IITM BS AI: Search Methods for Problem Solving (BSCS3003): Syllabus and Tips

AI: Search Methods for Problem Solving (BSCS3003) is a 4 credit core course at the BSc level of the IITM BS in Data Science. It is the classic side of AI: how a program finds a solution by searching, plays board games, plans actions and reasons with rules and constraints. There is no machine learning in it. The course page lists no prerequisite, and every student entering the BSc level must take it.

AI Search Methods at a glance

The instructor is Prof. Deepak Khemani of IIT Madras. The course page calls the type "Core Option II", and the handbook table tags it "Core_BD". Both give 4 credits and no prerequisite. The table (updated 18 March 2026) marks it as offered in May 2026, September 2026 and January 2027.

What you learn in this AI course

The page says 12 weeks of coursework but lists topics for weeks 1 to 11. Here they are in five groups.

• Weeks 1 and 2: what AI is and blind search. Some history and philosophy, the Turing Test and the Winograd Schema Challenge. Then state spaces and the basic search methods: depth first, breadth first and iterative deepening, with their analysis.
• Weeks 3 and 4: heuristics and local search. Using a heuristic to guide search, getting out of local optima, and random local search. Then methods that work with a population, like genetic algorithms and ant colony optimisation.
• Weeks 5 and 6: finding the best path. The A* algorithm, when it is guaranteed to find the optimal answer, the monotone condition, versions of A* that save memory, and sequence alignment as an example.
• Weeks 7 to 9: games, plans and goal trees. Minimax, alpha-beta pruning and SSS* for board games. Goal stack planning and partial order planning. Breaking a problem into parts with goal trees and the AO* algorithm.
• Weeks 10 and 11: rules and constraints. Rule based systems with forward chaining and the Rete algorithm. Then constraint processing: backtracking, arc consistency, the Waltz algorithm and model based diagnosis.

The main textbook is A First Course in Artificial Intelligence, written by the instructor himself (chapters 1 to 8 and parts of 9 and 10). The page also lists reference books, including Russell and Norvig.

How it is assessed

The course page lists weekly online assignments, 2 in-person invigilated quizzes and 1 in-person invigilated end term exam. No OPPE or project is listed.

Where it counts

This course is one of the five mandatory courses of the BSc degree level, next to Software Engineering, Software Testing, Deep Learning and Strategies for Professional Growth. It is not part of any minor.

Who finds it hard and how to prepare

If you expect neural networks, this course will surprise you. It is about algorithms you can trace step by step on paper. That makes it learnable, but only if you actually do the tracing.

• Revise PDSA first. Breadth first search, depth first search and priority queues from PDSA (BSCS2002) are the base for weeks 2 to 6.
• Trace search by hand. Take a small graph with 6 or 7 nodes and run A* on it. Write the open list, the closed list and the f, g and h values at every step.
• Draw game trees. Make a tree of depth 3 with made-up scores. Run minimax, then mark which branches alpha-beta would skip.
• Code two small problems. Solve the 8-puzzle with breadth first search and with A*, and write minimax for tic-tac-toe in Python. Count how many states each one explores.
• Learn the conditions, not only the steps. Know when A* is sure to find the best path, and why. These "why" points are easy to forget.

What to take before and after

The AI half of the BSc core pairs this course with Deep Learning (BSCS3004), which covers the learning side of AI. If you like the algorithm design here, Advanced Algorithms (BSCS4021) goes further into hard problems and ways to cope with them. For decision making over time, look at the Minor in Algorithmic Decision Making.

Common questions

Is there any machine learning in this AI course?

No. The syllabus is search, games, planning, rule based reasoning and constraints. The learning side of AI is in Deep Learning, the other AI related core course.

Do I need to write code in the exams?

The course page does not list an OPPE or a programming exam. The assessment is weekly assignments, two quizzes and an end term exam. Writing small programs is still the best way to understand the algorithms.

Are Prof. Khemani's NPTEL courses the same as this one?

The course page says he has NPTEL courses on search, knowledge representation and constraint satisfaction. It does not say the IITM BS course is the same. Use them as extra practice, and follow your own course portal for what is tested.

What happens if I exit with the BSc and rejoin later?

The handbook's re-entry rules say you must take final exams in Deep Learning, AI, Software Engineering and Software Testing to join the BS level again. So keep your notes.