hw03 Solution PDF

Fundamental Algorithms
CSCI-GA.1170-001/Summer 2016
Solution to Homework 3
Problem 1. (1 point) Illustrate the operation of heapsort on the array:
A = (19, 2, 11, 14, 7, 17, 4, 3, 5, 15)
By showing the values in array A after initial heapification and after each call to max-heapify.
Solution:
A
A
A
A
A
A
A
A
A
A
=
=
=
=
=
=
=
=
=
=
[19 , 15 , 17 , 14 , 7, 11 , 4, 3, 5, 2]
[17 , 15 , 11 , 14 , 7, 2, 4, 3, 5, 19]
[15 , 14 , 11 , 5, 7, 2, 4, 3, 17 , 19]
[14 , 7, 11 , 5, 3, 2, 4, 15 , 17 , 19]
[11 , 7, 4, 5, 3, 2, 14 , 15 , 17 , 19]
[7 , 5, 4, 2, 3, 11 , 14 , 15 , 17 , 19]
[5 , 3, 4, 2, 7, 11 , 14 , 15 , 17 , 19]
[4 , 3, 2, 5, 7, 11 , 14 , 15 , 17 , 19]
[3 , 2, 4, 5, 7, 11 , 14 , 15 , 17 , 19]
[2 , 3, 4, 5, 7, 11 , 14 , 15 , 17 , 19]
#
#
#
#
#
#
#
#
#
#
After
After
After
After
After
After
After
After
After
After
initial heapification
extracting 19
extracting 17
extracting 15
extracting 14
extracting 11
extracting 7
extracting 5
extracting 4
extracting 3
Problem 2. (1 point) Illustrate the operation of counting sort on the array:

A = (4, 6, 3, 5, 0, 5, 1, 3, 5, 5)
By showing the values in array C after each loop, and the final array B.
Solution:
C
C
C
C
=
=
=
=
0
[0 ,
[1 ,
[1 ,
[0 ,
1
0,
1,
2,
1,
2
0,
0,
2,
2,
3
0,
2,
4,
2,
4
0,
1,
5,
4,
5
0,
4,
9,
5,
6
0]
1]
10]
9]
#
#
#
#
After
After
After
After
initialization
counting elements
computing running sum
producing sorted B
1 2 3 4 5 6 7 8 9 10
B = [0 , 1, 3, 3, 4, 5, 5, 5, 5, 6] # Sorted B
Problem 3. (1 point) Illustrate the operation of radix sort on the array:
A = (392, 517, 364, 931, 726, 912, 299, 250, 600, 185)
By showing the values in array A after each intermediate sort.
Solution:
A = [392 ,
517 ,
364 ,
[250 ,
600 ,
931 ,
[600 ,
912 ,
517 ,
[185 ,
250 ,
299 ,
1
931 ,
726 ,
912 ,
299 ,
250 ,
600 ,
185]
->
392 ,
912 ,
364 ,
185 ,
726 ,
517 ,
299]
->
726 ,
931 ,
250 ,
364 ,
185 ,
392 ,
299]
->
364 ,
392 ,
517 ,
600 ,
726 ,
912 ,
931]
Problem 4. (1 point) Illustrate the operation of bucket sort on the array:

A = (0.88, 0.23, 0.25, 0.74, 0.18, 0.02, 0.69, 0.56, 0.57, 0.49)
By showing the final array B of sorted buckets.
Solution:
B = {0:
1:
2:
3:
4:
5:
6:
7:
8:
9:
[0.02]
[0.18]
[0.23 , 0.25]
[]
[0.49]
[0.56 , 0.57]
[0.69]
[0.74]
[0.88]
[]}
Problem 5. (3 points) Consider a d-ary heap a generalization of a binary heap, in which all
internal nodes (with at most one exception) have d children. Design a method to store a d-ary
heap in an array and derive the expressions for:
(a) Parent of i-th node in a d-ary heap.
(b) j-th child of i-th node in a d-ary heap.
(c) Maximum number of nodes of height h in any n-element d-ary heap.
(d) Maximum height of any n-element d-ary heap.
Solution:
(a) Parent of i-th node in a d-ary heap.
(b) j-th child of i-th node in a d-ary heap.
Assuming 1-based indexing, we can extend the standard level-wise method of storing
binary heaps to d-ary heaps by taking:
i1
,
par ent(i) =
d
child(i, j) = d(i 1) + j + 1.
2
(c) Maximum number of nodes of height h in any n-element d-ary heap.

