|
|
|
|
---
|
|
|
|
|
comments: true
|
|
|
|
|
---
|
|
|
|
|
|
|
|
|
|
# 13.1. 初探动态规划
|
|
|
|
|
|
|
|
|
|
动态规划(Dynamic Programming)是一种用于解决复杂问题的优化算法,它把一个问题分解为一系列更小的子问题,并把子问题的解存储起来以供后续使用,从而避免了重复计算,提升了解题效率。
|
|
|
|
|
|
|
|
|
|
在本节中,我们先从一个动态规划经典例题入手,学习动态规划是如何高效地求解问题的,包括:
|
|
|
|
|
|
|
|
|
|
1. 如何暴力求解动态规划问题,什么是重叠子问题。
|
|
|
|
|
2. 如何向暴力搜索引入记忆化处理,从而优化时间复杂度。
|
|
|
|
|
3. 从递归解法引出动态规划解法,以及如何优化空间复杂度。
|
|
|
|
|
|
|
|
|
|
!!! question "爬楼梯"
|
|
|
|
|
|
|
|
|
|
给定一个共有 $n$ 阶的楼梯,你每步可以上 $1$ 阶或者 $2$ 阶,请问有多少种方案可以爬到楼顶。
|
|
|
|
|
|
|
|
|
|
如下图所示,对于一个 $3$ 阶楼梯,共有 $3$ 种方案可以爬到楼顶。
|
|
|
|
|
|
|
|
|
|
![爬到第 3 阶的方案数量](intro_to_dynamic_programming.assets/climbing_stairs_example.png)
|
|
|
|
|
|
|
|
|
|
<p align="center"> Fig. 爬到第 3 阶的方案数量 </p>
|
|
|
|
|
|
|
|
|
|
**不考虑效率的前提下,动态规划问题理论上都可以使用回溯算法解决**,因为回溯算法本质上就是穷举,它能够遍历决策树的所有可能的状态,并从中记录需要的解。
|
|
|
|
|
|
|
|
|
|
对于本题,我们可以将爬楼梯想象为一个多轮选择的过程:从地面出发,每轮选择上 $1$ 阶或 $2$ 阶,每当到达楼梯顶部时就将方案数量加 $1$ 。
|
|
|
|
|
|
|
|
|
|
=== "Java"
|
|
|
|
|
|
|
|
|
|
```java title="climbing_stairs_backtrack.java"
|
|
|
|
|
/* 回溯 */
|
|
|
|
|
void backtrack(List<Integer> choices, int state, int n, List<Integer> res) {
|
|
|
|
|
// 当爬到第 n 阶时,方案数量加 1
|
|
|
|
|
if (state == n)
|
|
|
|
|
res.set(0, res.get(0) + 1);
|
|
|
|
|
// 遍历所有选择
|
|
|
|
|
for (Integer choice : choices) {
|
|
|
|
|
// 剪枝:不允许越过第 n 阶
|
|
|
|
|
if (state + choice > n)
|
|
|
|
|
break;
|
|
|
|
|
// 尝试:做出选择,更新状态
|
|
|
|
|
backtrack(choices, state + choice, n, res);
|
|
|
|
|
// 回退
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* 爬楼梯:回溯 */
|
|
|
|
|
int climbingStairsBacktrack(int n) {
|
|
|
|
|
List<Integer> choices = Arrays.asList(1, 2); // 可选择向上爬 1 或 2 阶
|
|
|
|
|
int state = 0; // 从第 0 阶开始爬
|
|
|
|
|
List<Integer> res = new ArrayList<>();
|
|
|
|
|
res.add(0); // 使用 res[0] 记录方案数量
|
|
|
|
|
backtrack(choices, state, n, res);
|
|
|
|
|
return res.get(0);
|
|
|
|
|
}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C++"
|
|
|
|
|
|
|
|
|
|
```cpp title="climbing_stairs_backtrack.cpp"
|
|
|
|
|
/* 回溯 */
|
|
|
|
|
void backtrack(vector<int> &choices, int state, int n, vector<int> &res) {
|
|
|
|
|
// 当爬到第 n 阶时,方案数量加 1
|
|
|
|
|
if (state == n)
|
|
|
|
|
res[0]++;
|
|
|
|
|
// 遍历所有选择
|
|
|
|
|
for (auto &choice : choices) {
|
|
|
|
|
// 剪枝:不允许越过第 n 阶
|
|
|
|
|
if (state + choice > n)
|
|
|
|
|
break;
|
|
|
|
|
// 尝试:做出选择,更新状态
|
|
|
|
|
backtrack(choices, state + choice, n, res);
|
|
|
|
|
// 回退
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* 爬楼梯:回溯 */
|
|
|
|
|
int climbingStairsBacktrack(int n) {
|
|
|
|
|
vector<int> choices = {1, 2}; // 可选择向上爬 1 或 2 阶
|
|
|
|
|
int state = 0; // 从第 0 阶开始爬
|
|
|
|
|
vector<int> res = {0}; // 使用 res[0] 记录方案数量
