An algorithm for extracting primes, on

the proof of Riemann hypothesis

By Samuel Bonaya Buya

Ngao girls’ secondary school

December 2017

Declaration
The article entitled “An algorithm for extracting primes” is my own original work and every effort has
been made to acknowledge any work cited by means of proper referencing.
Abstract
In this research a method of extracting primes from even numbers is proposed. The method is based on
the assumption that a composite odd prime number can always be factorized into two odd numbers
with each of the prime numbers having a value of greater than one. For many years mathematicians
have correctly suspected that the Zeta function has carries the key for proper understanding of prime
numbers. The reason for the suspicion is that the Euler product formula connected the zeta function to
prime numbers. Further research by Riemann on distribution of primes gave rise to the Riemann
hypothesis. In this research an equation connected a family of curves will be derived to enable
extraction of primes. The formula will then be used to prove the Riemann hypothesis.

Introduction
Prime numbers are of fundamental importance in mathematics and particularly number theory.
Properties of prime numbers have been studied by many great mathematicians. Euclid came up with a
proof of the infinity of the primes. Leonard Euler derived the Euler’s product formula that connected the
prime number to the zeta function. Gauss and Legendre formulated the prime number theorem, the
proof of which was formulated by Hadamard and de la Vallée Poussin. In 1859 Bernard Riemann
published an 8 page paper in which he made breakthrough new and previously unknown discoveries
about distribution of prime. In the paper Riemann formulated a hypothesis about the roots of the Zeta
function which are a mystery to date.

In this research a method of extracting the prime numbers will be presented and then the same method
will be used to account for the Riemann hypothesis

In researches involving prime numbers a lot of time has been spent in the search for those properties
that can efficiently determine primality of numbers. A correct, reliable and sufficient primality test is one
that can separate composite odd numbers from odd prime numbers. Fermat’s little theorem fails do so.

Many primality and compositeness tests have been proposed and used in mathematics. A primality test
is an algorithm of determining whether an input number is a prime number. A compositeness test is an
algorithm in which is an input number is composite.

When a compositeness test is properly defined and simplified it can be used to extract both composite
odd numbers and odd prime numbers.
We know the product of an odd number and an odd number is an odd number. This means at minimum
a composite odd number is a product of two odd numbers each greater than one. This mathematical
truth will be used to come up with a procedure for extracting prime numbers.

Method

Consider two natural numbers, x and y such that x  1 and y>1. Two odd numbers, 2 x  1 and 2 y  1
can be formed from these natural numbers. The product of these two odd numbers is a composite odd
number given by:

Oc  (2 x  1)(2 y  1) 1

Equation one can be expanded and simplified to:

4 xy  2( x  y)  Oc  1  e 2

Here, e is an even number.

Expanding and simplifying equation 2:

e
x
e  2x 2
y  3
4x  2 2x 1

Equation 3 can be viewed as a Diophantine relationship between e, x and y.

The Diophantine relationship above holds if ( e  1 ) is a composite odd number but does not hold if (
e  1) is a prime number. This is to say for a given even number (e), if x and y have positive integral
e
values then (e+1) is a composite number otherwise it is a prime number. In equation 3 if ( ) is even
2
then select an odd value of x to make the numerator value odd (notice the denominator value is odd). If
e
( ) is odd select an even value of x to make the numerator value odd.
2

The Upper bound value of x is obtained when y is 1 and is given by:

e2
xub  4
6
 e e
 The lower bound value of x is 2 if ( ) is odd and 1 if ( ) is even.
2 2
 The interval between successive test values of x is 2. The test values are the values used to solve
the Diophantine equation 3 above. Test values may be even or odd but cannot be
simultaneously odd and even for a given even number. The test values are odd if of half the
selected even results to another even number. The test values are on the other hand is even if
half of the selected even number results to an odd number.
e
 If ( ) is odd then the interval of x is given by:
2

e2
2 x 5
6

The test values of x extracted from the interval 5 above are even

e
 If ( ) is even then the interval of x is given by:
2

e2
1 x  6
6

 In general the family of curves of equation cut the x and y axes at all the positive integer values
(see figures 1 and 2)

The test values of x extracted from the interval 6 are odd.

If one of the above boundary values form a solution of the Diophantine equation 3 then (e+1) is a
composite odd number. If all the values of x in the given interval do not satisfy the Diophantine equation
3 above then (e+1) is an odd prime number.

Example 1

The odd number ( e  1  9  e  8 ), the upper bound value of x (from 4) is 1 while the lower bound
value is 1 Thus the two boundary values coincide. The test value of x is therefore equal to 1. On
substituting this test value in equation 3 we get y = 1. This implies than 9 is a composite odd number.

Example 2

The odd number ( e  1  11  e  10 ), the upper bound value of x (from 4) is 4 while the lower
3
boundary value is 2, since half the even number gives an odd number. Since the lower bound value is
less than the calculated upper bound value then the odd number that has no test value is an odd prime
number.
Example3

The odd number ( e  1  37  e  36 ), the upper bound value of x (from 4) is 17 while the lower
3
bound value is 1, since half the even number gives another even number. The test values of x are (1,3) .

On substituting these test values we obtain the values of y as (17 ,15 ) . These values of y are
3 7
fractional. Thus the odd number is an odd prime number.

