SOFT COMPUTING

ETCS-456

Faculty Name: Ms. Ruchi Goel Student name: Dhruv Goel

Roll No.: 44114802717
Semester: 8th
Group: 8-C-8

Maharaja Agrasen Institute of Technology, PSP Area,

Sector – 22, Rohini, New Delhi – 110085
 SOFT COMPUTING LAB

PRACTICAL RECORD

PAPER CODE : ETCS-456

Name of the student : Dhruv Goel

University Roll No. : 44114802717

Branch : CSE

Section/ Group : 8-C-8

PRACTICAL DETAILS

a) Experiments according to the list provided by GGSIPU

Exp. no Experiment Name Date of Date of Remarks Marks

performance checking (10)
Implementation of
Fuzzy Operations
1.
Implementation of
Fuzzy Relations
2. (Max-min
Composition)
Implementation of
3. Fuzzy Controller
(Washing Machine)
Implementation of
4. Simple Neural
Network
(McCullohPitts
model)
Implementation of
5. Perceptron
Learning Algorithm
Implementation of
6. Unsupervised
Learning Algorithm
Exp. no
Experiment Name Date of Date of Remarks Marks
performance checking (10)
7. Implementation of
Simple Genetic
Application
Study of ANFIS
8. Architecture
Study of
9. Derivative-free
Optimization
Study of research
10. paper on Soft
Computing
 MAHARAJA AGRASEN INSTITUTE OF TECHNOLOGY

VISION

To nurture young minds in a learning environment of high academic value and imbibe spiritual and
ethical values with technological and management competence.

MISSION

The Institute shall endeavour to incorporate the following basic missions in the teaching methodology:
❖ Engineering Hardware – Software Symbiosis: Practical exercises in all Engineering
and Management disciplines shall be carried out by Hardware equipment as well as the related
software enabling a deeper understanding of basic concepts and encouraging inquisitive nature.
❖ Life-Long Learning: The Institute strives to match technological advancements and
encourage students to keep updating their knowledge for enhancing their skills and inculcating
their habit of continuous learning
❖ Liberalization and Globalization: The Institute endeavors to enhance technical and
management skills of students so that they are intellectually capable and competent
professionals with Industrial Aptitude to face the challenges of globalization.
❖ Diversification: The Engineering, Technology and Management disciplines have
diverse fields of studies with different attributes. The aim is to create a synergy of the above
attributes by encouraging analytical thinking.
❖ Digitization of Learning Processes: The Institute provides seamless opportunities for
innovative learning in all Engineering and Management disciplines through digitization of
learning processes using analysis, synthesis, simulation, graphics, tutorials and related tools to
create a platform for multi-disciplinary approach.
❖ Entrepreneurship: The Institute strives to develop potential Engineers and Managers
by enhancing their skills and research capabilities so that they emerge as successful
entrepreneurs and responsible citizens.
 MAHARAJA AGRASEN INSTITUTE OF TECHNOLOGY

COMPUTER SCIENCE & ENGINEERING DEPARTMENT

VISION

To Produce “Critical thinkers of Innovative Technology”

MISSION

To provide an excellent learning environment across the computer science discipline to inculcate
professional behavior, strong ethical values, innovative research capabilities and leadership abilities
which enable them to become successful entrepreneurs in this globalized world.

❖ To nurture an excellent learning environment that helps students to enhance their

problem solving skills and to prepare students to be lifelong learners by offering a solid
theoretical foundation with applied computing experiences and educating them about their
professional, and ethical responsibilities.
❖ To establish Industry-Institute Interaction, making students ready for the industrial
environment and be successful in their professional lives.
❖ To promote research activities in the emerging areas of technology convergence.
❖ To build engineers who can look into technical aspects of an engineering solution
thereby setting a ground for producing successful entrepreneur
 Introduction to the Lab Lab
Objective The objective of this course is to provides necessary understanding, analyze and
applications of the basic mathematical elements of fuzzy sets, neural networks. This course
prepares the students for developing intelligent systems and provides knowledge of
learning algorithm in neural networks. Upon completion of this course, the student should
be able to understand the research trends in soft computing.

Course Outcomes At the end of the course, a student will be able to:
C456.1: Learn mathematical basis as well as the general principles of various soft
computing techniques.

C456.2: Identify, formulate and solve engineering problem.

C456.3: Understand the role of fuzzy sets, neural networks, ANN, Neuro-fuzzy and
genetic algorithms in engineering.

C456.4: Implement and validate the solution including both hardware and software.

C456.5: Use the current techniques and tools required for computing practice.

C456.6: Understand the importance of research in soft computing.

 LAB REQUIREMENTS
Hardware Detail
CentOS/Fedora Linux SCILAB Intel i3/C2D Processor/2 GB RAM/500GB
HDD/MB/Lan Card/ Key Board/ Mouse/CD Drive/15” Color Monitor/ UPS 24
Nos

LaserJet Printer 1 No

Software Detail
CentOS/Fedora Linux
SCILAB
LIST OF EXPERIMENTS

Experiment No. – 1: Implementation of Fuzzy Operations

Experiment No. – 2: Implementation of Fuzzy Relations (Max-

min Composition)

Experiment No. – 3: Implementation of Fuzzy Controller (Washing Machine)

Experiment No. – 4: Implementation of Simple Neural Network (McCulloh-

Pitts model)

Experiment No. – 5: Implementation of Perceptron Learning Algorithm

Experiment No. – 6: Implementation of Unsupervised Learning Algorithm

Experiment No. – 7: Implementation of Simple Genetic Application

Experiment No. – 8: Study of ANFIS Architecture

 4. LIST OF EXPERIMENTS
(Beyond the syllabus)

Experiment No. – 9: Study of Derivative-free Optimization

Experiment No. – 10: Study of research paper on Soft Computing

MARKING SCHEME FOR THE PRACTICAL EXAMS
There will be two practical exams in each semester.

i. Internal Practical Exam

ii. External Practical Exam

INTERNAL PRACTICAL EXAM

It is taken by the respective faculty of the batch.

