mirror of
https://github.com/krahets/hello-algo.git
synced 2024-12-26 11:56:30 +08:00
1170 lines
25 KiB
Markdown
Executable file
1170 lines
25 KiB
Markdown
Executable file
# 空间复杂度
|
||
|
||
「空间复杂度 Space Complexity」用于衡量算法占用内存空间随着数据量变大时的增长趋势。这个概念与时间复杂度非常类似,只需将“运行时间”替换为“占用内存空间”。
|
||
|
||
## 算法相关空间
|
||
|
||
算法运行过程中使用的内存空间主要包括以下几种:
|
||
|
||
- **输入空间**:用于存储算法的输入数据。
|
||
- **暂存空间**:用于存储算法运行过程中的变量、对象、函数上下文等数据。
|
||
- **输出空间**:用于存储算法的输出数据。
|
||
|
||
一般情况下,空间复杂度的统计范围是“暂存空间”加上“输出空间”。
|
||
|
||
暂存空间可以进一步划分为三个部分:
|
||
|
||
- **暂存数据**:用于保存算法运行过程中的各种常量、变量、对象等。
|
||
- **栈帧空间**:用于保存调用函数的上下文数据。系统在每次调用函数时都会在栈顶部创建一个栈帧,函数返回后,栈帧空间会被释放。
|
||
- **指令空间**:用于保存编译后的程序指令,在实际统计中通常忽略不计。
|
||
|
||
因此在分析一段程序的空间复杂度时,**我们通常统计暂存数据、输出数据、栈帧空间三部分**。
|
||
|
||
![算法使用的相关空间](space_complexity.assets/space_types.png)
|
||
|
||
=== "Java"
|
||
|
||
```java title=""
|
||
/* 类 */
|
||
class Node {
|
||
int val;
|
||
Node next;
|
||
Node(int x) { val = x; }
|
||
}
|
||
|
||
/* 函数 */
|
||
int function() {
|
||
// do something...
|
||
return 0;
|
||
}
|
||
|
||
int algorithm(int n) { // 输入数据
|
||
final int a = 0; // 暂存数据(常量)
|
||
int b = 0; // 暂存数据(变量)
|
||
Node node = new Node(0); // 暂存数据(对象)
|
||
int c = function(); // 栈帧空间(调用函数)
|
||
return a + b + c; // 输出数据
|
||
}
|
||
```
|
||
|
||
=== "C++"
|
||
|
||
```cpp title=""
|
||
/* 结构体 */
|
||
struct Node {
|
||
int val;
|
||
Node *next;
|
||
Node(int x) : val(x), next(nullptr) {}
|
||
};
|
||
|
||
/* 函数 */
|
||
int func() {
|
||
// do something...
|
||
return 0;
|
||
}
|
||
|
||
int algorithm(int n) { // 输入数据
|
||
const int a = 0; // 暂存数据(常量)
|
||
int b = 0; // 暂存数据(变量)
|
||
Node* node = new Node(0); // 暂存数据(对象)
|
||
int c = func(); // 栈帧空间(调用函数)
|
||
return a + b + c; // 输出数据
|
||
}
|
||
```
|
||
|
||
=== "Python"
|
||
|
||
```python title=""
|
||
class Node:
|
||
"""类"""
|
||
def __init__(self, x: int):
|
||
self.val: int = x # 节点值
|
||
self.next: Optional[Node] = None # 指向下一节点的引用
|
||
|
||
def function() -> int:
|
||
"""函数"""
|
||
# do something...
|
||
return 0
|
||
|
||
def algorithm(n) -> int: # 输入数据
|
||
A = 0 # 暂存数据(常量,一般用大写字母表示)
|
||
b = 0 # 暂存数据(变量)
|
||
node = Node(0) # 暂存数据(对象)
|
||
c = function() # 栈帧空间(调用函数)
|
||
return A + b + c # 输出数据
|
||
```
|
||
|
||
=== "Go"
|
||
|
||
```go title=""
|
||
/* 结构体 */
|
||
type node struct {
|
||
val int
|
||
next *node
|
||
}
|
||
|
||
/* 创建 node 结构体 */
|
||
func newNode(val int) *node {
|
||
return &node{val: val}
|
||
}
|
||
|
||
/* 函数 */
|
||
func function() int {
|
||
// do something...
