Académique Documents
Professionnel Documents
Culture Documents
FT ch5
FT ch5
X i , i N
n }
iN n
N n {1, 2,
, n}
R( X 1 , X 2 ,
, Xn)
i.e.,
$
|
a set
X 1 X 2
Xn
$
|
Characteristic function:
R ( x1 , x1 ,
xn )
xn ) R
1 if ( x1 , x1 ,
otherwise
0
Representation of a relation
R ( X 1 ,..., X n )
(ri1 ,i2 ,..,in )
ri1 ,i2 ,..,in
= 1
iff
( x1 ,..., xn ) R
otherwise
Example 5.1 :
R ( X Y ) Z )Y1Y2Y3Y4 Z 2 Z1Z 3 Z 4 Z 5
R ({ X 1 , X 3}
, Xi J
N nY
{ X i | j J N n } R2
1
1
1
1
0.8
0.8
0.8
0.8
y { X 1 , X 3}
Y j X j j J
[ R X Y ]:[ R2 { X 1 , X 3}]
xf y
Y j | j J X X j
jJ
.
a,* a,$ b,* b,$
0.9
0.4
1
0.7
0.8
R1,3
X , a X ,b Y , a Y ,b
X ,* X ,$ Y ,* Y ,$
1
0
0.6
0.9
0.7
0 R ( x1 , x2 ,..., xn ) 1
mark}
Z={US , France , Canada , Britain , Germary}
R( X Y Z )
France)
(English , dollar , Canada) , (French ,
dollar , Canada)
(English , pound , Britain)}
Y1
Dollar 1 0 1 0 0
Dollar
0 1
Pound
0 0
0 0
Y2
Pound 0 0 0 1 0
0 0
Y3
Franc
0 0 0 0 0
Franc
1 0
0 0 0 0 0
Mark
0 0 0 0
0 0
Y4
Mark
0
US Fran Can Brit Ger
US
Z2
Z3
Z4
Z5
Z1
Z2
Z3
Z4
Z5
English
Franch
X1
X2
Fuzzy Relations
Cartesian Product :
tuples :
X 1 X 2
Xn
( x1 , x2 ,..., xn )
membership grade :
0 R ( x1 , x2 ,..., xn ) 1
R( X Y )
1
0
0.6
0.9
0.7
0.3
Paris
1
0.9
0.7
0.6
0.3
set family X = { X
| i Nn}
Let X =
xi | i N n X X i
Let Y =
Yj | j J X X j
Where
jJ
jJ
J Nn
,|J|=r
Y a subsequence of X ,
iff Y
YX
X j j J
R ( X 1 , X 2 ,
, Xn)
Y= { X
[ R y]
: a relation
| j J Nn}
[ R y ](Y ) max R ( x)
xf y
Example 5.3
R( X 1 , X 2 , X 3 )
X1
0.9
0.4
0.7
0.8
Let R = [ R { X , X }] ,
ij
R1,2
Ri [ R { X i }]
0.9
0.4
1
0.8
X , a X ,b Y , a Y,b
R1,3
0.9
0
1
0.8
X ,* X ,$ Y ,* Y ,$
R2,3
R1
0.9 1
*
y
R2
1 0.8
a b
1 0.8
R3
* $
Cyclindric Extension
[R X Y ]
the CE of R into
X-Y
X-Y : sets
Xi
[ R X Y ]( x) R ( y )
R: a relation defined on Y
Let X = { X , X , X }
And R= R
y { X 1 , X 3}
1,2
X-Y =
= {*,$}
X3
1,2
{ X 3}]
0.9
0.4
1
0.8
X , a X ,b Y , a Y,b
R1,2
0.9
0.9
0.4
0.4
1
1
0.8
0.8
[ R2 { X 1 , X 3}] [ R3 { X 1 , X 2 }]
Consider
[R X Y ]
[ R2 { X 1 , X 3}]
{x,y} {x,$}
R= R
1 0.8
a b
[ R2 { X 1 , X 3}]
1
1
1
1
0.8
0.8
0.8
0.8
5-7
Cyclindric closure
-A relation may be exactly reconstructed from
several of its projections by taking the set
intersection of their cyclindric extensions
Pi | i I :a
Yi
Pi
is defined.
