Fuzzy Sets and Systems

Disclaimer: These notes have not been subjected to the usual scrutiny reserved for formal
publications. They may be distributed outside this class only with the permission of the
Instructor.

Classical Set and Fuzzy Set

The widely used term set will be called Ordinary/Crisp/Classical set in our discussion. This
will be done to distinguish between set and fuzzy set.
Set: Set is a well defined collection of objects. By the term well defined collection, we mean,
whether any object is included in this collection or not is certain. Any object included in this
collection is called a member of the set and any object not included in this collection, then
that object is not a member of this set.
Examples:
i) Suppose V is a collection of vowels in English alphabets. Here, the collection is well
defined, as which alphabet is vowel and which is not vowel, we can specify. So, the
collection, V is a set.

Non Examples:
i) Suppose B is a collection of tall buildings in Dhaka city. Here, the collection is not
well defined, as which building is tall and which is not tall, we can not specify. So, B
is not a set.

Characteristic function of a set: A set S can be described by a function, 𝒳𝑠 : 𝑋 → {0, 1},

usually called a characteristic function of S, that declares which elements of X are members
of the set S and which are not. Set S is defined by it’s characteristic function, 𝒳𝑠 , as follows:
1 𝑓𝑜𝑟 𝑥 ∈ 𝑆
𝒳𝑠 (𝑥) = {
0 𝑓𝑜𝑟 𝑥 ∈ 𝑆
Examples:
i) Suppose V is the set of vowels in English alphabets, then
𝒳𝑣 (𝑎) = 1 as 𝑎 ∈ 𝑉
𝒳𝑣 (𝑏) = 0 as 𝑏 ∈ 𝑉
𝒳𝑣 (𝑒) = 1 as 𝑒 ∈ 𝑉
𝒳𝑣 (𝑖) = 1 as 𝑖 ∈ 𝑉
𝒳𝑣 (𝑜) = 1 as 𝑜 ∈ 𝑉
𝒳𝑣 (𝑢) = 1 as 𝑢 ∈ 𝑉

Empty set: A set ∅ is called empty set specified by the function, 𝒳∅ : 𝑋 → {0, 1}, where 𝑋 is
universe and 𝒳∅ (𝑥) = 0; ∀ 𝑥 ∈ 𝑋.

Fuzzy set: As defined, the characteristic function of a crisp set, assigns a value either 1 or 0 to
each individual of the universe 𝑋, there by discriminating between members and non members of
the crisp set under consideration. This function can be generalized such that the values assigned
1 Fuzzy Sets and Systems
to the elements of the universe 𝑋, fall within a specified range and indicate the membership
grade of these elements in the set under consideration. Larger values denote higher degrees of
membership. This set specified by such grading is termed as Fuzzy set. LUTFI ABU ZADEH is
the first man who introduce the notion of fuzzy sets.

Definition: A function 𝜇𝐴 : 𝑋 → [0, 1]is called a membership function and the set 𝐴 defined by
it, a fuzzy set. The universe 𝑋 is called referential set, [0, 1] is called a valuation set and 𝜇(𝑥)
is called grade of 𝑥 for 𝑥 ∈ 𝑋.

Note-1: The most commonly used valuation set is the unit interval [0, 1]. In this case, each
membership function maps elements of given referential set 𝑋, which is always a crisp set, into
real numbers in [0, 1].

The membership function 𝜇𝐴 of a fuzzy set 𝐴 can be denoted by 𝐴 itself and has of course the
same form:

𝐴: 𝑋 → [0, 1].

Note-2: Representing [0, 1] by I, we denote the set of all fuzzy sets on 𝑋 by 𝐼 𝑋 .

Examples:

i) Suppose 𝐶(3, 5) ∈ ℝ2 . Then the set, 𝑆 = {𝑃 ∈ ℝ2 : 𝐶𝑃 ≤ 2}is a classical set but the
set,

𝑇 = {𝑃 ∈ ℝ2 : 𝑃 𝑖𝑠 𝑎 𝑛𝑒𝑎𝑟 𝑝𝑜𝑖𝑛𝑡 𝑜𝑓 𝐶} is a fuzzy set.

ii) Suppose 𝑌: 𝑋 → 𝐼 defines a fuzzy set “Youth” such that,

0 𝑤ℎ𝑒𝑛 𝑥 ≤ 10
𝑥 − 10
𝑤ℎ𝑒𝑛 10 < 𝑥 < 15
5
𝑌(𝑥) = 1 𝑤ℎ𝑒𝑛 15 ≤ 𝑥 ≤ 35
50 − 𝑥
𝑤ℎ𝑒𝑛 35 < 𝑥 < 50
15
{ 0 𝑤ℎ𝑒𝑛 𝑥 ≥ 50

Where 𝑋 is the set of ages of human beings.

