I'm making a C++ program for the game chopsticks.
It's a really simple game with only 625 total game states (and it's even lower if you account for symmetry and unreachable states). I have read up minimax and alpha-beta algorithms, mostly for tic tac toe, but the problem I was having was that in tic tac toe it's impossible to loop back to a previous state while that can easily happen in chopsticks. So when running the code it would end up with a stack overflow.
I fixed this by adding flags for previously visited states (I don't know if that's the right way to do it.) so that they can be avoided, but now the problem I have is that the output is not symmetric as expected.
For example in the start state of the game each player has one finger so it's all symmetric. The program tells me that the best move is to hit my right hand with my left but not the opposite.
My source code is -
#include <iostream>
#include <array>
#include <vector>
#include <limits>
std::array<int, 625> t; //Flags for visited states.
std::array<int, 625> f; //Flags for visited states.
int no = 0; //Unused. For debugging.
class gamestate
{
public:
gamestate(int x, bool t) : turn(t) //Constructor.
{
for (int i = 0; i < 2; i++)
for (int j = 0; j < 2; j++) {
val[i][j] = x % 5;
x /= 5;
}
init();
}
void print() //Unused. For debugging.
{
for (int i = 0; i < 2; i++) {
for (int j = 0; j < 2; j++)
std::cout << val[i][j] << "\t";
std::cout << "\n";
}
std::cout << "\n";
}
std::array<int, 6> canmove = {{ 1, 1, 1, 1, 1, 1 }}; //List of available moves.
bool isover() //Is the game over.
{
return ended;
}
bool won() //Who won the game.
{
return winner;
}
bool isturn() //Whose turn it is.
{
return turn;
}
std::vector<int> choosemoves() //Choose the best possible moves in the current state.
{
std::vector<int> bestmoves;
if(ended)
return bestmoves;
std::array<int, 6> scores;
int bestscore;
if(turn)
bestscore = std::numeric_limits<int>::min();
else
bestscore = std::numeric_limits<int>::max();
scores.fill(bestscore);
for (int i = 0; i < 6; i++)
if (canmove[i]) {
t.fill(0);
f.fill(0);
gamestate *play = new gamestate(this->playmove(i),!turn);
scores[i] = minimax(play, 0, std::numeric_limits<int>::min(), std::numeric_limits<int>::max());
std::cout<<i<<": "<<scores[i]<<std::endl;
delete play;
if (turn) if (scores[i] > bestscore) bestscore = scores[i];
if (!turn) if (scores[i] < bestscore) bestscore = scores[i];
}
for (int i = 0; i < 6; i++)
if (scores[i] == bestscore)
bestmoves.push_back(i);
return bestmoves;
}
private:
std::array<std::array<int, 2>, 2 > val; //The values of the fingers.
bool turn; //Whose turn it is.
bool ended = false; //Has the game ended.
bool winner; //Who won the game.
void init() //Check if the game has ended and find the available moves.
{
if (!(val[turn][0]) && !(val[turn][1])) {
ended = true;
winner = !turn;
canmove.fill(0);
return;
}
if (!(val[!turn][0]) && !(val[!turn][1])) {
ended = true;
winner = turn;
canmove.fill(0);
return;
}
if (!val[turn][0]) {
canmove[0] = 0;
canmove[1] = 0;
canmove[2] = 0;
if (val[turn][1] % 2)
canmove[5] = 0;
}
if (!val[turn][1]) {
if (val[turn][0] % 2)
canmove[2] = 0;
canmove[3] = 0;
canmove[4] = 0;
canmove[5] = 0;
}
if (!val[!turn][0]) {
canmove[0] = 0;
canmove[3] = 0;
}
if (!val[!turn][1]) {
canmove[1] = 0;
canmove[4] = 0;
}
}
int playmove(int mov) //Play a move to get the next game state.
