创建一个洪水涂料AI

Question

在水彩画游戏中，游戏的目标是使整个棋盘在尽可能少的转弯中保持相同的颜色。

游戏的开头是一块类似于这样的棋盘：

3 3 5 4 1 3 4 1 5
5 1 3 4 1 1 5 2 1
6 5 2 3 4 3 3 4 3
4 4 4 5 5 5 4 1 4
6 2 5 3[3]1 1 6 6
5 5 1 2 5 2 6 6 3
6 1 1 5 3 6 2 3 6
1 2 2 4 5 3 5 1 2
3 6 6 1 5 1 3 2 4

目前，板中心的数字(代表一种颜色)为3。每转一圈，中心的正方形就会改变颜色，所有从中心移动到同一颜色的方格(即在中心广场的洪流区)都会随着颜色的变化而改变颜色。因此，如果正方形改变颜色为5：

3 3 5 4 1 3 4 1 5
5 1 3 4 1 1 5 2 1
6 5 2 3 4 3 3 4 3
4 4 4 5 5 5 4 1 4
6 2 5 5[5]1 1 6 6
5 5 1 2 5 2 6 6 3
6 1 1 5 3 6 2 3 6
1 2 2 4 5 3 5 1 2
3 6 6 1 5 1 3 2 4

然后，中间3左边的3也会改变颜色。现在总共有7名5英尺的S可以从中间到达，所以如果我们把颜色改为4：

3 3 5 4 1 3 4 1 5
5 1 3 4 1 1 5 2 1
6 5 2 3 4 3 3 4 3
4 4 4 4 4 4 4 1 4
6 2 4 4[4]1 1 6 6
5 5 1 2 4 2 6 6 3
6 1 1 5 3 6 2 3 6
1 2 2 4 5 3 5 1 2
3 6 6 1 5 1 3 2 4

油漆区域的面积再次急剧增加。

您的任务是创建一个程序，该程序将以从1到6的19×19颜色网格作为输入，无论您选择何种形式：

4 5 1 1 2 2 1 6 2 6 3 4 2 3 2 3 1 6 3
4 2 6 3 4 4 5 6 4 4 5 3 3 3 3 5 4 3 4
2 3 5 2 2 5 5 1 2 6 2 6 6 2 1 6 6 1 2
4 6 5 5 5 5 4 1 6 6 3 2 6 4 2 6 3 6 6
1 6 4 4 4 4 6 4 2 5 5 3 2 2 4 1 5 2 5
1 6 2 1 5 1 6 4 4 1 5 1 3 4 5 2 3 4 1
3 3 5 3 2 2 2 4 2 1 6 6 6 6 1 4 5 2 5
1 6 1 3 2 4 1 3 3 4 6 5 1 5 5 3 4 3 3
4 4 1 5 5 1 4 6 3 3 4 5 5 6 1 6 2 6 4
1 4 2 5 6 5 5 3 2 5 5 5 3 6 1 4 4 6 6
4 6 6 2 6 6 2 4 2 6 1 5 6 2 3 3 4 3 6
6 1 3 6 3 5 5 3 6 1 3 4 4 5 1 2 6 4 3
2 6 1 3 2 4 2 6 1 1 5 2 6 6 6 6 3 3 3
3 4 5 4 6 6 3 3 4 1 1 6 4 5 1 3 4 1 2
4 2 6 4 1 5 3 6 4 3 4 5 4 2 1 1 4 1 1
4 2 4 1 5 2 2 3 6 6 6 5 2 5 4 5 4 5 1
5 6 2 3 4 6 5 4 1 3 2 3 2 1 3 6 2 2 4
6 5 4 1 3 2 2 1 1 1 6 1 2 6 2 5 6 4 5
5 1 1 4 2 6 2 5 6 1 3 3 4 1 6 1 2 1 2

