Cryptography Reference
In-Depth Information
S N ,then SM has determinant N , so there exists A S
SL 2 ( Z ) and a uniquely determined M S
PROOF
If S
S N such that A S M S = SM . f
S 1 ,S 2
S N and M S 1 = M S 2 ,then
A 1
S 1 S 1 M = M S 1 = M S 2 = A 1
S 2 S 2 M,
which implies that A S 2 A 1
S 1 S 1 = S 2 . By the uniqueness part of Lemma 10.10,
S 1 = S 2 . Therefore, the map S → M S is an injection on the finite set S N ,
hence is a permutation of the set. Since j
A = j for A
SL 2 ( Z ), we have
F N ( X, Mτ )=
S
( X
j ( SMτ ))
S N
=
S
( X
j ( A S M S τ ))
S N
=
S
( X − j ( M S τ ))
S N
=
S
( X − j ( ))
S N
= F N ( X, τ ) .
The next to last equality expresses the fact that S
M S is a permutation of
S N , hence does not change the product over all of S N .
Since F N is invariant under τ → Mτ , the same must hold for its coecients
a k ( τ ).
LEMMA 10.12
For each k ,there existsaninteger n su ch that
q −n Z [[ q ]] ,
a k ( τ )
where Z [[ q ]] denotes pow er series in q withinteger coe cients. In other w ords,
a k ( τ ) can be expressed as a Laurent series w ithon y finitelymanynegative
term s, and the coe cientsareintegers.
PROOF
The j -function has the expansion
=
j ( τ )= 1
c k q k = P ( q ) ,
q + 744 + 196884 q +
···
k =
1
where the coe cients c k are integers (see Exercise 9.1). Therefore,
j (( + b ) /d )=
c k ( ζ b e 2 πiaτ/d ) k = P ( ζ b e 2 πiaτ/d ) ,
k =
1
Search WWH ::




Custom Search