[ Roll No.:........................ .
:Nationa{ Institute ofr:lecn:noCo8J, (])eChi
Name of the Examination: [Link].
Ap~,
Re-Mid Semester Examination (Spring, JO.8b.-2025)
Branch : CSE Semester : IV
Title of the Course : Artificial Intelligence Course Code : CSBB 252
Time: 1.5 Hour Maximum Marks: 25
Note: 1. Attempt all questions.
2. Read all questions carefully.
3. Explain all the desired variables.
[Link] parameters or values may be assumed.
COURSE OUTCOMES COGNITIVE
LEVELS
COl Understand the basic concept of AI Ll, L2
CO2 Apply search strategies to solve AI problems L3
Apply knowledge representation and reasoning to solve real world AI L3
C03
problem
C04 Explore machine learning concepts and algorithms for real world applications L4
Att empit a II ques tiIons from eac h secfIOn.
Section-A
(lx 5 = 5 Marks)
Qs. Question CO BL PO
No.
1. Can a search algorithm be complete but not optimal? Provide an COl Ll PI
example.
2. In A *, what are the consequences of using a heuristic that is not COl L2 PI
admissible?
3. Define and give an example of the state space explosion problem. COl Ll P2
4 Why is greedy best-first search not guaranteed to find the optimal path? CO2 L3 P2
5 In what situation is backward chaining more efficient than forward COl Ll PI
chaining?
Section-B
(2.5 x 4 = 10 Marks)
Qs. Question CO BL PO
No.
6. Genetic algorithms are difficult to use well. Explain three reasons why it CO2 L3 P2
is difficult to solve a problem with a genetic algorithm.
7. Did the A * algorithm under the HI heuristic find the least-cost path? Give CO2 L3 P2
proper justifications.
8. Suppose that we run a greedy search algorithm with the following two CO2 L3 P2
heuristics:
(a) hen) = -g(n)
(b) hen) = g(n)
What sort of search will the greedy search emulate?
9. Simulated annealing will find a global optimum with probability CO2 L3 P2
approaching 1 if you lower
T slowly enough. Please explain what simulated annealing will do in each
of the following
cases
(a) What happens if you lower T very quickly?
(b) What happens if you never lower T?
Section-C
(5 x 2 = 10 Marks)
Qs. Question CO BL PO
No.
10. Consider the following graph, where nodes represent cities, edges CO2 L3 P2
represent road distances, and heuristic values (h(n») represent the
estimated straight-line distance to the goal node (G).
Node hen) Neighbors (Cost)
A 10 B (3), C (4)
B 8 D (5), E (6)
C 6 F (4)
D 5 G (7)
E 7 G (3)
F 3 G (8)
G 0 -
(a) Apply the A algorithm* to find the shortest path from A ---+ G.
Show the steps, including the priority queue (OPEN list),
expanded nodes (CLOSED list), and fen) calculations at each
step.
(b) Compare two heuristics:
0 hl(n) (given in the table)
0 h2(n)=0 for all nodes (i.e., uniform-cost search).
0 Explain why the A * algorithm with hl(n) is more efficient
than using h2(n) .
(c) Suppose that the heuristic function hen) is inconsistent.
0 Provide an example of an inconsistent heuristic for this
graph.
0 Explain how A * might fail in this case.
11. Explain Genetic algorithm. Here are three populations from a genetic CO2 L3 P2
algorithm:
(a) 01010000 10000001 (b) 1001000001000001 (c) 0101000010100001
Population b was created by population a reproducing (the crossover is
after the second digit). Population c was created by population a mutating
(the second string was mutated). Let the fitness function be the number of
1's in a string. What is the maximum fitness of each population? Is there
a better crossover point than the one used for creating population b? If so,
where is it and why is it better?
