Jump to content

  • Log In with Google      Sign In   
  • Create Account

asking for tic-tac-toe AI help


Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

  • You cannot reply to this topic
2 replies to this topic

#1 timothyjlaird   Members   -  Reputation: 404

Like
0Likes
Like

Posted 15 September 2013 - 01:44 PM

I've been writing a simple tic-tac-toe game to get a grounding in Python and AI (minimax algorithm). I've got something working but I'm stuck on the AI part.

 

I *think* that I understand the theory of minimax in this case...given the current game state...you run through every possible outcome...if you hit a leaf that results in a win you add +1 to the root node that represents the next player choice...if the leaf results in a loss you subtract 1...then you play out the root node with the highest value? Is that correct?

 

I know this probably requires recursion but that has never been something I've been good at.

 

My code is pretty simple...I keep an unordered set of human moves made and a set of ai moves made. When a move is made by either the human or ai, it gets added to the appropriate set. I check each set at the end of the turn for a win or a tie. I've gotten it to the point where it randomly plays against itself (given a human set and an ai set with the list of moves already made) but I'm not quite sure where to go from there.

 

I hate doing this but I've been on this for hours...can someone give me a shove in the right direction? The code listing is below. Calling 'playcycle()' runs the game with prompts for human input and randomly chose 'computer' moves. Calling 'playoutaiall()' plays 10 random games and prints out the result for each.

import random

#checks a player set (either human or ai) for a possible win
def checkforwin(moves):
    if '1' in moves:
        if '2' in moves and '3' in moves:
            return True
        if '4' in moves and '7' in moves:
            return True
        if '5' in moves and '9' in moves:
            return True
    if '9' in moves:
        if '7' in moves and '8' in moves:
            return True
        if '3' in moves and '6' in moves:
            return True
    if '4' in moves and '5' in moves and '6' in moves:
        return True
    if '2' in moves and '5' in moves and '8' in moves:
        return True
    if '7' in moves and '5' in moves and '3' in moves:
        return True
    return False

#given human and ai move sets, returns remaining valid moves
def getmovesleft(playermoves, aimoves):
    allmoves = set('123456789')
    return allmoves - (playermoves | aimoves)

#given human and ai move sets, prints board state
def printboard(playermoves, aimoves):
    count = 1
    while (count < 10):
        if str(count) in playermoves:
            print('x', end="")
        elif str(count) in aimoves:
            print('o', end="")
        else:
            print(str(count), end="")
        if count % 3 == 0:
            print('\n', end="")
        count = count + 1

#returns whether a particular move is valid, given human and ai moves sets
def isvalidmove(newmove, playermoves, aimoves):
    return newmove in getmovesleft(playermoves, aimoves)

#given a set of remaining valid moves, returns one at a random
def generateaimove(aimovesleft):
    return random.sample(aimovesleft, 1)

#given human and player move sets, gets a valid move from the human player and returns it
def makehumanplayermove(playermoves, aimoves):
    print('It is your move! Current board state...')
    printboard(playermoves, aimoves)
    movesleft = getmovesleft(playermoves, aimoves)
    while True:
        print('Possible moves' , movesleft)
        move = input('Enter a move: ')
        if isvalidmove(move, playermoves, aimoves) == True:
            return move

#given sets of human and ai moves already made, returns a valid move
def makeaiplayermove(playermoves, aimoves):
    move = generateaimove(getmovesleft(playermoves, aimoves))
    return move

#given a player move and sets of existing (already played) human and ai moves, play out one turn (or cycle)
def singlecycle(pmove, playermoves, aimoves):
    playermoves.update(pmove)
    if checkforwin(playermoves) == True:
        return 'pwin'
    if len(getmovesleft(playermoves, aimoves)) == 0:
        return 'tie'
    aimove = makeaiplayermove(playermoves, aimoves)
    aimoves.update(aimove)
    if checkforwin(aimoves) == True:
        return 'aiwin'
    if len(getmovesleft(playermoves, aimoves)) == 0:
        return 'tie'
    return 'continue'

#given an end of game result, print it out for the human's benefit
def printgameresult(endresult):
    if endresult == 'pwin':
        print('The human won!')
    elif endresult == 'aiwin':
        print('The AI won. No suprise there. Stupid human...')
    elif endresult == 'tie':
        print('A tie game? How boring!')

#given a list of existing player and ai move sets, play out a complete game at random (no human input)
def playoutai(playermoves, aimoves):
    while True:
        randomplayermove = makeaiplayermove(playermoves, aimoves)
        cycleresult = singlecycle(randomplayermove, playermoves, aimoves)
        if cycleresult != 'continue':
            printboard(playermoves, aimoves)
            print('ai moves:' , aimoves)
            print('player moves:', playermoves)
            printgameresult(cycleresult)
            break

#play out 10 complete games
def playoutaiall():
    count = 0
    while count < 10:
        player = set()
        ai = set()
        playoutai(player, ai)
        count = count + 1

#play out a complete game with human input
def playcycle():
    playermoves = set()
    aimoves = set()
    print('Welcome to a new game!')
    
    while True:
        move = makehumanplayermove(playermoves, aimoves)
        cycleresult = singlecycle(move, playermoves, aimoves)
        if cycleresult != 'continue':
            printgameresult(cycleresult)
            break
        
    #end of current game
    print('End board state...')
    printboard(playermoves, aimoves)

I think all my base logic is there for AI but I'm not sure where to go from here. Any help would be much appreciated.

 

Thanks,

John


Edited by timothyjlaird, 15 September 2013 - 01:46 PM.


Sponsor:

#2 dejaime   Crossbones+   -  Reputation: 4119

Like
1Likes
Like

Posted 15 September 2013 - 07:23 PM

This: http://www.ntu.edu.sg/home/ehchua/programming/java/JavaGame_TicTacToe_AI.html

 

Maybe it helps you, looks good.



#3 timothyjlaird   Members   -  Reputation: 404

Like
0Likes
Like

Posted 16 September 2013 - 07:12 PM

Thanks! That got me started in the right direction again.






Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.



PARTNERS