Chapter 23 Algorithm Efficiency Liang 8th | Time Complexity | Logarithm

Chapter 23 Algorithm Efficiency

23.1 introduction
Two algorithnms perform the same tasks, such as a search (linear search vs. binary
search), or sort (selection sort vs. insertion sort).
We want to know which algorithm is better?
We then implement these algorithms in Java and run them to get the execution times.
This approach is problematic for two reasons:
1. Many tasks run concurrently on a computer. Execution of a particular program
depends on system load.
2. Execution time depends on specific input. For example, linear search and binary
search, if an element to be searched happens to be the first in the list, linear
search will find the element quicker than binary search.

It is difficult to compare algorithms using their execution time.

A theoretical approach was developed to analyze algorithms independent of
computers and specific input. This approach approximates the effects of a change
on the size of the input. This allows us to see how fast an algorithm’s execution
time increases as the input size increases. Thus, you compare two algoritms by
examining their growth rates.

23.2 Big O Notation
The linear search algorithm compares the key with the elements in the array
sequentially until the key is found, or the array is exhausted.
If the key is not in the array, it requires n comparisons for an array of size n.
If they key is in the array, it requires n/2 comparisons on average.
The algorithm’s execution time is proportional to the size of the array.
Doubling the size of the array, doubles the number of comparisons, so the
algorithms grows at a linear rate. Growth rate has an order of magnitude of n.
Computer Scientists us Big O notation to represent “order of magnitude”.
Using this notation, the complexity of the linear search algorithm is O(n),
pronounced as “order of n.”
For the same input size, an algorithm’s execution time may vary, depending on the
input. Input resulting in the shortests execution time is called best-case inpu.
Input resulting in the longests execution time is the worst-case input.
Best-case, and worst-case are not representative, but worst-case analysis is very
useful. Algoritms will never be slower than the worst case.
An aeverage-case analysis attemtps to determine the average amount of time
among all possible inputs of the same size.
Average case analysis is ideal, but difficult to perform because for many problems, it
is difficult to determine the relative probabilities and distributions of various input
Worst case analysis is easier to perform, so the analysis is generally conducted for
the worst case.
Linear search algorithms requires n comparisons in the worst case, and n/2
comparisons in the average case if you are nearly always looking for something
known to be in the list.
Using Big O notation, both cases require O(n) time. The multiplicative constant ½
can be omitted.
Algorithms analysis is focused on growth rate, so multiplicative constants have no
impact on growth rates. Growth rate for n/2 or 100n is the same as for n. Thus,
O(n) = O(n/2) = O(100n).

Consider the algorithm to find the maximum number in an array of n elements. To
find the maximum number if n is 2, it takes one comparison; if n is 3 it takes two
comparisons. Generally, it takes n – 1 comparisons to find the maximum number in
a list of n elements.
Algorithm anslysis is for large input size. If the input size is small, there is no
significance in estimating an algorithm’s efficiency.
As n grows larger, the n part in the expression n – 1 dominates the complecity.
Big O notation allows to ignore the nondominanting part, e.g. the – 1 in the
expression n – 1, and highlights the important part, e.g. n in the expressio n – 1.
So the complecity of this algorithmis O(n).
Big O notation estimates execution time of an algorithm in relation to tehinput size.
If time is not reltaed to input size, algorithm is said to take constant time with
notation O(1). For example, a method that retrieves an element at a given index in
an array takes constant time because the time doesnot grow as the size of the array
increases. The following mathematical summations are often useful in algorithm
1 + 2 + 3 + … + (n – 1) + n = n(n+1)/2.
A^0 + a^1 + a^3 + … + a^(n-1) + a^n = a^n+1 – 1 / a-1
2^0 + 2^1 + 2^2 + 2^3 + … 2^(n-1) + 2^n = 2^(n+1 – ½ - 1 = 2^n+1 – 1.

23.3 Examples: Determining Big O
Section gives examples determining Big O for repetitions, sequences, and selection

Consider time complexity for the following loop:
For ( I = 1; I <= n; i++){
K = k + 5;
It is a constant time, c, for executing
K = k + 5;
Since the loop executes n times, time complecity for loop is T(n) = (a constant c)* n
= O(n).

Example 2
What is the time complexity of the following loop?
For (I = 1; i<= n; i++){
For (j = 1; j<= n; J++){
K = k + I + j;
It is a constant time c for executing
K = k + I + j;
Outer loop executes n times. For each iteration in the outer loop, the inner-loop is
executed n times. So, the time complexity for the loop is
T(n) = (a constant c) * n * n = O(n^2)

Algorithms with O(n^2) time complecity are called quadratic algorithms. Quagratic
algorithms grow quickly as the problem size increases. Doubling input size
quadruples the algorithm’s time. Algorithms with nested loops are often quadratic.

Example 3
Consider the followin gloops:
For( I = 1; I <= n; i++) {
For (j = 1; j<= I; j++){
k = k+1+j;

Outer loop executes n times. For I = 1, 2, …, inner loop executes one time, two
times, and n times. So time complexity for loop is
T(n) = c + 2c + 3c + 4c + … + nc = cn(n + 1)/2
= (c/2)n^2 + (c/2)n
= O(n^2)
Example 4
Consider the following loop
For ( I = 1; I <= n ; I ++) {
For (j = 1; j <=20; j++) {
K = k + I + j;
Inner loop executes 20 times, outer loop n times. So time complexity for loop is T(n)
= 20* c * n = O(n).