2 Fuzzy Sets and Systems

 1

0.8

0.6

0.4

0.2

15 35 60 80 100
Figure: Fuzzy Set “Youth”
iii) Suppose 𝑀: 𝑋 → 𝐼 defines a fuzzy set of “Middle aged people” where 𝑋 is the set of
ages of human beings such that

0 𝑤ℎ𝑒𝑛 𝑥 ≤ 20
𝑥 − 20
𝑤ℎ𝑒𝑛 20 < 𝑥 < 35
15
𝑀(𝑥) = 1 𝑤ℎ𝑒𝑛 35 ≤ 𝑥 ≤ 45
60 − 𝑥
𝑤ℎ𝑒𝑛 45 < 𝑥 < 60
15
{ 0 𝑤ℎ𝑒𝑛 𝑥 ≥ 60

0.8

0.6

0.4

0.2
20 40 60 80 100

Figure: Fuzzy Set “Middle aged people”

3 Fuzzy Sets and Systems

 iv) Suppose 𝑂: 𝑋 → 𝐼 defines a fuzzy set of “Old people” where 𝑋 is the set of ages of
human beings such that

0 𝑤ℎ𝑒𝑛 𝑥 < 45
𝑥 − 45
𝑂(𝑥) = { 𝑤ℎ𝑒𝑛 45 ≤ 𝑥 ≤ 60
15
1 𝑤ℎ𝑒𝑛 𝑥 > 60

0.8

0.6

0.4

0.2

0
20 40 60 80 100

Figure: Fuzzy Set “Old people”

Note: Observe that the referential set in each of the above cases is a subset of the real
numbers and so an infinite set. The referential set may be a discrete and finite set.

v) Suppose 𝑋 = {𝑎, 𝑏, 𝑐, 𝑑, 𝑒} and the fuzzy set 𝑓: 𝑋 → 𝐼 is defined by

𝑓(𝑎) = 0.8, 𝑓(𝑏) = 0.4, 𝑓(𝑐) = 0, 𝑓(𝑑) = 0.1, 𝑓(𝑒) = 1.

Note: This fuzzy set 𝑓 can also be written as

𝑓 = (0.8, 0.4, 0, 0.1, 1 )

Different ways to represent fuzzy sets:

Suppose X= {a, b, c, d} is a referential set and 𝑓 ∈ 𝐼 𝑥 then 𝑓 can be written in the following
ways:
1. 𝑓(𝑎) = 0.4, 𝑓(𝑏) = 1, 𝑓(𝑐) = 0.5, 𝑓(𝑑) = 𝑂.
0.4 1 0.5 0
2. 𝑓 = ( 𝑎 , 𝑏 , , 𝑑).
𝑐
3. 𝑓 = [0.4, 1, 0.5, 0].
4. 𝑓 = {(𝑎, 0.4) , (𝑏, 1) , (𝑐, 0.5), (𝑑, 0)}

4 Fuzzy Sets and Systems

𝜶 −cut: Suppose 𝑓: 𝑋 → 𝐼 defines a fuzzy set. For 𝛼 ∈ 𝐼, 𝑓 𝛼 is called 𝛼 −level set or 𝛼 −cut of
𝑓, where 𝑓 𝛼 = {𝑥 ∈ 𝑋: 𝑓(𝑥) ≥ 𝛼}. Clearly, 𝑓 𝛼 is a classical set and 𝑓 𝛼 ⊆ 𝑋.

That is, the 𝛼 −cut of a fuzzy set 𝑓 is the classical set 𝑓 𝛼 that contains all the elements of the
universal set 𝑋 whose membership grades in 𝑓 are greater than or equal to the specified value of
𝛼.

Examples:
i) Suppose 𝑋 = {𝑎, 𝑏, 𝑐, 𝑑, 𝑒} and the fuzzy set 𝑓: 𝑋 → 𝐼 is defined by
𝑓(𝑎) = 0.8, 𝑓(𝑏) = 0.4, 𝑓(𝑐) = 0, 𝑓(𝑑) = 0.1, 𝑓(𝑒) = 1.
Then 𝑓 0.4 = {𝑎, 𝑏, 𝑒}.
𝑓 0.5 = {𝑎, 𝑒}.
ii) In the example of “Middle aged people” 𝑀0.5 = [27.5, 52.5].

0.8

0.6

0.4

0.2
20 40 60 80 100

𝑀0.5

Figure: 𝛼 −cut (𝛼 = 0.5) of fuzzy set “Middle aged people”

Strong 𝜶 −cut: Suppose 𝑓: 𝑋 → 𝐼 defines a fuzzy set. For 𝛼 ∈ 𝐼, 𝑓 𝛼+ is called strong 𝛼 −level
set or strong 𝛼 −cut of 𝑓, where 𝑓 𝛼+ = {𝑥 ∈ 𝑋: 𝑓(𝑥) > 𝛼}. Clearly, 𝑓 𝛼+ is a classical set and
𝑓 𝛼+ ⊆ 𝑋.

