N, (a, b) = 1, then each divisor of (a 2 + ab + b2 ) will be as similar.

It should be mentioned that specialists of theory of numbers have found a way to solve a
number of Diophantine equations and found that many of them cannot be solved, but the general
method has not yet been found. By Hilbert were stated the Task on the possibility of the existence of
a universal method. In 1970, the famous Russian mathematician I. Matyasevich proved that there
could be no such general rule.
In the work also us cist in of our interesting study on Fibonacci numbers. Is introduced the
concept of Fibonacci number and is found some of their properties.
Basic part
Theorem 1. Let’s say that p and q are odd prime numbers and p = 2q + 1, from the various
pq − 1
prime divisors of number A = only one of them is less then p. A has at least two different
p −1
divisors.
Proof
Let’s firstly consider the equation
A = qt , (1)
pq − 1
where A = , p = 2q + 1, t  N and p, q – are odd prime numbers.
p −1
Due introduction we will obtain
(2q + 1)q − 1 = 2qt +1 . (2)
Due application of binomial theorem’s formula from (2) we will obtain
(2q)q + Cq1 (2q)q −1 + + Cqq −1  (2q ) = 2qt +1 . (3)

The left side of (3) would be contained on 2q 2 and is not divided on than 2q3 , thus t
+ 1  2 t  1. I.e. from (3) we will have
(2q + 1)q − 1  (2q)q  2q 2  2qt +1 . (4)
From (4) follows the presentation of A as (6.1) is impossible. From the system


 p − 1 = 2q
 p q − 1   A  0(mod q),
 , p − 1 = ( p − 1, q) = (2q, q) = q   (5)
 p − 1   A  2q.

 A = 1 + p + p 2 + + p q −1 = p − 1  p − 1 = 2q
q

 p −1
 4

From the (1) and (5) follow that A has at least two different prime divisors. Let’s now say that
A = p11 , p2 2 , , pk k , where 1,2 , ,k  N , p1, p2 , , pk − are different prime numbers and
p1  p2   pk .
Let’s say that to p belongs m exponent with module pi, where pi  q and i = 1, 2,..., k, then

 p q − 1 = ( p − 1)(1 + p + p 2 + + p q −1 ),  p q  1(mod pi ),
 q −1  m
1 + p + p + + p  0(mod pi )  p  1(mod pi ),
2

 m  
 p  1(mod pi ), m − is exponent m − is exponent
q − is odd prime number q − is odd prime number
 

q  0(mod m),  m = 1,
  (6)
q − is odd prime number  m = q.
As pi  q , are odd prime numbers and p − 1 = 2q  0(mod pi ) , thus m  1. I.e. m = q.

As exponent q of mutually prime p number at module pi will divide  ( pi ) = pi − 1 , so

pi − 1
pi − 1  0(mod q)  q  , i.e. q = p1 .
2
 pi − 1
 q 
2  pi − 1  p − 1  p  p ; i = 1, 2,..., k.

q = p − 1
i
2 2
 2
I.e. A has different from q all prime divisors greater than p. At the same time q is divisor of A.
Thus finally we have:
q = p1  p  p2  p3   pk . Q.E.D.
Theorem 2. Let’s say that p are q odd prime numbers and p < 2q + 1, then all prime divisors of
pq − 1
A= number is greater than p.
p −1

pq − 1
Let’s say that A = = p11 , p22 , , pkk , where p1, p2 , , pk − are different odd prime
p −1

numbers, p1  p2   pk and 1,2 , ,k  N . Let’s also say that p belongs to exponent m with
pi module i = 1, 2,..., k. We have
 p q − 1 = ( p − 1)(1 + p + p 2 + + p q −1 ),  p q  1(mod pi ),
 q −1  m
1 + p + p + + p  0(mod pi )  p  1(mod pi ),
2

 m  
 p  1(mod pi ), m − is exponent m − is exponent
q − is odd prime number q − is odd prime number
 
 5

q  0(mod m), m = 1
  
q − is odd prime number  m = q.
 p  2q + 1,  p − 1  2q,
  
 p, q − are odd prime numbers  p, q − are odd prime numbers

( p − 1, q) = 1,

 p −1
q  2 .

 p q − 1 
 , p − 1)  = ( p − 1, q) = 1,
 p − 1 
 q 
 p −1
 p − 1  0(mod pi ).

 p − 1  0(mod pi )  p  1(mod pi )  m  1 . I. e. m = q.
As exponent q of mutually prime p number at module pi will divide  ( pi ) = pi − 1 , thus
pi − 1
pi − 1  0(mod q)  q  .
2
We have
 p −1
q  2 ,
  p − 1 pi − 1
 pi − 1   ,  p  pi ,
q  ,  2 2  
 2  p1  p2   pk .  p1  p2 

 pk .
 p1  p2   pk .


 p  p1  p2   pk . Q.E.D.
Theorem 3. Let’s say that q is odd prime number and p  N \ {1} , p ]1; q] [q + 2;2q] , then
pq − 1
from the various prime divisors of number A = , taken separately, only one is more than
p −1
p.
Proof.
pq − 1
Let’s say that A = = p11 , p22 , , pkk , where p1, p2 , , pk − are different odd prime
p −1
divisors. At the same time, p1  p2   pk and 1,2 , ,k  N . Let’s alps say that p belongs to
exponent m with module p1 , then
 6

 p q − 1 = ( p − 1)(1 + p + p 2 + + p q −1 ),
 q  p q  1(mod pi ),
 p − 1  0(mod p ),  m
  p  1(mod pi ),
 p −1
i
 
 m m − is exponent,
 p  1(mod p1 ), m − is exponent q − is odd prime number.
q − is odd prime number 

q  0(mod m),  m = 1,
   m = q.
q − is odd prime number 
Accordingly of task conditions we have:
 p  N,  p  N,
 
 p  q + 1, p  2q + 1   p − 1  q, p − 1  2q, 
q − is odd prime number. q − is odd prime number.
 
( p − 1, q) = 1,

q − is odd prime number.

 p q − 1 
 , p − 1 = ( p − 1, q) = 1,
 p − 1 
 q
 p −1
  0(mod p1 ).  p − 1  0(mod p1)  p  1(mod p1)  m  1 .
 p − 1
 p1 − is odd prime number.



I.e. m = q.
As the exponent q of mutually prime number p at module p1 will be divide  ( p1) = p1 − 1 thus

 p1 − 1  0(mod q), p −1
  q 1 .
 p, q − is odd prime number 2
I.e. we have:
 p −1
q  2 ,
  p − 1 p1 − 1
 p1 − 1   ,  p  pi ,
q  ,  2 2  
 2  p1  p2   pk .  p1  p2 

 pk .
 p1  p2   pk .


 p  p1  p2   pk . Q.E.D.
 7

Theorem 4. Let’s say that q is odd prime number and p  {q + 1; 2q + 1}, then from the various

pq − 1
prime divisors of number A = , taken separately, only one is less than p. A has at least two
p −1
different prime divisors.
Proof.
Let’s firstly consider the equation
A = qt , (7)

pq − 1
where A = , p = kq + 1, k {1; 2}, t  N and q- is odd prime number.
p −1
By introduction we will obtain
(kq + 1)q − 1 = kqt +1 , k  {1; 2}. (8)
Due the formula of binomial theorem from (8) we will obtain
(kq)q + Cq1 (kq)q −1 + + Cqq −1(kq) = kqt +1 . (9)

The left side of (9) is divided on kq 2 and is not divided on kq 3 , so t + 1  2  t  1 t = 1.

I.e. we will have
(kq + 1)q − 1  (kq)2  kqt +1 . (10)
From (10) follows that representation of A as (7) is impossible.
From the system

 p − 1 = kq, k  {1;2},

 p q − 1   A  0(mod q),
 , p − 1 = ( p − 1, q) = (kq, q) = q,   (11)
 p − 1   A  kq.