Example 5
Consider the following sequences:
For (j = 1; j <= 10; j++){
K = k + 4;
For (I = 1; I <=n; I++){
K = k+i+j;
First loop executes 10 times, second loop 20* n times. Time complexity for loop is
T(n) = 10* c + 20* c* n = O(n)

Example 6
Consider the following selection statement:
If (list.contains(e)){
For (Object t: list){
Suppose list contains n elements. Execution time for list.contains(e) is O(n). Loop
in the else clause takes O(n) time. Time complexity for entire statement is
T9n) = if test time + worst case time (if clause, else clause)
= O(n) + O(n) = O(n).

Example 7
Consider the computation a^n for an integer n. A simple algoritm would multiply a
n times, as follows:
Result = 1;
For (int I = 1; I <= n ; i++)
Result *= a;
Algorithm takes O(n) time. Without loss of generality, assume n = 2^k. You can
improve the algorithm using the following scheme:
Result = a
For (int I = 1; i<=k i++)
Result = result * result;
Algorithm taes O(logn) time. For an arbitray n, you can revise algorithm and prove
that the complexity is still O(logn).

23.4 Analyzing Algorithm Time Complexity
Analyzing various algorithms will follow.
23.4.1 Analyzing Binary Search
Binary search searches a key in a sorted array. Each iteration in the algorithm
contains a fixed number of operations, denoted by c. Let T(N) denote the time
compelxity for a binary search on a listof n elements. Without loss of generaility,
assume n is a power of 2 and k = log n. Binary search eliminates half of input after
two comparisons,

T(n) = T(n/2) + c = T(n/2^2) + c + c = T(n/2^k) + kc
= T(1) + c log n = 1 + (log n)c
=O(log n).

Ignoring constants and nondominating terms, complexity for binary search algorithm
is O(log n). Algorithms with O(log n) time complexity are logarithmic algorithms.
Base of log is 2, but base does not affect logarithmic growth rate, so it can be
omitted. Logarithmic algorithms grows slowly as problem size increases. Squaring
input size results in double the time for the algorithm.
23.4.2 Analyzing Selection Sort
Selection sort finds the smallest number in a list, and places it first. Then, selection
sort finds the smallest number remaining and places it after the the first, and so
son, until list contains only a single number. The number of comparirons is n-1 for
the first iteration, n-2 for the secon iteration, and so on. Let T(n) denote the
complexity for selection sort, and c denote the total number of other operations
such as assignments and additional comparisons in each iteration. So,

T(n) = (n-1) + c + (n-2) + c + … + 2 + c + 1 + c
= (n-1)(n-1+1)/2 + c(n-1) = n^2/2 – n/2 + cn-c
= O(n^2).
Thus, complexity of the selection sort algorithm is O(n^2).
23.4.3 Analyzing Insertion Sort
Insertion sort algorithm sorts a lsit of values by repeatedly inserting a new element
into a sorted partial array until the whole array is sorted. At the kth iteration, to
insert an element into an array of size k, it may take k comparisons to find the
insertion position, and k moves to insert the element. Let T(n) denote complexity
for insertion sort and c denote total number of other operations such as
assignments and additional comparisons in each iteration. So,

T(n) = (2 + c) + (2 x 2 + c) + … + (2 x (n-1) + c)
= 2(1 + 2 + … + n-1) + c(n-1)
= 2(n-1)n/2 + cn – c = n^2 – n + cn-c
= O(n^2).
Thus, complexity for insertion sort algorithm is O(n^2). So selection sort and
insertion sort are same time complexity.

23.4.4 Analyzing Towers of Hanoi Problem
Towers of Hanoi recursively moves n disks from tower A to tower B with assistance
from tower C as follows:
1. Move the first n – 1 disks from A to C with assistance of tower B.
2. Move disks n from A to B.
3. Move n – 1 disks from C to B with assistance of tower A.

Let T(n) denote complexity for algorithm that moves n disks and c denote the
constant time to move on disk: i.e., T(1) is c. So,

T(n) = t(n – 1) + c + T(n – 1)
= 2T(n-1) + c
= 2(2T(n-2) + c ) + c
=2(2(2T(n-3) + c) + c) + c
2^n-1T(1) + 2^n-2c + … + 2c + c
= 2^n-1c + 2^n-2c + … + 2c + c = (2^n – 1)c = O(2^n).

Algorithms with O(2^n) are exponential algorithms. As input increases, time for
exponential algorithm grows exponentially. Exponential algoriwhtms are not
practical for large input sizes.

23.4.5 Comparing Common Growth Functions
The following are functions ordered from least to greatest showing various time
complexity arguments.

O(1) < O(log n) < O(n) < O(n log n) < O(n^2) < O(n^3) < O(2^n)

Function Name n = 25 n = 50 F(50)/f(25)
O(1) constant time 1 1 1
O(log n) Logarithmic time 4.64 5.64 1.21
O(n) Linear time 25 50 2
O(n log n) Long-linear time 116 282 2.43
O(n^2) Quadratic Time 625 2500 4
O(n^3) Cubic Time 15625 125000 8
O(2^n) Exponential Time 3.36 x 10^7 1.27 x 10^15 3.35 x 10^7

23.5 Case Studies: Finding Fibonacci Numbers
Here is a recursive method for finding the Fibonacci numbers…

Left off here.

Sign up to vote on this title
UsefulNot useful

Master Your Semester with Scribd & The New York Times

Special offer for students: Only $4.99/month.

Master Your Semester with a Special Offer from Scribd & The New York Times

Cancel anytime.