Assignment for Soft Computing

2 Marks
Write down the difference between ANN and BNN (Brain Neural Network).
Define soft computing with suitable example.
Define the term ‘Hard Computing’ with suitable
examples.
Define Hybrid computing with suitable examples.
State the meaning of the term with respect to
Computing: Antecedent, Consequent and Mapping.
State a few examples of biological inspired
methodologies used is soft computing.
Define Fuzzy set with suitable example.
State the laws of Excluded middle and contradiction for
a fuzzy relation.
Define: Multivalued logic.
List the operations defined on fuzzy propositions.
Define: T-norm operator.
State the Mamdani & Larsen FI Rules.
State Goguen’s and Boolean Fuzzy rules.
Define: Relative and Fuzzy cardinalities of a fuzzy set.
Define defuzzification method.
List the properties of  -cut for fuzzy sets.
Define idempotency and involution for any fuzzy set z.
Define a fuzzy binary relation.
Define the FAM table with a suitable example.
Discuss the limitations of defuzzification.
State the three most common inferring procedures used in proposition logic.
State the ways to interpret ‘A Entails B’.
Define: Activation function with respect to ANN.
Define: Threshold with respect to ANN.
Define: Bias with respect to ANN.
Define: Learning rate with respect to ANN.
Define population diversity with respect to GA.
State the Converging criteria of Gradient descent.
State the convergence criteria of McCulloch-Pitts neuron.
List the parameters of Gradient descent.
List the advantages of binary encoding.
List the disadvantages of binary encoding.
Describe the scenario where crisp set is preferred over
fuzzy set.
Explain a fuzzy set with an example.
Explain the term crossover points for fuzzy set.
 Assignment for Soft Computing

Explain Fuzzy Singleton.

Explain Fuzzy Bandwidth.
Discuss the limitation of fuzzy logic.
Discuss about the membership values that is obtained
for a crisp set Vs a fuzzy set.
Describe the fuzzy implication rule.
Explain GMP and GMT IN Fuzzy Logic.
Describe Sugeno Defuzzification method.
Explain the containment & identity operations on fuzzy
relations.
Let P be a fuzzy set defined as follows:
P= {(0,0.01), (9,0.54), (19,0.29), (29,0.13), (39,0.69),
(49,0.85), (59,0.9)}
Identify the scalar & relative cardinalities of P.
Explain First of Maxima and Last Maxima.
Discuss the canonical form of fuzzy proposition.
Explain the salient features of an ANN.
Describe Recurrent neural network.
Describe Identity activation function.
Describe ReLU activation function
Describe Ramp activation function.
Discuss the biological motivations behind GA.
Discuss about how do we check convergence of GA.
Discuss different genetic operators.
Discuss the reasons that roulette-wheel selection may
cause less diversity in population.
If the rank assigned to the ith individual is ri, then whatamount of area will be allotted to it in roulette
wheel incase of rank based selection?
Demonstrate how boundary is related to the support
and core of a fuzzy set.
Implement the truth table for three valued logic(0,1/2,1)
with the operations: Negation, Conjunction, equality and implication.
Sketch the graphical implementation of fuzzy
proposition.
Interpret the following propositions using Zadeh’s max-
min rule in terms of a fuzzy relation Rmm :
i. If x is A then y is B
ii. If x is A then y is B else y is C
Two fuzzy sets are defined as follows:
A= {(x1,0.25), (x2,0.75), (x3,0.68), (x4,0.98), (x5,0.49)} B= {(x1,0.39), (x3,0.88), (x4,0.09)} Implement (AUB)’
AND (A’UB’)
Demonstrate where A is any fuzzy set graphically. Hence, prove Law of contradiction.
Demonstrate A’ where A is any fuzzy set graphically. Hence, prove Law of contradiction.
 Assignment for Soft Computing
Let, P & Q be the propositions with their truth values
given below:
P: Jack is Strong; T(P) = 0.65 Q: Kate is Strong; T(Q) = 0.87
Compute the truth values of the following propositions using fuzzy connectives.
i. If Kate is Strong then Jack is also Strong
ii. Kate and Jack are equally strong
Let A be a fuzzy set defined as A= {(1,1), (2,0.8),
(3,0.6),(4,0.4), (5,0.2)}
Compute the concentration and dilation of A.
Two fuzzy set are given by:
A= {(-1,0.5), (0,1), (1,0.5), (2,0.3)}
B= {(-1,0.2), (0,0.8), (1,0.3), (2,0.9)}
Compute the vector product A & B and the scaler product of (0.7)B
Two fuzzy sets are defined on follows:
A={(a,0.2), (b,0.72), (c,0.49), (d,0.28)}
B={(a,0.29), (b,0.9), (c,0.54), (d,0.32)}
Justify: A×B B×A
Two fuzzy sets are defined as follows:
A={(y1,0.5), (y2,0.3), (y3,0.01), (y4,0.05)}
B={(y1,0.25), (y2,0.39), (y3,0.78), (y4,0.96)}
Evaluate: A-B and B-A
Justify: Fuzzy bandwidth in dependent on the
crossover points of fuzzy set.
Write down the Selection methods used in Genetic algorithm.
Show an example of Linear Real Coded crossover operator?
What is Hyper Parameter in Support Vector Machine?
How to select the chromosome from Initial Population in Genetic Algorithm?
What is fitness score?
What is the utility of Selection phase in Genetic Algorithm?
Why we use crossover point?
How many crossover operators are there? Write down their names?
Explain Multipoint crossover operator with an example.
Write down the condition for generating the offspring 1 using Three point crossover operator.
What is Soft Marginal Plane in Support Vector Machine?
Explain ‘Eta’ in cost function in Support Vector machine.
What we can infer from Pareto front?
 Assignment for Soft Computing

