Sei sulla pagina 1di 74

DIOPHANTINE EQUATIONS

By Abhishek Sanskrityayn 09MA4008

INTRODUCTION
One

of the important branches of Number Theory is finding Integer Solutions to equations Such equations are called Diophantine Diophantine Equations of great import are: 1. Linear Equations + = 2. Pythagorean Equation 2 + 2 = 2 3. Pell Equation 2 2 = 1 4. Fermat Equation + = for n > 2

LINEAR DIOPHANTINE EQUATIONS

LINEAR DIOPHANTINE EQUATIONS


The

easiest Diophantine equations to tackle are linear in two variables + = ; , , THEOREM 1: If = , , + = has an integral solution if and only if d divides c. If (0 , 0 ) is any solution, then (0 + ,0 ) forms the set of all integral solutions

Outline of Proof

If (0 , 0 ) is a solution, then 0 + 0 = and therefore d|c

Outline of Proof
If (0 , 0 ) is a solution, then 0 + 0 = and therefore d|c Conversely if , then = . = (, ) + = + = = and hence 0 , 0 = , is a solution

Outline of Proof
If (0 , 0 ) is a solution, then 0 + 0 = and therefore d|c Conversely if , then = . = (, ) + = + = = and hence 0 , 0 = , is a solution = (,) , = (,) is obviously a solution of the equation + = 0. Adding any solution 0 , 0 of + = to this gives another solution of + =

Continued
Conversely, if x,y is any solution of + = , then = 0 , = 0 satisfies + = 0 and therefore satisfies (, )( + ) = 0

Continued
Conversely, if x,y is any solution of + = , then = 0 , = 0 satisfies + = 0 and therefore satisfies (, )( + ) = 0 Since , = and

= 1, b must divide x, that is, hence = Therefore, = 0 + = 0 + = 0 + (,). Similarly, = 0 (,)

Continued
Conversely, if x,y is any solution of + = , then = 0 , = 0 satisfies + = 0 and therefore satisfies (, )( + ) = 0 Since , = and

= 1, b must divide x, that is, hence = Therefore, = 0 + = 0 + = 0 + (,). Similarly, = 0 (,)


Any

particular solution of + = may be found by using the Euclidean Algorithm

An Example
Let

us consider the linear equation 12 5 = 5 From the table we can see that , = 1, such that 12 3 5 7 = 1 Therefore, 12 15 5 35 = 5 Hence all the solutions are 15 5, 35 12 ;

Pair (12,-5)

Symbol (a,b)

(7,-5)
(2,-5) (2,-3) (2,-1) (1,-1)

(a+b,b)
(a+2b,b) (a+2b,a+3b) (a+2b,2a+5b) (3a+7b,2a+5b)

PYTHAGOREAN TRIPLETS

Pythagorean Theorem
An integer solution of 2 + 2 = 2 implies
2

=1
Q O X

We

are seeking rational points on the unit circle

Pythagorean Theorem
An integer solution of 2 + 2 = 2 implies 2 2 We
Y R Q O X

are seeking rational points on the unit circle We join the point Q (-1,0) to any rational point R on the circle We find the points of intersection of this line

=1

These happen 1; 2 2 be , 1: 2 1: 2

to

where

Continued
If

we substitute =

, ,

, then the
,

rational point becomes

2 ; 2 2 , 2 : 2 2 : 2

Therefore the general integer solution is 2 2 , 2, 2 + 2 , , , ,

the

first two may be interchanged To generalize, the equation <1 2 = 2 has integer solutions 1 = (1 2 2 ) <2 2 = 21 2 = 21 = 2 where , <1

THE PELL EQUATION

The Pell Equation


The

Pell equation 2 2 = 1 is one of the most important quadratic Diophantine equation It is not obvious that it has any integer solutions other than the trivial 1,0 In fact for n other than positive nonsquare integers there are only uninteresting solutions Therefore, we shall only consider positive nonsquare integer n

