Homework Policy: Homework is uploaded after each class. Submit your work in LMS. Homework correction will be given next session. Students who volunteer to do exercises will get extra points.

Submission: Assignments should be submitted through LMS (Learning Management System). Code should be uploaded in Kaggle platform.

Collaboration Policy: You may form study groups and discuss problems with your classmates. However, you must write up the assignment solutions and the code from scratch, without referring to notes from your joint session.

The course project constitutes a significant part of the midterm evaluation (20%). Each team selects one topic from the provided project list. The project consists of two main parts:

1. Presentation: 10–12 minutes per team to be delivered next week
2. Report: Due one week after the presentation date

Submit the presentation and report (all are .pdf files) via email or Zalo. There are awards for best presentation and best report.

Project Workflow:

1. Identify a real-world problem
2. Model it as a search problem (state space search)
3. Apply at least 2–3 search algorithms (BFS, DFS, Greedy, A*, etc.)
4. Implement and test with code
5. Compare results (expanded nodes, time, solution quality)
6. Relate findings back to the real-world context

Suggested Report Structure:

1. Introduction: Real-world background of the problem
2. Problem Formulation: Search problem representation
3. Algorithms: Description of applied algorithms
4. Experiments & Results: Comparison table and visual examples
5. Discussion: Analysis of results and insights
6. Response to Reviewers: Questions and answers from presentation
7. Conclusion: Summary and findings

Project Example: 8-Puzzle

Purely theoretical version: "Initial state → Goal state. Apply BFS, GBFS (heuristic = misplaced tiles), A* (heuristic = Manhattan distance). Compare number of expanded nodes and solution depth."

Real-world interpretation: The 8-Puzzle can be modeled as seat arrangement in an international conference. There are 9 chairs (8 delegates + 1 empty spot). The organizer must arrange guests so that each delegate sits in the correct assigned seat. Valid operation = swapping a delegate with an adjacent empty chair. Goal = correct seating arrangement with the fewest moves.

Sample introduction section for the report: "In international conferences, the correct arrangement of delegates according to diplomatic protocols is essential. We can model this seating arrangement task as the 8-Puzzle problem, where each tile represents a delegate and the empty tile represents an unoccupied chair. The objective is to reach the final correct arrangement with minimal movements. In this study, we apply classical search algorithms (BFS, GBFS, A*) and compare their performance in terms of solution depth and number of expanded nodes, thereby evaluating the effectiveness of AI search methods for real-world organizational tasks."

Important Notes

• Creativity is required: Every project must be grounded in a real-world scenario (politics, society, traffic, healthcare, environment, education, etc.), not only in abstract puzzles.
• Response to Reviewers section is mandatory: This is what makes your project unique compared to standard coursework.
• Visualization encouraged: Teams are encouraged to use visualizations, simulations, or examples to make the application clearer.

Project Submission