Scribd
Upload
59 views

1 Summary on Θ (n) Sorting Algorithms (Section 7.2) : Lecture Seven

The document summarizes the time complexities of various sorting algorithms like insertion sort, bubble sort, selection sort, and quicksort. It analyzes the best-case, average-case, and worst-case time complexities of insertion sort, which are Θ(n), Θ(n2), and Θ(n2) respectively. For quicksort, the worst-case time complexity is Θ(n2) but the average-case is Θ(n log n). Randomizing the pivot selection makes quicksort's expected worst-case time Θ(n log n) for any input.

Uploaded by

test
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
59 views

1 Summary on Θ (n) Sorting Algorithms (Section 7.2) : Lecture Seven

The document summarizes the time complexities of various sorting algorithms like insertion sort, bubble sort, selection sort, and quicksort. It analyzes the best-case, average-case, and worst-case time complexities of insertion sort, which are Θ(n), Θ(n2), and Θ(n2) respectively. For quicksort, the worst-case time complexity is Θ(n2) but the average-case is Θ(n log n). Randomizing the pivot selection makes quicksort's expected worst-case time Θ(n log n) for any input.

Uploaded by

test
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Lecture Seven

Winter 2015

ITEC2620

Summary on (n2) Sorting Algorithms (Section 7.2)

Insertion Bubble Selection


Best-case
(n)
(n2 )
(n2 )
2
2
Average-case
(n )
(n )
(n2 )
Worst-case
(n2 )
(n2 )
(n2 )

Insertion Sort

Let A[1 . . . n] be an array containing a sequence of n numbers to be sorted. The insertion


sort algorithm is presented below
InsertionSort(A)
1: for j <- 2 to length[A]
2:
key <- A[j]
3:
i <- j-1
4:
while i>0 and A[i] > key
A[i+1] <- A[i]
i <- i-1
5:
A[i+1] <- key

# of times the line executed


n-1+1 = n
n-1
n-1
?
?
?
n-1

Let the cost be the number of times each statement is executed.

2.1

Best-case Complexity

Let tj be the number of times the while loop in Line 4 is executed for that value of j, for
each j = 2, 3, . . . , n, where n = lenght[A], times of executions for line 4 is
n
X

tj = t2 + t3 + . . . + tn .

j=2

If Line 4 is executed tj times, then each of Line 5 and Line 6 will be executed (tj 1) times.
n
n
P
P
That is, times of executions for Line 5 is
(tj 1), and for Line 6 is also
(tj 1). For the
j=2

j=2

best case, we know that the array is already sorted. For each j = 2, 3, . . . , n, we then find
that A[i] key in Line 4 when i has its initial value of j 1. Thus, tj = 1 for j = 2, 3, . . . , n
and the best-case of the times of executions is
T (n) = 3(n 1) + n +

n
X
j=2

= 4n 3 +

n
X

n
n


X
X
tj + 2
(tj 1) note that
(tj 1) = 0
j=2

j=2

j=2

= 4n 3 + n 1
= 5n 4.
Thus T (n) is a linear function, i.e., T (n) O(n).
1

Lecture Seven

Winter 2015

2.2

ITEC2620

Worst-case Complexity

For worst-case, the array is in reversely sorted order (decreasing order). We must compare
each element A[j] with each element in the entirely sorted subarray A[1 . . . j 1], and so
tj = j for j = 2, 3, . . . , n. Hence, the worst-case running time is
T (n) = 3(n 1) + n +

n
X

n
X
tj + 2
(tj 1)

j=2

= 4n 3 +

n
X

j+2

j=2

j=2
n
X

(j 1)

j=2

(n 1)n
n(n + 1)
1 + 2(
)
2
2
n2 n
+ 1 + n2 n
= 4n 3 +
2
2
3n2 7n
=
+
4
2
2
= 4n 3 +

Hence, in worst-case, T (n) O(n2 ).

QuickSort (Section 7.5)

