DS1701 Neuro-Fuzzy Computing Department of ADS 2024 - 2025

UNIT I – ARTIFICIAL NEURAL NETWORK

Review of fundamentals – Biological neuron, artificial neuron, activation function, single layer perception
– Limitation – Multilayer perception – Back Propagation Algorithm (BPA) – Recurrent Neural Network
(RNN) – Adaptive Resonance Theory (ART) based network – Radial basis function network – online
learning algorithms, BP through time – RTRL algorithms –Reinforcement learning
Unit I / Part - A
1. What is soft computing?
2. What are the key components of soft computing?
3. What are the key characteristics of soft computing?
4. What are the key characteristics of soft computing?
5. Compare hard computing and soft computing.
6. List the structure and Components of biological neuron.
7. List the structure and Components of artificial neuron.
8. Compare biological neuron and artificial neuron
9. What is perceptron?
10. What are the basic components of a perceptron (or) structure of a perceptron?
11. List the types of Perceptron.
12. Specify the limitations of single layer perceptron.
13. How can you address the limitations of single layer perceptron?
14. What is activation function?
15. List the importance of activation function in the neural network.
16. How to choose an activation function?
17. Which function is commonly used as an activation function in the output layer of a i) binary
classification ii) Multi-Class Classification iii) regression neural network?
18. What is multilayer perceptron?
19. What is backpropagation algorithm?
20. What is Recurrent Neural Network (RNN)?
21. What are the key characteristics of RNN?
22. What is Adaptive Resonance Theory (ART)?
23. List the key concepts of ART.
24. What is Radial Basis Function Network (RBFN)?
25. Discuss Online Learning Algorithms:
26. What is Discuss Backpropagation Through Time (BPTT)?
27. What is Real-Time Recurrent Learning (RTRL)?
Unit I / Part - B
1. With diagram, illustrate the terms cell body, axon, synapse, dendrite and neuron with reference to a
biological neuron
2. What is linear inseparability in a data set and explain, using a two dimensional example how a single
layer perceptron model cannot solve this problem?
3. Assume that there are L input nodes, plus the bias, M hidden nodes, also plus a bias, and N output
nodes. How many weights are there i) between the input and the hidden layer and ii) between the
hidden layer and the output.

St. Joseph’s College of Engineering Page 1 of 10

 DS1701 Neuro-Fuzzy Computing Department of ADS 2024 - 2025
4. Explain multilayer perceptron algorithm in detail. What are the choices that can be made about the
network in order to use it for solving real problems?
5. Explain why a recurrent neural network would yield better results than a feed-forward network in
which your input is the summation (or average) of past and current data? Give two examples that
can be more naturally modeled with RNNs than with feedforward neural networks with
explanation?
6. RNNs are trained using an algorithm known as Backpropagation Through Time (BPTT). Explain
BPTT. How is the loss function computed in BPTT?
7. Explain what differentiates a Recurrent Neural Network (RNN) from a normal Feedforward Neural
Network, and for what type of NLP classification tasks the RNN model is particularly appropriate
8. Draw and explain the architecture of Adaptive Resonance Theory (ART)
9. Explain the basic structure of a Radial Basis Function (RBF) Network and "compare it to the basic
structure of a self-organizing map (SOM)
10. Discuss the following in detail: i) online learning algorithms ii) BP through time iii) RTRL algorithms
11. What are the benefits and use cases of reinforcement learning? How does reinforcement learning
work?
12. Discuss the types of reinforcement learning algorithms in detail.
13. Discuss the terms used in reinforcement learning in detail.
Unit I / Part - C
14. Consider a neuron with 2 inputs, 1 output, and a threshold activation function. If the two weights
are w1 = 1 and w2 = 1, and the bias is b = −1.5, then what is the output for input (0,0)? What about
for inputs (1,0), (0,1), and (1,1)? Draw the discriminant function for this function, and write down its
equation. Does it correspond to any particular logic gate?
15. Work out the Perceptrons that construct logical OR and logical AND of their inputs.
16. Work out the Perceptrons that construct logical NOT, NAND, and NOR of their inputs.
17. Below is a diagram if a single artificial neuron (unit):

