Soft Computing Question Bank

DR.
SIVANTHI ADIATANR COLLEGE OF ENGINEERING
TIRUCHENDUR
Department of Electrical and Electronics Engineering
QUESTION BANK
M.E.Power Electronics and Drives
SOFT COMPUTING TECHNIQUES CL72014
UNIT I : INTRODUCTION AND ARTIFICIAL NEURAL NETWORKS
Two Marks :
1. Define soft computing
In computer science, soft computing is the use of inexact solutions to computationally

hard tasks such as the solution of NP-complete problems, for which there is no known
algorithm that can compute an exact solution in polynomial time.
2. Differentiate hard and soft computing
Hard computing, i.e., conventional computing, requires a precisely stated analytical

model and often a lot of computation time. Soft computing differs from conventional
(hard) computing in that, unlike hard computing, it is tolerant of imprecision, uncertainty,
partial truth, and approximation. In effect, the role model for soft computing is the
human mind.
3. Name some soft computing techniques.
Neural Networks, Support Vector Machine, Fuzzy Logic, Evolutionary Algorithm,

Genetic Algorithm, Machine Learning (ML) and Probabilistic Reasoning (PR), Ant
colony optimization, Particle swarm optimization, cuckoo search algorithm
4. Define artificial neural network.
Artificial neural network (ANN) is an information processing system which is a

mathematical model of human brain system. ANNs are considered nonlinear statistical
data modeling tools where the complex relationships between inputs and outputs are
modeled or patterns are found
1
5. Draw the model of Artificial neuron
6. What is neural network architecture? Explain logistic sigmoid function with

example?
A neural network is characterized by its pattern of connections between the neurons called its
architecture. f(x) = 1/(1+exp(-x)).
7. Explain important characteristics and applications of artificial neural network?
Fault tolerance.Signal processing, Control, pattern recognition, Medicine, Speech

production, Speech recognition and Business etc.
8. Name some neural network activation functions
9. What is the classification of training? Explain supervised training?

Training is classified into Supervised and unsupervised training. Training is accomplished by
presenting a sequence of training vectors or patterns, each with an associated target output
2
vector. The weights are adjusted according to the learning algorithm. This process is known
as supervised training.
10. What is unsupervised training?
A sequence of input vector is provided, but no target vectors are specified. The net modifies
the weights so that the most similar input vectors are assigned to the same output unit.
11. Define identity function. Define binary step function with example ?
f(x) = x for all x.

For a given threshold T, the binary step function is:
f(x) = 1 if x>= T
f(x) = 0 if x< T
12. Define delta rule? Explain significance of delta rule in defining the weights?
The weights represent information being used by the net to solve problem. The delta rule
changes the weights of the neural connections so as to minimize the difference between the
net input to the output unit and the target value.
13. What is feedforward neural network. Draw its structure.
A feedforward neural network is an artificial neural network where connections between

the units do not form a directed cycle. In this network, the information moves in only one
direction, forward, from the input nodes, through the hidden nodes (if any) and to the output
nodes.
BIG QUESTIONS:
1. What is perceptron? Explain different layers of perceptron?
Book: Fundamentals of Neural Network by Laurene Fausett

Pages:60-76
3
2. Draw the structure and algorithm of ADALINE and MADALINE net.

Pages:81-88
3. Draw the structure and explain the algorithm of back propagation net.

Pages:290-300
UNIT II – ARTIFICIAL NEURAL NETWORKS
Two Marks:
1. What are the different types of neural network
MccullochPitts Neural network, Perceptron, Adaline, Madaline, Counter Propagation

Network, Hopfield Network, Radial Basis Function Network, Adaptive Resonant Theory
(ART) network, Probabilistic Neural Network, Self Organising Map, Learning Vector
Quantization etc
2. Give examples for supervised and unsupervised learning neural networks
Supervised Learning : Multi Layer Pereptron, Adaline,Madaline, Feed forward

backpropagation network, Radial Basis Function Network, Probabilistic Neural Network
Unsupervised Learning : Self Organising Map, Learning Vector Quantization, Counter

propagation network, Adaptive Resonance theory, Hidden Markov Model etc
3. What is the concept of hebbian learning?
Hebb proposed that learning occurs by modification of the synapse strengths (weights) in
a manner such that if two interconnected neurons are both on at the same time, then the
weights between these neurons should be increased.
4. What is a counter propagation network (CPN)?
The counterpropagation network is a hybrid unsupervised learning network. It consists of

an outstar network and a competitive filter network. It was developed in 1986 by Robert
Hecht-Nielsen. It is guaranteed to find the correct weights, unlike regular back
propagation networks that can become trapped in local minimums during training.
5. Define an ART network
Adaptive resonance theory (ART) is a theory developed by Stephen Grossberg and Gail
Carpenter on aspects of how the brain processes information. It describes a number of neural
network models which use supervised and unsupervised learning methods, and address
4
problems such as pattern recognition and prediction. It typically consists of a comparison
field and a recognition field composed of neurons, a vigilance parameter (threshold of
recognition), and a reset module.
6. Draw the structure of ART network
7. State the characteristics of counter propagation network (CPN)
o CPN is a multilayer network based on combination of input, clustering

and output layers
o CPN is trained in two stages
o CPN can be used to compress data, approximate patterns or to associate
patterns
o CPN approximates training vector pairs by constructing lookup table
8. What is the feature of boltzmann machine?
A Boltzmann machine is a network of symmetrically connected, neuron-like units that

make stochastic decisions about whether to be on or off. Boltzmann machines have a
simple learning algorithm (Hinton & Sejnowski, 1983) that allows them to discover
interesting features that represent complex regularities in the training data.
9. Differentiate ART1 and ART2
Art 1 is designed to work for continuous valued input vectors while ART1 is used for
binary valued input vectors. ART2 requires preprocessing of input vectors but ART1
does not require it. ART2 treats small components as noise and does not distinguish
between patterns that are merely scaled varieties of each other.
10. What is associative memory? Explain its types?
A neural net that is trained to associate a set of input vectors with a corresponding set of
output vectors is called associative memory. Associative memory neural networks are
nets in which the weights are determined in such a way that the net can store a set of P
pattern associations. If the desired output vector is same as the input vector, the net is an
auto associative memory. If the desired output vector is different from the input vector,
the net is a hetero associative memory.
5
11. Differentiate Hopfield and Boltzmann machine
Hopfield network is a deterministic neural network whereas Boltzmann is a stochastic

network. Boltzmann machine is not used for solving practical problems due to its
unconstrained connectivity.
12. State the properties of neural networks
 Adaptivity
Adapt weights to environment and retrained easily
 Generalization ability
Ability to work for unseen data
 Fault tolerance
Able to work even with partial and noisy data
BIG QUESTIONS
1. With neat diagram explain Hopfield/Recurrent network

