Using Minimax, implement an AI to play Tic-Tac-Toe optimally.
Complete the implementations of player
, actions
, result
, winner
, terminal
, utility
, and minimax
.
- The
player
function should take aboard
state as input, and return which player’s turn it is (eitherX
orO
). -In the initial game state,X
gets the first move. Subsequently, the player alternates with each additional move.- Any return value is acceptable if a terminal board is provided as input (i.e., the game is already over).
- The
actions
function should return aset
of all of the possible actions that can be taken on a given board.- Each action should be represented as a tuple
(i, j)
wherei
corresponds to the row of the move (0
,1
, or2
) andj
corresponds to which cell in the row corresponds to the move (also0
,1
, or2
). - Possible moves are any cells on the board that do not already have an
X
or anO
in them. - Any return value is acceptable if a terminal board is provided as input.
- Each action should be represented as a tuple
- The
result
function takes aboard
and anaction
as input, and should return a new board state, without modifying the original board.- If
action
is not a valid action for the board, your program should raise an exception. - The returned board state should be the board that would result from taking the original input board, and letting the player whose turn it is make their move at the cell indicated by the input action.
- Importantly, the original board should be left unmodified: since Minimax will ultimately require considering many different board states during its computation. This means that simply updating a cell in
board
itself is not a correct implementation of theresult
function. You’ll likely want to make a deep copy of the board first before making any changes.
- If
- The
winner
function should accept aboard
as input, and return the winner of the board if there is one.- If the X player has won the game, your function should return
X
. If the O player has won the game, your function should returnO
. - One can win the game with three of their moves in a row horizontally, vertically, or diagonally.
- You may assume that there will be at most one winner (that is, no board will ever have both players with three-in-a-row, since that would be an invalid board state).
- If there is no winner of the game (either because the game is in progress, or because it ended in a tie), the function should return None.
- If the X player has won the game, your function should return
- The
terminal
function should accept aboard
as input, and return a boolean value indicating whether the game is over.- If the game is over, either because someone has won the game or because all cells have been filled without anyone winning, the function should return True.
- Otherwise, the function should return False if the game is still in progress.
- The
utility
function should accept a terminalboard
as input and output the utility of the board.- If X has won the game, the utility is 1. If O has won the game, the utility is
-1
. If the game has ended in a tie, the utility is0
. - You may assume
utility
will only be called on aboard
ifterminal(board)
isTrue
.
- If X has won the game, the utility is 1. If O has won the game, the utility is
- The
minimax
function should take aboard
as input, and return the optimal move for the player to move on that board.- The move returned should be the optimal action
(i, j)
that is one of the allowable actions on the board. If multiple moves are equally optimal, any of those moves is acceptable. - If the
board
is a terminal board, theminimax
function should returnNone
.
- The move returned should be the optimal action
For all functions that accept a board
as input, you may assume that it is a valid board (namely, that it is a list that contains three rows, each with three values of either X
, O
, or EMPTY
). You should not modify the function declarations (the order or number of arguments to each function) provided.
Once all functions are implemented correctly, you should be able to run python runner.py
and play against your AI. And, since Tic-Tac-Toe is a tie given optimal play by both sides, you should never be able to beat the AI (though if you don’t play optimally as well, it may beat you!)