|
||
return 0
|
||
}
|
||
|
||
func algorithm(n int) int { // 输入数据
|
||
const a = 0 // 暂存数据(常量)
|
||
b := 0 // 暂存数据(变量)
|
||
newNode(0) // 暂存数据(对象)
|
||
c := function() // 栈帧空间(调用函数)
|
||
return a + b + c // 输出数据
|
||
}
|
||
```
|
||
|
||
=== "JS"
|
||
|
||
```javascript title=""
|
||
/* 类 */
|
||
class Node {
|
||
val;
|
||
next;
|
||
constructor(val) {
|
||
this.val = val === undefined ? 0 : val; // 节点值
|
||
this.next = null; // 指向下一节点的引用
|
||
}
|
||
}
|
||
|
||
/* 函数 */
|
||
function constFunc() {
|
||
// do something
|
||
return 0;
|
||
}
|
||
|
||
function algorithm(n) { // 输入数据
|
||
const a = 0; // 暂存数据(常量)
|
||
let b = 0; // 暂存数据(变量)
|
||
const node = new Node(0); // 暂存数据(对象)
|
||
const c = constFunc(); // 栈帧空间(调用函数)
|
||
return a + b + c; // 输出数据
|
||
}
|
||
```
|
||
|
||
=== "TS"
|
||
|
||
```typescript title=""
|
||
/* 类 */
|
||
class Node {
|
||
val: number;
|
||
next: Node | null;
|
||
constructor(val?: number) {
|
||
this.val = val === undefined ? 0 : val; // 节点值
|
||
this.next = null; // 指向下一节点的引用
|
||
}
|
||
}
|
||
|
||
/* 函数 */
|
||
function constFunc(): number {
|
||
// do something
|
||
return 0;
|
||
}
|
||
|
||
function algorithm(n: number): number { // 输入数据
|
||
const a = 0; // 暂存数据(常量)
|
||
let b = 0; // 暂存数据(变量)
|
||
const node = new Node(0); // 暂存数据(对象)
|
||
const c = constFunc(); // 栈帧空间(调用函数)
|
||
return a + b + c; // 输出数据
|
||
}
|
||
```
|
||
|
||
=== "C"
|
||
|
||
```c title=""
|
||
/* 函数 */
|
||
int func() {
|
||
// do something...
|
||
return 0;
|
||
}
|
||
|
||
int algorithm(int n) { // 输入数据
|
||
const int a = 0; // 暂存数据(常量)
|
||
int b = 0; // 暂存数据(变量)
|
||
int c = func(); // 栈帧空间(调用函数)
|
||
return a + b + c; // 输出数据
|
||
}
|
||
```
|
||
|
||
=== "C#"
|
||
|
||
```csharp title=""
|
||
/* 类 */
|
||
class Node {
|
||
int val;
|
||
Node next;
|
||
Node(int x) { val = x; }
|
||
}
|
||
|
||
/* 函数 */
|
||
int function() {
|
||
// do something...
|
||
return 0;
|
||
}
|
||
|
||
int algorithm(int n) { // 输入数据
|
||
const int a = 0; // 暂存数据(常量)
|
||
int b = 0; // 暂存数据(变量)
|
||
Node node = new Node(0); // 暂存数据(对象)
|
||
int c = function(); // 栈帧空间(调用函数)
|
||
return a + b + c; // 输出数据
|
||
}
|
||
```
|
||
|
||
=== "Swift"
|
||
|
||
```swift title=""
|
||
/* 类 */
|
||
class Node {
|
||
var val: Int
|
||
var next: Node?
|
||
|
||
init(x: Int) {
|
||
val = x
|
||
}
|
||
}
|
||
|
||
/* 函数 */
|
||
func function() -> Int {
|
||
// do something...
|
||
return 0
|
||
}
|
||
|
||
func algorithm(n: Int) -> Int { // 输入数据
|
||
let a = 0 // 暂存数据(常量)
|
||
var b = 0 // 暂存数据(变量)
|
||
let node = Node(x: 0) // 暂存数据(对象)
|
||
let c = function() // 栈帧空间(调用函数)
|
||
return a + b + c // 输出数据
|
||
}
|
||
```
|
||
|
||
=== "Zig"
|
||
|
||
```zig title=""
|
||
|
||
```
|
||
|
||
=== "Dart"
|
||
|
||
```dart title=""
|
||
/* 类 */
|
||
class Node {
|
||
int val;
|
||
Node next;
|
||
Node(this.val, [this.next]);
|
||
}
|
||
|
||
/* 函数 */
|
||
int function() {
|
||
// do something...