Pages:135-140
2. Explain Counter Propagation Algorithm

Pages:196-206
3. Write in detail about ART network

Pages:222-243
4. Explain Boltzmann machine algorithm

Pages:338-347
UNIT III FUZZY LOGIC SYSTEM
Two Marks:
1. What is meant by fuzzy logic?
Fuzzy logic is a form of many-valued logic which deals with reasoning that is
approximate rather than fixed and exact. Compared to traditional binary sets (where
variables may take on true or false values), fuzzy logic variables may have a truth
6
value that ranges in degree between 0 and 1. Fuzzy logic has been extended to handle
the concept of partial truth, where the truth value may range between completely true
and completely false
2. Differentiate classical and fuzzy set
A classical set is defined by crisp boundaries. i.e there is no uncertainty in the

prescription or location of the boundaries of the set. The boundaries of fuzzy set are
ambiguously specified. In crisp set an element either belongs to the set or it does not
belong to it. In fuzzy set an element belongs to one or more sets.
3. When is fuzzy logic used?
Fuzzy logic can be used where the system cannot be modelled by mathematical
equation and where human knowledge and perception needs to be involved.
4. What are the fuzzy set operations?
 Fuzzy Complement
 Fuzzy Union
 Fuzzy Intersection
 Combinations of Operations
 General Aggregation Operations
5. Define fuzzy Union and Intersection operation
UNION:
The membership function of the Union of two fuzzy sets A and B with membership
functions µA and µB respectively is defined as the maximum of the two individual
membership functions. This is called the maximum criterion.
INTERSECTION:
The membership function of the Intersection of two fuzzy sets A and B with
membership functions µA and µB respectively is defined as the minimum of the two
individual membership functions. This is called the minimum criterion.
6. Define Fuzzy complement operation
The membership function of the Complement of a Fuzzy set A with membership

function µA is defined as the negation of the specified membership function. This is
called the negation criterion.
7
7. State the properties of fuzzy set
8. Define linguistic variable
A linguistic variable is defined as a variable whose values are sentences in a natural

or artificial language. Thus, if tall, not tall, very tall, very very tall, etc. are values of
HEIGHT, then HEIGHT is a linguistic variable.
9. What is meant by fuzzy conditional statements and fuzzy algorithm?
Fuzzy conditional statements are expressions of the form IF A THEN B, where A

and B have fuzzy meaning, e.g., IF x is small THEN y is large, where small and large
are viewed as labels of fuzzy sets.
A fuzzy algorithm is an ordered sequence of instructions which may contain fuzzy
assignment and conditional statements, e.g., x =very small, IF x is small THEN y is
large. The execution of such instructions is governed by the compositional rule of
inference and the rule of the preponderant alternative.
10. Define fuzzy normal and subnormal set
A fuzzy subset of X is called normal if there exists at least one element χ∈X such that
µA (χ)=1. A fuzzy subset that is not normal is called subnormal.
11. Define height of a fuzzy subset
The height of a fuzzy subset A is the largest membership grade of an element in A

height(A) = max χ(µA(χ))
8
12. Define support and core in fuzzy set
Assume A is a fuzzy subset of X; the support of A is the crisp subset of X whose

elements all have non-zero membership grades in A:supp(A) = {x|µA(x) > 0 and
x∈X}
Assume A is a fuzzy subset of X; the core of A is the crisp subset of X consisting of
all elements with membership grade 1:Core(A) = {x|µA(x) = 1 and x∈X}
13. Name the fuzzy membership functions
 Triangular
 Trapezoidal
 Gaussian
 Generalized Bell
14. Define cartesian or cross product
The cartesian or cross product of fuzzy subsets A and B, of sets X and Y respectively
is denoted as A × B
This cross product relationship T on the set X × Y is denoted as T = A × B
15. Define cardinality of fuzzy set
The number of members of a finite discrete fuzzy set A is called cardinality of fuzzy
set A and is denoted by |A|
16. Define α-cut of fuzzy set
That is, the alpha-cut (or the strong alpha-cut) of a fuzzy set A is the crips set aA (or the
crisp set a+A) that contains all the elements of the universal set X whose membership
grades in A are greater than or equal to (or only greater than) the specified value of
alpha." aA and a+A are crisp sets because a particular value x either is or isn't in the set;
there is no partial membership.
17. What are the steps in fuzzy logic
 Specify the problem and define linguistic variables.

 Determine fuzzy sets.
 Elicit and construct fuzzy rules.
 Fuzzy Rules Evaluation
 Aggregation of rule outputs
9
 Defuzzification
18. Name the defuzzification methods
 Max-membership principle
 Centroid method
 Weighted average method
 Mean-max membership
 Centre of sums
 Centre of largest area
BIG QUESTIONS :
1. Explain fuzzy set operation and fuzzy properties

Text Book: Fuzzy Logic with Engineering Applications by Timothy J.Ross
Page : 17-27
2. Write in detail about the fuzzification process
Page : 90-91
3. Explain in detail the defuzzification methods

Page : 136-137
4. Explain fuzzy modelling and control schemes
Page : 470-474
UNIT IV – GENETIC ALGORITHM
TWO MARKS:
1. What is genetic algorithm?