{
auto newval = val;
switch (mov) {
case 0:
newval[!turn][0] = (newval[turn][0] + newval[!turn][0]);
newval[!turn][0] = (5 > newval[!turn][0]) ? newval[!turn][0] : 0;
break;
case 1:
newval[!turn][1] = (newval[turn][0] + newval[!turn][1]);
newval[!turn][1] = (5 > newval[!turn][1]) ? newval[!turn][1] : 0;
break;
case 2:
if (newval[turn][1]) {
newval[turn][1] = (newval[turn][0] + newval[turn][1]);
newval[turn][1] = (5 > newval[turn][1]) ? newval[turn][1] : 0;
} else {
newval[turn][0] /= 2;
newval[turn][1] = newval[turn][0];
}
break;
case 3:
newval[!turn][0] = (newval[turn][1] + newval[!turn][0]);
newval[!turn][0] = (5 > newval[!turn][0]) ? newval[!turn][0] : 0;
break;
case 4:
newval[!turn][1] = (newval[turn][1] + newval[!turn][1]);
newval[!turn][1] = (5 > newval[!turn][1]) ? newval[!turn][1] : 0;
break;
case 5:
if (newval[turn][0]) {
newval[turn][0] = (newval[turn][1] + newval[turn][0]);
newval[turn][0] = (5 > newval[turn][0]) ? newval[turn][0] : 0;
} else {
newval[turn][1] /= 2;
newval[turn][0] = newval[turn][1];
}
break;
default:
std::cout << "\nInvalid move!\n";
}
int ret = 0;
for (int i = 1; i > -1; i--)
for (int j = 1; j > -1; j--) {
ret+=newval[i][j];
ret*=5;
}
ret/=5;
return ret;
}
static int minimax(gamestate *game, int depth, int alpha, int beta) //Minimax searching function with alpha beta pruning.
{
if (game->isover()) {
if (game->won())
return 1000 - depth;
else
return depth - 1000;
}
if (game->isturn()) {
for (int i = 0; i < 6; i++)
if (game->canmove[i]&&t[game->playmove(i)]!=-1) {
int score;
if(!t[game->playmove(i)]){
t[game->playmove(i)] = -1;
gamestate *play = new gamestate(game->playmove(i),!game->isturn());
score = minimax(play, depth + 1, alpha, beta);
delete play;
t[game->playmove(i)] = score;
}
else
score = t[game->playmove(i)];
if (score > alpha) alpha = score;
if (alpha >= beta) break;
}
return alpha;
} else {
for (int i = 0; i < 6; i++)
if (game->canmove[i]&&f[game->playmove(i)]!=-1) {
int score;
if(!f[game->playmove(i)]){
f[game->playmove(i)] = -1;
gamestate *play = new gamestate(game->playmove(i),!game->isturn());
score = minimax(play, depth + 1, alpha, beta);
delete play;
f[game->playmove(i)] = score;
}
else
score = f[game->playmove(i)];
if (score < beta) beta = score;
if (alpha >= beta) break;
}
return beta;
}
}
};
int main(void)
{
gamestate test(243, true);
auto movelist = test.choosemoves();
for(auto i: movelist)
std::cout<<i<<std::endl;
return 0;
}
I'm passing the moves in a sort of base-5 to decimal system as each hand can have values from 0 to 4.
In the code I have input the state -
3 3
4 1
The output says I should hit my right hand (1) to the opponent's right (3) but it does not say I should hit it to my opponent's left (also 3)
I think the problem is because of the way I handled infinite looping.
What would be the right way to do it? Or if that is the right way, then how do I fix the problem?
Also please let me know how I can improve my code.
Thanks a lot.
Edit:
I have changed my minimax function as follows to ensure that infinite loops are scored above losing but I'm still not getting symmetry. I also made a function to add depth to the score
static float minimax(gamestate *game, int depth, float alpha, float beta) //Minimax searching function with alpha beta pruning.
{
if (game->isover()) {
if (game->won())
return 1000 - std::atan(depth) * 2000 / std::acos(-1);
else
return std::atan(depth) * 2000 / std::acos(-1) - 1000;
}
if (game->isturn()) {
for (int i = 0; i < 6; i++)
if (game->canmove[i]) {
float score;
if(!t[game->playmove(i)]) {
t[game->playmove(i)] = -1001;
gamestate *play = new gamestate(game->playmove(i), !game->isturn());
score = minimax(play, depth + 1, alpha, beta);
delete play;
t[game->playmove(i)] = score;
} else if(t[game->playmove(i)] == -1001)
score = 0;
else
score = adddepth(t[game->playmove(i)], depth);
if (score > alpha) alpha = score;
if (alpha >= beta) break;
}
return alpha;
} else {
for (int i = 0; i < 6; i++)
if (game->canmove[i]) {
float score;
if(!f[game->playmove(i)]) {
f[game->playmove(i)] = -1001;
gamestate *play = new gamestate(game->playmove(i), !game->isturn());
score = minimax(play, depth + 1, alpha, beta);
delete play;
f[game->playmove(i)] = score;
} else if(f[game->playmove(i)] == -1001)
score = 0;
else
score = adddepth(f[game->playmove(i)], depth);
if (score < beta) beta = score;
if (alpha >= beta) break;
}
return beta;
}
}
This is the function to add depth -
float adddepth(float score, int depth) //Add depth to pre-calculated score.
