当我看到问题陈述时，我想到了“Trie”。但看到其他一些海报使用了这种方法，我寻找另一种不同的方法。可惜的是，Trie方法表现更好。我在我的机器上运行了Kent的Perl解决方案，在调整它以使用我的字典文件后，它花了0.31秒运行。我自己的perl实现需要0.54秒才能运行。

这就是我的方法:

Create a transition hash to model the legal transitions. Iterate through all 16^3 possible three letter combinations. In the loop, exclude illegal transitions and repeat visits to the same square. Form all the legal 3-letter sequences and store them in a hash. Then loop through all words in the dictionary. Exclude words that are too long or short Slide a 3-letter window across each word and see if it is among the 3-letter combos from step 2. Exclude words that fail. This eliminates most non-matches. If still not eliminated, use a recursive algorithm to see if the word can be formed by making paths through the puzzle. (This part is slow, but called infrequently.) Print out the words I found. I tried 3-letter and 4-letter sequences, but 4-letter sequences slowed the program down.

在我的代码中，我使用/usr/share/dict/words作为我的字典。它是MAC OS X和许多Unix系统的标准配置。如果你愿意，你可以使用另一个文件。要破解不同的谜题，只需更改变量@puzzle。这将很容易适应更大的矩阵。你只需要改变%transitions哈希值和%legalTransitions哈希值。

这种解决方案的优点是代码短，数据结构简单。

下面是Perl代码(我知道它使用了太多的全局变量):

#!/usr/bin/perl
use Time::HiRes  qw{ time };

sub readFile($);
sub findAllPrefixes($);
sub isWordTraceable($);
sub findWordsInPuzzle(@);

my $startTime = time;

# Puzzle to solve

my @puzzle = ( 
    F, X, I, E,
    A, M, L, O,
    E, W, B, X,
    A, S, T, U
);

my $minimumWordLength = 3;
my $maximumPrefixLength = 3; # I tried four and it slowed down.

# Slurp the word list.
my $wordlistFile = "/usr/share/dict/words";

my @words = split(/\n/, uc(readFile($wordlistFile)));
print "Words loaded from word list: " . scalar @words . "\n";

print "Word file load time: " . (time - $startTime) . "\n";
my $postLoad = time;

# Define the legal transitions from one letter position to another. 
# Positions are numbered 0-15.
#     0  1  2  3
#     4  5  6  7
#     8  9 10 11
#    12 13 14 15
my %transitions = ( 
   -1 => [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],
    0 => [1,4,5], 
    1 => [0,2,4,5,6],
    2 => [1,3,5,6,7],
    3 => [2,6,7],
    4 => [0,1,5,8,9],
    5 => [0,1,2,4,6,8,9,10],
    6 => [1,2,3,5,7,9,10,11],
    7 => [2,3,6,10,11],
    8 => [4,5,9,12,13],
    9 => [4,5,6,8,10,12,13,14],
    10 => [5,6,7,9,11,13,14,15],
    11 => [6,7,10,14,15],
    12 => [8,9,13],
    13 => [8,9,10,12,14],
    14 => [9,10,11,13,15],
    15 => [10,11,14]
);

# Convert the transition matrix into a hash for easy access.
my %legalTransitions = ();
foreach my $start (keys %transitions) {
    my $legalRef = $transitions{$start};
    foreach my $stop (@$legalRef) {
        my $index = ($start + 1) * (scalar @puzzle) + ($stop + 1);
        $legalTransitions{$index} = 1;
    }
}

my %prefixesInPuzzle = findAllPrefixes($maximumPrefixLength);

print "Find prefixes time: " . (time - $postLoad) . "\n";
my $postPrefix = time;

my @wordsFoundInPuzzle = findWordsInPuzzle(@words);

print "Find words in puzzle time: " . (time - $postPrefix) . "\n";

print "Unique prefixes found: " . (scalar keys %prefixesInPuzzle) . "\n";
print "Words found (" . (scalar @wordsFoundInPuzzle) . ") :\n    " . join("\n    ", @wordsFoundInPuzzle) . "\n";

print "Total Elapsed time: " . (time - $startTime) . "\n";

