/// <summary>
/// An implementation of the alpha-beta pruning algorithm.
/// </summary>
/// <param name="game">The position to solve.</param>
/// <param name="depth">The amount of full moves to search through.</param>
/// <param name="alpha">The value of alpha (maximising player).</param>
/// <param name="beta">The value of beta (minimising player).</param>
/// <returns></returns>
private static (int?, int) AlphaBetaPruning(Game game, int depth, int alpha, int beta)
{
bool maxPlayer = game.Turn % 2 == 0;
List <int> children = new();
for (int col = 0; col < game.Grid.Length; col++)
{
if (!game.IsFilled(col))
{
children.Add(col);
}
}
if (game.Draw)
{
return(null, 0);
}
foreach (int child in children)
{
if (game.IsWinningMove(child))
{
return(child, maxPlayer ? int.MaxValue : int.MinValue);
}
}
if (depth == 0)
{
return(null, game.Evaluation(maxPlayer ? 0 : 1));
}
if (maxPlayer)
{
int value = int.MinValue;
Dictionary <int, int> scores = new();
foreach (int child in children)
{
Game opp = new(game);
opp.Play(child);
int score = AlphaBetaPruning(opp, depth - 1, alpha, beta).Item2;
scores.Add(child, score);
if (score > value)
{
value = score;
}
alpha = Max(alpha, value);
if (alpha >= beta)
{
break;
}
}
return(scores.Aggregate((l, r) => l.Value > r.Value ? l : r).Key, value);
}
else
{
int value = int.MaxValue;
Dictionary <int, int> scores = new();
foreach (int child in children)
{
Game opp = new(game);
opp.Play(child);
int score = AlphaBetaPruning(opp, depth - 1, alpha, beta).Item2;
scores.Add(child, score);
if (score < value)
{
value = score;
}
beta = Max(beta, value);
if (alpha >= beta)
{
break;
}
}
return(scores.Aggregate((l, r) => l.Value > r.Value ? l : r).Key, value);
}
}