{
int olddepth;
float newscore;
if(score > 0) {
olddepth = std::tan((1000 - score) * std::acos(-1) / 2000);
depth += olddepth;
newscore = 1000 - std::atan(depth) * 2000 / std::acos(-1);
} else {
olddepth = std::tan((1000 + score) * std::acos(-1) / 2000);
depth += olddepth;
newscore = std::atan(depth) * 2000 / std::acos(-1) - 1000;
}
return newscore;
}
Disclaimer: I don't know C++, and I frankly haven't bothered to read the game rules. I have now read the rules, and still stand by what I said...but I still don't know C++. Still, I can present some general knowledge of the algorithm which should set you off in the right direction.
Asymmetry is not in itself a bad thing. If two moves are exactly equivalent, it should choose one of them and not stand helpless like Buridan's ass. You should, in fact, be sure that any agent you write has some method of choosing arbitrarily between policies which it cannot distinguish.
You should think more carefully about the utility scheme implied by refusing to visit previous states. Pursuing an infinite loop is a valid policy, even if your current representation of it will crash the program; maybe the bug is the overflow, not the policy that caused it. If given the choice between losing the game, and refusing to let the game end, which do you want your agent to prefer?
Playing ad infinitum
If you want your agent to avoid losing at all costs -- that is, you want it to prefer indefinite play over loss -- then I would suggest treating any repeated state as a terminal state and assigning it a value somewhere between winning and losing. After all, in a sense it is terminal -- this is the loop the game will enter forever and ever and ever, and the definite result of it is that there is no winner. However, remember that if you are using simple minimax (one utility function, not two), then this implies that your opponent also regards eternal play as a middling result.
It may sound ridiculous, but maybe playing unto infinity is actually a reasonable policy. Remember that minimax assumes the worst case -- a perfectly rational foe whose interests are the exact opposite of yours. But if, for example, you're writing an agent to play against a human, then the human will either err logically, or will eventually decide they would rather end the game by losing -- so your agent will benefit from patiently staying in this Nash equilibrium loop!
Alright, let's end the game already
If you want your agent to prefer that the game end eventually, then I would suggest implementing a living penalty -- a modifier added to your utility which decreases as a function of time (be it asymptotic or without bound). Implemented carefully, this can guarantee that, eventually, any end is preferable to another turn. With this solution as well, you need to be careful about considering what preferences this implies for your opponent.
A third way
Another common solution is to depth-limit your search and implement an evaluation function. This takes the game state as its input and just spits out a utility value which is its best guess at the end result. Is this provably optimal? No, not unless your evaluation function is just completing the minimax, but it means your algorithm will finish within a reasonable time. By burying this rough estimate deep enough in the tree, you wind up with a pretty reasonable model. However, this produces an incomplete policy, which means that it is more useful for a replanning agent than for a standard planning agent. Minimax replanning is the usual approach for complex games (it is, if I'm not mistaken, the basic algorithm followed by Deep Blue), but since this is a very simple game you probably don't need to take this approach.
A side note on abstraction
Note that all of these solutions are conceptualized as either numeric changes to or estimations of the utility function. This is, in general, preferable to arbitrarily throwing away possible policies. After all, that's what your utility function is for -- any time you make a policy decision on the basis of anything except the numeric value of your utility, you are breaking your abstraction and making your code less robust.
Related
Pardon me if this question already exists, I've searched a lot but I haven't gotten the answer to the question I want to ask. So, basically, I'm trying to implement a Tic-Tac-Toe AI that uses the Minimax algorithm to make moves.
However, one thing I don't get is, that when Minimax is used on an empty board, the value returned is always 0 (which makes sense because the game always ends in a draw if both players play optimally).
