Documentos de Académico
Documentos de Profesional
Documentos de Cultura
(IJERA)
ISSN: 2248-9622
www.ijera.com
Vol. 2, Issue 6, November- December 2012, pp.001-030
Abstract
System NP Hardness problem is
addressed and Mathematical framework is
proposed for certificates and enhancement of
security systems
Keywords- Key Management, Certificates, NP
I. INTRODUCTION
The problem of computing minimum cost
delegation chains supporting all the attributes in
certificate cert is NP-hard, even for configurations
with no authority classes. In fact, the minimum set
cover problem reduces to it in polynomial time, as
formally stated by the following theorem.
THEOREM A.1 (NP-HARDNESS). The problem
of computing minimum-cost delegation chains
supporting all the attributes in a certificate is NPhard.
PROOF. The proof is a reduction from the NP-hard
problem of theminimum set cover, which can be
formulated as follows: given a collection S of subsets
of a finite set U, determine a subset S_ S such that
every element in U belongs to at least one set in S_
and the cardinality of S_ is minimized. Given a finite
set U and a collection S of subsets of U, the problem
of computing a minimum set cover for U can be
translated into an equivalent instance of the problem
of computing a minimum cost supporting chain, as
follows. Each element in the finite set U translates to
an attribute certified by cert and that belongs to TT.
Any subset s in S translates to a delegation certificate
in Deleg Certs, issued by an authority in the
authoritative clause of TT, supporting the attributes
in s, and with cost equal to 1. The minimum-cost
supporting chain for cert corresponds to a subset S_
of S that completely covers U and with minimum
cardinality.
We consider the following anycast field equations
defined over an open bounded piece of network and
d
/or feature space R . They describe the
dynamics of the mean anycast of each of p node
populations.
p
d
(
l
)
V
(
t
,
r
)
dt i i
J ij (r, r )S[(V j (t ij (r, r ), r ) h | j )]d r
j 1
I iext (r , t ),
t 0,1 i p,
V
(
t
,
r
)
(
t
,
r
)
t
[
T , 0]
i
i
S ( z)
1
1 e z
(2)
note
the
p dimensional vector
p function i , i 1,..., p,
S ( x) [ S ( 1 ( x1 h1 )),..., S ( p h p ))],
and
the
diagonal
matrix
p p
L0 diag (l1 ,..., l p ). Is the intrinsic dynamics of
by
d
d
li ) is replaced by ( li ) 2 to use
dt
dt
d
the alpha function response. We use ( li ) for
dt
transfer. (
1|Page
(r , r ) whose element
described by a matrix
ij (r , r )
Proposition 1.0
satisfied.
between
1.
J L2 (2 , R p p ),
2.
is
the
propagation
delay
ij (r, r ) ij (r, r ),
p p
thus
max i , j (r , r )
i , j ( r , r )
(3)
ext
C 0 ( R, F ),
C 0 (2 , Rp p ),sup m .
Then for any C , there exists a unique solution
V C 1 ([0, ), F ) C 0 ([ m , , F ) to (3)
3.
Notice that this result gives existence on R , finitetime explosion is impossible for this delayed
differential equation. Nevertheless, a particular
solution could grow indefinitely, we now prove that
this cannot happen.
B. Boundedness of Solutions
A valid model of neural networks should
only feature bounded packet node potentials.
Theorem 1.0 All the trajectories are ultimately
bounded by the same constant
R if
Hence we choose T m
A. Mathematical Framework
A convenient functional setting for the nondelayed packet field equations is to use the space
infinite, hence
I max tR I ext (t )
Proof
:Let
us
defined
f : R C R as
def
V ,U
Vi (r )U i (r )dr
i 1
(1)
C C ([ m , 0], F )
0
space
supt[
m ,0]
(t ) F ,
with
V0 C ,
(2)
Where
L1 J
linear
continuous
L2 ( 2 , R p p )
2
F
( p J
I ) V (t )
Thus, if
operator
V (t ) F 2
p . J F I def
lR2 def
R, f (t ,Vt )
0
l
2
L1 : C F ,
the
f (t ,Vt ) l V (t )
Is
1d V
2 dt
If
V0
and
set
Suppose
2|Page
2
F
d
V
dt
BR ,
to
2
F t T
we
also
have
f (T ,VT ) 0
V (T ) BR which
contradicts the definition of T. Thus T R and
Because
f<0
on
BR is stable.
BR ,V (0) BR implies that t 0, V (t ) BR
V (0) CBR .
t 0,V (t ) BR ,
that
d
t 0, V
dt
2
F
2 ,
V (t )
thus
then
is
our
assumption.