The node has three inputs x = (x1; x2; x3) that receive only binary signals (either 0 or 1). How many
different input patterns this node can receive? What if the node had four inputs? Five? Can you give
a formula that computes the number of binary input patterns for a given number of inputs?
18. Consider the unit shown on Previous Figure. Suppose that the weights corresponding to the three
inputs have the following values:

and the activation of the unit is given by the step function:

St. Joseph’s College of Engineering Page 2 of 10

 DS1701 Neuro-Fuzzy Computing Department of ADS 2024 - 2025

Calculate what will be the output value y of the unit for each of the following
input patterns:

19. Logical operators (i.e. NOT, AND, OR, XOR, etc.) are the building blocks of any computational
device. Logical functions return only two possible values, true or false, based on the truth or false
values of their arguments. For example, operator AND returns true only when all its arguments are
true, otherwise (if any of the arguments is false) it returns false. If we denote truth by 1 and false by
0, then logical function AND can be represented by the following table:

This function can be implemented by a single unit with two inputs:

if the weights are w1 = 1 and w2 = 1 and the activation function is:

Note that the threshold level is 2 (v ≥ 2).

a) Test how the neural AND function works.
b) Suggest how to change either the weights or the threshold level of this single unit in
order to implement the logical OR function (true when at least one of the arguments is
true):
c) Do you think it is possible to implement XOR function using a single unit? A network
of several units?
20. The following diagram represents a feed forward neural network with one hidden layer:

A weight on connection between nodes i and j is denoted by wij , such as w13 is the weight on the
connection between nodes 1 and 3. The following table lists all the weights in the network:

St. Joseph’s College of Engineering Page 3 of 10

 DS1701 Neuro-Fuzzy Computing Department of ADS 2024 - 2025

Each of the nodes 3, 4, 5 and 6 uses the following activation function:

where v denotes the weighted sum of a node. Each of the input nodes (1 and 2) can only receive
binary values (either 0 or 1). Calculate the output of the network (y5 and y6) for each of the input
patterns:

21. Assume that the neurons have a sigmoid activation function, perform a forward pass and a backward
pass on the network. Assume that the actual output of y is 5 and learning rate is 1.

22. Assume that the neurons have a sigmoid activation function, perform a forward pass and a backward
pass on the network. Assume that the actual output of y is 1 and learning rate is 0.9.

St. Joseph’s College of Engineering Page 4 of 10

 DS1701 Neuro-Fuzzy Computing Department of ADS 2024 - 2025
UNIT II NEURAL NETWORKS FOR MODELING AND CONTROL
Modelling of non-linear systems using ANN – Generation of training data – Optimal architecture–Model
validation – Control of non-linear systems using ANN – Direct and indirect Neuro control schemes –
Adaptive Neuro controller – Familiarization with neural network toolbox
Unit II / Part - A
1. List the approaches to generate training data for a non-linear system.
2. Specify the considerations for generating training data.
3. Write the step-by-step approach to controlling non-linear systems using ANN.
4. What the key Considerations in controlling non-linear systems using ANN?.
5. How can you determine the optimal architecture for a neural network?
6. What are the steps involved in finding optimal architecture for neural network?
7. Provide Guidelines for Finding Optimal Architecture
8. What are the issues involved in designing layers?
9. What are the issues in Hyperparameter Tuning?
10. What is the purpose of model validation?
11. Define training, validation and test data.
12. Define cross validation
13. Write k-Fold cross-validation
14. List the performance metrics.
15. Write notes on deep learning toolbox.
16. Provide some common MATLAB toolboxes.
17. List the important purposes of Toolboxes in MATLAB
18. Provide the key architectures you can work with using the toolbox
19. List the steps in training neural networks using the Deep Learning Toolbox in MATLAB.
20. What is the purpose of transfer functions in MATLAB neural network toolbox?
21. List the commonly used transfer functions in MATLAB
Unit II/ Part B
1. Explain the steps in modelling of non-linear systems using ANN with example.
2. Discuss the components of adaptive neuro controller in detail.
3. How to construct the optimal neural architecture for your machine learning task?
4. Explain the concept of cross-validation and its importance in model assessment. Compare k-fold
cross-validation and leave-one-out-cross validation.
5. Briefly explain the differences among the cross validation techniques.
6. Explain how cross-validation for classification problems can be adapted to be applicable to time
series prediction.
7. In artificial neural network, input datasets are needed to divide as training and validation sets.
Explain why we need training and validation sets to develop ANN model.
8. Discuss with relevant diagrams and mathematical expressions how a nonlinear system can be
identified and controlled using MATLAB Neural Network Tool box. Choose appropriate example.
9. Explain the case study of the application of neural network for stability analysis of interconnected
systems.