MARKING SCHEME FOR THIS EXAM IS:

Total Marks: 40

Division of 10 marks per practical is as follows:

Rubrics for : Laboratory (General)

Experiment Grading Rubrics

Max.
Sr No. Component Marks
(LAC)
2 marks 1 mark

Completeness of practical, Incomplete practical,

Practical
1 2 exhibits proficiency in using unformatted, lacks comments,
Performance
different types of inputs. Demonstrates no proficiency.

Output is free of errors and Output contains few logical

output is obtained. errors and/or no output is
Output and
2 2 Demonstrates excellent obtained. Demonstrates partial
Validation
understanding of the concepts understanding of the concepts
relevant to the experiment. relevant to the experiment.

Attendance 1. Four marks for answering more than 75% questions.

and Viva
3 4 2. Two marks for answering more then 50% questions.
Questions
Answered 3. One mark for answering less then 50% questions.

Timely
Submission
4 2 On time submission Late submission
of Lab
Record
Each experiment will be evaluated out of 10 marks. At the end of the semester average of 8
best performed practical will be considered as marks out of 40.
EXTERNAL PRACTICAL EXAM

It is taken by the concerned lecturer of the batch and by an external examiner. In this exam student needs to
perform the experiment allotted at the time of the examination, a sheet will be given to the student in which
some details asked by the examiner needs to be written and at the last viva will be taken by the external
examiner.

MARKING SCHEME FOR THIS EXAM IS:

Total Marks: 60

Division of 60 marks is as follows

1. Sheet filled by the student: 20

2. Viva Voice: 15

3. Experiment performance: 15

4. File submitted: 10

NOTE:

• Internal marks + External marks = Total marks given to the students

(40 marks) (60 marks) (100 marks)

Experiments given to perform can be from any section of the lab.

 EXPERIMENT-1

Aim: Implementation of Fuzzy Operations.

Objectives:
• Understanding of the basic mathematical elements of fuzzy sets.
• Analyze concepts of fuzzy set.
• To use fuzzy set operations to implement current computing techniques used in
fuzzy computing.
Theory:

Fuzzy Logic:
Fuzzy logic is an organized method for dealing with imprecise data. It is a multivalued logic that
allows intermediate values to be defined between conventional solutions.
In classical set theory, the membership of elements in a set is assessed in binary terms according to a
bivalent condition — an element either belongs or does not belong to the set. By contrast, fuzzy set
theory permits the gradual assessment of the membership of elements in a set; this is described with the
aid of a membership function valued in the real unit interval [0, 1].

Bivalent Set Theory can be somewhat limiting if we wish to describe a 'humanistic' Illustration
mathematically. For Illustration, Fig 1 below illustrates bivalent sets to characterize the temperature of
a room. The most obvious limiting feature of bivalent sets that can be seen clearly from the diagram is
that they are mutually exclusive - it is not possible to have membership of more than one set. Clearly, it
is not accurate to define a transition from a quantity such as 'warm' to 'hot' by the application of one
degree Fahrenheit of heat. In the real world a smooth (unnoticeable) drift from warm to hot would
occur.

Fuzzy Sets:

A fuzzy set is a pair where is a set and

For each the value is called the degree of membership of in

For a finite set the fuzzy set is often denoted by
 Let Then,
• is called not included in the fuzzy set if
• is called fully included if
• is called a fuzzy member if .
Illustration:

1. Union:

Union of two fuzzy sets and is denoted as and is defined as,

2. Intersection

3. Complement

Complement of a fuzzy set denoted by and is defined as,

 Illustration:

4. Algebraic Sum

Algebraic sum of two fuzzy sets and B is denoted by and is defined as,

Illustration:

5. Algebraic Product

Algebraic product of two fuzzy sets and B denoted by and is defined

as, Illustration:

Result/Conclusion

The concepts of union, intersection and complement are implemented using fuzzy sets which helped
to understand the differences and similarities between fuzzy set and classical set theories. It provides
the basic mathematical foundations to use the fuzzy set operations.
Viva Questions:

• What are the properties of Fuzzy set?

Commutativity
Associativity
Distributivity
Idempotency
Identity
Transitivity

• What is De Morgan‘s Law in Crisp Set?

For any two finite sets A and B;

(i) (A U B)' = A' ∩ B' (which is a De Morgan's law of union).
(ii) (A ∩ B)' = A' U B' (which is a De Morgan's law of intersection).

• What is the difference between the crisp set and fuzzy set?

FUZZY SET CRISP SET

Prescribed by vague or Defined by precise and certain
ambiguous properties. characteristics.
Elements are allowed to be Element is either the member
partially included in the set. of a set or not
Used in fuzzy controllers Digital design
Infinite-valued bi-valued

• List the different fuzzy set operations?

Union, intersection, complement, scalar product, vector product, Cartesian product and power.

• What is fuzzy logic?