 A = p − 1  p − 1 = kq.
q

 p −1

From the (11) follows that A has at least two different prime divisors. Let’s say that
A = p11 , p2 2 , , pk k , where p1, p2 , , pk − are different prime numbers, p1  p2   pk ,
1,2 , ,k  N . Let’s also say that p belongs to exponent m with module pi. There pi  q ,
i = 1,2, , k then we will have

 p q − 1 = ( p − 1)(1 + p + p 2 + + p q −1 ),  p q  1(mod pi ),
 q −1  m
1 + p + p + + p  0(mod pi ),  p  1(mod pi ),
2

 m  
 p  1(mod p 1 ), m − is exponent,  m − is exponent,
q − is odd prime number, q  p . q − is odd prime number.
 i 
 8

q  0(mod m),  m = 1,
 
q − is odd prime number.  m = q.
 pq − 1
  0(mod pi ),
 p − 1
 p q − 1 
 , p − 1)  = q,  p − 1  0(mod pi )  m  1 . I.e. m = q.
 p − 1 
 p  q.
 i

As at module pi the exponent of mutually prime number p will be divide  ( p2 ) = p2 − 1 , thus
 pi − 1  0(mod q), p −1
  q i . I. e. i  1 and q = p1 .
 pi , q − are odd prime numbers. 2

 pi − 1
q  2 , p − 1 pi − 1
    p  pi , i = 2, 3, ..., k.
q = p − 1 2 2
, k  {1; 2},
 k
Finally we have: q = p1  p  p2  p3   pk . Q.E.D.

Lemma 1. Let’s say that a, b, n  N and (a, b) = 1. Let’s prove that if a n  0 (mod | a − b |) ,

or bn  0 (mod | a − b |) , then | a − b |= 1 .
Proof.
a, b  N , a  b,(a, b) = 1 (| a − b |, a n ) = (| a − b |, b 4 ) = 1
   |a – b| = 1.
| a − b | a an | a − b | b a, b  N , a  b,(a, b) = 1
4 4

yz −1
Task 1. Let’s solve the equation 2 x = in natural numbers. In addition y will be the prime
y −1
number.

Solution:
Let’s consider the equation
yz −1
2x = , (12)
y −1
where y is prime number and x, z  N .
If z = 1, then from (12) 2 x = 1  x = 0  N ;
If z > 1 and y = 2, then the left side of (6.12) is even, and the right side is odd number. I.e. y
is odd prime number;
 9

If z > 1 is odd number and y is odd prime number, then the right side of (12) is odd number,
and the keft side – is even that is impossible;
If z is composite number that has the different from 1 odd divisor, i.e. let’s say that z = a b,
where a, b  N, b > 1 and b- is odd, then
( y a )b − 1 y a − 1
2 =
x
=  (1 + y a + ( y a )2 + + ( y a )b −1 ) . (13)
y −1 y −1
In the brackets of right side of (13) we have odd quantity of odd numbers sum. I.e. this
multiplier is odd and in the left side we have 2 x . This is impossible. Thus (12) has not in this case
the solution in natural numbers;
If z = 2n , where n  N \ {1}, then (12) will be as
n
y2 − 1 n −1
2 = x
= ( y + 1)( y 2 + 1)( y 4 + 1)   ( y2 + 1) . (14)
y −1

In this case (14) has at least two different multipliers (y + 1) and ( y 2 + 1) , for (12) having the

solution in natural numbers, for this (y + 1) and ( y 2 + 1) also simultaneously will represent the
natural powers of 2.
I. e.
 y + 1 = 2k ,

 2 (15)
y +1 = 2 ,

l

where k, l  N, k > 1, l >1.

From the first equation of (15) y – is odd prime number, thus by dividing ( y 2 + 1) on 4 in

remainder gives 2, and the second equation in right side of (15) 2l - is dividing without remainder on
4. I.e. in this case (12) has not the solution in natural numbers. It remains to consider the case when
z = 2.
In this case
y2 − 1
2x = = y + 1  y = 2x − 1. (16)
y −1
It would be easily shown that for a prime number y the x must necessarily be prime. I.e. (12)
has a solution according to the assignment if and only if x = P – is prime number and y = P – is a
Mersenne prime, or P = 2 p − 1 – is a prime number, and z is necessarily equal to 2.
From this it would be easily shown that if
 P1 − 1   P22 − 1   Pk − 1 
2n =  1    k ,
 P1 − 1   P2 − 1   Pk − 1 
 10

where P1, P2 , , Pk − are prime numbers and n, 1 , 2 , ,k  N .

Then obligatory:
1 = 2 = = k = 2 , P1 = 2 p1 − 1 , P2 = 2 p2 − 1, , Pk = 2 pk − 1 ,
where p1 , p2 , , pk − are prime numbers, p1 + p2 + + pk = n , P1, P2 , , Pk − are the Mersenne
primes,
yz −1
Task 2. Let’s solve the equation 3 = in natural x, y, z numbers. In addition y – must be
x

y −1
prime number.
Solution:
Let’s consider the equation
yz −1
3x = , (17)
y −1

where y – is a prime number and x, z  N.

If z = 1, then from (17) 3x = 1  x = 0  N .
It is obvious that y  3. Let’s say that y is odd prime number. Then it is obvious that it is
represented as 6a – 1 or 6a + 1, where a  N.
If y = 6a – 1 and z – is odd, then then at dividing of right side of (17) as remainder will be 1,
and the left side – is multiple to 3 that is impossible.
If y = 6a – 1 and z = 2n, n  N, then from (17) we will have:
(6a − 1) 2 n − 1 (6a − 1) 2 − 1
3x = =  ((6a − 1) 2( n −1) + (6a − 1) 2( n − 2) + + 1) =
(6a − 1) − 1 (6a − 1) − 1

= 6a((6a − 1)2( n −1) + (6a − 1)2( n −2) + + 1) ,

that is impossible, as the left side of this equation is not divided on 2.
If y = 6a + 1 and z – are odd, then from (17) we will obtain that z is dividing on 3. I.e. z=3q,
q  N. By introducing in (17) we will have
( y 3 )1 − 1
3x =  (( y 3 ) q −1 + ( y 3 ) q − 2 + + 1) = 3(3(4a 2 + 2a) + 1)(( y 3 ) q −1 + + 1) 
y −1

 3x  0(mod(3(4a2 + 2a) + 1) ,
that is impossible.
If y = 6a + 1 and z – are even, let’s say that z = 2n, n  N, then from (17) we have

(6a + 1)2 − 1
3x =  ((6a + 1)2( n−1) + (6a + 1)2( n−2) + + 1) =
(6a + 1) − 1
 11

= 2(3a + 1)((6a + 1)2( n −1) + + 1)  3x  0(mod2) ,

that is impossible;
It remains to consider the case y = 2. In this case, from (17) we will obtain
3x = 2z − 1. (18)
Let’s say that x > 1.
If x –odd, then from (18) we will have:
2z = 3x + 1 = (3 + 1)(3x−1 − 3x−2 + + 1) . (19)
In the second bracket of right side of (19) we have odd quantity of odd numbers sum, i.e. odd
number, thus (19) has not the sp; in natural numbers;
If x = 2n, n  N, then from (18) we will obtain:
2z = 32n + 1 = (4 − 1)2n + 1 = 4k + 2 , (20)
where k  N.
At dividing of right side of (20) on 4 the remainder gives 2, and the left side is dividing on 4,
thus also in this case will not the solution of (18):
By us was considered the case for x = 1. In this case we will have
yz −1
31 = . (21)
y −1
It is obvious that (21) has unique solution y = z = 2. Finally we will obtain that the solution of
(17), by stated condition is only: x = 1; y = z = 2.
Task 3. Let’s solve the equation
yz −1
Px =
,
y −1
where P  {5; 7; 11; ...} – is a prime number, x, z  N and y – are prime numbers.

Solution:
Let’s consider the equation
yz −1
Px = , (22)
y −1

where P  {5; 7; 11; ...} –is prime number, y – is prime and x, z  N.