Example
R( X , X , X )
in example
5.3.
It is not fully reconstructable from its projections
become of ignoramus of
R1
, R ,and
2
R3
5-8
X Y bipartite
graph
graph
X Y directed
Representations
i, matrices
R [ rij ]
, where
y1
y3
y4
y5
0
0
x1
.9
x2 0
1
.4
0
0
0
0
x3
.2
x4
x5
0
0
0
0
0
0
0
0
.4
.5
.2
x6
ii)
y2
rij R ( xi , y j )
5-9
Domaindom R
Crisp dom R = {x X | ( x, y) R, y Y }
Fuzzy dom R(x) =
max R ( x, y )
yY
Rangeran R
Crisp ran R = { y Y | ( x, y) R, x X }
Fuzzy ran R(y) =
max R ( x, y )
xX
e.g.,
h( R ) 1
xX
5-10
Inverse R
(Y , X )
R 1 ( y, x ) R ( x, y )
R 1 R T , ( R 1 ) 1 R
e.g.
0.3
R 0
0.6
0.3
R 1 R T
0.2
0.2
1
0.4
0.6
0.4
0
1
Composition R( X , Z ) P( X , Y ) Q(Y , Z )
R ( x, z ) [ P Q]( x, z ) max min[ P ( x, y ), Q ( y , z )]
yY
Max-min composition
PQ QP
Properties
1
( PQ ) 1 Q
P 1
) R P(QR )
( PQ
Matric form [r ] [ p
ij
ik
ik
] [ qkj ]
, qkj )
5-11
R ( x, z ) [ PQ]( x, z ) max[ P ( x, y
) Q ( y , z )]
yY
max-product composition
matrix form
[ rij ] [ pik
] [ qkj ]
Where [r ] max( p
ij
ik
, qkj )
Example
0.3
0.0
0.4
0.5
0.7
0.6
0.8 0.9
1.0 0.3
0.5 1.0
Max-min =
Max-prod =
0.8
1 .0
0.5
0.8
1.0
0.5
0.5
0.2
0.7
0.0
0.0
0.5
0.3
0.2
0.5
0.5
0.4
0.5
0.7
0.9
0.5
0 .5
0.7
0.6
0.15
0.14
0.4
0.5
0.2
0.28
0.45
0.63
0.54
5-12
Example
[ PQ]( x, z ) max[ P * Q]( x, y, z )
yY
5-14
characteristic Properties (Crisp case)
i,
reflexive
irrflexive
antiflexive
ii,
symmetric
asymmetric
antisymmetric
( x, y ) R, ( y, x) R x y
strictly antisymmetric
iii, transitive
nontransitive
x y, ( x, y ) R
or
( y, x) R
antitransitive
Fuzzy Relations
i, reflexive ---
irreflexive ---
R ( x, x ) 1
antiflexive ---
R ( x, x ) 1
-reflexive ---
R ( x, x )
R ( x, x ) 1
5-14
ii, symmetric
asymmetric
---
x, y , R ( x, x ) R ( y, x)
x , y ,
antisymmetric --
R ( x, x ) R ( y , x )
R ( x, x ) 0
R ( y, x) 0
x y
x, z
R ( x, z ) max
min R( x, y ), R( y, z )
yY
max-product transitive --
x , z
R ( x, z ) max
min R( x, y ) R( y, z )
yY
nontransitive --
x , z
R ( x, z ) max
min R( x, y ), R( y, z )
yY
antitransitive --
x, z
R ( x, z ) max
min R( x, y ), R( y, z )
yY
5-15
Crisp:
equivalence;
Fuzzy:
similarity
Quasiequivalence
Compatibilit
y or
Tolerance
Partial
ordering
Preordering
or
Transitive
Antisymmetric
Symmetric
Antireflexive
Reflexive
Summary
Quasiordering
Strict
ordering
Figure3.6 Some important types of binary relation
R(X,X)
transitive closure:
RT ( X )
RT
R / R ( R oR )
2. If
R/ R
, Let
R/ R
, go to step1
3. Stop,
RT R /
Where
component-wise max
5-16
Example 5.8
0.7
0.0
R
0.0
0.0
Step1:
R oR
R ( R oR )
Step2: Q R
0.7
0.0
0.0
0.0
/
0.5
0.0
0.4
0.0
0.7
0.0
0.0
0.0
0.5
0.0
0.4
0.4
0.0
0.0
0.0
0.8
0.5
0.0
0.0
0.4
0.0
0.8
0.0
0.8
, Let
0.0
1.0
0.0
0.0
0.0
0.8
0.0
0.0
0.5
1.0
0.4
0.0
0.5
0.0
0.4
0.0
R/
R R/
repeat step1
0.7 0.5 0.5 0.5
0.0 0.4 0.8 0.4
R oR
0.0 0.4 0.4 0.4
Step3: Q R
, Let
R R/
repeat step1
0.7
0.0
'
R
0.0
0.0
0.5
0.4
0.4
0.4
Step4: Stop
0.5
0.8
0.4
0.8
0.5
1.0
R
0.4
0.4
RT R /
5-17
5.5 Fuzzy Equivalence Relation
Crisp binary relation
equivalence: reflexive, symmetric, and
transitive
equivalence classes
partition: X/R
Example 5.9:
X 1, 2,L ,10
R( X X )
divided by 3}
5-18
Fuzzy Binary Relation
Fuzzy
Similarity relation
equivalence relation
Similarity classes
equivalence classes
2.
x X ,
Ax
defined on X .
a fuzzy relation
[0,1]
(2.1)(2.2))
If R: Similarity relation,
R :
Let
equivalence relation
( R) :
the partition of
w.r.t.
( R ) ( R) | 0,1
( R) :
nested, i.e.,
( R) :
a redefinement of
( R)
iff
R: X X
is a similarity
is a equivalent relation
Pf : R : a similar relation
R : reflexive, i.e.,
symmetric, i.e.,
transitive, i.e.,
i,
x X , R ( x, x) 1
x, y X , R ( x, y ) R ( y, x )
x, z X 2 , R( x, z ) max[ R ( x, y ), R ( y, z )]
: reflexive
Q x X , R ( x, x) 1 [0,1], ( x, x) R
ii,
: reflexive
: symmetric
R : symmetric
x, y Z , R ( x, y ) R ( y, x)
Let
R ( x, y ) R ( y , x )
Then
iii,
or
a, if
=>
( x, y ), ( y, x) R
b, if
=>
( x, y ), ( y, x) R
: transitive
R : transitive
yY
x, z X 2 , R( x, z ) max[ R ( x, y ), R ( y, z )]
yY
Let
R ( x, y ) 1
Assume
R ( y, z ) 2
1 2
1 2
Then
a. if
=>
, or
1 2
( x, y ) R , ( y , z ) R
--- (A)
Q R ( x, z ) max[ R ( x, y ), R ( y, z )] min[ 1 , 2 ] 1
yY
R ( x , z ) , ( x, z ) R
--- (B)
R : transitive
1 2
( x, y ) R, ( y, z ) R
c. if
, dont care
( x, z )
, dont care
( x, z )
1 2
( x, y ) R, ( y, z ) R
Example 5.10 :
R( X , X )
: a fuzzy relation
Q R ' R ( R oR ) R
level set :
's
0 0 1 0 1 0.9 0.5
a b c d e
f
g
For e :
0 0 1 0 1 0.9 0.5
a b c d e
f
g
tolerance
proximity
Crisp case :
relation
fuzzy compatibility
relation
maximal ---- compatibility classes
complete-cover
Example 5.11 :
R( X , X )
: a fuzzy relation
R : reflexive , symmetric
a compatibility relation
5-25
0.5
0.7 0.9
b
d
a 1 0.7
b 0
1
c 0.5 0.7
d 0
0
e 0 0.1
e.q.