|
|
|
|
|
backtrack(choices, state, n, res);
|
|
|
|
|
return res[0];
|
|
|
|
|
}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Python"
|
|
|
|
|
|
|
|
|
|
```python title="climbing_stairs_backtrack.py"
|
|
|
|
|
def backtrack(choices: list[int], state: int, n: int, res: list[int]) -> int:
|
|
|
|
|
"""回溯"""
|
|
|
|
|
# 当爬到第 n 阶时,方案数量加 1
|
|
|
|
|
if state == n:
|
|
|
|
|
res[0] += 1
|
|
|
|
|
# 遍历所有选择
|
|
|
|
|
for choice in choices:
|
|
|
|
|
# 剪枝:不允许越过第 n 阶
|
|
|
|
|
if state + choice > n:
|
|
|
|
|
break
|
|
|
|
|
# 尝试:做出选择,更新状态
|
|
|
|
|
backtrack(choices, state + choice, n, res)
|
|
|
|
|
# 回退
|
|
|
|
|
|
|
|
|
|
def climbing_stairs_backtrack(n: int) -> int:
|
|
|
|
|
"""爬楼梯:回溯"""
|
|
|
|
|
choices = [1, 2] # 可选择向上爬 1 或 2 阶
|
|
|
|
|
state = 0 # 从第 0 阶开始爬
|
|
|
|
|
res = [0] # 使用 res[0] 记录方案数量
|
|
|
|
|
backtrack(choices, state, n, res)
|
|
|
|
|
return res[0]
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Go"
|
|
|
|
|
|
|
|
|
|
```go title="climbing_stairs_backtrack.go"
|
|
|
|
|
[class]{}-[func]{backtrack}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsBacktrack}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "JavaScript"
|
|
|
|
|
|
|
|
|
|
```javascript title="climbing_stairs_backtrack.js"
|
|
|
|
|
[class]{}-[func]{backtrack}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsBacktrack}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "TypeScript"
|
|
|
|
|
|
|
|
|
|
```typescript title="climbing_stairs_backtrack.ts"
|
|
|
|
|
[class]{}-[func]{backtrack}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsBacktrack}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C"
|
|
|
|
|
|
|
|
|
|
```c title="climbing_stairs_backtrack.c"
|
|
|
|
|
[class]{}-[func]{backtrack}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsBacktrack}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C#"
|
|
|
|
|
|
|
|
|
|
```csharp title="climbing_stairs_backtrack.cs"
|
|
|
|
|
[class]{climbing_stairs_backtrack}-[func]{backtrack}
|
|
|
|
|
|
|
|
|
|
[class]{climbing_stairs_backtrack}-[func]{climbingStairsBacktrack}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Swift"
|
|
|
|
|
|
|
|
|
|
```swift title="climbing_stairs_backtrack.swift"
|
|
|
|
|
[class]{}-[func]{backtrack}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsBacktrack}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Zig"
|
|
|
|
|
|
|
|
|
|
```zig title="climbing_stairs_backtrack.zig"
|
|
|
|
|
[class]{}-[func]{backtrack}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsBacktrack}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Dart"
|
|
|
|
|
|
|
|
|
|
```dart title="climbing_stairs_backtrack.dart"