Genetic algorithms are search algorithms based on the mechanics of natural selection and
genetics. As such they represent an intelligent exploitation of a random search used to solve
optimization problems. The genetic algorithm repeatedly modifies a population of individual
solutions. At each step, the genetic algorithm selects individuals at random from the current
population to be parents and uses them to produce the children for the next generation. Over
successive generations, the population "evolves" toward an optimal solution.
2. Name the three operators of genetic algorithm
 Reproduction
 Crossover
 Mutation
10
Reproduction is a process in which individual strings are copied according to their objective
function values f.
3. Define fitness function

The fitness function is the function to be optimized. For standard optimization algorithms, this is
known as the objective function.
4. Define individual and population

An individual is any point to which you can apply the fitness function. The value of the fitness
function for an individual is its score. An individual is sometimes referred to as a genome and
the vector entries of an individual as genes.
A population is an array of individuals. At each iteration, the genetic algorithm performs a series
of computations on the current population to produce a new population. Each successive
population is called a new generation.
5. What are the components of genetic algorithm?
Encoding technique - (gene, chromosome)

Initialization - ( population creation)
Evaluation function - (environment)
Selection of parents - (reproduction)
Genetic operators - (mutation, recombination)
Parameter settings - ( By practice)
6. Define cross over

The process of combining two individuals or parents in the current population to create new
individuals for possible inclusion in next generation. It is the main operator for local search. At
each coordinate of the child vector, the default crossover function randomly selects an entry, or
gene, at the same coordinate from one of the two parents and assigns it to the child.
7. Define mutation
Each component of every individual is modified with the probability pm to have random walk
over the search space. i.e random bits of genes are changed from 1 to 0 and vice versa. It is the
main operator for global search. It helps to restore lost information to the population.
8. Define diversity
Diversity refers to the average distance between individuals in a population. A population has
high diversity if the average distance is large; otherwise it has low diversity.
9. Define elite children
Elite children are the individuals in the current generation with the best fitness values. These
individuals automatically survive to the next generation.
11
10. Name the stopping criteria used in genetic algorithm
Generations — The algorithm stops when the number of generations reaches the value of
Generations.
Time limit — The algorithm stops after running for an amount of time in seconds equal to Time
limit.
Fitness limit — The algorithm stops when the value of the fitness function for the best point in
the current population is less than or equal to Fitness limit.
Stall generations — The algorithm stops when the weighted average change in the fitness
function value over Stall generations is less than Function tolerance.
Stall time limit — The algorithm stops if there is no improvement in the objective function
during an interval of time in seconds equal to Stall time limit.
Function Tolerance — The algorithm runs until the weighted average change in the fitness
function value over Stall generations is less than Function tolerance
11. What is meant by tabu search?
Tabu search is a higher level heuristic procedure for solving optimization problems, designed to
guide other methods to escape the trap of local optimality.
12. What is ant colony optimization algorithm?
The ant colony optimization algorithm (ACO) is a probabilistic technique for solving
computational problems which can be reduced to finding good paths through graphs. This
algorithm is a member of the ant colony algorithms family, in swarm intelligence methods, and
it constitutes some meta heuristic optimizations. It is based on the behavior of ants seeking a
path between their colony and a source of food.
BIG QUESTIONS:
1. Explain genetic algorithm in detail
Book : Genetic Algorithms by David E.Goldberg
Pages: 61-67
2. Expalin with an example the adjustment of genetic algorithm parameters
Pages: 97-103
3. Explain solution of typical control problems using GA

Pages: 125-135
12
UNIT V APPLICATIONS
TWO MARKS:
1. What are the applications of fuzzy logic?
• Consumer products, Industrial systems

• Automotive and Decision analysis
• Medicine, Geology
• Pattern recognition
• Robotics
• Computational theory of perceptions
• Natural language processing
• Financial engineering
• Biomedicine and Forecasting
2. What are the applications of genetic algorithm

 Control, designing
 Scheduling
 Robotics
 Machine learning
 Signal Processing
3. Name some applications of soft computing
• Automotive systems and Manufacturing

• Bioinformatics and biomedicine
• Handwriting recognition, speech recognition
• Signal, Image processing and data compression
• Architecture design
• decision-support systems
• Power systems
• Neurofuzzy systems
• Fuzzy logic control
• Weather forecasting, Stock market prediction
• Electrical load forecasting
4. Name some power system optimization problems
Preparation and planning

Prioritizing investments in distribution network
Optimal protection and switching device placement
Generation scheduling
Maintenance scheduling
Power mix planning
Operative control
13
Constrained load flow
Power plant operation optimizer
Unit commitment – economic dispatch
Optimal power flow
FACTS (Flexible AC Transmission System) control
Voltage/VAr and loss reduction
Dynamic load modeling
Short-Term load forecast
Network reconfiguration and load reduction
5. What is meant by stability in power system?

Stability of a power system is its ability to return to normal or stable operating conditions
after having been subjected to some form of disturbance. Conversely, instability means a
condition denoting loss of synchronism or falling out of step.
6. Define linear dynamic system

Linear dynamical systems are dynamical systems whose evaluation functions are linear.
While dynamical systems in general do not have closed-form solutions, linear dynamical
systems can be solved exactly, and they have a rich set of mathematical properties.
Linear systems can also be used to understand the qualitative behavior of general
dynamical systems, by calculating the equilibrium points of the system and
approximating it as a linear system around each such point.
7. Name some optimization techniques used in power system
 Genetic algorithm
 Particle swarm optimization
 Ant Colony optimization
 Cuckoo search algorithm
 Simulated annealing
8. What are the features of fuzzy logic control system
 Fixed and uniform input output scaling factors
 Flat single partition rule base with fixed and noninteractive rules
 Fixed membership functions
 Limited rules
 Low level control
9. What are the three categories of tools in fuzzy logic toolbox?