1
1
0
0.5
R 0
0
0
0
0
1
1
0
0
0
0
0
0
0
0 1 0.7
0 0.9 0
1 1 0.8
0 1
0
0 0.9 1
0 0 0 0
0 0 0 0
1 1 1 0
1 1 1 1
1 1 1 1
0 1 1 1
0 1 1 0
0 0 1 0
0 0 0 0
0
0
0
1
1
0
1
0
0
0
0
0
0
1
0
0
1
0
0
0
0
0
0
0
0
0
1
x y xx yy xx X ( x, y ) XS {x1 , x2 } Xy Ax y
(x,y)y X ( x y, or y xAx XA Xx yR[ x ] ( y ) R ( y , x )
x U ( R, A)( x ) I R[ x ]y
xA
1
1
0
0.4
R 0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
1
1
1
0
0
0
0
0
0
1
1
1
1
1
0
0
0
0
1
1
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
0
1
1
1
1
0
0
0
0
0
0
1
0
0
1
0
0
0
0
0
0
0
0
0
1
0.4
(1,2) , (3,4,5),(4,5,6,7),(5,8),(9)
(34,),(4,5,6),(4,5,7),(3,5),(5,6)
(4,5),(5,6,7),(4,6,7),(4,6),(6,7)
cover):
(1,2),(3,4,5),(4,5,6,7),(5,8),(9)
These do not partition X.
5-26
e.q.
0.5
1
1
0
0.5
R 0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
1
1
1
0
0
0
0
0
0
1
1
1
1
1
0
0
0
0
1
1
1
1
1
1
0
0
0
0
1
1
1
0
0
0
0
0
0
1
1
0
1
0
0
0
0
0
0
1
0
0
1
0
0
0
0
0
0
0
0
0
1
5-27
5.7. Ordering Relations
partial ordering: reflexive , antisummetric ,
transitive
X Y
: X : predecessor
precedes Y : successor
if exist
First member : if
x y y X
(minimum)
unique
Last member : if
y x y X
(maximam)
yx x y
properties :
1, if , at most one first member
if , at most one last member
2, There may be several maximal and minimal
member
3, if a first member X ,
inverse
ordering
partial ordering
5-28
In a partial ordering , it does not guarantee that
,
(x,y)
If ,
( x y, or y x
).
A X
If
x X
, and
x: lower bound of A on X
If ,
y A
x y
x y
x : upper bound of A on X
S {x1 , x2 } X
5-29
Connected a partial ordering is said to be
connected
iff
x, y X
x y
x<y or y>x
( x, y )
comparable
Hasse diagrams representing partial orderings in
which
indicates
5-30
Fuzzy partial ordering
- reflexive , antisymmetric , and transitive under
some form of transitivity.
any fuzzy partial ordering can be resolved into a
series of crisp partial ordering .
i.e. taking a series of cut that produce increasing
levels of refinement
In a fuzzy partial ordering , R
x X
R[ x ]
: dominating class
R[ x ] ( y ) R ( x, y )
R[ x ]
: dominated class
R[ x ] ( y ) R ( y , x )
5-31
x undominated iff R(x,y) = 0
y x
A X
y x
is a fuzzy set
U ( R, A) I R[ x ]
xA
2,
U ( R, A)( x)
y
x U ( R, A)
>0
R(x,y) > 0 ,
support [ U(R,A) ]
Example 5.13
a b c d
Fuzzy partial ordering
R:
e
a
b
c
d
e
1 0.7 0 1 0.7
0
1 0 0.9 0
0.5 0.7 1 1 0.8
0
0 0 1
0
0 0.1 0 0.9 1
2. d : undominated , C : undominating
3. For A = {a,b} , U(R,A) = the intersection of
The dominating classes of a and b =
4, LUB(A) =b
0.7 0.9
b
d
5-32
5. Crisp ordering captured by the fuzzy ordering
e.g. 0.5
1
0
R 1
0
0
1
1
1
0
0
0
0
1
0
0
1 1
1 0
1 1
1 0
1 1
# is 2 3 1 5 3
The ordering become weaken with the increasing
5-33
Fuzzy preordering reflexive and transitive
Fuzzy weak ordering
i, an ordering satisfying the proportion of a fuzzy
total ordering except antisymmetry.