###########################################

sub readFile($) {
    my ($filename) = @_;
    my $contents;
    if (-e $filename) {
        # This is magic: it opens and reads a file into a scalar in one line of code. 
        # See http://www.perl.com/pub/a/2003/11/21/slurp.html
        $contents = do { local( @ARGV, $/ ) = $filename ; <> } ; 
    }
    else {
        $contents = '';
    }
    return $contents;
}

# Is it legal to move from the first position to the second? They must be adjacent.
sub isLegalTransition($$) {
    my ($pos1,$pos2) = @_;
    my $index = ($pos1 + 1) * (scalar @puzzle) + ($pos2 + 1);
    return $legalTransitions{$index};
}

# Find all prefixes where $minimumWordLength <= length <= $maxPrefixLength
#
#   $maxPrefixLength ... Maximum length of prefix we will store. Three gives best performance. 
sub findAllPrefixes($) {
    my ($maxPrefixLength) = @_;
    my %prefixes = ();
    my $puzzleSize = scalar @puzzle;

    # Every possible N-letter combination of the letters in the puzzle 
    # can be represented as an integer, though many of those combinations
    # involve illegal transitions, duplicated letters, etc.
    # Iterate through all those possibilities and eliminate the illegal ones.
    my $maxIndex = $puzzleSize ** $maxPrefixLength;

    for (my $i = 0; $i < $maxIndex; $i++) {
        my @path;
        my $remainder = $i;
        my $prevPosition = -1;
        my $prefix = '';
        my %usedPositions = ();
        for (my $prefixLength = 1; $prefixLength <= $maxPrefixLength; $prefixLength++) {
            my $position = $remainder % $puzzleSize;

            # Is this a valid step?
            #  a. Is the transition legal (to an adjacent square)?
            if (! isLegalTransition($prevPosition, $position)) {
                last;
            }

            #  b. Have we repeated a square?
            if ($usedPositions{$position}) {
                last;
            }
            else {
                $usedPositions{$position} = 1;
            }

            # Record this prefix if length >= $minimumWordLength.
            $prefix .= $puzzle[$position];
            if ($prefixLength >= $minimumWordLength) {
                $prefixes{$prefix} = 1;
            }

            push @path, $position;
            $remainder -= $position;
            $remainder /= $puzzleSize;
            $prevPosition = $position;
        } # end inner for
    } # end outer for
    return %prefixes;
}

# Loop through all words in dictionary, looking for ones that are in the puzzle.
sub findWordsInPuzzle(@) {
    my @allWords = @_;
    my @wordsFound = ();
    my $puzzleSize = scalar @puzzle;
WORD: foreach my $word (@allWords) {
        my $wordLength = length($word);
        if ($wordLength > $puzzleSize || $wordLength < $minimumWordLength) {
            # Reject word as too short or too long.
        }
        elsif ($wordLength <= $maximumPrefixLength ) {
            # Word should be in the prefix hash.
            if ($prefixesInPuzzle{$word}) {
                push @wordsFound, $word;
            }
        }
        else {
            # Scan through the word using a window of length $maximumPrefixLength, looking for any strings not in our prefix list.
            # If any are found that are not in the list, this word is not possible.
            # If no non-matches are found, we have more work to do.
            my $limit = $wordLength - $maximumPrefixLength + 1;
            for (my $startIndex = 0; $startIndex < $limit; $startIndex ++) {
                if (! $prefixesInPuzzle{substr($word, $startIndex, $maximumPrefixLength)}) {
                    next WORD;
                }
            }
            if (isWordTraceable($word)) {
                # Additional test necessary: see if we can form this word by following legal transitions
                push @wordsFound, $word;
            }
        }

    }
    return @wordsFound;
}

# Is it possible to trace out the word using only legal transitions?
sub isWordTraceable($) {
    my $word = shift;
    return traverse([split(//, $word)], [-1]); # Start at special square -1, which may transition to any square in the puzzle.
}

# Recursively look for a path through the puzzle that matches the word.
sub traverse($$) {
    my ($lettersRef, $pathRef) = @_;
    my $index = scalar @$pathRef - 1;
    my $position = $pathRef->[$index];
    my $letter = $lettersRef->[$index];
    my $branchesRef =  $transitions{$position};
BRANCH: foreach my $branch (@$branchesRef) {
            if ($puzzle[$branch] eq $letter) {
                # Have we used this position yet?
                foreach my $usedBranch (@$pathRef) {
                    if ($usedBranch == $branch) {
                        next BRANCH;
                    }
                }
                if (scalar @$lettersRef == $index + 1) {
                    return 1; # End of word and success.
                }
                push @$pathRef, $branch;
                if (traverse($lettersRef, $pathRef)) {
                    return 1; # Recursive success.
                }
                else {
                    pop @$pathRef;
                }
            }
        }
    return 0; # No path found. Failed.
}