5 Marks
Consider the fuzzy propositions:
U = Brad is tall; T(U)=0.72
V = Brad us intelligent; T(V)=0.48
Compute the truth values of the following statements:
i. Brad is tall or intelligent
ii. Brad is neither tall not intelligent
iii. If Brad is intelligent then he is tall too
Brad is either tall or intelligent
State the feature of Hard computing and soft
Computing.
Define: Universal fuzzy sets, Empty fuzzy sets and
Equal fuzzy sets with proper examples.
Define selection. List the different selection methods
used in GA.
Discuss the idea behind fuzzy logic. How does it differ
from the crisp logic? Explain with a suitable example.
Describe the feature of fuzzy logic.
Explain convex fuzzy set with examples and diagrams.
Describe Triangular and Trapezoidal Membership
Functions with suitable diagrams.
Describe the Bell-shaped Membership Function with
suitable diagram.
Describe the Gaussian and Sigmoid Membership
Function with suitable diagrams.
Describe the industrial applications of fuzzy logic
controller.
Describe Sigmoid activation function along with its
significance to obtain network output.
Discuss the drawbacks of simple genetic algorithm.
Demonstrate the support, core and boundary of a fuzzy
set with proper diagrams.
Sketch the simple architecture of ANN with detailed
notations and the calculation of net input.
Sketch the architectural design of a single-layer feed-
forward neural network.
Sketch the architectural design of a multi-layer feed-
forward neural network.
Implement NOT function using McCulloch-Pitts neuron
taking binary data.
Interpret the Delta rule for single output unit.
Interpret the training rule of Gradient Descent. Hence,
implement the steps of Gradient descent method.
 Assignment for Soft Computing
Differentiate with suitable example: Fuzzy Relation Vs
Classical relation.
Contrast between a normal and a subnormal fuzzy set.
Also, State how height of the fuzzy set acts as adetermining factor of normality.
Differentiate: Supervised Learning vs. Unsupervised
Learning with suitable diagram/examples.
Differentiate: Biological Neural Network vs Artificial
Neural Network.
Justify how Ramp function differs from the ReLU
function.
Justify: Bipolar inputs are more effective than binary
inputs.
The membership functions for a transistor(T) and a
resistor (R) are given below:
= {(0,0), (1,0.2), (2,0.7), (3,0.8), (4,0.9), (5,1)}
= {(0,0), (1,0.1), (2,0.3), (3,0.2), (4,0.4), (5,0.5)}
Evaluate the following:
a. Algebraic Sum
b. Algebraic Difference
c. Algebraic Product
d. Bounded Sum
Bounded Difference
Consider the fuzzy relation:

Evaluate the  -cut relations for  = 0, 0.1, 0.5, 0.7.