并返回一个颜色序列，中心广场将改变到每一个回合，再次以您选择的格式：

263142421236425431645152623645465646213545631465

在每个移动序列的末尾，19乘19网格中的方格必须都是相同的颜色。

您的程序必须是完全确定性的；允许伪随机解决方案，但程序每次必须为相同的测试用例生成相同的输出。

获胜的程序将采取最少的步骤来解决在这个文件中找到的所有100,000个测试用例(压缩文本文件，14.23MB)。如果两个解决方案采取相同的步骤(例如，如果它们都找到了最优策略)，那么较短的程序将获胜。

BurntPizza用Java编写了一个程序来验证测试结果。若要使用此程序，请运行提交文件并将输出输送到一个名为steps.txt的文件中。然后，使用steps.txt和floodtest文件在同一个目录下运行这个程序。如果条目有效，并为所有文件生成正确的解决方案，则应通过所有测试并返回All boards solved successfully.。

import java.io.*;
import java.util.*;

public class PainterVerifier {

    public static void main(String[] args) throws FileNotFoundException {

        char[] board = new char[361];

        Scanner s = new Scanner(new File("steps.txt"));
        Scanner b = new Scanner(new File("floodtest"));

        int lineNum = 0;

        caseloop: while (b.hasNextLine()) {

            for (int l = 0; l < 19; l++) {
                String lineb = b.nextLine();
                if (lineb.isEmpty())
                    continue caseloop;
                System.arraycopy(lineb.toCharArray(), 0, board, l * 19, 19);
            }

            String line = s.nextLine();
            if (line.isEmpty())
                continue;
            char[] steps = line.toCharArray();

            Stack<Integer> nodes = new Stack<Integer>();

            for (char c : steps) {
                char targetColor = board[180];
                char replacementColor = c;

                nodes.push(180);

                while (!nodes.empty()) {
                    int n = nodes.pop();
                    if (n < 0 || n > 360)
                        continue;
                    if (board[n] == targetColor) {
                        board[n] = replacementColor;
                        if (n % 19 > 0)
                            nodes.push(n - 1);
                        if (n % 19 < 18)
                            nodes.push(n + 1);
                        if (n / 19 > 0)
                            nodes.push(n - 19);
                        if (n / 19 < 18)
                            nodes.push(n + 19);
                    }
                }
            }
            char center = board[180];
            for (char c : board)
                if (c != center) {
                    s.close();
                    b.close();

                    System.out.println("\nIncomplete board found!\n\tOn line " + lineNum + " of steps.txt");
                    System.exit(0);
                }

            if (lineNum % 5000 == 0)
                System.out.printf("Verification %d%c complete...\n", lineNum * 100 / 100000, '%');

            lineNum++;
        }
        s.close();
        b.close();
        System.out.println("All boards solved successfully.");
    }
}

另外，一个记分板，因为结果实际上不是按分数排序的，在这里，它实际上很重要：

1. 一千九百八千零七十八 - smack42，Java
2. 2,075,452 - user1502040，C
3. 2,098,382 -蒂格鲁，C#
4. 2,155,834 - CoderTao，C#
5. 2,201,995 - MrBackend，Java
6. 二千三百三十三千五百六十九 - CoderTao，C#
7. 2,384,020 - Herjan，C
8. 二四三一八九 -原始人，爪哇
9. 2,445,761 - Herjan，C
10. 二千四百七十五千零五十六 - Jeremy List，Haskell
11. 2,480,714 - SteelTermite，C (2,395字节)
12. 2,480,714 - Herjan，Java (4,702字节)
13. 2,588,847 - BurntPizza，Java (2,748字节)
14. 2,588,847 - Gero3，node.js (4,641字节)
15. 2,979,145 - Teun Pronk，Delphi XE3
16. 4,780,841 - BurntPizza，Java
17. 10 800 000 - Joe .，Python