ii, a fuzzy preordering in which
x y
, either
R(x,y)>0 or R(y,x)>0
Fuzzy strict ordering
Antireflexive
Antisymmetric
Transitive
5.8. Morphisms
Crisp homomorphism h from (X,R) to (Y,Q)
Where R(X,X), Q(Y,Y)binary relations
( x1 , x2 ) R ( h( x1 ), h( x2 )) Q
Fuzzy homomorphism h
If R(X,X), Q(Y,Y)Fuzzy binary relations
And
R ( x1 , x2 ) Q[h( x1 ), h( x2 )]
( h( x1 ), h( x2 )) Q
which
5-34
Crisp strong homomorphism h
If
( x1 , x2 ) R (h( x1 ), h( x2 )) Q
And
( y1 , y 2 ) Q (h 1 ( y1 ), h 1 ( y 2 )) R
h 1 ( y )
Xs
Fuzzy strong homomorphism h
H imposes a partition
on X
Let
A {a1 , a 2 , , a n }
B {b1 , b2 , , bn } h
R,Qfuzzy relations
h : strong homomorphism
iff
max ( R( ai , b j )) Q( y1 , y 2 )
i, j
y1 h( ai )ai A
y 2 h(b j )b j B
where
contains a set of
5-35
Example 5.14
R(X,X)
0
0
R
1
0 .5
0
0
0 .6
0
0 .9
0
0
0
0
0 .5
Q(Y,Y)
0.5
Q 1
1
0.9
0
0.9
0
0.9
0
But
i,e,
Q ( , ) 0.9 R ( d , c ) 0
( , ) Q
while
(d , c) R
where
h ( d ) , h (c )
5-36
R(X,X)
0.8
0
0
0
0 .4
0.5
0
0.5
0
0
0
0.3
0
0
0
0 .7
0
0
1
0
0
0
0 .9
0
1
0
0
0
0.5
0
0.8
Q(Y,Y)
0.7
0.4
0
0.8
0
0.9
0
1
5-37
Q represents a simplification of R
Isomorphism : (congruence)
h:1-1, onto
X Y
Endomorphism : (subgraph)
h:XY, Y X
Automorphism :
Isomorphism and End Endomorphism
i.e.m X=Y nad R=Q
5-38
5.9 SUP-i Compositions of Fuzzy Relations
Generalize max-min Composition
i : t-norm
sup : t-conorm
P(X,Y), Q(Y,Z)fuzzy relations
i
P o Q( X Z )
:sup-i composition
[ P o Q]( X , Z ) sup i[ P ( x, y ), Q ( y , z )]
yY
Properties
1.
( P o Q) o R P o( Q o R)
2.
P o ( Q j ) ( P o Q j )
3.
P o ( Q j ) ( P o Q j )
4.
( Pj ) o Q ( Pj o Q)
5.
( Pj ) o Q ( Pj o Q)
6.
( P o Q ) 1 Q 1 o R 1
5-39
3. Show Eq.(5.16), i.e.,
Where
P( X , Y )
and
P o( I Q j ) I ( P oQ j ),
jJ
Q(Y , Z )
jJ
P
o
Q ( x, z ) sup i P ( x, y ), Q( y, z )
yY
P o( I Q j ) ( x, z ) sup i P ( x, y ), I Q j ( y, z )
jJ
jJ
yY
Let
Q I Qj
jJ
Q Q1 , Q Q2 ,L , Q Q J
i.e.,
( y, z ), Q( y, z ) Q1 ( y, z ),L , Q( y, z ) Q J ( y, z )
i is monotonically increasing
i[ P ( x, y ), I Q j ( y , z )] i[ P( x, y ), Q1 ( y , z )]
jJ
...........
