TERM PAPER

ARTIFICIAL
INTELLIGENCE
(CSE-402)
TOPIC: FUZZY EXPERT
SYSTEM

Submitted to: Submitted by:

Mr. G.S.M THAKUR ARUN

PANTA

Roll no.
RD1804A14

Reg.No.1
0810737

Sec:
D1804
 B.TECH –
MBA (INT) CSE

ACKNOWLEDGEMENT

This is a humble effort to express our sincere gratitude towards those

who have guided and helped us to complete this project.

A project reported is major milestone during the study period of

a student.

We could have faced many problems but our teacher’s kind response
to our needs and requirement, their patient approach and their
positive criticism helped us in making our project.

Very warm thanks to our project-in-charge “Mr. G.S.M THAKUR”

with his support and constant encouragement AND LPU
LIBRARY it was not very easy without whose support to finish
our project.

With the motivation of our parent it was very easy to finish our
project successfully and satisfactorily in short span of time.

ARUN PANTA.
Contents:
• Fuzzy control system

• Overview
• History and applications
• Fuzzy sets
• Fuzzy control in detail
• Example
 Antilock breaks

• Fuzzy expert system

• What is fuzzy expert
system?
• The inference process
• Fuzzification
• Where fuzzy expert
system is used?
• Use of fuzzy expert
system

• An example of fuzzy expert system.

 cannot be expressed as "true" or
Fuzzy control "false" but rather as "partially
true". Although genetic
system algorithms and neural networks
can perform just as well as fuzzy
" logic in many cases , fuzzy logic
has the advantage that the
A fuzzy control system is a solution to the problem can be
control system based on fuzzy cast in terms that human
logic—a mathematical system operators can understand, so that
that analyzes analog input values their experience can be used in
in terms of logical variables that the design of the controller. This
take on continuous values makes it easier to mechanize
between 0 and 1, in contrast to tasks that are already
classical or digital logic, which successfully performed by
operates on discrete values humans.
of either 0 or 1 (true or
false).