Answer 1

Java - 1,985,078步

https://github.com/smack42/ColorFill

又来晚了一次。包含1,985,078个步骤的结果文件可以找到这里。

一些背景信息：

几年前，我发现了这个挑战，当时我开始编写自己的克隆版游戏Flood-it。

“最佳不完全”DFS与A*算法

从一开始，我就想为这个游戏创建一个很好的算法。随着时间的推移，我可以通过包括几种不同的不完全搜索策略(类似于这里的其他程序中使用的策略)和对每个解决方案使用这些策略的最佳结果来改进我的解决方案。我甚至用Java重新实现了蒂格鲁氏A*算法，并将它添加到我的解决程序中，以获得比tigrou的结果更好的整体解决方案。

穷举DFS算法

然后，我重点讨论了一种总是能找到最优解的算法。我花了很多精力来优化我的深度优先搜索策略。为了加快搜索速度，我包括了一个hashmap，它存储所有已探测到的状态，这样搜索就可以避免再次搜索它们。虽然这个算法工作得很好，并且足够快地解决了所有14x14个谜题，但是它占用了太多的内存，并且在这个代码挑战中的19x19谜题中运行非常慢。

Puchert A*算法

几个月前，有人联系我看洪水解决者亚伦和西蒙普切特。该程序使用A*型算法和允许的启发式(相对于tigrou)和移动剪枝类似于跳点搜索。我很快注意到这个程序非常快，并找到了最优的解决方案！

当然，我不得不重新实现这个伟大的算法，并将其添加到我的程序中。我努力优化我的Java程序，使其运行速度与Puchert兄弟最初的C++程序一样快。然后我决定尝试这个挑战的100,000个测试用例。在我的机器上，程序运行了超过120个小时，使用我实现的Puchert A*算法找到了1,985,078个步骤。

这将是这个挑战的最佳解决方案，除非程序中有一些错误会导致次优解决方案。欢迎任何反馈！

编辑:在这里添加了代码的相关部分：

类AStarPuchertStrategy

/**
 * a specific strategy for the AStar (A*) solver.
 * <p>
 * the idea is taken from the program "floodit" by Aaron and Simon Puchert,
 * which can be found at <a>https://github.com/aaronpuchert/floodit</a>
 */
public class AStarPuchertStrategy implements AStarStrategy {

    private final Board board;
    private final ColorAreaSet visited;
    private ColorAreaSet current, next;
    private final short[] numCaNotFilledInitial;
    private final short[] numCaNotFilled;

    public AStarPuchertStrategy(final Board board) {
        this.board = board;
        this.visited = new ColorAreaSet(board);
        this.current = new ColorAreaSet(board);
        this.next = new ColorAreaSet(board);
        this.numCaNotFilledInitial = new short[board.getNumColors()];
        for (final ColorArea ca : board.getColorAreasArray()) {
            ++this.numCaNotFilledInitial[ca.getColor()];
        }
        this.numCaNotFilled = new short[board.getNumColors()];
    }

    /* (non-Javadoc)
     * @see colorfill.solver.AStarStrategy#setEstimatedCost(colorfill.solver.AStarNode)
     */
    @Override
    public void setEstimatedCost(final AStarNode node) {

        // quote from floodit.cpp: int State::computeValuation()
        // (in branch "performance")
        //
        // We compute an admissible heuristic recursively: If there are no nodes
        // left, return 0. Furthermore, if a color can be eliminated in one move
        // from the current position, that move is an optimal move and we can
        // simply use it. Otherwise, all moves fill a subset of the neighbors of
        // the filled nodes. Thus, filling that layer gets us at least one step
        // closer to the end.

        node.copyFloodedTo(this.visited);
        System.arraycopy(this.numCaNotFilledInitial, 0, this.numCaNotFilled, 0, this.numCaNotFilledInitial.length);
        {
            final ColorAreaSet.FastIteratorColorAreaId iter = this.visited.fastIteratorColorAreaId();
            int nextId;
            while ((nextId = iter.nextOrNegative()) >= 0) {
                --this.numCaNotFilled[this.board.getColor4Id(nextId)];
            }
        }