|
||
return 0;
|
||
}
|
||
|
||
int algorithm(int n) { // 输入数据
|
||
const int a = 0; // 暂存数据(常量)
|
||
int b = 0; // 暂存数据(变量)
|
||
Node node = Node(0); // 暂存数据(对象)
|
||
int c = function(); // 栈帧空间(调用函数)
|
||
return a + b + c; // 输出数据
|
||
}
|
||
```
|
||
|
||
=== "Rust"
|
||
|
||
```rust title=""
|
||
|
||
```
|
||
|
||
## 推算方法
|
||
|
||
空间复杂度的推算方法与时间复杂度大致相同,只需将统计对象从“计算操作数量”转为“使用空间大小”。
|
||
|
||
而与时间复杂度不同的是,**我们通常只关注「最差空间复杂度」**。这是因为内存空间是一项硬性要求,我们必须确保在所有输入数据下都有足够的内存空间预留。
|
||
|
||
观察以下代码,最差空间复杂度中的“最差”有两层含义。
|
||
|
||
1. **以最差输入数据为准**:当 $n < 10$ 时,空间复杂度为 $O(1)$ ;但当 $n > 10$ 时,初始化的数组 `nums` 占用 $O(n)$ 空间;因此最差空间复杂度为 $O(n)$ 。
|
||
2. **以算法运行中的峰值内存为准**:例如,程序在执行最后一行之前,占用 $O(1)$ 空间;当初始化数组 `nums` 时,程序占用 $O(n)$ 空间;因此最差空间复杂度为 $O(n)$ 。
|
||
|
||
=== "Java"
|
||
|
||
```java title=""
|
||
void algorithm(int n) {
|
||
int a = 0; // O(1)
|
||
int[] b = new int[10000]; // O(1)
|
||
if (n > 10)
|
||
int[] nums = new int[n]; // O(n)
|
||
}
|
||
```
|
||
|
||
=== "C++"
|
||
|
||
```cpp title=""
|
||
void algorithm(int n) {
|
||
int a = 0; // O(1)
|
||
vector<int> b(10000); // O(1)
|
||
if (n > 10)
|
||
vector<int> nums(n); // O(n)
|
||
}
|
||
```
|
||
|
||
=== "Python"
|
||
|
||
```python title=""
|
||
def algorithm(n: int):
|
||
a = 0 # O(1)
|
||
b = [0] * 10000 # O(1)
|
||
if n > 10:
|
||
nums = [0] * n # O(n)
|
||
```
|
||
|
||
=== "Go"
|
||
|
||
```go title=""
|
||
func algorithm(n int) {
|
||
a := 0 // O(1)
|
||
b := make([]int, 10000) // O(1)
|
||
var nums []int
|
||
if n > 10 {
|
||
nums := make([]int, n) // O(n)
|
||
}
|
||
fmt.Println(a, b, nums)
|
||
}
|
||
```
|
||
|
||
=== "JS"
|
||
|
||
```javascript title=""
|
||
function algorithm(n) {
|
||
const a = 0; // O(1)
|
||
const b = new Array(10000); // O(1)
|
||
if (n > 10) {
|
||
const nums = new Array(n); // O(n)
|
||
}
|
||
}
|
||
```
|
||
|
||
=== "TS"
|
||
|
||
```typescript title=""
|
||
function algorithm(n: number): void {
|
||
const a = 0; // O(1)
|
||
const b = new Array(10000); // O(1)
|
||
if (n > 10) {
|
||
const nums = new Array(n); // O(n)
|
||
}
|
||
}
|
||
```
|
||
|
||
=== "C"
|
||
|
||
```c title=""
|
||
void algorithm(int n) {
|
||
int a = 0; // O(1)
|
||
int b[10000]; // O(1)
|
||
if (n > 10)
|
||
int nums[n] = {0}; // O(n)
|
||
}
|
||
```
|
||
|
||
=== "C#"
|
||
|
||
```csharp title=""
|
||
void algorithm(int n) {
|
||
int a = 0; // O(1)
|
||
int[] b = new int[10000]; // O(1)
|
||
if (n > 10) {
|
||
int[] nums = new int[n]; // O(n)
|
||
}
|
||
}
|
||
```
|
||
|
||
=== "Swift"
|
||
|
||
```swift title=""
|
||
func algorithm(n: Int) {
|
||
let a = 0 // O(1)
|
||
let b = Array(repeating: 0, count: 10000) // O(1)
|
||
if n > 10 {
|
||
let nums = Array(repeating: 0, count: n) // O(n)
|
||
}
|
||
}
|
||
```
|
||
|
||
=== "Zig"
|
||
|
||
```zig title=""
|
||
|
||
```
|
||
|
||
=== "Dart"
|
||
|
||
```dart title=""
|
||
void algorithm(int n) {
|
||
int a = 0; // O(1)
|
||