Fuzzy logic is a form of many-valued logic in which the truth values of variables may be any real
number between 0 and 1 both inclusive. It is employed to handle the concept of partial truth,
where the truth value may range between completely true and completely false.
 EXPERIMENT-2

Aim: Implementation of fuzzy relations (Max-Min Composition)

Objectives:
• Understanding of the basic mathematical elements of the theory of fuzzy sets.
• To introduce the ideas of fuzzy sets, fuzzy logic.
• To implement applications using the fuzzy set operations.

Theory:
Max-Min Composition of fuzzy Relations
Fuzzy relation in different product space can be combined with each other by the operation
called ―Composition . There are many composition methods in use , e.g. max- product method, max-
average method and max-min method. But max-min composition method is best known in fuzzy logic
applications.

Definition:
Composition of Fuzzy Relations
• Consider two fuzzy relation; R (X ×Y) and S (Y ×Z), then a relation T (X × Z ), can
be expressed asmax-min composition

T=RoS
µT (x, z) = max-min [µR (x, y), µS (y, z)]
= V [µR (x, y) ^ µS (y, z)]

• If algebraic product is adopted, then max-product compositionis adopted:

T=RoS
µT (x, z) = max [µR (x, y) . µS (y, z)]

= V [µR (x, y) . µS (y, z)]

• The max-min composition can be interpreted as indicating the strength of the existence of
relation between the elements of X and Z.

• Calculations of (R o S) are almost similar to matrix multiplication.

Crisp relation:
Crisp relation is defined on the Cartesian product of two sets. Consider,

The relation on this Cartesian product will be,

 Illustration: Let X={1,4,5} and Y={3,6,7} then for relation R=x<y,

Fuzzy relation:

Let be universal sets then,

Is called a fuzzy relation in

Illustration: Let X={1,2,3} and Y={1,2}

If , then

Max-Min Composition:

Let X,Y and Z be universal sets and let R and Q be relations that relate them as,

Then S will be a relation that relates elements of X with elements of Z as,

Max min composition is then defined as,

Illustration: and

Result/Conclusion: With the use of fuzzy logic principles max min composition of fuzzy set
is calculated which describes the relationship between two or more fuzzy sets.
Viva Questions:

• What is the main difference between the probability and fuzzy logic?

Fuzzy Logic is all about degree of truth. Probability theory has nothing to reason about things that
are't entirely true or false. In short, we can say that Fuzzy Logic captures the meaning of partial
truth whereas Probability theory captures partial knowledge.

• What are the types of fuzzy logic sets?

L-fuzzy sets
Neutrosophic fuzzy sets
Pythagorean fuzzy sets

• Who is the founder of fuzzy logic?

Lotfi Zadeh

• What is fuzzy arithmetic?

Fuzzy arithmetic or arithmetic of fuzzy numbers is generalisation of interval arithmetic,

where rather than considering intervals at one constant level only, several levels are
considered in [0, 1].
 EXPERIMENT-3

Aim: Implementation of fuzzy controller (Washing Machine)

Objectives:

• Cover fuzzy logic inference with emphasis on their use in the design of
intelligent or humanistic systems.
• Prepare the students for developing intelligent systems.

Theory:

Control System:

Any system whose outputs are controlled by some inputs to the system is called control
system.
Fuzzy Controller:

Fuzzy controllers are the most important applications of fuzzy theory. They
work different than conventional controllers as:
Expert knowledge is used instead of differential equations to describe a system.
This expert knowledge can be expressed in very natural way using linguistic
variables, which are described by fuzzy sets.
The fuzzy controllers are constructed in following three stages:
1. Create the membership values (fuzzify).
2. Specify the rule table.
3. Determine your procedure for defuzzifying the result.

Washing Machine Controller:

To design a system using fuzzy logic, input & output is necessary part of the system.
Main function of the washing machine is to clean cloth without damaging the cloth. In
order to achieve it, the output parameters of fuzzy logic, which are the washing
parameters, must be given more importance. The identified input & output parameters
are:
Input: 1. Degree of dirt
2. Type of dirt
Output: Wash time
 Fuzzy sets:

The fuzzy sets which characterize the inputs & output are given as follows:

1. Dirtiness of clothes
 2. Type of dirt

3. Wash time

Procedure:
Step1: Fuzzification of inputs

For the fuzzification of inputs, that is, to compute the membership for the
antecedents, the formula used is,
 MAX

Degree of
membership

Point 1 x Point 2

Step 2: Defining set of rules

S M L

NG VS S M

M M M L

G L L VL

1. If Dirtness of clothes is Large and Type of dirt is Greasy then Wash Time is Very Long;

2. If Dirtness of clothes is Medium and Type of dirt is Greasy then Wash Time is Long;
3. If Dirtness of clothes is Small and Type of dirt is Greasy then Wash Time is Long;

4. If Dirtness of clothes is Large and Type of dirt is Medium then Wash Time is Long;

5. If Dirtness of clothes is Medium and Type of dirt is Medium then Wash Time is Medium

6. If Dirtness of clothes is Small and Type of dirt is Medium then Wash Time is Medium;

7. If Dirtness of clothes is Large and Type of dirt is Not Greasy then Wash Time is Medium;

8. If Dirtness of clothes is Medium and Type of dirt is Not Greasy then Wash Time is Short;

9. If Dirtness of clothes is Small and Type of dirt is Not Greasy then Wash Time is Very Short;

Fuzzy Output

Fuzzy output of the system is the ‗fuzzy OR‘ if all the fuzzy outputs of the rules with non zero
rule strengths.

