63. Unique Paths II（不同路径 II）两种解法（C++ & 注释）

1. 题目描述

一个机器人位于一个 m x n 网格的左上角 （起始点在下图中标记为“Start” ）。

机器人每次只能向下或者向右移动一步。机器人试图达到网格的右下角（在下图中标记为“Finish”）。

现在考虑网格中有障碍物。那么从左上角到右下角将会有多少条不同的路径？

网格中的障碍物和空位置分别用 1 和 0 来表示。

说明：m 和 n 的值均不超过 100。

示例 1:

输入: [ [0,0,0], [0,1,0], [0,0,0] ]
输出: 2
解释: 3x3 网格的正中间有一个障碍物。
从左上角到右下角一共有 2 条不同的路径：

1. 向右 -> 向右 -> 向下 -> 向下
2. 向下 -> 向下 -> 向右 -> 向右

题目链接：中文题目；英文题目

2. 回溯法（Backtracking, Time Limit Exceeded）

2.1 解题思路

这题和62. Unique Paths（不同路径）的回溯法思路基本相似，只是这题会遇到障碍物，所以当遇到障碍物，我们则跳过当前坐标，返回上一步继续向右或向下寻找抵达终点的路径，不过这方法超时预订。

具体的讲解可以参考：62. Unique Paths（不同路径）

2.2 实例代码

class Solution {
    int count = 0;

    void move(int i, int j, int iFinish, int jFinish, vector<vector<int>>& obstacleGrid) {
        if (i >= iFinish + 1 || j >= jFinish + 1 || obstacleGrid[i][j]) return;
        if (i == iFinish && j == jFinish) { count++; return; }
        move(i, j + 1, iFinish, jFinish, obstacleGrid);
        move(i + 1, j, iFinish, jFinish, obstacleGrid);
    }
public:
    int uniquePathsWithObstacles(vector<vector<int>>& obstacleGrid) {
        if (!obstacleGrid.size()) return this->count;
        move(0, 0, obstacleGrid.size() - 1, obstacleGrid[0].size() - 1, obstacleGrid);
        return this->count;
    }
};

3. 动态规划（Dynamic Programming）

3.1 解题思路

同样，对回溯法进行空间的优化思路也是和上题是一致的，遇到障碍物返回0即可。这里可以创建一个额外的二维数组存储结果，也可以直接修改题目给的obstacleGrid二维数组。

如果使用第二种方法，需要把obstacleGrid预处理一下，把里面表示障碍物的1，全部变成-1，如果还是保留之前的1，后面遇到重复情况时，会把障碍物误认为有一个路径数的坐标。

最后就是用迭代实现了，迭代的思路还是和上一题基本类似，只有两点不同：

1. 处理右边界和下边界，或上边界和左边界时，需要先预处理终点（[m - 1][n - 1]）或起点（[0][0]）是否为1,，然后操作边界时，当前坐标的值需要保证不为1（障碍物），而且前一个坐标值不能0，只有满足这两个条件，这个坐标值才能等于1，也就是保证了这条路径是通向终点的；
2. 处理中间部分的坐标时，如果原值为1（障碍物）时，直接赋值0；反之，按照前一题的公式进行计算即可：dp[i][j] = dp[i + 1][j] + dp[i][j + 1] 或 dp[i][j] = dp[i - 1][j] + dp[i][j - 1]

具体的讲解可以参考：62. Unique Paths（不同路径）

番外：3.2.3 迭代（Iteration） - 从上到下，这个方法的代码不知道为什么线上有一个测试案例无法通过，显示int类型越界，但是笔者本地测试可以得到正确答案，也不知道为什么…测试案例详情，我已经写在这部分代码里面了，如果有童鞋发现了原因，麻烦告知一下啦，Thanks♪(･ω･)ﾉ

3.2 实例代码

3.2.1 递归（Recursion） - 需要额外空间

class Solution {
    int move(int i, int j, int iFinish, int jFinish, vector<vector<int>>& obstacleGrid, vector<vector<int>>& memory) {
        if (i >= iFinish + 1 || j >= jFinish + 1 || obstacleGrid[i][j]) return 0;
        if (memory[i][j]) return memory[i][j];
        if (i == iFinish && j == jFinish) return 1;
        memory[i][j] = move(i, j + 1, iFinish, jFinish, obstacleGrid, memory) + move(i + 1, j, iFinish, jFinish, obstacleGrid, memory);
        return memory[i][j];
    }
public:
    int uniquePathsWithObstacles(vector<vector<int>>& obstacleGrid) {
        if (!obstacleGrid.size()) return 0;
        int m = obstacleGrid.size(), n = obstacleGrid[0].size();
        vector<vector<int>> memory(m, vector<int>(n, 0));
        memory[0][0] = move(0, 0, m - 1, n - 1, obstacleGrid, memory);
        return memory[0][0];
    }
};

3.2.2 递归（Recursion） - 不需要额外空间

class Solution {
    int move(int i, int j, int iFinish, int jFinish, vector<vector<int>>& obstacleGrid) {
        if (i >= iFinish + 1 || j >= jFinish + 1 || obstacleGrid[i][j] == -1) return 0;
        if (obstacleGrid[i][j]) return obstacleGrid[i][j];
        if (i == iFinish && j == jFinish) return 1;
        obstacleGrid[i][j] = move(i, j + 1, iFinish, jFinish, obstacleGrid) + move(i + 1, j, iFinish, jFinish, obstacleGrid);
        return obstacleGrid[i][j];
    }
public:
    int uniquePathsWithObstacles(vector<vector<int>>& obstacleGrid) {
        if (!obstacleGrid.size()) return 0;
        int m = obstacleGrid.size(), n = obstacleGrid[0].size();

