0% found this document useful (0 votes)
3 views3 pages

Soft Computing Techniques and Applications

The document outlines various topics related to Soft Computing, Fuzzy Logic, Rough Set Theory, Neural Networks, and Genetic Algorithms. It includes questions and tasks that require explanations, comparisons, and computations related to these areas, emphasizing their applications in real-world scenarios. The content is structured into sections, each focusing on different aspects of these computational techniques and their advantages over classical methods.

Uploaded by

utkarshd9990
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
3 views3 pages

Soft Computing Techniques and Applications

The document outlines various topics related to Soft Computing, Fuzzy Logic, Rough Set Theory, Neural Networks, and Genetic Algorithms. It includes questions and tasks that require explanations, comparisons, and computations related to these areas, emphasizing their applications in real-world scenarios. The content is structured into sections, each focusing on different aspects of these computational techniques and their advantages over classical methods.

Uploaded by

utkarshd9990
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

SECTION-A: Soft Computing Foundations & Fuzzy Logic

1. Describe a scenario in which classical deterministic computing provides an inadequate solution


but a Soft Computing technique produces a robust outcome. Explain why.

2. Compare the roles of Fuzzy Logic, ANN, and Evolutionary Algorithms in solving real-world
optimization or control problems.

3. Give two examples where uncertainty or imprecise input makes Soft Computing a better choice
than Hard Computing.

4. Explain “approximate reasoning” in Fuzzy Logic with a small real-life example.

5. List any four engineering applications where fuzzy systems outperform classical rule-based
systems. Give justification for two.

6. Write four fuzzy rules for an autonomous room heater that adjusts heating power using
parameters “Room Temperature” and “Humidity”.

7. Define a fuzzy linguistic variable. Using traffic density, construct 3 linguistic terms and sketch a
membership curve for each.

8. Explain α-cut and strong α-cut.

Given fuzzy set


A = {(0,0), (2,0.3), (4,0.6), (6,1)}
Find A₀.4 and A₀.7⁺.

9. A Mamdani fuzzy controller has rules:

• If Speed is High AND BrakePressure is Low → Risk is High

• If Speed is Medium AND BrakePressure is Medium → Risk is Moderate


For Speed=70 km/h (μ=0.8 High, 0.2 Medium) and BrakePressure=40% (μ=0.3 Low, 0.6
Medium), compute rule firing strengths.

10. For a TSK system with rules:

• If A is Large and B is Small → z = 3A + 2B

• If A is Small and B is Large → z = A + 5B


Given μ values for A,B, compute the weighted average output for A=3, B=4 where
membership values are provided by you.

SECTION-B: Rough Set Theory

11. Given a decision table with attributes {Color, Size, Shape}, define the indiscernibility relation
for each attribute.

12. Explain lower and upper approximation using an example of medical diagnosis.

13. Define:

(a) Positive region


(b) Boundary region
(c) Negative region
Give a simple dataset to illustrate all three.

14. A dataset contains attributes {A,B,C}. Show how reduct and core are computed and explain
why core attributes cannot be removed.

15. How can Rough Sets assist in dimensionality reduction in high-dimensional healthcare data
such as ECG or MRI features?

16. Describe one hybrid Soft Computing model where Rough Sets and Fuzzy Logic are combined.
Explain its advantage over classical RST alone.

SECTION-C: Neural Networks & Learning

17. Draw and describe the structure of an artificial neuron. Compare signal processing in biological
and artificial neurons.

18. A perceptron receives inputs x₁=0.5, x₂=1.0 with weights w₁=0.4, w₂=−0.2 and bias b=0.1.

Compute:
(i) Net input
(ii) Output using a step function threshold at 0.

19. For a perceptron classifier with three inputs, explain how decision boundaries are formed in
geometric space.

20. Using Hebbian learning, update weights:

Initial w=[0.2, −0.1], input x=[1, 1], target y=1, learning rate η=0.3.

21. Using Delta rule, update a single weight:

x=3, target=1, output=0.6, η=0.05.

22. In competitive learning:

Input vector = [0.7, 0.4]


Weight vectors: W₁=[0.9,0.1], W₂=[0.2,0.8]
Determine winning neuron using Euclidean distance.

23. Explain why competitive learning is suitable for clustering tasks. Give one real-world example.

SECTION-D: Genetic Algorithms & Optimization

24. Define: gene, chromosome, genotype, phenotype, fitness landscape. Explain their
interrelationships in GA.

25. A GA uses binary encoding for x ∈ [0,31].

Chromosomes:
P=[10110], Q=[01001]
Perform a single-point crossover at position 3.

26. For values x={4,7,10,12}, fitness f(x)=x³.


Compute selection probabilities and cumulative roulette wheel ranges.

27. Apply bit-flip mutation at bit-position 2 to chromosome 11010. Provide new chromosome.

28. Compare GA, PSO, and ACO in terms of convergence and search strategy.

29. A GA has population {3,8,14} (decimal).

Compute fitness for f(x)=2x+5 and identify the fittest individual.

30. Give one real-life engineering scenario where PSO outperforms GA. Provide technical
justification.

You might also like