Alpha-Beta Pruning Demonstration

In class, we examined some example game tries to explore the use of the minimax algorithm for computing the optimal utility of a game tree, as well as the more efficient alpha-beta pruning algorithm. A copy of the examples we used can be downloaded here.

You may see an animated demonstration of both algorithms on these very same examples on turing as follows. Go to the directory /Public/goldwasser/362/games. Then you may execute any of the following commands.

• For the standard minimax algorithm, the five examples can be seen as:

  python solve.py random -v all -z --vs=1000 -t all --vd=0 -A minimax -b 2 --oddB=3 -s 5
  python solve.py random -v all -z --vs=1000 -t all --vd=0 -A minimax -b 2 --oddB=3 -s 9
  python solve.py random -v all -z --vs=1000 -t all --vd=0 -A minimax -b 2 --oddB=3 -s 21
  python solve.py random -v all -z --vs=1000 -t all --vd=0 -A minimax -b 3 --oddB=2 -s 21
  python solve.py random -v all -z --vs=1000 -t all --vd=0 -A minimax -b 3 --oddB=2 -s 40

  Note that this puts you in manual mode where you must hit return to advance each step. You can instead give a fixed sleep time for it to advance with argument such as --vd=0.1.


• To get the more advanced alpha-beta algorithm (which is the default), simply omit the -A minimax argument from the above commands.