Thus
T 0 | V (T ) BR .
functions on
be measured simple
X . for E M , define
(E) s d
(2)
i 1
i 1
r 1
r 1 i 1
r 1
Eij Ai B j ,
and
Eij
z1 and
z2
are
f ( z2 ) f ( z1 ) Where
subject
have lim ( ) 0
to
the
condition
(2)
f ( z ) ( z ) ( ),
2 ( )
(3)
(4)
And
1 ()( )
d d
( i ),
(5)
z
X
Where X is the set of all points in the support of
whose distance from the complement of K does
not . (Thus X contains no point which is far
within K .) We construct as the convolution of
f with a smoothing function A. Put a(r ) 0 if
r , put
a(r )
sd td i ( Eij ) j ( Eij )
disjoint
the
Eij
j } If
( s t )d ( i j ) ( Eij )
( z )
i ( Ai Er ) ( Er )
Eij
()( z )
shows that
( E ) i ( Ai E ) i ( Ai Er )
that
such
f ( z) P( z) 10,000 ( ) ( z K )
polynomial
Then is a measure on M .
(s t )d s d td
X
(1)
exists
there
union
of
X . Since X is
the
sets
(1
r2
(0 r ),
)2
(6)
And define
A( z) a( z )
(7)
3|Page
A 1,
R
f ( z rei )d
A 0,
R
( z ) a(r )rdr
(8)
2 f ( z ) a(r )rdr f ( z ) A f ( z )
(9)
R3
For all z
24 2
,
15
(10)
0,
a ,
( z ) f ( z ) Ad d A( z ) f ( )d d
(11)
R2
Since f and
. Since
Q j ( , z )
( z ) f ( z )
[ f ( z ) f ( z )] A( )d d (12)
R
Q j ( , z )
And A( ) 0 if
'
Now define
R2
50
(16)
1
4, 000 2
2
z
z
Qj ( , z) g j ( z) ( bj ) g 2j ( z)
Let
()( z ) (A)( z ) f ( )d d
R2
f ( z )(A)( )d d
R2
[ f ( z ) f ( z )](A)( )d d
(13)
R2
K exceeds . We shall do
( z ) f ( z )
( z G);
(14)
Note that f 0 in G , since f is holomorphic
there. Now if z G, then z is in the interior of
(17)
D j , if
(15)
R2
(18)
Define
R ( , z ) Q j ( , z )
( X j , z )
(19)
And
X
( z )
F ( z)
()( ) R( , z ) d d
(20)
Since,
F ( z)
j 1
()( )Q j ( , z )d d ,
(21)
4|Page
F ( z ) ( z )
2 ( )
1
| d d
z
| R( , z )
( z ) (22)
(23)
4, 000 2
1
2
1 2
dr
( , )d
r
2 0
As 0, ( , ) f ( z) uniformly.
gives (2)
(5)
This
X k X1,...X n , then
X a and
If
And
, with period
d 2, 000 .
(24)
F ( z) ( z) 6,000( )
( z )
(25)
monomial ideals in
H (), K , and S 2 K
is connected, Runges theorem shows that F can be
uniformly approximated on K by polynomials.
Since F
1
i
2 x y
df
(f )( )
d d
R2 z
(r , ) f ( z rei ), r 0,
i
If z re , the chain rule gives
some S
(2)
Put
A n | n ,
( i )
ideal corresponding to
(1)
f ( z)
real
Clearly
X , | S
A satisfies
, and
a X | A . Conversely, if A , then
for
X X X
some
and
n
from the fact that X | X . Let
1 i
(f )( ) ei
(r , )
2 r r
(3)
1
2
d dr
r r
(4)
S 1 ,... s
of
such
A n | i 2 , some i S
(The
i ' s
that
A ) Moreover,
5|Page
DEFINITION 1.0.
generated by
LT ( f ) | f a
k[ X 1 ,... X n 1 ][ X n ] k[ X 1 ,... X n ]
is
an
isomorphism this simply says that every
polynomial f in n variables X 1 ,... X n can be
expressed uniquely as a polynomial in X n with
LEMMA 1.2
g1 ,..., g n a .
PROOF. Since ( LT (a)) can also be described as
coefficients in k[ X 1 ,..., X n ] :
some
1.2.
Every
k X 1 ,..., X n is
finitely
ideal
generated;
in
LT (a)
Let
find
f a1g1 ... as gs r,
ai , r k X1,..., X n
, where either r 0 or no monomial occurring in it
LT ( g i ) . But
is
divisible
by
any
r f ai g i a
and
therefore
LT (r ) LT (a) ( LT ( g1 ),..., LT ( g s ))
implies that every monomial occurring in
,
is
LT ( g i ) . Thus r 0 , and
divisible by one in
g ( g1 ,..., g s ) .