        // visit the first layer of neighbors, which is never empty, i.e. the puzzle is not solved yet
        node.copyNeighborsTo(this.current);
        this.visited.addAll(this.current);
        int completedColors = 0;
        {
            final ColorAreaSet.FastIteratorColorAreaId iter = this.current.fastIteratorColorAreaId();
            int nextId;
            while ((nextId = iter.nextOrNegative()) >= 0) {
                final byte nextColor = this.board.getColor4Id(nextId);
                if (--this.numCaNotFilled[nextColor] == 0) {
                    completedColors |= 1 << nextColor;
                }
            }
        }
        int distance = 1;

        while(!this.current.isEmpty()) {
            this.next.clear();
            final ColorAreaSet.FastIteratorColorAreaId iter = this.current.fastIteratorColorAreaId();
            int thisCaId;
            if (0 != completedColors) {
                // We can eliminate colors. Do just that.
                // We also combine all these elimination moves.
                distance += Integer.bitCount(completedColors);
                final int prevCompletedColors = completedColors;
                completedColors = 0;
                while ((thisCaId = iter.nextOrNegative()) >= 0) {
                    final ColorArea thisCa = this.board.getColorArea4Id(thisCaId);
                    if ((prevCompletedColors & (1 << thisCa.getColor())) != 0) {
                        // completed color
                        // expandNode()
                        for (final int nextCaId : thisCa.getNeighborsIdArray()) {
                            if (!this.visited.contains(nextCaId)) {
                                this.visited.add(nextCaId);
                                this.next.add(nextCaId);
                                final byte nextColor = this.board.getColor4Id(nextCaId);
                                if (--this.numCaNotFilled[nextColor] == 0) {
                                    completedColors |= 1 << nextColor;
                                }
                            }
                        }
                    } else {
                        // non-completed color
                        // move node to next layer
                        this.next.add(thisCaId);
                    }
                }
            } else {
                // Nothing found, do the color-blind pseudo-move
                // Expand current layer of nodes.
                ++distance;
                while ((thisCaId = iter.nextOrNegative()) >= 0) {
                    final ColorArea thisCa = this.board.getColorArea4Id(thisCaId);
                    // expandNode()
                    for (final int nextCaId : thisCa.getNeighborsIdArray()) {
                        if (!this.visited.contains(nextCaId)) {
                            this.visited.add(nextCaId);
                            this.next.add(nextCaId);
                            final byte nextColor = this.board.getColor4Id(nextCaId);
                            if (--this.numCaNotFilled[nextColor] == 0) {
                                completedColors |= 1 << nextColor;
                            }
                        }
                    }
                }
            }

            // Move the next layer into the current.
            final ColorAreaSet tmp = this.current;
            this.current = this.next;
            this.next = tmp;
        }
        node.setEstimatedCost(node.getSolutionSize() + distance);
    }

}

AStarSolver类的一部分

private void executeInternalPuchert(final ColorArea startCa) throws InterruptedException {
    final Queue<AStarNode> open = new PriorityQueue<AStarNode>(AStarNode.strongerComparator());
    open.offer(new AStarNode(this.board, startCa));
    AStarNode recycleNode = null;
    while (open.size() > 0) {
        if (Thread.interrupted()) { throw new InterruptedException(); }
        final AStarNode currentNode = open.poll();
        if (currentNode.isSolved()) {
            this.addSolution(currentNode.getSolution());
            return;
        } else {
            // play all possible colors
            int nextColors = currentNode.getNeighborColors(this.board);
            while (0 != nextColors) {
                final int l1b = nextColors & -nextColors; // Integer.lowestOneBit()
                final int clz = Integer.numberOfLeadingZeros(l1b); // hopefully an intrinsic function using instruction BSR / LZCNT / CLZ
                nextColors ^= l1b; // clear lowest one bit
                final byte color = (byte)(31 - clz);
                final AStarNode nextNode = currentNode.copyAndPlay(color, recycleNode, this.board);
                if (null != nextNode) {
                    recycleNode = null;
                    this.strategy.setEstimatedCost(nextNode);
                    open.offer(nextNode);
                }
            }
        }
        recycleNode = currentNode;
    }
}