Firstly let’s show that if ( y n − 1) is dividing on ( y m − 1) , where y  1 and m, n  N, then

obligatory n is dividing on m.
Proof.
Let’s say that n = am + r, where m, n, a, r  N and 0  r < m. Then we will have
 12

y n − 1 = y am  y r − 1 = y r ( y am − 1) + ( y r − 1) = y r ( y m − 1)( y m( a −1) + + 1) + ( y r − 1) . (23)

For fulfilment of condition of task, is necessary to divide ( y r − 1) on ( y m − 1) , but it is

impossible, when r = 0 and then we will have n = am, where a , m  N. I.e. finally we will obtain
provable or n  0(mod n). Q.E.D.
From the proved and taking into account that P- is a prime number, it is easy to show that (22)
has the solution by stated condition, if z- is a prime number, or z = 2n , where n  N.
If z = 2n , n  N, then (22) will be as
n
y2 − 1
P =
x
. (24)
y −1

y2 − 1
For n = 1 P x = = y + 1  y = P x − 1 if we take into account that y and P- are prime
y −1

numbers, then P = 2 and x = P1 , where P1 − is a prime number and y = 2 p1 − 1 – is Messene prime;

If n > 1, then (22) will be as

n
y2 − 1 n−1
P =x
= (1 + y )(1 + y 2 )   (1 + y 2 ) . (25)
y −1

In this case in the right side (25) has at least two multipliers (y + 1) and ( y 2 + 1) . For having

to (25) the solution in natural numbers, for this either (y + 1) and ( y 2 + 1) simultaneously will be
presented by any natural power of P. I.e.

 y + 1 = P a ,
 2  Pb = ( y + 1)2 − 2 y = P2a − 2 y . (26)
 y + 1 = P
b

As (P, y) = 1, thus from (26) follows that P  0(mod 2). But P- is prime number, so accordingly
of task P =2 and (26) would be written down as

2b = 22a − 2 y  y = 2b −1 (22a −b − 1) . (27)

y- is a prime number, thus from the (27) we will have: b =1, 2a − 1 = P1 – are prime numbers

and y = 2 p1 − 1 – represents a Messene prime.

It would be mentioned that
x
yP z −1
 y + 1, (28)
x
Therefore if  N , then (22) has not the solution!
z −1
 13

Let’s now the case, when z – is a prime number.

I.e. we will have
yz −1
Px = ,
y −1
where x  N and P, z, y – are prime numbers.
We have
 y z − 1 
 , y − 1 = ( y − 1, z ).
 y − 1 

 y, z , P − prime numbers,  P − 1  0(mod z)  P  1(mod z)  P = 2kz + 1 ,
 z
 y − 1  0(mod P),
 y −1

where P, z – are prime numbers and k  N.
Let’s consider the several cases
1 If P = 5 = 2  2 + 1, i.e. z = 2 , then from (22) we will have
y2 − 1
5x = = y + 1  y = 5 x − 1  0(mod 4) .
y −1
This is impossible, as y – is a prime number.
2 If P = 7 = 2  3 + 1 , i.e. z = 3 . Then from (22) we have
y3 − 1
7 =
x
= y2 + y + 1 . (29)
y −1
Let’s consider the identities:
(a2 + ab + b2 )(c2 + cd + d 2 ) = (ac − bd )2 + (ac − bd )(ad + bc + bd ) +
+(ad + bc + bd )2 = (ad − bc)2 + (ad − bc)(ac + bd + bc) + (ac + bd + bc)2 ; (30)
(a2 + ab + b2 )2 = (a2 − b2 )1 + (b2 + 2ab)(a2 − b2 ) + (b2 + 2ab)2 =
= (b2 − a2 )2 + (b2 − a2 )(a2 + 2ab) + (a2 + 2ab)2 ; (31)
(a2 + ab + b2 )3 = (a3 − 3b2a − b3 )2 + (a3 − 3b2a − b3 )(3ab(a + b)) + (3ab(a + b))2 . (32)
72 = 12 + 1  2 + 22 .
By this formulae we will have:
(12 + 1 2 + 22 )2 = 32 + 3  5 + 52 , (12 + 1 2 + 22 )3 = 12 + 118 + 182 .
(12 + 1 2 + 22 )5 = (12 + 118 + 182 )(32 + 3  5 + 52 ) = 72 + 7 126 + 1262 .
Let’s say that we have:
 14

|1  b − 2  a |= 1
1 + p + p 2 = (1 + 2 + 22 )(a 2 + ab + b 2 ) a  1 + 2  b + 2  a = p
 2 
a + ab + b = (1 + 2 + 2 )(c + cd + d )  |1  c − 2  d |= a
2 2 2 2

a, b, c, d  N , b  a, c  d 1  d + 2  c + 2  d = b
 
a, b, c, d  N , b  a, c  d

b = 2a  1, p = 7 a  2 b = 2a  1, p = 7a  2
c = 2d + a c = 2d − a  d
 
From this we obtain:  or  , from that
 7 d = 1 7 d = 4 a  1  4 d + 1
a, b, c, d  N , b  a, c  d a, b, c, d  N , b  a,

1 + p + p2
no one has the solution in natural numbers. I.e.  N . Thus last presentation of
(12 + 1  2 + 22 )3

(1 + p + p2 ) as (12+12+22)m for m = 3 (12+12+22)3 = 12 + 118 + 182.

I.e. the solutions of (29) are only: x = 1; y = 2; z = 3.
Task 4. Let’s solve the equation

x+ y
z

 =x −y
z z
 (33)
 2 
in natural numbers.
Solution:
Let’s rewrite (33) as
( x + y) z = 2 z ( x z − y z ) . (34)

For z > 1, from (6.34) we will obtain x + y = 2x – 2y  x = 3y. I.e. the solutions of (33) are:
x = 3t, y = t, where t  N;
for z  1 let’s say that sols of (34) are: x0 , y0 , z1 and x0 = x1d , y0 = y1d , where d = ( x0 , y0 ) .

By introduction in (34) and dividing of two sides on d z1 we will obtain

( x1 + y1 ) z1 = 2 z1 ( x1z1 − y1z1 ) . (35)

The right side of (25) is a multiple of 2. Therefore x1 and y1 numbers are odd.

I.e.
x1 = 2a − 1 end y1 = 2b − 1 , where a, b  N. (36)

By introduction of (36) in (35) and simplification we will obtain

(a + b − 1) z1 = (2a − 1) z1 − (2b − 1) z1 . (37)

(2a −1) z1 − (2b −1) z1 = ((2a −1) − (2b −1))(2a −1) z1 −1 + + (2b −1) z1 −1 ) =
 15

= 2(a − b)((2a − 1) z1 −1 + + (2b − 1) z1 −1) , (38)

(2b − 1) z1 = (2a − 1) z1 − (a + b − 1) z1 = ((2a − 1) − (a + b − 1))((2a − 1) z1 −1 +
(
+ (a + b − 1) z1 −1 ) = (a − b) (2a − 1) z1 −1 + + (a + b − 1) z1 −1 ) (39)

From (37), (38) and (39) we have that (a − b) \ (2a − 1) z1 and (a − b) \ (2b − 1) z1 , but ((2a
– 1), (2b – 1)) = 1, therefore accordingly of Lemma 1 by introduction in a – b = 1 or a = b + 1 (37).
(2b) z1 = (2b + 1) z1 − (2b − 1) z1  (2b − 1) z1 + (2b) z1 = (2b + 1) z1 . (40)
If z1 = 1 , then from (40)

(
(2b) z1 = ( (2b + 1) − (2b − 1) ) (2b + 1) z1 −1 + )
+ (2b − 1) z1 −1 =

(
= 2 (2b + 1) z1 −1 + )
+ (2b − 1) z1 −1  2 z1 −1  b z1 = (2b + 1) z1 −1 + + (2b − 1) z1 −1 . (41)

If z1  1 is an odd natural number, then the right side of (41) is an odd amount of the sum of

odd numbers, or an odd number, and the left side is even, so for odd z1 numbers other than 1, this
equation has no solution in natural numbers x, y, z.
If z1 is even, i.e. z1 = 2t , then (40) will be as

(2b − 1)2t + (2b)2t = (2b + 1)2t . (42)

If t = 1, then we have (2b − 1)2 + (2b)2 = (2b + 1)2 , from that z1 = b = 2 .

