[LeetCode] Minimum Path Sum-阿里云开发者社区

[LeetCode] Minimum Path Sum

2015-06-02 848

版权

本文内容由阿里云实名注册用户自发贡献，版权归原作者所有，阿里云开发者社区不拥有其著作权，亦不承担相应法律责任。具体规则请查看《阿里云开发者社区用户服务协议》和《阿里云开发者社区知识产权保护指引》。如果您发现本社区中有涉嫌抄袭的内容，填写侵权投诉表单进行举报，一经查实，本社区将立刻删除涉嫌侵权内容。

简介： This is a typical DP problem. Suppose the minimum path sum of arriving at point (i, j) is S[i][j], then the state equation is S[i][j] = min(S[i - 1][j], S[i][j - 1]) + grid[i][j].

This is a typical DP problem. Suppose the minimum path sum of arriving at point (i, j) is S[i][j], then the state equation is S[i][j] = min(S[i - 1][j], S[i][j - 1]) + grid[i][j].

Well, some boundary conditions need to be handled. The boundary conditions happen on the topmost row (S[i - 1][j] does not exist) and the leftmost column (S[i][j - 1] does not exist). Suppose grid is like [1, 1, 1, 1], then the minimum sum to arrive at each point is simply an accumulation of previous points and the result is [1, 2, 3, 4].

Now we can write down the following (unoptimized) code.

 1 class Solution {
 2 public:
 3     int minPathSum(vector<vector<int>>& grid) {
 4         int m = grid.size();
 5         int n = grid[0].size(); 
 6         vector<vector<int> > sum(m, vector<int>(n, grid[0][0]));
 7         for (int i = 1; i < m; i++)
 8             sum[i][0] = sum[i - 1][0] + grid[i][0];
 9         for (int j = 1; j < n; j++)
10             sum[0][j] = sum[0][j - 1] + grid[0][j];
11         for (int i = 1; i < m; i++)
12             for (int j = 1; j < n; j++)
13                 sum[i][j]  = min(sum[i - 1][j], sum[i][j - 1]) + grid[i][j];
14         return sum[m - 1][n - 1];
15     }
16 };

As can be seen, each time when we update sum[i][j], we only need sum[i - 1][j] (at the current column) and sum[i][j - 1] (at the left column). So we need not maintain the full m*nmatrix. Maintaining two columns is enough and now we have the following code.

 1 class Solution {
 2 public:
 3     int minPathSum(vector<vector<int>>& grid) {
 4         int m = grid.size();
 5         int n = grid[0].size();
 6         vector<int> pre(m, grid[0][0]);
 7         vector<int> cur(m, 0);
 8         for (int i = 1; i < m; i++)
 9             pre[i] = pre[i - 1] + grid[i][0];
10         for (int j = 1; j < n; j++) { 
11             cur[0] = pre[0] + grid[0][j]; 
12             for (int i = 1; i < m; i++)
13                 cur[i] = min(cur[i - 1], pre[i]) + grid[i][j];
14             swap(pre, cur); 
15         }
16         return pre[m - 1];
17     }
18 };

Further inspecting the above code, it can be seen that maintaining pre is for recovering pre[i], which is simply cur[i] before its update. So it is enough to use only one vector. Now the space is further optimized and the code also gets shorter.

 1 class Solution {
 2 public:
 3     int minPathSum(vector<vector<int>>& grid) {
 4         int m = grid.size();
 5         int n = grid[0].size();
 6         vector<int> cur(m, grid[0][0]);
 7         for (int i = 1; i < m; i++)
 8             cur[i] = cur[i - 1] + grid[i][0]; 
 9         for (int j = 1; j < n; j++) {
10             cur[0] += grid[0][j]; 
11             for (int i = 1; i < m; i++)
12                 cur[i] = min(cur[i - 1], cur[i]) + grid[i][j];
13         }
14         return cur[m - 1];
15     }
16 };