Fuzzy Sets, Logic and Systems & Applications

Week 2: Lectures 06 - 10

Membership Functions
There are some commonly employed membership functions (MFs) in fuzzy theory, as
listed below.

1. Triangular Membership Function:

A triangular membership can be defined by three parameters [a, b, c] with (a < b <
c). These parameters determine the coordinates of three corners of a triangular
membership function. The membership function is defined as follows:



 0 x≤a



 x−a a≤x≤b


b−a
triangle(x; a, b, c) =
c−x
b≤x≤c




 c−b



0 c≤x

An alternate expression for the triangular function using min and max can be
written as:
   
x−a c−x
triangle(x; a, b, c) = max min , ,0
b−a c−b

2. Trapezoidal Membership Function:

A trapezoidal MF is specified by four parameters [a, b, c, d] with a < b ≤ c < d as
follows:


0 x≤a



x−a
a≤x≤b



 b−a



trapezoid(x; a, b, c, d) = 1 b≤x≤c



d−x




 d−c
c≤x≤d



d≤x

0

1
 An alternate expression for the trapezoidal function using min and max can be
written as:
   
x−a d−x
trapezoid(x; a, b, c, d) = max min , 1, ,0
b−a d−c
3. Gaussian Membership Function:
A Gaussian MF is specified by two parameters [c, σ] as:

1 x − c !2
−
gaussian(x; c, σ) = e 2 σ

where c represents the MF’s center and σ determines the MF’s width.
4. Generalized Bell-shaped Membership Function:
A generalized bell-shaped MF (or bell MF) is specified by three parameters [a, b, c]
as:

1
bell(x; a, b, c) =
x−c 2b
1+ a

Here:

• a defines the width of the membership function i.e., larger value creates a
wider membership function.
• b defines the shape of the curve on either side of the central plateau i.e., larger
value creates a steeper transition.
• c defines the center of the membership function.

5. Sigmoidal Membership Function:

The sigmoidal membership function is given by:

1
Sig(x; a, c) =
1 + e−a(x−c)
Here,

• The magnitude of a controls the width of the transition area, and

• c defines the center of the transition area.

6. Left-right (L-R) Membership Function:

A left-right MF or L-R MF is specified by three parameters [α, β, c] as:

c−x


FL

α
x≤c
LR(x; c, α, β) =  
x−c
FR

β
x≥c

where FL (.) and FR (.) are monotonically decreasing functions, defined on [0, ∞)
with FL (0) = FR (0) = 1 and limx→∞ FL (.) = limx→∞ FR (.) = 0.

2
 • c defines the point where the value of the membership function is 1.
• α controls the width of the left region for x ≤ c.
• β controls the width of the right region for x ≥ c.

7. Pi-shaped (π) Membership Function:

There are two π membership functions. The first π membership function is defined
with two parameters [a, b]. The function has a membership value of 1 at point a
and a membership value of 0.5 at a − b and a + b, respectively. Unlike the function,
π function decreases towards zero asymptotically as we move away from point a.
The first π membership function is defined as:

1
π1 (x; a, b) =
x−a 2


1+ b

The another four-variable π membership function with four parameters is defined

as:

lw

 lp+lw−x
x < lp



π2 (x; lw, lp, rp, rw) = 1 lp ≤ x ≤ rp



 rw

x > rp
x−rp+rw

Parameters lw and rw define the feet of the membership function, and lp and rp
define its shoulders, respectively. This membership function was used to define
the fuzzy set “Good Stopping Accuracy” in the Sendai Subway Train Control
System.

8. Open Right Membership Function:

The open right membership function is defined by two parameters [α, β] as follows:



 0 x<α


x−α
OpenR(x; α, β) = β−α
α≤x≤β




1 x>β

9. Open Left Membership Function:

The open left membership function is defined by two parameters [α, β] as follows:



1 x<α


β−x
OpenL(x; α, β) = β−α
α≤x≤β




0 x>β

3
 10. S-shaped Membership Function:
A S-shaped membership function is defined by two parameters [a, b] as follows:



 0 x≤a


x−a 2

a+b


2 a≤x≤


b−a 2
f (x; a, b) =
x−b 2 a+b


1−2 ≤x≤b




 b−a 2



1 x≥b

The parameter a defines the feet of the membership function, and b defines its
shoulder.

Nomenclatures used in Fuzzy Set Theory

There are some commonly used nomenclatures in fuzzy set theory as listed below.

