Implemented a expectiminimax agent (2-ply search) with alpha – beta pruning and forward pruning (to reduce the branching factor in the game tree) to determine the best move give the state of the board.
Dependencies:
Python3
copy
numpy
To execute:
python backgammon.py
Input Format:
The first line of the input is 1 or 2 indicating its either white or black who is playing the game. 26 lines follow. Each line contains 1 or 2 integers (separated by a single space) indicating the number of checkers on the point and the type of the checker (1 for 1st player's checker and 2 for 2nd player's checker)
• 5 1 indicates that there are 5 checkers of the first player.
• 6 2 indicates that there are 6 checkers of the second player.
• 0 indicates that the point has no checkers. The first and the last line are bear off points for 2nd player and the 1st player respectively.
2 lines follow, each line indicating the number of coins in the bar for the 1st player and the 2nd player.
The next line contains an integer (2 or 4) indicating the number of dice rolls played on behalf of player by the computer. 4 – both the dice rolled the same number. 2 or 4 lines follow each indicating a number between 1-6 which is the number on the dice.
Output format:
For every possible legal dice roll possible, print a checker move in a new line. Each move has two integers (start point and end).
• 1 2 a checker is moved from 1 to 2.
• -1 5 a checker is removed from the bar and is placed at 5 on a dice roll of 5.
1 0 a coin is being bared off.
Example Input:
1
0
2 1
0
0
0
0
5 2
0
3 2
0
0
0
5 1
5 2
0
0
0
3 1
0
5 1
0
0
0
0
2 2
0
0
0
2
5
4