Information Technology Reference
In-Depth Information
Theorem 2. Let f, g : V n F 2 be two Boolean functions. Suppose that f and
g are related by
g ( x )= f ( Ax + b )+ c
·
x + d, where
A
O (2 ,n ) ,b,c
V n ,d
V 1 .
Then, if f is bent-negabent, g is also bent-negabent.
Proof. As discussed above, g is bent if f is bent. It remains to show that g is
negabent. From [3, Lem. 2] we know that, if f ( Ax ) is negabent, so is f ( Ax +
b )+ c
x + d . It is therefore sucient to assume that b and c are all-zero vectors
and d =0.Observethat
·
wt( x )= x T Ix,
where I is the n
×
n identity matrix and the matrix operations are over
Z
.We
therefore have
( g )( u )=2 2
x∈V n
1) f ( Ax )+ u·x i x T Ix .
N
(
Now, since A is invertible by assumption, there exists B such that AB = I .
Moreover, when x ranges over V n ,sodoes Bx .Thus,
( g )( u )=2 2
x∈V n
1) f ( x )+ u·Bx i ( Bx ) T I ( Bx ) .
N
(
O (2 ,n )andso B T IB = I . Hence,
Since A
O (2 ,n ), we have B
( Bx ) T I ( Bx )= x T ( B T IB ) x = x T Ix.
We conclude
( g )( u )=2 2
x
1) f ( x )+ u·Bx i x T Ix
N
(
V n
=2 2
x∈V n
1) f ( x )+ B T u·x i wt( x )
(
( f )( B T u ) ,
=
N
which proves the theorem.
It is known that the dual of a bent function is again a bent function, and it
was proved in [3, Thm. 11] that the dual of a bent-negabent function is also
bent-negabent. The following theorem generalizes this concept by showing that,
if a bent-negabent function is bent with respect to certain subspaces, then the
corresponding partial duals are also bent-negabent.
Theorem 3. Write V n = U
W ,where dim W = k and k
n , so that dim U =
n
k .Let f : V n F 2 be a bent-negabent function that is bent with respect to
U and bent with respect to W .Then f W is also bent-negabent.
Search WWH ::




Custom Search