1. Support:
The support of a fuzzy set A is the set of all points x in the universe of discourse
X for any associated membership function such that:

µA (x) > 0 (1)

It can be represented mathematically as:

Support(A) = {x | µA (x) > 0} (2)

where x belongs to the universe of discourse X. The support of A consists of all

values of x where the membership function is greater than zero.
Additionally, this condition can be expressed as:

µA (x) > 0, ∀x ∈ (x1 , x2 ) in the universe of discourse X (3)

Outside this range:

µA (x) = 0, ∀x ∈
/ (x1 , x2 ) (4)

2. Core:
The core of fuzzy set A is the set of all points x in the universe of discourse X such
that:

µA (x) = 1 (5)

It can be represented mathematically as:

4
 Core(A) = {x | µA (x) = 1} (6)

For the given fuzzy set A with universe of discourse X, we can express this as:

µA (x) = 1, ∀x ∈ [x1 , x2 ] (7)

3. Crossover Points:
The crossover points of a fuzzy set A are the points x in the universe of discourse
X at which:

µA (x) = 0.5 (8)

It can be represented mathematically as:

Crossover(A) = {x | µA (x) = 0.5} (9)

For the given fuzzy set A with universe of discourse X, the crossover points are:

Crossover(A) = {x1 , x2 } (10)

4. Height of a Fuzzy Set:

• The height of a fuzzy set A is the maximum value of the membership function.
It is represented as:

hgt(A) = max{µA (x)} (11)

If hgt(A) < 1, then the fuzzy set A is called a Subnormal fuzzy set.
The hgt(A) is viewed as the degree of validity or credibility of information expressed
by the fuzzy set A.
For example, given fuzzy sets A, B, and C:

hgt(A) = 1, hgt(B) = 0.40, hgt(C) = 0.75 (12)

5. Normality of a Fuzzy Set:

A fuzzy set A is normal if its core is non-empty.
In other words, we can always find a point x in the universe of discourse X such
that:

µA (x) = 1 (13)

The core of the fuzzy set A is given by:

Core(A) = {x2 } (14)

Since the core of A is non-empty, the fuzzy set A is a normal fuzzy set.

5
6. Subnormal Fuzzy Sets:
A fuzzy set A is called normal if there exists at least one element x in the universe
of discourse X such that:

µA (x) = 1 (15)

A fuzzy set that does not satisfy this condition, meaning:

max µA (x) < 1 (16)

is called a subnormal fuzzy set.

A normal fuzzy set reaches a membership degree of 1 for at least one element in X.
A subnormal fuzzy set** never reaches a membership value of 1. The height of a
fuzzy set A is given by:

hgt(A) = max{µA (x)} (17)

If hgt(A) < 1, then A is a subnormal fuzzy set.

Subnormal fuzzy sets are useful in cases where uncertainty does not fully support
the membership of any element to the set with full confidence.
Normalization of a Fuzzy Set:
The fuzzy set A′ is the normalized version of a fuzzy set A, which is represented by:

A′ = Norm(A) (18)

The normalization is carried out as follows:

X µA (x)
A′ = Norm(A) = /x (19)
X
hgt(A)

where:

• A′ is the normalized fuzzy set.

• hgt(A) is the height of the fuzzy set A, defined as max{µA (x)}.
• The normalization ensures that the membership values are scaled in such a
way that the highest membership degree reaches 1.

7. Fuzzy Singleton:
A fuzzy set whose support is a single point in the universe of discourse X with
µA (x) = 1 is called a Fuzzy Singleton. In other words, we can say that:

“A fuzzy singleton has a core with only one element in its set.”

6
 Mathematically, it is represented as:

Support(A) = {x | µA (x) > 0} = {x1 } (20)

Core(A) = {x | µA (x) = 1} = {x1 } (21)

This means that a fuzzy singleton has exactly one element where the membership
function is equal to 1.

8. α-Cut of a Fuzzy Set:

The α-Cut or α-level set of a fuzzy set A is a crisp set defined as:

Aα = {x | µA (x) ≥ α} (22)

where:

• Aα represents the subset of elements in X where the membership function

µA (x) is at least α.
• The α-Cut transforms a fuzzy set into a crisp set by including only those
elements with membership values above the threshold α.
• This concept is useful in fuzzy logic applications, such as decision-making and
approximate reasoning.

9. Strong α-Cut of a Fuzzy Set:

The strong α-Cut or strong α-level set of a fuzzy set A is a crisp set defined as:

A′α = {x | µA (x) > α} (23)

where:

• A′α represents the subset of elements in X where the membership function

µA (x) is strictly greater than α.
• Unlike the standard α-Cut, which includes elements where µA (x) ≥ α, the
strong α-Cut excludes elements where µA (x) = α.
• This distinction is important in fuzzy logic applications where strict inclusion
criteria are required.

α-Cut and Strong α-Cut of a Fuzzy Set:

Using the notation for a level set, we can represent an α-Cut and Strong α-Cut for
a fuzzy set A as:

Aα = {x ∈ X | A(x) ≥ α} A′α = {x ∈ X | A(x) > α}

Core(A) = {x | µA (x) = 1} Support(A) = {x | µA (x) > 0}

Additionally, the relationships between the core and support of a fuzzy set can be
written as:

7
 Core(A) = A1 Support(A) = A′0

where:

• Aα represents the set of elements with membership values greater than or equal
to α.
• A′α represents the set of elements with membership values strictly greater than
α.
• Core(A) contains elements where membership is exactly 1.
• Support(A) contains elements where membership is positive.