So Minimax always chooses the first tile as the best move when AI is X (since all moves return 0 as value). Same happens for the second move and it always chooses the second tile instead. How can I fix this problem to make my AI pick the move with the higher probability of winning? Here is the evaluation and Minimax function I use (with Alpha-Beta pruning):
int evaluate(char board[3][3], char AI)
{
for (int row = 0; row<3; row++)
{
if (board[row][0] != '_' && board[row][0] == board[row][1] && board[row][1] == board[row][2])
{
if (board[row][0]==AI)
{
return +10;
}
else
{
return -10;
}
}
}
for (int col = 0; col<3; col++)
{
if (board[0][col] != '_' && board[0][col] == board[1][col] && board[1][col] == board[2][col])
{
if (board[0][col]==AI)
{
return +10;
}
else
{
return -10;
}
}
}
if (board[1][1] != '_' && ((board[0][0]==board[1][1] && board[1][1]==board[2][2]) || (board[0][2]==board[1][1] && board[1][1]==board[2][0])))
{
if (board[1][1]==AI)
{
return +10;
}
else
{
return -10;
}
}
return 0;
}
int Minimax(char board[3][3], bool AITurn, char AI, char Player, int depth, int alpha, int beta)
{
bool breakout = false;
int score = evaluate(board, AI);
if(score == 10)
{
return score - depth;
}
else if(score == -10)
{
return score + depth;
}
else if(NoTilesEmpty(board))
{
return 0;
}
if(AITurn == true)
{
int bestvalue = -1024;
for(int i = 0; i < 3; i++)
{
for(int j = 0; j<3; j++)
{
if(board[i][j] == '_')
{
board[i][j] = AI;
bestvalue = max(bestvalue, Minimax(board, false, AI, Player, depth+1, alpha, beta));
alpha = max(bestvalue, alpha);
board[i][j] = '_';
if(beta <= alpha)
{
breakout = true;
break;
}
}
}
if(breakout == true)
{
break;
}
}
return bestvalue;
}
else if(AITurn == false)
{
int bestvalue = +1024;
for(int i = 0; i < 3; i++)
{
for(int j = 0; j<3; j++)
{
if(board[i][j] == '_')
{
board[i][j] = Player;
bestvalue = min(bestvalue, Minimax(board, true, AI, Player, depth+1, alpha, beta));
beta = min(bestvalue, beta);
board[i][j] = '_';
if(beta <= alpha)
{
breakout = true;
break;
}
}
}
if(breakout == true)
{
break;
}
}
return bestvalue;
}
}
Minimax assumes optimal play, so maximizing "probability of winning" is not a meaningful notion: Since the other player can force a draw but cannot force a win, they will always force a draw. If you want to play optimally against a player who is not perfectly rational (which, of course, is one of the only two ways to win*), you'll need to assume some probability distribution over the opponent's moves and use something like ExpectMinimax, where with some probability the opponent's move is overridden by a random mistake. Alternatively, you can deliberately restrict the ply of the minimax search, using a heuristic for the opponent's play beyond a certain depth (but still searching the game tree for your own moves.)
* The other one is not to play.
Organize your code into smaller routines so that it looks tidier and easier to debug. Apart from the recursive minimax function, an all-possible-valid-move generation function and a robust evaluation sub-routine are essential ( which seems lacking here).
For example, at the beginning of the game, the evaluation algorithm should return a non-zero score, every position should have a relative scoring index ( eg middle position may have slightly higher weightage than the corners).
Your minimax boundary condition - return if there is no empty cell positions ; is flawed as it will evaluate even when a winning/losing move occurred in the preceding ply. Such conditions will aggravate in more complex AI games.
If you are new to minimax, you can find plenty of ready to compile sample codes on CodeReview
I'm trying to implement NegaMax ai for Connect 4. The algorithm works well some of the time, and the ai can win. However, sometimes it completely fails to block opponent 3 in a rows, or doesn't take a winning shot when it has three in a row.
The evaluation function iterates through the grid (horizontally, vertically, diagonally up, diagonally down), and takes every set of four squares. It then checks within each of these sets and evaluates based on this.
I've based the function on the evaluation code provided here: http://blogs.skicelab.com/maurizio/connect-four.html
My function is as follows:
//All sets of four tiles are evaluated before this
//and values for the following variables are set.
if (redFoursInARow != 0)
{
redScore = INT_MAX;
}
else
{
redScore = (redThreesInARow * threeWeight) + (redTwosInARow * twoWeight);
}
int yellowScore = 0;
if (yellowFoursInARow != 0)
{
yellowScore = INT_MAX;
}
else
{
yellowScore = (yellowThreesInARow * threeWeight) + (yellowTwosInARow * twoWeight);
}
int finalScore = yellowScore - redScore;
return turn ? finalScore : -finalScore; //If this is an ai turn, return finalScore. Else return -finalScore.
