ISRO CS 2025 Q.20: solution and answer

Official answer: A or CSource: ISRO's final answer key of 15 May 2026

1. The question asks which algorithm is used for search or optimization in AI. Read literally, an option qualifies if it is a strategy that AI systems use to search a space of states or to optimize an objective.

2. (c) Hill Climbing starts from a state and repeatedly moves to the neighbouring state with the best objective value, stopping at a peak. It is the textbook local search and optimization method in AI, so it qualifies.

3. (a) Greedy Algorithm also qualifies. A greedy strategy commits at each step to the choice that looks best by a local criterion. AI search uses it directly: greedy best-first search expands the node with the lowest heuristic value, and hill climbing is itself a greedy local search. Greedy methods are also used as optimization heuristics. The question only asks for a method used for search/optimization, so (a) satisfies it as printed.

4. (b) Recursive Algorithm describes how a procedure is written (it calls itself). It is an implementation technique, not a search or optimization strategy, so it does not qualify.

5. (d) Hot Potato Algorithm is a routing technique in networks: a router forwards a packet out immediately instead of queueing it. It has nothing to do with AI search, so it does not qualify.

6. Both (a) and (c) are valid readings of the question. The most characteristic AI answer is (c) Hill Climbing Algorithm.