CHEMICAL ENGINEERING

SOFT COMPUTING CATEGORY L T P CREDIT

CHT397
TECHNIQUES VAC 3 1 0 4

Preamble:This course gives an introduction to some new fields in soft computing with its
principal components of fuzzy logic, NN, GA and Fuzzy genetic hybrid systems.

Pre-requisite: Nil

Course Outcomes: After the completion of the course the student will be able to:
CO 1 Explain the concepts involved in fuzzy logic systems.
CO 2 Design fuzzy logic controllers for various applications
CO 3 Train neural networks for various applications
CO 4 Solve optimisation problems using Genetic Algorithms
CO 5 Understand the concepts of Hybrid Soft computing techniques.

Mapping of course outcomes with program outcomes

PO PO PO PO PO PO PO PO PO PO PO PO
1 2 3 4 5 6 7 8 9 10 11 12
CO 1 3 3 3 3 3
CO 2 3 3 3 3 3
CO 3 3 3 3 3 3
CO 4 3 3 3 3 3
CO 5 3 3 3 3 3

Assessment Pattern
Continuous Assessment Tests
Bloom’s Category End Semester Examination
1 2
Remember 10 10 10
Understand 20 20 20
Apply 20 20 70
Analyse
Evaluate
Create

Mark distribution
Total Marks CIE ESE ESE Duration

150 50 100 3 hours

 CHEMICAL ENGINEERING

Continuous Internal Evaluation Pattern:

Attendance : 10 marks
Continuous Assessment Test (2 numbers) : 25 marks
Assignment/Quiz/Course project : 15 marks

End Semester Examination Pattern: There will be two parts; Part A and Part B. Part A
contain 10 questions with 2 questions from each module, having 3 marks for each question.
Students should answer all questions. Part B contains 2 questions from each module of which
student should answer any one. Each question can have maximum 2 sub-divisions and carry
14 marks.

Course Level Assessment Questions

Course Outcome 1 (CO1):

1. Explain the features of fuzzy membership functions with proper diagrams.
2. Discuss fuzzy equivalence relations and list out its properties?
3. For the fuzzy sets A = {0.2/x1 + 0.9/x2} and B = {0.3/y1 + 0.5/ y2 + 1/y3}, find the
fuzzy relation A×B

Course Outcome 2 (CO2)

1. Defuzzify the following output membership function using centroid method:
0/0, 0.3/1, 0.3/3.5, 0.5/4, 0.5/5.5, 1/6, 1/7, 0/8.
2. Explain any five defuzzification methods.

Course Outcome 3(CO3):

1. List the stage involved in Back Propagation Algorithm?

Course Outcome 4 (CO4):

1. Explain different types of Encoding Techniques.

Course Outcome 5 (CO5):

1. Explain the characteristics and different classifications of a neuro-fuzzy hybrid
system.
 CHEMICAL ENGINEERING

Model Question paper

Name :Reg.No.
APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY
FIFTH SEMESTER B. TECH DEGREE EXAMINATION,
Month& Year
CHEMICAL ENGINEERING
CHT 397 Soft Computing Techniques
(2019 Scheme)
Duration: 3hours Max. Marks: 60

PART A

(Answer all questions. Each question carries 3 marks)

1. Distinguish between fuzzy and probability with example.
2. State the relevance of fuzzification. Explain different types.
3. What is unsupervised learning in ANN? Explain
4. Explain any training algorithm that can be used for training a multilayer feed forward
neural network.
5. Explain radial basis function network.
6. State the significance of error portion ᵟkand ᵟjin Back Propagation Network.
7. What is the concept of crossover in Genetic Algorithm?
8. Write short note on application of GA.
9. Explain Genetic Neuro - Hybrid systems
10. Discuss the classification of Neuro Fuzzy Hybrid System?