The toolbox provides three categories of tools:
 Command line functions

 Graphical interactive tools
 Simulink blocks and examples
14
10. Name some fuzzy application areas in power system
 Contingency analysis
 Diagnosis/monitoring
 Distribution planning
 Load frequency control
 Generator maintenance scheduling
 Generation dispatch
 Load flow computations
 Load forecasting
 Load management
 Reactive power/voltage control
 Security assessment
 Stabilization control (PSS)
 Unit commitment
BIG QUESTIONS
1. Explain the application of genetic algorithm for power system

optimization problem
Pages: 61-67
2. Explain fuzzy logic controller with an example

Page : 475-477
3. Write in detail about the stability analysis of fuzzy control systems

Page : 477-483
15
ANNA UNIVERSITY QUESTION PAPER
MAY/JUNE 2015
PART A
1. Distinguish between artificial neuron and biological neuron
Artificial Neuron Biological neuron

The individual computational elements A biological neuron is a single nerve cell
that make up most artificial neural system in a biological neural network like brain
models are called artificial neurons
Each neuron sums up the input signals Each neuron sums up the input signals
and the output of neuron varies depending and the output is binary ie either ON or
upon the input and the activation function OFF depending on the threshold value
applied
The connections varies depending on the The connection is very complicated
architecture and it is not complicated to
make processing easier
2. What are the factors affecting the back propagation training

 Amount of training data
 Number of hidden layers and number of neurons
 Initialization of weights
 Learning rate parameters
3. Distinguish between recurrent and non recurrent networks
Recurrent Network Non Recurrent Network
Feed forward network with atleast one Feed forward network without any
feedback connection feedback connection
Activations around a loop No any looping activation
Perform temporal processing No temporal processing
Have some form of memory and perform No memory
non linear mapping
4. What is the purpose of Hopfield network? Give an example

Hopfield is a deterministic neural network with fully connected neurons. The idea behind
the Hopfield network is that patterns are stored in the weight matrix. The input must
contain part of these patterns. The dynamics of the network then retrieve the patterns
stored in the weight matrix. This is called Content Addressable Memory (CAM). The
network can also be used for auto-association. Generally they are used for : – Associative
memories : the network is able to memorize some states, patterns. – Combinatorial
optimization : if the problem is correctly modelled , the network can give some minimum
and some solution but can rarely find the optimal solution. Example: Travelling salesman
problem
16
5. What is approximate reasoning?
By approximate reasoning, we mean a type of reasoning which is neither very exact nor
very inexact. it is the process or processes by which a possible imprecise conclusion is
deduced from a collection of imprecise premises. In a simplest way, we can say that,
fuzzy logic plays a key role in approximate reasoning.
6. What are the methods of defuzzification?
 Max-membership principle
 Centroid method
 Weighted average method
 Mean-max membership
 Centre of sums
 Centre of largest area
7. What are the basic genetic algorithm operators?

 Mutation
 Cross over
 Elitism
8. What is Roulette Wheel selection in GA?
Fitness proportionate selection, also known as roulette wheel selection, is a genetic

operator used in genetic algorithms for selecting potentially useful solutions for
recombination. In fitness proportionate selection, as in all selection methods, the fitness
function assigns a fitness to possible solutions or chromosomes. This fitness level is used to
associate a probability of selection with each individual chromosome. If 𝑓𝑖 is the fitness of
𝑓
individual 𝑖 in the population, its probability of being selected is 𝑝𝑖 = ∑𝑁 𝑖 𝑓
𝑗=1 𝑗
where 𝑁 is the number of individuals in the population. This could be imagined similar to a
Roulette wheel in a casino. Usually a proportion of the wheel is assigned to each of the
possible selections based on their fitness value. This could be achieved by dividing the
fitness of a selection by the total fitness of all the selections, thereby normalizing them to 1.
Then a random selection is made similar to how the roulette wheel is rotated.
9. What are the applications of GA in Power Systems

 Economic load despatch problem
 Stability analysis
 Optimal Power flow analysis
 Generator scheduling
 Cost minimization in wins power generation and loss minimization
 FACTS controller
 Hydro thermal coordination problem
 Reactive power optimization
 Sensor network optimization problem
17
10. What are the applications of neural network in Power systems?
 Contingency analysis
 Diagnosis/monitoring
 Distribution planning
 Load frequency control
 Generator maintenance scheduling
 Generation dispatch
 Load flow computations
 Load forecasting
 Load management
 Reactive power/voltage control
 Security assessment
 Stabilization control (PSS)
 Unit commitment
PART B
11. (a) (i) Explain with a neat diagram, the neural network architecture of multilayer feed
forward network (10)

Pages:52-56
(ii) Draw and describe the structure of a biological neuron (6)

Pages:20-24
(b) (i) Draw and describe the different activation functions used in ANN (8)
(ii) Explain the various steps of back propagation algorithm used in training ANN(8)
18
Pages:290-306
12. (a) (i) Develop and describe with a neat diagram the Counter Propagation network
learning algorithm (8)

Pages:196-206
(ii) Describe with a neat diagram the architecture of recurrent network to perform XOR
task with two inputs (8)
Pages:135-140
(b) (i) Describe the structure of an Associative memory (6)

Pages:212-223
(ii) Explain Adaptive Resonance theory with an example (10)

Pages:222-243
13. (a) (i) Explain with a neat block diagram, the various components of a Fuzzy Logic
system (10)
Book: Fuzzy Logic with Engineering Applications by Timothy J.Ross

Page : 90-91
(ii) Describe shortly on Centroid method
Book: Fuzzy Logic with Engineering Applications by Timothy J.Ross

Page : 136-137
(b) (i) Explain the Self organising Fuzzy Logic control scheme with a suitable example
(8)
Page : 470-474
(ii) Explain the implementation of Fuzzy Logic control for nonlinear time delay system
(8)
Page : 475-483
19
14. (a) (i) Describe the basic steps of Genetic Algorithm used for solving optimization
problems (10)

Pages: 61-67
(ii) Compare the features of Genetic Algorithm with other optimization techniques(6)

Pages: 125-136
(b) (i) Describe the concept of tabu search method for solving optimization problems(8)

Pages: 114-122
(ii) Describe the concept of ant colony search method for solving optimization problems

Pages: 106-114
15. (a) Explain the case study of GA application to solve Power system optimization
problems (16)

Pages: 125-135
(b) Explain the case study of the application of Neural network for stability analysis of
interconnected systems (16)

Pages:310-320
ANNA UNIVERSITY QUESTION PAPER
MAY/JUNE 2016
ANSWER ALL QUESTIONS
PART A
1. Compare soft computing vs hard computing

20
human mind.
2. What are supervised learning and unsupervised learning?