•
History and
Overview applications
Fuzzy logic is widely used in Fuzzy logic was first proposed
machine control. The term itself by Lotfi A. Zadeh of the
inspires a certain skepticism, University of California at
sounding equivalent to "half- Berkeley in a 1965 paper. He
baked logic" or "bogus logic", elaborated on his ideas in a 1973
but the "fuzzy" part does not paper that introduced the
refer to a lack of rigour in the concept of "linguistic variables",
method, rather to the fact that which in this article equates to a
the logic involved can deal with variable defined as a fuzzy set.
fuzzy concepts—concepts that Other research followed, with
the first industrial application, a maintained stability in both
cement kiln built in Denmark, cases. Yamakawa eventually
coming on line in 1975. went on to organize his own
fuzzy-systems research lab to
Fuzzy systems were largely help exploit his patents in the
ignored in the U.S. because they field.
were associated with artificial
intelligence, a field that Following such demonstrations,
periodically oversells itself, Japanese engineers developed a
especially in the mid-1980s, wide range of fuzzy systems for
resulting in a lack of credibility both industrial and consumer
within the commercial domain. applications. In 1988 Japan
established the Laboratory for
The Japanese did not have this International Fuzzy Engineering
prejudice. Interest in fuzzy (LIFE), a cooperative
systems was sparked by Seiji arrangement between 48
Yasunobu and Soji Miyamoto of companies to pursue fuzzy
Hitachi, who in 1985 provided research.
simulations that demonstrated
the superiority of fuzzy control Japanese consumer goods often
systems for the Sendai railway. incorporate fuzzy systems.
Their ideas were adopted, and Matsushita vacuum cleaners use
fuzzy systems were used to microcontrollers running fuzzy
control accelerating, braking, algorithms to interrogate dust
and stopping when the line sensors and adjust suction power
opened in 1987. accordingly. Hitachi washing
machines use fuzzy controllers
Another event in 1987 helped to load-weight, fabric-mix, and
promote interest in fuzzy dirt sensors and automatically
systems. During an international set the wash cycle for the best
meeting of fuzzy researchers in use of power, water, and
Tokyo that year, Takeshi detergent.
Yamakawa demonstrated the use
of fuzzy control, through a set of As a more specific example,
simple dedicated fuzzy logic Canon developed an
chips, in an "inverted pendulum" autofocusing camera that uses a
experiment. This is a classic charge-coupled device (CCD) to
control problem, in which a measure the clarity of the image
vehicle tries to keep a pole in six regions of its field of view
mounted on its top by a hinge and use the information provided
upright by moving back and to determine if the image is in
forth. focus. It also tracks the rate of
change of lens movement during
Observers were impressed with focusing, and controls its speed
this demonstration, as well as to prevent overshoot.
later experiments by Yamakawa
in which he mounted a wine The camera's fuzzy control
glass containing water or even a system uses 12 inputs: 6 to
live mouse to the top of the obtain the current clarity data
pendulum. The system provided by the CCD and 6 to
measure the rate of change of simulations show that a fuzzy
lens movement. The output is control system can greatly
the position of the lens. The reduce fuel consumption. Firms
fuzzy control system uses 13 such as Boeing, General Motors,
rules and requires 1.1 kilobytes Allen-Bradley, Chrysler, Eaton,
of memory. and Whirlpool have worked on
fuzzy logic for use in low-power
As another example of a refrigerators, improved
practical system, an industrial air automotive transmissions, and
conditioner designed by energy-efficient electric motors.
Mitsubishi uses 25 heating rules
and 25 cooling rules. A In 1995 Maytag introduced an
temperature sensor provides "intelligent" dishwasher based
input, with control outputs fed to on a fuzzy controller and a "one-
an inverter, a compressor valve, stop sensing module" that
and a fan motor. Compared to combines a thermistor, for
the previous design, the fuzzy temperature measurement; a
controller heats and cools five conductivity sensor, to measure
times faster, reduces power detergent level from the ions
consumption by 24%, increases present in the wash; a turbidity
temperature stability by a factor sensor that measures scattered
of two, and uses fewer sensors. and transmitted light to measure
the soiling of the wash; and a
The enthusiasm of the Japanese magnetostrictive sensor to read
for fuzzy logic is reflected in the spin rate. The system determines
wide range of other applications the optimum wash cycle for any
they have investigated or load to obtain the best results
implemented: character and with the least amount of energy,
handwriting recognition; optical detergent, and water. It even
fuzzy systems; robots, including adjusts for dried-on foods by
one for making Japanese flower tracking the last time the door
arrangements; voice-controlled was opened, and estimates the
robot helicopters, this being no number of dishes by the number
mean feat, as hovering is a of times the door was opened.
"balancing act" rather similar to
the inverted pendulum problem; Research and development is
control of flow of powders in also continuing on fuzzy
film manufacture; elevator applications in software, as
systems; and so on. opposed to firmware, design,
including fuzzy expert systems
Work on fuzzy systems is also and integration of fuzzy logic
proceeding in the US and with neural-network and so-
Europe, though not with the called adaptive "genetic"
same enthusiasm shown in software systems, with the
Japan. The US Environmental ultimate goal of building "self-
Protection Agency has learning" fuzzy control systems.
investigated fuzzy control for
energy-efficient motors, and Fuzzy sets
NASA has studied fuzzy control
for automated space docking:
The input variables in a fuzzy • All the rules that apply
control system are in general are invoked, using the
mapped into by sets of membership functions
membership functions similar to and truth values obtained
this, known as "fuzzy sets". The from the inputs, to
process of converting a crisp determine the result of
input value to a fuzzy value is the rule.
called "fuzzification". • This result in turn will be
mapped into a
A control system may also have membership function and
various types of switch, or "ON- truth value controlling
OFF", inputs along with its the output variable.
analog inputs, and such switch • These results are
inputs of course will always combined to give a
have a truth value equal to either specific ("crisp") answer,
1 or 0, but the scheme can deal the actual brake pressure,
with them as simplified fuzzy a procedure known as
functions that happen to be "defuzzification".
either one value or another.
This combination of fuzzy
Given "mappings" of input operations and rule-based
variables into membership "inference" describes a "fuzzy
functions and truth values, the expert system".
microcontroller then makes
decisions for what action to take Traditional control systems are
based on a set of "rules", each of based on mathematical models
the form: in which the control system is
described using one or more
IF brake temperature IS differential equations that define
warm AND speed IS not very
the system response to its inputs.
fast
THEN brake pressure IS Such systems are often
slightly decreased. implemented as "PID
controllers" (proportional-
In this example, the two input integral-derivative controllers).
variables are "brake They are the products of decades
temperature" and "speed" that of development and theoretical
have values defined as fuzzy analysis, and are highly
sets. The output variable, "brake effective.
pressure", is also defined by a
fuzzy set that can have values If PID and other traditional
like "static", "slightly increased", control systems are so well-
"slightly decreased", and so on. developed, why bother with
This rule by itself is very fuzzy control? It has some
puzzling since it looks like it advantages. In many cases, the
could be used without bothering mathematical model of the
with fuzzy logic, but remember control process may not exist, or
the decision is based on a set of may be too "expensive" in terms
rules: of computer processing power
and memory, and a system based
on empirical rules may be more the required range of an input
effective. value, or the "universe of
discourse" in fuzzy jargon.
Furthermore, fuzzy logic is well
suited to low-cost As discussed earlier, the
implementations based on cheap processing stage is based on a
sensors, low-resolution analog- collection of logic rules in the
to-digital converters, and 4-bit or form of IF-THEN statements,
8-bit one-chip microcontroller where the IF part is called the
chips. Such systems can be "antecedent" and the THEN part
easily upgraded by adding new is called the "consequent".
rules to improve performance or Typical fuzzy control systems
add new features. In many cases, have dozens of rules.
fuzzy control can be used to
improve existing traditional Consider a rule for a thermostat:
controller systems by adding an
extra layer of intelligence to the IF (temperature is
current control method. "cold") THEN (heater is
"high")