|
|
|
|
|
[class]{}-[func]{backtrack}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsBacktrack}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
## 13.1.1. 方法一:暴力搜索
|
|
|
|
|
|
|
|
|
|
然而,爬楼梯并不是典型的回溯问题,更适合从分治的角度进行解析。在分治算法中,原问题被分解为较小的子问题,通过组合子问题的解得到原问题的解。例如,归并排序将一个长数组从顶至底地划分为两个短数组,再从底至顶地将已排序的短数组进行排序。
|
|
|
|
|
|
|
|
|
|
对于本题,设爬到第 $i$ 阶共有 $dp[i]$ 种方案,那么 $dp[i]$ 就是原问题,其子问题包括:
|
|
|
|
|
|
|
|
|
|
$$
|
|
|
|
|
dp[i-1] , dp[i-2] , \cdots , dp[2] , dp[1]
|
|
|
|
|
$$
|
|
|
|
|
|
|
|
|
|
由于每轮只能上 $1$ 阶或 $2$ 阶,因此当我们站在第 $i$ 阶楼梯上时,上一轮只可能站在第 $i - 1$ 阶或第 $i - 2$ 阶上,换句话说,我们只能从第 $i -1$ 阶或第 $i - 2$ 阶前往第 $i$ 阶。因此,**爬到第 $i - 1$ 阶的方案数加上爬到第 $i - 2$ 阶的方案数就等于爬到第 $i$ 阶的方案数**,即:
|
|
|
|
|
|
|
|
|
|
$$
|
|
|
|
|
dp[i] = dp[i-1] + dp[i-2]
|
|
|
|
|
$$
|
|
|
|
|
|
|
|
|
|
![方案数量递推公式](intro_to_dynamic_programming.assets/climbing_stairs_state_transfer.png)
|
|
|
|
|
|
|
|
|
|
<p align="center"> Fig. 方案数量递推公式 </p>
|
|
|
|
|
|
|
|
|
|
基于此递推公式,我们可以写出递归代码:以 $dp[n]$ 为起始点,**从顶至底地将一个较大问题拆解为两个较小问题**,直至到达最小子问题 $dp[1]$ 和 $dp[2]$ 时返回。其中,最小子问题的解是已知的,即爬到第 $1$ , $2$ 阶分别有 $1$ , $2$ 种方案。
|
|
|
|
|
|
|
|
|
|
以下代码与回溯解法一样,都属于深度优先搜索,但它比回溯算法更加简洁,这体现了从分治角度考虑这道题的优势。
|
|
|
|
|
|
|
|
|
|
=== "Java"
|
|
|
|
|
|
|
|
|
|
```java title="climbing_stairs_dfs.java"
|
|
|
|
|
/* 搜索 */
|
|
|
|
|
int dfs(int i) {
|
|
|
|
|
// 已知 dp[1] 和 dp[2] ,返回之
|
|
|
|
|
if (i == 1 || i == 2)
|
|
|
|
|
return i;
|
|
|
|
|
// dp[i] = dp[i-1] + dp[i-2]
|
|
|
|
|
int count = dfs(i - 1) + dfs(i - 2);
|
|
|
|
|
return count;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* 爬楼梯:搜索 */
|
|
|
|
|
int climbingStairsDFS(int n) {
|
|
|
|
|
return dfs(n);
|
|
|
|
|
}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C++"
|
|
|
|
|
|
|
|
|
|
```cpp title="climbing_stairs_dfs.cpp"
|
|
|
|
|
/* 搜索 */
|
|
|
|
|
int dfs(int i) {
|
|
|
|
|
// 已知 dp[1] 和 dp[2] ,返回之
|
|
|
|
|
if (i == 1 || i == 2)
|
|
|
|
|
return i;
|
|
|
|
|
// dp[i] = dp[i-1] + dp[i-2]
|
|
|
|
|
int count = dfs(i - 1) + dfs(i - 2);
|
|
|
|
|
return count;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* 爬楼梯:搜索 */
|
|
|
|
|
int climbingStairsDFS(int n) {
|
|
|
|
|
return dfs(n);
|
|
|
|
|
}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Python"
|
|
|
|
|
|
|
|
|
|
```python title="climbing_stairs_dfs.py"
|
|
|
|
|
def dfs(i: int) -> int:
|
|
|
|
|
"""搜索"""
|
|
|
|
|
# 已知 dp[1] 和 dp[2] ,返回之
|
|
|
|
|
if i == 1 or i == 2:
|
|
|
|
|
return i
|
|
|
|
|
# dp[i] = dp[i-1] + dp[i-2]
|
|
|
|
|
count = dfs(i - 1) + dfs(i - 2)
|
|
|
|
|
return count
|
|
|
|
|
|
|
|
|
|
def climbing_stairs_dfs(n: int) -> int:
|
|
|
|
|
"""爬楼梯:搜索"""
|
|
|
|
|
return dfs(n)
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Go"
