Cryptography Reference
In-Depth Information
Conditions for which monoidic implies Cauchy.
Theorem 1. Let M ( h ) be A -adic for a sequence h of length N over
F
.Then M
is Cauchy iff
(1)
h ( a i ) are distinct and invertible in
F
for all 0
i<N ,and
a j )) 1 =( h ( a i )) 1 +( h (
a j )) 1
( h (0)) 1
(2)
( h ( a i
for all 0
i, j < N .
In this case M ( h )= C ( β, γ ) ,where β ( a i )=( h ( a i )) 1
and γ ( a i )=( h (0)) 1
a i )) 1 .
( h (
Proof. We start by showing that our conditions indeed imply that M is Cauchy.
For the disjointness, assume that there are indices i and j , such that β ( a i )=
γ ( a j ). In this case we get 0 = β ( a i )
a j ), which is a con-
tradiction. Finally we compare the matrices M ( h )and C ( β, γ ) resulting in the
equality
γ ( a j )=1 /h ( a i
M i,j = h ( a i
a j )=1 / (1 /h ( a i )+1 /h (
a j )
1 /h (0)) = 1 / ( β ( a i )
γ ( a j )) = C i,j .
We continue by showing that if M is Cauchy, i.e., M ( h )= C ( β ), then indeed
our conditions must hold. Since C ( β )= C ( β + ω, γ + ω ) for any ω
,we
can choose the sequences in such a way that γ (0) = 0. Now, M i, 0 = C i, 0 for all
i ,whichmeans h ( a i )=1 ( a i ). By the properties of β this gives us condition
(1), i.e., that all h ( a i ) are distinct and invertible, as well as β = β .Weuse
similarly that M 0 ,i = C 0 ,i which implies h (
F
γ ( a i )). Solving for
γ revealsthatitequals γ .Since β = β and γ = γ ,wegetthat M ( h )= C ( β, γ )
implying condition (2).
a i )=1 / ( β (0)
Note that if the A -adic matrix of a sequence h is also Cauchy, then the se-
quence of r -th powers, i.e., h r =( h 0 ,h a 1 ,...,h a n 1 ) yields the correspond-
ing Cauchy power matrix. In other words, for any number r> 0wehave
M ( h )= C ( β, γ )=
M ( h r )= C ( β, γ, r ).
Now, we will show how to construct random monoidic Cauchy matrices.
Construction of monoidic Cauchy matrices.
Corollary 1. Let A be a finite, abelian group with set of generators b 1 ,...,b d
and M ( h ) be A -adic and Cauchy for a sequence h over
F
, then for all c 1 ,...,c d
Z
,
+ c d b d )) 1 = c 1 ( h ( b 1 )) 1 +
+ c d ( h ( b d )) 1
1)( h (0)) 1 .
( h ( c 1 b 1 +
···
···
( c 1 +
···
+ c d
Furthermore, the field characteristic char( F ) divides the order of any element in
A
\{
0
}
.
Proof. By Theorem 1, we know that for all a, a
A the following holds
( h ( a + a )) 1 =( h ( a )) 1 +( h ( a )) 1
( h (0)) 1 .
 
Search WWH ::




Custom Search