DEFINITION
finite
subset
(Gr obner )
bases
for
if
A[ X ]
PROOF.
For a polynomial
f ( X ) a0 X r a1 X r 1 ... ar ,
ai A,
a0 0,
suppose
has
degree
s r , say,
'
ai leading coefficient of gi
Now
f bi gi X
1.1.
more
LEMMA 1.3. If
s ri
gi .
is
a
of
degree
can
be
written
d
f d f d 1
mod( g d ,1 ,...g d ,md ) With f d 1
6|Page
Such that
: D D D
: D D D
D D
f as an element
of D, it makes sense to require that ( ( f )) f ,
that is, o id DD Furthermore, we often
intended to represent the function
( (d )) d
f o f R idY
f g
Where
means
that
1.3:
Let
be
category and
pair
(A,a),
where
is
K-object
and
a : F ( A) A is an isomorphism. A prefixed
f1
f2
Do D1 D2 .....
is a Karrows
o id D
: X as a
' s components
: X is a cocone with
'
if : X is also a cocone
Similarly, a colimit
f R o f id X
or
suppose
are
D D D .Let us
working
with
the untyped
the equation
is
some
D D ,
such that
f1
f2
Do D1 D2 .....
cone
:X
op
of an chain is a K-object X
and a collection of K-arrows i : Di | i 0 such
i 0, i fi o i 1 . An op -limit of
7|Page
k : X' X
such that for all i 0, i o k i . We write k
(or just ) for the distinguish initial object of K,
when it has one, and A for the unique arrow
from to each K-object A. It is also convenient to
there exists a unique mediating arrow
write D1
f1
f2
Do and f 0 . By analogy, is i | i 1 .
For the images of and under F we write
except
F ( fo )
F ( f1 )
F ( f2 )
F ( ) F ( Do ) F ( D1 )F ( D2 ) .....
and
F ( ) F (i ) | i 0
i
that
is,
F o ( f ) f , F 1 ( f ) F ( f ), F 2 ( f ) F ( F ( f ))
of
! F ( )
F ()
F (! F ( ))
F (! F ( ))
2
F ()
2
.........
If both : D and F ( ) : F () F ( D)
are colimits, then (D,d) is an intial F-algebra, where
d : F ( D) D is the mediating arrow from
F ( ) to the cocone
NDk X 1 , X 2 ,.... X k 1
Let
Dk X k 1 , X k 2 ,.... X n
follows
dk
1 k m, (k , m) 1,
where
k ( x) mk ,
m being
mth root of
we can write
m x. One advantage of
8|Page
th
Q( m ) Q( n ) Q( ( m,n ) );
here
m, n and m, n denote
Q( m ) is
Q ( n )
where the
ei or fi to be zero)
Q( m ) Q( p )Q( p )...Q( p )
2
m Q( n ),
Q(
max( e1, f )
1
p1
).......Q(
max( ek , f )
k
p1
max( e1, f )
max( ek , f )
1 ........ p
k
1
p1
Q(m ,n );
An entirely similar computation shows that
Q( m ) Q( n ) Q( ( m ,n ) )
Mutual
information
I ( xi , yi ) log 2
(Recall the
Q( m ) Q( n ) Q
Q(
th
ek
and
m n is a primitive
Thus
Q( m , n ) Q( nm )
Q( m , n ) is the compositum of
Q( m ) and Q( n ) )
e2
and
Q( n ) Q( p f1 )Q( p f2 )...Q( p fk )
Since
Write
e1
1
If m divides n , then
P(
measures
the
xi is sent and y i is
xi
)
yi
bits
P( xi )
(1)
y i is uniquely
Q( m , n ) Q( nm ) We know that Q( m , n )
has degree (mn) over Q , so we must have
xi , and so they
constitute an input output pair ( xi , yi ) for which
1
x
P( i ) 1 and I ( xi , y j ) log 2
bits;
yj
P( xi )
and
Q( m ) Q( n ) Q
Q(m , n ) Q(m,n )
And
xi In a
P(
xi
yj
9|Page
yj
I ( X , Y ) P( xi , y j ) I ( xi , y j ) P( xi , y j ) log 2
i. j
i. j
P( xi )
P( xi , y j ) P(
xi
P( y j ) P(
yj
P( xi ) P(
i
xi
) P( y j ) P(
yj
xi
yj
yj
xi
) P( xi )
) P( xi )
) P( y j )
Then
I ( X , Y ) P ( xi , y j )
I ( X , Y ) I (Y , X )
I ( X , Y ) H (Y ) H (Y
Where
H (Y ) P ( y j ) log 2
j
1
P( y j )
I ( X ,Y ) H ( X ) H ( X )
Y
i. j
1
P ( xi , y j ) log 2
i. j
P ( xi )
P ( xi , y j ) log 2
x
i. j
P( i )
yj
1
P ( xi , y j ) log 2
i. j
P ( xi )
yj, H (X / yj )
P( xi )
