Vagueness

• The Oxford Companion to Philosophy (1995):

“Words like smart, tall, and fat are vague since in most contexts of
use there is no bright line separating them from not smart, not tall,
and not fat respectively …”

1
07 December 2009
 Vagueness
• Imprecision vs. Uncertainty:

The bottle is about half-full.

vs.

It is likely to a degree of 0.5 that the bottle is full.

2
07 December 2009
 Fuzzy Sets
• Zadeh, L.A. (1965). Fuzzy Sets
Journal of Information and Control

3
07 December 2009
Fuzzy Sets

4
07 December 2009
 Fuzzy Set Definition
A fuzzy set is defined by a membership function that
maps elements of a given domain (a crisp set) into values
in [0, 1].

µ : U→ [0, 1] 1
A
young
µA↔ A
0.5

0
20 30 40 Age

5
07 December 2009
 Fuzzy Set Representation
• Discrete domain:
high-dice score: {1:0, 2:0, 3:0.2, 4:0.5, 5:0.9, 6:1}
• Continuous domain:
A(u) = 1 for u∈[0, 20]
A(u) = (40 - u)/20 for u∈[20, 40] 1
A(u) = 0 for u∈[40, 120]
0.5

0
20 30 40 Age

6
07 December 2009
 Fuzzy Set Representation
• α-cuts:
Aα = {u | A(u)≥α}
Aα+ = {u | A(u)>α} strongα-cut

1
0.5
A = [0, 30]
0.5

0
20 30 40 Age
7
07 December 2009
 Fuzzy Set Representation
• α-cuts:
α
A = {u | A(u)≥α}
Aα+ = {u | A(u)>α} strongα-cut
A(u) = sup {α | u∈ A }

1
0.5
A = [0, 30]
0.5

0
20 30 40 Age
8
07 December 2009
 Fuzzy Set Representation
• Support: 0+
supp(A) = {u | A(u) > 0} = A

• Core:
core(A) = {u | A(u) = 1} = A 1

• Height:
h(A) = supU A(u)

9
07 December 2009
 Fuzzy Set Representation
• Normal fuzzy set: h(A) = 1
• Sub-normal fuzzy set: h(A) < 1

10
07 December 2009
 Membership Degrees
• Subjective definition

11
07 December 2009
 Membership Degrees
• Voting model:
Each voter has a subset of U as his/her own crisp definition of
the concept that A represents.

A(u) is the proportion of voters whose crisp definitions include u.

A defines a probability distribution on the power set of U across

the voters.

12
07 December 2009
 Membership Degrees
• Voting model:

P1 P2 P3 P4 P5 P6 P7 P8 P9 P10
1
2

3 x x
4 x x x x x
5 x x x x x x x x x
6 x x x x x x x x x x

13
07 December 2009
Fuzzy Subset Relations

A ⊆ B iff A(u) ≤ B(u) for every u∈U

14
07 December 2009
Fuzzy Subset Relations

A ⊆ B iff A(u) ≤ B(u) for every u∈U

A is more specific than B

15
07 December 2009
Fuzzy Subset Relations

A ⊆ B iff A(u) ≤ B(u) for every u∈U

A is more specific than B

“X is A” entails “X is B”

16
07 December 2009
 Fuzzy Set Operations
• Standard definitions:

Complement: A(u) = 1− A(u)

Intersection: (A∩B)(u) = min[A(u), B(u)]

Union: (A∪B)(u) = max[A(u), B(u)]

17
07 December 2009
 Fuzzy Set Operations
• Example:

not young = young

not old = old

middle-age = not young∩not old

old =¬young

18
07 December 2009
 Fuzzy Numbers
• A fuzzy number A is a fuzzy set on R:
A must be a normal fuzzy set
Aα must be a closed interval for everyα∈(0, 1]
supp(A) = 0+A must be bounded

19
07 December 2009
Basic Types of Fuzzy Numbers

1 1

0 0

1 1

0 0

20
07 December 2009
Basic Types of Fuzzy Numbers

1 1

0 0

21
07 December 2009
 Operations of Fuzzy Numbers
• Extension principle for fuzzy sets:
f: U1 ×...×Un → V

induces
~ ~ ~
g: U1×...×Un→ V

22
07 December 2009
 Operations of Fuzzy Numbers
• Extension principle for fuzzy sets:
f: U1 ×...×Un → V

induces
~ ~ ~
g: U1×...×Un→ V

23
07 December 2009
 Operations of Fuzzy Numbers
• EP-based operations:
(A + B)(z) = sup{(x,y) | z = x+y}min{A(x),B(y)}

(A − B)(z) = sup{(x,y) | z = x-y}min{A(x),B(y)}

(A * B)(z) = sup{(x,y) | z = x*y}min{A(x),B(y)}

(A / B)(z) = sup{(x,y) | z = x/y}min{A(x),B(y)}

24
07 December 2009
 Operations of Fuzzy Numbers
• EP-based operations:

about 2 about 3 more or less 6 = about 2.about 3

1

2 3 6 +
∞

25
07 December 2009
 Operations of Fuzzy Numbers
• Discrete domains:

A = {x
i
: A(x
i
)} Bi = {y
i
: B(y )}

26
07 December 2009
 Operations of Fuzzy Numbers
• Interval-based operations:

27
07 December 2009
 Operations of Fuzzy Numbers
• Arithmetic operations on intervals:
[a, b] °[d, e] = {f °g | a ≤ f≤ b, d ≤ g ≤ e}

28
07 December 2009
 Operations of Fuzzy Numbers
• Arithmetic operations on intervals:
[a, b] °[d, e] = {f° g | a ≤ f≤ b, d ≤ g≤ e}

[a, b] + [d, e] = [a + d, b + e]

[a, b]− [d, e] = [a− e, b− d]

[a, b]*[d, e] = [min(ad, ae, bd, be), max(ad, ae, bd, be)]

[a, b]/[d, e] = [a, b]*[1/e, 1/d] 0∉[d, e]

29
07 December 2009
Operations of Fuzzy Numbers

about 2 about 3
1
about 2 + about 3 = ?

about 2× about 3 = ?
0

2 3 +∞

30
07 December 2009
 Operations of Fuzzy Numbers
• Discrete domains:

A = {x
i : A(x
i
)} Bi = {yi : B(y )}

31
07 December 2009
 Possibility Theory
• Possibility vs. Probability
• Possibility and Necessity

32
07 December 2009
 Possibility Theory
• Zadeh, L.A. (1978). Fuzzy Sets as a Basis for a Theory
of Possibility
Journal of Fuzzy Sets and Systems

33
07 December 2009
 Possibility Theory
• Membership degree = possibility degree

34
07 December 2009
 Possibility Theory
• Axioms:

0 ≤ Pos(A)≤ 1
Pos(Ω) = 1 Pos(∅) = 0
Pos(A ∪ B) = max[Pos(A), Pos(B)]
Nec(A) = 1 – Pos(A)

35
07 December 2009
 Possibility Theory
• Derived properties:

Nec(Ω) = 1 Nec(∅) = 0

Nec(A ∩ B) = min[Nec(A), Nec(B)]

max[Pos(A), Pos(A)] = 1
min[Nec(A), Nec(A)] = 0
Pos(A) + Pos(A) ≥ 1
Nec(A) + Nec(A) ≤ 1
Nec(A) ≤ Pos(A)

36
07 December 2009
Possibility Theory

37
07 December 2009
 Fuzzy Relations
• Crisp relation:
R(U1, ..., U n ) ⊆ U1 × ...×Un

R(u1, ..., u n ) = 1 iff (u1 , ..., u n )∈ R or = 0 otherwise

38
07 December 2009
 Fuzzy Relations
• Crisp relation:
R(U1 , ..., U n ) ⊆ U1 × ...×Un
R(u1, ..., un) = 1 iff (u1 , ..., un)∈ R or = 0 otherwise

• Fuzzy relation is a fuzzy set on U1 × ...×U n

39
07 December 2009
 Fuzzy Relations
• Fuzzy relation:
U1 = {New York, Paris}, U2 = {Beijing, New York, London}

R = very far
NY Paris
Beijing 1 .9
NY 0 .7
London .6 .3

R = {(NY, Beijing): 1, ...}

40
07 December 2009
 Multivalued Logic
• Truth values are in [0, 1]

• Lukasiewicz:
¬a = 1 − a
a ∧ b = min(a, b)
a ∨ b = max(a, b)
a ⇒ b = min(1, 1− a + b)

41
07 December 2009
Fuzzy Logic
if x is A then y is B
x is A*
------------------------
y is B*

42
07 December 2009
 Fuzzy Logic
• View a fuzzy rule as a fuzzy relation

if x is A then y is B R(u, v)≡ A(u) ⇒ B(v)

x is A* A*(u)
------------------------ ----------------------------------------
y is B* B*(v)

43
07 December 2009
 Fuzzy Logic
• Measure similarity of A and A*

if x is A then y is B
x is A*
------------------------
y is B*

B* = B +∆(A/A*)

44
07 December 2009
 Fuzzy Controller
• As special expert systems

• When difficult to construct mathematical models

• When acquired models are expensive to use

45
07 December 2009
 Fuzzy Controller
IF the temperature is very high
AND the pressure is slightly low
THEN the heat change should be sligthly negative

46
07 December 2009
 Fuzzy Controller
FUZZY
actions Defuzzification CONTROLLER
model

Controlled Fuzzy inference Fuzzy rule

process engine base

Fuzzification
model
conditions

47
07 December 2009
 Fuzzification

x0

48
07 December 2009
 Defuzzification
• Center of Area:

x = (∑A(z).z)/ ∑A(z)

49
07 December 2009
 Defuzzification
• Center of Maxima:

M = {z | A(z) = h(A)}

x = (min M + max M)/2

50
07 December 2009
 Defuzzification
• Mean of Maxima:

M = {z | A(z) = h(A)}

x =∑z/|M|

51
07 December 2009
 Exercises
• In Klir’s FSFL: 1.9, 1.10, 2.11, 4.5, 5.1 (a)-(b), 8.6, 12.1.

52
07 December 2009