I.e. the solution of (35) are:

 x1 = 3,  x1 = 5,
 
 y1 = 1, or  y1 = 3,
 z = 1.  z = 2.
1  1
Thus the solutions of (33) will be:
 x = 3k ,  x = 5k ,
 
 y = k, or  y = 3k , (43)
 z = 1.  z = 2,
 
where k  N.
If t > 1, then from the binomial theorem formula from (42) is easy to see that b \ t.
By the dividing of both sides of (42) on (2b)2t we will obtain:
2t 2t
 1   1   1  2t t t
2  1 −  + 1 = 1 +   1 + C21t   = 1 + =1+ 1   t  b,
 2b   2b   2b  2b b b
but b \ t, that is impossible.
Finally we have that all sons of (33) gives (43).
Task 5. Let’s solve the equation
 16

( x + y) z = (2x) z + y z , (44)
in natural numbers.
Solution:

Firstly let’s say that z > 2 and solutions of (44) are: x0 , y0 , z1  N , z1  2 . At the same time

x0 = x1d , y0 = y1d , where d = ( x0 ; y0 ) . By dividing of both sides of (44) on d z1 we will obtain:

( x1 + y1 ) z1 = (2 x1 ) z1 + y1z1  y z1 = ( x1 + y1 ) z1 − (2 x1 ) z1 =

(
= ( ( x1 + y1 ) − 2 x1 ) ( x1 + y1 ) z1 −1 + ( x1 + y1 ) z1 − 2 (2 x1 ) + )
+ (2 x1 ) z1 −1 =

(
= ( y1 − x1 ) ( x1 + y1 ) z1 −1 + ( x1 + y1 ) z1 − 2  (2 x1 ) + )
+ (2 x1 ) z1 − 2  ( y1 − x1 ) \ y1z1 ,

In addition, ( x1, y1 ) = 1 . Thus accordingly of Lemma 6.1 y1 − x1 = 1 or y1 = x1 + 1.

By introduction in (44) we have

(2 x1 + 1) z1 = (2 x1 ) z1 + ( x1 + 1) z1 . (45)

The left side of (45) is odd, there fire ( x1 + 1) will be odd, i.e. x1 = 2k , k  N. By introduction in

(45) we will obtain

(4k + 1) z1 = (4k ) z1 + (2k + 1) z1 . (46)

If z1 is odd, then from (46)

(
(4k + 1) z1 = (4k + (2k + 1)) (4k ) z1 −1 − (4k ) z1 − 2 (2k + 1) + + (2k + 1) z1 −1  )
 (6k + 1) \ (4k + 1) z1 ,

but (6k + 1; 4k + 1) = 1 (indeed 6k + 1 = (4k + 1) + 2k)). I.e. z1 is even or z1 = 2n , where n  N, n

 1. Due taking it into account (46) will be as:

( (4k + 1) ) = ( (4k ) ) + ((2k + 1) )

n 2 n 2 n 2
. (47)

(4k + 1), 4k and (2k + 1) pairwise are coprime numbers. Thus would be found such p, q  N,
p > q and (p, q) = 1, that:

(4k + 1)n = p2 + q2 , (4k )n = 2 pq , (2k + 1)n = p2 − q2 . (48)

From (48) we have
2q2 = ( p2 + q2 ) − ( p2 − q2 ) = (4k + 1)n − (2k + 1)n =
= ( (4k + 1) − (2k + 1) ) ( (4k + 1) n −1 + + (2k + 1) n −1 ) =

= 2k ( (4k + 1) n −1 + + (2k + 1) n −1 )  k \ q 2 . (49)

 17

As (p, q) = 1 and k \ q2 , therefore (k, p) = 1.

(4k )n = 2 pq  22n −1 k n = pq . (60)

If k = 1 , then (46) will be as:

5z1 = 4z1 + 3z1 or

 4  z1  3  z1
  +   = 1,
 5   5 
 z1 z1 2 2
 z1  2   4  +  3    4  +  3  = 1,
 5 5  5 5
 z1 z1 2 2
 z1 = 2 (51)
    
4 3    
4 3
 z1 = 2   5  +  5  =  5  +  5  = 1

 z1
 4 3
z1 2
 4 3
2

 z1  2    +      +   = 1.
 5 5 5 5
If k  1 , then from (49) and (50) k is odd and p = 22n −1 , q = k n . From (48) we will obtain
(4k + 1)n = (22n −1 )2 + (k n )2 .
From that
24 n − 2 = (4k + 1) n − (k 2 ) n = (4k + 1 − k 2 ) ( (4k + 1) n −1 + (4k + 1) n − 2 k 2 + + (k 2 ) n −1 ) . (52)

From (52) follow that 4k + 1 − k 2  1  k ]0;4[ . At the same time k is other than 1 odd
number. I.e. k = 3. By introduction in (47) we will obtain
132n = 122n + 72n . (53)
If n = 1, then 132n = 132  122 + 72 = 122n + 72n .
If n > 1, then from (53) we have
122n = (132 )n − (72 )n = 169n − 49n = (169 − 49)(169n−1 + 160n−249 + + 49n−1 ) =
= 120(169n −1 + + 49n −1 ) . (54)
The right side of (54) is multiple to 5, and the left side not, thus (54) has no solution in natural
numbers.
Finally we have that if z > 2, then (44) has no solution in natural numbers.
If z = 1, then from (44) we will obtain x + y = 2x + y  x = 0, that is impossible, as x  N .
If z = 2, then
( x + y)2 = (2x)2 + y 2  3x = 2 y , (55)
From that we will obtain that all natural solutions of (44) are:
x = 2t, y = 3t, z = 2,
where t  N.
 18

Task 6. Let’s solve the equation

( x + y) z = (3x) z + y z , (56)
in natural numbers.
Solution:
Let’s say that the solutions of (56) are
x1, y1, z0  N and ( x1, y1 ) = d , x1 = x0d , y1 = y0d . (57)

By introduction of (57) in (56) and simplification we will obtain:

( x0 + y0 ) z0 = (3x0 ) z0 + y0z0 . (58)

From (58) if z0  1 we will have:

(
y0z0 = ( y0 − 2 x0 ) ( x0 + y0 ) z0 −1 + + (3x0 ) z0 −1 . ) (59)

From that
y0z0  0(mod( y0 − 2 x0 )) . (60)

If y0 - is odd, then accordingly of Lemma 1 from (60) we have:

y0 − 2 x0 = 1 or y0 = 2 x0 + 1 . (61)

By introduction of (61) in (58) we will obtain:

(3x0 + 1) z0 = (3x0 ) z0 + (2 x0 + 1) z0 . (62)

If z0 - is odd, then from (62) we will have:

(
(3x0 + 1) z0 = (5 x0 + 1) (3x0 ) z0 −1 − + (2 x0 + 1) z0 −1 ,) (63)

( ) (
5 x0 + 1 = (3x0 + 1) z0 ,5 x0 + 1 = (5 x0 + 1 − 2 x0 ) z0 ,5 x0 + 1 = )
= ( (2 x )
0
z0
)
,5 x0 + 1 = (2 z0 ,5 x0 + 1) . (64)

I.e. if z0 - is odd ( z0  1) , then for (62) to have a solution, then x0 also will be odd. From (63)

and (64) follows that 5 x0 + 1 = 2 z0 , as in the second bracket on the right side of (63) would have an
odd amount of sums of odd numbers, or odd number. At the same time
(3x0 + 1) z0 = (3(2k + 1) + 1) z0 = 2z0 (3k + 2) z0 and in the right side even is only (5x0 + 1) multiplier.
I.e.
5 x0 = 2 z0 − 1 . (65)

For odd z0 the right side of (65) is ended by 1 or 7, thus (65) will not be executed for odd x0

and z0 . I.e. z0 - is an even.

 19

As the equation (an )4 + (bn )4 = (cn )4 has no solution in a, b, c, n  N (this know fact is proven

due infinite assumptions method), thus even z0 would be as

z0 = 4(n – 1) + 2, n  N. (66)
From (62) and (66) we will have
(9 x02 + 6 x0 + 1)2 n −1 = (9 x02 )2 n −1 + (4 x02 + 4 x0 + 1)2 n −1 . (67)
If n > 1, n  N, then from (67) we will have:
(6 x0 + 1)(9 x02 + 6 x0 + 1)2 n − 2 + (9 x02 + 6 x0 + 1)2 n −3  (9 x02 ) + + (9 x02 )2 n − 2 ) =

= (2 x0 + 1)4n − 2 . (68)
As (6x0+1, 2x0 + 1)=(6x0 + 1, 6x0 +2) = (2,6x0 + 3) = 1, thus (68) has no solution accordingly of
above mentioned condition.
If z0 = 1 , then from (68)

( x0 + y0 ) = (3x0 ) + y0  x0 = 0  N .

