Files
algorithm-homework/AlgorithmHomework6.md
T
2020-04-16 18:28:04 +08:00

3.8 KiB
Executable File

Algorithm Homework 6

姓名:阙建明

学号:1901213139

作业要求:

Leetcode: 5, 64, 120

题解:

  • Leetcode problem 5

    最长回文子串

    • 算法思路:

      • 使用动态规划的思想;

      • 状态转移方程如下:

        P(i, j) = \begin{cases}True &\text{if } P(i + 1, j - 1)==True , S_i == S_j \\False &\text{if 其它 } \end{cases}

      • 首先判断长度为1的子串,再判断长度为2的子串,再判断长度为3的子串,以此类推。

    • Leetcode提交结果:

      image-20200407102849523

    • 代码:

      class Solution:
          def longestPalindrome(self, s: str) -> str:
              arr = [[False] * len(s) for i in range(len(s))]
              for i in range(len(s)):
                  arr[i][i] = True
              maxLen, maxStrStart, maxStrEnd = 1, 0, 1
              for span in range(1, len(s)):
                  left, right = 0, span
                  while right < len(s):
                      arr[left][right] = True if s[left] == s[right] and (span == 1 or arr[left + 1][right - 1]) else False
                      if arr[left][right] and right - left + 1 > maxLen:
                          maxLen, maxStrStart, maxStrEnd = right - left + 1, left, right + 1
                      left, right = left + 1, right + 1
              return s[maxStrStart: maxStrEnd]
      
  • Leetcode problem 64

    最小路径和

    • 算法思路:

      • 采用动态规划的思想,自右下角向左上脚遍历。

      • 首先将最下面一排的路径开销自右往左以此叠加,把最右边一排的路径开销自下往上以此叠加,接着其它情况使用与下述状态转换方程:

        P(i, j) = \begin{cases}P(i, j) + P(i + 1, j) &\text{if } P(i, j + 1) >= P(j + 1) \\P(i, j) + P(i, j + 1) &\text{if }P(i, j + 1) < P(j + 1) \end{cases}

    • Leetcode提交结果:

      image-20200407105454180

    • 代码:

      class Solution:
          def minPathSum(self, grid: List[List[int]]) -> int:
              rowNum, colNum = len(grid), len(grid[0])
              for col in range(colNum - 2, -1, -1):
                  grid[rowNum - 1][col] = grid[rowNum - 1][col] + grid[rowNum - 1][col + 1]
              for row in range(rowNum - 2, -1, -1):
                  grid[row][colNum - 1] = grid[row][colNum - 1] + grid[row + 1][colNum - 1]
              for col in range(colNum - 2, -1, -1):
                  for row in range(rowNum - 2, -1, -1):
                      grid[row][col] = grid[row][col] + min(grid[row + 1][col], grid[row][col + 1])
              return grid[0][0]
      
  • Leetcode problem 120

    三角形最小路径和

    • 算法思路:

      • 用动态规划的思想,自底向上一次计算到当前访问节点到叶子节点的最短路径;

      • 状态转移方程如下:

        P(row, index) = \begin{cases}P(row, index) &\text{if } row == len(triangle) - 1 \\P(row, index) + min(P(row + 1, index), P(row + 1, index + 1)) &\text{if }row < len(triangle) - 1 \end{cases}

    • Leetcode提交结果:

      image-20200407122505839

    • 代码:

      class Solution:
        def minimumTotal(self, triangle: List[List[int]]) -> int:
              for row in range(len(triangle) - 2, -1, -1):
                  for index in range(len(triangle[row])):
                      triangle[row][index] = triangle[row][index] + min(triangle[row + 1][index], triangle[row + 1][index + 1])
              return triangle[0][0]