P ( xi ) P ( y j )
i, j
i, j
Then
P( xi ) P( y j )
1
x
P ( i ) P ( y j ) log 2
y
j
P ( xi )
i
1
i P( xi ) log 2 P( x ) H ( X )
i
I ( X ,Y ) H ( X ) H ( X )
Y
Where H ( X
) i , j P( xi , y j ) log 2
I ( X , Y ) P( xi , y j )
i, j
P log
i 1
P(
1
xi
factor
yj
y j provides H ( X ) H ( X ) bits of
Y
P(
xi
yj
) P( y j ) P(
yj
xi
P( xi , y j )
Qi
)0
Pi
Q 1 . It can be concluded
i
that the average mutual information is a nonnegative number. It can also be equal to zero, when
the input and the output are independent of each
other. A related entropy called the joint entropy is
defined as
) P( xi )
10 | P a g e
P ( xi , y j ) log 2
i, j
1
P ( xi , y j )
P ( xi ) P ( y j )
tea P( X , Y ) A P X F
P ( xi , y j )
1
P ( xi , y j ) log 2
P ( xi ) P ( y j )
i, j
We
proceed as follows.
Let
B j 1 B j . ( If t 0, take B ). Then
t
probabilities
p ( x)
p ( x)
and P( y3 x ) 0
1
y
1
and P( x ) p
2
y
3
and P( x ) 1 p
2
p( y | x)
A ( x, y) : log
a
p( y )
(1)
uea P( X , Y ) A P X F
p( x, y )dx dy
p( y | x)dy dx
y Ax
yB Ax
p( y | x)dy dx p( x)
x
Algorithms
Ideals. Let A be a ring. Recall that an ideal a in A
is a subset such that a is subgroup of A regarded as a
group under addition;
a a, r A ra A
r s
i i
with
a b | a a, b b
set
is an ideal, denoted by
Where
P X F ... p ( x)dx
Proof: A sequence
and
where Ax y : ( x, y ) A;
Choose the decoding set
chosen
( x , y ) A
P( y3 x ) 1 p and P( y2 x ) 0,
2
P ( X , Y ) A
B1 to be Ax . Having
(1)
x k F such that
k 1
P Y Ax( k ) Bi | X x( k ) 1 ;
i 1
Set Bk Ax( k )
k 1
i 1
11 | P a g e
shows
that
and
b b | (a, b) c some a a
Let A be a ring. An A -algebra is a ring B
together with a homomorphism iB : A B . A
homomorphism of A -algebra B C is a
homomorphism of rings : B C such that
(iB (a)) iC (a) for all a A . An A -algebra
B is said to be finitely generated ( or of finite-type
over A) if there exist elements x1 ,..., xn B such
that every element of B can be expressed as a
polynomial in the xi with coefficients in i ( A) , i.e.,
such that the homomorphism
A X1 ,..., X n B
X i to
xi is surjective.
A ring
homomorphism A B is finite, and B is finitely
generated as an A-module. Let k be a field, and let
A be a k -algebra. If 1 0 in A , then the map
k A is injective, we can identify k with its
image, i.e., we can regard k as a subring of A . If
1=0 in a ring R, the R is the zero ring, i.e., R 0 .
sending
k be a field. A monomial
a . We sometimes abbreviate it by
.
polynomial ring
a1 ....an
q, r k X such that
f qg r with either r 0 or deg r < deg g .
unique
polynomials
(b1 ,...bn )
be two elements of
a1
1
X ...X ,
ca1....an k ,
aj
; then
XY 2 Y 3 Z 4 ; X 3Y 2 Z 4 X 3Y 2 Z . Note that
this isnt quite how the dictionary would order them:
it would put XXXYYZZZZ after XXXYYZ .
Graded reverse lexicographic order (grevlex). Here
monomials are ordered by total degree, with ties
broken by reverse lexicographic ordering. Thus,
if ai bi , or ai bi and in
the right most nonzero entry is negative. For
example:
the monomials in
an element f of
k X1 ,... X n in a canonical
f 4 XY 2 Z 4Z 2 5 X 3 7 X 2 Z 2
as
f 5 X 3 7 X 2 Z 2 4 XY 2 Z 4Z 2 (lex)
or
12 | P a g e
f 4 XY 2 Z 7 X 2 Z 2 5 X 3 4Z 2 ( grevlex)
ideal a (Y X ) contains Y X
2
Let
a X k X1,..., X n , in decreasing
order:
f a0 X 0 1 X 1 ...,
= X 0;
The leading term of f to be LT( f ) =
a0 X
but not
Y or X .
