战舰!

早在2003年(当时我17岁),我参加了一场《战舰》AI编码比赛。尽管我输了那场比赛,但我从中获得了很多乐趣,也学到了很多东西。

现在,我想恢复这个比赛,在搜索最好的战舰AI。

下面是这个框架,现在托管在Bitbucket上。

获胜者将获得+450声望奖励!比赛将于2009年11月17日开始。17号零时之前的投稿和编辑将不被接受。(中央标准时间) 尽早提交你的作品,这样你就不会错过机会!

为了保持这个目标,请遵循比赛的精神。

游戏规则:

游戏在10x10的网格上进行。 每个参赛者将5艘船(长度为2、3、3、4、5)中的每一艘放在他们的网格上。 没有船只可以重叠,但它们可以相邻。 然后选手们轮流向对手射击。 游戏的一个变体允许每次齐射多次,每艘幸存的船一次。 如果击球沉、命中或未命中,对手将通知选手。 当任何一名玩家的所有船只都沉没时,游戏就结束了。

比赛规则:

The spirit of the competition is to find the best Battleship algorithm. Anything that is deemed against the spirit of the competition will be grounds for disqualification. Interfering with an opponent is against the spirit of the competition. Multithreading may be used under the following restrictions: No more than one thread may be running while it is not your turn. (Though, any number of threads may be in a "Suspended" state). No thread may run at a priority other than "Normal". Given the above two restrictions, you will be guaranteed at least 3 dedicated CPU cores during your turn. A limit of 1 second of CPU time per game is allotted to each competitor on the primary thread. Running out of time results in losing the current game. Any unhandled exception will result in losing the current game. Network access and disk access is allowed, but you may find the time restrictions fairly prohibitive. However, a few set-up and tear-down methods have been added to alleviate the time strain. Code should be posted on stack overflow as an answer, or, if too large, linked. Max total size (un-compressed) of an entry is 1 MB. Officially, .Net 2.0 / 3.5 is the only framework requirement. Your entry must implement the IBattleshipOpponent interface.

得分:

Best 51 games out of 101 games is the winner of a match. All competitors will play matched against each other, round-robin style. The best half of the competitors will then play a double-elimination tournament to determine the winner. (Smallest power of two that is greater than or equal to half, actually.) I will be using the TournamentApi framework for the tournament. The results will be posted here. If you submit more than one entry, only your best-scoring entry is eligible for the double-elim.

好运!玩得开心!


编辑1: 多亏弗里德,他在飞船上发现了一个错误。是否是可用的函数。问题已经解决了。请下载框架的更新版本。

编辑2: 由于人们对将统计数据持久化到磁盘等非常感兴趣,所以我添加了一些非计时设置和删除事件,它们应该能够提供所需的功能。这是一个半突破性的变化。也就是说:修改了接口,添加了功能,但不需要body。请下载框架的更新版本。

编辑3: 错误修复1:GameWon和GameLost只在超时的情况下被调用。 错误修复2:如果引擎在每一款游戏中都暂停计时,那么竞争将永远不会结束。 请下载框架的更新版本。

编辑4: 比赛结果:


当前回答

交火中更新。 我知道它不能与Farnsworth或Dreadnought竞争,但它比后者快得多,而且如果有人想尝试的话,它很简单。 这依赖于我的库的当前状态,包括在这里使它易于使用。

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Drawing;
using System.IO;
using System.Collections.ObjectModel;

namespace Battleship.ShuggyCoUk
{
    public class Simple : IBattleshipOpponent
    {
        BoardView<OpponentsBoardState> opponentsBoard = new BoardView<OpponentsBoardState>(new Size(10,10));
        Rand rand = new Rand();
        int gridOddEven;
        Size size;

        public string Name { get { return "Simple"; } }

        public Version Version { get { return new Version(2, 1); }}

        public void NewMatch(string opponent) {}

        public void NewGame(System.Drawing.Size size, TimeSpan timeSpan)
        {
            this.size = size;
            this.opponentsBoard = new BoardView<OpponentsBoardState>(size);
            this.gridOddEven = rand.Pick(new[] { 0, 1 });
        }