St. Joseph’s College of Engineering Page 5 of 10

 DS1701 Neuro-Fuzzy Computing Department of ADS 2024 - 2025
10. Using MATLAB Neural Network tool box discuss how will you identify and control the linear and
nonlinear dynamic system.
11. With a suitable example demonstrate the perceptron learning law with its decision regions using
MATLAB. Give the output in graphical form.
UNIT III FUZZY SET THEORY
Fuzzy set theory – Fuzzy sets – Operation on fuzzy sets – Scalar cardinality, fuzzy cardinality, union and
intersection, complement (Yager and Sugeno), equilibrium points, aggregation, projection, composition,
cylindrical extension, fuzzy relation – Fuzzy membership functions
Unit III / Part - A
1. What is fuzzy set?
2. List the application of fuzzy set.
3. What is fuzzy logic?
4. List the key concepts of fuzzy logic.
5. What are the advantages and disadvantages of fuzzy logic?
6. Define complement operation
7. Define union operation.
8. Define Intersection operation.
9. Define Alpha-Cut operation
10. Define Support operation.
11. Define Core operation
12. Define Height operation
13. Define Cartesian Product
14. Define Fuzzy Relations
15. Assume X= {1,2,3,4,5}. Let A be a fuzzy set with the membership function μA given by:
μA= {(1,0.2), (2,0.5), (3,0.7), (4,0.9), (5,1.0)}.
Let B be another fuzzy set with the membership function μB given by:
μB = {(1,0.3), (2,0.4), (3,0.6), (4,0.8), (5,0.9)}.
Find Complement of A, Union of A and B & Intersection of A and B
16. Define scalar cardinality.
17. Define fuzzy cardinality.
18. Define Yager’s complement.
19. Define Sugeno’s complement.
20. What is fuzzy membership function?
21. List the applications of fuzzy membership function.
22. Define Fuzzy relation.
Unit III/ Part B
1. Define a fuzzy set and explain how it differs from a classical set. Provide an example of a fuzzy set
in a real-world context, illustrating how its elements have degrees of membership rather than binary
inclusion or exclusion as in classical sets
2. Define main operations of fuzzy sets. Provide examples. How are fuzzy set operations, their
properties and hedges used to obtain a variety of fuzzy sets from the existing ones?