DEFINITION 1.5. An ideal a is monomial if
all with c 0 .
PROPOSITION 1.3. Let a be a monomial ideal,
0 1 ..., 0 0 c X a X a
Then we define.
and let A
condition
| X
A, n
And a is the
k -subspace of k X1,..., X n
()
satisfying
a of k X1 ,..., X n
A is
generated by
| A
is a monomial ideal.
f 4 XY Z ..., the
For the polynomial
multidegree is (1,2,1), the leading coefficient is 4,
2
the leading monomial is XY Z , and the leading
2
term is 4XY Z . The division algorithm in
2
r such
k X1,..., X n
X k X1 ,..., X n
X a
and
.
If a permutation is chosen uniformly and at
random from the n ! possible permutations in S n ,
then the counts
f a1 g1 ... as g s r
Where either
r 0 or no monomial in r is divisible by any of
LT ( g1 ),..., LT ( g s )
Step
1:
If
P[C ( n ) c]
that
(n)
( n)
n
n 1 c 1
1
N (n, c) 1 jc j n ( ) j
,
n!
j 1
j 1 j c j !
for c .
n
Lemma1.7
For
nonnegative
integers
m1,..., mn ,
Proof.
This can be established directly by
exploiting
cancellation
of
the
form
with
r2
for
f a1 g1 ... as g s LT (r1 ) r3
:
(different
n 1 m j
n
[m ]
E (C (j n ) ) j
j 1
j 1 j
[mj ]
cj
1 jm j n
j 1
(1.4)
13 | P a g e
(1.1)
d j c j m j , we have
n
n
[m ]
[m ]
E (C (j n ) ) j P[C ( n ) c] (c j ) j
j 1
j 1
c
n
n (c j ) j
1
jc
j
cj
c:c j m j for all j j 1
j 1 j c j !
[m ]
j 1
n
n
1
1
1
jd
d j
j
mj
j d j 1
j 1 j (d j )!
1(m n),
corresponding to the fact that if n m 0, then
d j 0 for j n m, and a random permutation
This last sum simplifies to the indicator
in
Snm
must
have
some
cycle
structure
follow
immediately as
(1.2)
[m ]
E (C (j n ) ) j E Z j j 1 jm j n ,
j 1
j 1
j 1
(1.3)
variables
that
satisfy
E ( Z j ) 1/ j
The marginal distribution of cycle counts provides
a formula for the joint distribution of the cycle
counts
Proof.
length
j k
k!
with
the
(1)l
l 0
j l
l!
(1.1)
1,2,...n,
so that
n[ rj ] 1
1
r
1(rj n) r
j r ! n!
j r!
Finally, the inclusion-exclusion series for the
number of permutations having exactly k properties
is
k l
Sk l ,
l
(1)
l
1 nk
1
(1)l ,
(1.2)
k ! l 0
l!
(n)
and the moments of C1 follow from (1.2) with
j 1. In particular, for n 2, the mean and
(n)
variance of C1 are both equal to 1. The joint
P[C1( n ) k ]
(n)
j n,
[ n / j ] k
Sr (n rj )!1(rj n)
l0
E(C(j n) )[ r ] j r 1 jr n
distribution
[ rj ]
I n[ j ]/ j . For each I ,
that is,
li
b
b 1 ci 1
1 1
l1 ... lb
(
1)
i 1 i li !
i 1 i ci ! l 0 with
ili n m
The
(n)
1
joint
moments
(n)
b
C ,..., C
of
the
b counts
first
P[C (j n ) k ]
j k 1/ j
e , k 0,1, 2,...,
k!
14 | P a g e
(1.3)
( n)
j
d Z j
'
n ,
(1.1)
Z j , j 1, 2,...,
the
are
1
with E ( Z j )
j
as i , for
P[T0 m ( Z ' ) n]
P[T0 m ( Z ) n]
P[ An (C ( n ) )]
and
1i n
ri' 1 j ri
1 (1 Ei 0 )
iri
(1.1)
P[T0 n ( Z ' ) n]
exp [log(1 i 1 d ) i 1 d ]
n
i 1
1 O(n
'1,2,7 (n))
(1.2)
and
P[T0 n ( Z ' ) n]
exp [log(1 i 1 d ) i 1 d ]
n
i 1
1 O(n
Error rates
The proof of Theorem says nothing about
the rate of convergence. Elementary analysis can be
used to estimate this rate when b 1 . Using
properties of alternating series with decreasing
terms, for k 0,1,..., n,
Where
1
1
1
(
) P[C1( n ) k ] P[ Z1 k ]
k ! (n k 1)! (n k 2)!