10. Convexity of a Fuzzy Set:

A fuzzy set A is convex if and only if for any x1 , x2 ∈ X and any λ ∈ [0, 1], the
following condition holds:

µA (λx1 + (1 − λ)x2 ) ≥ min[µA (x1 ), µA (x2 )] (24)

where:

• Alternatively, A is convex if all its α-level sets are convex.

• A convex fuzzy set is described by a membership function whose values are:
– Strictly monotonically increasing.
– Strictly monotonically decreasing.
– Strictly monotonically increasing and then strictly monotonically decreas-
ing with increasing values for elements in the universe of discourse.
• For any elements x1 , x2 , x3 in a fuzzy set A with x1 < x2 < x3 , the condition
for convexity is defined as:

µA (x2 ) ≥ min[µA (x1 ), µA (x3 )] (25)

Properties of Convex Fuzzy Sets:

• A special property of convex fuzzy sets is that the intersection of two convex
fuzzy sets is also convex.
• If fuzzy sets A and B are convex, then the fuzzy set A ∩ B is also convex.

Comparison with Convexity of Crisp Sets and Functions:

• A crisp set C in Rn is convex if and only if for any two points x1 ∈ C and
x2 ∈ C, their convex combination λx1 +(1−λ)x2 is still in C, where 0 ≤ λ ≤ 1.
• The definition of convexity of a fuzzy set is not as strict as the common def-
inition of convexity for functions. For comparison, the definition of convexity
of a function f (x) is given as:

f (λx1 + (1 − λ)x2 ) ≥ λf (x1 ) + (1 − λ)f (x2 ) (26)

8
11. Cardinality of a Fuzzy Set:
In a crisp set, the cardinality of a set is a measure of the ”number of elements
of the set”. For example, the set:

A = {5, 10, 15, 20}

contains 4 elements, and therefore A has a cardinality of 4.

In contrast, for a continuous fuzzy set, the universe of discourse contains an infinite
number of elements.
Thus, the Cardinality of a Continuous Fuzzy Set is Infinite.

12. Fuzzy Number:

• A fuzzy number A is a fuzzy set in the real number space (R) that satisfies
the condition for normality and convexity.
• Most (non-composite) fuzzy sets used in the literature satisfy the condition for
normality and convexity, so fuzzy numbers are the most basic type of fuzzy
sets.

13. Bandwidth of Normal and Convex Fuzzy Sets:

For a normal and convex fuzzy set A, the bandwidth or width is defined as the
distance between the two unique crossover points.

Width(A) = |x1 − x2 |

Here,
µA (x1 ) = µA (x2 ) = 0.5

14. Fuzzy Symmetry:

A fuzzy set A is symmetric if its membership function is symmetric around a certain
point x = c, i.e., it satisfies the condition given below for the universe of discourse
X:
µA (c + x) = µA (c − x) ∀x ∈ X

15. Open Left/Open Right Fuzzy Set:

• A fuzzy set A is open left if

lim µA (x) = 1 and lim µA (x) = 0

x→−∞ x→+∞

• A fuzzy set B is open right if

lim µB (x) = 0 and lim µB (x) = 1

x→−∞ x→+∞

9
Set Theoretic Operations on Fuzzy Sets
• Set Theory Notations:
For two sets A and B, which consist of the collection of some elements in the
universe of discourse X, the set theory notations are defined as follows:

x∈X → x belongs to X
x∈A → x belongs to A
x∈
/A → x does not belong to A
A⊂B → A is fully contained in B
A⊆B → A is contained in B and equivalent to B
A=B → A is equivalent to B, i.e., A ⊆ B and B ⊆ A
ϕ → Null/empty set i.e. the set contains no element

• Equality of Fuzzy Sets:

Let A and B be two fuzzy sets. They are equal if and only if the following conditions
are satisfied:

µA (x) = µB (x), ∀x ∈ X
Here, X is the universe of discourse, and µA (x) and µB (x) are membership degrees
for an element x in the fuzzy sets A and B, respectively.
• Containment (Subset) of a Fuzzy Set:
Fuzzy set B contains another fuzzy set A if and only if the following condition
satisfies:

µA (x) ≤ µB (x), ∀x ∈ X
It can be denoted by:

A⊆B
• Complement of a Classical Set:
The complement of a set A, denoted by Ā, represents all elements in the universe
of discourse X that do not belong to set A.
It can be represented as:

Ā = {x | x ∈
/ A and x ∈ X}
• Complement of a Fuzzy Set:
The complement of a fuzzy set A is a fuzzy set Ā in the universe of discourse X
whose membership function is defined as:

µĀ (x) = 1 − µA (x) ∀x ∈ X

The complement of a fuzzy set should contain all elements along with their mem-
bership values from the universe of discourse.

10