St. Joseph’s College of Engineering Page 6 of 10

 DS1701 Neuro-Fuzzy Computing Department of ADS 2024 - 2025
3. Explain the operations on crisp sets by using Venn diagrams?
4. Differentiate between Mamdani FIS and Sugeno FIS.
5. An athletic race was conducted. The following membership functions are defined based on the speed
of athletes:
Low = {0/100 + 0.1/200 + 0.3/300}
Medium = {0.5/100 + 0.57/200 + 0.6/300}
High = {0.8/100 + 0.9/200 + 1.0/300}
Find the following:
R = Low x Medium
S = Medium x High
T = R ₀ S using max-min composition
6. Consider the following two fuzzy sets.
A = {0.2/1 + 0.3/2 + 0.4/3 + 1.0/4}
B = {0.1/1 + 0.2/2 + 0.3/3 + 0.5/4}
Find the algebraic sum, algebraic product, bounded sum & bounded difference.
7. Discuss the common types of fuzzy membership functions used in practice, each with different
characteristics and applications.
8. Discuss the types and applications of fuzzy relation.
9. Discuss the various operations on fuzzy relations with example.
10. Explain how aggregation and composition differ. Give examples to support your explanation.
11. Provide a numerical example demonstrating the composition of fuzzy relations.
12. Write a short notes on aggregation of fuzzy rules and explain about determination
of aggregation strategy.
13. In a computer engineering different logic families are often compared on the basis of their power-
delay product. The fuzzy set F is the logic families F= {NMOS, CMOS, TTL, ECL, JJ} . The range of
delay time D= {0.1, 1, 10, 100} in Nano seconds. The power dissipation in micro watts P= {0.01, 0.1, 1,
10, 100}. By using max-min composition to obtain a fuzzy relation between delay time and Power
dissipation.
14. Given the fuzzy expressions for A and B are:
A =0.3/1+0.5/2+0.3/3+1.0/4+0.9/5+0.5/6+0.2/7+0.3/8+0.9/9
B =0.4/1+0.8/2+1.0/3+0.9/4+0.9/5+0.4/6+0.7/8+0.9/9+0.7/10
Solve the following fuzzy relation: i) 𝐴 ∩ 𝐵 ii) 𝐴 ∩ 𝐵
UNIT IV FUZZY LOGIC FOR MODELING AND CONTROL
Modelling of non-linear systems using fuzzy models – TSK model – Fuzzy logic controller – Fuzzification
– Knowledge base – Decision making logic – Defuzzification – Adaptive fuzzy systems – Familiarization
with fuzzy logic toolbox
Unit IV / Part - A
1. Write the steps for modeling nonlinear systems.
2. What are the challenges using fuzzy models for nonlinear systems?
3. List the benefits of fuzzy models for nonlinear systems.
4. What is TSK model?
5. List the key components of the TSK Model.
6. What is fuzzy logic controller?.

St. Joseph’s College of Engineering Page 7 of 10

 DS1701 Neuro-Fuzzy Computing Department of ADS 2024 - 2025
7. List the 4 modules of a general fuzzy controller.
8. How does fuzzy controller operate?
9. What are the two main approaches to designing a Fuzzy Logic Controller (FLC)?
10. How Mamdani approach FLC differ from the Takagi and Sugeno’s approach of FLC?
11. What is Fuzzification?
12. What is the purpose of decision making logic (Inference)?
13. What is defuzzification?
14. List the different defuzzification Methods
15. Define Center of Sums Method (COS).
16. The aggregated fuzzy set of two fuzzy sets C1 and C2 is shown in Figure. Let the area of this two
fuzzy sets are A1 and A2. Calculate the defuzzified value using COS method.

17. Define Center of gravity (COG) / Centroid of Area (COA) Method.

18. Write the defuzzified value using COG for continuous and discrete membership value?
19. Write formula for Center of Area / Bisector of Area Method (BOA)
20. Define weighted average method
Unit IV/ Part B
1. Explain how fuzzy models can be used to model nonlinear systems with example.
2. Discuss the key concepts and components of TSK model in detail.
3. TSK fuzzy systems provide a powerful framework for modeling and controlling complex systems
with high accuracy and efficiency-Elucidate.
4. Briefly discuss the fuzzy logic controller concept with an aided of a diagram.
5. Design and simulate a fuzzy logic controller for a temperature control system using MATLAB
6. Design a washing machine controller using Fuzzy Logic (FL). You need to fulfill the following
criteria: (i)Fuzzification based three (3) membership functions using three (3) input variables [weight
of the clothes load, dirt of the clothes, temperature of the water] (ii)One (1) output variable [wash
time] (iii)Establish 11 fuzzy rule sets (iv)Use Mamdani method
7. List and define different methods of defuzzification in detail.
8. What is defuzzification? What are the most popular defuzzification methods? Explain graphically,
giving an example, how we determine the final output of a fuzzy system using these methods.
9. For the logical union of the membership functions shown below find the defuzzified value using
different defuzzification methods.