Step 3: Defuzzification

The Centre of Gravity method is applied go defuzzify the output,

Result/Conclusion

Fuzzy controller for washing machine application is implemented using the fuzzy logic
which also defines the rules for fuzzification and de-fuzzification. We have analyzed how
outputs are controlled by some inputs of the system through this experiment.
 Viva Questions:

• What is the reason that logic function has rapidly become one of the most successful
technologies for developing sophisticated control systems?
There are mainly two reasons:
(i)Fuzzy logic applies the concept of ‘certain degree’ which is similar to the way human beings
think. Instead of just being either true or false, fuzzy logic can be true partially and also false
partially at the same time. This is similar to the human mind.
(ii)Fuzzy logic can uses exact points representing to what degree an event occurs and with fuzzy
rules it generates precise outcomes.

• What is sequence of steps taken in designing a fuzzy logic machine?

Following is the sequence for the designing a fuzzy logic machine:
Fuzzification ->Rule Evaluation->Defuzzification
when designing a fuzzy logic,we first have to define the fuzzy sets and make appropriate
member function.The rule evaluation comes in which matches the sets to its corresponding rules.

• What is the fuzzy inference system (FIS)?

A fuzzy inference system (FIS) is defined as a system that uses fuzzy membership functions to make
a decision.

• What is defuzzification and fuzzy controller?

Defuzzification is the process of producing a quantifiable result in Crisp logic, given fuzzy sets
and corresponding membership degrees. It is the process that maps a fuzzy set to a crisp set.
 EXPERIMENT-4

Aim: To implement Mc-Culloch pits Model using XOR

Objectives:

• The student will be able to obtain the fundamentals and different

architecture of neural networks.
• The student will have a broad knowledge in developing the
different algorithms for neural networks.

Theory:

Neural network was inspired by the design and functioning of human brain
and components.
Definition:
―Information processing model that is inspired by the way biological nervous
system (i.e the brain) process information, is called Neural Network.
Neural Network has the ability to learn by Illustrations. It is not designed to perform
fix /specific task, rather task which need thinking (e.g. Predictions).
ANN is composed of large number of highly interconnected processing
elements(neurons) working in unison to solve Illustrations. It mimic human brain. It is
configured for special application such as pattern recognition and data classification
through a learning process. ANN is 85-90% accurate.

Basic Operation of a Neural Network:

X1 and X2 – input neurons.

Y- output neuron

Weighted interconnection links- W1 and W2.

Net input calculation is :

Yin= x1w1+x2w2

Output is :

y=f(Yin)

Output= function

The McCulloch-Pitts Model of Neuron:

The early model of an artificial neuron is introduced by Warren McCulloch and Walter
Pitts in 1943. The McCulloch-Pitts neural model is also known as linear threshold gate. It
is a neuron of a set of inputs I1,I2,I3…Im and one output y . The linear threshold gate
simply classifies the set of inputs into two different classes. Thus the output y is binary.
Such a function can be described mathematically using these equations:
 W1,W2…Wm are weight values normalized
in the range of either (0,1) or (-1,1) and
associated with each input line, Sum is the
weighted sum, and T is a threshold constant.
The function f is a linear step function at
threshold T as shown in figure

A simple M-P neuron is shown in the figure.

It is excitatory with weight (w>0) / inhibitory with weight –p (p<0).
In the Fig., inputs from x1 to xn possess excitatory weighted connection and Xn+1 to xn+m has
inhibitory weighted interconnections.
Since the firing of neuron is based on threshold, activation function is defined as

For inhibition to be absolute, the threshold with the activation function should satisfy the
following condition:
θ >nw –p
Output will fire if it receives ―k or more excitatory inputs but no inhibitory inputs where
kw≥θ>(k-1) wθ>(k-1) w
- The M-P neuron has no particular training algorithm.
- An analysis is performed to determine the weights and the threshold.
- It is used as a building block where any function or phenomenon is modelled based on
a logic function.
Illustration Statement: Implement XOR function using MP model
Truth table for XOR function is:

X1 X2 Y
0 0 0

0 1 1
1 0 1

1 1 0

Activation function Yin is as follows:

Yin=x1w1+x2w2

As we know,

Let Z1= and Z2=

X1 X2 Z1
0 0 0
0 1 0

1 01
1 10

For Z1,

W11=1 and W12=-1

Θ=1
 X1 X2 Z2
0 0 0
0 1 1
1 0 0
1 1 0

For Z2,

W11=-1 and W12=1

Θ=1

Y=Z1+Z2

Z1 Z2 Y
0 0 0
0 1 1
1 0 1
1 1 0

For Y,

W11=1 and W12=1

Θ=1

Result/Conclusion:
Mc-Culloch pits Model is implemented for XOR function by using the thresholding
activation function. Activation of M-P neurons is binary (i.e) at any time step the neuron may
fire or may not fire. Threshold plays major role here.
Viva Questions:

• What are Neural Networks? What are the types of neural networks?

A neural network is a series of algorithms that endeavors to recognize underlying

relationships in a set of data through a process that mimics the way the human brain operates.
Types:
Feedforward Neural Network – Artificial Neuron.
Radial Basis Function Neural Network.
Multilayer Perceptron.
Convolutional Neural Network.

• How are Artificial Neural Networks different from Normal Computers?

Difference between traditional computers and artificial neural networks is the way in which
they function. While computers function logically with a set of rules and calculations,
artificial neural networks can function via images, pictures, and concepts

• What is simple Artificial Neuron?