|
|
|
|
|
|
|
|
|
|
```go title="climbing_stairs_dfs.go"
|
|
|
|
|
[class]{}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsDFS}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "JavaScript"
|
|
|
|
|
|
|
|
|
|
```javascript title="climbing_stairs_dfs.js"
|
|
|
|
|
[class]{}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsDFS}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "TypeScript"
|
|
|
|
|
|
|
|
|
|
```typescript title="climbing_stairs_dfs.ts"
|
|
|
|
|
[class]{}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsDFS}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C"
|
|
|
|
|
|
|
|
|
|
```c title="climbing_stairs_dfs.c"
|
|
|
|
|
[class]{}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsDFS}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C#"
|
|
|
|
|
|
|
|
|
|
```csharp title="climbing_stairs_dfs.cs"
|
|
|
|
|
[class]{climbing_stairs_dfs}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{climbing_stairs_dfs}-[func]{climbingStairsDFS}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Swift"
|
|
|
|
|
|
|
|
|
|
```swift title="climbing_stairs_dfs.swift"
|
|
|
|
|
[class]{}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsDFS}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Zig"
|
|
|
|
|
|
|
|
|
|
```zig title="climbing_stairs_dfs.zig"
|
|
|
|
|
[class]{}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsDFS}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Dart"
|
|
|
|
|
|
|
|
|
|
```dart title="climbing_stairs_dfs.dart"
|
|
|
|
|
[class]{}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsDFS}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
下图展示了该方法形成的递归树。对于问题 $dp[n]$ ,递归树的深度为 $n$ ,时间复杂度为 $O(2^n)$ 。指数阶的运行时间增长地非常快,如果我们输入一个比较大的 $n$ ,则会陷入漫长的等待之中。
|
|
|
|
|
|
|
|
|
|
![爬楼梯对应递归树](intro_to_dynamic_programming.assets/climbing_stairs_dfs_tree.png)
|
|
|
|
|
|
|
|
|
|
<p align="center"> Fig. 爬楼梯对应递归树 </p>
|
|
|
|
|
|
|
|
|
|
实际上,**指数阶的时间复杂度是由于「重叠子问题」导致的**。例如,问题 $dp[9]$ 被分解为子问题 $dp[8]$ 和 $dp[7]$ ,问题 $dp[8]$ 被分解为子问题 $dp[7]$ 和 $dp[6]$ ,两者都包含子问题 $dp[7]$ ,而子问题中又包含更小的重叠子问题,子子孙孙无穷尽也,绝大部分计算资源都浪费在这些重叠的问题上。
|
|
|
|
|
|
|
|
|
|
## 13.1.2. 方法二:记忆化搜索
|
|
|
|
|
|
|
|
|
|
为了提升算法效率,**我们希望所有的重叠子问题都只被计算一次**。具体来说,考虑借助一个数组 `mem` 来记录每个子问题的解,并在搜索过程中这样做:
|
|
|
|
|
|
|
|
|
|
- 当首次计算 $dp[i]$ 时,我们将其记录至 `mem[i]` ,以便之后使用;
|
|
|
|
|
- 当再次需要计算 $dp[i]$ 时,我们便可直接从 `mem[i]` 中获取结果,从而将重叠子问题剪枝;
|
|
|
|
|
|
|
|
|
|
=== "Java"
|
|
|
|
|
|
|
|
|
|
```java title="climbing_stairs_dfs_mem.java"
|
|
|
|
|
/* 记忆化搜索 */
|
|
|
|
|
int dfs(int i, int[] mem) {
|
|
|
|
|
// 已知 dp[1] 和 dp[2] ,返回之
|
|
|
|
|
if (i == 1 || i == 2)
|
|
|
|
|
return i;
|
|
|
|
|
// 若存在记录 dp[i] ,则直接返回之
