Game Playing in Artificial Intelligence

Game Playing in AI

It is a mixture of reasoning and creativity. Game playing is often called as Adversarial Search. Game playing is generally deemed to require intelligence. Game playing is a search problem with the following components:

1. The initial state of the game
2. Operators defining legal moves
3. Successor function
4. Terminal test
5. Goal test
6. Path cost

Game playing has been one of the most publicized areas of AI research. With the success of Deep Blue in 1997, another landmark was reached, a computer program that could defeat the best chess player in the world. Game playing happens to be a good domain to explore machine intelligence because of two main reasons:

i. They provide a structured task in which it is very easy to measure success or failure.

ii. They did not require a large amount of knowledge. They are solvable by straight forward search from starting state to the final (winning) state.

Game Tree

A game tree is defined as an instance of a tree in which the root node represents the state before any moves have been made, the nodes in the tree represent possible states of the games and arcs in the tree represent moves.