List<int> b = List.filled(10000, 0); // O(1)
|
||
if (n > 10) {
|
||
List<int> nums = List.filled(n, 0); // O(n)
|
||
}
|
||
}
|
||
```
|
||
|
||
=== "Rust"
|
||
|
||
```rust title=""
|
||
|
||
```
|
||
|
||
**在递归函数中,需要注意统计栈帧空间**。例如以下代码:
|
||
|
||
- 函数 `loop()` 在循环中调用了 $n$ 次 `function()` ,每轮中的 `function()` 都返回并释放了栈帧空间,因此空间复杂度仍为 $O(1)$ 。
|
||
- 递归函数 `recur()` 在运行过程中会同时存在 $n$ 个未返回的 `recur()` ,从而占用 $O(n)$ 的栈帧空间。
|
||
|
||
=== "Java"
|
||
|
||
```java title=""
|
||
int function() {
|
||
// do something
|
||
return 0;
|
||
}
|
||
/* 循环 O(1) */
|
||
void loop(int n) {
|
||
for (int i = 0; i < n; i++) {
|
||
function();
|
||
}
|
||
}
|
||
/* 递归 O(n) */
|
||
void recur(int n) {
|
||
if (n == 1) return;
|
||
return recur(n - 1);
|
||
}
|
||
```
|
||
|
||
=== "C++"
|
||
|
||
```cpp title=""
|
||
int func() {
|
||
// do something
|
||
return 0;
|
||
}
|
||
/* 循环 O(1) */
|
||
void loop(int n) {
|
||
for (int i = 0; i < n; i++) {
|
||
func();
|
||
}
|
||
}
|
||
/* 递归 O(n) */
|
||
void recur(int n) {
|
||
if (n == 1) return;
|
||
return recur(n - 1);
|
||
}
|
||
```
|
||
|
||
=== "Python"
|
||
|
||
```python title=""
|
||
def function() -> int:
|
||
# do something
|
||
return 0
|
||
|
||
def loop(n: int):
|
||
"""循环 O(1)"""
|
||
for _ in range(n):
|
||
function()
|
||
|
||
def recur(n: int) -> int:
|
||
"""递归 O(n)"""
|
||
if n == 1: return
|
||
return recur(n - 1)
|
||
```
|
||
|
||
=== "Go"
|
||
|
||
```go title=""
|
||
func function() int {
|
||
// do something
|
||
return 0
|
||
}
|
||
|
||
/* 循环 O(1) */
|
||
func loop(n int) {
|
||
for i := 0; i < n; i++ {
|
||
function()
|
||
}
|
||
}
|
||
|
||
/* 递归 O(n) */
|
||
func recur(n int) {
|
||
if n == 1 {
|
||
return
|
||
}
|
||
recur(n - 1)
|
||
}
|
||
```
|
||
|
||
=== "JS"
|
||
|
||
```javascript title=""
|
||
function constFunc() {
|
||
// do something
|
||
return 0;
|
||
}
|
||
/* 循环 O(1) */
|
||
function loop(n) {
|
||
for (let i = 0; i < n; i++) {
|
||
constFunc();
|
||
}
|
||
}
|
||
/* 递归 O(n) */
|
||
function recur(n) {
|
||
if (n === 1) return;
|
||
return recur(n - 1);
|
||
}
|
||
```
|
||
|
||
=== "TS"
|
||
|
||
```typescript title=""
|
||
function constFunc(): number {
|
||
// do something
|
||
return 0;
|
||
}
|
||
/* 循环 O(1) */
|
||
function loop(n: number): void {
|
||
for (let i = 0; i < n; i++) {
|
||
constFunc();
|
||
}
|
||
}
|
||
/* 递归 O(n) */
|
||
function recur(n: number): void {
|
||
if (n === 1) return;
|
||
return recur(n - 1);
|
||
}
|
||
```
|
||
|
||
=== "C"
|
||
|
||
```c title=""
|
||
int func() {
|
||
// do something
|
||
return 0;
|
||
}
|
||
/* 循环 O(1) */
|
||
void loop(int n) {
|
||
for (int i = 0; i < n; i++) {
|
||
func();
|
||
}
|
||
}
|
||
/* 递归 O(n) */
|
||
void recur(int n) {
|
||
if (n == 1) return;
|
||
return recur(n - 1);
|
||
}
|
||
```
|
||
|
||
=== "C#"
|
||
|
||
```csharp title=""
|
||
int function() {
|
||
// do something
|
||
return 0;
|
||
}
|
||