n1
As par ent(i) = i1
,
the
parent
of
the
last
node
is
par
ent(n)
=
. Defining nh as
d
d
the number of nodes of height h, and observing that the parent of the last node is the
last non-leaf node, we can express the number of leaf nodes as:
n1
.
n0 = n
d
Removing all leaf nodes from the heap of size n would result in a heap of size n0 :
n1
0
n = n n0 =
d
0
with n0 leaves:
0
n 1
0
0
n0 = n
d
n1
1
n1
d
=
d
d
n1
n1d
=
.
d
d2
Removing all leaf nodes from the heap of size n0 would result in a heap of size n00 :
n00 = n0 n00

n1
n1d
n1
+
=
d
d
d2
n1d
=
d2
with n000 leaves:
n00 1
=n
d
n1d
1
n1d
d2
=
d2
d

n1d
n 1 d d2
,
=
d2
d3
suggesting a general expression for a number of leaves in a heap obtained from a heap
of size n by removing all leaf nodes k times:

n 1 d ... d k1
n 1 d ... d k
(k)
n0 =
dk
d k+1
Pk
Pk1 i
n i=0 d
n i=0 d i
=
dk
d k+1
k
k+1
1
n dd1
n d d11
=
for d > 1.
dk
d k+1
n000
00
We now observe that the leaves in heap H 0 , obtained from heap H by removing all leaf
nodes, had height 1 in the original heap H. The leaves in heap H 00 , obtained from heap
H 0 by removing all leaf nodes, had height 1 in H 0 and height 2 in H. In general, nh the
number of nodes of height h in an n-element d-ary heap H is the number of leaves in
H (h) , obtained from H by removing all leaf nodes h times:

h+1
h
1
n d d11
n dd1
for d > 1.
nh =
dh
d h+1
(d) Maximum height of any n-element d-ary heap.
We notice that a maximum number of nodes in a d-ary heap of height h is observed when
the last level is full, and a minimum number of nodes when the last level has only one
node:
d h+1 1
for d 6= 1,
d 1
dh 1
nmin (h) = d 0 + d 1 + ... + d h1 + 1 =
+ 1 for d 6= 1.
d 1
nma x (h) = d 0 + d 1 + ... + d h1 + d h =
This allows us to provide a tight bound on the size of a d-ary heap of height h (we assume
d 6= 1 for the rest of this problem):
d h+1 1
dh 1
+1 n
.
d 1
d 1
With integer n we can get rewrite and then reduce the inequality as follows:
dh 1
d h+1 1
<n
,
d 1
d 1
d h 1 < n(d 1) d h+1 1,
d h < n(d 1) + 1 d h+1 ,
h < logd (n(d 1) + 1) h + 1.
For integer n and real x:
n 1 < x n iff
n = dxe,
and so:

h = logd (n(d 1) + 1) 1.
Problem 6. (4 points) Consider a min-priority queue representing a set of integers and supporting the following operations: insert(k) to insert element with value k, get-min() to return
the minimum element, and extract-min() to remove and return the minimum element.
Give pseudocode and worst-case running time for each operation, assuming the priority queue
is implemented with the following data structures:
4
(a) Unordered array.

(b) Ordered array.
(c) Unordered linked list.
(d) Ordered linked list.
(e) Min-heap.
Solution:
(a) Unordered array of size n referenced by variable array:
insert(k): Without the need to maintain order, we can insert in O(1) by adding
the new element to the end of the array.
def insert (k ):
array . append (k )
We note that the true worst case is O(n) and is observed when the array needs to
be reallocated (though amortized running time is still constant).
get-min(): With no order to rely upon, we have to resort to a O(n) linear scan.
def get - min ():

assert len ( array ) > 0
index = find - min ( array ) # O(n)
return array [ index ]
extract-min(): As with get-min(), we have no way around a O(n) linear scan.
We dont, however, need to spend another O(n) closing the gap left after removing
the element, as with no order to worry about we can just fill the hole with the last
element.
def extract - min ():