That is, the strong 𝛼 −cut of a fuzzy set 𝑓 is the classical set 𝑓 𝛼+ that contains all the elements of
the universal set 𝑋 whose membership grades in 𝑓 are only greater than to the specified value of
𝛼.

Examples:

5 Fuzzy Sets and Systems

 i) Suppose 𝑋 = {𝑎, 𝑏, 𝑐, 𝑑, 𝑒} and the fuzzy set 𝑓: 𝑋 → 𝐼 is defined by
𝑓(𝑎) = 0.8, 𝑓(𝑏) = 0.4, 𝑓(𝑐) = 0, 𝑓(𝑑) = 0.1, 𝑓(𝑒) = 1.
Then 𝑓 0.4+ = {𝑎, 𝑒}.
𝑓 0.5+ = {𝑎, 𝑒}.
ii) In the example of “Middle aged people” 𝑀0.5+ = (27.5, 52.5).

0.8

0.6

0.4

0.2
20 40 60 80 100

𝑀0.5+

Figure: Strong 𝛼 −cut (𝛼 = 0.5) of fuzzy set “Middle aged people”

N.B. 𝛼 −cut is an interval but not always bounded, e.g. in the example of “Old people”, 𝛼 −cut
does not have upper bound.

Theorem: Suppose 𝑓: 𝑋 → 𝐼, then show that 𝛼 ≤ 𝛽 implies 𝑓 𝛼 ⊇ 𝑓 𝛽 and 𝑓 𝛼+ ⊇ 𝑓𝛽+ ;∀ 𝛼, 𝛽 ∈

[0, 1].

Proof: For any 𝑥 ∈ 𝑓 𝛽 , we have 𝑓(𝑥) ≥ 𝛽 ≥ 𝛼 and, hence 𝑓(𝑥) ≥ 𝛼. This means that 𝑥 ∈ 𝑓 𝛼 .
Therefore, 𝑓 𝛼 ⊇ 𝑓 𝛽 . Again, for any 𝑥 ∈ 𝑓𝛽+ , we find 𝑓(𝑥) > 𝛽 ≥ 𝛼. This implies that 𝑓(𝑥) > 𝛼
which means that 𝑥 ∈ 𝑓 𝛼+ . Thus 𝑓 𝛼+ ⊇ 𝑓𝛽+ .

Note: all 𝛼 −cuts and strong 𝛼 −cuts of a fuzzy set form a family of nested crisp set.

6 Fuzzy Sets and Systems

Support of fuzzy set: Suppose 𝑓: 𝑋 → 𝐼 is a fuzzy set, then support of 𝑓 is defined as follows:
Supp(𝑓) or 𝑆(𝑓) = {𝑥 ∈ 𝑋: 𝑓(𝑥) > 0} i.e. the support of a fuzzy set 𝑓 within a universal set 𝑋 is
the crisp set that contains all the elements of 𝑋 that have nonzero membership grades in 𝑓.
Clearly, the support of 𝑓 is exactly the same as the strong 𝛼 −cut of 𝑓 for 𝛼 = 0.

Examples:
i) Suppose 𝑋 = {𝑎, 𝑏, 𝑐, 𝑑, 𝑒} and the fuzzy set 𝑓: 𝑋 → 𝐼 is defined by
𝑓(𝑎) = 0.8, 𝑓(𝑏) = 0.4, 𝑓(𝑐) = 0, 𝑓(𝑑) = 0.1, 𝑓(𝑒) = 1.
Then 𝑆𝑢𝑝𝑝(𝑓) = {𝑎, 𝑏, 𝑑, 𝑒}.
ii) For previous examples of “Young People”, “Middle aged People” and “Old People”,
we find the supports as follows: 𝑆(𝑌) = [10, 50), 𝑆(𝑀) = (20, 60), 𝑆(𝑂) = [45, ∞).

Core of fuzzy set: Suppose 𝑓: 𝑋 → 𝐼 is a fuzzy set, then the Core of 𝑓 is defined as follows:

𝐶(𝑓) = {𝑥 ∈ 𝑋: 𝑓(𝑥) = 1}. The 1-cut, 𝑓 1 , is often called the core of 𝑓.

Examples:
i) Suppose 𝑋 = {𝑎, 𝑏, 𝑐, 𝑑, 𝑒} and the fuzzy set 𝑓: 𝑋 → 𝐼 is defined by
𝑓(𝑎) = 0.8, 𝑓(𝑏) = 0.4, 𝑓(𝑐) = 0, 𝑓(𝑑) = 0.1, 𝑓(𝑒) = 1.
Then 𝐶(𝑓) = {𝑒}.
ii) For previous examples, we find

𝐶(𝑌) = [15, 35]

𝐶(𝑀) = [35, 45]

𝐶(𝑂) = (60, ∞)

Height of fuzzy set: The height, ℎ(𝑓), of a fuzzy set 𝑓 is the largest membership grade obtained
by any element in the set. Formally, ℎ(𝑓) = sup𝑥∈𝑋 𝑓(𝑥).

