Leetcode 120. 三角形最小路径和

Leetcode 120. 三角形最小路径和

题目描述

给定一个三角形，找出自顶向下的最小路径和。每一步只能移动到下一行中相邻的结点上。
例如，给定三角形：
[
[2],
[3,4],
[6,5,7],
[4,1,8,3]
]
自顶向下的最小路径和为 11（即，2 + 3 + 5 + 1 = 11）。

解题思路

动态规划，用$dp[i][j]$表示从顶点到达$(i,j)$点的最短路径和。状态转移方程如下：
d p [ i ] [ j ] = { t r i a n g l e [ i ] [ j ]                                                                     i = 0 t r i a n g l e [ i ] [ j ] + d p [ i − 1 ] [ j ]                                             j = 0 t r i a n g l e [ i ] [ j ] + d p [ i − 1 ] [ j − 1 ]                                      j = t r i a n g l e [ i ] . l e n g t h − 1 t r i a n g l e [ i ] [ j ] + m i n { d p [ i − 1 ] [ j ] , d p [ i − 1 ] [ j − 1 ] }      t r i a n g l e [ i ] . l e n g t h − 1 > j > 0 dp[i][j]= \left \{ \begin{array}{lr} triangle[i][j]\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ i=0\\ triangle[i][j]+dp[i-1][j]\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ j=0\\ triangle[i][j]+dp[i-1][j-1]\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ j=triangle[i].length-1\\ triangle[i][j]+min\{dp[i-1][j],dp[i-1][j-1]\}\ \ \ \ triangle[i].length-1>j>0 \end{array} \right . dp[i][j]=⎩⎪⎪⎨⎪⎪⎧​triangle[i][j]                                                                   i=0triangle[i][j]+dp[i−1][j]                                           j=0triangle[i][j]+dp[i−1][j−1]                                    j=triangle[i].length−1triangle[i][j]+min{dp[i−1][j],dp[i−1][j−1]}    triangle[i].length−1>j>0​

可以看出来，更新第$i$行只需要用到第$i-1$行，所以可以进一步优化空间。
用$dp[j]$表示从顶点到达$(i,j)$点的最短路径和,优化之后的公式如下：
d p [ j ] = { t r i a n g l e [ i ] [ j ]                                               i = 0 t r i a n g l e [ i ] [ j ] + d p [ j ]                                  j = 0 t r i a n g l e [ i ] [ j ] + d p [ j − 1 ]                           j = t r i a n g l e [ i ] . l e n g t h − 1 t r i a n g l e [ i ] [ j ] + m i n { d p [ j ] , d p [ j − 1 ] }      t r i a n g l e [ i ] . l e n g t h − 1 > j > 0 dp[j]= \left \{ \begin{array}{lr} triangle[i][j]\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ i=0\\ triangle[i][j]+dp[j]\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ j=0\\ triangle[i][j]+dp[j-1]\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ j=triangle[i].length-1\\ triangle[i][j]+min\{dp[j],dp[j-1]\}\ \ \ \ triangle[i].length-1>j>0 \end{array} \right . dp[j]=⎩⎪⎪⎨⎪⎪⎧​triangle[i][j]                                             i=0triangle[i][j]+dp[j]                                j=0triangle[i][j]+dp[j−1]                         j=triangle[i].length−1triangle[i][j]+min{dp[j],dp[j−1]}    triangle[i].length−1>j>0​
其中， i = { 0 , 1 , 2 , . . . , n − 1 } i=\{0,1,2,...,n-1\} i={0,1,2,...,n−1}
j = { t r i a n g l e [ i ] . l e n g t h − 1 , t r i a n g l e [ i ] . l e n g t h − 2 , . . . , 0 } j=\{triangle[i].length-1, triangle[i].length-2,...,0\} j={triangle[i].length−1,triangle[i].length−2,...,0}
注意，$j$必须是降序，这是因为求$dp[j]$需要用到$dp[j-1]$。如果求$dp[j]$需要用到$dp[j+1]$，则$j$需要是升序。

代码实现

不带空间优化

class Solution {
public:
    int minimumTotal(vector<vector<int>>& triangle) {
        int dp[1050][1050] = {0};
        int n = triangle.size();
        if (n == 0) return 0;
        if (triangle[0].size() == 0) return 0;
        
        dp[0][0] = triangle[0][0];
        
        for (int i = 1; i < n; i++) {
            for (int j = 0; j < triangle[i].size(); j++) {
                if (j == 0) {
                    dp[i][j] = dp[i-1][j] + triangle[i][j];
                }
                else if (j == triangle[i].size()-1) {
                    dp[i][j] = dp[i-1][j-1] + triangle[i][j];
                }
                else {
                    dp[i][j] = triangle[i][j] + min(dp[i-1][j], dp[i-1][j-1]);
                }
            }
        }
        int ans = dp[n-1][0];
        for (int j = 1; j < triangle[n-1].size(); j++) {
            ans = min(ans, dp[n-1][j]);
        }
        return ans;
    }
};

空间优化

class Solution {
public:
    int minimumTotal(vector<vectot<int>>& triangle) {
        int dp[1050] = {0};
        int n = triangle.size();
        if (n == 0) return 0;
        if (triangle[0].size() == 0) return 0;
        
        dp[0] = triangle[0][0];
        
        for (int i = 1; i < n; i++) {
            for (int j = triangle[i].size()-1; j >= 0; j--) {
                if (j == 0) {
                    dp[j] = dp[j] + triangle[i][j];
                }
                else if (j == triangle[i].size()-1) {
                    dp[j] = dp[j-1] + triangle[i][j];
                }
                else {
                    dp[j] = triangle[i][j] + min(dp[j], dp[j-1]);
                }
            }
        }
        int ans = dp[0];
        for (int j = 1; j < triangle[n-1].size(); j++) {
            ans = min(ans, dp[j]);
        }
        return ans;
    }
};