If z0 = 2 , then from (62)

(3x0 + 1)2 = (3x0 )2 + (2 x0 + 1)2  9 x02 + 6 x0 + 1 =

= 9 x02 + 4 x02 + x0 + 1  x0 = 0  N .
Finally we have that (56) accordingly of this condition has no solution.
Task 7. Let’s solve the equation
( y − x) x + y = x y , ( y  x) (69)
in natural numbers.
Solution:
Let’s say that solutions of stated equation are x0 , y0 and at the same time x0 = x1d , y0 = y1d ,

where d = ( x0 , y0 ) . Due introduction in the stated equation, we will obtain:

y1d
 x 
d x1d
( y1 − x1 ) x1d
= 1  . (70)
 y1 − x1 
From this
y1
 x 
d ( y1 − x1 ) =  1  .
x1 x1
(71)
 y1 − x1 
From (71) ( y1 − x1 ) \ x1 . Accordingly of Lemma 1

y1 − x1 = 1  y1 = x1 + 1 . (72)

By introduction of (72) in (71) and simplification we will obtain d x1 = x1x1 +1 or

 20

d = x1  x .1
x1
(73)
1
It is easy to show that if n n – is a rational number, then it obligatory will be natural.
Let’s now say that.
1
k , n  N, n  1 and n n = k . (74)
1

From (74) n = k  2 . But the last is not valid. I.e. x − is an irrational if x1  N , x1  1 .

n n
1
x1

For x1 = 1 we will assume that (71) has unique pair of solutions:

 x1 = 1

 y1 = 2.
290  d =1 therefore the solutions of (69) are (1, 2).
Task 8. Let’s solve the equation
( y − x) x + y = y x ( y  x) (75)
in natural numbers.
Solution:
Let’s say that solutions of (75) are x0 , y0 and at the same time

x0 = x1d , y0 = y1d , (76)

where d0 = ( x0 , y0 ) .

By introduction of (76) in (75) we will obtain:

d x1d + y1d ( y1 − x1 ) x1d + y1d = d x1d  y1x1d . (77)

From (77) we have:

x1
 y 
d ( y1 − x1 ) =  1  .
y1 x1
(78)
 y1 − x1 
The left side of (78) is natural number, thus accordingly coif Lemma 1
y1 = x1 + 1, from that y1 = x1 + 1. (79)

By introduction of (79) in (6.78) we will obtain:

 x +1
1 x
x1 +1
d = ( x1 + 1)  d =  1
x1
 . (80)
 d 
In order for (80) to have a solution, it is necessary to find k  N \ {1}, for that x1 + 1 = kd . By

taking into account (80) we obtain

 21

d = k kd −1 . (81)
But as for each n  N n  22n −1 , thus (80) has no solution and hence also (75) will have the
solution in natural numbers.
Task 9. Let’s solve the equation
( x + y) x − y = x y (82)
in natural numbers.

solution:
Let’s say that solutions of (82) are x0 , y0 and at the same time

x0 = x1d , y0 = y1d , (83)

where d = ( x0 , y0 ) .

By introduction of (83) in (82) we will obtain:

d ( x1 − y1 ) d ( x1 + y1 )( x1 − y1 ) d = x1 y1d d y1d . (84)

From (84) we will obtain:

( x1 + y1 ) x1 − y1
y1
= d 2 y1 − x1 . (85)
x1

Let’s say that x1  1 . Then we will have:

(
( x1 , y1 ) = 1  ( x1 + y1 , x1 ) = 1  ( x1 + y1 ) x1 − y1 , x1y1 = 1 . ) (86)

The right side of (85) or its reciprocal number is natural. From (6.86) follows that neither left
side of (85) and nor its reciprocal value is possible to be natural number. I.e. for x1  1 also (85) has

no solution in natural numbers.

If x1 = 1 , then from (85) easy will be obtained x1 = y1 = d = 1 .

I.e. (82) has the unique solution in natural numbers: x0 = y0 = 1.

Task 10. Let’s solve the equation

( x + y) x − y = y x (87)
in natural numbers.
solution:
Let’s say that solutions of (87) are x0 , y0 and at the same time

x0 = x1d , y0 = y1d , (88)

where d = ( x0 , y0 ) .
 22

By introduction of (88) in (87) we will obtain:

d ( x1 − y1 ) d ( x1 + y1 )( x1 − y1 ) d = y1x1d d x1d , (89)

From that we will obtain:

d x1 − y1 ( x1 + y1 ) x1 − y1 = y1x1 d x1 . (90)
From (90) we will obtain
( x1 + y1 ) x1 − y1
d y1
= . (91)
y1x1

(
( x1 , y1 ) = 1  ( x1 + y1 , y1 ) = 1  ( x1 + y1 ) x1 − y1 , y1y1 = 1 . ) (92)

The left side of (91) is a natural number, as well as right side, to be natural, from (92) follows
that y1 necessary will be equal to 1. I.e. y1 = 1 .

By introduction of d = ( x1 + 1) x1 −1 in (91) we finally have that all pairs of solution of (87) will

be from
 x = t (t + 1)t −1 ,
 t −1
 y = (t + 1)
t  N

Task 11. Let’s solve the equation
( x + y) y = ( x − y) x , ( x  y) (93)
in natural numbers.
Solution:

Let’s say that solution of (93) is x0 and y0 , where ( x0 , y0 ) = d , x0 = x1d and y0 = y1d . I.e.

( x1, y1 ) = 1 , x1, y1  N .
By introduction in (93) and simplification
( x1 + y1 ) y1 = ( x1 − y1 ) x1 d x1 − y1 . (94)

I.e.
 x1 + y1   2 y1  2 y1
   N1 or 1 +  N  N .
 x1 − y1   x1 − y1  x1 − y1

From the last accordingly of Lemma 1,

x1 − y1 = 1 , (95) or x1 − y1 = 2 . (96)

Let’s take into account the case (95)

x1 − y1 = 1  y1 = x1 − 1 .
 23

It is obvious that x1  1 .

By introduction in (94) and simplification we will obtain:

d = (2 x1 − 1) x1 −1 .
I.e. the solutions of (93) are
x0 = x1 (2 x1 − 1) x1 −1 and y0 = ( x1 − 1)(2 x1 − 1) x1 −1 , (97)

where x1  N \ {1} .
Let’s take into account the case (96)
x1 − y1 = 2  y1 = x1 − 2 .
It is obvious that x1  N \{1} .
By introduction in (94) and simplification, we will obtain
( x1 − 1) x1 −1 = (2d )2 . (98)

From (98) is clear that x1 is impossible to be even number, thus x1 − 1 = (2n)2 , where n  N ,

or x1 = 4n2 + 1 .
2
−1
(2n) 4 n
From (98), d = ; i.e. in this case:
2
2 2
−1 −1
(2n)4 n (2n)4 n
x0 = (4n + 1) 
2
; y0 = (4n − 1) 
2
. (99)
2 2
Finally we have that all solutions of (93) are stated by (97) and (99) formula.
Task 12. Let’s solve the equation
( x − y) x + y = x x − y ( x  y) (100)
in natural numbers.
Solution:
Let’s say that solution of equation (100) are x0 and y0 , x0 = x1d , y0 = y1d , where d = ( x0 , y0 )

, x1, y1, d  N . By introduction in (100) we will obtain

d ( x1 + y1 )  ( x1 − y1 )( x1 + y1 ) d = x1( x1 − y1 ) d  d ( x1 + y1 ) d ,
From that
x1 − y1
 x 
d 2 y1
( x1 − y1 ) 2 y1
= 1  . (101)
 x1 − y1 
 x1 
The left side of (101) is a natural number. Thus natural will be also   . Accordingly of
 x1 − y1 
this Lemma 1 x1 − y1 = 1  y1 = x1 − 1 . By introduction in (101) and simplification we will obtain
 24

d 2( x1 −1) = x1 . (102)
By application of method of mathematical induction, easy is to prove that if x1  N \ {1} , then
22( x1 −1)  x1 , thus from (102) will be obtained that d = 1 and x1 = 1 , that is impossible as x1, y1  N
and x1  y1 . I.e. (100) has no solution in natural numbers.
Task 13. Let’s solve the equation
( x + y) x − y = ( x − y) x ( x  y) (103)
in natural numbers
Solution:
Let’s say that solutions of (103) are x0 and y0 , x0 = x1d , y0 = y1d , where x1, y1, d  N ;

( x0 , y0 ) = d . By introduction in (103) and simplification we will obtain:

(d ( x1 + y1 ))( x1 − y1 ) d = (d ( x1 − y1 )) x1d , (104)

from that we obtain
d x1 − y1 ( x1 + y1 ) x1 − y1 = d x1 ( x1 − y1 ) x1d , (105)

Hence we have
x − y1
 x1 + y1  1
  = d y1 ( x1 − y1 ) y1 . (106)
 x1 − y1 
x1 + y1 2 y1
I.e. =1+  N , or accordingly of Lemma 1 we have
x1 − y1 x1 − y1

 x1 − y1 = 1,  y1 = x1 − 1, (107)
 x − y = 2.   y = x − 2. (108)
 1 1  1 1

By introduction of (107) in (108) we will obtain d x1 −1 = 2 x1 − 1 , from that d = 1 and x1 = 1 ,

as in contrary
d x1 −1  2 x1 −1  2 x1 − 1 . (109)

Due the mathematical induction method would is easy to be proven the inequality
2 x1 −1  2 x1 − 1 , where x1  N \ {1} .