1
k !(n k 1)!
It follows that
n
2n 1
n
2n 1 1
P[C1( n ) k ] P[ Z1 k ]
(n 1)! n 2 k 0
( n 1)!
(1.11)
1,2,7 (n))
'1,2,7 (n)
Z
from
'
(n)
P[ An (C )] Kn
depending on
It
thus
(1 d )
P[ Z1 n]
e 1
1
1
1
(1
...)
,
(n 1)!
n 2 (n 2)(n 3)
(n 1)!
(n)
and the
the
asymptotics
of
where
An (C ( n ) )
( n)
ij
1i n
ri' 1 j ri
'
that
and computable
n1 if g ' 1.
distribution L ( Z1 ) of Z1
Establish
follows
for a constant
and the ri
approximation is of order
Since
(1.3)
0 ,
Where
under Conditions
15 | P a g e
implied by
r A
P[Tbn ( Z ) n r ]
1
P[T0 n ( Z ) n]
(1.4)
l 2
l il O(i a1 ) to
i1
P[T0 n n]
r 0
[ n /2]
P[T0b r ]
P[T0b r ]
r n /2
r 0 P[T0 b n ]
order
r n /2
[ n /2]
P[T0b r ] P[T0b r ]
r 0
[ n /2]
P[Tbn n s] P[Tbn n r ]
s 0
P[T0 n n]
P[T0b s]
[ n /2]
P[T0b r ]
s 0
s [ n /2]1
3n
,
P [0,1]
12
10.8
(n) ET0b
P [0,1] n
Hence we may take
(n / 2, b)
P [0,1] 10.5(1)
(n)
P
P [0,1]
10.8
(1.5)
replaced by 10.11
10.10 (n),
The
factor
1 (n) u1 (n),
contribution of order O (n
[ n /2]
[ n /2]
3n
1
4n210.8 (n) P[T0b r ] P[T0b s] r s
P [0,1]
2
r 0
s 0
210.5(1) (n / 2, b)
P[Tbn( m ) s ] P[Tbn( m ) s 1] .
1 a1
) in the estimate
of the difference P[Tbn s ] P[Tbn s 1],
which, in the remainder of the proof, is translated
1 a
bn a1 for any 0 in
defining
the
would seem to be
function
by
16 | P a g e
[ n /2]
r 0
[ n /2]
[ n /2]
( s r )(1 )
P[T0b s ]
P[T0 n n] )
n 1
s 0
1
dTV ( L(C[1, b]), L( Z [1, b])) (n 1) 1 1 E T0b ET0b
2
7,8 (n, b),
12
]) for any
s [ n /2]1
(1.2)
0b
[ n /2]
(s r )(1 )
r ] P[T0b s]
n 1
s 0
( s r )(1 )
P[T0b s]
n 1
s 0
[ n /2]
P[T
0b
r]
s [ n /2]
P[T0b s ]
(s r ) 1
n 1
1 n E (T0b1T0b n / 2) 2 1 n 2 ET02b ,
1
P[T0b n / 2]
P [0,1]
r 0
r 0
n /2 s n
3 10.5(2) (n / 2, b)
P[T
8n 2 ET02b
3
)
nP [0,1]
[ n /2]
r
]
1
0b
P[T0 n n] r 0 P[T0 n n]
r 0
6
ET
(n, b)
nP [0,1] 0b 10.14
P[T0 n n]
r 0
1
P[T0b s ] s r
P[T0b r ]
n P[T0 n n] r 0
s 0
2
Where
P[T0b r ]
P[T0 n n]
(1.1)
(1.3)
P[T0b r ] P[T0b s]
n 1
r [ n /2]
s 0
1
n 1 ( ET0b P[T0b n / 2] E (T0b1T0b n / 2))
We have
4 1 n 2 ET02b
find
(1.4)
17 | P a g e
7.8 (n, b)
directions:
(1.5)
( B12 )
, provided that
supplementary condition can
10.8
(n) is replaced by
S () ;
definition of
7.8 (n, b) ,
this
removed if
be
10.11
(n)
in the
P[T
r 0
0b
1
1 E T0b ET0b
2
Example
1.0.
Consider
coordinates
the
point
of
the
vector
n as an
Remark 1.0.
Euclidean geometry.