|
|
|
|
|
if (mem[i] != -1)
|
|
|
|
|
return mem[i];
|
|
|
|
|
// dp[i] = dp[i-1] + dp[i-2]
|
|
|
|
|
int count = dfs(i - 1, mem) + dfs(i - 2, mem);
|
|
|
|
|
// 记录 dp[i]
|
|
|
|
|
mem[i] = count;
|
|
|
|
|
return count;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* 爬楼梯:记忆化搜索 */
|
|
|
|
|
int climbingStairsDFSMem(int n) {
|
|
|
|
|
// mem[i] 记录爬到第 i 阶的方案总数,-1 代表无记录
|
|
|
|
|
int[] mem = new int[n + 1];
|
|
|
|
|
Arrays.fill(mem, -1);
|
|
|
|
|
return dfs(n, mem);
|
|
|
|
|
}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C++"
|
|
|
|
|
|
|
|
|
|
```cpp title="climbing_stairs_dfs_mem.cpp"
|
|
|
|
|
/* 记忆化搜索 */
|
|
|
|
|
int dfs(int i, vector<int> &mem) {
|
|
|
|
|
// 已知 dp[1] 和 dp[2] ,返回之
|
|
|
|
|
if (i == 1 || i == 2)
|
|
|
|
|
return i;
|
|
|
|
|
// 若存在记录 dp[i] ,则直接返回之
|
|
|
|
|
if (mem[i] != -1)
|
|
|
|
|
return mem[i];
|
|
|
|
|
// dp[i] = dp[i-1] + dp[i-2]
|
|
|
|
|
int count = dfs(i - 1, mem) + dfs(i - 2, mem);
|
|
|
|
|
// 记录 dp[i]
|
|
|
|
|
mem[i] = count;
|
|
|
|
|
return count;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* 爬楼梯:记忆化搜索 */
|
|
|
|
|
int climbingStairsDFSMem(int n) {
|
|
|
|
|
// mem[i] 记录爬到第 i 阶的方案总数,-1 代表无记录
|
|
|
|
|
vector<int> mem(n + 1, -1);
|
|
|
|
|
return dfs(n, mem);
|
|
|
|
|
}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Python"
|
|
|
|
|
|
|
|
|
|
```python title="climbing_stairs_dfs_mem.py"
|
|
|
|
|
def dfs(i: int, mem: list[int]) -> int:
|
|
|
|
|
"""记忆化搜索"""
|
|
|
|
|
# 已知 dp[1] 和 dp[2] ,返回之
|
|
|
|
|
if i == 1 or i == 2:
|
|
|
|
|
return i
|
|
|
|
|
# 若存在记录 dp[i] ,则直接返回之
|
|
|
|
|
if mem[i] != -1:
|
|
|
|
|
return mem[i]
|
|
|
|
|
# dp[i] = dp[i-1] + dp[i-2]
|
|
|
|
|
count = dfs(i - 1, mem) + dfs(i - 2, mem)
|
|
|
|
|
# 记录 dp[i]
|
|
|
|
|
mem[i] = count
|
|
|
|
|
return count
|
|
|
|
|
|
|
|
|
|
def climbing_stairs_dfs_mem(n: int) -> int:
|
|
|
|
|
"""爬楼梯:记忆化搜索"""
|
|
|
|
|
# mem[i] 记录爬到第 i 阶的方案总数,-1 代表无记录
|
|
|
|
|
mem = [-1] * (n + 1)
|
|
|
|
|
return dfs(n, mem)
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Go"
|
|
|
|
|
|
|
|
|
|
```go title="climbing_stairs_dfs_mem.go"
|
|
|
|
|
[class]{}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsDFSMem}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "JavaScript"
|
|
|
|
|
|
|
|
|
|
```javascript title="climbing_stairs_dfs_mem.js"
|
|
|
|
|
[class]{}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsDFSMem}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "TypeScript"
|
|
|
|
|
|
|
|
|
|
```typescript title="climbing_stairs_dfs_mem.ts"
|
|
|
|
|
[class]{}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsDFSMem}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C"
