Game Playing in AI and minimax algorithm.pptx

Game playing in AI and minimax algorithm

1. UNIT II Game Playing

2. Introduction Game playing has been a classic problem in AI since the 1950s. Provides a controlled environment to test AI techniques. Games involve: Players (human or machine) Rules (define valid moves) Strategies (plans to win) Outcomes (win/lose/draw, or points)

3. Why AI in Games? Testbed for search algorithms and decision-making . Simulates real-world problem-solving in a simplified form. Helps in developing techniques like: Search (DFS, BFS, A*) Heuristics Minimax & Alpha-Beta pruning Reinforcement Learning

4. Examples of Games in AI Chess → Deep Blue (IBM) defeating Garry Kasparov. Go → AlphaGo (DeepMind) beating world champions. Tic-Tac-Toe, Checkers, Pac-Man → teaching search & strategy. Game playing in AI = searching through possible moves to choose the best strategy under given rules.

5. Minimax Algorithm 🌟 What is Minimax? : It is a backtracking technique which is used in Game theory and decision making. A decision-making algorithm used in two-player games (e.g., Chess, Tic-Tac-Toe). Players: MAX → tries to maximize score (AI). MIN → tries to minimize score (opponent). Assumes both players play optimally .

6. How it Works Represent the game as a tree of moves . At each node: MAX picks the highest value move . MIN picks the lowest value move . Continue until reaching leaf nodes (game-over states). Backtrack values up the tree → root node gives best move for MAX .

7. Key Features Adversarial Search → AI vs Opponent. Works with zero-sum games (one’s gain = other’s loss). Guarantees optimal decision if search goes to the end.

8. Limitations Slow for complex games such as chess 35 moves Depth: 100 35 100 very big

9. Numerical Problem No. 1

10. Problem No.2

11. Problem No. 3

12. Problem no. 4