An artificial neuron is a mathematical function conceived as a model of biological neurons, a
neural network. ... Usually each input is separately weighted, and the sum is passed through
a non-linear function known as an activation function or transfer function.

• What is meant by training of artificial neural networks?

Once a network has been structured for a particular application, that network is rained. To
start this process the initial weights are chosen randomly. Then, the training, or learning,
begins. There are two approaches to training - supervised and unsupervised.
 EXPERIMENT-5

Aim: Implementation of Single layer Perceptron Learning Algorithm.

Objectives:
• To become familiar with neural networks learning algorithms from available Illustrations.
• Provide knowledge of learning algorithm in neural networks.

Theory: Neural networks are a branch of ―Artificial Intelligence". Artificial Neural

Network is a system loosely modelled based on the human brain. Neural networks are a
powerful technique to solve many real world Illustrations. They have the ability to learn from
experience in order to improve their performance and to adapt themselves to changes in the
environment. In addition to that they are able to deal with incomplete information or noisy
data and can be very effective especially in situations where it is not possible to define the
rules or steps that lead to the solution of a Illustration. In a nutshell a Neural network can be
considered as a black box that is able to predict an output pattern when it recognizes a given
input pattern. Once trained, the neural network is able to recognize similarities when
presented with a new input pattern, resulting in a predicted output pattern.

Algorithm: The algorithm is as follows:

1. Initialize the weights and threshold to small random numbers.
2. Present a vector x to the neuron inputs and calculate the output.
3. Update the weights according to:
wj(t+1)= wj(t)+ η (d-y)x
where  d is the desired output,
• t is the iteration number, and
• eta is the gain or step size, where 0.0 < n < 1.0
4. Repeat steps 2 and 3 until:
1. the iteration error is less than a user-specified error threshold or
2. a predetermined number of iterations have been completed.

Result/Conclusion: Single layer perceptron learning algorithm is implemented for AND

function. It is used for train the iterations of neural network. Neural network mimics the
human brain and perceptron learning algorithm trains the neural network according to the
input given.
Viva Questions:
• What is feed forward network?
Deep feedforward networks, also often called feedforward neural networks,
or multilayer perceptrons(MLPs), are the quintessential deep learning models. The
goal of a feedforward network is to approximate some function f*. These models are
called feedforward because information flows through the function being evaluated
from x, through the intermediate computations used to define f, and finally to the
output y.
• Write the logistic sigmoid function? The
logistic sigmoid has the following form:

• Why use Artificial Neural Networks? What are its advantages?

ANNs have some key advantages that make them most suitable for certain
problems and situations:
1. ANNs have the ability to learn and model non-linear and complex
relationships, which is really important because in real-life, many of the
relationships between inputs and outputs are non-linear as well as complex.
2. ANNs can generalize — After learning from the initial inputs and their
relationships, it can infer unseen relationships on unseen data as well, thus
making the model generalize and predict on unseen data.
3. Unlike many other prediction techniques, ANN does not impose
any restrictions on the input variables.

• List some commercial practical applications of Artificial Neural Networks.

ANNs, due to some of its wonderful properties have many applications:
1. Image Processing and Character recognition.
2. Forecasting

• What are the disadvantages of Artificial Neural Networks?

1. Hardware dependence
2. Unexplained behavior of the network
3. Determination of proper network structure
 EXPERIMENT-6

Aim: Implementation of unsupervised learning algorithm – Hebbian Learning

Objectives:
• To become familiar with neural networks learning algorithms from available
Illustrations.
• To give design methodologies for artificial neural networks.
• Provide knowledge of un-supervised learning in neural networks.

Theory: Unsupervised Learning Algorithm:

These types of model are not provided with the correct results during the training. It can be
used to cluster the input data in classes on the basis of their statistical properties only. The
labelling can be carried out even if the labels are only available for a small number of objects
represented of the desired classes. All similar input patters are grouped together as clusters.
If matching pattern is not found, a new cluster is formed. In contrast to supervised learning,
unsupervised or self-organized learning does not require an external teacher. During the
training session, the neural network receives a number of different patterns & learns how to
classify input data into appropriate categories. Unsupervised learning tends to follow the
neuro-biological organization of brain. It aims to learn rapidly & can be used in real-time.
Hebian Learning:
Hebbian Learning is inspired by the biological neural weight adjustment mechanism.
It describes the method to convert a neuron an inability to learn and enables it to
develop cognition with response to external stimuli. These concepts are still the basis
for neural learning today.

Algorithm: The algorithm is as follows:

1. Set initial synaptic weights and thresholds to small random values, say in an interval [0,1].
2.Compute the neuron output at iteration p n
(p)=∑ (p) (p)-
i=1

Where n is number of neuron inputs, & is the threshold value of neuron j. 3.

Update the weights in the network
( + 1) = +∆ ( )

Where ( ) is the weight correction at iteration p.

∆ ( ) = (p)[λ ( ) − ( )] 4. Increase iteration p by one, go

back to step 2.
Result/Conclusion: Unsupervised Hebbian learning algorithm is implemented which
does not require supervisor. It update the weights the accordingly if error comes and train
the network.
Viva Questions:
• What is unsupervised and supervised training?
• In Supervised learning, you train the machine using data which is well
"labeled."

• Unsupervised learning is a machine learning technique, where you do not

need to supervise the model.

• How Artificial Neurons learns?