AStarNode类的一部分

/**
 * check if this color can be played. (avoid duplicate moves)
 * the idea is taken from the program "floodit" by Aaron and Simon Puchert,
 * which can be found at <a>https://github.com/aaronpuchert/floodit</a>
 * @param nextColor
 * @return
 */
private boolean canPlay(final byte nextColor, final List<ColorArea> nextColorNeighbors) {
    final byte currColor = this.solution[this.solutionSize];
    // did the previous move add any new "nextColor" neighbors?
    boolean newNext = false;
next:   for (final ColorArea nextColorNeighbor : nextColorNeighbors) {
        for (final ColorArea prevNeighbor : nextColorNeighbor.getNeighborsArray()) {
            if ((prevNeighbor.getColor() != currColor) && this.flooded.contains(prevNeighbor)) {
                continue next;
            }
        }
        newNext = true;
        break next;
    }
    if (!newNext) {
        if (nextColor < currColor) {
            return false;
        } else {
            // should nextColor have been played before currColor?
            for (final ColorArea nextColorNeighbor : nextColorNeighbors) {
                for (final ColorArea prevNeighbor : nextColorNeighbor.getNeighborsArray()) {
                    if ((prevNeighbor.getColor() == currColor) && !this.flooded.contains(prevNeighbor)) {
                        return false;
                    }
                }
            }
            return true;
        }
    } else {
        return true;
    }
}

/**
 * try to re-use the given node or create a new one
 * and then play the given color in the result node.
 * @param nextColor
 * @param recycleNode
 * @return
 */
public AStarNode copyAndPlay(final byte nextColor, final AStarNode recycleNode, final Board board) {
    final List<ColorArea> nextColorNeighbors = new ArrayList<ColorArea>(128);  // constant, arbitrary initial capacity
    final ColorAreaSet.FastIteratorColorAreaId iter = this.neighbors.fastIteratorColorAreaId();
    int nextId;
    while ((nextId = iter.nextOrNegative()) >= 0) {
        final ColorArea nextColorNeighbor = board.getColorArea4Id(nextId);
        if (nextColorNeighbor.getColor() == nextColor) {
            nextColorNeighbors.add(nextColorNeighbor);
        }
    }
    if (!this.canPlay(nextColor, nextColorNeighbors)) {
        return null;
    } else {
        final AStarNode result;
        if (null == recycleNode) {
            result = new AStarNode(this);
        } else {
            // copy - compare copy constructor
            result = recycleNode;
            result.flooded.copyFrom(this.flooded);
            result.neighbors.copyFrom(this.neighbors);
            System.arraycopy(this.solution, 0, result.solution, 0, this.solutionSize + 1);
            result.solutionSize = this.solutionSize;
            //result.estimatedCost = this.estimatedCost;  // not necessary to copy
        }
        // play - compare method play()
        for (final ColorArea nextColorNeighbor : nextColorNeighbors) {
            result.flooded.add(nextColorNeighbor);
            result.neighbors.addAll(nextColorNeighbor.getNeighborsIdArray());
        }
        result.neighbors.removeAll(result.flooded);
        result.solution[++result.solutionSize] = nextColor;
        return result;
    }
}