This rule uses the truth value of

the "temperature" input, which is
Fuzzy control in detail some truth value of "cold", to
generate a result in the fuzzy set
Fuzzy controllers are very for the "heater" output, which is
simple conceptually. They some value of "high". This result
consist of an input stage, a is used with the results of other
processing stage, and an output rules to finally generate the crisp
stage. The input stage maps composite output. Obviously,
sensor or other inputs, such as the greater the truth value of
switches, thumbwheels, and so "cold", the higher the truth value
on, to the appropriate of "high", though this does not
membership functions and truth necessarily mean that the output
values. The processing stage itself will be set to "high", since
invokes each appropriate rule this is only one rule among
and generates a result for each, many. In some cases, the
then combines the results of the membership functions can be
rules. Finally, the output stage modified by "hedges" that are
converts the combined result equivalent to adjectives.
back into a specific control Common hedges include
output value. "about", "near", "close to",
"approximately", "very",
The most common shape of "slightly", "too", "extremely",
membership functions is and "somewhat". These
triangular, although trapezoidal operations may have precise
and bell curves are also used, but definitions, though the
the shape is generally less definitions can vary
important than the number of considerably between different
curves and their placement. implementations. "Very", for
From three to seven curves are one example, squares
generally appropriate to cover membership functions; since the
membership values are always favors the rule with the output of
less than 1, this narrows the greatest area, while the height
membership function. method obviously favors the rule
"Extremely" cubes the values to with the greatest output value.
give greater narrowing, while
"somewhat" broadens the The diagram below
function by taking the square demonstrates max-min
root. inferencing and centroid
defuzzification for a system with
In practice, the fuzzy rule sets input variables "x", "y", and "z"
usually have several antecedents and an output variable "n". Note
that are combined using fuzzy that "mu" is standard fuzzy-logic
operators, such as AND, OR, nomenclature for "truth value":
and NOT, though again the
definitions tend to vary: AND, in
one popular definition, simply
uses the minimum weight of all
the antecedents, while OR uses
the maximum value. There is
also a NOT operator that
subtracts a membership function
from 1 to give the
"complementary" function.

There are several different ways

to define the result of a rule, but
one of the most common and
simplest is the "max-min"
inference method, in which the
output membership function is Notice how each rule provides a
given the truth value generated result as a truth value of a
by the premise. particular membership function
for the output variable. In
Rules can be solved in parallel in centroid defuzzification the
hardware, or sequentially in values are OR'd, that is, the
software. The results of all the maximum value is used and
rules that have fired are values are not added, and the
"defuzzified" to a crisp value by results are then combined using
one of several methods. There a centroid calculation.
are dozens in theory, each with
various advantages and Fuzzy control system design is
drawbacks. based on empirical methods,
basically a methodical approach
The "centroid" method is very to trial-and-error. The general
popular, in which the "center of process is as follows:
mass" of the result provides the
crisp value. Another approach is • Document the system's
the "height" method, which takes operational specifications
the value of the biggest and inputs and outputs.
contributor. The centroid method
 • Document the fuzzy sets into a range of "states": "cold",
for the inputs. "cool", "moderate", "warm",
• Document the rule set. "hot", "very hot". The transition
• Determine the from one state to the next is hard
defuzzification method. to define.
• Run through test suite to
validate system, adjust An arbitrary static threshold
details as required. might be set to divide "warm"
• Complete document and from "hot". Like at exactly 90
release to production. degrees, warm ends and hot
begins. But this would result in a
discontinuous change when the
input value passed over that
threshold. The transition
wouldn't be smooth, as would be
Example: required in braking situations.