The smallest and most important unit of the artificial neural network is the neuron. As
in biological neural systems, these neurons are connected with each other and together
they have the great processing power. Every neuron has input connections and output
connections. These connections simulate the behavior of the synapses in the brain.
The same way that synapses in the brain transfer the signal from one neuron to
another, connections pass information between artificial neurons. These connections
have weights, meaning that value that is sent to every connection is multiplied by this
factor.
• What is the difference between neural network and fuzzy
logic? Fuzzy Logic
It is an extension of convention set theory (also called crisp set)
In conventional set logic true is 1(in general) and false is 0. But in fuzzy there
is no concept of exact true or exact false they are given membership.
Neural Network
It is nothing but trying to simulate our brain in electronic domain so that it
can learn like we do.

• What is Unsupervised Hebbian learning algorithm 

It is a linear feedforward neural network model for unsupervised learning with applications
primarily in principal components analysis.
 EXPERIMENT-7
Aim: Implementation Genetic Application – Match Word Finding.

Objectives:
• To familiarize with Mathematical foundations for Genetic algorithm, operator.
• To study the Applications of Genetic Algorithms.

Theory:
Genetic algorithm:
• Genetic algorithm is a search technique used in computing to find true or
approximate solutions to approximate solutions to optimization & search Illustrations.
• Genetic algorithms are inspired by Darwin's theory about evolution. Solution to
an Illustration solved by genetic algorithms is evolved.
• Algorithm is started with a set of solutions (represented by chromosomes) called
population. Solutions from one population are taken and used to form a new population.
This is motivated by a hope, that the new population will be better than the old one.
Solutions which are selected to form new solutions (offspring) are selected according to their
fitness - the more suitable they are the more chances they have to reproduce.
• This is repeated until some condition (for Illustration number of populations
or improvement of the best solution) is satisfied.

Outline of the Basic Genetic Algorithm

1. [Start] Generate random population of n chromosomes (suitable solutions for the
Illustration)
2. [Fitness] Evaluate the fitness f(x) of each chromosome x in the population
3. [New population] Create a new population by repeating following steps until the new
population is complete
1. [Selection] Select two parent chromosomes from a population according to
their fitness (the better fitness, the bigger chance to be selected)
2. [Crossover] With a crossover probability cross over the parents to form a new
offspring (children). If no crossover was performed, offspring is an exact copy of
parents.
3. [Mutation] With a mutation probability mutate new offspring at each
locus (position in chromosome).
4. [Accepting] Place new offspring in a new population
5. [Replace] Use new generated population for a further run of algorithm
6. [Test] If the end condition is satisfied, stop, and return the best solution in current
population
7. [Loop] Go to step 2
Flowchart
Illustration Statement:
Match Word Finding Here we try to guess a word from the given population of word.

Algorithm: Match Word Finding Algorithm:

Step 1: Select the word to be guessed
This value is taken through user input.
Step 2: Initialize the population
User inputs the population.
Step 3: Evaluate the population
Fitness is assigned based on number of correct letters in correct
place. Step 4: Select breeding population
Selection is done on the basis of fitness.
Step 5: Create new population
Population is created by using uniform crossover between
breeding populations.
Step 6: Check for stopping condition
Here maximum fitness value in population is checked. If it is 60%.
Step 7: If stopping condition is not true, goto Step 3;
Else return the offspring with highest fitness value.

Result/Conclusion:
The match word finding algorithm is implemented using the genetic algorithms which
include all the genetic algorithm operators. Genetic algorithm includes the selection,
crossover, mutation operators along with fitness function.
Viva Questions:
• Name some of the existing search methods.
1. Calculus based Search
2. Enumerative Search
3. Random Search

• What are the operators involved in a simple genetic algorithm?

There are three main types of operators which must work in conjunction with one another in
order for the algorithm to be successful –
1. Mutation
2. Crossover
3. Selection

• What is reproduction?
Genetic algorithm reproduction methods for distribution system loss reduction and load
balancing problems. Selected fittest parents create new child.

• What is crossover?
Crossover is a genetic operator used to combine the genetic information of two parents
to generate new offspring.
 EXPERIMENT-8
Aim: Study of ANFIS Architecture.

Objectives:
• Study of hybrid systems.
• Prepare the students for developing intelligent system.

Theory:
The adaptive network-based fuzzy inference systems (ANFIS) is used to solve Illustrations
related to parameter identification. This parameter identification is done through a hybrid
learning rule combining the back-propagation gradient descent and a least-squares method.

ANFIS is basically a graphical network representation of Sugeno-type fuzzy systems

endowed with the neural learning capabilities. The network is comprised of nodes with
specific functions collected in layers. ANFIS is able to construct a network realization of
IF / THEN rules.

Consider a Sugeno type of fuzzy system having the rule base

1. If x is A1 and y is B1, then f1= c11x+c12y+c10
2. If x is A2 and y is B2, then f2= c21x+c22y+c20

Let the membership functions of fuzzy sets Ai, Bi, i=1,2, be , AiBi.
In evaluating the rules, choose product for T-norm (logical and).
1. Evaluating the rule premises results in

2. Evaluating the implication and the rule consequences gives

Or leaving the arguments out

This can be separated to phases by first defining

Then f can be written as

All computations can be presented in a diagram form. ANFIS normally

has 5 layers of neurons of which neurons in the same layer are of the
same function family.
 Layer 1 (L1): Each node generates the membership grades of a linguistic label.