/* 循环 O(1) */
|
||
void loop(int n) {
|
||
for (int i = 0; i < n; i++) {
|
||
function();
|
||
}
|
||
}
|
||
/* 递归 O(n) */
|
||
int recur(int n) {
|
||
if (n == 1) return 1;
|
||
return recur(n - 1);
|
||
}
|
||
```
|
||
|
||
=== "Swift"
|
||
|
||
```swift title=""
|
||
@discardableResult
|
||
func function() -> Int {
|
||
// do something
|
||
return 0
|
||
}
|
||
|
||
/* 循环 O(1) */
|
||
func loop(n: Int) {
|
||
for _ in 0 ..< n {
|
||
function()
|
||
}
|
||
}
|
||
|
||
/* 递归 O(n) */
|
||
func recur(n: Int) {
|
||
if n == 1 {
|
||
return
|
||
}
|
||
recur(n: n - 1)
|
||
}
|
||
```
|
||
|
||
=== "Zig"
|
||
|
||
```zig title=""
|
||
|
||
```
|
||
|
||
=== "Dart"
|
||
|
||
```dart title=""
|
||
int function() {
|
||
// do something
|
||
return 0;
|
||
}
|
||
/* 循环 O(1) */
|
||
void loop(int n) {
|
||
for (int i = 0; i < n; i++) {
|
||
function();
|
||
}
|
||
}
|
||
/* 递归 O(n) */
|
||
void recur(int n) {
|
||
if (n == 1) return;
|
||
return recur(n - 1);
|
||
}
|
||
```
|
||
|
||
=== "Rust"
|
||
|
||
```rust title=""
|
||
|
||
```
|
||
|
||
## 常见类型
|
||
|
||
设输入数据大小为 $n$ ,常见的空间复杂度类型有(从低到高排列):
|
||
|
||
$$
|
||
\begin{aligned}
|
||
O(1) < O(\log n) < O(n) < O(n^2) < O(2^n) \newline
|
||
\text{常数阶} < \text{对数阶} < \text{线性阶} < \text{平方阶} < \text{指数阶}
|
||
\end{aligned}
|
||
$$
|
||
|
||
![空间复杂度的常见类型](space_complexity.assets/space_complexity_common_types.png)
|
||
|
||
!!! tip
|
||
|
||
部分示例代码需要一些前置知识,包括数组、链表、二叉树、递归算法等。如果你遇到看不懂的地方,可以在学习完后面章节后再来复习。
|
||
|
||
### 常数阶 $O(1)$
|
||
|
||
常数阶常见于数量与输入数据大小 $n$ 无关的常量、变量、对象。
|
||
|
||
需要注意的是,在循环中初始化变量或调用函数而占用的内存,在进入下一循环后就会被释放,即不会累积占用空间,空间复杂度仍为 $O(1)$ 。
|
||
|
||
=== "Java"
|
||
|
||
```java title="space_complexity.java"
|
||
[class]{space_complexity}-[func]{function}
|
||
|
||
[class]{space_complexity}-[func]{constant}
|
||
```
|
||
|
||
=== "C++"
|
||
|
||
```cpp title="space_complexity.cpp"
|
||
[class]{}-[func]{func}
|
||
|
||
[class]{}-[func]{constant}
|
||
```
|
||
|
||
=== "Python"
|
||
|
||
```python title="space_complexity.py"
|
||
[class]{}-[func]{function}
|
||
|
||
[class]{}-[func]{constant}
|
||
```
|
||
|
||
=== "Go"
|
||
|
||
```go title="space_complexity.go"
|
||
[class]{}-[func]{function}
|
||
|
||
[class]{}-[func]{spaceConstant}
|
||
```
|
||
|
||
=== "JS"
|
||
|
||
```javascript title="space_complexity.js"
|
||
[class]{}-[func]{constFunc}
|
||
|
||
[class]{}-[func]{constant}
|
||
```
|
||
|
||
=== "TS"
|
||
|
||
```typescript title="space_complexity.ts"
|
||
[class]{}-[func]{constFunc}
|
||
|
||
[class]{}-[func]{constant}
|
||
```
|
||
|
||
=== "C"
|
||
|
||
```c title="space_complexity.c"
|
||
[class]{}-[func]{func}
|
||
|
||
[class]{}-[func]{constant}
|
||
```
|
||
|
||
=== "C#"
|
||
|
||
```csharp title="space_complexity.cs"
|
||
[class]{space_complexity}-[func]{function}
|
||
|
||
[class]{space_complexity}-[func]{constant}
|
||
```
|
||
|
||
=== "Swift"
|
||
|
||
```swift title="space_complexity.swift"
