第12课-动态规划

分治 + 回溯 + 递归 + 动态规划

递归代码模版

public void recur(int level, int param) {
     // terminator
     if (level > MAX_LEVEL) {
       // process result
	 return; 
	 }
     // process current logic
     process(level, param);
     // drill down
     recur( level: level + 1, newParam);
     // restore current status
}

分治 Divide & Conquer

递归状态树

分治代码模板

def divide_conquer(problem, param1, param2, ...):
  # recursion terminator
  if problem is None:
    print_result
return
  # prepare data
  data = prepare_data(problem)
  subproblems = split_problem(problem, data)
  # conquer subproblems
  subresult1 = self.divide_conquer(subproblems[0], p1, ...)
  subresult2 = self.divide_conquer(subproblems[1], p1, ...)
  subresult3 = self.divide_conquer(subproblems[2], p1, ...)
  ...
  # process and generate the final result
  result = process_result(subresult1, subresult2, subresult3, ...)
  # revert the current level states

感触

1. 人肉递归低效、很累
2. 找到最近最简方法，将其拆解成可重复解决的问题
3. 数学归纳法思维(抵制人肉递归的诱惑)

本质:寻找重复性 —> 计算机指令集

动态规划 Dynamic Programming

1. Wiki 定义: https://en.wikipedia.org/wiki/Dynamic_programming
2. “Simplifying a complicated problem by breaking it down into simpler sub-problems”
   (in a recursive manner)
3. Divide & Conquer + Optimal substructure 分治 + 最优子结构

关键点:

1. 动态规划 和 递归或者分治 没有根本上的区别(关键看有无最优的子结构)
2. 共性:找到重复子问题
3. 差异性:最优子结构、中途可以淘汰次优解

实战例题一 斐波拉契数列

fib (n) = fib (n-1) + fib (n-2) 
fib (0) = 0
fib (1) = 1

int fib (int n) {
  if (n <= 0) {
    return 0;
  } else if (n == 1) {
    return 1;
  } else {
    return fib (n - 1) + fib (n - 2);
  }
}

F[n] = F[n-1] + F[n-2]

a[0] = 0, a[1] = 1;
for (int i = 2; i <= n; ++i) {
	a[i] = a[i-1] + a[i-2]; 
}

---

实战例题二 路径计数

状态转移方程(DP 方程)

opt[i , j] = opt[i + 1, j] + opt[i, j + 1] 
完整逻辑:
if a[i, j] = ‘空地’:
	opt[i , j] = opt[i + 1, j] + opt[i, j + 1] 
else:
	opt[i , j] = 0

动态规划关键点

1. 最优子结构 opt[n] = best_of(opt[n-1], opt[n-2], …)
2. 储存中间状态:opt[i]
3. 递推公式(美其名曰:状态转移方程或者 DP 方程)
   Fib: opt[i] = opt[n-1] + opt[n-2]
   二维路径:opt[i,j] = opt[i+1][j] + opt[i][j+1] (且判断a[i,j]是否空地)

实战例题三 最长公共子序列

https://leetcode-cn.com/problems/longest-common-subsequence/

给定两个字符串 text1 和 text2，返回这两个字符串的最长公共子序列。 “ABAZDC” , “BACBAD”

字符串问题

1. S1 = “”
   S2 = 任意字符串
2. S1 =“A” S2 = 任意
3. S1 =“…A” S2 = “…A”

子问题

• S1 = “ABAZDC” S2 = “BACBAD”
• If S1[-1] != S2[-1]: LCS[s1, s2] = Max(LCS[s1-1, s2], LCS[s1, s2-1]) 
LCS[s1, s2] = Max(LCS[s1-1, s2], LCS[s1, s2-1], LCS[s1-1, s2-1])
• If S1[-1] == S2[-1]: LCS[s1, s2] = LCS[s1-1, s2-1] + 1
LCS[s1, s2] = Max(LCS[s1-1, s2], LCS[s1, s2-1], LCS[s1-1, s2-1], LCS[s1-1][s2-1] + 1)

DP 方程

• If S1[-1] != S2[-1]: LCS[s1, s2] = Max(LCS[s1-1, s2], LCS[s1, s2-1]) 
• If S1[-1] == S2[-1]: LCS[s1, s2] = LCS[s1-1, s2-1] + 1

动态规划小结

1. 打破自己的思维惯性，形成机器思维
2. 理解复杂逻辑的关键
3. 也是职业进阶的要点要领

MIT algorithm course

B 站搜索: mit 动态规划

https://www.bilibili.com/video/av53233912? from=search&seid=2847395688604491997

实战题目

1. https://leetcode-cn.com/problems/climbing-stairs/description/
2. https://leetcode-cn.com/problems/triangle/description/
   (https://leetcode.com/problems/triangle/discuss/38735/Python-easy-to-understand-solutions-(top-down-bottom-up).)
3. https://leetcode-cn.com/problems/maximum-subarray/
   https://leetcode-cn.com/problems/maximum-product-subarray/description/
4. https://leetcode-cn.com/problems/coin-change/description/

实战题目

1. https://leetcode-cn.com/problems/house-robber/
2. https://leetcode-cn.com/problems/house-robber-ii/description/
3. https://leetcode-cn.com/problems/best-time-to-buy-and-sell-stock/
4. https://leetcode-cn.com/problems/best-time-to-buy-and-sell-stock-ii/
5. https://leetcode-cn.com/problems/best-time-to-buy-and-sell-stock-iii/
6. https://leetcode-cn.com/problems/best-time-to-buy-and-sell-stock-with-cooldown/
7. https://leetcode-cn.com/problems/best-time-to-buy-and-sell-stock-iv/
8. https://leetcode-cn.com/problems/best-time-to-buy-and-sell-stock-with-transaction- fee/
9. https://leetcode-cn.com/problems/best-time-to-buy-and-sell-stock/solution/yi-ge- fang-fa-tuan-mie-6-dao-gu-piao-wen-ti-by-l-3/

实战题目

1. https://leetcode-cn.com/problems/perfect-squares/
2. https://leetcode-cn.com/problems/edit-distance/ (重点)
3. https://leetcode-cn.com/problems/jump-game/
4. https://leetcode-cn.com/problems/jump-game-ii/
5. https://leetcode-cn.com/problems/unique-paths/
6. https://leetcode-cn.com/problems/unique-paths-ii/
7. https://leetcode-cn.com/problems/unique-paths-iii/
8. https://leetcode-cn.com/problems/coin-change/
9. https://leetcode-cn.com/problems/coin-change-2/

Homework

1.https://leetcode-cn.com/problems/longest-valid-parentheses/
2.https://leetcode-cn.com/problems/minimum-path-sum/
3.https://leetcode-cn.com/problems/edit-distance/
4.https://leetcode-cn.com/problems/decode-ways
5.https://leetcode-cn.com/problems/maximal-square/
6.https://leetcode-cn.com/problems/max-sum-of-rectangle-no-larger-than-k/
7.https://leetcode-cn.com/problems/frog-jump/
8.https://leetcode-cn.com/problems/split-array-largest-sum
9.https://leetcode-cn.com/problems/student-attendance-record-ii/
10.https://leetcode-cn.com/problems/task-scheduler/
11.https://leetcode-cn.com/problems/palindromic-substrings/
12. https://leetcode-cn.com/problems/minimum-window-substring/
13. https://leetcode-cn.com/problems/burst-balloons/