Time Complexity

Copyright © 直通硅谷

http://www.zhitongguigu.com/

What is Time Complexity?

  • Find a Number from a Array
    • If the Array has 10 numbers?
    • If the Array has 100 numbers?
    • If the Array has 1,000,000 numbers?
  • We consider the input n is closing to ∞

Copyright © 直通硅谷

http://www.zhitongguigu.com/

Calculate Prime

Given a number n. Is n a prime?

Copyright © 直通硅谷

http://www.zhitongguigu.com/

A Prime Number can be divided evenly only by 1, or itself.
And it must be a whole number greater than 1.

Input? Output?

2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, 43, 47, 53, 59, 61, 67, 71, 73, 79, 83, 89, 97

Calculate Prime

Given a number n. Is n a prime?

Copyright © 直通硅谷

http://www.zhitongguigu.com/

public boolean isPrime(int n) {
    // implement you method here
}

Calculate Prime

Naive Idea

Copyright © 直通硅谷

http://www.zhitongguigu.com/

public boolean isPrime(int n) {
    // Sudo code
    int i from 2 -> n-1;
        if n mod i == 0
        return false
    return true
}

How many calculations?

n - 2

Calculate Prime

Better Idea

Copyright © 直通硅谷

http://www.zhitongguigu.com/

public boolean isPrime(int n) {
    // Sudo code
    int i from 2 -> n / 2;
        if n mod i == 0
        return false
    return true
}

How many calculations?

n / 2 - 2

Calculate Prime

Even Better

Copyright © 直通硅谷

http://www.zhitongguigu.com/

public boolean isPrime(int n) {
    // Sudo code
    int i from 2 -> Math.sqrt(n);
        if n mod i == 0
        return false
    return true
}

How many calculations?

sqrt(n) - 1

Calculate Prime

Why Square Root Works

Copyright © 直通硅谷

http://www.zhitongguigu.com/

If n = a * b, then n is not a prime

If we cannot find such a (b), then n is a prime

If Math.sqrt(n) = m = 3.316, then a <= m <= b

Do we really care about b?

Conclusion: check from 2 to Math.sqrt(n)

Calculate Prime

Copyright © 直通硅谷

http://www.zhitongguigu.com/

Suppose 1 calculation is 1ms

input n (n-2) sqrt(n) - 1: 
10 8ms 2ms
10^10 10^10ms = 115 days 10^5ms = 1.66mins

Calculate Prime

Homework:

Implement the above 2

Copyright © 直通硅谷

http://www.zhitongguigu.com/

Time Complexity

\Theta
Θ\Theta
\Omega
Ω\Omega
O
OO

Copyright © 直通硅谷

http://www.zhitongguigu.com/

Big O notation

Time Complexity

f(n) = n^2 + 2n + 5
f(n)=n2+2n+5f(n) = n^2 + 2n + 5
f(n) \in O(n^3)
f(n)∈O(n3)f(n) \in O(n^3)

Copyright © 直通硅谷

http://www.zhitongguigu.com/

If both        and         are right, we need to give

f(n) \in O(n^2)
f(n)∈O(n2)f(n) \in O(n^2)
O(n^3)
O(n3)O(n^3)
O(n^2)
O(n2)O(n^2)

Different Time Complexity

  • Best Case
  • Worst Case
  • Average Case

Copyright © 直通硅谷

http://www.zhitongguigu.com/

Best Example: Quick Sort

avg: O(n log n)
avg:O(nlogn)avg: O(n log n)
worst: O(n^2)
worst:O(n2)worst: O(n^2)

Different Time Complexity

Copyright © 直通硅谷

http://www.zhitongguigu.com/

Recurrence Equation

  • Merge Sort: T(n) = 2T(n/2) + cn 
  • Binary Search: T(n) = T(n/2) + c
  • Selection Sort: T(n) = T(n-1) + cn
  • Factorial: T(n) = T(n-1) + c

Copyright © 直通硅谷

http://www.zhitongguigu.com/

Recursive deduction

T(n) = T(n-1) + cn

T(n-1) = T(n-2) + c(n-1)

...