“ANN structure is inspired by the parts of a biological
neuron”-Justify.
Justify the importance of softmax activation function.
Given two Universe of discourses:
X= {1,3,5}
Y= {2,4,6,8}
Let, R and S be two relations defined on X Y as:R={(x, y) | y=2x }
S={(x, y) | [(y =x-1) ] }
Construct R and S in matrix form and hence find RUS .
 Assignment for Soft Computing

Explain how Genetic algorithm works with a flow chart.

Find the formula of Final Constraint from the equation of straight line.
Solve a problem f(x)=x3, where 0 <= x<= 15. Find some stages of application of genetic algorithm.
Deduce the formula of final constraint from the formula of straight line.
Explain Precision and Recall in Confusion Matrix and also write down the formula for both of them.

On the basis of dataset given classify the given set is Underweight or Normal.

Create offspring for P1 and P2 using Cycle Crossover technique.

If a student studies for 5 hours whether he will pass or not? Use Logistic Regression algorithm.
Write down the difference between Logistic Regression and Linear Regression.
 Assignment for Soft Computing

10 Marks
List the different mutation operations available for
binary GA. Explain them.
Discuss the application domain of ANN.
Describe the training of McCulloch-Pitts neuron.
Discuss the advantages and the limitations of an ANN.
Describe the features of a perceptron with a proper
structure diagram.
Apply the Gradient Descent training rule to find the
minimum value of the function f(x)=x2+x+1, given the initial value of x=1, learning rate(  ) =
0.01, tolerance
= 0.25, precision upto 4 decimal places.
Design a Computer Software to perform image
processing to locate objects without a scene. The two fuzzy sets representing a Plane and a Train
image are:Plane = { (train, 0.2), (bike, 0.5), (boat, 0.3), (plane, 0.8), (house, 0.1) }
Train = { (train, 1), (bike, 0.2), (boat, 0.4), (plane, 0.5),
(house, 0.2) }
Determine the Following:
a. Plane * Train (Algebraic/Vector Product)
b. Plane + Train (Algebraic Sum)
c. Plane – Train (Algebraic Difference)
d. Plane Train (Bounded Sum)
e. Plane Train (Bounded Difference)
f. Prove the Associativity for Union and
Intersection for both the sets.
Examine the DeMorgan’s law on both the sets.
Consider the following proposition:
P: If job is risky then compensation is highQ: compensation is quite low
The universe of discourses are given as:
job = {1, 2, 3, 4}
compensation = {a, b, c, d}
All the elements are arranged in ascending order.The fuzzy sets for the propositions are
given as: risky job = {(1, 0.3) (2, 0.8), (3, 0.7), (4, 0.9)}
high compensation = {(a, 0.2) (b, 0.4), (c, 0.6), (d,0.8)}
Apply Zadeh’s interpretation to find the fuzzy relationfor the Rule P.
Hence, find the conclusion caused by the premise Q
using appropriate Fuzzy Inference rule, given,
quite-low = {(a, 0.1) (b, 0.2), (c, 0.3), (d, 0.5)}
Implement AND function using McCulloch-Pitts neuron
taking binary data.
Implement OR function using McCulloch-Pitts neuron
taking binary data.
 Assignment for Soft Computing
Implement AND-NOT function using McCulloch-Pitts
neuron taking binary data.
Implement XOR function using McCulloch-Pitts neuron
taking binary data.
Implement NOR function using McCulloch-Pitts neuron
taking binary data.
Implement NAND function using McCulloch-Pitts
neuron taking binary data.
Compare between different types of computing after
defining the term ‘computing’ and stating it's vitalfeatures.
 Assignment for Soft Computing
A1, A2, and A3 are three fuzzy sets as shown in the
following figures. Examine the aggregation of theaforementioned three sets and then
defuzzify by:
(i) weighted average method
(ii) centre of sums method
(iii) mean of max method
 Assignment for Soft Computing

