You are on page 1of 7

204568 Fuzzy Logic: Theory and Applications Lecture 1

Solutions for homework 1 (Due October 5, 2003)


1. Excluded middle laws:
Given a fuzzy set A%with a triangular membership function as follows:

Fig 1. Triangular membership function.

a. Derive m A%( x ) as a mathematical equation.


By using a generic linear equation in the form of y = ax + b , we can easily have
m A%( x ) as follows:
�x - 2, 2 �x < 3

m A%( x ) = �- x + 4, 3 �x < 4
� 0, otherwise

b. Draw a new membership function based on Fig. 1 to show that


excluded middle laws are not valid for fuzzy sets.
Law of excluded middle:

m A%�A%( x ) ��ȹ
1 A% A% U

1
204568 Fuzzy Logic: Theory and Applications Lecture 1

Law of contradiction:

m A%�A%( x ) ��ǹ�
0 A% A%

2. Convexity:
Given a fuzzy set A%with a following membership function:

Fig 2. Nonconvex membership function.

a. Derive m A%( x ) as a mathematical equation.


Similar to 1 (a), we can derive the membership function as:
� x - 2, 2 �x < 3
�-0.4 x + 2.2, 3 �x < 4

m A%( x ) = �
�0.2 x - 0.2, 4 �x < 6

� - x + 7, 6 �x < 7

b. Show that this membership function is nonconvex based on:


m A%( l x1 + ( 1 - l ) x2 ) �min { m A%( x1 ) , m A%( x2 ) } , x1 , x2 �X , l �[ 0,1]
1
To show that this function is nonconvex, let x1 = 3, x2 = 6, l = .
3
m A%( x1 ) = 1, m A%( x2 ) = 1 and m A%( l x1 + ( 1 - l ) x2 ) = 0.8
Since m A%( l x1 + ( 1 - l ) x2 ) = 0.8 < min { 1,1} , thus m A%( x ) is nonconvex.

c. If you define “medium speed of a car” as a fuzzy set, should the


membership function of “medium speed” be convex or nonconvex ?
Explain your idea.
This membership function should be convex because

2
204568 Fuzzy Logic: Theory and Applications Lecture 1

3. Form of a membership function:


Given a general shape of a trapezoidal membership function as follows:

Fig. 3. Trapezoidal membership function.


Derive a closed (general) form of m A%( x ) .

� �x - a �
a�
� ,
� a �x �c
c-a �
��
� a, c �x �d
m A%( x ) = �
� �x - b �
a� , d �x �b
� �d - b �


� 0, Otherwise

4. You are asked to select an implementation technology for a numerical


processor. Computation throughput is directly related to clock speed. Assume
that all implementations will be in the same family (e.g., CMOS). You are
considering whether the design should be implemented using medium-scale
integration (MSI) with discrete parts, field-programmable array parts (FPGA),
or multichip modules (MCM). Define the universe of potential clock speeds as
X = { 1,10, 20, 40,80,100} MHz; and define MSI, FPGA, and MCM as fuzzy
sets of clock frequencies that should be implemented in each of these
technologies. The following table defines the membership values for each of
the three fuzzy sets. Note: The assignments made in this table reflect only the
operational speed capability per cost ratio and do not include other factors.

Clock frequency, MHz MSI FPGA MCM


1 1 0.3 0
10 0.7 1 0
20 0.4 1 0.5
40 0 0.5 0.7
80 0 0.2 1
100 0 0 1

Representing the three sets as MSI = M%, FPGA = F%, and MCM = C%, find the
following:
�1 1 1 0.5 0.2 0 �
a. M%�F%= � + + + + + �
�1 10 20 40 80 100
�0.3 0.7 0.4 0 0 0 �
b. M%�F%= � + + + + + �
�1 10 20 40 80 100
�0 0.3 0.6 1 1 1 �
c. M%= � + + + + + �
�1 10 20 40 80 100

3
204568 Fuzzy Logic: Theory and Applications Lecture 1

�0.7 0 0 0.5 0.8 1 �


d. F%= � + + + + + �
�1 10 20 40 80 100
�0.7 0 0.5 0.7 1 1 �
e. C%�F%= � + + + + + �
�1 10 20 40 80 100
�0 0.3 0.5 0.3 0 0 �
f. M%�C%= � + + + + + �
�1 10 20 40 80 100
�0.7 1 1 0.5 0.8 1 �
g. F%�F%= � + + + + + �
�1 10 20 40 80 100
�1 1 0.5 0.5 0.8 1 �
h. C%�F%= � + + + + + �
�1 10 20 40 80 100
� 1 1 0.5 0.3 0 0 �
i. M%�C%= � + + + + + �
� 1 10 20 40 80 100
� 1 0.7 0.6 1 1 1 �
j. M%�M%= � + + + + + �
� 1 10 20 40 80 100

5. Consider a local area network of interconnected workstations that


communicate using Ethernet protocols at a maximum rate of 10 Mbit/s. Traffic
rates on the network can be expressed as the peak value of the total bandwidth
(BW) used; and the two fuzzy variables, “Quiet” and “Congested,” can be
used to describe the perceived loading of the LAN. If the discrete universal set
X = { 0,1, 2,5, 7,9,10} represents bandwidth usage, then the membership
grades of these elements in the fuzzy sets Quiet Q%and Congested C%are as
shown in the table and in Fig. 4.

x ( BW ) , Mbit/s mQ%( x ) mC%( x )


0 1.0 0.0
1 1.0 0.0
2 0.8 0.0
5 0.3 0.4
7 0.1 0.6
9 0.0 0.8
10 0.0 1.0

Fig. 4. Membership functions of “Quiet” and “Congested”.


For these two fuzzy sets, find the union, intersection, complement of each,
difference Q%% C , both De Morgan’s laws, and both excluded middle laws, (a)
graphically and (b) numerically.

4
204568 Fuzzy Logic: Theory and Applications Lecture 1

�1 1 0.8 0.4 0.6 0.8 1 �


Q%�C%= � + + + + + + �
�0 1 2 5 7 9 10

�0 0 0 0.3 0.1 0 0 �
Q%�C%= � + + + + + + �
�0 1 2 5 7 9 10

�0 0 0.2 0.7 0.9 1 1 �


Q%= � + + + + + + �
�0 1 2 5 7 9 10

�1 1 1 0.6 0.4 0.2 0 �


C%= � + + + + + + �
�0 1 2 5 7 9 10

�1 1 0.8 0.3 0.1 0 0 �


Q%%
C = Q%�C%= � + + + + + + �
�0 1 2 5 7 9 10

5
204568 Fuzzy Logic: Theory and Applications Lecture 1

0 0 0.2 0.6 0.4 0.2 0 �


( Q%�C%) �= Q%��C%�= �
�+ +
�0 1 2
+
5
+
7
+ + �
9 10

1 1 1 0.7 0.9 1 1 �
( Q%�C%) �= Q%��C%�= �
�+ + +
�0 1 2 5
+ + + �
7 9 10

�0 0 0.2 0.3 0.1 0 0 �


Q%�Q%
�=� + + + + + + �
�0 1 2 5 7 9 10

�1 1 0.8 0.7 0.9 1 1 �


Q%�Q%
�=�+ + + + + + �
�0 1 2 5 7 9 10

6
204568 Fuzzy Logic: Theory and Applications Lecture 1

�0 0 0 0.4 0.4 0.2 0 �


C%�C%
�=�+ + + + + + �
�0 1 2 5 7 9 10

�1 1 1 0.6 0.6 0.8 1 �


C%�C%
�=�+ + + + + + �
�0 1 2 5 7 9 10

You might also like