我创建了一个用于所有方向的单词求解器。它可以在水平方向、垂直方向和反向方向查找单词。然而,我在让它向所有方向发展时遇到了问题。所以把"hello“放在下面:
H E i l
x L p q
c L O m有谁能教我怎么做吗?以下是我搜索单词的算法(在C++中):
/*
* For loops that search each row, each column in all 8 possible directions.
*/
void Scramble::solve() {
cout << "Output:" << endl;
for (int row = 0; row < getRows(); row++) {
for (int col = 0; col < getCols(); col++)
for (int rowDir = -1; rowDir <= 1; rowDir++)
for (int colDir = -1; colDir <=1; colDir++)
if (rowDir != 0 || colDir != 0)
findWords(row, col, rowDir, colDir);
}
}
/*
* Finds the matches in a given direction. Also calls verifyWord() to verify that the
* current sequence of letters could possibly form a word. If not, search stops.
*/
void Scramble::findWords(int startingRow, int startingCol, int rowDir, int colDir) {
int searchResult;
string sequence = "";
sequence = sequence + wordsArr[startingRow][startingCol];
for (int i = startingRow + rowDir, j = startingCol + colDir; i >= 0 && j >= 0
&& i < getRows() && j < getCols(); i = i + rowDir, j = j + colDir) {
sequence = sequence + wordsArr[i][j];
if (sequence.length() >= 3) {
searchResult = verifyWord(words, sequence);
if ((unsigned int)searchResult == words.size())
break;
if (words[searchResult].rfind(sequence) > words[searchResult].length())
break;
if (words[searchResult] == (sequence))
cout << sequence << endl;
}
}
}
/*
* Performs the verifyWord search method.
* Searches the word to make sure that so far, there is possibly that the current sequence
* of letter could form a word. That is to avoid continuing to search for a word
* when the first sequence of characters do not construct a valid word in the dictionary.
*
* For example, if we have 'xzt', when this search is done it prevents the search
* to continue since no word in the dictionary starts with 'xzt'
*/
int Scramble::verifyWord(vector<string> words, string str) {
int low = 0;
int mid = 0;
int high = words.size();
while (low < high) {
mid = (low + high) / 2;
if (str > words[mid]) {
low = mid + 1;
}
else if (str < words[mid]) {
high = mid - 1;
}
else
return mid;
}
}发布于 2011-05-01 06:51:27
1)目前,您的solve()函数在直线上查找从每个点开始的单词:这是您想要的吗?我之所以问这个问题,是因为'hello‘在你的样本矩阵中不是直线:
H E i l
x L p q
c L O m如果你确实只想要直线单词,那好(这就是我一直认为these puzzles的工作方式),但如果你实际上想以蛇形的方式查找单词,那么像Zilchonum和BlueRaja建议的递归搜索将是一个很好的选择。只需小心,不要在已经使用过的字母上循环返回。
2)无论哪种情况,您的verifyWord()函数都有一些问题:在您退出while (low < high)循环的情况下,它至少需要返回一些值。
即便如此,它仍然不能完全满足您的需要:例如,假设您的字典包含{"ant", "bat" "hello", "yak", "zoo"},您使用str="hel"调用verifyWord(),您希望返回值为2,但目前它是这样做的:
step low mid high
0 0 0 5 // initialise
1 0 2 5 // set mid = (0+5)/2 = 2... words[2] == "hello"
2 0 2 1 // "hel" < "hello" so set high = mid - 1
3 0 0 1 // set mid = (0+1)/2 = 0... words[0] == "ant"
4 1 0 1 // "hel" > "ant" so set low = mid + 1
5 // now (low<high) is false, so we exit the loop with mid==0与其将"hello“与”hello“进行比较,不如将字典中的单词截断为与str相同的长度:即将str与word[mid].substr(0,str.length())进行比较
发布于 2011-05-01 06:23:12
这里有一个有趣的思考方式:找到这个单词类似于解决一个迷宫。“start”和“end”对应于您正在查找的单词的开头和结尾,“死胡同”对应于路径和单词之间的不匹配,而“成功”是指路径上的字符串匹配。
这里的好消息是,有很多关于迷宫求解算法的资源。我熟悉的一种特殊算法是recursion with backtracking,它实现起来并不太困难。
显然,为了解决您的问题,必须进行一些更改。例如,您不知道起点在哪里,但幸运的是,这无关紧要。你可以检查每个可能的起始位置,其中许多在第一步就会因为不匹配而被丢弃。
发布于 2011-05-01 06:29:21
只需将其视为一个图,其中每个字母都连接到所有相邻的字母,并从每个字母开始进行深度/广度优先搜索,只接受其字母等于您要查找的下一个字母的节点。
https://stackoverflow.com/questions/5844861
复制相似问题