My negamax function looks like this:
inline int NegaMax(char g[6][7], int depth, int &bestMove, int row, int col, bool aiTurn)
{
{
char c = CheckForWinner(g);
if ('E' != c || 0 == depth)
{
return EvaluatePosition(g, aiTurn);
}
}
int bestScore = INT_MIN;
for (int i = 0; i < 7; ++i)
{
if (CanMakeMove(g, i)) //If column i is not full...
{
{
//...then make a move in that column.
//Grid is a 2d char array.
//'E' = empty tile, 'Y' = yellow, 'R' = red.
char newPos[6][7];
memcpy(newPos, g, sizeof(char) * 6 * 7);
int newRow = GetNextEmptyInCol(g, i);
if (aiTurn)
{
UpdateGrid(newPos, i, 'Y');
}
else
{
UpdateGrid(newPos, i, 'R');
}
int newScore = 0; int newMove = 0;
newScore = NegaMax(newPos, depth - 1, newMove, newRow, i, !aiTurn);
newScore = -newScore;
if (newScore > bestScore)
{
bestMove = i;
bestScore = newScore;
}
}
}
}
return bestScore;
}
I'm aware that connect four has been solved are that there are definitely better ways to go about this, but any help or suggestions with fixing/improving this will be greatly appreciated. Thanks!
void generate_moves(int gameBoard[9], list<int> &moves)
{
for (int i = 0; i < 9; i++)
{
if (gameBoard[i] == 0){
moves.push_back(i);
}
}
}
int evaluate_position(int gameBoard[9], int playerTurn)
{
state currentGameState = checkWin(gameBoard);
if (currentGameState != PLAYING)
{
if ((playerTurn == 1 && currentGameState == XWIN) || (playerTurn == -1 && currentGameState == OWIN))
return +infinity;
else if ((playerTurn == -1 && currentGameState == XWIN) || (playerTurn == 1 && currentGameState == OWIN))
return -infinity;
else if (currentGameState == DRAW)
return 0;
}
return -1;
}
int MinMove(int gameBoard[9], int playerTurn)
{
//if (checkWin(gameBoard) != PLAYING) { return evaluate_board(gameBoard); };
int pos_val = evaluate_position(gameBoard, playerTurn);
if (pos_val != -1) return pos_val;
int bestScore = +infinity;
list<int> movesList;
generate_moves(gameBoard, movesList);
while (!movesList.empty())
{
gameBoard[movesList.front()] = playerTurn;
int score = MaxMove(gameBoard, playerTurn*-1);
if (score < bestScore)
{
bestScore = score;
}
gameBoard[movesList.front()] = 0;
movesList.pop_front();
}
return bestScore;
}
int MaxMove(int gameBoard[9], int playerTurn)
{
//if (checkWin(gameBoard) != PLAYING) { return evaluate_board(gameBoard); };
int pos_val = evaluate_position(gameBoard, playerTurn);
if (pos_val != -1) return pos_val;
int bestScore = -infinity;
list<int> movesList;
generate_moves(gameBoard, movesList);
while (!movesList.empty())
{
gameBoard[movesList.front()] = playerTurn;
int score = MinMove(gameBoard, playerTurn*-1);
if (score > bestScore)
{
bestScore = score;
}
gameBoard[movesList.front()] = 0;
movesList.pop_front();
}
return bestScore;
}
int MiniMax(int gameBoard[9], int playerTurn)
{
int bestScore = -infinity;
int index = 0;
list<int> movesList;
vector<int> bestMoves;
generate_moves(gameBoard, movesList);
while (!movesList.empty())
{
gameBoard[movesList.front()] = playerTurn;
int score = MinMove(gameBoard, playerTurn);
if (score > bestScore)
{
bestScore = score;
bestMoves.clear();
bestMoves.push_back(movesList.front());
}
else if (score == bestScore)
{
bestMoves.push_back(movesList.front());
}
gameBoard[movesList.front()] = 0;
movesList.pop_front();
}
index = bestMoves.size();
if (index > 0) {
time_t secs;
time(&secs);
srand((uint32_t)secs);
index = rand() % index;
}
return bestMoves[index];
}
I created a tic tac toe program in C++ and tried to implement a MiniMax algorithm with exhaustive search tree.