Examples:
i) Suppose 𝑋 = {𝑎, 𝑏, 𝑐, 𝑑, 𝑒} and the fuzzy set 𝑓: 𝑋 → 𝐼 is defined by
𝑓(𝑎) = 0.8, 𝑓(𝑏) = 0.4, 𝑓(𝑐) = 0, 𝑓(𝑑) = 0.1, 𝑓(𝑒) = 0.5.
Then ℎ(𝑓) = 0.8.
ii) For previous examples, we have
ℎ(𝑌) = 1

ℎ(𝑀) = 1

ℎ(𝑂) = 1

Normal fuzzy set: A fuzzy set 𝑓 is called normal when ℎ(𝑓) = 1 equivalently, 𝑓 is called
normal if ∃ 𝑥 ∈ 𝑋, such that 𝑓(𝑥) = 1.
7 Fuzzy Sets and Systems
Examples:
i) For previous examples, all the fuzzy sets 𝑌, 𝑀, 𝑂 are normal.
ii) Suppose 𝑋 = {𝑎, 𝑏, 𝑐, 𝑑, 𝑒} and the fuzzy set 𝑓: 𝑋 → 𝐼 is defined by
𝑓(𝑎) = 0.8, 𝑓(𝑏) = 0.4, 𝑓(𝑐) = 0, 𝑓(𝑑) = 0.1, 𝑓(𝑒) = 0.5.
Clearly, the fuzzy set 𝑓 is not normal, here.

Subnormal fuzzy set: A fuzzy set 𝑓 is called Subnormal when ℎ(𝑓) < 1 equivalently, 𝑓 is
called Subnormal iff 𝑓(𝑥) < 1; ∀𝑥 ∈ 𝑋.

Examples:
i) For previous examples, none the fuzzy sets 𝑌, 𝑀, 𝑂 are subnormal.
ii) Suppose 𝑋 = {𝑎, 𝑏, 𝑐, 𝑑, 𝑒} and the fuzzy set 𝑓: 𝑋 → 𝐼 is defined by
𝑓(𝑎) = 0.8, 𝑓(𝑏) = 0.4, 𝑓(𝑐) = 0, 𝑓(𝑑) = 0.1, 𝑓(𝑒) = 0.5.
Clearly, the fuzzy set 𝑓 is subnormal, here.

Standard complement of fuzzy set: Suppose 𝑓: 𝑋 → 𝐼 is a fuzzy set, then the standard
complement of 𝑓 is another fuzzy set 𝑓:̅ 𝑋 → 𝐼, defined by 𝑓(̅ 𝑥) = 1 − 𝑓(𝑥).

Examples:

i) Suppose 𝑋 = {𝑎, 𝑏, 𝑐, 𝑑, 𝑒}, 𝑓 ≔ {0.3, 0.7, 0, 1, 0.5}, then 𝑓 ̅ = {0.7, 0.3, 1, 0, 0.5}.

Equilibrium point of fuzzy set: Suppose 𝑓: 𝑋 → 𝐼 is a fuzzy set, then 𝑥 ∈ 𝑋 is called an

̅
equilibrium point of 𝑓 if 𝑓(𝑥) = 𝑓 (𝑥).

Examples:

i) Suppose 𝑋 = {𝑎, 𝑏, 𝑐, 𝑑, 𝑒}, 𝑓 ≔ {0.3, 0.7, 0, 1, 0.5}, then 𝑓 ̅ = {0.7, 0.3, 1, 0, 0.5}.

Therefore, equilibrium point of 𝑓 is 𝑒.

Constant fuzzy set: A fuzzy set 𝑓: 𝑋 → 𝐼is called a Constant fuzzy set if
𝑓(𝑥) = 𝑘; ∀𝑥 𝜖 𝑋; where 𝑘 is a particular number in I. For convenience such a fuzzy set will be
denoted by k.

Convex set: A classical set 𝑆 is called a convex set iff ∀ 𝑥, 𝑦 ∈ 𝑆 and ∀ 𝜆 ∈ 𝐼;

𝜆𝑥 + (1 − 𝜆)𝑦 ∈ 𝑆.
i.e. A set is convex if, given any two points in the set, the line segment connecting
them lies entirely inside the set.

Examples:
i) The empty set, The singleton set, The complete space ℝ𝑛 , Line segments, a circle, a
cube.
8 Fuzzy Sets and Systems
 ii) The set of points that form the boundary of a circle is not a convex set.

iii)

Figure: Examples of convex and non-convex sets in ℝ2 .

Convex Fuzzy Set: An important property of fuzzy sets defined on ℝ𝑛 (for some 𝑛 ∈ ℕ) is their
convexity. This property is viewed as a generalization of the classical concept of convexity of
crisp sets. In order to make the generalized convexity consistent with the classical definition of
convexity, it is required that a-cuts of a convex fuzzy set be convex for all 𝛼 ∈ (0, 1]in the
classical sense (0 −cut is excluded here since it is always equal to ℝ𝑛 in this case and thus
includes −∞ to +∞).