|
|
|
|
|
|
|
|
|
|
```c title="climbing_stairs_dfs_mem.c"
|
|
|
|
|
[class]{}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsDFSMem}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C#"
|
|
|
|
|
|
|
|
|
|
```csharp title="climbing_stairs_dfs_mem.cs"
|
|
|
|
|
[class]{climbing_stairs_dfs_mem}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{climbing_stairs_dfs_mem}-[func]{climbingStairsDFSMem}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Swift"
|
|
|
|
|
|
|
|
|
|
```swift title="climbing_stairs_dfs_mem.swift"
|
|
|
|
|
[class]{}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsDFSMem}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Zig"
|
|
|
|
|
|
|
|
|
|
```zig title="climbing_stairs_dfs_mem.zig"
|
|
|
|
|
[class]{}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsDFSMem}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Dart"
|
|
|
|
|
|
|
|
|
|
```dart title="climbing_stairs_dfs_mem.dart"
|
|
|
|
|
[class]{}-[func]{dfs}
|
|
|
|
|
|
|
|
|
|
[class]{}-[func]{climbingStairsDFSMem}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
观察下图,**经过记忆化处理后,所有重叠子问题都只需被计算一次,时间复杂度被优化至 $O(n)$** ,这是一个巨大的飞跃。实际上,如果不考虑递归带来的额外开销,记忆化搜索解法已经几乎等同于动态规划解法的时间效率。
|
|
|
|
|
|
|
|
|
|
![记忆化搜索对应递归树](intro_to_dynamic_programming.assets/climbing_stairs_dfs_memo_tree.png)
|
|
|
|
|
|
|
|
|
|
<p align="center"> Fig. 记忆化搜索对应递归树 </p>
|
|
|
|
|
|
|
|
|
|
## 13.1.3. 方法三:动态规划
|
|
|
|
|
|
|
|
|
|
**记忆化搜索是一种“从顶至底”的方法**:我们从原问题(根节点)开始,递归地将较大子问题分解为较小子问题,直至解已知的最小子问题(叶节点);最终通过回溯将子问题的解逐层收集,得到原问题的解。
|
|
|
|
|
|
|
|
|
|
**我们也可以直接“从底至顶”进行求解**,得到标准的动态规划解法:从最小子问题开始,迭代地求解较大子问题,直至得到原问题的解。
|
|
|
|
|
|
|
|
|
|
由于没有回溯过程,动态规划可以直接基于循环实现。我们初始化一个数组 `dp` 来存储子问题的解,从最小子问题开始,逐步求解较大子问题。在以下代码中,数组 `dp` 起到了记忆化搜索中数组 `mem` 相同的记录作用。
|
|
|
|
|
|
|
|
|
|
=== "Java"
|
|
|
|
|
|
|
|
|
|
```java title="climbing_stairs_dp.java"
|
|
|
|
|
/* 爬楼梯:动态规划 */
|
|
|
|
|
int climbingStairsDP(int n) {
|
|
|
|
|
if (n == 1 || n == 2)
|
|
|
|
|
return n;
|
|
|
|
|
// 初始化 dp 列表,用于存储子问题的解
|
|
|
|
|
int[] dp = new int[n + 1];
|
|
|
|
|
// 初始状态:预设最小子问题的解
|
|
|
|
|
dp[1] = 1;
|
|
|
|
|
dp[2] = 2;
|
|
|
|
|
// 状态转移:从较小子问题逐步求解较大子问题
|
|
|
|
|
for (int i = 3; i <= n; i++) {
|
|
|
|
|
dp[i] = dp[i - 1] + dp[i - 2];
|
|
|
|
|
}
|
|
|
|
|
return dp[n];
|
|
|
|
|
}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C++"
|
|
|
|
|
|
|
|
|
|
```cpp title="climbing_stairs_dp.cpp"
|
|
|
|
|
/* 爬楼梯:动态规划 */
|
|
|
|
|
int climbingStairsDP(int n) {
|
|
|
|
|
if (n == 1 || n == 2)
|
|
|
|
|
return n;
|
|
|
|
|
// 初始化 dp 列表,用于存储子问题的解
|
|
|
|
|
vector<int> dp(n + 1);
|
|
|
|
|
// 初始状态:预设最小子问题的解
|
|
|
|
|
dp[1] = 1;
|
|
|
|
|
dp[2] = 2;
|
|
|
|
|
// 状态转移:从较小子问题逐步求解较大子问题
