jacob5567 / connect-4-ai

A simple AI that can play Connect 4

Connect 4 AI

There is no lookup table for this project; all the work is done in the PlayerX and PlayerO files. This allows for less work to be done overall. The parameters for the algorithm can be changed by changing the constants on the top of each file.

SCORE_BONUS_4 is the bonus given to the score if the AI player has a 4-in-a-row on a given board.

SCORE_DETRACTION_4 is the amount of points taken away from the score score if the other player has a 4-in-a-row on a given board.

SCORE_BONUS_3 is the bonus given to the score if the AI player has a 3-in-a-row on a given board.

SCORE_BONUS_2 is the bonus given to the score if the AI player has a 3-in-a-row on a given board.

SCORE_DETRACTION_3 is the amount of points taken away from the score score if the other player has a 3-in-a-row on a given board.

SCORE_DETRACTION_2 is the amount of points taken away from the score score if the other player has a 2-in-a-row on a given board.

PREFER_HIGHER_ROWS_MULTIPLIER is inverse to the amount by which the bonus points for a 3-in-a-row is multiplied for successively higher.

PREFER_CENTRAL_COLUMN_MULTIPLIER is the amount by which the AI prefers a central move to an edge move.

PREFER_NEARBY_COLUMNS_MULTIPLIER is the amount by which the AI prefers a move in a column adjacent to the center column.

ALPHA_BETA_DEPTH_LEVEL is the number of levels deeps the alpha-beta pruning tree goes. I found that 6 terminates in less than 3 seconds at the most, and 7 terminates in less than 8 seconds at the most.

The most optimal parameters are the ones that are preset in the files.

The "Connect4.py" file that has been provided must be in the folder with the PlayerX and PlayerO files for them to work.