PART B
(Answer one full question from each module. Each question carries 14 marks)
Module 1
11. a) If membership functions A is {1/2 +0.5/3 + 0.6/4 + 0.2/5 + 0.6/6} and B is {0.5/2 +
0.8/3+0.4/4+0.7/5 + 0.3/6} find A/B and B/A ( 4 marks)
b) Two triangular membership functions are given as A = triang(1,3,4) and B=triang(3,5,6).
Draw membership functions of A ∩ B, A U B and A’. (5 marks)
c) For the fuzzy sets A = {0.2/x1 + 0.9/x2} and B = {0.3/y1 + 0.5/ y2 + 1/y3}, find the fuzzy
relation A×B (5 marks)
12. a) Using the intuition method develop fuzzy membership functions for the following
shapes. (a) Trapezoid. (b) Gaussian function. (c) Isosceles triangle. (4 marks)
b) Define Fuzzy Propositions. Explain different fuzzy propositions. (10 marks)

Module II
13. a)Defuzzify the following output membership function using centroid method:
0/0, 0.3/1 , 0.3/3.5 , 0.5/4 , 0.5/5.5 , 1/6 , 1/7 , 0/8. (4 marks)
b) In the following fuzzy system to predict power demand of a state, the following input
membership functions are used:
 CHEMICAL ENGINEERING

Temperature :triang(0 10 20), triang(10 25 40), triang(30 40 50). (for low, medium and high)
Month :triang(0 2 4), triang(2 6 10), triang(8 10 12), (for summer, rainy and winter)
The oputput membership function is
Power demand :triang(0 250 400), triang(250 500 750), triang( 600 750
1000), (for low, medium and high)
The rules are
1) If temp is low and month is winter, demand is low
2) If temp is medium and month is rainy, demand is medium
3) If temp is medium and month is winter, demand is high.
For a temperature of 16 and for a month of 9.5, find the power demand using centroid
defuzzification method. (5 marks)
c) Do problem 2b using weighted average defuzzification method. (5 marks)

14. a) Explain any five defuzzification methods. (5 marks)

b) What are the different steps involved in the design of fuzzy logic controller? Explain with
a typical example. (9 marks)

Module III
15. a) A perceptron is used to classify the inputs and targets mentioned below:
{ p1 = [1;1], t1 =[0;0]}, {p2 = [1;2], t2 = [0;0]}, { p3 = [2;-1], t3 = [0;1]}, p4 = [2;0], t4 =
[0;1]},
{p5= [[-1;2], t5 =[1 0]}, {p6=[-2;1], t6=[1;0]}, {p7=[-1;-1], t7=[1;1], {p8 = [-2;-2], t8=[1;1]}

Train the network with the initial weight and bias matrices of [1 0;0 1] and [1;1] respectively.
(4 marks)
b) A 1-2-1 network with log sigmoid transfer function in the hidden layer and linear transfer
function in the output layer is used to model a function which gives an output value of 1.261
for an input of 1. For the initial weight and bias matrices

of
find the weight and biases after one iteration of back propagation. (5 marks)
c) Realize an Exclusive-OR gate using a combination of Perceptron neurons. (5 marks)
16. a) State the significance of error portion ᵟkand ᵟjin Back Propagation Network.
(5 marks)
b) Explain the stage involved in Back Propagation Algorithm? (9 marks)

Module IV
17. a) Draw the flow chart of Genetic Algorithm based optimization process and enlist the
steps involved and the methods available in each step. (4 marks)
b) Do one iteration of finding the maximum of the following function using Genetic
Algorithm:
Maximize Y = 800 – 62.83(2D + 0.91D-0.2), 0 < D < 6.3
 CHEMICAL ENGINEERING

Choose (0.3, 1.5, 3, 5) as initial population and make 5th position in single point crossover.
Do mutations at 5th, 17th and 23rd positions. (5 marks)
c) It is required to solve the equation a+2b+3c+4d = 30 after converting this to a
maximization problem using Genetic Algorithm. Do one iteration. (5 marks)
18. a) Explain various coding techniques in GA with examples? (5 marks)
b) Explain Genetic Fuzzy Rule Based systems? (9 marks)

Module V
19. a) Explain the use of Genetic Algorithm in training a neural network (4 mark)
b) Explain how the performance of a PI controller can be improved with fuzzy system.
(5 marks)
c) Explain the concept of Neuro Genetic Systems (5 marks)
20. a) Explain the characteristics and different classifications of a neuro-fuzzy hybrid
system. (10 marks)
b) Explain Genetic Fuzzy Rule Based systems? (4 marks)

