B.TECH - Semester 8 soft computing techniques Question Paper 2019 (nov)
Practice authentic previous year university questions for better exam preparation.
Sample Questions
- (a) What are the different methods of encoding in Genetic Algorithm? 5
- Explain the definitions of Strong Minimum, Global minimum and Weak Minimum.
- What is meant by Simulated Annealing?
- What are the different reproduction techniques used in Genetic Algorithm?
- What is meant by decision boundary with respect to a Perceptron network?
- Design a Perceptron network which acts as an AND gate.
- Explain the LMS algorithm for training neural network.
- What is meant by reinforcement learning?. Explain with an example.
- Explain the centroid method of defuzzification.
- What is meant by fuzzy composition operations?
- What is meant by competitive learning? $$ (10 \times 4=40 \text { Marks) } $$ Answer any one full questions from each Module.
- (b) Do two iterations to minimize $\mathrm{Y}=800-62.83\left(2 \mathrm{D}+0.91 \mathrm{D}^{-0.2}\right)$ using Genetic Algorithm. Take four appropriate initial populations.
- (a) Explain the concept of conjugate gradient in optimization.
- (b) Do two iterations to minimize the function $F(x)=\left(x_{1}-x_{2}\right)^{4}+8 x_{1} x_{2}-x_{1}+x_{2}+$ 3 with an initial guess of [1.50] ${ }^{\top}$ using Newton's method.
- (a) What's the role of sensitivity matrix in training a multilayer neural network?
- (b) Derive the expressions for backpropagation learning rule. OR
- (a) Describe the architecture and learning methods of a Radial Basis Function network.
- (b) Explain the concept of modular networks.
- (a) Consider two fuzzy sets defined by
- (i) Approximately $3: 0.5 / 1,1 / 3.0 .5 / 4$ (ii) Approximately 4:0.812,0.9/3,07/5 Find the fuzzy sets for "approximately 12 " and "approximately 7 "
- (b) Two triangular membership functions are given as $\mathrm{A}=$ triang $(1,3,4)$ and $B=$ triang $(3,5,6)$. Draw membership functions of $A \cap B, A \cup B$ and $A^{\prime} \quad \mathbf{1 0}$
- (c) What are fuzzy modifiers? If $A=$ triangular ( $1,3,5$ ), draw the membership functions
- (i) very A and (ii) more or less A
- Defuzzify the output membership function ( $0 / 0,0.3 / 1,0.3 / 3.5,0.5 / 4,0.5 / 5.5,1 / 6,1 / 7,0 / 8$ ) using (a) Centroid method 12 (b) Centre of Sums method 8 ( $\mathbf{3} \boldsymbol{\times} \mathbf{2 0} \boldsymbol{=} \mathbf{6 0}$ Marks)