St. Joseph’s College of Engineering Page 8 of 10

 DS1701 Neuro-Fuzzy Computing Department of ADS 2024 - 2025

10. Consider the following aggregated output membership function of a Mamdani-type inference
system. Using the COG defuzzification method determine the crisp output of the output system.

11. Explain the influence of choosing type of defuzzification method on fuzzy controller.
12. Discuss the key concepts and components of adaptive fuzzy systems in detail.
13. Discuss the benefits, key challenges and applications of adaptive fuzzy systems in detail.
14. Using MATLAB ANFIS tool box discuss how will you identify and control the linear and nonlinear
dynamic system.
UNIT V HYBRID CONTROL SCHEMES
Fuzzification and rule base using ANN – Neuro fuzzy systems – ANFIS – Fuzzy neuron– Introduction to
GA – Optimization of membership function and rule base using Genetic Algorithm – Introduction to
support vector machine – Particle swarm optimization – Case study – Familiarization with ANFIS toolbox
Unit V / Part - A
1. What do you mean by fuzzification and rule base using ANN?
2. What is ANFIS?
3. List the components of ANFIS.
4. List the applications of ANFIS.
5. List the advantages and disadvantages of ANFIS..
6. What is fuzzy neuron?
7. What is genetic algorithm?
8. Define the basic terminologies used in genetic algorithm
9. What is selection process?
10. How Genetic Algorithm Work?
11. What is crossover?
12. List the different types of crossover?

St. Joseph’s College of Engineering Page 9 of 10

 DS1701 Neuro-Fuzzy Computing Department of ADS 2024 - 2025
13. What are Support Vector Machines?
14. What is the goal of SVM?
15. What are support vectors? Give example.
16. What are types of SVM?
17. Define Kernel.
18. What is the purpose of particle Swarm Optimization (PSO)?
19. What are the Key advantages of Particle Swarm Optimization?
20. List the applications of PSO.
21. What are the main equations in the PSO algorithm?
Unit V / Part - B
1. From the behaviors of an intelligent systems, List all differences between Mamdani-type and Sugeno-
type fuzzy inferences.
2. Construct an ANFIS that is equivalent to a two-input two-rule Mamdani fuzzy model with max-min
composition and centroid defuzzification? Specify how to convert the defuzzification function into
node functions in the resulting ANFIS.
3. Explain the different types of mutation methods with example.
4. Explain the Roulette wheel technique for traditional GA selection.
5. With the help of examples, explain the various crossover techniques employed in genetic algorithms
6. Compare and contrast traditional algorithm and genetic algorithm
7. Explain the terms chromosome, fitness function, crossover and mutation as used in genetic
algorithms with illustrations. Also illustrate how genetic algorithm works in pseudocode.
8. A budget airline company operates 3 plains and employs 5 cabin crews. Only one crew can operate
on any plain on a single day, and each crew cannot work for more than two days in a row. The
company uses all planes every day. A Genetic Algorithm is used to work out the best combination
of crews on any particular day. Answer the following: (Any three)
i) Suggest what chromosome could represent an individual in this algorithm?
ii) Suggest what could be the alphabet of this algorithm? What is its size?
iii) Suggest a fitness function for this problem.
How many solutions are in this problem? Is it necessary to use Genetic Algorithms for solving it?
What if the company operated more plains and employed more crews?
9. Compare the applications fuzzy logic and genetic algorithm.
10 What is support vector machine (SVM)? How is SVM used for classification problem? Illustrate with
example.
11. Explain briefly how Particle Swarm Optimization (PSO) works.
11. Particle swarm optimization (PSO) is a population-based optimization algorithm that is inspired by
the social behavior of birds. It has been applied to a variety of optimization problems, including
finding the global minimum of a function.
Let us imagine you wanted to minimize a function, such as
f(x) = x2 + 2x + 1, where x is a real-valued variable. High might you design a PSO to find this?
12. Give an example where using a greedy algorithm would fail but PSO may work.

St. Joseph’s College of Engineering Page 10 of 10