只是为了好玩，我在bash中实现了一个。 它不是超级快，但很合理。

http://dev.xkyle.com/bashboggle/

你可以把这个问题分成两部分:

某种搜索算法可以在网格中列举出可能的字符串。 一种测试字符串是否是有效单词的方法。

理想情况下，(2)还应该包括一种测试字符串是否是有效单词前缀的方法——这将允许您精简搜索并节省大量时间。

亚当·罗森菲尔德(Adam Rosenfield)的Trie是(2)的一个解决方案。它很优雅，可能是算法专家的首选，但有了现代语言和现代计算机，我们可能会更懒一点。此外，正如Kent所建议的，我们可以通过丢弃网格中没有字母的单词来减少字典的大小。这是一些蟒蛇:

def make_lookups(grid, fn='dict.txt'):
    # Make set of valid characters.
    chars = set()
    for word in grid:
        chars.update(word)

    words = set(x.strip() for x in open(fn) if set(x.strip()) <= chars)
    prefixes = set()
    for w in words:
        for i in range(len(w)+1):
            prefixes.add(w[:i])

    return words, prefixes

哇;常数时间前缀测试。加载你链接的字典需要几秒钟，但只有几秒钟:-)(注意words <= prefixes)

现在，对于第(1)部分，我倾向于用图表来思考。所以我将创建一个像这样的字典:

graph = { (x, y):set([(x0,y0), (x1,y1), (x2,y2)]), }

例如，graph[(x, y)]是你从位置(x, y)可以到达的坐标集。我还将添加一个虚拟节点None，它将连接到所有东西。

构建它有点笨拙，因为有8个可能的位置，你必须做边界检查。下面是一些相应笨拙的python代码:

def make_graph(grid):
    root = None
    graph = { root:set() }
    chardict = { root:'' }

    for i, row in enumerate(grid):
        for j, char in enumerate(row):
            chardict[(i, j)] = char
            node = (i, j)
            children = set()
            graph[node] = children
            graph[root].add(node)
            add_children(node, children, grid)

    return graph, chardict

def add_children(node, children, grid):
    x0, y0 = node
    for i in [-1,0,1]:
        x = x0 + i
        if not (0 <= x < len(grid)):
            continue
        for j in [-1,0,1]:
            y = y0 + j
            if not (0 <= y < len(grid[0])) or (i == j == 0):
                continue

            children.add((x,y))

这段代码还建立了一个字典映射(x,y)到相应的字符。这让我把一个位置列表转换成一个单词:

def to_word(chardict, pos_list):
    return ''.join(chardict[x] for x in pos_list)

最后，我们进行深度优先搜索。基本程序是:

搜索到达一个特定的节点。 检查到目前为止的路径是否可能是单词的一部分。如果不是，就不要进一步探索这个分支。 检查到目前为止的路径是否是一个单词。如果是，则添加到结果列表中。 探索迄今为止所有孩子未走的路。

Python:

def find_words(graph, chardict, position, prefix, results, words, prefixes):
    """ Arguments:
      graph :: mapping (x,y) to set of reachable positions
      chardict :: mapping (x,y) to character
      position :: current position (x,y) -- equals prefix[-1]
      prefix :: list of positions in current string
      results :: set of words found
      words :: set of valid words in the dictionary
      prefixes :: set of valid words or prefixes thereof
    """
    word = to_word(chardict, prefix)

    if word not in prefixes:
        return

    if word in words:
        results.add(word)

    for child in graph[position]:
        if child not in prefix:
            find_words(graph, chardict, child, prefix+[child], results, words, prefixes)

运行代码如下:

grid = ['fxie', 'amlo', 'ewbx', 'astu']
g, c = make_graph(grid)
w, p = make_lookups(grid)
res = set()
find_words(g, c, None, [], res, w, p)

检查保留区，看看答案。下面是为你的例子找到的单词列表，按大小排序:

 ['a', 'b', 'e', 'f', 'i', 'l', 'm', 'o', 's', 't',
 'u', 'w', 'x', 'ae', 'am', 'as', 'aw', 'ax', 'bo',
 'bu', 'ea', 'el', 'em', 'es', 'fa', 'ie', 'io', 'li',
 'lo', 'ma', 'me', 'mi', 'oe', 'ox', 'sa', 'se', 'st',
 'tu', 'ut', 'wa', 'we', 'xi', 'aes', 'ame', 'ami',
 'ase', 'ast', 'awa', 'awe', 'awl', 'blo', 'but', 'elb',
 'elm', 'fae', 'fam', 'lei', 'lie', 'lim', 'lob', 'lox',
 'mae', 'maw', 'mew', 'mil', 'mix', 'oil', 'olm', 'saw',
 'sea', 'sew', 'swa', 'tub', 'tux', 'twa', 'wae', 'was',
 'wax', 'wem', 'ambo', 'amil', 'amli', 'asem', 'axil',
 'axle', 'bleo', 'boil', 'bole', 'east', 'fame', 'limb',
 'lime', 'mesa', 'mewl', 'mile', 'milo', 'oime', 'sawt',
 'seam', 'seax', 'semi', 'stub', 'swam', 'twae', 'twas',
 'wame', 'wase', 'wast', 'weam', 'west', 'amble', 'awest',
 'axile', 'embox', 'limbo', 'limes', 'swami', 'embole',
 'famble', 'semble', 'wamble']

代码需要(字面上的)几秒钟来加载字典，但其余的在我的机器上是立即完成的。

我也用Java解决了这个问题。我的实现有269行，非常容易使用。首先，您需要创建Boggler类的一个新实例，然后用网格作为参数调用solve函数。在我的电脑上加载5万个单词的字典大约需要100毫秒，它在大约10-20毫秒内找到单词。找到的单词存储在一个数组列表中，即foundWords。

import java.io.BufferedReader;
import java.io.File;
import java.io.FileInputStream;
import java.io.FileNotFoundException;
import java.io.IOException;
import java.io.InputStreamReader;
import java.net.URISyntaxException;
import java.net.URL;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Comparator;

public class Boggler {
    private ArrayList<String> words = new ArrayList<String>();      
    private ArrayList<String> roundWords = new ArrayList<String>(); 
    private ArrayList<Word> foundWords = new ArrayList<Word>();     
    private char[][] letterGrid = new char[4][4];                   
    private String letters;                                         

    public Boggler() throws FileNotFoundException, IOException, URISyntaxException {
        long startTime = System.currentTimeMillis();

        URL path = GUI.class.getResource("words.txt");
        BufferedReader br = new BufferedReader(new InputStreamReader(new FileInputStream(new File(path.toURI()).getAbsolutePath()), "iso-8859-1"));
        String line;
        while((line = br.readLine()) != null) {
            if(line.length() < 3 || line.length() > 10) {
                continue;
            }

            this.words.add(line);
        }
    }

    public ArrayList<Word> getWords() {
        return this.foundWords;
    }

    public void solve(String letters) {
        this.letters = "";
        this.foundWords = new ArrayList<Word>();

        for(int i = 0; i < letters.length(); i++) {
            if(!this.letters.contains(letters.substring(i, i + 1))) {
                this.letters += letters.substring(i, i + 1);
            }
        }

        for(int i = 0; i < 4; i++) {
            for(int j = 0; j < 4; j++) {
                this.letterGrid[i][j] = letters.charAt(i * 4 + j);
            }
        }

        System.out.println(Arrays.deepToString(this.letterGrid));               

        this.roundWords = new ArrayList<String>();      
        String pattern = "[" + this.letters + "]+";     

        for(int i = 0; i < this.words.size(); i++) {

            if(this.words.get(i).matches(pattern)) {
                this.roundWords.add(this.words.get(i));
            }
        }

        for(int i = 0; i < this.roundWords.size(); i++) {
            Word word = checkForWord(this.roundWords.get(i));

            if(word != null) {
                System.out.println(word);
                this.foundWords.add(word);
            }
        }       
    }

    private Word checkForWord(String word) {
        char initial = word.charAt(0);
        ArrayList<LetterCoord> startPoints = new ArrayList<LetterCoord>();

        int x = 0;  
        int y = 0;
        for(char[] row: this.letterGrid) {
            x = 0;

            for(char letter: row) {
                if(initial == letter) {
                    startPoints.add(new LetterCoord(x, y));
                }

                x++;
            }

            y++;
        }

        ArrayList<LetterCoord> letterCoords = null;
        for(int initialTry = 0; initialTry < startPoints.size(); initialTry++) {
            letterCoords = new ArrayList<LetterCoord>();    

            x = startPoints.get(initialTry).getX(); 
            y = startPoints.get(initialTry).getY();