Learning the input vectors in the presence of target vector is called supervised learning.
In the case of unsupervised learning, a sequence of input vector is provided, but no target
vectors are specified. The net modifies the weights so that the most similar input vectors
are assigned to the same output unit.
3. Why Hopfield network is called as recurrent neural network

Hopfield is a deterministic neural network with fully connected neurons. Since it
contains feedback connection it is a recurrent network. The idea behind the Hopfield
network is that patterns are stored in the weight matrix. The input must contain part of
these patterns. The dynamics of the network then retrieve the patterns stored in the
weight matrix. This is called Content Addressable Memory (CAM). The network can
also be used for auto-association. Since Hopfield network has some form of memory and
perform non linear mapping it is called a recurrent network
4. What are the properties of adaptive resonance theory?

Adaptive Resonance Networks are plastic and stable. These nets achieve stability by
gradually reducing the learning rate as the same set of training patterns is presented many
times. These nets respond to new pattern at any stage of learning and hence plastic.
5. What is an empty fuzzy set and what do you mean by height of a fuzzy set?
An empty fuzzy set is a function that for every u in the set U assumes the value of zero.
Using the notation just described, we can write this set in the following way.
Empty Set = {{u , 0}}, u in the set U

6. State the principle centre of gravity method of fuzzification

The centre of gravity method is also called centroid method. It is given by the algebraic
expression
∫ 𝜇𝐶 (𝑧). 𝑧 𝑑𝑧
𝑍∗ =
∫ 𝜇𝐶 (𝑧) 𝑑𝑧
7. What are the advantages of GA over conventional algorithm?

Genetic algorithms are search algorithms based on the mechanics of natural selection and
genetics. It applies random search and reaches the optimal solution faster than
conventional methods. The genetic algorithm starts with a set of initial solutions called
21
population and hence does not fall in local maxima or minima compared to other
conventional methods.
8. What is TABU search?

Tabu search is a higher level heuristic procedure for solving optimization problems,
designed to guide other methods to escape the trap of local optimality. It is a
metaheuristic search method employing local search methods used for mathematical
optimization.
9. A modern fully automatic washing machine is fixed with what type of intelligent
control? Define the input and output of the system
A modern fully automatic washing machine is fixed with neuro fuzzy control since it
involves human knowledge. The input is water, detergent and the panel setting the type
of cloth. The output is clean washed clothes.
Load Auto Sensor: This sensor will use to determine the size of a load of dirty clothes,
and then add the appropriate amount of water to the washing machine.
Temperature Control: It adjust the water to the correct temperature for the cycle you
are using. Automatic Dispensers: Automatic dispensers for bleach, detergent at the
appropriate time in the cycle.
Fuzzy Control: It will contain load sensors, which indicates the water and detergent
amount that is required.
Neuro-fuzzy Control: It features sensors that automatically detects the fabric type and
determine the detergent and water needed.
10. What is a hybrid intelligent control?
Hybrid intelligent control is a combination of many soft computing techniques like

fuzzy, neural and genetic algorithm to design an intelligent control system
PART B
11. a. i. Write short notes on adaline and madaline networks

Pages: 81-88
ii. Explain the single perceptron with its learning algorithm and its separability
and convergence property

Pages: 60-76
22
b. Explain back propagation for feed forward network to find the gradient vector in a
network

Pages: 290-300
12. a. i. Describe briefly the architecture of Hopfield network

Pages:135-140
ii. Construct and test a heteroassociative memory net to store the given vector pairs
S(1)=(0 0 0 1), t(1)= (0 1)
S(2)=(0 0 11) ,t(2)=(0 1)
S(3)= (0 1 0 0) , t(3) = (1 0)
S(4) = (1 1 0 0) , t(4) = (1 0)
Also test the network with noisy input patterns
b. Explain the adaptive resonance theory network with an example

Pages:222-243
13. a. i. The discretized membership functions for a transistor and resisitor are given
below:
µT = {0/0 + 0.2/1 +0.7/2 + 0.8/3 + 0.9/4 + 1/5}
µR = { 0/0 + 0.1/1 +0.3/2+0.2/3+0.4/4+0.5/5}
Find the following (a) Algebraic sum (b) Algebraic product (c) bounded sum and (d)
bounded difference
Algebraic sum =
{max(0,0),max(0.2,0.1),max(0.7,0.3),max(0.8,0.2),max(0.9,0.4),max(1,0.5)}
23
={0,0.2,0.7,0.8,0.9,1}
Algebraic Difference =
{min(0,0),min(0.2,0.1),min(0.7,0.3),min(0.8,0.2),min(0.9,0.4),min(1,0.5)}
={0,0.1,0.3,,0.2,0.4,,0.5}
Bounded Sum=max (min)=
Bounded Difference = min (max)=
ii. Explain self organising FLC. Mention its advantages over fuzzy logic controller

Page : 475-477
b. i. If x is A, then y is B is S where ‘S’ is fuzzy truth qualifier and a fact is in the form
“x is A”, we want to make an inference in the form “y is B”. Develop a method based on
the truth value restrictions for getting the inference

Page : 194-195
ii. Sketch the block diagram of fuzy logic controller for a nonlinear process

Page : 275-278 (Pg:497 example)
13. a. Explain the major components of GA and apply GA to maximize the “Peaks”
functions

Pages: 61-67
b. Explain in detail the ANT colony algorithm

24
Pages: 106-114
15. a. Using Matlab Neural network tool box discuss how will you identify and control
the linear and nonlinear dynamic system
Book :
Page :
b. i. Describe the optimization of travelling salesman problem using genetic algorithm

approach

Pages: 120-134
ii. Write short notes on stability analysis of fuzzy control systems

Page : 477-483
PART C
16. a. Consider four travel packages offered by Celtic Club Mahindra, Metro and
Himalaya Travels. We want to choose one. Their costs are INR 100.000, INR
200.000, INR 150.000 INR 175.000. Their travel time in hours are 150,200,,100 and
125 respectively. They are viewed as interesting with degrees 0.4,0.3,0.6 and 0.5.
Define your own fuzzy set of acceptable travel times. Then determine the fuzzy set
of interesting travel packages whose cost and travel times are acceptable and use
this set to choose one of your own packages.
Solution attached as Annexure
b. Assume a typical control problem of yours and explain the various steps involved
in finding a solution using GA
Notes
CL 7204 – SOFT COMPUTING
NOVEMBER/DECEMBER 2016
PART A
1. What are supervised learning and unsupervised learning?
Learning the input vectors in the presence of target vector is called supervised learning.
In the case of unsupervised learning, a sequence of input vector is provided, but no target
vectors are specified. The net modifies the weights so that the most similar input vectors
are assigned to the same output unit.
25
2. Draw the basic model of a Madaline network
Contains n units of input layer, m units of adaline layers, 1 unit of madaline layer
3. What is a counter propagation network?