These are the functions I have written using wiki and with the help of some websites. But the AI just doesn't work right and at times doesn't play its turn at all.
Could someone have a look and please point out if there is anything wrong with the logic?
This is how I think it works:
Minimax : This function starts with very large -ve number as best score and goal is to maximize that number. It calls minMove function. If new score > best score, then best score = new score...
MinMove : This function evaluates game board. If game over then it returns -infinity or +infinity depending on who won. If game is going on this function starts with max +infinity value as best score and goal is to minimize it as much possible. It calls MaxMove with opponent player's turn. (since players alternate turns).
If score < best score then best score = score. ...
MaxMove : This function evaluates game board. If game over then it returns -infinity or +infinity depending on who won. If game is going on this function starts with least -infinity value as best score and goal is to maximize it as much possible. It calls MinMove with opponent player's turn. (since players alternate turns).
If score > best score then best score = score. ...
Minmove and MaxMove call each other mutually recursively, MaxMove maximizing the value and MinMove minimizing it. Finally it returns the best possible moves list.
If there are more than 1 best moves, then a random of them is picked as the computer's move.
In MiniMax, MinMove(gameBoard, playerTurn) should be MinMove(gameBoard, -playerTurn) as you do in MaxMove.
As you use MinMove and MaxMove, your evaluation function should be absolute. I mean +infinity for XWIN
and -infinity for OWIN. And so MinMove can only be use when player == -1 and MaxMove when player == 1, thus the parameter become useless. And so MiniMax can only be used by player == 1.
I have done some changes in your code and it works (https://ideone.com/Ihy1SR).
I'm trying to implement a minimax algorithm for tic tac toe with alpha-beta pruning. Right now I have the program running, but it does not seem to be working. Whenever I run it it seems to input garbage in all the squares. I've implemented it so that my minimax function takes in a board state and modifies that state so that when it is finished, the board state contains the next best move. Then, I set 'this' to equal the modified board. Here are my functions for the minimax algorithm:
void board::getBestMove() {
board returnBoard;
miniMax(INT_MIN + 1, INT_MAX -1, returnBoard);
*this = returnBoard;
}
int board::miniMax(int alpha, int beta, board childWithMaximum) {
if (checkDone())
return boardScore();
vector<board> children = getChildren();
for (int i = 0; i < 9; ++i) {
if(children.empty()) break;
board curr = children.back();
if (curr.firstMoveMade) { // not an empty board
board dummyBoard;
int score = curr.miniMax(alpha, beta, dummyBoard);
if (computerTurn && (beta > score)) {
beta = score;
childWithMaximum = *this;
if (alpha >= beta) break;
} else if (alpha < score) {
alpha = score;
childWithMaximum = *this;
if (alpha >= beta) break;
}
}
}
return computerTurn? alpha : beta;
}
vector<board> board::getChildren() {
vector<board> children;
for (int i = 0; i < 3; ++i) {
for (int j = 0; j < 3; ++j) {
if (getPosition(i, j) == '*') { //move not made here
board moveMade(*this);
moveMade.setPosition(i, j);
children.push_back(moveMade);
}
}
}
return children;
}
And here are my full files if someone wants to try running it:
.cpp : http://pastebin.com/ydG7RFRX
.h : http://pastebin.com/94mDdy7x
There may be many issues with your code... you sure posted a lot of it. Because you are asking your question it is incumbent on you to try everything you can on your own first and then reduce your question to the smallest amount of code necessary to clarify what is going on. As it is, I don't feel that you've put much effort into asking this question.
But maybe I can still provide some help:
void board::getBestMove() {
board returnBoard;
miniMax(INT_MIN + 1, INT_MAX -1, returnBoard);
*this = returnBoard;
}
See how you are saying *this = returnBoard.
That must mean that you want to get a board back from miniMax.
But look at how miniMax is defined!
int board::miniMax(int alpha, int beta, board childWithMaximum)
It accepts childWithMaximum via pass by value so it cannot return a board in this way.
what you wanted to say was probably:
int board::miniMax(int alpha, int beta, board & childWithMaximum)
Edit: to clarify, the problem is with the second algorithm.