index = find - min ( array ) # O(n)
value = array [ index ]
exchange ( array [ index ], array [ len ( array ) - 1])
array . delete ( len ( array ) - 1) # O(1)
return value
(b) Ordered array of size n referenced by variable array:
insert(k): One way to insert the element into a sorted array is to use modified
binary search to find correct position in O(lg n) time, then insert and shift the elements in O(n). An alternative is to "drag" the element into its correct position by
successive exchanges (not unlike insertion sort). Both ways are O(n) in the worst
case.
def insert (k ):
index = find - insert - position ( array , k) # O(lg n)
5
array . insert (k , index ) # O(n)

get-min(): It is more convenient to represent a min-heap with an array sorted
in non-increasing order (so we can delete the minimum element in O(1)). The
minimum element is then in the last position and can be accessed in O(1).
def get - min ():

return array [ len ( array ) - 1] # O(1)
extract-min(): Deleting the last element does not require closing any gaps and
can be done in O(1):

value = array [ len ( array ) - 1] # O(1)
array . delete ( len ( array ) - 1) # O(1)
return value
(c) Unordered linked list of size n referenced by variable list:
insert(k): Without the need to maintain order, we can insert in O(1) by adding
the new element to the head of the list.
def insert (k ):
list . insert (k , list . head ())
We note that the linked list data does not have to be contiguous and we are unaffected by the arrays "true worst case" when the storage needs to be reallocated.
get-min(): As with the unordered array, we have to resort to a O(n) linear scan.
def get - min ():

return find - min ( list ) # O(1)
extract-min(): As with get-min(), we have no way around a O(n) linear scan.
There are, however, no gaps to be closed we can delete from any position in O(1).

# O(n) to find , O(1) to remove
return find - and - remove - min ( list )
(d) Ordered linked list of size n referenced by variable list:
insert(k): Linked lists do not support indexing, so we cannot use binary search to
find the insertion position and have to resort to a linear scan. Once the position is
found, however, we can insert in O(1) (assuming the position is given as a reference,
not an index):
def insert (k ):
ref = find - insert - position ( list , k) # O(n)
list . insert (k , ref ) # O(1)
6
get-min(): Assuming the list is sorted in a non-decreasing order, the minimum

element is the head of the list and can be accessed in O(1):
def get - min ():

return list . head () # O(1)
extract-min(): We can delete at any position in O(1):

min = list . head () # O(1)
remove - head ( list ) # O(1)
return min
(e) Min-heap referenced by variable heap:
insert(k): We can insert an element into the heap in O(lg n) by adding it as the last
node and then percolating it up until the min-heap property is no longer broken:
def insert (k ):
++ heap . size
i = heap . size
heap [i] = k
while i > 0 and heap [ parent (i )] > heap [i ]: # O(lg n)
exchange ( heap [ parent (i )] , heap [i ])
i = parent (i)
get-min(): The minimum element in a min-heap is at the root and can be accessed
in O(1):
def get - min ():

return heap [0] # O(1)
extract-min(): We can remove the minimum element from the heap by replacing
the root element with the last element, excluding the last element from the heap,
and rebuilding the heap, all in O(lg n):

min = heap [0]
heap [0] = heap [ heap . size - 1]
-- heap . size
min - heapify ( heap , 0) # O(lg n)
return min

hw03 Solution PDF

Uploaded by

Copyright:

Available Formats

hw03 Solution PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

hw03 Solution PDF

Uploaded by

Copyright:

Available Formats

Fundamental Algorithms

Problem 2. (1 point) Illustrate the operation of counting sort on the array:

Problem 4. (1 point) Illustrate the operation of bucket sort on the array:

(c) Maximum number of nodes of height h in any n-element d-ary heap.

nma x (h) = d 0 + d 1 + ... + d h1 + d h =

(a) Unordered array.

def get - min ():

def extract - min ():

array . insert (k , index ) # O(n)

def get - min ():

def extract - min ():

def get - min ():

def extract - min ():

get-min(): Assuming the list is sorted in a non-decreasing order, the minimum

def get - min ():

def extract - min ():

def get - min ():

def extract - min ():

You might also like