I.e. in this case (103) has no solution because y1 = 1 − 1 = 0  N .

By introduction of (108) in (106) we will obtain

 2 x1 − 2 
2
x1 − 2
  =d  2 x1 − 2  ( x1 − 1)2 = (2d ) x1 − 2 . (110)
 2 
 d  1,
If  then by the method of mathematical induction it is easy to prove that (2d ) x1 − 2  ( x1 − 1)2 .
 x1  2,
 25

If d = 1 , then from (110)

2x1 − 2 = ( x1 − 1)2 . (111)

From (111) either ( x1 − 1) and ( x1 − 2) simultaneously will be even, that is impossible. I.e.

(111) has no solution in natural numbers and therefore, (103) also will have the solution in natural
numbers.
Theorem 5. If (a, b) = 1, then a2 + ab + b2 and each divisor of (112) will be as (112).
Firstly let’s prove the Lemma 2 and Lemma 3.
Lemma 2. If (a, b) = 1 and a2 + ab + b2 type number is divided on c2 + cd + d 2 = p (a, b, c, d
 N) prime number, then the divider also would be present as following type of number.
Proof.
Let’s consider the identities:
(a 2 + ab + b 2 )(c 2 + cd + d 2 ) = (ac − bd )2 + (ac − bd )(ad + bc + bd ) + (ad + bc + bd )2 ; (113)
 2
(a + ab + b )(c + cd + d ) = (ad − bc) + (ad − bc)(ac + bd + bc) + (ac + bd + bc) ;
2 2 2 2 2
(114)

(ac − bd )(ac + bd + bc) = c (a + ab + b ) − b (c + cd + d ), (115)
2 2 2 2 2 2

 (116)
(ad − bc)(ad + bc + bd ) = d (a + ab + b ) − b (c + cd + d ).
2 2 2 2 2 2

Let’s say that a2 + ab + b2  0(mod(c2 + cd + d 2 )) , where (a, b, c, dN) and c2 + cd + d 2 = p

is prime number; it is obvious that (c, d) = 1.

As it is clear from (115) and (116) is clear are possible the following cases. Let’s consider
each of them:
ac − bd  0(mod p),
ad − bc  0(mod p), (c − d )(a + b)  0(mod(c 2 + cd + d 2 )),
 
I.  2  a 2 + ab + b 2  0(mod(c 2 + cd + d 2 )). 
c + cd + d = p,
2
(a, b) = 1, (c, d ) = 1.
a 2 + ab + b 2  0(mod p). 


a + b  0(mod p ), a 2 + 2ab + b 2  0(mod p), ab  0(mod p ),

  
 a 2 + ab + b 2  0(mod p),  a 2 + ab + b 2  0(mod p),  a 2 + ab + b 2  0(mod p ), 
( a, b) = 1 ( a , b ) = 1 ( a, b) = 1
 
 

a  0(mod p),

 b  0(mod p), p = 1. this is impossible.
(a, b) = 1.


ac − bd  0(mod p),

II.  From (113) it is obvious that presentation is possible;
ad + bc + bd  0(mod p).
 26

ac + bd + bc  0(mod p ),
III.  From (114) it is obvious that presentation is possible;
ad − bc  0(mod p ).
ac + bd + bc  0(mod p ),
ad + bc + bd  0(mod p), a(c − d )  0(mod(c + cd + d )),
2 2

 
IV.  2  a 2 + ab + b 2  0(mod(c 2 + cd + d 2 )), 
a + ab + b  0(mod p),
2
(a, b) = 1, (c, d ) = 1.
(a, b) = 1,(c, d ) = 1. 

a  0(mod p ), a  0(mod p ),
 2 
a + ab + b  0(mod p ),  b  0(mod p ),  p = 1 . This is impossible.
2

(a, b) = 1. 
 (a, b) = 1.

Proceeding from all this, by dividing (113) and (114) on (c2 + cd + d 2 )2 we will obtain the
provable.
I.e. if a2 + ab + b2  0(mod(c2 + cd + d 2 )) , where (a,b,c,dN), (a,b)=1) and c2 + cd + d 2 = p

are prime numbers, then the dividend also will be as (i.e. it would be represented as A2 + AB + B 2 ,
where A, B  N. Q.E.D.
this form, the dividend has a divisor that cannot be represented in the form (118).
Lemma 10. If a2 + ab + b2 (117) type number, where a, b  N and (a, b) = 1, is divisible by
a prime numbers that would not be represented in such way, the dividend has a divisor that is
presented as (118).
Proof.
Let’s assume that
a 2 + ab + b2 = x  P1 , P2 , , Pn . (119)

If P1 , P2 , , Pn all prime divisor will be as (118), then by successive division of (119) on P1 ,

P2 and so on Pn by virtue of the previous theorem, each divisor including x must be as (118). And

this contradicts the condition of Lemma 3. Therefore, any of Pi i = 1,, n does not have such form.

Q.E.D.
Now let's prove
Theorem 5. If a, b  N and (a, b) = 1, then each divisors of (a2 + ab + b2 ) will be as (120).
Proof:
Let’s say that (a, b) = 1, a, b  N and divisor x of (a2 + ab + b2 ) is not as (120). In addition

let’s say that a2 + ab + b2 is east between such numbers. It is obvious that a and b would be presented
as following:
 27

a = mx + (−1)k c ; b = nx + (−1) d , (121)

x x
where m, n  Z0 ; c, d , k ,  N ; 0  x  , 0d  .
2 2
then
a2 + ab + b2 = (mx + (−1)k c)2 + (mx + (−1)k c)(nx + (−1) d ) + (nx + (−1) d )2 =

= (m2 + mn + n2 ) x 2 + (2(−1)k mc + (−1) md + (−1) k nc + 2(−1) nd ) x +

+(c2 + (−1)k +1 cd + d 2 ) . (122)

Let’s mention that if k and ℓ simultaneously are odd, then mn  0. (123)

Let’s show that by conditions of (121) and (122) in the right side of (40) the sum of first two
summands is positive. Indeed
(m2 + mn + n2 ) x2 + (2(−1)k mc + (−1) md + (−1)k nc + 2(−1) md ) x = (m2 + mn + n2 ) x 2 +

 x x x x
+(−2mc − md − nc − 2nd ) x  (m 2 + mn + n 2 ) x 2 +  −2m − m − n − 2n  x =
 2 2 2 2

x2 x2
= (2m + 2mn + 2n − 3m − 3n) = (m(2m + n − 3) + n(2n + m − 3))  0 ,
2 2
(124)
2 2
As from m  1; n  1 and x N, i.e.

c2 + (−1)k + cd + d 2  a2 + ab + b2 and c2 + (−1)k + cd + d 2  0(mod x) . (125)

Let’s mention that
c2 − cd + d 2 = (c − d )2 + d (c − d ) + d 2 , (126)

c2 − cd + d 2 = (d − c)2 + d (d − c) + c2 , (127)

 x
 0c ,
 2 | c − d | x .
 (128)
0  d  x . 2

 2
Let’s say that k and ℓ are equally even or odd, then from (125) we have
c2 + cd + d 2 = yx . (129)
If c and d has more than 1 common divisors, then it is not dividing on x, as then it will divide
also a and b, but (a, b) = 1. By dividing of (129) on square of common divisor we will obtain
e2 + ef + f 2 = zx , (130)
In addition
3x 2
e2 + ef + f 2  c 2 + cd + d 2   x 2  a 2 + ab + b 2 .
4
 28

I.e. we obtain e2 + ef + f 2  a2 + ab + b2 and it also has divisor that is not as (120). This would
lead us to the infinite assumption.
I.e. if (a, b) = 1, then (120) has only and only (120) type divisors.
If k and ℓ are various odd and even numbers then we have:
c2 + cd + d 2 = yx and if c > d, then by virtue of (126)