I have a bit of C++ code that samples cards from a 52 card deck, which works just fine:
void sample_allcards(int table[5], int holes[], int players) {
int temp[5 + 2 * players];
bool try_again;
int c, n, i;
for (i = 0; i < 5 + 2 * players; i++) {
try_again = true;
while (try_again == true) {
try_again = false;
c = fast_rand52();
// reject collisions
for (n = 0; n < i + 1; n++) {
try_again = (temp[n] == c) || try_again;
}
temp[i] = c;
}
}
copy_cards(table, temp, 5);
copy_cards(holes, temp + 5, 2 * players);
}
I am implementing code to sample the hole cards according to a known distribution (stored as a 2d table). My code for this looks like:
void sample_allcards_weighted(double weights[][HOLE_CARDS], int table[5], int holes[], int players) {
// weights are distribution over hole cards
int temp[5 + 2 * players];
int n, i;
// table cards
for (i = 0; i < 5; i++) {
bool try_again = true;
while (try_again == true) {
try_again = false;
int c = fast_rand52();
// reject collisions
for (n = 0; n < i + 1; n++) {
try_again = (temp[n] == c) || try_again;
}
temp[i] = c;
}
}
for (int player = 0; player < players; player++) {
// hole cards according to distribution
i = 5 + 2 * player;
bool try_again = true;
while (try_again == true) {
try_again = false;
// weighted-sample c1 and c2 at once
// h is a number < 1325
int h = weighted_randi(&weights[player][0], HOLE_CARDS);
// i2h uses h and sets temp[i] to the 2 cards implied by h
i2h(&temp[i], h);
// reject collisions
for (n = 0; n < i; n++) {
try_again = (temp[n] == temp[i]) || (temp[n] == temp[i+1]) || try_again;
}
}
}
copy_cards(table, temp, 5);
copy_cards(holes, temp + 5, 2 * players);
}
My problem? The weighted sampling algorithm is a factor of 10 slower. Speed is very important for my application.
Is there a way to improve the speed of my algorithm to something more reasonable? Am I doing something wrong in my implementation?
Thanks.
edit: I was asked about this function, which I should have posted, since it is key
inline int weighted_randi(double *w, int num_choices) {
double r = fast_randd();
double threshold = 0;
int n;
for (n = 0; n < num_choices; n++) {
threshold += *w;
if (r <= threshold) return n;
w++;
}
// shouldn't get this far
cerr << n << "\t" << threshold << "\t" << r << endl;
assert(n < num_choices);
return -1;
}
...and i2h() is basically just an array lookup.
Your reject collisions are turning an O(n) algorithm into (I think) an O(n^2) operation.
There are two ways to select cards from a deck: shuffle and pop, or pick sets until the elements of the set are unique; you are doing the latter which requires a considerable amount of backtracking.
I didn't look at the details of the code, just a quick scan.
you could gain some speed by replacing the all the loops that check if a card is taken with a bit mask, eg for a pool of 52 cards, we prevent collisions like so:
DWORD dwMask[2] = {0}; //64 bits
//...
int nCard;
while(true)
{
nCard = rand_52();
if(!(dwMask[nCard >> 5] & 1 << (nCard & 31)))
{
dwMask[nCard >> 5] |= 1 << (nCard & 31);
break;
}
}
//...
My guess would be the memcpy(1326*sizeof(double)) within the retry-loop. It doesn't seem to change, so should it be copied each time?
Rather than tell you what the problem is, let me suggest how you can find it. Either 1) single-step it in the IDE, or 2) randomly halt it to see what it's doing.
That said, sampling by rejection, as you are doing, can take an unreasonably long time if you are rejecting most samples.
Your inner "try_again" for loop should stop as soon as it sets try_again to true - there's no point in doing more work after you know you need to try again.
for (n = 0; n < i && !try_again; n++) {
try_again = (temp[n] == temp[i]) || (temp[n] == temp[i+1]);
}
Answering the second question about picking from a weighted set also has an algorithmic replacement that should be less time complex. This is based on the principle of that which is pre-computed does not need to be re-computed.
In an ordinary selection, you have an integral number of bins which makes picking a bin an O(1) operation. Your weighted_randi function has bins of real length, thus selection in your current version operates in O(n) time. Since you don't say (but do imply) that the vector of weights w is constant, I'll assume that it is.
You aren't interested in the width of the bins, per se, you are interested in the locations of their edges that you re-compute on every call to weighted_randi using the variable threshold. If the constancy of w is true, pre-computing a list of edges (that is, the value of threshold for all *w) is your O(n) step which need only be done once. If you put the results in a (naturally) ordered list, a binary search on all future calls yields an O(log n) time complexity with an increase in space needed of only sizeof w / sizeof w[0].