Cryptography Reference
In-Depth Information
Exercise 7.9.6 Prove that the following algebraic formulae for the chord-and-tangent
rule are correct. Let P 1 ,P 2
E (
k
), we want to compute S
=
P 1 +
P 2 .If P 1 = O E then
S
=
P 2 and if P 2 = O E then S
=
P 1 . Hence, we may now assume that P 1 =
( x 1 ,y 1 ) and
P 2 =
( x 2 ,y 2 ) are affine. If y 2 =−
y 1
H ( x 1 ) then S
= O E . Otherwise, set λ to be as in
λ 2
Definition 7.9.1 and compute x 3 =
+
a 1 λ
a 2
x 1
x 2 and y 3 =−
λ ( x S
x 1 )
y 1 .
The sum is S
=
( x 3 ,y 3 ).
Before proving the main theorem, we state the following technical result, whose proof
is postponed to the next chapter (Corollary 8.6.5 ).
Theorem 7.9.7 Let P 1 ,P 2
k
) be a points on an elliptic curve such that P 1 =
E (
P 2 . Then
( P 1 )
( P 2 ) is not a principal divisor.
We now consider the divisor class group Pic 0
k
( E ). The following result is usually obtained
as a corollary to the Riemann-Roch theorem, but we give an ad-hoc proof for elliptic curves.
One can consider this result as the Abel-Jacobi map in the case of genus 1 curves.
) and Pic 0
k
Theorem 7.9.8 There is a one-to-one correspondence between E (
k
( E ) , namely
P
( P )
(
O E ) .
Proof We first show that the map is injective. Suppose ( P 1 )
(
O E )
( P 2 )
(
O E ). Then
( P 1 )
( P 2 ) is principal, and so Theorem 7.9.7 implies P 1 =
P 2 .
= P n P ( P ) be any effective divisor
on E . We prove that D is equivalent to a divisor of the form
It remains to show that the map is surjective. Let D
( P )
+
(deg( D )
1)(
O E ) .
(7.10)
We will do this by replacing any term ( P 1 )
+
( P 2 ) by a term of the form ( S )
+
(
O E )for
some point S .
The key equations are ( P )
+
( ι ( P ))
=
2(
O E )
+
div( v ( x )) where v ( x ) is as in Defini-
tion 7.9.1 ,or,if P 1 =
ι ( P 2 ), ( P 1 )
+
( P 2 )
=
( S )
+
(
O E )
+
div( l ( x,y ) /v ( x )). The first equa-
tion allows us to replace any pair ( P )
+
( ι ( P )), including the case P
=
ι ( P ), by 2(
O E ). The
second equation allows us to replace any pair ( P 1 )
+
( P 2 ), where P 1 =
ι ( P 2 ) (but including
the case P 1 =
O E ). It is clear that any pair of affine points is included in
one of these two cases, and so repeating these operations a finite number of times reduces
any effective divisor to the form in equation ( 7.10 ).
Finally, let D be a degree zero divisor on E . Write D
P 2 ) with ( S )
+
(
D 2 where D 1 and D 2
are effective divisors of the same degree. By the above argument, we can write D 1
( S 1 )
=
D 1
( S 2 ).
Finally, adding the divisor of the vertical line function through S 2 and subtracting the divisor
of the line between S 1 and ι ( S 2 )gives D
+
(deg( D 1 )
1)(
O E ) and D 2
( S 2 )
+
(deg( D 1 )
1)(
O E ). Hence, D
( S 1 )
( S )
(
O E ) for some point S as required.
) is in bijection with the group Pic 0
k
) is a group, with the
group law coming from the divisor class group structure of E . It remains to show that the
group law is just the chord-and-tangent rule. In other words, this result shows that the chord-
and-tangent rule is associative. Note that many texts prove that both E (
Since E (
k
( E ) it follows that E (
k
) and Pic 0
k
k
( E )are
Search WWH ::




Custom Search