Answer 2

C# - 2,098,382步

我尝试了很多事情，大多数都失败了，只是直到最近才起作用。我有一件有趣的事可以贴出答案。

当然还有更多的方法来进一步改善这一点。我想走两百万步可能是可能的。

需要大约7 hours才能生成结果。下面是一个包含所有解决方案的txt文件，以防有人想要检查它们：results.zip

using System;
using System.Collections.Generic;
using System.IO;
using System.Linq;
using System.Threading.Tasks;

namespace FloodPaintAI
{
    class Node
    {   
        public byte Value;             //1-6
        public int Index;              //unique identifier, used for easily deepcopying the graph
        public List<Node> Neighbours;  
        public List<Tuple<int, int>> NeighboursPositions; //used by BuildGraph() 

        public int Depth;    //used by GetSumDistances() 
        public bool Checked; // 

        public Node(byte value, int index)
        {
            Value = value;      
            Index = index;          
        }

        public Node(Node node)
        {           
            Value = node.Value; 
            Index = node.Index;                     
        }
    }

    class Board
    {
        private const int SIZE = 19;
        private const int STARTPOSITION = 9;

        public Node Root;         //root of graph. each node is a set of contiguous, same color square
        public List<Node> Nodes;  //all nodes in the graph, used for deep copying


        public int EstimatedCost; //estimated cost, used by A* Pathfinding
        public List<byte> Solution;

        public Board(StreamReader input)
        {                   
            byte[,] board = new byte[SIZE, SIZE];
            for(int j = 0 ; j < SIZE ; j++)
            {
                string line = input.ReadLine();
                for(int i = 0 ; i < SIZE ; i++)         
                {                                       
                    board[j, i] = byte.Parse(line[i].ToString());
                }               
            }
            Solution = new List<byte>();
            BuildGraph(board);  
        }

        public Board(Board boardToCopy)
        {               
            //copy the graph            
            Nodes = new List<Node>(boardToCopy.Nodes.Count);
            foreach(Node nodeToCopy in boardToCopy.Nodes)
            {
                Node node = new Node(nodeToCopy);
                Nodes.Add(node);
            }

            //copy "Neighbours" property
            for(int i = 0 ; i < boardToCopy.Nodes.Count ; i++)
            {
                Node node = Nodes[i];
                Node nodeToCopy = boardToCopy.Nodes[i];

                node.Neighbours = new List<Node>(nodeToCopy.Neighbours.Count);
                foreach(Node neighbour in nodeToCopy.Neighbours)
                {
                    node.Neighbours.Add(Nodes[neighbour.Index]);
                }
            }

            Root = Nodes[boardToCopy.Root.Index];
            EstimatedCost = boardToCopy.EstimatedCost;          
            Solution = new List<byte>(boardToCopy.Solution);            
        }

        private void BuildGraph(byte[,] board)
        {                       
            int[,] nodeIndexes = new int[SIZE, SIZE];
            Nodes = new List<Node>();

            //check how much sets we have (1st pass)
            for(int j = 0 ; j < SIZE ; j++)
            {
                for(int i = 0 ; i < SIZE ; i++)         
                {               
                    if(nodeIndexes[j, i] == 0) //not already visited                    
                    {
                        Node newNode = new Node(board[j, i], Nodes.Count);                      
                        newNode.NeighboursPositions = new List<Tuple<int, int>>();
                        Nodes.Add(newNode);

                        BuildGraphFloodFill(board, nodeIndexes, newNode, i, j, board[j, i]);
                    }
                }       
            }

            //set neighbours and root (2nd pass)
            foreach(Node node in Nodes)
            {
                node.Neighbours = new List<Node>();
                node.Neighbours.AddRange(node.NeighboursPositions.Select(x => nodeIndexes[x.Item2, x.Item1]).Distinct().Select(x => Nodes[x - 1]));
                node.NeighboursPositions = null;                
            }
            Root = Nodes[nodeIndexes[STARTPOSITION, STARTPOSITION] - 1];            
        }

        private void BuildGraphFloodFill(byte[,] board, int[,] nodeIndexes, Node node, int startx, int starty, byte floodvalue)
        {
            Queue<Tuple<int, int>> queue = new Queue<Tuple<int, int>>();
            queue.Enqueue(new Tuple<int, int>(startx, starty));

            while(queue.Count > 0)
            {
                Tuple<int, int> position = queue.Dequeue();
                int x = position.Item1;
                int y = position.Item2;

                if(x >= 0 && x < SIZE && y >= 0 && y < SIZE)
                {
                    if(nodeIndexes[y, x] == 0 && board[y, x] == floodvalue)
                    {
                        nodeIndexes[y, x] = node.Index + 1;