        public void PlaceShips(System.Collections.ObjectModel.ReadOnlyCollection<Ship> ships)
        {
            BoardView<bool> board = new BoardView<bool>(size);
            var AllOrientations = new[] {
                ShipOrientation.Horizontal,
                ShipOrientation.Vertical };

            foreach (var ship in ships)
            {
                int avoidTouching = 3;
                while (!ship.IsPlaced)
                {
                    var l = rand.Pick(board.Select(c => c.Location));
                    var o = rand.Pick(AllOrientations);
                    if (ship.IsLegal(ships, size, l, o))
                    {
                        if (ship.IsTouching(ships, l, o)&& --avoidTouching > 0)
                            continue;
                        ship.Place(l, o);
                    }
                }
            }
        }
        protected virtual Point PickWhenNoTargets()
        {
            return rand.PickBias(x => x.Bias,
                opponentsBoard
                // nothing 1 in size
                .Where(c => (c.Location.X + c.Location.Y) % 2 == gridOddEven)
                .Where(c => c.Data == OpponentsBoardState.Unknown))
                .Location;
        }

        private int SumLine(Cell<OpponentsBoardState> c, int acc)
        {
            if (acc >= 0)
                return acc;
            if (c.Data == OpponentsBoardState.Hit)
                return acc - 1;
            return -acc;
        }

        public System.Drawing.Point GetShot()
        {
            var targets = opponentsBoard
                .Where(c => c.Data == OpponentsBoardState.Hit)
                .SelectMany(c => c.Neighbours())
                .Where(c => c.Data == OpponentsBoardState.Unknown)
                .ToList();
            if (targets.Count > 1)
            {
                var lines = targets.Where(
                    x => x.FoldAll(-1, SumLine).Select(r => Math.Abs(r) - 1).Max() > 1).ToList();
                if (lines.Count > 0)
                    targets = lines;
            }
            var target = targets.RandomOrDefault(rand);
            if (target == null)
                return PickWhenNoTargets();
            return target.Location;
        }

        public void OpponentShot(System.Drawing.Point shot)
        {
        }

        public void ShotHit(Point shot, bool sunk)
        {
            opponentsBoard[shot] = OpponentsBoardState.Hit;
            Debug(shot, sunk);
        }

        public void ShotMiss(Point shot)
        {
            opponentsBoard[shot] = OpponentsBoardState.Miss;
            Debug(shot, false);
        }

        public const bool DebugEnabled = false;

        public void Debug(Point shot, bool sunk)
        {
            if (!DebugEnabled)
                return;
            opponentsBoard.WriteAsGrid(
                Console.Out,
                x =>
                {
                    string t;
                    switch (x.Data)
                    {
                        case OpponentsBoardState.Unknown:
                            return " ";
                        case OpponentsBoardState.Miss:
                            t = "m";
                            break;
                        case OpponentsBoardState.MustBeEmpty:
                            t = "/";
                            break;
                        case OpponentsBoardState.Hit:
                            t = "x";
                            break;
                        default:
                            t = "?";
                            break;
                    }
                    if (x.Location == shot)
                        t = t.ToUpper();
                    return t;
                });
            if (sunk)
                Console.WriteLine("sunk!");
            Console.ReadLine();
        }

        public void GameWon()
        {
        }

        public void GameLost()
        {
        }

        public void MatchOver()
        {
        }

        #region Library code
        enum OpponentsBoardState
        {
            Unknown = 0,
            Miss,
            MustBeEmpty,
            Hit,
        }

        public enum Compass
        {
            North, East, South, West
        }

        class Cell<T>
        {
            private readonly BoardView<T> view;
            public readonly int X;
            public readonly int Y;
            public T Data;
            public double Bias { get; set; }

            public Cell(BoardView<T> view, int x, int y)
            {
                this.view = view; this.X = x; this.Y = y; this.Bias = 1.0;
            }

            public Point Location
            {
                get { return new Point(X, Y); }
            }

            public IEnumerable<U> FoldAll<U>(U acc, Func<Cell<T>, U, U> trip)
            {
                return new[] { Compass.North, Compass.East, Compass.South, Compass.West }
                    .Select(x => FoldLine(x, acc, trip));
            }