The way around this is to make

Antilock brakes the states fuzzy. That is, allow
them to change gradually from
As a first example, consider an one state to the next. In order to
anti-lock braking system, do this there must be a dynamic
directed by a microcontroller relationship established between
chip. The microcontroller has to different factors.
make decisions based on brake
temperature, speed, and other We start by defining the input
variables in the system. temperature states using
"membership functions":
The variable "temperature" in
this system can be subdivided

membership function while

gaining value in the next. In
With this scheme, the input other words, its ranking in the
variable's state no longer jumps category of cold decreases as it
abruptly from one state to the becomes more highly ranked in
next. Instead, as the temperature the warmer category.
changes, it loses value in one
At any sampled timeframe, the data, however, in this case,
"truth value" of the brake temperature.
temperature will almost always
be in some degree part of two Adding additional sophistication
membership functions: i.e.: '0.6 to this braking system, could be
nominal and 0.4 warm', or '0.7 done by additional factors such
nominal and 0.3 cool', and so on. as traction, speed, inertia, set up
in dynamic functions, according
The above example to the designed fuzzy system.
demonstrates a simple
application, using the abstraction
of values from multiple values.
This only represents one kind of

low is a membership function

Fuzzy Expert (fuzzy subset) defined on x, high
is a membership function
Systems defined on y, and medium is a
membership function defined on
What is a Fuzzy Expert z. The part of the rule between
System? the "if" and "then" is the rule's
_premise_ or _antecedent_. This
Put as simply as possible, a is a fuzzy logic expression that
fuzzy expert system is an expert describes to what degree the rule
system that uses fuzzy logic is applicable. The part of the rule
instead of Boolean logic. In following the "then" is the rule's
other words, a fuzzy expert _conclusion_ or _consequent_.
system is a collection of This part of the rule assigns a
membership functions and rules membership function to each of
that are used to reason about one or more output variables.
data. Unlike conventional expert Most tools for working with
systems, which are mainly fuzzy expert systems allow
symbolic reasoning engines, more than one conclusion per
fuzzy expert systems are rule.
oriented toward numerical
processing. A typical fuzzy expert system
has more than one rule. The
The rules in a fuzzy expert entire group of rules is
system are usually of a form collectively known as a
similar to the following: _rulebase_ or _knowledge
base_.
if x is low and y
is high then z = The Inference Process
medium
With the definition of the rules
and membership functions in
where x and y are input variables
hand, we now need to know how
(names for know data values), z
to apply this knowledge to
is an output variable (a name for
specific values of the input
a data value to be computed),
variables to compute the values
of the output variables. This and vice-versa. This is also
process is referred to as not required, but fairly
_inferencing_. In a fuzzy expert common.
system, the inference process is 3. The same membership
a combination of four functions are used for all
variables. This isn't
subprocesses: _fuzzification_,
required, and is also *not*
_inference_, _composition_, and common.
_defuzzification_. The
defuzzification subprocess is
optional.
Fuzzification:

For the sake of example in the In the fuzzification sub process,

following discussion, assume the membership functions
that the variables x, y, and z all defined on the input variables
take on values in the interval [ 0, are applied to their actual values,
10 ], and that we have the to determine the degree of truth
following membership functions for each rule premise. The
and rules defined. degree of truth for a rule's
premise is sometimes referred to
low(t) = 1 - t / 10 as its _alpha_. If a rule's premise
high(t) = t / 10 has a nonzero degree of truth (if
the rule applies at all...) then the
Rule 1: if x is low and rule is said to _fire_.

y is low then z is high For example:

Rule 2: if x is low and
x y low(x)
y is high then z is low high(x) low(y) high(y)
Rule 3: if x is high alpha1 alpha2 alpha3
alpha4
and y is low then z is --------------------------
--------------------------
low --------------------------
Rule 4: if x is high 0.0 0.0 1.0
0.0 1.0 0.0
and y is high then z is 1.0 0.0 0.0
high 0.0
0.0 3.2 1.0
0.0 0.68 0.32
Notice that instead of assigning 0.68 0.32 0.0
a single value to the output 0.0
variable z, each rule assigns an 0.0 6.1 1.0
0.0 0.39 0.61
entire fuzzy subset (low or high). 0.39 0.61 0.0
0.0
Notes: 0.0 10.0 1.0
0.0 0.0 1.0
1. In this example, low(t) 0.0 1.0 0.0
0.0
+high(t)=1.0 for all t. This 3.2 0.0 0.68
is not required, but it is 0.32 1.0 0.0
fairly common. 0.68 0.0 0.32
2. The value of t at which 0.0
low(t) is maximum is the 6.1 0.0 0.39
same as the value of t at 0.61 1.0 0.0
which high(t) is minimum,
0.39 0.0 0.61
0.0
10.0 0.0 0.0
1.0 1.0 0.0
0.0 0.0 1.0
0.0
3.2 3.1 0.68
0.32 0.69 0.31
0.68 0.31 0.32
0.32 • Use of Fuzzy
3.2 3.3 0.68
0.32 0.67 0.33 Expert System:
0.67 0.33 0.32
0.32 THE USE OF EXPERT
10.0 10.0 0.0
1.0 0.0 1.0 SYSTEMS HAS IMPROVED
0.0 0.0 0.0 STUDENTS LEARNING IN
1.0 ZIMBABWE