An Example
Let

us consider the equation 2 2 2 = 1 By trial we see that (3,2) is a solution We claim that if 1 , 1 2 , 2 are solutions then another solution 3 , 3 is given by 1 + 1 2 2 + 2 2 = 3 + 3 2, . . , 3 = 1 2 + 21 2, 3 = 1 2 + 1 2
This

is true since 1 2 + 21 2 2 2 1 2 + 1 2 1 2 21 2 2 2 22 2 =1

Group of Solutions
Thus,

by composing (3,2) with itself we obtain another solution (17,12) This product rule gives the set of solutions a group structure since 1. (1,0) acts as identity 2. every solution (x,y) has the inverse (x,-y) In fact, the group of positive solutions, where + 2 is positive is the infinite cyclic group of powers of 3+2 2

Powers of 3 + 2 2
Taking

log of all the positive solutions, since log = log + log , the resulting numbers form a group under addition
2) is the least positive element here

log(3 + 2

since 1. (3,2) is the least solution corresponding to x,y>0, and 2. the solutions (x,-y) with y>0 are inverses of solutions (x,y) with x,y>0 but they have negative logarithms

Continued
Such

a group contains the integer multiples of the least positive element, say, m If any element k lies between multiples of m, i.e., mn < k <m(n+1) 0 < k-mn < m Thus k-mn is a group element smaller than m contradicting the minimality of m Also since for any solution, either + 2 or 2 is positive. Thus the remaining solutions are the negatives of the powers of 3 + 2 2

The general Pell equation


In

the general case, the set of solutions with positive + forms a cyclic group whose generator is the smallest positive solution (1 , 1 ) with positive 1 , 1 Brahmagupta (600 C.E.) found that if (1 , 1 ) and (2 , 2 ) are solutions then a third solution is given by (1 2 + 1 2 , 2 1 + 1 2 ) We have not confirmed the existence of the generator of the solutions

Existence of Generator
For

any irrational and integer B>0, there are integers a, b with 0<b<B and 1 <

Let

us consider the B-1 numbers , 2 , , ( 1) . For each k choose integer such that 0 < <1 All of these are distinct Consider the B+1 numbers, 0, 1 , 2 2 , , ;1 + ( 1) ,1distributed in the [0,1] interval

Continued
Divide

the [0,1] interval into subintervals of length 1/B. These are B in number, therefore, by the pigeonhole principle, one of these contains two of the constructed numbers Their difference which is of the form < 1/ as required
b

is the difference of two positive integers less than B, therefore b<B

Continued
The

above result holds for any B>0, increasing B yields more values of a and b. Thus there are infinitely many a and b 1 satisfying < and since 0<b<B,

<

+ + 2 3 2
1 3

= 3

Thus

we have infinitely many a and b with norm less than 3

Continued
Hence

there are infinitely many with the same norm N Infinitely many a in the same congruence class mod N Infinitely many b in the same congruence class mod N We have two numbers 1 1 and 2 2 with these properties

Finally
1 ;1 Let = 2 ;2 1 2 ;1 2 ; + 1 2 1 2

1 ;1 2 :2 2 2 ;2 2

=
Pell Equation satisfied

This

quotient has norm 1 since the two numbers have equal norms Since the two numbers are unequal, their quotient cannot be 1 or -1
0

1 2 1 2 1 1 1 1 1 2 a is an integer 1 2 ( ) and so is b 1 2 and 2 1 , Multiplying we obtain 1 2 2 1

Generator
After

Diophantus, the task of solving equations in integers passed to Indians Around 600 C.E., Brahmagupta found the composition law of solutions of Diophantine Equations He found the smallest solution (1151,120) for the equation 2 92 2 = 1 and claimed that a person solving this problem within a year is a mathematician In 1150 C.E., Bhaskara II extended Brahmaguptas rule and found the solution (1766319049,226153980) for 2 61 2 = 1

Finding The Generator


Let us find the generator for the group of solutions of the equation 2 17 2 = 1 We find the continued fraction representation of 17 If the period s is even, the convergent ;1 , ;1 forms the smallest solution If the period s is odd, the convergent 2;1 , 2;1 forms the smallest solution 17 = 4 +
1
1

8:8+

, Here the period is 1, therefore


1 8 33 , 8

the solution is 4 + =

since 332 17. 82 = 1

Finding the Generator Graphically (1,0)


We

(0,1)

make this map where starting with (1,0) and (0,1), each successive label is the sum of vectors on its right and left