            public U FoldLine<U>(Compass direction, U acc, Func<Cell<T>, U, U> trip)
            {
                var cell = this;
                while (true)
                {
                    switch (direction)
                    {
                        case Compass.North:
                            cell = cell.North; break;
                        case Compass.East:
                            cell = cell.East; break;
                        case Compass.South:
                            cell = cell.South; break;
                        case Compass.West:
                            cell = cell.West; break;
                    }
                    if (cell == null)
                        return acc;
                    acc = trip(cell, acc);
                }
            }

            public Cell<T> North
            {
                get { return view.SafeLookup(X, Y - 1); }
            }

            public Cell<T> South
            {
                get { return view.SafeLookup(X, Y + 1); }
            }

            public Cell<T> East
            {
                get { return view.SafeLookup(X + 1, Y); }
            }

            public Cell<T> West
            {
                get { return view.SafeLookup(X - 1, Y); }
            }

            public IEnumerable<Cell<T>> Neighbours()
            {
                if (North != null)
                    yield return North;
                if (South != null)
                    yield return South;
                if (East != null)
                    yield return East;
                if (West != null)
                    yield return West;
            }
        }

        class BoardView<T> : IEnumerable<Cell<T>>
        {
            public readonly Size Size;
            private readonly int Columns;
            private readonly int Rows;

            private Cell<T>[] history;

            public BoardView(Size size)
            {
                this.Size = size;
                Columns = size.Width;
                Rows = size.Height;
                this.history = new Cell<T>[Columns * Rows];
                for (int y = 0; y < Rows; y++)
                {
                    for (int x = 0; x < Rows; x++)
                        history[x + y * Columns] = new Cell<T>(this, x, y);
                }
            }

            public T this[int x, int y]
            {
                get { return history[x + y * Columns].Data; }
                set { history[x + y * Columns].Data = value; }
            }

            public T this[Point p]
            {
                get { return history[SafeCalc(p.X, p.Y, true)].Data; }
                set { this.history[SafeCalc(p.X, p.Y, true)].Data = value; }
            }

            private int SafeCalc(int x, int y, bool throwIfIllegal)
            {
                if (x < 0 || y < 0 || x >= Columns || y >= Rows)
                {
                    if (throwIfIllegal)
                        throw new ArgumentOutOfRangeException("[" + x + "," + y + "]");
                    else
                        return -1;
                }
                return x + y * Columns;
            }

            public void Set(T data)
            {
                foreach (var cell in this.history)
                    cell.Data = data;
            }

            public Cell<T> SafeLookup(int x, int y)
            {
                int index = SafeCalc(x, y, false);
                if (index < 0)
                    return null;
                return history[index];
            }

            #region IEnumerable<Cell<T>> Members

            public IEnumerator<Cell<T>> GetEnumerator()
            {
                foreach (var cell in this.history)
                    yield return cell;
            }

            System.Collections.IEnumerator System.Collections.IEnumerable.GetEnumerator()
            {
                return this.GetEnumerator();
            }

            public BoardView<U> Transform<U>(Func<T, U> transform)
            {
                var result = new BoardView<U>(new Size(Columns, Rows));
                for (int y = 0; y < Rows; y++)
                {
                    for (int x = 0; x < Columns; x++)
                    {
                        result[x, y] = transform(this[x, y]);
                    }
                }
                return result;
            }

            public void WriteAsGrid(TextWriter w)
            {
                WriteAsGrid(w, "{0}");
            }

            public void WriteAsGrid(TextWriter w, string format)
            {
                WriteAsGrid(w, x => string.Format(format, x.Data));
            }

            public void WriteAsGrid(TextWriter w, Func<Cell<T>, string> perCell)
            {
                for (int y = 0; y < Rows; y++)
                {
                    for (int x = 0; x < Columns; x++)
                    {
                        if (x != 0)
                            w.Write(",");
                        w.Write(perCell(this.SafeLookup(x, y)));
                    }
                    w.WriteLine();
                }
            }