|
||
[class]{}-[func]{function}
|
||
|
||
[class]{}-[func]{constant}
|
||
```
|
||
|
||
=== "Zig"
|
||
|
||
```zig title="space_complexity.zig"
|
||
[class]{}-[func]{function}
|
||
|
||
[class]{}-[func]{constant}
|
||
```
|
||
|
||
=== "Dart"
|
||
|
||
```dart title="space_complexity.dart"
|
||
[class]{}-[func]{function}
|
||
|
||
[class]{}-[func]{constant}
|
||
```
|
||
|
||
=== "Rust"
|
||
|
||
```rust title="space_complexity.rs"
|
||
[class]{}-[func]{function}
|
||
|
||
[class]{}-[func]{constant}
|
||
```
|
||
|
||
### 线性阶 $O(n)$
|
||
|
||
线性阶常见于元素数量与 $n$ 成正比的数组、链表、栈、队列等。
|
||
|
||
=== "Java"
|
||
|
||
```java title="space_complexity.java"
|
||
[class]{space_complexity}-[func]{linear}
|
||
```
|
||
|
||
=== "C++"
|
||
|
||
```cpp title="space_complexity.cpp"
|
||
[class]{}-[func]{linear}
|
||
```
|
||
|
||
=== "Python"
|
||
|
||
```python title="space_complexity.py"
|
||
[class]{}-[func]{linear}
|
||
```
|
||
|
||
=== "Go"
|
||
|
||
```go title="space_complexity.go"
|
||
[class]{}-[func]{spaceLinear}
|
||
```
|
||
|
||
=== "JS"
|
||
|
||
```javascript title="space_complexity.js"
|
||
[class]{}-[func]{linear}
|
||
```
|
||
|
||
=== "TS"
|
||
|
||
```typescript title="space_complexity.ts"
|
||
[class]{}-[func]{linear}
|
||
```
|
||
|
||
=== "C"
|
||
|
||
```c title="space_complexity.c"
|
||
[class]{hashTable}-[func]{}
|
||
|
||
[class]{}-[func]{linear}
|
||
```
|
||
|
||
=== "C#"
|
||
|
||
```csharp title="space_complexity.cs"
|
||
[class]{space_complexity}-[func]{linear}
|
||
```
|
||
|
||
=== "Swift"
|
||
|
||
```swift title="space_complexity.swift"
|
||
[class]{}-[func]{linear}
|
||
```
|
||
|
||
=== "Zig"
|
||
|
||
```zig title="space_complexity.zig"
|
||
[class]{}-[func]{linear}
|
||
```
|
||
|
||
=== "Dart"
|
||
|
||
```dart title="space_complexity.dart"
|
||
[class]{}-[func]{linear}
|
||
```
|
||
|
||
=== "Rust"
|
||
|
||
```rust title="space_complexity.rs"
|
||
[class]{}-[func]{linear}
|
||
```
|
||
|
||
以下递归函数会同时存在 $n$ 个未返回的 `algorithm()` 函数,使用 $O(n)$ 大小的栈帧空间。
|
||
|
||
=== "Java"
|
||
|
||
```java title="space_complexity.java"
|
||
[class]{space_complexity}-[func]{linearRecur}
|
||
```
|
||
|
||
=== "C++"
|
||
|
||
```cpp title="space_complexity.cpp"
|
||
[class]{}-[func]{linearRecur}
|
||
```
|
||
|
||
=== "Python"
|
||
|
||
```python title="space_complexity.py"
|
||
[class]{}-[func]{linear_recur}
|
||
```
|
||
|
||
=== "Go"
|
||
|
||
```go title="space_complexity.go"
|
||
[class]{}-[func]{spaceLinearRecur}
|
||
```
|
||
|
||
=== "JS"
|
||
|
||
```javascript title="space_complexity.js"
|
||
[class]{}-[func]{linearRecur}
|
||
```
|
||
|
||
=== "TS"
|
||
|
||
```typescript title="space_complexity.ts"
|
||
[class]{}-[func]{linearRecur}
|
||
```
|
||
|
||
=== "C"
|
||
|
||
```c title="space_complexity.c"
|
||
[class]{}-[func]{linearRecur}
|
||
```
|
||
|
||
=== "C#"
|
||
|
||
```csharp title="space_complexity.cs"