(1,1)

(2,1)

(1,2)

(3,1)

(3,2)

(2,3)

(1,3)

Finding the Generator Graphically (1,0)


We

(0,1)

make this map where starting with (1,0) and (0,1), each successive label is the sum of vectors on its right and left This way we obtain all possible pairs of relatively prime natural numbers

(1,1)

(2,1)

(1,2)

(3,1)

(3,2)

(2,3)

(1,3)

We

make this map where starting with (1,0) and (0,1), Map for each successive 2 3 2 = 1 label is the sum of vectors on its right and left (2,1) This way we obtain all possible pairs of relatively prime (3,1) natural numbers

Finding the Generator Graphically (1,0) 1

(0,1)

-3

(1,1)

(1,2)

(3,2)

(2,3)

(1,3)

We

make this map where starting with (1,0) and (0,1), Map for each successive 2 3 2 = 1 label is the sum of vectors on its right and left (2,1) This way we obtain all possible pairs of relatively prime (3,1) natural numbers

Finding the Generator Graphically (1,0) 1

(0,1)

-3

(1,1) -2

(1,2)

(3,2)

(2,3)

(1,3)

We

make this map where starting with (1,0) and (0,1), Map for each successive 2 3 2 = 1 label is the sum of vectors on its right and left (2,1) This way we obtain 1 all possible pairs of relatively prime (3,1) natural numbers

Finding the Generator Graphically (1,0) 1

(0,1)

-3

(1,1) -2

(1,2)

-11
(3,2) (2,3) (1,3)

We

make this map where starting with (1,0) and (0,1), Map for each successive 2 3 2 = 1 label is the sum of vectors on its right and left (2,1) This way we obtain 1 all possible pairs of relatively prime (3,1) natural numbers 6

Finding the Generator Graphically (1,0) 1

(0,1)

-3

(1,1) -2

(1,2)

-11
(3,2) -3 (2,3) -23 (1,3) -26

Continued
When

we proceed further, we obtain a pattern of repeating values 1,-3,-2 This we call the river It is found that all the solutions are on this river On the other side, the value of the function appears to increase in absolute value without bound
(0,1) (1,1) (1,0) (2,1) (3,2) (5,3) (12,7) (19,11)

(7,4)

(26,15)

Continued
When

we proceed further, we obtain a pattern of repeating values 1,-3,-2 This we call the river It is found that all the solutions are on this river On the other side, the value of the function appears to increase in absolute value without bound
(0,1) (1,0)

-3
(1,1) 1 (2,1) (3,2) (5,3) (12,7) (19,11)

(7,4)

(26,15)

Continued
When

we proceed further, we obtain a pattern of repeating values 1,-3,-2 This we call the river It is found that all the solutions are on this river On the other side, the value of the function appears to increase in absolute value without bound
(0,1) (1,0)

-3
(1,1) 1

-2 (3,2) (2,1) (5,3) (12,7) (19,11)

(7,4)

(26,15)

Continued
When

we proceed further, we obtain a pattern of repeating values 1,-3,-2 This we call the river It is found that all the solutions are on this river On the other side, the value of the function appears to increase in absolute value without bound
(0,1) (1,0)

-3
(1,1) 1

-2 (3,2) (2,1) 1 (5,3) (12,7) (19,11)

(7,4)

(26,15)

Continued
When

we proceed further, we obtain a pattern of repeating values 1,-3,-2 This we call the river It is found that all the solutions are on this river On the other side, the value of the function appears to increase in absolute value without bound
(0,1) (1,0)

-3
(1,1) 1

-2 (3,2) (2,1) 1

-3 (5,3) (12,7) (19,11)

(7,4)

(26,15)

Continued
When

we proceed further, we obtain a pattern of repeating values 1,-3,-2 This we call the river It is found that all the solutions are on this river On the other side, the value of the function appears to increase in absolute value without bound
(0,1) (1,0)

-3
(1,1) 1

-2 (3,2) (2,1) 1

-3 -2 (5,3) (12,7) (19,11) (26,15)

(7,4)