Suppose 𝑓: 𝑋 → 𝐼 is a fuzzy set, then 𝑓 is called convex fuzzy set if

𝑓(𝜆𝑥 + (1 − 𝜆)𝑦) ≥ min(𝑓(𝑥), 𝑓(𝑦)).

Examples:
𝑥 − 1; 𝑥 ∈ [1, 2)
i) Suppose 𝜇: ℝ → 𝐼 is a fuzzy set such that 𝜇(𝑥) = { 3 − 𝑥; 𝑥 ∈ [2, 3]
0; 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

9 Fuzzy Sets and Systems

 𝜇(𝑥)

0 𝑥
1 2 3
Clearly, 𝜇 is a convex fuzzy set, here.

ii)

𝑓(𝑦)

𝑓(𝑍)

𝑓(𝑥)

0 𝑥 𝑍 𝑦

Here, 𝑍 = 𝜆𝑥 + (1 − 𝜆)𝑦 and 𝑓(𝜆𝑥 + (1 − 𝜆)𝑦) ≥ 𝑓(𝑥) which means that 𝑓 is a convex fuzzy
set.
iii)

10 Fuzzy Sets and Systems

iv)

11 Fuzzy Sets and Systems

Theorem: A fuzzy set 𝐴 on ℝ is convex iff 𝐴(𝜆𝑥 + (1 − 𝜆)𝑦) ≥ min(𝐴(𝑥), 𝐴(𝑦)) for all
𝑥1 , 𝑥2 ∈ ℝ and all 𝜆 ∈ [0, 1], where min denotes the minimum operator.

Theorem: Suppose 𝑓 ∈ 𝐼 𝑥 is a fuzzy set. Then 𝑓 is fuzzy convex iff,

∀𝛼 ∈ 𝐼 ; 𝑓 𝛼 is convex subset of X.

Proof: First suppose that,𝑓is a convex fuzzy set and 𝛼 ∈ 𝐼 then we have to prove that 𝑓 𝛼 is
convex subset of 𝑋. Since 𝑓 is a convex fuzzy set, so ∀𝑥, 𝑦 ∈ 𝑓 𝛼 and ∀𝜆 ∈ 𝐼;

12 Fuzzy Sets and Systems

𝑓(𝜆𝑥 + (1 − 𝜆)𝑦) ≥ 𝑓(𝑥) ∧ 𝑓(𝑦).
Since 𝑥, 𝑦 ∈ 𝑓 𝛼 ,so
𝑓(𝑥) ≥ 𝛼and𝑓(𝑦) ≥ 𝛼.
Therefore 𝑓(𝑥) ∧ 𝑓(𝑦) ≥ 𝛼.
So,𝑓(𝜆𝑥 + (1 − 𝜆)𝑦) ≥ 𝛼
𝜆𝑥 + (1 − 𝜆)𝑦 ∈ 𝑓 𝛼
Hence 𝑓 𝛼 is a convex subset of 𝑋.
Conversely, suppose ∀ 𝛼 ∈ 𝐼, 𝑓 𝛼 is a classical convex set. We have to show that, 𝑓 is a fuzzy
convex set.
Here 𝑓 𝛼 is a crisp set, So∀𝑥, 𝑦 ∈ 𝑓 𝛼 without loss of generality suppose, 𝑓(𝑥) ≤ 𝑓(𝑦).
Then 𝑓(𝑥) ∧ 𝑓(𝑦) = 𝑓(𝑥)
If 𝑓(𝑥) = 0, then 𝑓(𝑥) ∧ 𝑓(𝑦) = 0.
Since 𝑓 is a fuzzy set, so, clearly 𝑓(𝜆𝑥 + (1 − 𝜆)𝑦) ≥ 0
Therefore, 𝑓(𝜆𝑥 + (1 − 𝜆)𝑦) ≥ 0 = 𝑓(𝑥) ∧ 𝑓(𝑦)
⇒ 𝑓(𝜆𝑥 + (1 − 𝜆)𝑦) ≥ 𝑓(𝑥) ∧ 𝑓(𝑦)
Now, let 𝑓(𝑥) = 𝛽 > 0; where 𝛽 ∈ 𝐼,then
𝑓(𝑥) ∧ 𝑓(𝑦) = 𝛽.
By the hypothesis, 𝑓 is a convex set, then ∀𝑥, 𝑦 ∈ 𝑓𝛽 and ∀𝜆 ∈ 𝐼;
𝛽

𝜆𝑥 + (1 − 𝜆)𝑦 ∈ 𝑓 𝛽
⟹ 𝑓(𝜆𝑥 + (1 − 𝜆)𝑦) ≥ 𝛽
⟹ 𝑓(𝜆𝑥 + (1 − 𝜆)𝑦) ≥ 𝑓(𝑥) ∧ 𝑓(𝑦).