In
considered as an affine space we can already do a
good deal of geometry. For example, we can
consider lines and planes, and quadric surfaces like
an ellipsoid. However, we cannot discuss such
things as lengths, angles or areas and
volumes. To be able to do so, we have to introduce
n
some more definitions, making a Euclidean
space. Namely, we define the length of a vector
a (a1 ,..., a n ) to be
a : (a1 )2 ... (a n )2
(1)
as an Euclidean space
n
24 1 10.8 (n)
2n 2 ET02b 4 3 1
P [0,1]
n : n=
= {vectors}
{points},
Operations with vectors: multiplication by a
number, addition. Operations with points and
vectors: adding a vector to a point (giving a point),
n
subtracting two points (giving a vector). treated
in this way is called an n-dimensional affine space.
(An abstract affine space is a pair of sets , the set
of points and the set of vectors so that the operations
as above are defined axiomatically). Notice that
d ( A, B) : AB
(2)
(3)
cos :
( a, b)
a b
(4)
18 | P a g e
(a, b)2 a b
2
(5)
Example 1.1.
dx (the
differential of x ) applied to an arbitrary vector h
i
is simply h .From these examples follows that we
can rewrite df as
f
f
df 1 dx1 ... n dx n ,
(1)
x
x
(the i-th coordinate). The linear function
i
f 1
h
x1
(2)
f ( x(t0 t ) f ( x0 )
df ( x0 )( .t (t )t )
( .t (t )t ). t (t )t
df ( x0 )( ).t (t )t
For a certain
(t )
such that
(t ) 0
df ( x(t )) f 1
f
1 x ... n x n
dt
x
x
To calculate the value Of
(2)
df at a point x0
f ( x(t )) at t t0 .
Theorem 1.8.
For functions
U ,
d ( f g ) df dg
d ( fg ) df .g f .dg
f , g :U ,
(1)
(2)
Hence
Suppose we have a parametrized
curve
d ( f g )( x0 )( )
at t t0 and
d ( fg )( x0 )( )
d
( f ( x(t )) g ( x(t )))
dt
d
( f ( x(t )) g ( x(t )))
dt
F : U m at a point x
n
will be a linear function taking vectors in to
m
vectors in (instead of ) . For an arbitrary
n
vector h | ,
differential of a map
F ( x h) F ( x) dF ( x)(h)
19 | P a g e
(h) 0
(3)
h 0 . We have
when
(4)
which vector in
map
F ( x h) F ( x) dF ( x)(h) (h) h
(3)
Where (h) 0 when h 0 . we obtain
.
h
.
F ( x) dF ( x)( x(t )t (t )t )
.
( x(t )t (t )t ). x(t )t (t )t
.
it is taken by
d (GoF ) . the
dF of
under
the
vector
x (t ) .
Hence
for an arbitrary
vector x .
Corollary 1.0.
If we denote coordinates in by
F ( x) dF ( x)( x(t )t (t )t
For some
(t ) 0
when t 0 . This
.
G1 G1
1 .... m
y
y
d (GoF ) ... ... ...
p
p
G ... G
y1 y m
F 1 F 1
1 .... n
x
x
... ... ...
m
m
F ... F
n
x1
x
dx1
...
dx n
(4)
20 | P a g e
y1 y1
1 .... n
x
x
... ... ...
m
m
y ... y
1
x n
x
Here
depending on
, x
the vectors x
(6)
, x
. Explicitly we have
x
(cos ,sin )
r
x
( r sin , r cos )
(3)
(4)
.
n
, denoted for
F 1 ( x) ny
In other words,
F : ( y1..., y n ) x x( y1..., y n )
(1)
n
specified
(1)
, x
are,
x(r , ) (r r0 fixed )
. We can conclude
that for an arbitrary curve given in polar coordinates
.
, e : x
x er r e
x(t ) : r r (t ), (t )
m
z
z y i
x a i 1 y i x a
(2)
(5)
Or
dx x dr x d x . x .
dt r dt dt r
(5)
x i
(originally, a
xi x( x1 ,..., x n )
21 | P a g e
dF ( x0 ) :
dx(t )
dF ( x(t ))
(t0 )
(t0 )
dt
dt
(1)
attached to a point x0
the point F ( x )
, the integral
to vectors attached to
parametrization of
remains the same.
F 1
F
dx ... n dx n
1
x
x
1
F F 1
1 ... n dx1
x
x
Proof:
(2)
f
f
dx1 ... n dx n ,
(3)
1
x
x
i
Where x are arbitrary coordinates. The form of the
df
have
Substituting into
i
j
(2)
x
dxi (e j ) dxi j
x
every point x .