                        queue.Enqueue(new Tuple<int, int>(x + 1, y));
                        queue.Enqueue(new Tuple<int, int>(x - 1, y));
                        queue.Enqueue(new Tuple<int, int>(x, y + 1));
                        queue.Enqueue(new Tuple<int, int>(x, y - 1));                                           
                    }               
                    if(board[y, x] != floodvalue)
                        node.NeighboursPositions.Add(position);                         
                }       
            }
        }

        public int GetEstimatedCost()
        {       
            Board current = this;

            //copy current board and play the best color until the end.
            //number of moves required to go the end is the heuristic
            //estimated cost = current cost + heuristic
            while(!current.IsSolved())
            {
                int minSumDistance = int.MaxValue;
                Board minBoard = null;

                //find color which give the minimum sum of distance from root to each other node
                foreach(byte i in current.Root.Neighbours.Select(x => x.Value).Distinct())
                {
                    Board copy = new Board(current);
                    copy.Play(i);                   

                    int distance = copy.GetSumDistances();                  

                    if(distance < minSumDistance)
                    {
                        minSumDistance = distance;
                        minBoard = copy;
                    }
                }
                current = minBoard;
            }           
            return current.Solution.Count;
        }

        public int GetSumDistances()
        {
            //get sum of distances from root to each other node, using BFS
            int sumDistances = 0;           

            //reset marker
            foreach(Node n in Nodes)
            {
                n.Checked = false;                                  
            }

            Queue<Node> queue = new Queue<Node>();
            Root.Checked = true;
            Root.Depth = 0; 
            queue.Enqueue(Root);

            while(queue.Count > 0)
            {
                Node current = queue.Dequeue();                             
                foreach(Node n in current.Neighbours)
                {
                    if(!n.Checked)          
                    {                                   
                        n.Checked = true;                                               
                        n.Depth = current.Depth + 1;
                        sumDistances += n.Depth;            
                        queue.Enqueue(n);   
                    }               
                }
            }
            return sumDistances;
        }       

        public void Play(byte value)            
        {
            //merge root node with other neighbours nodes, if color is matching
            Root.Value = value;
            List<Node> neighboursToRemove = Root.Neighbours.Where(x => x.Value == value).ToList();
            List<Node> neighboursToAdd = neighboursToRemove.SelectMany(x => x.Neighbours).Except((new Node[] { Root }).Concat(Root.Neighbours)).ToList();

            foreach(Node n in neighboursToRemove)
            {
                foreach(Node m in n.Neighbours)
                {
                    m.Neighbours.Remove(n);
                }
                Nodes.Remove(n);
            }   

            foreach(Node n in neighboursToAdd)
            {
                Root.Neighbours.Add(n);         
                n.Neighbours.Add(Root); 
            }           

            //re-synchronize node index
            for(int i = 0 ; i < Nodes.Count ; i++)
            {
                Nodes[i].Index = i;
            }           
            Solution.Add(value);
        }

        public bool IsSolved()
        {           
            //return Nodes.Count == 1;
            return Root.Neighbours.Count == 0;  
        }           
    }


    class Program
    {       
        public static List<byte> Solve(Board input)
        {
            //A* Pathfinding            
            LinkedList<Board> open = new LinkedList<Board>();       
            input.EstimatedCost = input.GetEstimatedCost();
            open.AddLast(input);            

            while(open.Count > 0)
            {                       
                Board current = open.First.Value;
                open.RemoveFirst();

                if(current.IsSolved())
                {
                    return current.Solution;                
                }
                else
                {
                    //play all neighbours nodes colors
                    foreach(byte i in current.Root.Neighbours.Select(x => x.Value).Distinct())
                    {                       
                        Board newBoard = new Board(current);
                        newBoard.Play(i);           
                        newBoard.EstimatedCost = newBoard.GetEstimatedCost();   