Hence, 𝑓 is a fuzzy convex set. Thus the theorem is proved.

Standard intersection of fuzzy sets: Suppose 𝑓 and 𝑔 are two fuzzy sets of 𝑋, then standard
intersection of 𝑓 and 𝑔 denoted by " 𝑓 ∩ 𝑔 " and defined by
(𝑓 ∩ 𝑔)(𝑥) = 𝑚𝑖𝑛{𝑓(𝑥), 𝑔(𝑥)}; ∀𝑥 ∈ 𝑋.
Examples:
0.2 0.4 1 0.8 0
i) Suppose, 𝑓 = { , ,𝑐, ,𝑒}
𝑎 𝑏 𝑑

0 0.8 0.3 0.2 𝑜. 1

𝑔={ , , , , }
𝑎 𝑏 𝑐 𝑑 𝑒

0 0.4 0.3 0.2 0

Then, 𝑓 ∩ 𝑔 = { 𝑎 , , , ,𝑒 }
𝑏 𝑐 𝑑

Standard union of fuzzy sets: Suppose 𝑓 and 𝑔 are two fuzzy sets of 𝑋, then standard union of
𝑓 and 𝑔 denoted by " 𝑓 ∪ 𝑔 " and defined by

13 Fuzzy Sets and Systems

 (𝑓 ∪ 𝑔)(𝑥) = 𝑚𝑎𝑥{𝑓(𝑥), 𝑔(𝑥)}; ∀𝑥 ∈ 𝑋.
Examples:
0.2 0.4 1 0.8 0
i) Suppose, 𝑓 = { , ,𝑐, ,𝑒}
𝑎 𝑏 𝑑

0 0.8 0.3 0.2 𝑜. 1

𝑔={ , , , , }
𝑎 𝑏 𝑐 𝑑 𝑒

0.2 0.8 1 0.8 0.1

Then, 𝑓 ∪ 𝑔 = { , ,𝑐, , }
𝑎 𝑏 𝑑 𝑒

Fuzzy subset: Suppose 𝑓: 𝑋 → 𝐼and 𝑔: 𝑋 → 𝐼 are two fuzzy sets. Then we say, f is a fuzzy subset
of g and write 𝑓 ⊆ 𝑔 iff 𝑓(𝑥) ≤ 𝑔(𝑥); ∀𝑥 ∈ 𝑋.
Examples:
i) Suppose 𝑋 = { 1, 2, 3, 4, 5, 6, 7, 8, 9 } and 𝑓, 𝑔 ∈ 𝑙 𝑥 , where
1 1
𝑓(𝑥) = 2 , 𝑔(𝑥) =
𝑥 𝑥
𝑥 1 2 3 4 5 6 7 8 9

𝑓(𝑥) 1 0.25 0.11 0.062 0.04 0.027 0.02 0.015 0.012

𝑔(𝑥) 1 0.5 0.33 0.25 0.20 0.166 0.14 0.125 0.111

Clearly, ∀𝑥 ∈ 𝑋, 𝑓(𝑥) ≤ 𝑔(𝑥). Therefore, 𝑓 ⊆ 𝑔.

Figure: 𝑓 ⊆ 𝑔.

Theorem: Suppose 𝑓, 𝑔 ∈ 𝐼 𝑋 , then for all 𝛼 ∈ 𝐼, show that i) (𝑓 ∩ 𝑔)𝛼 = 𝑓 𝛼 ∩ 𝑔𝛼

ii) (𝑓 ∪ 𝑔)𝛼 = 𝑓 𝛼 ∪ 𝑔𝛼
14 Fuzzy Sets and Systems
Proof of (i): Suppose 𝑓, 𝑔 ∈ 𝐼 𝑋 , then ∀ 𝑥 ∈ (𝑓 ∩ 𝑔)𝛼 ; we have

(𝑓 ∩ 𝑔)(𝑥) ≥ 𝛼

⇒ min[𝑓(𝑥), 𝑔(𝑥)] ≥ 𝛼

⇒ 𝑓(𝑥) ≥ 𝛼 and 𝑔(𝑥) ≥ 𝛼

⇒ 𝑥 ∈ 𝑓 𝛼 and 𝑥 ∈ 𝑔𝛼

⇒ 𝑥 ∈ (𝑓 𝛼 ∩ 𝑔𝛼 )

∴ (𝑓 ∩ 𝑔)𝛼 ⊆ (𝑓 𝛼 ∩ 𝑔𝛼 )… … … (i)

Again, ∀ 𝑥 ∈ (𝑓 𝛼 ∩ 𝑔𝛼 ), we find

𝑥 ∈ 𝑓 𝛼 and 𝑥 ∈ 𝑔𝛼

⇒ 𝑓(𝑥) ≥ 𝛼 and 𝑔(𝑥) ≥ 𝛼

⇒ min[𝑓(𝑥), 𝑔(𝑥)] ≥ 𝛼

⇒ (𝑓 ∩ 𝑔)(𝑥) ≥ 𝛼

⇒ 𝑥 ∈ (𝑓 ∩ 𝑔)𝛼

∴ (𝑓 𝛼 ∩ 𝑔𝛼 ) ⊆ (𝑓 ∩ 𝑔)𝛼 … … … (ii)
Therefore, from (i) and (ii), we get

(𝑓 ∩ 𝑔)𝛼 = 𝑓 𝛼 ∩ 𝑔𝛼 .