************

Syllabus
Module 1
Introduction to fuzzy logic: Fuzzy Sets-Fuzzy set operations-Fuzzy Relations-Cardinality of
fuzzy relations-operations on Fuzzy Relations-Properties of Fuzzy relations-properties of
fuzzy relations-Membership functions-features of membership functions-Fuzzification-
methods of membership value assignments-fuzzy rule base

Module 2
Defuzzification-Defuzzification methods-Fuzzy logic controller (Block diagram).artificial
neural networks. Basic concepts-Neural network architecture-Single layer feed forward
network-Multilayer feed forward network-Recurrent networks

Module 3
Characteristics of Neural networks- Learning methods. Perceptron networks-Back
propagation networks-Radial base function network-Hopfield network-Kohonenself-
organising maps.

Module 4
Fundamentals of genetic algorithms: Basic concepts-working principle-encoding-different
methods-fitness function-reproduction-different methods. Genetic modelling- inheritance-
cross over mutation-convergence of genetic algorithm.

Module 5
Hybrid systems: Neural network, fuzzy logic and genetic algorithm hybrids-Neuro fuzzy
hybrids-neuro genetic hybrids-Fuzzy genetic hybrids.
 CHEMICAL ENGINEERING

Text Books:
1. S. Haykins, Neural networks a comprehensive foundation, Pearson Education.
2. L. Fausett, Fundamentals of Neural Networks, Prentice Hall 1994.
3. Timothy J Ross, Fuzzy Logic with Engineering Applications, McGrawHill, New
York.

Reference Books
1. S. N. Sivanandam, S. N. Deepa, Principles of Soft Computing, Wiley India Pvt. Ltd.
2. S. Rajasekharanand and G.A. VijayalakshmiPai, Neural Network, Fuzzy Logic and
Genetic Algorithms- Synthesis and Applications, Prentice Hall of India.
3. D.E. Goldberg, Genetic Algorithms in search Optimization and Machine Learning,
Pearson Education.
4. AmitKonar, Artificial Intelligence and Soft Computing, First Edition,CRC Press,
2000.
5. John Yen, Reza Lengari, Fuzzy Logic- Intelligence, Control and Information, Pearson
Education.

Course Contents and Lecture Schedule

No. of
No Topic Lectures
(Hrs.)
1 Fuzzy logic (9 Hrs)
1.1 Introduction to fuzzy logic: Fuzzy Sets-Fuzzy set operations 2
1.2 Fuzzy relations-Cardinality of fuzzy relations-operations on Fuzzy 3
Relations-Properties of Fuzzy relations-properties of fuzzy relations-
1.3 Membership functions-features of membership functions 2
1.4 Fuzzification-methods of membership value assignments-fuzzy rule base 2
2 Defuzzification (9 Hrs)
2.1 Defuzzification methods 2
2.2 Fuzzy logic controller (Block diagram). 2
2.3 Artificial neural networks. Basic concepts 2
2.4 Neural network architecture-Single layer feed forward network- 3
Multilayer feed forward network-Recurrent networks
3 Characteristics of Neural networks (9 Hrs)
3.1 Characteristics of Neural networks- Learning methods. 2
3.2 Perceptron networks-Back propagation networks 2
3.3 Radial base function network-Hopfield network 3
3.4 Kohonen self-organising maps 2
4 Fundamentals of genetic algorithms (10 Hrs)
4.1 Fundamentals of genetic algorithms- Basic concepts-working principle 2
4.2 Encoding-different methods 2
4.3 Fitness function-reproduction-different methods. 3
4.4 Genetic modelling- inheritance-cross over mutation-convergence of 3
 CHEMICAL ENGINEERING

genetic algorithm.
5 Hybrid systems (8 Hrs)
5.1 Hybrid systems: Neural network, fuzzy logic and genetic algorithm 3
hybrids
5.2 Neuro fuzzy hybrids-neuro genetic hybrids 3
5.3 Fuzzy genetic hybrids. 2