            #endregion
        }

        public class Rand
        {
            Random r;

            public Rand()
            {
                var rand = System.Security.Cryptography.RandomNumberGenerator.Create();
                byte[] b = new byte[4];
                rand.GetBytes(b);
                r = new Random(BitConverter.ToInt32(b, 0));
            }

            public int Next(int maxValue)
            {
                return r.Next(maxValue);
            }

            public double NextDouble(double maxValue)
            {
                return r.NextDouble() * maxValue;
            }

            public T Pick<T>(IEnumerable<T> things)
            {
                return things.ElementAt(Next(things.Count()));
            }

            public T PickBias<T>(Func<T, double> bias, IEnumerable<T> things)
            {
                double d = NextDouble(things.Sum(x => bias(x)));
                foreach (var x in things)
                {
                    if (d < bias(x))
                        return x;
                    d -= bias(x);
                }
                throw new InvalidOperationException("fell off the end!");
            }
        }
        #endregion
    }

    public static class Extensions
    {
        public static bool IsIn(this Point p, Size size)
        {
            return p.X >= 0 && p.Y >= 0 && p.X < size.Width && p.Y < size.Height;
        }

        public static bool IsLegal(this Ship ship,
            IEnumerable<Ship> ships,
            Size board,
            Point location,
            ShipOrientation direction)
        {
            var temp = new Ship(ship.Length);
            temp.Place(location, direction);
            if (!temp.GetAllLocations().All(p => p.IsIn(board)))
                return false;
            return ships.Where(s => s.IsPlaced).All(s => !s.ConflictsWith(temp));
        }

        public static bool IsTouching(this Point a, Point b)
        {
            return (a.X == b.X - 1 || a.X == b.X + 1) &&
                (a.Y == b.Y - 1 || a.Y == b.Y + 1);
        }

        public static bool IsTouching(this Ship ship,
            IEnumerable<Ship> ships,
            Point location,
            ShipOrientation direction)
        {
            var temp = new Ship(ship.Length);
            temp.Place(location, direction);
            var occupied = new HashSet<Point>(ships
                .Where(s => s.IsPlaced)
                .SelectMany(s => s.GetAllLocations()));
            if (temp.GetAllLocations().Any(p => occupied.Any(b => b.IsTouching(p))))
                return true;
            return false;
        }

        public static ReadOnlyCollection<Ship> MakeShips(params int[] lengths)
        {
            return new System.Collections.ObjectModel.ReadOnlyCollection<Ship>(
                lengths.Select(l => new Ship(l)).ToList());
        }

        public static IEnumerable<T> Shuffle<T>(this IEnumerable<T> source, Battleship.ShuggyCoUk.Simple.Rand rand)
        {
            T[] elements = source.ToArray();
            // Note i > 0 to avoid final pointless iteration
            for (int i = elements.Length - 1; i > 0; i--)
            {
                // Swap element "i" with a random earlier element it (or itself)
                int swapIndex = rand.Next(i + 1);
                T tmp = elements[i];
                elements[i] = elements[swapIndex];
                elements[swapIndex] = tmp;
            }
            // Lazily yield (avoiding aliasing issues etc)
            foreach (T element in elements)
            {
                yield return element;
            }
        }

        public static T RandomOrDefault<T>(this IEnumerable<T> things, Battleship.ShuggyCoUk.Simple.Rand rand)
        {
            int count = things.Count();
            if (count == 0)
                return default(T);
            return things.ElementAt(rand.Next(count));
        }
    }

}

其他回答

这是一个供人们玩的对手:

http://natekohl.net/files/FarnsworthOpponent.cs

与其使用固定的几何启发策略,我认为尝试估计任何特定的未探索空间拥有一艘船的潜在概率会很有趣。

为了做到这一点,你需要探索所有符合你当前世界观的船的可能配置,然后基于这些配置计算概率。你可以把它想象成探索一棵树:

扩大可能的战列舰国家http://natekohl.net/media/battleship-tree.png

在考虑了所有与你所了解的世界相冲突的树叶后(游戏邦注:例如,船只不能重叠,所有被击中的方块都必须是船只等),你可以计算船只在每个未探索位置出现的频率,从而估算船只位于那里的可能性。

这可以可视化为热图,其中热点更有可能包含船只:

每个未开发位置的概率热图http://natekohl.net/media/battleship-probs.png

我喜欢这个战列舰比赛的一个原因是上面的树几乎小到可以强制使用这种算法。如果这5艘船每艘都有150个可能的位置,那就是1505 = 750亿种可能性。这个数字只会越来越小,特别是如果你可以排除整艘船。

我上面链接的对手并没有探索整棵树;750亿美元仍然太大,无法在一秒钟内进入。不过,在一些启发式的帮助下,它确实试图估计这些概率。

当然,我们也有可能在游戏的基础上运行一系列类似的游戏。

添加3d飞机或能够移动一艘船而不是射击回合等内容可能会改变游戏。

这是我的入口!(最天真的解决方案)

“随机1.1”

namespace Battleship
{
    using System;
    using System.Collections.ObjectModel;
    using System.Drawing;

    public class RandomOpponent : IBattleshipOpponent
    {
        public string Name { get { return "Random"; } }
        public Version Version { get { return this.version; } }

        Random rand = new Random();
        Version version = new Version(1, 1);
        Size gameSize;

        public void NewGame(Size size, TimeSpan timeSpan)
        {
            this.gameSize = size;
        }

        public void PlaceShips(ReadOnlyCollection<Ship> ships)
        {
            foreach (Ship s in ships)
            {
                s.Place(
                    new Point(
                        rand.Next(this.gameSize.Width),
                        rand.Next(this.gameSize.Height)),
                    (ShipOrientation)rand.Next(2));
            }
        }

        public Point GetShot()
        {
            return new Point(
                rand.Next(this.gameSize.Width),
                rand.Next(this.gameSize.Height));
        }

        public void NewMatch(string opponent) { }
        public void OpponentShot(Point shot) { }
        public void ShotHit(Point shot, bool sunk) { }
        public void ShotMiss(Point shot) { }
        public void GameWon() { }
        public void GameLost() { }
        public void MatchOver() { }
    }
}

我现在没有时间写一个完整的算法,但我有一个想法:如果你的对手随机放置船只,放置概率不会是一个以(5.5,5.5)为中心的简单分布吗?例如,战列舰(5个单位长)在x维度的放置可能性如下:

x    1 2 3 4 5  6  7 8 9 10
P(x) 2 4 6 8 10 10 8 6 4 2

同样的计算也适用于y。其他船只的分布不会那么陡峭,但你最好的猜测仍然是中心。在此之后,数学方法将慢慢地向中心外辐射对角线(可能是平均船的长度,17/5)。例:

...........
....x.x....
.....x.....
....x.x....
...........

显然,这个想法需要添加一些随机性,但我认为从纯粹的数学角度来看,这是可行的。

一秒钟的游戏时间是机器特定的。与比赛机器相比,我的机器上每秒的CPU操作将是不同的。如果我优化Battle Ship算法,让它在1秒内利用最多CPU时间,那么它就会运行在可能较慢的比赛机器上,它总是会输。

我不确定如何绕过框架的这一限制,但它应该得到解决。

...

一个想法是做在这个比赛http://www.bcsstudentcontest.com/

每个回合有最大时间,而不是最大总游戏时间。这样我可以限制算法以适应已知的转弯时间。一款游戏可能会持续50到600多个回合,如果我的算法管理它的总游戏时间,它可能没有足够的时间来发挥最佳效果,或者它可能会花费太多时间而失败。在战舰算法中管理游戏总时间是非常困难的。

我建议修改规则,限制回合时间,而不是总游戏时间。

Edit

如果我写了一个算法,枚举所有可能的镜头,然后对它们进行排名,然后选择排名最高的镜头。生成所有可能的镜头需要很长时间,所以我会让算法运行一段时间,然后停止它。

如果有基于回合的限制,我可以让算法运行0.9秒并返回排名最高的镜头,并且在回合时间限制内很好。

如果我的游戏总时间被限制在1秒,那么就很难确定算法在每个回合中应该运行多长时间。我想要最大化我的CPU时间。如果游戏持续500轮,我可以将每个回合限制在0.002秒,但如果游戏持续100轮,我可以给每个回合0.01秒的CPU时间。

对于一个算法来说,使用半穷举搜索镜头空间来找到当前限制的最佳镜头是不切实际的。

1秒的游戏总时间限制了可以有效用于游戏竞争的算法类型。