i[ P ( x, y ), Q ( y , z )] i[ P( x, y ), Q ( y, z )]
IjJ j
J
( x, y )
i[ P ( x, y ), ( I Q j )( y, z )] I i[ P( x, y ), Q j ( y, z )]
jJ
jJ
jJ
yY
yY jJ
I sup i[ P( x, y ), Q j ( y, z )], ( x, y ), ( y, z )
jJ yY
P o( I Q j ) ( x, z )
jJ
i,e.,
i
( P oQ j ) ( x, z ), ( x, z )
jJ
P o( I Q j ) I ( P oQ j )
jJ
jJ
5-40
Sup-i composition monotonically increases
i
i.e.,
Q1 oP Q2 oP (5.21)
Identity of
E oP P oE P
Relation R on
iff
Q1 Q2
1, x y
E ( x, y )
0, x y
i.e.,
if
: i-transitive
R ( x, z ) i R ( x, y ), R ( y, z ) , x, y, z X
R oR R
i-transitive closure
RT (i )
RT ( i ) U R ( n )
n 1
R ( n ) R oR ( n 1)
, where
5-41
By (5.15) (5.17)
proof:
i,
RT (i ) oRT (i )
i
U R o
n 1
(n)
k 2
k 1
(m)
R U
U
m 1
n 1
m 1
(n )
oR U R ( n m )
n , m 1
(m)
U R ( k ) U R ( k ) RT (i )
i.e.,
RT (i )
: i-transitive (Q R
T (i )
oRT (i ) RT ( i )
RS
R (2) R oR S oS S
If
i-transitive
(n)
mathematical
induction
S,
i
R ( n 1) R oR ( n ) S oS S
R ( k ) S , k
RT ( i ) U R ( k ) S
k 1
i.e.,
RT (i )
: smallest
X n
R ( m ) R m 1
RT ( i ) R ( n 1) n
m
n 1
R
5-42
proof : i, Q R : reflexive,
i
E R, R E oR R oR R (2)
(By repetition)
R ( n 1) R ( n )
ii, show
R ( n 1) R ( n )
proof: If
If
x y, R ( n 1) ( x, x ) 1
reflexive
x y,
Extension of definition
i
R ( n ) ( x, y ) sup i R( x, z1 ), R( z1 , z 2 ),L , R( z n 1 , y )
Z1 ,L , Z n1
Q X n
X Z 0 , Z1 ,L , Z n y
contains
Z r Z s (r s )
i R ( x, z1 ),L , R ( zr 1 , zr ), L , R ( zs , z s 1 ),L , R ( zn 1 , y ) R ( k ) ( x, y ), (k n 1)
x, y X , R ( n ) ( x, y ) R ( n 1) ( x, y ),
R ( n ) R ( n 1) ( B )
RT ( i ) R ( n 1)
R ( n ) R ( n 1) ( A, B )
5-43
5.10 INF- w Compositions of Fuzzy Relations
i
a b b
a b 1
w operation:
i
wi ( a, b) sup x [0,1] | i( a, x) b
where
a, b [0,1]
, i : continuous t-norm
Theorem 5.3
1,
i ( a, b) d
iff
2,
wi ( wi (a, b), b) a
3,
4,
a b, wi (a, d ) wi (b, d )
wi (a, b) b
wi (d , a ) wi (d , b)
---- i
--- ii
5,
6,
wi (inf a j , b) sup wi ( a j , b)
7,
wi (sup a j , b) inf wi (a j , b)
8,
9,
, if then)
10,
5-44
proof: (1) i , If
i ( a, b) d , b x | i ( a, x ) d
( )
ii, If
b wi (a, d )
i: continuous monotone
i: monotone increasing
()
By (1)<=
bi (a, b) abid(a, w
i ( a, d )) i ( a,sup x | i ( a, x ) d ) sup i ( a, x ) | i ( a, x ) d d
(3)
Q i ( a, xa) wbi (b, dd )
