Escolar Documentos
Profissional Documentos
Cultura Documentos
Tony Thomas
Abstract
This paper uses a geometrical construction to model Fermats Conjecture (FLT). The contrary
assumption (FLT) is made and propositions derived from this assumption with the aim of
revealing inconsistencies under the conditions applicable to FLT.
Introduction
The aim of this paper is to establish a more direct proof of Fermats Last Theorem than the proof
published by Andrew Wiles in The Annals of Mathematics 142 (1995). The main idea is that the
three terms an, bn and cn all lie on the curve xn which facilitates the construction of diagrams
showing the relation between these terms and several auxiliary variables used in the proof. The
two diagrams used are shown in the annex.
11d41
Fermats Conjecture
Fermats Conjecture: there are no natural numbers a, b, c, n such that an + bn = cn when n > 2.
FLT may be expressed formally as: abcn an + bn = cn
subject to the following conditions:
(1) a,b,c and n are distinct natural numbers*
abcn
abc NCF
n>2
a<b<c
Parity limitations
Condition (2) implies that a, b and c cannot all be even numbers otherwise they would have the
common factor 2. Furthermore, if any two variables are even numbers then they have the common
factor 2 and the third variable must also be even. Consequently, two of the variables must be odd
and the third even. It will be shown later that b must be an odd variable given condition (3) so
either a or c is even.
Diagram 1 shows the main elements of the model. Given any two integers a and c there will be two
points (a, an) and (c, cn) on the curve xn. A line drawn through these points intersects the
x-axis at . Parallel to this secant there is a unique tangent that meets the curve at (K, Kn) and
intersects the x-axis at .
Let the gradient of this tangent be g such that:
= g,
= g and
=g
=g
Henceforth the antithesis of FLT will be assumed to be true: abcn an + bn = cn subject to the
given conditions (1) through (5).
g=
= consequently nKn-1 =
are of opposite parities; therefore g is an odd integer so b must also be an odd integer. Since n is
defined as a prime number greater than 2 it must be odd, so g must also be odd.
Superficially, it may seem from the equation bn = (ca)nKn-1 that b must be irrational, however by
setting K = n(ca) the integer solution b = K results as a special case.
Rearranging; K = 1 and =
(1)
g | bn (g | an g | cn )
Condition (2) stipulates a, b and c have no common factors.
Since bn = g(c - a) and g it follows that g | bn.
Applying modus ponens proves the Lemma
From Diagram 1 it can be seen that a =
So = a
and = c
and c =
+
2
because
2 2
(+)()
therefore c + a = 2K = g.
This special characteristic explains why Pythagorean triples are possible because a and c always
lie symmetrically about K. It will be shown below that this is not the case when n >2 so the secant
can never simultaneously intersect integer values of (a, an) and (c, cn) when K.
The variables a and c can be defined in terms of K as follows:
( )
and q =
( )
, where Z = Kn + g( )= g(K ) .
Parities of variables
Case 1
Case 2
Function
even
odd
odd
even
odd
odd
nKn-1
odd
odd
g(K )
even
odd
cK
odd
even
Ka
even
even
(n-1)Kn
even
even
ga an
odd
odd
odd
odd
fraction fraction
(o-o) =
( )
()()
cK=
()( )
e=
The result is a disparity. The variables on the RHS are assumed to be integers so the conclusion is
that p cannot be an integer.
Case 2: q =
(o-o) =
( )
()()
Ka=
e=
()( )
This is also a disparity showing that q is not an integer assuming the variables on the right are
integers.
Case 1 assigned an even parity to a and an odd parity to c. Case 2 assigned an odd parity to a and
an even parity to c. This reversal had no effect on the results.
Corollary: Either p is not an integer or q is not an integer under the given assumptions.
So (p + q) and since p + q = c a it follows that (c a).
This conclusion is a consequence of assuming that K is an integer and also that Z, as a function of K,
is an integer. The remaining possibilities are that K is a rational fraction or that it is an irrational
number. These possibilities will now be examined.
Assuming K is a rational fraction
g = nKn-1 = (1 )
1
c, a because
and g =
so n =
(1)
()1
=nnn-1.
1 1
(1)1
This is intuitively evident since a secant passing through the integer points (a,an) and (c,cn) must
be associated with a tangent which can be irrational without affecting the rationality of (c a).
= g and m = Kn-1 so = g
but n = so = mn consequently mn = g
but g =
Now n cannot have a rational nth root because it is a prime and the nth root of Kn-1 is also
irrational.
rational. Given that the secant passes through the integer points (a, an) and (c, cn) when K
the conclusion must be that (c a ) and that b and consequently that cn an bn .
Tony Thomas
7 Sunblest Crt
Eatons Hill
Queensland 4037
15th November 2015
Annex