- Notifications
You must be signed in to change notification settings - Fork2
petetetete/halma-ai
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
An AI that uses an adversarial minimax search with alpha-beta purning todetermine the best subsequent move. The program also supports player vs.player games.
usage: halma <b-size> <t-limit> [<h-player>]
The final project for this course was structured around creating thisadversarial AI so that the teams in class could compete in one final tournamenton the last day of class.
Our AI worked its way through 4 games of this single-elimination tournament towin first place.
- The move-finding method is under the assumption that pieces cannot re-enterfriendly goals or leave enemy goals even during a jump chain.
- Computation is currently fixed to a ply-depth of 3 rather than usingdynamic depth.
- Turn time exhaustion will simply break the recursion with its current max.
- Peter Huettl -ph289@nau.edu
- Garrison Smith -gts35@nau.edu