Continued
When

we proceed further, we obtain a pattern of repeating values 1,-3,-2 This we call the river It is found that all the solutions are on this river On the other side, the value of the function appears to increase in absolute value without bound
(0,1) (1,0)

-3
(1,1) 1

-2 (3,2) (2,1) 1

-3 -2 (5,3) (12,7) (19,11)

(7,4)

(26,15)

Continued
When

we proceed further, we obtain a pattern of repeating values 1,-3,-2 This we call the river It is found that all the solutions are on this river On the other side, the value of the function appears to increase in absolute value without bound
(0,1) (1,0)

-3
(1,1) 1

-2 (3,2) (2,1) 1

-3 -2 (5,3)

-3
(12,7) (19,11)

(7,4)

(26,15)

Continued
When

we proceed further, we obtain a pattern of repeating values 1,-3,-2 This we call the river It is found that all the solutions are on this river On the other side, the value of the function appears to increase in absolute value without bound
(0,1) (1,0)

-3
(1,1) 1

-2 (3,2) (2,1) 1

-3 -2 (5,3)

-3
(12,7) (19,11)

-2

(7,4)

(26,15)

Continued
When

we proceed further, we obtain a pattern of repeating values 1,-3,-2 This we call the river It is found that all the solutions are on this river On the other side, the value of the function appears to increase in absolute value without bound
(0,1) (1,0)

-3
(1,1) 1

-2 (3,2) (2,1) 1

-3 -2 (5,3)

-3
(12,7) (19,11)

-2

(7,4)

(26,15)

RATIONAL POINTS ON ELLIPTIC CURVES

Elliptic Curves
The

graphical approach led us to solutions of Pythagorean triplets and the Pell equation We now hope to exploit the same ideas for cubic curves This will also lead to a group structure

2 = 3 2 + 4

Elliptic Curves

A cubic curve is the graph of a degree three polynomial equation in 2 variables An elliptic curve is a special type of cubic curve which we shall define later We will focus on cubic curves of the form 2 = 3 + 2 2 + 1 + 0 This curve is usually symmetric about the xaxis

2 = 3 4

Intersection of Lines and Curves


If

f(x) is a monic degree 3 rational polynomial, then the cubic curve 2 = () and the line y=mx+b intersect in three points counting multiplicities If coordinates of 2 of the intersection points are rational then so are the coordinates of the third

2 = 3 + 17

Singularities
A

point (0 , 0 ) on a curve F(x,y)=0 is singular if 0 , 0 = (0 , 0 )=0

curve 2 = 3 + + is nonsingular if and only if 43 27 2 0 A singularity is always of the form (0 , 0) where 0 is a root of () = 3 + 2 2 + 1 + 0 of multiplicity 2 or 3

A Node- a double root


2 = 3 3 + 2

Singularities
A

point (0 , 0 ) on a curve F(x,y)=0 is singular if 0 , 0 = (0 , 0 )=0

curve 2 = 3 + + is nonsingular if and only if 43 27 2 0 It may be proved that for curves of the form 2 = , the tangent will intersect the curve at a point of multiplicity at least 2

A Cusp-A triple root


2 = 3 2 2 4 + 8

Group Law
Let

O be the point at infinity Let , . Let L be the line joining P and Q, and R the third point of intersection of L with E. Let L be the line joining R and O Then P+Q is the point such that L intersects E at R, O, P+Q We may prove that this law of composition gives a group structure on the points of the curve and the points with rational coordinates form a subgroup

S Q

P
P+Q

2 = 3 + 2

R Q

P
P+Q

2 = 3 4

2 = 3 4 2 + 1

Elliptic Curves
A

rational elliptic curve is the group consisting of the set of solutions to a nonsingular cubic curve of the form 2 = () together with the point at infinity under the group operation defined earlier Ramanujan once told Hardy that 1729= 13 + 123 = 93 + 103 is the smallest number that can be written as the sum of two cubes in two different ways Now we are going to find integers which can be written as sum of two cubes in k different ways

Applications to Diophantine Equations


We

can often transform an equation to the equation for a cubic curve The equation 3 + 3 = transforms to 2 = 3 36: 36; 4322 under the rule = , =
6 6