|
||
[class]{space_complexity}-[func]{linearRecur}
|
||
```
|
||
|
||
=== "Swift"
|
||
|
||
```swift title="space_complexity.swift"
|
||
[class]{}-[func]{linearRecur}
|
||
```
|
||
|
||
=== "Zig"
|
||
|
||
```zig title="space_complexity.zig"
|
||
[class]{}-[func]{linearRecur}
|
||
```
|
||
|
||
=== "Dart"
|
||
|
||
```dart title="space_complexity.dart"
|
||
[class]{}-[func]{linearRecur}
|
||
```
|
||
|
||
=== "Rust"
|
||
|
||
```rust title="space_complexity.rs"
|
||
[class]{}-[func]{linear_recur}
|
||
```
|
||
|
||
![递归函数产生的线性阶空间复杂度](space_complexity.assets/space_complexity_recursive_linear.png)
|
||
|
||
### 平方阶 $O(n^2)$
|
||
|
||
平方阶常见于矩阵和图,元素数量与 $n$ 成平方关系。
|
||
|
||
=== "Java"
|
||
|
||
```java title="space_complexity.java"
|
||
[class]{space_complexity}-[func]{quadratic}
|
||
```
|
||
|
||
=== "C++"
|
||
|
||
```cpp title="space_complexity.cpp"
|
||
[class]{}-[func]{quadratic}
|
||
```
|
||
|
||
=== "Python"
|
||
|
||
```python title="space_complexity.py"
|
||
[class]{}-[func]{quadratic}
|
||
```
|
||
|
||
=== "Go"
|
||
|
||
```go title="space_complexity.go"
|
||
[class]{}-[func]{spaceQuadratic}
|
||
```
|
||
|
||
=== "JS"
|
||
|
||
```javascript title="space_complexity.js"
|
||
[class]{}-[func]{quadratic}
|
||
```
|
||
|
||
=== "TS"
|
||
|
||
```typescript title="space_complexity.ts"
|
||
[class]{}-[func]{quadratic}
|
||
```
|
||
|
||
=== "C"
|
||
|
||
```c title="space_complexity.c"
|
||
[class]{}-[func]{quadratic}
|
||
```
|
||
|
||
=== "C#"
|
||
|
||
```csharp title="space_complexity.cs"
|
||
[class]{space_complexity}-[func]{quadratic}
|
||
```
|
||
|
||
=== "Swift"
|
||
|
||
```swift title="space_complexity.swift"
|
||
[class]{}-[func]{quadratic}
|
||
```
|
||
|
||
=== "Zig"
|
||
|
||
```zig title="space_complexity.zig"
|
||
[class]{}-[func]{quadratic}
|
||
```
|
||
|
||
=== "Dart"
|
||
|
||
```dart title="space_complexity.dart"
|
||
[class]{}-[func]{quadratic}
|
||
```
|
||
|
||
=== "Rust"
|
||
|
||
```rust title="space_complexity.rs"
|
||
[class]{}-[func]{quadratic}
|
||
```
|
||
|
||
在以下递归函数中,同时存在 $n$ 个未返回的 `algorithm()` ,并且每个函数中都初始化了一个数组,长度分别为 $n, n-1, n-2, ..., 2, 1$ ,平均长度为 $\frac{n}{2}$ ,因此总体占用 $O(n^2)$ 空间。
|
||
|
||
=== "Java"
|
||
|
||
```java title="space_complexity.java"
|
||
[class]{space_complexity}-[func]{quadraticRecur}
|
||
```
|
||
|
||
=== "C++"
|
||
|
||
```cpp title="space_complexity.cpp"
|
||
[class]{}-[func]{quadraticRecur}
|
||
```
|
||
|
||
=== "Python"
|
||
|
||
```python title="space_complexity.py"
|
||
[class]{}-[func]{quadratic_recur}
|
||
```
|
||
|
||
=== "Go"
|
||
|
||
```go title="space_complexity.go"
|
||
[class]{}-[func]{spaceQuadraticRecur}
|
||
```
|
||
|
||
=== "JS"
|
||
|
||
```javascript title="space_complexity.js"
|
||
[class]{}-[func]{quadraticRecur}
|
||
```
|
||
|
||
=== "TS"
|
||
|
||
```typescript title="space_complexity.ts"
|
||
[class]{}-[func]{quadraticRecur}
|
||
```