The equation 3 above is a family of curves that cut both the x and y axis at natural number points. This is
a very interesting result which needs some further scrutiny given previous attempts at methods of
extracting primes and accounting for prime gaps.

In the graphs below a few curves are drawn to illustrate.

Figure 1: Odd primes don’t have solution to the Diophantine equation 2.

Figure 2: composite prime have integer solution of the Diophantine equation 3 labeled on the graph.

Thus to determine the primality of a given odd number:

 Determine the interval of x

 Identify the test values of x
 Use the test values of x to evaluate y
 If for a given set of test values an integral value of y is obtained then the odd number is
composite, otherwise it is a prime number.

With these results we can come up with a simple algorithm for extracting prime numbers.

Algorithm for extracting primes:

Step1: Start

Step2: Declare the variables e, x1 , x2 ...xn , y1 , y2 ,... yn

Step 3: Calculate y

e  2 x1
y1 
4 x1  2

e  2 x2
y2 
4 x2  2

.
.
.
e  2 xm
yn 
4 xn  2
Step 4: If y is fractional

Display e  1 is a prime number

Step 5: Stop

The above algorithm can be improved bearing in mind all the issues discussed above.

Thus the problem of extracting primes from a set of natural numbers is easily solvable and easily
verifiable.
The AKS primality test placed PRIMES in P. The above method of extracting PRIMES makes them to be P-
complete. An easy algorithm exists for verifying primality of an odd number, which NP since they are
easily checkable in polynomial time in which case for Primes P=NP

Discussion of the above results and way forward

An easy method for extracting primes form a given set of odd prime numbers has been found. Given an
even number the odd number immediately after it can be established to be composite or prime using an
algebraic equation which can be used to simultaneously solve the problem in polynomial time and verify
the solution in polynomial time. Thus primes are both P and NP complete.

The equation used to extract primes is a family of curves cutting both the x and y at natural number
points, suggesting a strong connection to the zeta function. This connection needs some investigation.

The Zeta function the Euler product and proof of Riemann Hypothesis
The real valued zeta function is given for y and n, two real numbers:


1 1 1 1
  n)   n
 1  n  n  ...  n 7
y 1 y 2 3 y
In 1737 in a paper entitled Variae observations circa series infinitas, established a connection between
the zeta function and primes p:

1 1
 y  1 p
y
s
p
s
8

Riemann was the first to consider the zeta function  ( s) for a complex variable s for which ( s    it ).


1 1 1 1 1
  s)   s
 s  s  s  ...  s 9
y 1 y 1 2 3 y

The Riemann Hypothesis

The non-trivial zeros of the Riemann zeta function  ( s) have the real part Re( s)  1 .
2

Proof

For the purpose of proof of the Riemann Hypothesis y will be written in the form given in equation 3

e
x
e  2 x
ys  ( )s  ( 2 )s 10
4x  2 2x 1

Substituting 10 into 9:
 
1 
1 4x  2 x 1 2 s
  s)      ( ) 11
y 1 y s
y 1
e x e  2x e x
( 4 2 )s 4 2
x 1 2

  s)  0  ( x  1 2)s  0 12

We shall consider the case:

s  1  it 13
2

Substituting 13 into 12:

3 3
 1 1 ( 1  it ) x 2
1  1 1 ( 1  it ) x 2 ...( 1  n  2) x1/2 n 1
( x  1 2)1/2  it
 x  1 ( 1  it ) x 
1/2 2 2 2 2  ... 2 2 2 2
2 2 2! n!
   
 n 3  1 1 ( 1  it ) x n 5/2
 1 1 ( 1  it ) x n 5/2
...( 1  n  2) x1/2
n 1
 x (1  (  it ) x
1 1 2
 2 2 2  ... 2 2 2 2 )0
2 2 2! n!
14

Thus when s  1  it then either:

2

x0 15

Or:

 n 3  1 1 ( 1  it ) x  n5/2  1 1 ( 1  it ) x  n5/2 ...( 1  n  2) x1/2

1  1 ( 1  it ) x 2
 2 2 2  ... 2 2 2 2 0
2 2 2! n!

16

Given the fractional nature of the powers of the x terms, the roots of equation 16 are likely to be
complex. Our interest is on the real roots.

The above results confirm the Riemann hypothesis.

Conclusion
An N and PN complete method for extracting primes has been derived. The equation used to extract the
primes is a family of curves that cuts the x and y axes at natural number points.

The equation used to extract primes can be modified to account for the non-trivial zeros of Riemann
hypothesis. Riemann hypothesis is thus verified to be true.
Bibliography
Conrey, J. B. (2003, March). Riemann Hypothesis. Retrieved 12 20, 2017, from
https://www.aimath.org/WWWNrh/

Lehman, D. J. (May 1982). On Primality tests. SIAM Journal on Computing , 374-375.

Manindra Agrawal, N. K. (Sep., 2004). Primes in P. Annals of Mathematics vol. 160, No.2 , 781-793.

Shanks, W. A. (1982). Strong Primality tests that are not sufficient. Math.Comp 39 , 255-300.

Veisaidal, J. (2016, August 22). Medium. Retrieved December 19, 2017, from Medium.com:
https:/medium/com/@JorgenVeisal?source=Post