(c − d )2 + d (c − d ) + d 2 = yx , (131)
And if c > d, then from (127)
(d − c)2 + c(d − c) + c2 = xy . (132)
By similar reasoning, as for (129), we will obtain the proof or if (a, b) = 1, (120) has only
(a2 + ab + b2 ) type divisors.
Application of this theorem is possible in lot of tasks. Let's state one example. Let's solve the
equation x 2 + xy + y 2 = n . Or prior solving this equation, we can first determine whether this
equation is solvable. In the canonic expansion of n from prime numbers with odd powers, if any of
them does not have the first form, then this equation has no solution.
Here we would like to briefly considered our studies in scope of Fibonacci numbers.
A lot of works has been written on Fibbonacci numbers and its properties should probably be
almost completely studied, but as Morley's theorem, the magnificent property that any triangle at the
intersection of trisections of any angle (both internal and external angles) will always turn out to be
an equilateral triangle, and from the discovery of that less than 90 years has passed, it makes one
wonder day it is possible to discover Fibonacci numbers with currently unknown amazing properties.
For our part, we want to draw attention to several facts.
It is known that Bines, when deriving his formula
It is known that Bines at deriving his formula
n n
 1+ 5   1− 5 
  − 
 2   2 
un = , (133)
5
Is applied the equation
x2 = x + 1 , (134)
The question is raised why exactly the equation (134)? Formula (134) itself raises the question of the
expediency of considering the following equations. These equations are:
x3 = x 2 + x + 1 ; (135)
x 4 = x3 + x 2 + x + 1 ; (136)
x5 = x4 + x3 + x2 + x + 1 , (137) ; xn = xn−1 + xn−2 + + x2 + x + 1 . (138)
 29

The diagrams of sides of equation (135) are easily understandable for us, and even for (136)
after much torment it will become more or less clear. And we would lose all interest in the subsequent
equations.
The situation is different if we proceed as follows:
n  N ; n  N ;
 n n −1 n−2 
x = x + x + + x 2 + x + 1;  x n ( x − 1) = ( x − 1)( x n −1 + n n − 2 + + x 2 + x + 1); 
 x  1. x  1
 

n  N ; n  N ;
n  N ;  
 n  n 1  1
  x ( x − 1) = x − 1;   x =
n
;   xn = ;
 x  1.  2− x  2− x
  x  1, 2 − x  0.  x  1; 2 ) .

Now it becomes clear what we are dealing with.

1
Let’s consider x n = n =1 x =1
2− x
 1− 5
x =  x = −0, 7748
 2
x = 1
n = 2  x = 1 n =3  n = 4   x = 1
  x = 1,83929  x = 1,9276
x = 1 + 5
 1, 6180
 2
 x = −0,8403
x = 1 x = 1
n =5  n = 6   x = 1 n=7
 x = 1,96595  x = 1,98358  x = 1,99196

 x = −0,87629
x = 1
n = 8   x = 1 n=9 
 x = 1,99603  x = 1,99803

Here we have obtained interesting characteristic numbers. The number that characterizes the
golden ratio q = 1,6180 and it is probably necessary to pay attention to the obtained values: 1;
1.83929; 1.9276; 1.96595; 1.98358; 1.99603; 1.99196; 1.99803 and so on.
In a few words, let's touch on some construction problems.
Let’s say that is given any MN segment, | MN |= a  1 , then in order to construct segments

with lengths: a 2 ; a 3 ; a4 ; ; an ; , is necessary to do the following.

1) Let’s on arbitrary straight line p take the points D and B in such manner that: | DB |= 1 ,

BDC = 90 and | CD |= a . Then if from C point we draw a perpendicular to CB segment up to
intersection of circumference p , we will obtain that
 30

| CD |2 =| AD || BD | a2 =| AD | 1 | AD |= a2 .
2) Then similarly let’s construct AC
1 1 B1 , where | D1B1 |= a ; | C1 D1 |= a 2 ;
 
C
1 1B1 = C1 D1 B1 = 90 , then similarly | C1D1 | =| A1D1 |  | D1B1 |
2
AC
 (a 2 )2 =| A1D1 | a | A1D1 |= a3 and so on.

A D B Let’s construct: An−1Cn−1Bn−1 , where

 
| Dn −1Bn −1 |= a n − 2 ; | Cn −1Dn −1 |= a n −1 ; An−1Cn−1Bn−1 = Cn−1Dn−1Bn−1 = 90 .

Here | Cn−1Dn−1 |2 =| An−1Dn−1 |  | Dn−1Bn−1 | (a n−1 )2 =| An−1Dn−1 | a n−2 | An−1Dn−1 |= a n .

I.e. if is given any [ MN ] , such that | MN |= a  1 , then is easy to construct the segment with

length of a n . It is known that also is easy to construct the segment with length of (2 – a).
Let’s consider the following sequences:
1) u1 = u2 = u3 = 1 , un = un−1 + un−2 + un−3 ;
2) u1 = u2 = u3 = u4 = 1 , un = un−1 + un−2 + un−3 + un−4
3) u1 = u2 = uk = 1 , un = un−1 + + un−k .
1) Pseudo-Fibbonacci numbers are sequences of order 3; 2) Pseudo-Fibbonacci numbers are
sequences of order 4; 3) sequences of order k. Let's call these sequences as pseudo-Fibbonacci
numbers: Let's call these sequences as pseudo-Fibbonacci numbers: 1) a sequence of 3 orders is
pseudo-Fibbonacci numbers; 2) pseudo-Fibbonacci numbers is a sequence of 4 orders; 3) order
sequence.
Now consider the first sequence
Let’s now consider the Let's call these sequences pseudo-Fibbonacci numbers: 1) a sequence
of 3 orders is pseudo-Fibbonacci numbers; 2) pseudo-Fibbonacci numbers is a sequence of 4 orders;
3) order sequence.
Now consider the first sequence
u1 = u2 = u3 = 1 , u4 = 3 , u5 = 5 , u6 = 9 , u7 = 17 , u8 = 31 , u9 = 57 ,

u10 = 105 , u11 = 193 . It is easy to show that Let’s prove

un = 2un −1 − un − 4 (139)

S1 = 1 , S2 = 2 , S3 = 3 , S4 = 6 , S5 = 11 , S6 = 20 , S7 = 37 , S8 = 68 .
Sn = 2Sn −1 − Sn − 4 (140)
It is easy to notice that
u5 = 2u4 − 1 = 2u4 − u1
 31

u6 = 2u5 − 1 = 2u5 − u2
u7 = 2u6 − 1 = 2u6 − u3
u8 = 2u7 − 3 = 2u7 − u4
u9 = 2u8 − 5 = 2u8 − u5
u10 = 2u9 − 9 = 2u9 − u6
..................................
un = 2un −1 − un − 4 by summation.

u5 + u6 + + un = 2(u4 + u5 + + un −1 ) − (u1 + u2 + u3 + + un −4 ) 
 6 + u5 + u6 + + un = 2(3 + u4 + u5 + + un −1 ) − (u1 + u2 + + un −4 ) 
 Sn = 2Sn −1 − Sn − 4 Q.E.D.

It would be easily shown that if S1 + S2 + + Sn  Sn , then similarly we will obtain

Sn = 2Sn −1 − Sn −4 for each n  N \ {1;2;3;4} and so infinitely.
Pseudo-Fibbonacci numbers were determined recursively, individually, according to their
numbers. It turns out that we would directly find any term of this sequence.
Let’s study different sequences u1, u2 , , un , that satisfy the dependency

un = 2un −1 − un − 4 . (141)
All such sequences are called by us as solutions of equation (141).
Let’s accordingly designate them as V , V  , V  and V  sequences
v1, v2 , v3 ,
v1, v2 , v3 ,
v1, v2, v3,
v1, v2, v3,
Firstly let’s prove two lemmas.
Lemma 1. If V is the solution of (141), and c  R, then cV (or the sequence cv1, cv2 , cv3 , )
also is the solution of (141).
Let’s multiply the dependency vn = 2vn −1 − vn − 4 term wise c , we will obtain

cvn = 2cvn −1 − cvn −4 Q.E.D.