            LetterCoord initialCoord = new LetterCoord(x, y);
            letterCoords.add(initialCoord);

            letterLoop: for(int letterIndex = 1; letterIndex < word.length(); letterIndex++) {
                LetterCoord lastCoord = letterCoords.get(letterCoords.size() - 1);  
                char currentChar = word.charAt(letterIndex);                        

                ArrayList<LetterCoord> letterLocations = getNeighbours(currentChar, lastCoord.getX(), lastCoord.getY());

                if(letterLocations == null) {
                    return null;    
                }       

                for(int foundIndex = 0; foundIndex < letterLocations.size(); foundIndex++) {
                    if(letterIndex != word.length() - 1 && true == false) {
                        char nextChar = word.charAt(letterIndex + 1);
                        int lastX = letterCoords.get(letterCoords.size() - 1).getX();
                        int lastY = letterCoords.get(letterCoords.size() - 1).getY();

                        ArrayList<LetterCoord> possibleIndex = getNeighbours(nextChar, lastX, lastY);
                        if(possibleIndex != null) {
                            if(!letterCoords.contains(letterLocations.get(foundIndex))) {
                                letterCoords.add(letterLocations.get(foundIndex));
                            }
                            continue letterLoop;
                        } else {
                            return null;
                        }
                    } else {
                        if(!letterCoords.contains(letterLocations.get(foundIndex))) {
                            letterCoords.add(letterLocations.get(foundIndex));

                            continue letterLoop;
                        }
                    }
                }
            }

            if(letterCoords != null) {
                if(letterCoords.size() == word.length()) {
                    Word w = new Word(word);
                    w.addList(letterCoords);
                    return w;
                } else {
                    return null;
                }
            }
        }

        if(letterCoords != null) {
            Word foundWord = new Word(word);
            foundWord.addList(letterCoords);

            return foundWord;
        }

        return null;
    }

    public ArrayList<LetterCoord> getNeighbours(char letterToSearch, int x, int y) {
        ArrayList<LetterCoord> neighbours = new ArrayList<LetterCoord>();

        for(int _y = y - 1; _y <= y + 1; _y++) {
            for(int _x = x - 1; _x <= x + 1; _x++) {
                if(_x < 0 || _y < 0 || (_x == x && _y == y) || _y > 3 || _x > 3) {
                    continue;
                }

                if(this.letterGrid[_y][_x] == letterToSearch && !neighbours.contains(new LetterCoord(_x, _y))) {
                    neighbours.add(new LetterCoord(_x, _y));
                }
            }
        }

        if(neighbours.isEmpty()) {
            return null;
        } else {
            return neighbours;
        }
    }
}

class Word {
    private String word;    
    private ArrayList<LetterCoord> letterCoords = new ArrayList<LetterCoord>();

    public Word(String word) {
        this.word = word;
    }

    public boolean addCoords(int x, int y) {
        LetterCoord lc = new LetterCoord(x, y);

        if(!this.letterCoords.contains(lc)) {
            this.letterCoords.add(lc);

            return true;
        }

        return false;
    }

    public void addList(ArrayList<LetterCoord> letterCoords) {
        this.letterCoords = letterCoords;
    } 

    @Override
    public String toString() {
        String outputString = this.word + " ";
        for(int i = 0; i < letterCoords.size(); i++) {
            outputString += "(" + letterCoords.get(i).getX() + ", " + letterCoords.get(i).getY() + ") ";
        }

        return outputString;
    }

    public String getWord() {
        return this.word;
    }

    public ArrayList<LetterCoord> getList() {
        return this.letterCoords;
    }
}

class LetterCoord extends ArrayList {
    private int x;          
    private int y;          

    public LetterCoord(int x, int y) {
        this.x = x;
        this.y = y;
    }

    public int getX() {
        return this.x;
    }

    public int getY() {
        return this.y;
    }

    @Override
    public boolean equals(Object o) {
        if(!(o instanceof LetterCoord)) {
            return false;
        }