Differentiate: Multi-layer feed-forward neural network

vs. Single-layer feed-forward neural network.
In a simple vision recognition system, we may want to
locate specific objects within a scene containing manyobjects. Let X be a universe of general,
well-known objects such as X = {car, boat, house, bike, tree, mountain}, - and let Y be a
universe of simple geometric shapes such as Y= {square, octagon, triangle, circle, ellipse}.
We now define simple fuzzy set of objects, such asthose for 'car' and 'square'.
A = car= {1.0/car+0.4/boat + 0.1/house +0.6/bike
+0.1/tree +0/mountain}
B= square= {1.0/square +0.5/octagon +0.4/triangle
+0/circle + 0.1/ellipse}
(a) Evaluate the relation between 'car' and 'square'. Now suppose we define another fuzzy set
of objects,called 'shapes with corners' as S= 'Shapes with corners' = {0.6/square +0.9/octagon
+0.4/triangle + 0/circle +0.2/ellipse}
Using an appropriate composition operation, find the fuzzy set T= 'objects using corners' using
the following:
a. Max-min composition
Max-Product composition
Let us say we build a temperature controller, but
instead of building a thermostatic system that controls temperature to a desired specific value of,
say, 72°F or85°F, we would like to indicate linguistic comfort ranges such as 'slightly cold' and
'not too hot’. The membership functions for the atomic terms defined on a universe of
temperatures in degree Fahrenheit are asfollows.
'Hot'= {0/50+0/60+0.1/70+0.5/80 +0.9/90 +1/100}‘Cold'= {1/50+0.9/60+0.3/70+0/80
+0/90+0/100}
Evaluate the membership functions for
a. Not very Hot
b. Slightly cold or slightly hot
c. Not very cold and not very hot
Very very hot and not cold
Justify- "Implementation of XOR function is not
possible by Hebb net".
Evaluate the weights required to perform the following
classification using perceptron network. The vectors (1,1, 1, 1) and (-1, 1, -1, -1) are belonging to
the class (so
have target value 1), vectors (1, 1, 1, -1) and (1, 1, - 1,
1) are not belonging to the class (so have target value -1). Assume learning rate as 1 and initial
weights as 0.
 Assignment for Soft Computing
Consider a universe of aircraft speed near the speed of
sound as
X = {0.72, 0.725, 0.75, 0.775, 0.78} and a fuzzy set onthis universe for the speed “near 0.75” =
M, where
M = { (0.72, 0), (0.725, 0.8), (0.75, 1), (0.775, 0.8),
(0.78, 0) }
Another universe is defined for altitudes as Y = {21, 22,23, 24, 25, 26, 27} in thousands of feet
and a fuzzy seton this universe for the altitude fuzzy set “approximately 24000 ft” = N, where
N = { (21, 0), (22, 0.2), (23, 0.7), (24, 1), (25, 0.7), (26,
0.2), (27, 0) }
a. Construct a relation R = M × N
b. For another aircraft speed, say P, where
P = { (0.72, 0), (0.725, 0.8), (0.75, 1), (0.775, 0.6),
(0.78, 0) }
Find relations:
i. A = P  R, using Max-min composition
B = P . R, using Max-product composition
Let X={a, b, c, d} and Y= {1, 2, 3, 4}
A = {(a, 0) ,(b, 0.8), (c, 0.6), (d, 1)}, B= {(1, 0.2), (2, 1),
(3, 0.8), (4, 0)}
C={(1, 0), (2, 0.4), (3, 1), (4, 0.8)}
Construct the implication relations
(i) IF x is A Then y is B
IF x is A Then y is B else y is C
Construct a feed-forward network with 5 input nodes, 3
hidden nodes and 4 output nodes. Describe the notations and clearly formulate the connection
matriceswith net inputs at each hidden and output layer, and the outputs generated.
Construct a recurrent network with 4 input notes 3
hidden notes 2 output notes that has feedback links from the hidden layer to the input layer.
Describe thenotations and clearly state the net inputs at each hidden and output layer.
 Assignment for Soft Computing

i) If the weather is Sunny and Wind is Weak then what will be probability of Playing Football and
Not Playing Football?
ii) If the Weather is Rain, Temperature is Cool, Humidity is Normal and Wind is Weak then what
will be probability of Playing Football and Not Playing Football?
Apply Naïve Bayes classifier.
What is support vector machine .Discuss in detail?

Give the Decision Trees for the above set of training example.
Explain different cross over operations performed in GA
How does the choice of kernel function in a Support Vector Machine (SVM) affect its performance in
classification tasks, and what are the advantages and disadvantages of popular kernel functions such as
linear, polynomial, radial basis function (RBF), and sigmoid kernels?