AI · Theory

Adversarial Search

Theory

25/215

Adversarial Search

Adversarial search is for games with an opponent who tries to make you lose. Chess, tic-tac-toe, Go. You cannot assume the other player will help. Minimax and alpha-beta live here. Chance games add dice (expectimax).

Not the same as finding a path on a map with no enemy. Say “opponent” in the first sentence.

Adversarial Search — output — three exam lines. Graph on paper first.

Diagram

start
    ├── BFS  queue
    ├── DFS  stack
    └── A*   g+h
    │
    ▼
   goal

Exam tip

Why BFS is the wrong tool for chess.

Example

# Adversarial Search
print("draw 6 nodes")
print("name the DS: queue / stack / priority")
print("say if the path is optimal")

Adversarial Search — output: three exam lines. Graph on paper first.

Short notes

DefSearch vs an opponent.
Tools : minimax, alpha-beta.

AI · Theory

Adversarial Search

← All stacks

Theory

25/215

Adversarial Search

Not the same as finding a path on a map with no enemy. Say “opponent” in the first sentence.

Adversarial Search — output — three exam lines. Graph on paper first.

Diagram

start
    ├── BFS  queue
    ├── DFS  stack
    └── A*   g+h
    │
    ▼
   goal

Exam tip

Why BFS is the wrong tool for chess.

Example

# Adversarial Search
print("draw 6 nodes")
print("name the DS: queue / stack / priority")
print("say if the path is optimal")

Adversarial Search — output: three exam lines. Graph on paper first.

Short notes

DefSearch vs an opponent.
Tools : minimax, alpha-beta.

Adversarial Search

Example

Short notes

Questions

Adversarial Search

Example

Short notes

Questions