|
|
|
|
|
for (int i = 3; i <= n; i++) {
|
|
|
|
|
dp[i] = dp[i - 1] + dp[i - 2];
|
|
|
|
|
}
|
|
|
|
|
return dp[n];
|
|
|
|
|
}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Python"
|
|
|
|
|
|
|
|
|
|
```python title="climbing_stairs_dp.py"
|
|
|
|
|
def climbing_stairs_dp(n: int) -> int:
|
|
|
|
|
"""爬楼梯:动态规划"""
|
|
|
|
|
if n == 1 or n == 2:
|
|
|
|
|
return n
|
|
|
|
|
# 初始化 dp 列表,用于存储子问题的解
|
|
|
|
|
dp = [0] * (n + 1)
|
|
|
|
|
# 初始状态:预设最小子问题的解
|
|
|
|
|
dp[1], dp[2] = 1, 2
|
|
|
|
|
# 状态转移:从较小子问题逐步求解较大子问题
|
|
|
|
|
for i in range(3, n + 1):
|
|
|
|
|
dp[i] = dp[i - 1] + dp[i - 2]
|
|
|
|
|
return dp[n]
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Go"
|
|
|
|
|
|
|
|
|
|
```go title="climbing_stairs_dp.go"
|
|
|
|
|
[class]{}-[func]{climbingStairsDP}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "JavaScript"
|
|
|
|
|
|
|
|
|
|
```javascript title="climbing_stairs_dp.js"
|
|
|
|
|
[class]{}-[func]{climbingStairsDP}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "TypeScript"
|
|
|
|
|
|
|
|
|
|
```typescript title="climbing_stairs_dp.ts"
|
|
|
|
|
[class]{}-[func]{climbingStairsDP}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C"
|
|
|
|
|
|
|
|
|
|
```c title="climbing_stairs_dp.c"
|
|
|
|
|
[class]{}-[func]{climbingStairsDP}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C#"
|
|
|
|
|
|
|
|
|
|
```csharp title="climbing_stairs_dp.cs"
|
|
|
|
|
[class]{climbing_stairs_dp}-[func]{climbingStairsDP}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Swift"
|
|
|
|
|
|
|
|
|
|
```swift title="climbing_stairs_dp.swift"
|
|
|
|
|
[class]{}-[func]{climbingStairsDP}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Zig"
|
|
|
|
|
|
|
|
|
|
```zig title="climbing_stairs_dp.zig"
|
|
|
|
|
[class]{}-[func]{climbingStairsDP}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Dart"
|
|
|
|
|
|
|
|
|
|
```dart title="climbing_stairs_dp.dart"
|
|
|
|
|
[class]{}-[func]{climbingStairsDP}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
与回溯算法一样,动态规划也使用“状态”概念来表示问题求解的某个特定阶段,每个状态都对应一个子问题以及相应的局部最优解。例如对于爬楼梯问题,状态定义为当前所在楼梯阶数。**动态规划的常用术语包括**:
|
|
|
|
|
|
|
|
|
|
- 将 $dp$ 数组称为「状态列表」,$dp[i]$ 代表第 $i$ 个状态的解;
|
|
|
|
|
- 将最简单子问题对应的状态(即第 $1$ , $2$ 阶楼梯)称为「初始状态」;
|
|
|
|
|
- 将递推公式 $dp[i] = dp[i-1] + dp[i-2]$ 称为「状态转移方程」;
|
|
|
|
|
|
|
|
|
|
![爬楼梯的动态规划过程](intro_to_dynamic_programming.assets/climbing_stairs_dp.png)
|
|
|
|
|
|
|
|
|
|
<p align="center"> Fig. 爬楼梯的动态规划过程 </p>
|
|
|
|
|
|
|
|
|
|
细心的你可能发现,**由于 $dp[i]$ 只与 $dp[i-1]$ 和 $dp[i-2]$ 有关,因此我们无需使用一个数组 `dp` 来存储所有状态**,而只需两个变量滚动前进即可。如以下代码所示,由于省去了数组 `dp` 占用的空间,因此空间复杂度从 $O(n)$ 降低至 $O(1)$ 。
|
|
|
|
|
|
|
|
|
|
=== "Java"
|
|
|
|