+ 3 = 35 becomes 2 = 3 529200 We know (2,3) is a solution to the former which is P=(84,252) after transformation Now by the theory we developed, we find 2P=(1596,-63756) and 3P=(1000/9,25327/27)
3

Continued
In

to 124 3 129 3 3 3 3 + 2 = 35, ( ) +( ) =


35
19 19 59347 3 8693 3 ( ) +( ) = 35 18162 18162

the original equation this corresponds

Multiplying

by the largest denominator we find three ways of writing the integer 1438201910159509320 = 10352343 + 6901563 = (2252088)3 +23428983 = 11275933 + 1651673

Encryption using Elliptic Curves


An

elliptic curve over , p a prime, is the set of solutions to 2 = ( ) and this forms a group under the same operation described before Hasses Theorem tell us that the number of points N on an elliptic curve over satisfies ( + 1) < 2 This encryption scheme uses the fact that it is easy to compute kP, for k an integer and P a point on the elliptic curve, by the group operation , but difficult to solve kP=Q for k

Curve E, prime p, and the number of points on E, n are known

Curve E, prime p, and the number of points on E, n are known The message is made to correspond with a point on the curve

Curve E, prime p, and the number of points on E, n are known The message is made to correspond with a point on the curve Sender chooses a < n such that gcd(a,n)=1

Curve E, prime p, and the number of points on E, n are known The message is made to correspond with a point on the curve Sender chooses a < n such that gcd(a,n)=1

Sender transmits aP to receiver

Curve E, prime p, and the number of points on E, n are known The message is made to correspond with a point on the curve Sender chooses a < n such that gcd(a,n)=1 Receiver also chooses b < n such that gcd(b,n)=1

Sender transmits aP to receiver

Curve E, prime p, and the number of points on E, n are known The message is made to correspond with a point on the curve Sender chooses a < n such that gcd(a,n)=1

Receiver sends back to the sender b(aP)

Receiver also chooses b < n such that gcd(b,n)=1

Sender transmits aP to receiver

Curve E, prime p, and the number of points on E, n are known The message is made to correspond with a point on the curve Sender chooses a < n such that gcd(a,n)=1

Receiver sends back to the sender b(aP)

Sender solves 1( ) and transmits a(baP)=bP

Receiver also chooses b < n such that gcd(b,n)=1

Sender transmits aP to receiver

Curve E, prime p, and the number of points on E, n are known The message is made to correspond with a point on the curve Sender chooses a < n such that gcd(a,n)=1

Receiver sends back to the sender b(aP)

Sender solves 1( ) and transmits a(baP)=bP

Receiver also chooses b < n such that gcd(b,n)=1

Receiver solves 1( ) and recovers bbP=P

Sender transmits aP to receiver

Fermats Last Theorem


Gerhard

Frey showed that if there was a nontrivial solution to the equation + = , > 2, then a certain semistable elliptic curve will not be modular However the Taniyama-Shimura Conjecture held that all rational elliptic curves are modular Andrew Wiles proved in 1995 that all semistable elliptic curves are modular Thus the equation + = , > 2 has no nontrivial solutions

OPEN PROBLEMS

The abc conjecture


Define
It

rad(n)=

was found that for the integer solutions of the equation a+b=c, c < rad(abc) is true for most cases Therefore it was conjectured that > 0, 1: < for all pairwise relatively prime positive integers a, b, c satisfying a+b=c This is called the abc Conjecture

Continued
In

fact no counterexamples are known for


2

the conjecture <


Let

= , = = , if the conjecture is true, we have = < (() )2 < ()2 < 6 Therefore, ;6 < 1, which is possible only when n<6 since z>1 Thus this would prove Fermats last Theorem for n>5

Hilberts tenth problem


David

Hilbert discussed 23 unsolved problems in mathematics at the 1900 ICM The tenth problem asked to give a universal procedure/algorithm, which for any Diophantine Equation, would tell whether it is solvable or unsolvable over In 1973, Yuri Matiyasevich demonstrated that there is no such procedure

Continued
This

led to some beautiful mathematics along the way Polynomials were found whose positive values are exactly the prime numbers (Of course, the number of variables is huge) The rational counterpart of Hilberts tenth problem is still an active open problem

The end

Potrebbero piacerti anche