Proof of (ii): Suppose 𝑓, 𝑔 ∈ 𝐼 𝑋 , then ∀ 𝑥 ∈ (𝑓 ∪ 𝑔)𝛼 ; we have

(𝑓 ∪ 𝑔)(𝑥) ≥ 𝛼

⇒ max[𝑓(𝑥), 𝑔(𝑥)] ≥ 𝛼
⇒ 𝑓(𝑥) ≥ 𝛼 or 𝑔(𝑥) ≥ 𝛼

⇒ 𝑥 ∈ 𝑓 𝛼 or 𝑥 ∈ 𝑔𝛼

⇒ 𝑥 ∈ (𝑓 𝛼 ∪ 𝑔𝛼 )

∴ (𝑓 ∪ 𝑔)𝛼 ⊆ (𝑓 𝛼 ∪ 𝑔𝛼 )… … … (i)

Again, ∀ 𝑥 ∈ (𝑓 𝛼 ∪ 𝑔𝛼 ), we find

𝑥 ∈ 𝑓 𝛼 or 𝑥 ∈ 𝑔𝛼

⇒ 𝑓(𝑥) ≥ 𝛼 or 𝑔(𝑥) ≥ 𝛼

15 Fuzzy Sets and Systems

 ⇒ max[𝑓(𝑥), 𝑔(𝑥)] ≥ 𝛼

⇒ (𝑓 ∪ 𝑔)(𝑥) ≥ 𝛼

⇒ 𝑥 ∈ (𝑓 ∪ 𝑔)𝛼

∴ (𝑓 𝛼 ∪ 𝑔𝛼 ) ⊆ (𝑓 ∪ 𝑔)𝛼 … … … (ii)
Therefore, from (i) and (ii), we get

(𝑓 ∪ 𝑔)𝛼 = 𝑓 𝛼 ∪ 𝑔𝛼 .

Theorem: Suppose 𝑓, 𝑔 ∈ 𝐼 𝑋 , then for all 𝛼 ∈ 𝐼, show that

i) (𝑓 ∩ 𝑔)𝛼+ = 𝑓 𝛼+ ∩ 𝑔𝛼+
ii) (𝑓 ∪ 𝑔)𝛼+ = 𝑓 𝛼+ ∪ 𝑔𝛼+
Proof: Try yourself.

Theorem: Suppose 𝑓 ∈ 𝐼 𝑋 , then show that for all 𝛼 ∈ 𝐼, (𝑓)𝛼 = (𝑓)(1−𝛼)+ .

Proof: Since 𝑓 ∈ 𝐼 𝑋 , then ∀𝑥 ∈ (𝑓)𝛼 , we get

𝑓(𝑥) ≥ 𝛼

⇔ 1 − 𝑓(𝑥) ≥ 𝛼

⇔ −𝑓(𝑥) ≥ 𝛼 − 1

⇔ 𝑓(𝑥) ≤ 1 − 𝛼
⇔ 𝑓(𝑥) > 1 − 𝛼 is false

⇔ 𝑥 ∉ (𝑓)(1−𝛼)+

⇔ 𝑥 ∈ (𝑓)(1−𝛼)+

Hence, (𝑓)𝛼 = (𝑓)(1−𝛼)+ ; ∀ 𝛼 ∈ 𝐼.

Theorem: Suppose 𝐴, 𝐵 ∈ 𝐼 𝑋 , then ∀𝛼 ∈ (0, 1], prove that

i) 𝐴 ⊆ 𝐵 iff 𝐴𝛼 ⊆ 𝐵 𝛼
ii) 𝐴 ⊆ 𝐵 iff 𝐴𝛼+ ⊆ 𝐵 𝛼+

Proof of i): Let 𝐴 ⊆ 𝐵, we shall prove that 𝐴𝛼 ⊆ 𝐵 𝛼 , ∀𝛼 ∈ (0, 1]. If possible, suppose that
there exists 𝛼0 ∈ (0, 1] such that 𝐴𝛼0 ⊈ 𝐵 𝛼0 .

Then ∃ 𝑥0 ∈ 𝑋 such that 𝑥0 ∈ 𝐴𝛼 and 𝑥0 ∉ 𝐵 𝛼 .

⟹ 𝐴( 𝑥0 ) ≥ 𝛼 and 𝐵( 𝑥0 ) < 𝛼
16 Fuzzy Sets and Systems
 ⟹ 𝐵( 𝑥0 ) < 𝐴( 𝑥0 )

This contradicts with 𝐴 ⊆ 𝐵 and hence 𝐴𝛼 ⊆ 𝐵 𝛼 whenever 𝐴 ⊆ 𝐵; ∀𝛼 ∈ (0, 1].