The counter propagation network is a hybrid unsupervised learning network. It consists
of an outstar network and a competitive filter network. It was developed in 1986
by Robert Hecht-Nielsen. It is guaranteed to find the correct weights, unlike regular back
4. Mention the principle of adaptive resonance theory

Carpenter on aspects of how the brain processes information. It describes a number
of neural network models which use supervised and unsupervised learning methods, and
address problems such as pattern recognition and prediction. It typically consists of a
comparison field and a recognition field composed of neurons, a vigilance parameter
(threshold of recognition), and a reset module.
5. What is an empty fuzzy set and define the height of a fuzzy set?
An empty fuzzy set is a function that for every u in the set U assumes the value of zero.
Using the notation just described, we can write this set in the following way.
Empty Set = {{u , 0}}, u in the set U

6. What is the principle of centre of gravity method of defuzzification?

The centre of gravity method is also called centroid method. It is the commonly used
defuzzification method. It is given by the algebraic expression
∫ 𝜇𝐶 (𝑧). 𝑧 𝑑𝑧
𝑍∗ =
∫ 𝜇𝐶 (𝑧) 𝑑𝑧
7. Mention the role of fitness function in GA

The fitness function is the function to be optimized. For standard optimization
algorithms, this is known as the objective function which needs to be maximized or
minimized.
26
8. State two features of Ant colony algorithm
The ant colony optimization algorithm (ACO) is a probabilistic technique for solving
computational problems which can be reduced to finding good paths through graphs.
This algorithm is a member of the ant colony algorithms family, in swarm
intelligence methods, and it constitutes some meta heuristic optimizations. It is based on
the behavior of ants seeking a path between their colony and a source of food.
9. What is two point crossover?
Two-point crossover calls for two points to be selected on the parent organism strings.
Everything between the two points is swapped between the parent organisms, rendering
two child organisms
Example
Parent 1: 10|00000|01 Child 1: 101111101
Parent 2: 01|11111|10 Child 2: 010000010
10. What is hybrid intelligent control?

Hybrid intelligent control is a combination of many soft computing techniques like
fuzzy, neural and genetic algorithm to design an intelligent control system
Part B
11. a. Explain the working of back propagation neural network with neat architecture
and flowchart
Pages:290-306
b. What is self organising map and discuss the algorithm and features of Kohonen’s
map?

Pages: 169-172
12. a. With a neat sketch explain the operation (Training and Testing) of recurrent
neural network?
b. Draw the architecture Hopfield network. Design Hopfield net for 4 bit bipolar
pattern. The training patterns are:
I Sample S1 (1,1,-1,-1)
II Sample S2(-1 1 -1 1)
III Sample S3(-1 -1 -1 1) find the weight matrix and energy for 3 input samples.
Xerox attached
13. a. Explain different types of defuzzification with suitable example
27
Text Book : Fuzzy Logic with Engineering Applications by Timothy J.Ross
Page : 136-137
b. Explain the working of self organising FLC. Mention its advantages over fuzzy
logic controller

Page : 470-474
14. a. Explain the various phases of GA to control a nonlinear time delay system

Pages: 61-67
b. Explain compare and contrast Ant colony and Tabu search algorithm

Pages: 114-122
15. a. How stability is ensured in Fuzzy control system? Analyze with reference to
global network computation
Page : 477-483
b. Explain in detail the optimization of load regulation problem using GA

Pages: 125-135
16. a. Describe briefly modelling and implementation of FLC for inverter pendulum
Book: Neural Networks and Fuzzy Logic by Bart Kosko
Pages: 484-489
b. i. Using Matlab neural network tool box discuss how will you identify and control
the linear and nonlinear dynamic system
Notes
ii. Briefly explain the use of GA assuming an application in daily life

(Traveling Saleman Problem or Power system problem)

Pages: 120-134
28
ME DEGREE EXAMINATION, MAY/JUNE 2017
CL 7204 SOFT COMPUTING TECHNIQUES
1. Define an artificial neural network

Artificial neural network (ANN) is an information processing system which is a mathematical
model of human brain system. ANNs are considered nonlinear statistical data modeling tools
where the complex relationships between inputs and outputs are modeled or patterns are found
2. State the function of synapse

Synapse represents the connection weight between two neurons. It refers to the strength or
amplitude of a connection between two nodes, corresponding to the amount of influence the
firing of one neuron has on another.
3. State the adaptive resonance theory

Carpenter on aspects of how the brain processes information. It describes a number of neural
network models which use supervised and unsupervised learning methods, and address problems
such as pattern recognition and prediction. It typically consists of a comparison field and a
recognition field composed of neurons, a vigilance parameter (threshold of recognition), and a
reset module.
4. What is meant by associative memory

A neural net that is trained to associate a set of input vectors with a corresponding set of output
vectors is called associative memory. Associative memory neural networks are nets in which the
weights are determined in such a way that the net can store a set of P pattern associations. If the
desired output vector is same as the input vector, the net is an auto associative memory. If the
desired output vector is different from the input vector, the net is a hetero associative memory.
5. Compare crisp and fuzzy sets

A classical set is defined by crisp boundaries. i.e there is no uncertainty in the prescription or
location of the boundaries of the set. The boundaries of fuzzy set are ambiguously specified. In
crisp set an element either belongs to the set or it does not belong to it. In fuzzy set an element
belongs to one or more sets.
6. What is fuzzy inference system

Fuzzy inference (reasoning) is the actual process of mapping from a given input to an output
using fuzzy logic
29
7. List the learning methods of back propagation network
Levenberg Marquardt, Bayesian regularization, Gradient descent, gradient descent with

momentum, Variable learning rate gradient descent, scaled conjucate gradient, resilient
backpropagation, BFGS Quasi Newton
8. List the various types of crossover and mutation techniques