		// 预处理obstacleGrid
        for (int i = 0; i < m; i++)
            for (int j = 0; j < n; j++)
                if (obstacleGrid[i][j]) obstacleGrid[i][j] = -1;

        obstacleGrid[0][0] = move(0, 0, m - 1, n - 1, obstacleGrid);
        return obstacleGrid[0][0];
    }
};

3.2.3 迭代（Iteration） - dp[i][j] = dp[i + 1][j] + dp[i][j + 1]

class Solution {
public:
    int uniquePathsWithObstacles(vector<vector<int>>& obstacleGrid) {
        if (!obstacleGrid.size()) return 0;
        int m = obstacleGrid.size(), n = obstacleGrid[0].size();
        if (obstacleGrid[m - 1][n - 1]) return 0;
        obstacleGrid[m - 1][n - 1] = 1;

        for (int i = m - 2; i >= 0; i--) 
        	obstacleGrid[i][n - 1] = (!obstacleGrid[i][n - 1] && obstacleGrid[i + 1][n - 1]) ? 1 : 0; // 右边界
        for (int i = n - 2; i >= 0; i--) 
        	obstacleGrid[m - 1][i] = (!obstacleGrid[m - 1][i] && obstacleGrid[m - 1][i + 1]) ? 1 : 0; // 下边界

        for (int i = m - 2; i >= 0; i--)
            for (int j = n - 2; j >= 0; j--)
                obstacleGrid[i][j] = obstacleGrid[i][j] ? 0 : obstacleGrid[i + 1][j] + obstacleGrid[i][j + 1];

        return obstacleGrid[0][0];
    }
};

int main()
{
    Solution s;
    // 线上无法通过测试的案例，
    vector<vector<int>> nums = {{0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,1,0,1,1,0,0,1,0,1,1,0,1,0,0,1,0,0,0,1,0,0},{0,0,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,1,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0},{1,0,0,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,1,0,1,0,1,0,0,1,0,0,0},{1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,0,0,1,0,0,0,0,0,0,1,0,0,1,0,1},{0,0,0,1,0,0,0,0,0,0,0,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0},{0,0,1,0,0,0,1,1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,1,1,0,0,1,0,0,1,0,0},{0,0,0,0,0,0,1,0,0,0,1,1,0,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0},{1,1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,1,1,0,0,0,1},{0,0,0,0,1,0,0,1,0,1,1,1,0,0,0,1,0,0,1,0,1,0,0,0,0,1,1,1,0,0,0,0,0,0,0,0},{0,0,0,1,0,0,0,0,0,0,0,0,0,1,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,1,1,0,0,0},{1,0,1,0,1,1,0,1,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,1,1,1,0,0,0,0,1,0,0,0,1,0},{0,0,0,0,0,0,1,0,0,1,1,0,0,1,0,0,0,0,1,0,0,1,1,0,0,0,0,0,1,0,0,1,0,0,0,1},{0,0,0,0,1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,0,1,0,0,1,0,0,0},{1,0,0,0,0,0,1,0,0,0,0,0,0,1,0,0,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0},{0,0,1,0,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0},{0,1,0,0,1,0,0,0,0,0,1,0,1,1,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,1,0,0,0,0,0,0},{0,0,0,0,1,0,0,0,1,0,0,0,0,1,0,0,1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0},{0,0,0,0,0,1,0,0,1,1,0,0,0,0,0,0,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,1,0},{0,0,1,0,1,0,1,0,0,0,0,0,1,0,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0},{0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,1,1,0,0,0,1,0,0,0,1,0,0,0,1,0,0,0},{0,0,0,1,0,1,0,0,1,0,0,0,0,0,1,1,1,0,1,1,1,0,0,1,0,1,0,1,1,1,0,0,0,0,0,0},{0,0,1,0,0,0,0,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0},{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0},{0,0,0,0,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0},{1,1,0,0,0,0,1,0,0,1,1,0,0,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0},{0,1,0,0,0,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,0,1,0,0,0,0,0,0,1,0,0,1,0,0},{0,0,0,0,1,0,0,1,0,0,0,0,0,0,1,0,0,1,0,1,1,1,0,0,0,0,0,0,1,0,0,0,1,1,0,0},{0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0} };
    cout << s.uniquePathsWithObstacles(nums);
 }

3.2.4 迭代（Iteration） - dp[i][j] = dp[i - 1][j] + dp[i][j - 1]

class Solution {
public:
    int uniquePathsWithObstacles(vector<vector<int>>& obstacleGrid) {
        if (!obstacleGrid.size()) return 0;
        int m = obstacleGrid.size(), n = obstacleGrid[0].size();
        if (obstacleGrid[0][0]) return 0;
        obstacleGrid[0][0] = 1;

        for (int i = 1; i < m; i++) obstacleGrid[i][0] = (!obstacleGrid[i][0] && obstacleGrid[i - 1][0]) ? 1 : 0; // 右边界
        for (int i = 1; i < n; i++) obstacleGrid[0][i] = (!obstacleGrid[0][i] && obstacleGrid[0][i - 1]) ? 1 : 0; // 上边界

        for (int i = 1; i < m; i++)
            for (int j = 1; j < n; j++)
                obstacleGrid[i][j] = obstacleGrid[i][j] ? 0 : obstacleGrid[i - 1][j] + obstacleGrid[i][j - 1];

        return obstacleGrid[m - 1][n - 1];
    }
};

---