We first review what is lower bounds and upper bound. Remember that worst-case running time T (n) = max{t(x) : x has size n}. An upper bound is a function g(n) s.t.
f (n) O(g(n)) (essentially, T (n) g(n)). A lower bound on T (n) is a function f (n)
s.t. T (n) (f (n) (essentially, T (n) f (n)).Generally, to prove that g(n) is an upper
bound on T (n), we argue that for all x of size n, t(x) g(n). Generally, to prove that f (n)
is a lower bound on T (n), we exhibit one specific input x, for which we can show t(x) f (n).
The average-case running time of an algorithm will always be less than or equal to the worstcase by definition. we have already shown an example (ListSearch in Lecture Two) where the
two are asymptotically the same. As we will see below, that they can also be asymptotically
different, with the average-case becoming much better.
QuickSort(S) (S is the list to sort)
1: if |S| <=1 return S
2: select pivot p in S
3: partition elements of S into:
L = elements of S less than p
E = elements of S equal to p
U = elements of S greater than p
4: return QuickSort(L); E; QuickSort(U)
Standard QuickSort: we select the pivot by simply letting p be the first element in S.
Example 1. S = [15, 25, 12, 2, 37].
p = 15, L =[12, 2]; M =15; U =[25, 37]
p = 12, L=[2]; M=12; U=[];
p = 25, L=[]; M=[25]; U=[37];
2, 12, 15, 25, 37.
2

Lecture Seven

Winter 2015

3.1

ITEC2620

Worst-case Complexity

How many comparisons are performed in the worst-case on a list of size n?


Upper bound: We show T (n) O(n2 ). All of the comparisons occur between the pivot and
other elements in the list. Each element will be compared at most once with every other
element. Hence, there can be no more comparisons than there are pairs of elements in the
list, i.e.,
 
n
n(n 1)
T (n)
=
2
2
Lower bound: We show T (n) (n2 ). Consider the input S = [n, n 1, n 2, . . . , 2, 1]. 1st
call: n 1 comparisons with pivot n;
2nd call: n 2 comparisons with pivot n-1;
3rd call: n 3 comparisons with pivot n-2;
...
...
...
...
. . .;
(n 1)st call: 1 comparisons with pivot 2.
.
Total is (n 1) + (n 2) + . . . + 1 = n(n1)
2
2
Conclusion: T (n) (n ), more precisely, T (n) =

n(n1)
.
2

Where is the quickness in QuickSort?

3.2

Average-case Complexity of QuickSort (Section 14.2.4)

First we define the sample space Sn of all inputs of size n:


Sn = {all permutation of [1, 2, 3, . . . n]},
assuming a uniform probability distribution on Sn .
Next let tn (S) be the number of comparisons performed by QuickSort on input S Sn . We
want to compute T 0 (n) = E(tn (S); S Sn ].
Observation:
1. In the first call, each i {1, 2, . . . , n}; in equally likely to be the pivot.
2. After the partition, L is a random permutation of [1, 2, . . . , i 1] and u is a random
permutation of [i + 1, i + 2, . . . , n].
This allow us to write: if the pivot is i, then
T 0 (n) = n 1 + T 0 (n 1) + T 0 (n i)
n1

2X 0
T (j) ()
T (n) = n 1 +
n j=0
0

n2

2 X 0
T (n 1) = n 2 +
T (j)()
n 1 j=0
0

Lecture Seven

Winter 2015

ITEC2620

n1

nT 0 (n) = n(n 1) + n

2X 0
T (j)(from )
n j=0
n2

2 X 0
(n 1)T (n 1) = (n 1)(n 2) + (n 1)
T (j)(from )
n 1 j=0
0

nT 0 (n) (n 1)T 0 (n 1) = 2(n 1) + 2T 0 (n 1),


nT 0 (n) = (n + 1)T 0 (n 1) + 2(n 1),
Divide both sides by n(n + 1)
T 0 (n 1) 2(n 1)
T 0 (n)
=
+
n+1
n
n(n + 1)
let A(n) =

T 0 (n)
,
n+1

we have
A(n) = A(n 1) +

and
A(0) =

2(n 1)
n(n + 1),

T 0 (0)
= 0.
1

hence,
A(n) =

n
X
2(i 1)
i=1

=2

i(i + 1)
n
X
i=1

n
X
i=1

X
2i
2

i(i + 1) i=1 i(i + 1)


n

X 1
1
1
2
(
)
(i + 1)
i
i
+
1
i=1

= 2(log n) 2(1

1
= (log n).
n + 1)

Hence,
T 0 (n) = (n 1)A(n) = (n log n).
Note that, The sum of reciprocals from 1 to n, is called the Harmonic Series and written as
Hn , has a value of loge n + O(1).

3.3

Randomized QuickSort

How can we make QuickSort quick for all inputs?


Average-case analysis works only if each permutation equally likely. In practice, this
may not be true.
Instead of relying on unknown distribution of inputs, randomize algorithm by picking
random element as pivot. This way, random behaviour of algorithm on any fixed
input is equivalent to fixed behaviour of algorithm on a uniformly random input, i.e.,
expected worst-case time of randomized algorithm on every single input is (n log n).
In general, randomized algorithms are good when there are many good choices but it
is difficult to find one choice that is guaranteed to be good.

You might also like