An Illustration of a membership function is the generalised bell function: where {a, b,

c} is the parameter set. As the values of the parameters change, the shape of the bell-
shaped function varies. Parameters in that layer are called premise parameters.

Layer 2 (L2): Each node calculates the firing strength of each rule using the min
or prod operator. In general, any other fuzzy AND operation can be used.

Layer 3 (L3): The nodes calculate the ratios of the rule are firing strength to the sum
of all the rules firing strength. The result is a normalised firing strength.

Layer 4 (L4): The nodes compute a parameter function on the layer 3 output.
Parameters in this layer are called consequent parameters.

Layer 5 (L5): Normally a single node that aggregates the overall output as
the summation of all incoming signals.

Algorithm:
When the premise parameters are fixed, the overall output is a linear combination of the
consequent parameters. In symbols, the output f can be written as which is linear in the
consequent parameters cij (i = 1,2¸ j = 0,1,2). A hybrid algorithm adjusts the consequent
parameters cij in a forward pass and the premise parameters {ai, bi, ci} in a backward pass
(Jang et al., 1997). In the forward pass the network inputs propagate forward until layer 4,
where the consequent parameters are identified by the least-squares method. In the
backward pass, the error signals propagate backwards and the premise parameters are
updated by gradient descent.

Because the update rules for the premise and consequent parameters are decoupled in
the hybrid learning rule, a computational speedup may be possible by using variants of
the gradient method or other optimisation techniques on the premise parameters.

Result/Conclusion:
This study experiments describe the architecture of neuro fuzzy systems. Fuzzy rule
based system includes the model like sugenor type fuzzy which is having neural learning
capabilities.
Viva Questions:
• What are the hybrid systems?
A hybrid system is a dynamical system that exhibits both continuous and discrete dynamic
behaviour – a system that can both flow (described by a differential equation) and jump
(described by a state machine or automaton)

• What are fuzzy inference systems?

Fuzzy Inference Systems take inputs and process them based on the pre specified rules
to produce the outputs. Both the inputs and outputs are real valued, whereas the internal
processing is based on fuzzy rules and fuzzy arithmetic.

• How neuro fuzzy inference systems work?

A neuro-fuzzy system is based on a fuzzy system which is trained by a learning algorithm
derived from neural network theory. The learning procedure of a neuro-fuzzy system takes
the semantical properties of the underlying fuzzy system into account. This results in
constraints on the possible modifications applicable to the system parameters.

• What is ANFIS architecture?

An adaptive neuro-fuzzy inference system or adaptive network-based fuzzy inference system
(ANFIS) is a kind of artificial neural network that is based on Takagi–Sugeno fuzzy
inference system. Since it integrates both neural networks and fuzzy logic principles, it has
potential to capture the benefits of both in a single framework. Its inference system
corresponds to a set of fuzzy IF–THEN rules that have learning capability to approximate
nonlinear functions. Hence, ANFIS is considered to be a universal estimator.
 EXPERIMENT-9

Aim: Study of Derivative-free Optimization.

Objectives: From this experiment, the student will be able to,

• Study of hybrid systems.

• Prepare the students for developing intelligent

system. The learner will be able to,

• Aware of the use of neuro fuzzy inference systems in the design of intelligent
or humanistic systems.
• To become knowledgeable about neuro fuzzy inference systems.
• An ability to apply knowledge of computing and use of current computing
techniques appropriate to the discipline.

Theory:

Optimization Illustrations defined by functions for which derivatives are unavailable or

available at a prohibitive cost are appearing more and more frequently in computational
science and engineering. Increasing complexity in mathematical modelling, higher
sophistication of scientific computing, and abundance of legacy codes are some of the
reasons why derivative-free optimization is currently an area of great demand.
Difficulties and challenges arise from multiple sources: expensive function evaluation,
black-box/legacy codes, noise and uncertainty, unknown a priori function domains, and
hidden constraints.

• Derivative free Optimization

Derivative free optimization is a subject of mathematical optimization. It may refer to

Illustrations for which derivative information is unavailable, unreliable or impractical to
obtain (derivative-free optimization Illustrations), or methods that do not use derivatives
(derivative-free optimization methods)

• Derivative free Optimization algorithm

 O Genetic algorithms (GA) O
Simulated Annealing (SA)

Genetic algorithms (GA)

In the field of artificial intelligence, a genetic algorithm (GA) is a search heuristic that
mimics the process of natural selection. This heuristic (also sometimes called a
metaheuristic) is routinely used to generate useful solutions to optimization and search
Illustrations.[1] Genetic algorithms belong to the larger class of evolutionary algorithms
(EA), which generate solutions to optimization Illustrations using techniques inspired by
natural evolution, such as inheritance, mutation, selection and crossover.

Optimization Illustrations

In a genetic algorithm, a population of candidate solutions (called individuals, creatures,

or phenotypes) to an optimization Illustration is evolved toward better solutions. Each
candidate solution has a set of properties (its chromosomes or genotype) which can be
mutated and altered; traditionally, solutions are represented in binary as strings of 0s and
1s, but other encodings are also possible.[2]

The evolution usually starts from a population of randomly generated individuals, and is
an iterative process, with the population in each iteration called a generation. In each
generation, the fitness of every individual in the population is evaluated; the fitness is
usually the value of the objective function in the optimization Illustration being solved.
The more fit individuals are stochastically selected from the current population, and each
individual's genome is modified (recombined and possibly randomly mutated) to form a
new generation. The new generation of candidate solutions is then used in the next
iteration of the algorithm. Commonly, the algorithm terminates when either a maximum
number of generations has been produced, or a satisfactory fitness level has been reached
for the population.

