(SEM III) THEORY EXAMINATION 2021-22 INTRODUCTION TO SOFT COMPUTING

This question paper is from the B.Tech Semester III examination for the subject Introduction to Soft Computing (KOE036).
It carries 100 marks and covers the foundational concepts of artificial neural networks, fuzzy logic, genetic algorithms, neuro-fuzzy systems, and machine-learning-based decision-making.
The exam is structured into three sections – A, B, and C, progressing from basic conceptual questions to in-depth analytical and application-based problems.

SECTION A – Short Conceptual Questions (20 Marks)

This section includes 10 questions (2 marks each) focusing on basic definitions and core ideas of soft computing, such as:

Counting layers in neural networks

Purpose of single-layer perceptron

Human-like intelligence through fuzzy logic

Improving decision-making systems

Working of ANFIS in MATLAB

How classification & regression trees operate

Most expensive genetic algorithm operation

Fitness calculation in GA

Searching technique used in GA

Technologies behind neuro-fuzzy hybrid systems

This section tests fundamental understanding and recall of soft computing concepts.

SECTION B – Analytical & Application Questions (30 Marks)

Students must attempt any three out of five questions. Topics include:

Definition & characteristics of Artificial Neural Networks

Operations in fuzzy sets with examples

Applications of Kohonen’s Self-Organizing Map & properties of ART

Travelling Salesman Problem using genetic algorithms

Solving a nonlinear optimization problem using GA

These questions check understanding of ANN behaviour, fuzzy operations, unsupervised learning, GA-based search, and problem-solving techniques.

SECTION C – Long Answer Questions (50 Marks)

Section C contains five groups (Q3 to Q7), each with two choices. Students must attempt one question per group.

Q3 – Neural Networks & Machine Learning (10 Marks)

Why neural networks are called parallel distributed processing

OR applications of supervised machine learning in businesses

Q4 – Multilayer Networks & Fuzzy Algebra (10 Marks)

Multilayer feed-forward architecture with diagram

OR algebraic sum of two fuzzy sets

Q5 – ANN vs ANFIS / Neuro-Fuzzy Regression (10 Marks)

Difference between ANN & ANFIS

OR neuro-fuzzy modeling for regression test case prioritization

Q6 – Mutation & Evolution (10 Marks)

Meaning of “survival of the fittest” with example

OR types of mutation techniques in soft computing

Q7 – Fuzzy Control & MATLAB Simulation (10 Marks)

Using fuzzy logic control + GA for structural optimization

OR how to calculate MATLAB simulation computational time

This section assesses deep reasoning, technical understanding, system design, and hybrid soft-computing techniques.

OVERALL SUMMARY

The Introduction to Soft Computing (KOE036) paper evaluates a student's understanding of:

Neural networks (structure, learning, ANN vs ANFIS)

Fuzzy logic (sets, operations, control systems)

Genetic algorithms (fitness, mutation, optimization, TSP)

Neuro-fuzzy hybrid systems

Decision-making & machine learning concepts

Soft computing applications in business, engineering, and optimization

The paper combines theory, diagrams, derivations, computational questions, and real-world applications, ensuring a complete assessment of soft computing fundamentals.