T(1) = T(0) + c

T(0) = 0

Copyright © 直通硅谷

http://www.zhitongguigu.com/

Recursive deduction

T(n) = T(n-1) + cn

T(n-1) = T(n-2) + c(n-1)

...

T(1) = T(0) + c

T(0) = 0

T(n) = c(0+1+2+...+n) = O(n^2)
T(n)=c(0+1+2+...+n)=O(n2)T(n) = c(0+1+2+...+n) = O(n^2)

Copyright © 直通硅谷

http://www.zhitongguigu.com/

Master Method

T(n) = aT(\frac{n}{b})+n^c
T(n)=aT(nb)+ncT(n) = aT(\frac{n}{b})+n^c
\text{Compare } log_ba \text{ and } c
Compare logba and c\text{Compare } log_ba \text{ and } c

Copyright © 直通硅谷

http://www.zhitongguigu.com/

https://en.wikipedia.org/wiki/Master_theorem

Master Method

T(n) = aT(\frac{n}{b})+n^c
T(n)=aT(nb)+ncT(n) = aT(\frac{n}{b})+n^c
\begin{cases} & \text{ if } log_ba > c, T(n) = n^{log_ba} \\ & \text{ if } log_ba = c, T(n) = n^clogn \\ & \text{ if } log_ba < c, T(n) = n^c \end{cases}
{ if logba>c,T(n)=nlogba if logba=c,T(n)=nclogn if logba<c,T(n)=nc\begin{cases} & \text{ if } log_ba > c, T(n) = n^{log_ba} \\ & \text{ if } log_ba = c, T(n) = n^clogn \\ & \text{ if } log_ba < c, T(n) = n^c \end{cases}

Copyright © 直通硅谷

http://www.zhitongguigu.com/

Master Method

T(n) = aT(\frac{n}{b})+n^c
T(n)=aT(nb)+ncT(n) = aT(\frac{n}{b})+n^c
\begin{cases} & \text{ if } log_ba > c, T(n) = n^{log_ba} \\ & \text{ if } log_ba = c, T(n) = n^clogn \\ & \text{ if } log_ba < c, T(n) = n^c \end{cases}
{ if logba>c,T(n)=nlogba if logba=c,T(n)=nclogn if logba<c,T(n)=nc\begin{cases} & \text{ if } log_ba > c, T(n) = n^{log_ba} \\ & \text{ if } log_ba = c, T(n) = n^clogn \\ & \text{ if } log_ba < c, T(n) = n^c \end{cases}

Copyright © 直通硅谷

http://www.zhitongguigu.com/

Example: Binary Search T(n) = T(n/2) + m

a = 1, b = 2, c = 0

It fits the second condition. So T(n) = logn

Recurrence Equation

  • Merge Sort: T(n) = 2T(n/2) + cn 
  • Binary Search: T(n) = T(n/2) + c
  • Selection Sort: T(n) = T(n-1) + cn
  • Factorial: T(n) = T(n-1) + c

Copyright © 直通硅谷

http://www.zhitongguigu.com/

O(n log n)
O(nlogn)O(n log n)
O(log n)
O(logn)O(log n)
O(n^2)
O(n2)O(n^2)
O(n)
O(n)O(n)

How to get the recursive equation

  • How do you solve the problem?
  • How do you divide the problem into smaller part? 
  • Before going into smaller parts, how many operation do you need?

Copyright © 直通硅谷

http://www.zhitongguigu.com/

Homework

Count Primes

Copyright © 直通硅谷

http://www.zhitongguigu.com/

Count the number of prime numbers less than a non-negative number, n.

[GoValley-201612] Time Complexity

By govalley201612

Source: ZhiTongGuiGu

[GoValley-201612] Time Complexity

Week 1 Lec 2

  • 697

More from govalley201612