- Created Wednesday, August 17th 2016 @ 08:21:44
I used alpha bita pruning algorithm for certain depth (maximum 10 depth) and after over depth, used a heuristic evaluation for each last child node. I think my current heuristic evaluation is not good enough. Now can anyone give me a better heuristic strategy evaluation or better idea. Any tutorial link will be good too . Thank you
- Created Wednesday, August 17th 2016 @ 10:44:05
Look in theme "How deep do the top ais search?" There are good examples of technics you can use at the start.
- Updated Wednesday, August 17th 2016 @ 14:55:57
Hi, I found these to be a nice read :
and generally any kinda google search to do with the normal tic tac toe.
For evaluation you might want to play a bit of the game yourself and come up with an idea because there is no set stone for a perfect evaluation.
Also remember to find a balance between heavy evaluations and light evaluations. Since it will affect the depth you reach in x time (mentioned by ghooo below).
From what ghooo said on my thread ( http://theaigames.com/discussions/ultimate-tic-tac-toe/57893ed25d203c1f6ccc86c7/how-deep-do-the-top-ais-search-/2/show ):
Fast Board Representation Negamax & AlphaBeta Pruning Opening Book Transposition Table Horizon Effect & Quiescent Search Endgame Database Move Ordering Negascout Killer Moves History Heuristics Heavy or Light Evaluation Functions (will affect your search depth). Iterative Deepining And much more, but this list is good enough to get you on the right track
and another discussion :
You might also find watching the games in which your bot gets demolished by better bots a way to learn how to evaluate.
Or even playing the game against other players online. (human vs human or human vs bot)
- Created Thursday, August 18th 2016 @ 10:10:37
Thanks a lot for these suggestions :)