Types of mutation
Mutation Gaussian, mutationuniform, mutation adaptivefeasible, insert, inverse, scramble, creep
mutation
Types of crossover
Crossoverscattered, crossoversinglepoint, crossovertwopoint, crossoverintermediate,
crossoverheuristic, crossoverarithmetic, order crossover, cycle crossover
9. Name any two search techniques used for solving optimization problems
Particle swarm optimization, Genetic algorithm, Ant Colony optimization, Tabu search
10. What is the purpose of tool boxes in matlab
Matlab toolbox are interactive applications written to perform technical computing tasks. They
are designed to create and test several applications without writing full coding. Toolboxes allow
you to learn and apply specialized technology. Toolboxes are comprehensive collections of
MATLAB functions (M-files) that extend the MATLAB environment to solve particular classes
of problems. Areas in which toolboxes are available include signal processing, control systems,
neural networks, fuzzy logic, wavelets, simulation, and many others.
30
M.E/M.TECH DEGREE EXAMINATION, JANUARY 2018
IN 5091 SOFT COMPUTING TECHNIQUES
PART A
1. Compare soft computing and hard computing

human mind.
2. List the various types of soft computing techniques and mention some application
areas for neural network
Neural Networks, Support Vector Machine, Fuzzy Logic, Evolutionary Algorithm,
Genetic Algorithm, Machine Learning (ML) and Probabilistic Reasoning (PR), Ant
colony optimization, Particle swarm optimization, cuckoo search algorithm
Applications of neural network
 Pattern recognition, Satellite image classification, Weather forecasting
 Stock prediction, Data compression, Data Mining
 Contingency analysis , Diagnosis/monitoring, Distribution planning
 Load frequency control, Generator maintenance scheduling
 Generation dispatch, Load flow computations
 Load forecasting, Load management, Unit commitment
3. What is counter propagation network

The counterpropagation network is a hybrid unsupervised learning network. It consists of
an outstar network and a competitive filter network. It was developed in 1986 by Robert
Hecht-Nielsen. It is guaranteed to find the correct weights, unlike regular back
4. Mention the working principle of adaptive resonance theory

Carpenter on aspects of how the brain processes information. It describes a number
of neural network models which use supervised and unsupervised learning methods, and
address problems such as pattern recognition and prediction. It typically consists of a
comparison field and a recognition field composed of neurons, a vigilance parameter
(threshold of recognition), and a reset module. Adaptive Resonance Networks are plastic
and stable. These nets achieve stability by gradually reducing the learning rate as the
same set of training patterns is presented many times. These nets respond to new pattern
at any stage of learning and hence plastic.
5. How does a fuzzy set differ from crisp set

A classical set is defined by crisp boundaries. i.e there is no uncertainty in the
prescription or location of the boundaries of the set. The boundaries of fuzzy set are
31
ambiguously specified. In crisp set an element either belongs to the set or it does not
belong to it. In fuzzy set an element belongs to one or more sets.
6. Draw the block diagram of fuzzy logic system
7. State the operators of genetic algorithm

Reproduction
Crossover
Mutation
8. List the mutation operators in GA

Commonly used mutation operators are Mutation Gaussian, mutationuniform, mutation
adaptivefeasible
1) Insert Mutation- It is used in Permutation encoding. First of all, pick two allele
values at random. Then move the second allele to follow the first, shifting the rest along
to accommodate. Note that this preserves most of the order and the adjacency
information
2) Inversion Mutation- Inversion mutation is used for chromosomes with permutation
encoding. In order to perform inversion, pick two alleles at random and then invert the
substring between them. It preserves most adjacency information and only breaks two
links but it leads to the disruption of order information.
3) Scramble Mutation- Scramble mutation is also used with permutation encoded
chromosome. In this mutation, one has to pick a subset of genes at random and then
randomly rearrange the alleles in those positions. Subset does not have to be contiguous.
4) Swap Mutation- It is also used in Permutation encoding. To perform swap mutation
select two alleles at random and swap their positions. It preserves most of the adjacency
information but links broken disrupts order more.
5) Flip Mutation- Based on a generated mutation chromosome, flipping of a bit involves
changing 0 to 1 and 1 to 0. A parent is considered and a mutation chromosome is
randomly generated . For a 1 in mutation chromosome, the corresponding bit in parent
chromosome is flipped (0 to 1 and 1 to 0) and child chromosome is produced. It is
32
commonly used in binary encoding. Interchanging Mutation- Two random positions of
the string are chosen and the bits corresponding to those positions are interchanged.
7) Reversing Mutation- In case of reversing mutation applied for binary encoded
chromosome, random position is chosen and the bits next to that position are reversed
and child chromosome is produced .
8) Uniform Mutation- The Mutation operator changes the value of chosen gene with
uniform random value selected between the user specified upper and lower bound for
that gene. It is used in case of real and integer representation.
9) Creep Mutation- In creep mutation, a random gene is selected and its value is
changed with a random value between lower and upper bound. It is used in case of real
representation .
9. What is meant by optimization? List few optimization techniques