• Where are Fuzzy ABSTRACT

The introduction of the e-learning,
Expert Systems both multimedia and conventional,
Used? has resulted in drastic and
inevitable improvements in the
learning environment for both
To date, fuzzy expert systems learners and tutors. This has been
are the most common use of facilitated by the use of internet and
fuzzy logic. They are used in has attracted more organizations
several wide-ranging fields, and among them, academic
including: institutions for communication as
well as to share crucial information,
• Linear and nonlinear such as research works and new
control. knowledge discovered. All these
• Pattern recognition. efforts created the learning
• Financial systems. environment for students favorable,
as learning resources becomes
abundant in addition to the books
and many others I can't think of.
and the tutors they already have.
It's late. I'm going home! :-) The current economic hardships
have affected Zimbabwean
---
education system having negative
Erik Horstkotte, Togai
InfraLogic, Inc. repercussions to the attraction of
The World's Source for both local and foreign
Fuzzy Logic Solutions (The professionals. The current
company, not me!) experienced professionally
[email protected], gordius!til! qualified tutors and technicians are
erik - (714) 975-8522 being drained out in search of
[email protected] for info, greener pastures leaving the
[email protected] for Universities with only senior
fuzzy mail-server
faculty heads and a pool of junior
staff. However, this is resulting in
inevitable deterioration of
\ educational standards as semi-
qualified staff is dominating
throughout the whole industry .
INTRODUCTION
In spite of the Internet and the
World Wide Web, which provides Expert System
abundant educational resources,
various computer application Expert systems are applicable to
packages, e-portfolios and the various trades, professions and
cropping of open systems, other sections that involve human
proceeded further to enhanced ideas, deductions and reasoning
student work through powerful which implies that any fields that
documentations, presentations, require human expertise can use it
calculations, publications, as well to minimize risks associated with
as formalized, integrated and doing the business, or improve
logical storage of student consistence of solutions, or improve
information. All these efforts completeness, or improve accuracy
proved effective but left the or all at once while appropriate
tutoring side unabated. documentation of the steps
followed is compiled for reference
and explanations (Darlington,
Only human intellectuals are
2000).In addition, today’s weather
domination the whole scene.
forecasting is inevitably done by
Therefore this gives room for
expert systems. These systems do
attitudes like biasness,
the actual prediction of the weather
favoritism, negligence, laziness and
accurately, quickly and consistently
incompetence, which all
unlike the case of human beings
deteriorates the quality of
whose reasoning is sometimes
professional work
unpredictable, slow and
delivered to the students in context.
inconsistent
In spite of all these human
In spite of these areas, expert
weaknesses, the Zimbabwean
systems are an essential useful tool
education system is
for instructing or training which
currently suffering in adequacy of
makes them ideal to aid academic
professionally qualified staff to
tutors to deliver quality data to their
execute the lecturing and examining
resective students.
duties as
the existing ones are leaving in
search of greener pastures in
neighboring nations.
Bulks of semi-qualified staff have
been inevitably considered to fill
the gap by the ministry of higher
METHODOLOGY
and tertiary
Although so many expert systems
education. The efforts rendered by
have been developed, some are
these staff are definitely below the
even being open systems; one can
expected professional level and
be designed
hence there
these expert systems for the
is absolute need to have expert
academic industry. In this case, a
system in brain drained sections of
knowledge base is developed to
profession such as medical and
capture all the
information
fundamental aspects pertaining to
technology so that the systems can
each particular module. Various
assume much of duties of tutors
expert ideas are coaxed out of
through provision of expert advice
human experts
and
and encoded for use by expert
decisions while lecturing.
systems for forward or backward
reasoning strategies in problem control group. A final analysis on
solving. All the their
course outlines, lecturing materials final examinations performance for
and tutorials are expertly the sections where the questions
formulated by professionals, emanated from the binary converter
analyzed and formally and text
represented into rules or frames or books/lecture based was done.
cases depending on the method
suggested by the knowledge Design of the expert system
engineer. The expert system (Binary
The author used an introductory Converter) performed number
course offered in nearly all the conversion from decimal to binary
Universities and Colleges in number system or
Zimbabwe named vice-versa. Under normal
Introduction to computer science circumstances, the conversion of
which is designed to strategically any number from decimal number
introduce computer science system to binary is
concepts to all done using the long division
students in their first academic year. mathematical principle, noting
The course covers computer basics down the remainders.
concepts, importance and The binary equivalent is the
applications as concatenation of the remainders
an aid to student learning, starting with the last calculated as
organizational and business the highest unit as
operational needs for profit or non- shown by the example below.
profit reasons. 7
However, the author designed a Question 1
sample expert system (Binary Convert 1010 into Binary
Converter) that performs number equivalent.
conversion from Answer
decimal to binary number system or However, the concept of recursive
vice-versa. Under normal long division while noting the
circumstances, the conversion of remainder until the dividend is 0
any number from has been
decimal number system to binary is considered and embedded in the
done using the long division design of Binary Converter Expert
mathematical principle, noting system. This system aids students in
down the understanding the division process
remainders. and how the actual binary
6 equivalent value can be deduced
After implementing the binary from the results of
converter and given to students the overall computations.
doing Introduction to Computers,
students used the
tool covering the concept. The Binary Converter
group of students was divided into Entering a valid digit from decimal
two groups, one using the expert number system will be followed by
system and clicking the "Compute" command
another using the textbooks and which begins the conversion
lecture based. Students were given process by drawing the long-
a questionnaire to determine their division horizontal and vertical
satisfaction lines as well as positioning the
using the binary converter and the divisor , dividend, remainder
text books/lecture based through the
 alternative to the private tutorial
Benefits to the students (tutee) and individual training. However,
The systems better performs expert systems are the best in those
simulations and aid practices better situations where there is a structure
than the teaching side. However, that is noted as previously existing
the expert or can be elicited just like the
systems are normally used in academic environment. They also
support of other learning / teaching work perfectly depending on the
activities such as the problem based technology used and need to be
learning designed suiting the students
(PBL). The students learn by background. The results show that
repeated use of the concept and they are suitable to the
they understand even the slow Zimbabwean educational needs and
learner can visualize. bring in high satisfaction and
Expert systems are reproducible positive outcomes to students
and thus making them readily learning and results. Most
available for consultation by developing countries are faced with
students at every severe brain drain, which leads to
11 loss of knowledgeable experts in
stage and permits (indeed always various sectors like; education,
should permit) the student to medical, engineering, mining,
interrogate and analyze the agriculture etc. This leaves a
reasoning process. country with absolutely minimum
Finally, worked examples, and all resources to support economic,
forms of guidelines are readily social and manpower development
available for revision purposes. since the remaining professional
will definitely fail to meet the
CONCLUSION requirements. The training of the
If implemented, expert systems can remaining manpower to meet the
utilize their ability to adaptively necessary knowledge will be
adjust the training for each inevitable. This means that relying
particular student on the bases of on knowledgeable human experts is
his/her own pace of learning which unsustainable for economic, social
allows students to gain deep and human development in the
understanding of fundamentals to country since more and more
be able to follow the more continue to trick 12 out with the
advanced topics within the need for greener pastures.
specified field. In addition, expert
systems will provide the excellent

An Example Fuzzy plans. However this example

does demonstrate the
Expert System mechanics of developing a
fuzzy expert system. A
The problem this expert mathematica notebook for the
system addresses is example is available.
supplemental feeding for range
cows. The example is Step 1 -- Define Fuzzy Sets
extremely simple and needs
additional knowledge
Step 2 -- Relate Observations
incorporated before it would be
to Fuzzy Sets
useful for developing actual
range cow supplementation
Step 3 -- Define Fuzzy Rules
Step 4 -- Evaluate Each Case Step 5 -- Combine Information
for all Fuzzy Rules from Rules

Step 6 -- Defuzzify Results

Thanks