i(b, i( a, x)) d
By (1)=>
Associativity
communitation
wi (i ( a, b), d )
wi (a, wi (b, d )) sup x | i (a, x ) wi (b, d ) sup x | x wi (i( a, b), d ) wi (i( a, b), d )
(7) Let
By(4)
S sup a j
j
---(B)
a j s, j
wi ( s, b) wi (a j , b), j
wi ( s, b) inf wi (a j , b)
j
---- (C)
Q inf wi (a j , b) wi (a j0 , b), j0 J
By(1) j
i (a j0 ,inf wi (a j , b)) b, j0
j
j0
By(1)
wi ( s, b) inf wi (a j , b)
j
By(B)(C)(D)
--- (D)
wi (sup a j , b) wi ( s, b) inf wi ( a j , b)
j
5-45
(2)Show
wi ( wi (a, b), b) a
3.10
imin (a, b) i (a, b) min( a, b)
i, If a>b
wi ( a, b) Sup x [0,1] | i (a, x ) b Sup x [0,1] | min(a, x ) b b
i ( wi ( a, b), a )
i (b, a )
i (b,1)
b
By Axiom i2 wi (a, b) b
By Axiom i2 ( a 1 )
By Axiom i1
i ( wi (a, b), a ) b
i, If
ab
By Axiom i2 wi (a, b) 1
By Axiom i2
By Axiom i1
i ( wi (a, b), a ) b
i ( a, b) d
iff
ab
ii, W (d , a) W (d , b)
i
Wi (a, d ) sup{x | i( a, x) d }
---- (A)
---- (B)
a, if
d ab
ad b
abd
ii, see i
5. show
Proof : if
if
A, if
a b Wi (a, b) b
a b Wi (a, b) 1
a b i (Wi (a, b),Wi (b, d )) i (b, Wi (b, d ))
=>
Wi (a, d ) d
=>
=>
B, if
=>
Wi (b, d ) d
Wi (a, d ) d
Wi (b, d ) d
Wi (a, d ) 1
Wi (b, d ) 1
Wi (a, d ) 1
10. show
i (a, Wi (a, b) b
Proof :
a b Wi (a, b) b
a b Wi (a, b) 1
A, if
ab
B, if
ab
inf W composition
i
Wi
inf
( P oQ)( x, z ) y Y Wi ( P ( x, y ), Q( y, z ))
Theorem 5.4 :
Wi
Wi
(1)( P oQ R ) (Q P 1 o R ) ( P (Q o R 1 ) 1 )
Wi
Wi
Wi
Theorem 5.5 :
Wi
Wi
( Pj ) oQ ( Pj oQ)
j
Wi
Wi
( Pj ) oQ) ( Pj oQ)
j
Wi
Wi
P o( Q j ) ( Pj oQ j )
j
Wi
Wi
P o( Q j ) ( Pj oQ j )
j
Theorem 5.6 : if
=>
Q1 Q2
Wi
Wi
Wi
P oQ1 P oQ2
Wi
Q1 o R Q2 o R
Proof :
=>
Q1 Q2
Q1 Q2 Q1
Q1 Q2 Q2
( P oQ ) ( P oQ ) P o(Q Q ) P oQ
Wi
Wi
Wi
=>
Wi
Wi
Wi
P oQ1 P oQ2
(Q o R) (Q
Wi
Wi
Wi
Wi
Wi
oR
Theorem 5.7 :
1.
P 1 o( P oQ) Q
2.
R P o( P 1 oR)
3.
P ( P oQ) oQ 1
4.
R ( R oQ 1 ) oQ
Wi
Wi
Wi
Wi
Wi
Wi
Proof :
(1) Q P oQ ( P
Wi
1 1
Wi
) oQ
---- (A)
Q (5.26) (5.25)
i.e.,
let
r r Wi r
ri r r
(Q P 1 o R) ( P oQ R )
r
P oQ Q
Wi
Wi
o R ) (Q1 Q2 ) o R Q2 o R
=> Q o R Q
r
P 1 P
r
QR
( A) P 1 ( P oQ) Q
(2)
Q P 1 oR P 1 oR
Let
r
P 1 P
---- (B)
r
QR
Wi
i
r
P 1 oR R
( B) R P o( P 1 oR )
(3) by (5.33) , [ P
Wi
( P oQ) 1 oP Q 1
Let
Wi
o( P oQ)]1 Q 1
---- (C)
Wi
r
r
r
P oQ P, P Q, Q 1 R
Wi
Wi
(C ) P ( P oQ ) 1 oQ 1