                        //insert board to open list
                        bool inserted = false;
                        for(LinkedListNode<Board> node = open.First ; node != null ; node = node.Next)
                        {                               
                            if(node.Value.EstimatedCost > newBoard.EstimatedCost)
                            {
                                open.AddBefore(node, newBoard);
                                inserted = true;
                                break;
                            }
                        }       
                        if(!inserted)
                            open.AddLast(newBoard);                                                 
                    }   
                }   
            }
            throw new Exception(); //no solution found, impossible
        }   

        public static void Main(string[] args)
        {                   
            using (StreamReader sr = new StreamReader("floodtest"))
            {   
                while(!sr.EndOfStream)
                {                               
                    List<Board> boards = new List<Board>();
                    while(!sr.EndOfStream && boards.Count < 100)
                    {
                        Board board = new Board(sr);                        
                        sr.ReadLine(); //skip empty line
                        boards.Add(board);
                    }                                           
                    List<byte>[] solutions = new List<byte>[boards.Count];                                          
                    Parallel.For(0, boards.Count, i => 
                    {                               
                        solutions[i] = Solve(boards[i]); 
                    });                                         
                    foreach(List<byte> solution in solutions)
                    {
                        Console.WriteLine(string.Join(string.Empty, solution));                                             
                    }       
                }               
            }
        }
    }
}

关于其工作方式的更多详细信息：

它采用一个*寻路算法。

困难的是找到一个好的heuristic。如果heuristic低估了成本，它的性能就像迪克斯特拉算法一样，而且由于一个6种颜色的19x19板的复杂性，它将永远运行。如果它高估了成本，它将很快收敛到一个解决方案上，但根本不会给出好的解决方案(大约有26次移动，19次是可能的)。找到完美的heuristic，给出达到解决方案所需步骤的确切剩余量将是最好的(它将是快速的，并且会给出最佳的可能解决方案)。这是不可能的(除非我错了)。它实际上需要首先解决董事会本身，而这正是你真正想要做的(鸡和蛋问题)。

我尝试了很多东西，下面是heuristic的工作原理：

• 我建立了一个图表的当前董事会来评估。每个node表示一组相邻的、相同颜色的正方形。使用该graph，我可以很容易地计算出从中心到任何其他节点的精确最小距离。例如，从中间到左上角的距离是10，因为至少有10种颜色将它们分开。
• 计算heuristic：我玩当前的板，直到结束。对于每一步，我选择颜色，这将最小化从根到所有其他节点的距离之和。
• 达到这一目的所需的移动数是heuristic。
• Estimated cost (由A*使用)= moves so far + heuristic

它并不完美，因为它稍微高估了成本(因此找到了非最优解)。无论如何，它是快速的计算和提供良好的结果。

通过使用图形来执行所有的操作，我能够获得巨大的速度提升。开始时，我有一个2-dimension数组。当需要时，我给它注入洪水并重新计算图(例如:计算启发式)。现在所有的事情都是用图表来完成的，它只是在一开始计算出来的。为了模拟步骤，不再需要洪水填充，而是合并节点。这要快得多。

Answer 3

Python - 10,800,000步

作为最后的参考解决方案，请考虑以下顺序：

print "123456" * 18

通过所有的颜色，n时间，意味着每一个平方n的步骤，将保证相同的颜色与中心广场。每个正方形最多离中心18步，所以18个周期将保证所有的方块都是相同的颜色。最有可能的情况下，它将在较少的完成，但程序不需要停止，只要所有的方块是相同的颜色，它只是更有益于这样做。这个不变的过程是每个测试用例108个步骤，总共10,800,000步。