A , we get
at
( x(t )),
Consider
dx
dt '
and
dx
dt '
dx
dx dt
'
= ( x(t (t ))), ' . ' ,
'
dt
dt dt
As
r 2 (sin 2 cos 2 )d r 2 d
Hence A r d is the formula for A in the
polar coordinates. In particular, we see that this is
again a 1-form, a linear combination of the
differentials of coordinates with functions as
coefficients. Secondly, in a more conceptual way,
we can define a 1-form in a domain U as a linear
function on vectors at every point of U :
2
( ) 1 1 ... n n ,
If
ei i , where ei x
(1)
x i
22 | P a g e
23 | P a g e
24 | P a g e
AND
visited yet, that is, a p attrsToCheck. The repeatuntil loop follows the corresponding path by
exploiting the predecessor global variable, and the
corresponding delegation certificates are added to
ver list. Since function FindChain correctly builds
the delegation chains through variable predecessor
(Lemma A.4) and the graph is acyclic, at each
iteration the repeat-until loop visits an edge of the
delegation chain until cert.issuer is reached.
Consequently, the repeat-until loop terminates, and
if ToCheck is empty, the function returns the set ver
list of certificates supporting attributes in
cert.attributesAttributes(E).
THEOREM
4.6
(TERMINATION
AND
CORRECTNESS). Given a finite set of delegation
certificates Deleg Certs, a finite set of authority
certificates Authority Certs, and a certificate cert
compatible with an entity E, function Satisfy
terminates and determines a correct set of delegation
chains for cert, if such delegation chains exist.
PROOF. We first prove that the function terminates
and then that if it returns a nonempty set of
certificates, they correspond to delegation chains
that
support
all
attributes
in
cert.attributesAttributes(E).
The
preliminary
operations of function Satisfy check whether or not
cert.issuer
appears
in
Except(E)
and
Authoritative(E). Since such sets are finite, the
preliminary operations terminate. Analogously, since
both function FindChain and function
BuildVerificationList terminate (Lemma A.4 and
Lemma A.5); both Phase 1 and Phase 2 of function
Satisfy also terminate. We now prove by
contradiction that if there are delegation chains
supporting
all
attributes
in
cert.attributesAttributes(E),
function
Satisfy
returns them. Suppose that all attributes in
cert.attributesAttributes(E) are supported by a
unique delegation chain and let s be the set of
delegation certificates forming the delegation chains
for such attributes. Suppose also that function
Satisfy returns an empty set. In this case, solution s
cannot coincide with cert, since Satisfy checks if cert
represents a solution to the problem. Then set s has
to include more than one certificate. Since function
Satisfy returns the empty set, function FindChain
returned an empty Candidates list. Therefore, the
while loop in function FindChain terminates because
Queue becomes empty, while ToCheck includes at
least one attribute a. By assumption, we know that
there is a supporting path p for attribute a also, going
from a valid authority for E to cert.issuer. By
Lemma A.3, the edge in p entering cert.issuer and
represented by element [from, cert.issuer, p attrs, p
cost] is inserted into Queue before the while loop.
Since, at the end of the while loop, Queue is empty,
within an iteration of the while loop this edge is
extracted from Queue, and therefore any edge
25 | P a g e
Let
p, then j is a
26 | P a g e
th
, 2 ,...,
p 1
. Therefore
TrK / ( j ) 2 ... p1 p ( ) 1 1
has
If
only
one
conjugate
1,
and
find that
p 1
OK [ p ] [ x] / ( p ( x));
) p
p 1
p 2
... 1 p ( x)
( x )( x )...( x
Plugging in x 1 shows that
p (1 )(1 2 )...(1
2
p 1
);
p 1
PROOF.
Let
OK and write
With ai .
Then
(1 ) a0 (1 ) a1 ( 2 ) ...
a p 2 (
Thus
a0 a1 ... a p2 p2
p 2
p 1
LEMMA 1.9
(1 )OK p
Proof. We saw above that p is a multiple of
OK ,
in
so
the
inclusion
(1 )
(1 )OK p is immediate.
Suppose
now that the inclusion is strict. Since
(1 )OK is an ideal of containing p
and p is a maximal ideal of
, we must have
1.1
For
any
OK ,
TrK / ((1 ) ) p.
PROOF.
We have
p 1
Example 1.4
( p)
a OK
of
OK .
a (a OK )OK , p ;
We
claim
that
1 (1 ) 1 ( ) ... p 1 (1 ) p 1 ( )
(1 ) 1 ( ) ... (1
p 1
) p 1 ( )
27 | P a g e
that
/ (a OK )OK , p ,
as
a (a OK )OK , p p bOK , p p OK , p
n
(OK ) OK ,
Since
/ OK , p
m 1 m 1
N K' / ( OK ) N K / ( )
(OK ) / ( ) OK / ;
this
shows
since
that
that
[1]
[2]
28 | P a g e
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[31]
[32]
29 | P a g e
[34]
[35]
[36]
[37]
30 | P a g e