A typical genetic algorithm requires:

a genetic representation of the solution domain,

a fitness function to evaluate the solution domain.
A standard representation of each candidate solution is as an array of bits.[2] Arrays of
other types and structures can be used in essentially the same way. The main property
that makes these genetic representations convenient is that their parts are easily aligned
due to their fixed size, which facilitates simple crossover operations. Variable length
representations may also be used, but crossover implementation is more complex in this
case. Tree-like representations are explored in genetic programming and graph-form
representations are explored in evolutionary programming; a mix of both linear
chromosomes and trees is explored in gene expression programming.

Once the genetic representation and the fitness function are defined, a GA proceeds to
initialize a population of solutions and then to improve it through repetitive application of
the mutation, crossover, inversion and selection operators.

Initialization

The population size depends on the nature of the Illustration, but typically contains
several hundreds or thousands of possible solutions. Often, the initial population is
generated randomly, allowing the entire range of possible solutions (the search space).
Occasionally, the solutions may be "seeded" in areas where optimal solutions are likely to
be found.

Selection

During each successive generation, a proportion of the existing population is selected to

breed a new generation. Individual solutions are selected through a fitness-based process,
where fitter solutions (as measured by a fitness function) are typically more likely to be
selected. Certain selection methods rate the fitness of each solution and preferentially
select the best solutions. Other methods rate only a random sample of the population, as
the former process may be very time-consuming.

The fitness function is defined over the genetic representation and measures the quality of
the represented solution. The fitness function is always Illustration dependent. For
instance, in the knapsack Illustration one wants to maximize the total value of objects that
can be put in a knapsack of some fixed capacity. A representation of a solution might be
an array of bits, where each bit represents a different object, and the value of the bit (0 or
1) represents whether or not the object is in the knapsack. Not every such representation
is valid, as the size of objects may exceed the capacity of the knapsack. The fitness of the
solution is the sum of values of all objects in the knapsack if the representation is valid,
or 0 otherwise.

In some Illustrations, it is hard or even impossible to define the fitness expression; in

these cases, a simulation may be used to determine the fitness function value of a
phenotype (e.g. computational fluid dynamics is used to determine the air resistance of a
vehicle
whose shape is encoded as the phenotype), or even interactive genetic algorithms are
used.

Simulated annealing (SA)

is a probabilistic technique for approximating the global optimum of a given function.

Specifically, it is a metaheuristic to approximate global optimization in a large search
space. It is often used when the search space is discrete (e.g., all tours that visit a given
set of cities). For Illustrations where finding the precise global optimum is less important
than finding an acceptable local optimum in a fixed amount of time, simulated annealing
may be preferable to alternatives such as brute-force search or gradient descent.

Simulated annealing interprets slow cooling as a slow decrease in the probability of

accepting worse solutions as it explores the solution space. Accepting worse solutions is a
fundamental property of metaheuristics because it allows for a more extensive search for
the optimal solution.

Terminology:

1. Objective function E(x): function to be optiimized

2. Move set: set of next points to explore
3. Generating function: to select next point
4. Acceptance function h(DE, T): to determine if the selected point should be
accept or not. Usually h(DE, T) = 1/(1+exp(DE/(cT)).
5. Annealing (cooling) schedule: schedule for reducing the temperature T
Flowchart

Result/Conclusion:

This study experiments describe the various techniques used for derivative
free optimization. It also describes how to use optimization techniques in soft
computing domain.
Viva Questions:

• What is the meaning of optimization?

In computing, optimization is the process of modifying a system to make some
features of it work more efficiently or use fewer resources. For instance, a
computer program may be optimized so that it runs faster, or to run with less
memory requirement or other resources ,or to consume less energy .

• What do you mean by simulated annealing?

Simulated annealing is a method for solving unconstrained and bound-constrained

optimization problems. The method models the physical process of heating a
material and then slowly lowering the temperature to decrease defects, thus
minimizing the system energy.

• What are the algorithms used for derivative free optimization?

Algorithm used for derivative free

optimization--Genetic algorithm
-Simulated annealing

• What are the algorithms used for derivative based optimization?

Algorithm used for derivative based optimization-

-Steepest Descent: Determine search direction according to an
objective function’s derivative information
-Newton’s Method: Determine search direction according to an
objective function’s second derivative
-Peak functions and Peak derivative
 EXPERIMENT-10

Aim: Study of research paper on Soft Computing and give a review report consists
of abstract, introduction, state of the art, methodology, results, conclusion, reference.

Objectives: From this experiment, the student will be able to,

• Understand the research trends in soft computing.

• Crate awareness among the students towards the recent trends in soft
computing.

The learner will be able to,

• Understand the importance of research in soft computing.

• To become knowledgeable for developing soft computing applications.

• An ability to match the industry requirements in the domains of Programming

with the required management skills.to analyze the local and global impact
of computing on individuals, organizations, and society.

Theory:

Students can find the research papers based on the artificial neural
network, hybrid systems, genetic algorithm, fuzzy system, fuzzy logic,
fuzzy inference system etc.
Students need to search recent papers on any of the above mentioned topics,
study it and prepare presentation on the same

Result/Conclusion:
Through this experiment, we have understood the recent advancements and
applications of various subdomains of soft computing.

References:
1. www.ieeeexplore.com
2. www.Scicencedirect.com
3. Any open access journal