Lemma 2. If V  , V  and V  are the solutions of (141), then their sum V  + V  + V  (or the
sequence v1 + v1 + v1 , v2 + v2 + v2 , v3 + v3 + v3, ) also represents the solution of (141).
From the conditions of lemma we have
vn = 2vn −1 − vn − 4
and
 32

vn = 2vn−1 − vn− 4 , vn = 2vn−1 − vn− 4 .

Due term wise summation of this equalities we will obtain
vn + vn + vn = 2(vn −1 + vn−1 + vn−1 ) − (vn −4 + vn−4 + vn−4 ) .
Due this Lemma is proven.
Let’s say that now V  , V  , V  are three non-proportional solutions of (141) (or are two such
v
solutions of (141), that for arbitrary c would be found such number n, for that n  c ). Let’s show
vn
that each sequence V that represents the solution of (141), is possible to represent as
cV
1
 + c2V  + c3V  , (142)

where c1 and c2 – are certain constants. Therefore, it is customary to say that (142) represents the

general solution of (141).

Now let's take certain solution V of (141). As already was mentioned, a sequence is considered
as given if its first three members v1 , v2 and v3 are given.

Let’s find such c1 , c2 and c3 , for that

c1v1 + c2v1 + c3v1= v1


c1v2 + c2v2 + c3v2 = v2 (143)
c v + c v + c v = v
13 2 3 3 3 3
From the condition of (141) is solvable the system with respect of (143) c1 , c2 and c3

v1 v1 v1 v1 v1 v3

c1 c
c1 = c1 = v2 v2 v2 c2 = 2 c2 = v2 v2 v2
 
v3 v3 v3 v3 v3 v3

v1 v1 v1 v1 v1 v1

c
c3 = 3 c3 = v2 v2 v2  = v2 v2 v2 .

v3 v3 v3 v3 v3 v3

The condition (141) mean that   0 . If we introduce the obtains c1 , c2 and c3 in (142), we

obtain the desired sequence V.

Therefore, to obtain all solutions of equation (141), is sufficient to find its three non-
proportional solutions.
Let's start the search among geometric progressions. According to Lemma 2, it is sufficient to
limit by the consideration of such progressions, the first term of that is 1. Thus, we take the
progression
2 3
1, q, q , q , ...
 33

n  4
In order for it to be a solution (141) for this, it is necessary for each  to fulfill
n  N
qn = 2qn −1 − qn − 4 ,
Or as q  1, we obtain

q3 − q 2 − q − 1 = 0 . (144)

Let’s say that solutions of (144) are q1 , q2 , q3 , then

19 + 3 33 + 3 19 − 3 33 + 1
3
q1 =
3
3
19 + 3 33 + 3 19 − 3 33 − 2 ( 3 19 + 3 33 − 3 19 − 3 33 )
q2 = − +i 3
6 6
3
19 + 3 33 + 3 19 − 3 33 − 2 ( 3 19 + 3 33 − 3 19 − 3 33 )
q3 = − −i 3 .
6 6
Let’s mention that for each root would be fulfilled
1 + q1 + q12 = q13 , 1 + q2 + q22 = q23 , 1 + q3 + q32 = q33 and q1q2q3 = −1 .
We obtain three geometric progressions, that represents the solutions of (141). Thus the
progression
c1 + c2 + c3 , c1q1 + c2q2 + c2q2 , c1q12 + c1q22 + c3q32 , (145)
Also represents the solution of (141). As the found progressions have different values, they are
therefore disproportionate. Formula (142) for various c1 , c2 and c3 will give all solutions of (141).

In particular, for any c1 , c2 and c3 formula (145) gives a pseudo-Fibbonacci series. To do

this, we rewrite (145) in this way .
c1 + c2 + c3 = 1

c1q1 + c2 q2 + c3q3 = 1
 2
c1q1 + c2 q2 + c3q3 = 1
2 2

c1 c2 c
c1 = , c2 = , c3 = 3 .
  
1 1 1 1 1 1 1 1 1 1 1 1
c1 = 1 q2 q3 , c2 = q1 1 q3 , c3 = q1 q2 1 ,  = q1 q2 q3 .
1 q22 q32 q12 1 q32 q12 q22 1 q12 q22 q32
(1 − q1 )(1 − q2 ) (1 − q1 )(q3 − 1) (1 − q3 )(1 − q2 )
c3 = , c2 = , c1 = .
(q3 − q1 )(q3 − q2 ) (q2 − q1 )(q3 − q2 ) (q3 − q1 )(q2 − q1 )
un = c1q1n −1 + c2 q2n −1 + c3q3n −1 .
Similarly we behave in the next case when
 34

u1 = u2 = u3 = u4 = 1 and so on.

Conclusions

We have shown that the entire class of tasks would be easily solved using one simple lemma;
pq − 1
in addition, is stated the task to study type expressions, where p and q are prime numbers,
p −1
with properties; some properties of the (a2 + ab + b2 ) type expression were studied, and so on.
The following issues are considered in the work:
– Obviously, for prime numbers p and q, it is of great interest to determine the quantity of those
pq − 1
prime divisors of the A = number that are less than p. To do this, we have considered:
p −1
Theorem 1. Let's say that p and q are odd prime numbers and p = 2q + 1. Then from the various
pq − 1
prime divisors of number A = , taken separately, only one is less than p. A has at least two
p −1
different prime divisors.
Theorem 2. Let's say that p and q are odd prime numbers and p < 2q + 1. Then all prime divisors
pq − 1
of the number A = are greater than p;
p −1
Theorem 3. Let's say that q is an odd prime number and, p  N \ {1} p ]1; q] [q + 2;2q] , then

pq − 1
each of the various prime divisors of the number A = , taken separately, is greater than p;
p −1
Theorem 4. Let's say that q is a prime odd number and p  {q + 1; 2q + 1}, then from different
pq − 1
prime divisors of number A = , taken separately, only one of them is less than p. A has at least
p −1
two different prime divisors.
yz −1
Task 1. Let’s solve the equation 2 x = in natural x, y, z numbers. At the same time y
y −1
must be a prime number.
yz −1
Task 2. Let’s solve the equation 3x = in natural x, y, z numbers. At the same time y must
y −1
be a prime number.
yz −1
Task 3. Let’s solve the equation p x = , where p{5; 7; 11; 13; ...} prime number x,
y −1
y  N and y – is prime number.
 – is stated the Lemma, by that easily would be solved the class of tasks:
 35

Lemma 1. Let’s say that a, b, n  N and (a, b) = 1. Let’s prove that if a n  0 (mod | a − b |) ,
or bn  0 (mod | a − b |) , then |a – b| = 1.
Let’s solve the equations ( – ) in natural x, y numbers:
z
x+ y
I.   =x −y ;
z z
VI. ( x + y) x − y = x y ;
 2 
II. ( x + y) z = (2 x) z + y z ; VII. ( x + y) x − y = y x ;
III. ( x + y) z = (3x) z + y z ; VIII. ( x + y) y = ( x − y) x , ( x  y) ;
IV. ( y − x) x + y = x y , ( y  x) ; IX. ( x − y ) x + y = x x − y ;
V. ( y − x) x + y = y x , ( y  x) ; X. ( x + y) x − y = ( x − y) x , ( y  x) .
theorem 5. If a, b  N, (a, b) = 1, then each divisor of (a 2 + ab + b2 ) will be as similar.

Refferences
1. Aghdgomelashvili Zurab. Diophantine geometric figures (Diophantine, Bidiophantine,
Pseudodiophantine and Pseudodiophantine planar geometric (Figures) LAP LAMBERT
Academic Publishing. 2020 (978-720-2-52393-6).
2. Aghdgomelashvili Zurab. On a Fundamental Task of Diophantine Geometric Figures. Cornell
University. Arxiv > org>arxiv. 2003. 02652.
3. Aghdgomelashvili Zurab. On a Fundamental Task of Bidiophantine Geometric Figures. Cornell
University. Ardxiv > org>arxiv. 2003. 10 846.
4. Zurab Agdgomelashvili. About on fundamental task on diophantine geometric Figures. Georgian
Technical University. N 4(514), 2019, 141-173.
5. Zurab Agdgomelashvili. Some interesting tasks from the classical number theory. Georgian
Technical University. N 4(518), 2020, 150-188.
6. Matiyasevich Y, Hilbert’s Tenth Problem. MLT Press Cambridge, Massachusetts, 1993.
7. Yaglom А.M. and Yaglom I.М. Non elementary tasks in the elementary statement. State
Publishing of Technical-theoretical literature. Moscow, 1954.