|
||
|
||
=== "C"
|
||
|
||
```c title="space_complexity.c"
|
||
[class]{}-[func]{quadraticRecur}
|
||
```
|
||
|
||
=== "C#"
|
||
|
||
```csharp title="space_complexity.cs"
|
||
[class]{space_complexity}-[func]{quadraticRecur}
|
||
```
|
||
|
||
=== "Swift"
|
||
|
||
```swift title="space_complexity.swift"
|
||
[class]{}-[func]{quadraticRecur}
|
||
```
|
||
|
||
=== "Zig"
|
||
|
||
```zig title="space_complexity.zig"
|
||
[class]{}-[func]{quadraticRecur}
|
||
```
|
||
|
||
=== "Dart"
|
||
|
||
```dart title="space_complexity.dart"
|
||
[class]{}-[func]{quadraticRecur}
|
||
```
|
||
|
||
=== "Rust"
|
||
|
||
```rust title="space_complexity.rs"
|
||
[class]{}-[func]{quadratic_recur}
|
||
```
|
||
|
||
![递归函数产生的平方阶空间复杂度](space_complexity.assets/space_complexity_recursive_quadratic.png)
|
||
|
||
### 指数阶 $O(2^n)$
|
||
|
||
指数阶常见于二叉树。高度为 $n$ 的「满二叉树」的节点数量为 $2^n - 1$ ,占用 $O(2^n)$ 空间。
|
||
|
||
=== "Java"
|
||
|
||
```java title="space_complexity.java"
|
||
[class]{space_complexity}-[func]{buildTree}
|
||
```
|
||
|
||
=== "C++"
|
||
|
||
```cpp title="space_complexity.cpp"
|
||
[class]{}-[func]{buildTree}
|
||
```
|
||
|
||
=== "Python"
|
||
|
||
```python title="space_complexity.py"
|
||
[class]{}-[func]{build_tree}
|
||
```
|
||
|
||
=== "Go"
|
||
|
||
```go title="space_complexity.go"
|
||
[class]{}-[func]{buildTree}
|
||
```
|
||
|
||
=== "JS"
|
||
|
||
```javascript title="space_complexity.js"
|
||
[class]{}-[func]{buildTree}
|
||
```
|
||
|
||
=== "TS"
|
||
|
||
```typescript title="space_complexity.ts"
|
||
[class]{}-[func]{buildTree}
|
||
```
|
||
|
||
=== "C"
|
||
|
||
```c title="space_complexity.c"
|
||
[class]{}-[func]{buildTree}
|
||
```
|
||
|
||
=== "C#"
|
||
|
||
```csharp title="space_complexity.cs"
|
||
[class]{space_complexity}-[func]{buildTree}
|
||
```
|
||
|
||
=== "Swift"
|
||
|
||
```swift title="space_complexity.swift"
|
||
[class]{}-[func]{buildTree}
|
||
```
|
||
|
||
=== "Zig"
|
||
|
||
```zig title="space_complexity.zig"
|
||
[class]{}-[func]{buildTree}
|
||
```
|
||
|
||
=== "Dart"
|
||
|
||
```dart title="space_complexity.dart"
|
||
[class]{}-[func]{buildTree}
|
||
```
|
||
|
||
=== "Rust"
|
||
|
||
```rust title="space_complexity.rs"
|
||
[class]{}-[func]{build_tree}
|
||
```
|
||
|
||
![满二叉树产生的指数阶空间复杂度](space_complexity.assets/space_complexity_exponential.png)
|
||
|
||
### 对数阶 $O(\log n)$
|
||
|
||
对数阶常见于分治算法和数据类型转换等。
|
||
|
||
例如“归并排序”算法,输入长度为 $n$ 的数组,每轮递归将数组从中点划分为两半,形成高度为 $\log n$ 的递归树,使用 $O(\log n)$ 栈帧空间。
|
||
|
||
再例如“数字转化为字符串”,输入任意正整数 $n$ ,它的位数为 $\log_{10} n$ ,即对应字符串长度为 $\log_{10} n$ ,因此空间复杂度为 $O(\log_{10} n) = O(\log n)$ 。
|
||
|
||
## 权衡时间与空间
|
||
|
||
理想情况下,我们希望算法的时间复杂度和空间复杂度都能达到最优。然而在实际情况中,同时优化时间复杂度和空间复杂度通常是非常困难的。
|
||
|
||
**降低时间复杂度通常需要以提升空间复杂度为代价,反之亦然**。我们将牺牲内存空间来提升算法运行速度的思路称为“以空间换时间”;反之,则称为“以时间换空间”。
|
||
|
||
选择哪种思路取决于我们更看重哪个方面。在大多数情况下,时间比空间更宝贵,因此以空间换时间通常是更常用的策略。当然,在数据量很大的情况下,控制空间复杂度也是非常重要的。
|