72. Edit Distance

Problem

Given two words word1 and word2, find the minimum number of steps required to convert word1 to word2. (each operation is counted as 1 step.)

You have the following 3 operations permitted on a word:

a) Insert a character
b) Delete a character
c) Replace a character

Solution

本来打算用LCS（最长公共子序列－－可以不连续，不是最长公共子串－－连续，LCS我看的他的）来解决来着，写完才发现，程序最后只有最长相同的序列，对齐即可了，那么剩余的元素，有可能元素对空（也就是添加元素）情况：

比如：

所有，最后还是换种方式来解决：

class Solution {
public:
    int minDistance(string word1, string word2) {
        int n1 = word1.size();
        int n2 = word2.size();

        if (n1 == 0 || n2 == 0) {
            return max(n1, n2);
        }
        // vector<vector<int>> e(n1, vector<int>(n2)), diff(n1, vector<int>(n2));
        vector<vector<int>> e(n1+1, vector<int>(n2+1));

        for (int i = 0; i <= n1; ++i) {
            e[i][0] = i;
        }
        for (int j = 1; j <= n2; ++j) {
            e[0][j] = j;
        }

        for (int i = 1; i <= n1; ++i) {
            for (int j = 1; j <= n2; ++j) {
                e[i][j] = min(min(e[i-1][j]+1, e[i][j-1]+1), e[i-1][j-1] + (word1[i-1] == word2[j-1] ? 0 : 1));
                // if (word1[i-1] == word2[j-1]) {
                    // diff[i][j] == 0;
                    // e[i][j] = min(min(e[i-1][j]+1, e[i][j-1]+1), e[i-1][j-1]);
                // }
                // else {
                    // diff[i][j] == 1;
                    // e[i][j] = min(min(e[i-1][j]+1, e[i][j-1]+1), e[i-1][j-1]+1);
                // }
            }
        }

        return e[n1][n2];
    }
};

这里，我做了优化，不用保存diff[i][j]

思路

动态规划，Edit distance问题

通过学习berkeley dynamic programming章节，然后做了leetcode这个题目，整理如下：

考虑三种情况：

1. x[1...i-1] <--> y[1...j] --> e[i-1][j] --> e[i][j] = e[i-1][j]+1?
2. x[1...i] <--> y[1...j-1] --> e[i][j-1] --> e[i][j] = e[i][j-1]+1
3. x[1....i-1] <--> y[1...j-1] --> e[i-1][j-1] + diff[i][j] --> e[i][j] = e[i-1][j-1] + diff[i][j]?

diff[i][j]:

• when x[i] == y[j], diff = 0;
• when x[i] != y[j], diff = 1;

E(i,j) = min{1+E(i−1,j), 1+E(i,j−1), diff(i,j)+E(i−1,j−1)}

结果如下：

算法：

for i = 0,1,2,...,m: E(i,0) = i
for j = 1,2,...,n: E(0,j) = j
for i = 1,2,...,m: for j = 1,2,...,n:
E(i, j) = min{E(i − 1, j) + 1, E(i, j − 1) + 1, E(i − 1, j − 1) + diff(i, j)} return E(m, n)

算法生成矩阵：

最后选择的元素（DAG）

之后，等把动态规划常见问题整理的差不多了，就做一个动态规划（DP）的总结。