|
|
|
|
|
|
```java title="climbing_stairs_dp.java"
|
|
|
|
|
/* 爬楼梯:状态压缩后的动态规划 */
|
|
|
|
|
int climbingStairsDPComp(int n) {
|
|
|
|
|
if (n == 1 || n == 2)
|
|
|
|
|
return n;
|
|
|
|
|
int a = 1, b = 2;
|
|
|
|
|
for (int i = 3; i <= n; i++) {
|
|
|
|
|
int tmp = b;
|
|
|
|
|
b = a + b;
|
|
|
|
|
a = tmp;
|
|
|
|
|
}
|
|
|
|
|
return b;
|
|
|
|
|
}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C++"
|
|
|
|
|
|
|
|
|
|
```cpp title="climbing_stairs_dp.cpp"
|
|
|
|
|
/* 爬楼梯:状态压缩后的动态规划 */
|
|
|
|
|
int climbingStairsDPComp(int n) {
|
|
|
|
|
if (n == 1 || n == 2)
|
|
|
|
|
return n;
|
|
|
|
|
int a = 1, b = 2;
|
|
|
|
|
for (int i = 3; i <= n; i++) {
|
|
|
|
|
int tmp = b;
|
|
|
|
|
b = a + b;
|
|
|
|
|
a = tmp;
|
|
|
|
|
}
|
|
|
|
|
return b;
|
|
|
|
|
}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Python"
|
|
|
|
|
|
|
|
|
|
```python title="climbing_stairs_dp.py"
|
|
|
|
|
def climbing_stairs_dp_comp(n: int) -> int:
|
|
|
|
|
"""爬楼梯:状态压缩后的动态规划"""
|
|
|
|
|
if n == 1 or n == 2:
|
|
|
|
|
return n
|
|
|
|
|
a, b = 1, 2
|
|
|
|
|
for _ in range(3, n + 1):
|
|
|
|
|
a, b = b, a + b
|
|
|
|
|
return b
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Go"
|
|
|
|
|
|
|
|
|
|
```go title="climbing_stairs_dp.go"
|
|
|
|
|
[class]{}-[func]{climbingStairsDPComp}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "JavaScript"
|
|
|
|
|
|
|
|
|
|
```javascript title="climbing_stairs_dp.js"
|
|
|
|
|
[class]{}-[func]{climbingStairsDPComp}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "TypeScript"
|
|
|
|
|
|
|
|
|
|
```typescript title="climbing_stairs_dp.ts"
|
|
|
|
|
[class]{}-[func]{climbingStairsDPComp}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C"
|
|
|
|
|
|
|
|
|
|
```c title="climbing_stairs_dp.c"
|
|
|
|
|
[class]{}-[func]{climbingStairsDPComp}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C#"
|
|
|
|
|
|
|
|
|
|
```csharp title="climbing_stairs_dp.cs"
|
|
|
|
|
[class]{climbing_stairs_dp}-[func]{climbingStairsDPComp}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Swift"
|
|
|
|
|
|
|
|
|
|
```swift title="climbing_stairs_dp.swift"
|
|
|
|
|
[class]{}-[func]{climbingStairsDPComp}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Zig"
|
|
|
|
|
|
|
|
|
|
```zig title="climbing_stairs_dp.zig"
|
|
|
|
|
[class]{}-[func]{climbingStairsDPComp}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Dart"
|
|
|
|
|
|
|
|
|
|
```dart title="climbing_stairs_dp.dart"
|
|
|
|
|
[class]{}-[func]{climbingStairsDPComp}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
**我们将这种空间优化技巧称为「状态压缩」**。在许多动态规划问题中,当前状态仅与前面有限个状态有关,不必保存所有的历史状态,这时我们可以应用状态压缩,只保留必要的状态,通过“降维”来节省内存空间。
|