Information Technology Engineering Sem 8 Soft Computing00068990cbsgs Question Paper PDF 2026 - Mumbai University | munotes
Loading PDF...
Questions asked in this paper
N.B.: 1. Question no. 1 is compulsory
2. Solve any Three questions out of remaining Five questions
Qu-1 Attempt any FOUR of the following
a) Demonstrate/outline the working of Roulette-wheel selection. 5
b) A single-layer neural network has the weights w = [0.2 0.45] with bias b=0.3. 5
It is given an input of I = [0.5 0.8 0.1 0.36]
Find/estimate the output if the sigmoidal activation function is used (slope = 0.3)
c) Demonstrate/Outline the excluded middle axioms, extended for fuzzy sets. 5
d) How do genetic Algorithms differ from conventional optimization algorithms? 5
Let us consider the discrete fuzzy set A = + + + + using
Zadeh’s notation, defined on universe. X = {a, d, e, f }
Qu-2 a) Using Mamdani fuzzy model design a fuzzy logic controller to determine the wash 10
time of a domestic washing machine. Assume that the inputs are dirt and grease on
cloths. Use three descriptors for each input variables and five descriptors for the output
variable. Derive a set of rules for control action and defuzzification. The design should
be supported by figures wherever possible. Show/Defend that if the clothes are soiled
to a larger degree the wash time will be more and vice-versa
b) Explain McCulloch Pitts neuron model with example. 10
Qu-3 a) Determine the weights after one iteration for Hebbian learning of a single neuron 10
network starting with initial weights w = [1 -1]. The inputs are X; -2], X2 = [2
3}, X3 = [1,-1] and learning rate c=1
b) What are Neuro-Fuzzy Systems? Explain the steps in Neuro-Fuzzy Hybrid System. 10
Qu-4 is Linear Separability? Explain with example why single layer perceptron is not 10
capable of solving Linearly Inseparable problems
b) Using the binary input/output row matrix shown in table-1 train a hetero-associative 10
network to store the input row vectors s = {s1,s2,s3,s4} to the output row vector t
= {tl,t2}. Obtain/predict the final weight matrix
Table-1: Input row vectors and output row vector t2}
Qu-5 Consider a 2-2-2 three-layer network as shown in figure-1. Perform calculations (upto 20
4 decimal places) assuming back-propagation learning for one iteration on the input,
and desired output patterns given in Table-2
a) Learning rate n = 0.8, Momentum o = 0.8 and Sigmoidal activation function
Figure-1: The MLP with initial weights
Qu-6 a) Describe Genetic Algorithms considering: Encoding, Selection, Crossover, Mutation, 10
and Stopping Condition for Genetic Algorithms
b) Let R and S be two fuzzy relations defined as: 10
a) Compute/Infer the result of using max-min composition
b) Compute/Infer the result of R:S using max-product composition
Read from the scan above, so a character or two may differ. The scan is the original.
Something wrong on this page? Report it and we will check it against the scan.
Information Technology Engineering Sem 8 Paper Path
Use this paper as one timed INFT mock, then compare Environmental Management, Digital Business Management, Project Management, and Big Data papers from nearby Mumbai University year sets.
Step: Open this Information Technology Engineering paper in the free viewer first.
Step: Compare the year set and nearby INFT subjects for repeated paper patterns.
Step: Use notes only for topics where marks were missed during practice.
Quick Help
Related Resources
Something wrong with this paper? Report it.