        LetterCoord lc = (LetterCoord) o;

        if(this.x == lc.getX() &&
                this.y == lc.getY()) {
            return true;
        }

        return false;
    }

    @Override
    public int hashCode() {
        int hash = 7;
        hash = 29 * hash + this.x;
        hash = 24 * hash + this.y;
        return hash;
    }
}

最快的解决方案可能是将字典存储在一个trie中。然后，创建一个三元组队列(x, y, s)，其中队列中的每个元素对应于一个可以在网格中拼写的单词的前缀s，结束于位置(x, y)。初始化队列中有N x N个元素(其中N是网格的大小)，网格中的每个正方形都有一个元素。然后，算法进行如下:

While the queue is not empty:
  Dequeue a triple (x, y, s)
  For each square (x', y') with letter c adjacent to (x, y):
    If s+c is a word, output s+c
    If s+c is a prefix of a word, insert (x', y', s+c) into the queue

如果将字典存储在trie中，则可以在常数时间内测试s+c是否是单词或单词的前缀(前提是还在每个队列数据中保留一些额外的元数据，例如指向trie中当前节点的指针)，因此此算法的运行时间为O(可拼写的单词数量)。

[编辑]下面是我刚刚编写的Python实现:

#!/usr/bin/python

class TrieNode:
    def __init__(self, parent, value):
        self.parent = parent
        self.children = [None] * 26
        self.isWord = False
        if parent is not None:
            parent.children[ord(value) - 97] = self

def MakeTrie(dictfile):
    dict = open(dictfile)
    root = TrieNode(None, '')
    for word in dict:
        curNode = root
        for letter in word.lower():
            if 97 <= ord(letter) < 123:
                nextNode = curNode.children[ord(letter) - 97]
                if nextNode is None:
                    nextNode = TrieNode(curNode, letter)
                curNode = nextNode
        curNode.isWord = True
    return root

def BoggleWords(grid, dict):
    rows = len(grid)
    cols = len(grid[0])
    queue = []
    words = []
    for y in range(cols):
        for x in range(rows):
            c = grid[y][x]
            node = dict.children[ord(c) - 97]
            if node is not None:
                queue.append((x, y, c, node))
    while queue:
        x, y, s, node = queue[0]
        del queue[0]
        for dx, dy in ((1, 0), (1, -1), (0, -1), (-1, -1), (-1, 0), (-1, 1), (0, 1), (1, 1)):
            x2, y2 = x + dx, y + dy
            if 0 <= x2 < cols and 0 <= y2 < rows:
                s2 = s + grid[y2][x2]
                node2 = node.children[ord(grid[y2][x2]) - 97]
                if node2 is not None:
                    if node2.isWord:
                        words.append(s2)
                    queue.append((x2, y2, s2, node2))

    return words

使用示例:

d = MakeTrie('/usr/share/dict/words')
print(BoggleWords(['fxie','amlo','ewbx','astu'], d))

输出:

['fa', 'xi', 'ie', 'io', 'el', 'am', 'ax', 'ae', 'aw', 'mi', 'ma', 'me', 'lo', 'li', 'oe', 'ox', 'em', 'ea', 'ea', 'es', 'wa', 'we', 'wa', 'bo', 'bu', 'as', 'aw', 'ae', 'st', 'se', 'sa', 'tu', 'ut', 'fam', 'fae', 'imi', 'eli', 'elm', 'elb', 'ami', 'ama', 'ame', 'aes', 'awl', 'awa', 'awe', 'awa', 'mix', 'mim', 'mil', 'mam', 'max', 'mae', 'maw', 'mew', 'mem', 'mes', 'lob', 'lox', 'lei', 'leo', 'lie', 'lim', 'oil', 'olm', 'ewe', 'eme', 'wax', 'waf', 'wae', 'waw', 'wem', 'wea', 'wea', 'was', 'waw', 'wae', 'bob', 'blo', 'bub', 'but', 'ast', 'ase', 'asa', 'awl', 'awa', 'awe', 'awa', 'aes', 'swa', 'swa', 'sew', 'sea', 'sea', 'saw', 'tux', 'tub', 'tut', 'twa', 'twa', 'tst', 'utu', 'fama', 'fame', 'ixil', 'imam', 'amli', 'amil', 'ambo', 'axil', 'axle', 'mimi', 'mima', 'mime', 'milo', 'mile', 'mewl', 'mese', 'mesa', 'lolo', 'lobo', 'lima', 'lime', 'limb', 'lile', 'oime', 'oleo', 'olio', 'oboe', 'obol', 'emim', 'emil', 'east', 'ease', 'wame', 'wawa', 'wawa', 'weam', 'west', 'wese', 'wast', 'wase', 'wawa', 'wawa', 'boil', 'bolo', 'bole', 'bobo', 'blob', 'bleo', 'bubo', 'asem', 'stub', 'stut', 'swam', 'semi', 'seme', 'seam', 'seax', 'sasa', 'sawt', 'tutu', 'tuts', 'twae', 'twas', 'twae', 'ilima', 'amble', 'axile', 'awest', 'mamie', 'mambo', 'maxim', 'mease', 'mesem', 'limax', 'limes', 'limbo', 'limbu', 'obole', 'emesa', 'embox', 'awest', 'swami', 'famble', 'mimble', 'maxima', 'embolo', 'embole', 'wamble', 'semese', 'semble', 'sawbwa', 'sawbwa']

Notes: This program doesn't output 1-letter words, or filter by word length at all. That's easy to add but not really relevant to the problem. It also outputs some words multiple times if they can be spelled in multiple ways. If a given word can be spelled in many different ways (worst case: every letter in the grid is the same (e.g. 'A') and a word like 'aaaaaaaaaa' is in your dictionary), then the running time will get horribly exponential. Filtering out duplicates and sorting is trivial to due after the algorithm has finished.

    package ProblemSolving;

import java.util.HashSet;
import java.util.Set;

/**
 * Given a 2-dimensional array of characters and a
 * dictionary in which a word can be searched in O(1) time.
 * Need to print all the words from array which are present
 * in dictionary. Word can be formed in any direction but
 * has to end at any edge of array.
 * (Need not worry much about the dictionary)
 */
public class DictionaryWord {
    private static char[][] matrix = new char[][]{
            {'a', 'f', 'h', 'u', 'n'},
            {'e', 't', 'a', 'i', 'r'},
            {'a', 'e', 'g', 'g', 'o'},
            {'t', 'r', 'm', 'l', 'p'}
    };
    private static int dim_x = matrix.length;
    private static int dim_y = matrix[matrix.length -1].length;
    private static Set<String> wordSet = new HashSet<String>();

    public static void main(String[] args) {
        //dictionary
        wordSet.add("after");
        wordSet.add("hate");
        wordSet.add("hair");
        wordSet.add("air");
        wordSet.add("eat");
        wordSet.add("tea");

        for (int x = 0; x < dim_x; x++) {
            for (int y = 0; y < dim_y; y++) {
                checkAndPrint(matrix[x][y] + "");
                int[][] visitedMap = new int[dim_x][dim_y];
                visitedMap[x][y] = 1;
                recursion(matrix[x][y] + "", visitedMap, x, y);
            }
        }
    }

    private static void checkAndPrint(String word) {
        if (wordSet.contains(word)) {
            System.out.println(word);
        }
    }

    private static void recursion(String word, int[][] visitedMap, int x, int y) {
        for (int i = Math.max(x - 1, 0); i < Math.min(x + 2, dim_x); i++) {
            for (int j = Math.max(y - 1, 0); j < Math.min(y + 2, dim_y); j++) {
                if (visitedMap[i][j] == 1) {
                    continue;
                } else {
                    int[][] newVisitedMap = new int[dim_x][dim_y];
                    for (int p = 0; p < dim_x; p++) {
                        for (int q = 0; q < dim_y; q++) {
                           newVisitedMap[p][q] = visitedMap[p][q];
                        }
                    }
                    newVisitedMap[i][j] = 1;
                    checkAndPrint(word + matrix[i][j]);
                    recursion(word + matrix[i][j], newVisitedMap, i, j);
                }
            }
        }
    }

}