Conversely, suppose 𝐴𝛼 ⊆ 𝐵 𝛼 but 𝐴 ⊈ 𝐵, then ∃ 𝑥0 ∈ 𝑋 such that 𝐴( 𝑥0 ) > 𝐵( 𝑥0 ).

Let 𝛼 = 𝐴( 𝑥0 ), then 𝑥0 ∈ 𝐴𝛼 and 𝑥0 ∉ 𝐵 𝛼 , which demonstrates that 𝐴𝛼 ⊆ 𝐵 𝛼 is not

satisfied for all 𝛼 ∈ (0, 1]. Hence, 𝐴𝛼 ⊆ 𝐵 𝛼 implies 𝐴 ⊆ 𝐵.

Thus, 𝐴 ⊆ 𝐵 iff 𝐴𝛼 ⊆ 𝐵 𝛼 .

Proof of ii): Suppose 𝐴𝛼+ ⊆ 𝐵 𝛼+ and 𝐴 ⊈ 𝐵, then ∃ 𝑥0 ∈ 𝑋 such that 𝐴( 𝑥0 ) > 𝐵( 𝑥0 ). Let 𝛼
be any number between 𝐴( 𝑥0 ) and 𝐵( 𝑥0 ) i.e. 𝐵( 𝑥0 ) < 𝛼 < 𝐴( 𝑥0 ). It follows that 𝑥0 ∈ 𝐴𝛼+
and 𝑥0 ∉ 𝐵 𝛼+ . Hence, 𝐴𝛼+ ⊈ 𝐵 𝛼+ , which demonstrates that 𝐴𝛼+ ⊆ 𝐵 𝛼+ is not satisfied for all
𝛼 ∈ [0, 1]. Therefore, 𝐴𝛼+ ⊆ 𝐵 𝛼+ implies 𝐴 ⊆ 𝐵… … … (i)

Conversely, Suppose 𝐴 ⊆ 𝐵 and 𝐴𝛼+ ⊈ 𝐵 𝛼+ , then ∃ 𝑥0 ∈ 𝑋 such that

𝑥0 ∈ 𝐴𝛼+ and 𝑥0 ∉ 𝐵 𝛼+

⟹ 𝐴( 𝑥0 ) > 𝛼 and 𝐵( 𝑥0 ) < 𝛼

⟹ 𝐵( 𝑥0 ) < 𝐴( 𝑥0 )
⟹ 𝐴 ⊈ 𝐵; which is a contradiction.

Therefore, 𝐴 ⊆ 𝐵 iff 𝐴𝛼+ ⊆ 𝐵 𝛼+ .

Theorem: Suppose 𝐴 ∈ 𝐼 𝑋 , then 𝐴𝛼+ ⊆ 𝐴𝛼 .

Proof: Since 𝐴 ∈ 𝐼 𝑋 , so ∀ 𝑥 ∈ 𝐴𝛼+ , we get 𝐴(𝑥) > 𝛼

⟹ 𝐴(𝑥) ≥ 𝛼 [‘or’ property]

⟹ 𝑥 ∈ 𝐴𝛼
∴ 𝐴𝛼+ ⊆ 𝐴𝛼 .

Theorem: On any referential set 𝑋, prove that

i) 0 is the smallest fuzzy set

ii) 1 is the largest fuzzy set.

Proof of i): Suppose 𝑓: 𝑋 → 𝐼 is a fuzzy set. Now ∀ 𝑥 ∈ 𝑋, 𝑓(𝑥) is grade of 𝑥 under 𝑓.

∴ 0 ≤ 𝑓(𝑥) ≤ 1.

Again, ∀ 𝑥 ∈ 𝑋, 0(𝑥) = 0

17 Fuzzy Sets and Systems

Hence, 0(𝑥) = 0 ≤ 𝑓(𝑥); ∀ 𝑥 ∈ 𝑋

⟹ 0(𝑥) ≤ 𝑓(𝑥); ∀ 𝑥 ∈ 𝑋

Therefore, 0 ⊆ 𝑓.

Proof of ii): Suppose 𝑓: 𝑋 → 𝐼 is a fuzzy set. Now ∀ 𝑥 ∈ 𝑋, 𝑓(𝑥) is grade of 𝑥 under 𝑓.

∴ 0 ≤ 𝑓(𝑥) ≤ 1.

Again, ∀ 𝑥 ∈ 𝑋, 1(𝑥) = 1 .

Thus, ∀ 𝑥 ∈ 𝑋, 𝑓(𝑥) ≤ 1 = 1(𝑥)

⟹ 𝑓(𝑥) ≤ 1(𝑥); ∀ 𝑥 ∈ 𝑋

∴ 𝑓 ⊆ 1.

**********************************************************************

18 Fuzzy Sets and Systems