Optimization is the process of finding alternative parameters with the most effective or
highest achievable performance under the given constraints, by maximizing desired
factors and minimizing undesired ones. Optimization needs a fitness or objective
function to be maximized or minimized under certain constraints.
The optimization techniques are
(i) Genetic Algorithm
(ii) Particle Swarm optimization
(iii) Ant Colony Optimization
(iv) Tabu Search
10. State any one stability analysis method for neural network
Lyapunov Stability Analysis : Lyapunov analysis consists of the selection of a positive
function candidate V which leads to the computation of an adaptation law insuring its
decrescence, that is 𝑉 ≤ 0, for continuous systems and for
discrete-time systems. Under these assumptions, the function V is called Lyapunov
function and guarantees the stability of the system.
Part B
If x is A, then y is B is S where ‘S’ is fuzzy truth qualifier and a fact is in the form “x is
A”, we want to make an inference in the form “y is B”. Develop a method based on the
truth value restrictions for getting the inference
Fuzzy Propositions Fuzzy logic deals with the degree of truth of fuzzy propositions,
which can be conditional, qualified, or both, as well as simple. A simple, unconditional
and unqualified, proposition states that a variable element belongs to a fuzzy set, as in
“the age of x is old”. For a particular element, the degree of truth of the proposition is
interpreted as the degree of membership of this element to the fuzzy set. In this way, any
fuzzy proposition can be interpreted as a possibility distribution function that is equal to
the membership function of the fuzzy set. A simple proposition p has the form: p : V is F
, p : V(i) is F, when it is important to discuss the individual element referred to by the
proposition, as in “The age of John is old”. It is possible to apply hedges to propositions
33
the way that we apply adverbs to adjectives, as in “the age of x is very old”. A qualified
fuzzy proposition is a simple fuzzy proposition modified by a fuzzy truth qualifier or a
fuzzy probability qualifier, as in “ ‘the age of x is old’ is likely”3 or “ ‘the probability
that the age of x is old’ is high”. They describe the degree of truth or the probability of
the original proposition. Simple fuzzy propositions are equivalent to qualified
propositions where the truth qualifier is “is true”. Qualified fuzzy propositions, p, take
the form: p: V is F is S or p: Pro{V is F} is P. Conditional propositions discuss the
implication of one fuzzy proposition from another, as in “If age is old, then strength is
feeble”. Conditional propositions, i.e. fuzzy implications, are very important to develop
fuzzy systems, since fuzzy rules can be seen as conditional propositions. They take the
form: p: If X is A, then Y is B. Propositions can also be quantified, i.e., we can also make
statements about to how many elements of a set a proposition are true. For example:
“most engineering students have good knowledge of math” or “about one sixth of the
students have no knowledge of it”. Quantified propositions are handy tools in datamining
engines.
The classical inference rules of modus ponens, modus tollens, and hypothetical
syllogism can be generalized for fuzzy logic and used to support approximate reasoning.
The Modus ponens deduction rule allows us to deduce the truth of the consequent of an
implication based on the truth of the antecedent.
Since most fuzzy systems use if-then rules (implications) and a single forward
pass, the generalization of modus ponens is of extreme importance to fuzzy logic. Given
an implication “if p then q” and the fact that p is true, modus ponens deduces q. If p and
q are fuzzy propositions (X is A) and (Y is B), and A * and B * are fuzzy sets in X and
Y, then the following schema represents generalized modus ponens: IF X IS A, then Y is
B X is A* — Y is B* Since Y is B* is a fuzzy proposition, it must be possible to deduce
its possibilistic distribution, i.e., the membership function of B *. This is performed using
one of many available compositional rules of inference: B * = A* o R, such as mB
*(y)=sup(min(mA *(x),mR(x,y)) where R is the vector describing the relationship
between sets A and B. There are four important classes of fuzzy implication
operators[4,5]: 1. S-implications, based on the formalism p®q = Ø p Ú q; 2. QL-
implications, based on p®q = Øp Ú (pÙq); 3. R-implications, based on p®q = sup{r Î
[0;1] | (p Ù r) £q }, and 4. t-norm-implications, based on p®q = pÙq (false in boolean
logic).
34
35

Soft Computing Question Bank

Uploaded by

Copyright:

Available Formats

Soft Computing Question Bank

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Soft Computing Question Bank

Uploaded by

Copyright:

Available Formats

DR.

SIVANTHI ADIATANR COLLEGE OF ENGINEERING

Department of Electrical and Electronics Engineering

M.E.Power Electronics and Drives

SOFT COMPUTING TECHNIQUES CL72014

UNIT I : INTRODUCTION AND ARTIFICIAL NEURAL NETWORKS

1. Define soft computing

In computer science, soft computing is the use of inexact solutions to computationally

2. Differentiate hard and soft computing

Hard computing, i.e., conventional computing, requires a precisely stated analytical

3. Name some soft computing techniques.

Neural Networks, Support Vector Machine, Fuzzy Logic, Evolutionary Algorithm,

4. Define artificial neural network.

Artificial neural network (ANN) is an information processing system which is a

6. What is neural network architecture? Explain logistic sigmoid function with

7. Explain important characteristics and applications of artificial neural network?

Fault tolerance.Signal processing, Control, pattern recognition, Medicine, Speech

8. Name some neural network activation functions

9. What is the classification of training? Explain supervised training?

10. What is unsupervised training?

f(x) = x for all x.

13. What is feedforward neural network. Draw its structure.

A feedforward neural network is an artificial neural network where connections between

1. What is perceptron? Explain different layers of perceptron?

Book: Fundamentals of Neural Network by Laurene Fausett

Book: Fundamentals of Neural Network by Laurene Fausett

Book: Fundamentals of Neural Network by Laurene Fausett

UNIT II – ARTIFICIAL NEURAL NETWORKS

1. What are the different types of neural network

MccullochPitts Neural network, Perceptron, Adaline, Madaline, Counter Propagation

2. Give examples for supervised and unsupervised learning neural networks

Supervised Learning : Multi Layer Pereptron, Adaline,Madaline, Feed forward

Unsupervised Learning : Self Organising Map, Learning Vector Quantization, Counter

3. What is the concept of hebbian learning?

4. What is a counter propagation network (CPN)?

The counterpropagation network is a hybrid unsupervised learning network. It consists of

6. Draw the structure of ART network

7. State the characteristics of counter propagation network (CPN)

o CPN is a multilayer network based on combination of input, clustering

8. What is the feature of boltzmann machine?

A Boltzmann machine is a network of symmetrically connected, neuron-like units that

9. Differentiate ART1 and ART2

10. What is associative memory? Explain its types?

Hopfield network is a deterministic neural network whereas Boltzmann is a stochastic

12. State the properties of neural networks

1. With neat diagram explain Hopfield/Recurrent network

Book: Fundamentals of Neural Network by Laurene Fausett

2. Explain Counter Propagation Algorithm

Book: Fundamentals of Neural Network by Laurene Fausett

3. Write in detail about ART network

Book: Fundamentals of Neural Network by Laurene Fausett

4. Explain Boltzmann machine algorithm

Book: Fundamentals of Neural Network by Laurene Fausett

UNIT III FUZZY LOGIC SYSTEM

1. What is meant by fuzzy logic?

2. Differentiate classical and fuzzy set

A classical set is defined by crisp boundaries. i.e there is no uncertainty in the

3. When is fuzzy logic used?

4. What are the fuzzy set operations?

5. Define fuzzy Union and Intersection operation