Data Structures

Introduction to Data Structures
Complexity, Rate of Growth,

Big O Notation
Arrays
Linked List
Stacks, Queues, Recursion
Sorting and Searching
Techniques
Hashing Techniques
MSc. In Software
1
INTRODUCTION TO
DATA Structures
!
The Need to learn Data structures
Coding, Testing And Refinement
Pitfalls
Basic Terminology; Elementary Data
Data Structures
Data Structures Operations
Complexity, Time-Space Tradeof
Summary
MSc. In Software
THE NEED TO LEARN DATA STRUCTURES
ome of the rules for good programming are:

1. Always name your variables and functions with the greatest care and
explain them thoroughly.
2. The reading time for programs is much more than the writing time. Write
clearly to enable easy reading.
3. Each function should do only one task. Ensure that it is done well.
4. Each function should hide something.
5. Keep your connections simple. Avoid global variables whenever possible. If
you must use global variables as input, document them thoroughly.
6. Never code until the specifications are precise and complete.
This subject describes programming methods and tools that will prove effective
for projects.
Food for thought:
How do you rewrite the following function so that it accomplishes the same result
in a less tricky way?
Doessomething(int *first,int *second)
{
*first = *second - *first;
*second = *second - *first;
*first = * second + *first;
}
A probable solution is like this:
Doessomething(int *first, int *second)
{
* first = 2(*second - *first);
}
By now we know that the computer understands logical commands. The
instructions given to the computer must be in a very structured form. This
structured form is called Algorithm in computer jargon. The algorithm is a
representation of any event in a stepwise manner. In some cases these sequence
of activities are quite simple and the algorithm can be easily constructed. But
when the problem at hand is quite complex, and a lot of different activities have
to be considered within a single problem, keeping track of all these events and the
variables they involve becomes a very tedious task. To manage and handle these
MSc. In Software
events and variables in a more structured and orderly manner we take the aid of
data structures.
Food for thought:
Rewrite the following function with meaningful variables, with better format and
without unnecessary variables.
#define MAXINT 100
int calculate(int apple, int orange)
{ int peach, lemon;
peach = 0; lemon = 0; if (apple < orange) {
peach = orange;} else if (orange <= apple) {
peach = apple;} else { peach = MAXINT; lemon = MAXINT;
} if (lemon != MAXINT) {return (peach);}}
A probable solution:
Do it yourself. If you cannot, email to us.
The need for using data structures arises from the fact that it teaches how to
write a function with meaningful variable names, without extra variables that
contribute nothing to the understanding, with a better layout and without
redundant and useless statements.
CODING, TESTING AND REFINEMENT

The three processes in the title above go hand-in-hand and must be done
together. Yet it is important to keep them separate in our thinking, since each
requires its own approach and method.
Coding is the process of writing an algorithm in the correct syntax (grammar) of
a computer language like C.
Testing is the process of running the program on simple data to find errors if
there are any.
Refinement is done on the basic program.
After coding the main program, most programmers wish to complete the writing
and coding of the functions as soon as possible, to check if the full project works.
But even for small projects, there are good reasons for debugging functions one
at a time.
PITFALLS
1. Be sure you understand your problem before you decide how to solve it.
2. Be sure you understand the algorithmic method before you start to
program.
3. In case of difficulty, divide the problem into pieces and think of each part
separately.
MSc. In Software
4. Keep your functions short and simple.

5. Keep your programs well formatted as you write them it will make
debugging much easier.
6. Keep your documentation consistent with your code, and when reading a
program make sure that you debug the code and not just the comments.
7. Explain your program to somebody else. Doing so will help you understand
it better.
8. To compile the program correctly, there must be something in the place of
each function that is used, and hence we must put in short, dummy
functions, called stubs. This makes debugging easier.
BASIC TERMINOLOGY; ELEMENTARY DATA

Try to remember some basic terminologies, which we will be using throughout the
subject.
Data Values or sets of values
Data-items
Refers to single unit of values
Group-items
These are data-items that can be divided into sub-items
Elementaryitems - These are data items that cannot be divided into sub-items
An entity is something that has certain attributes or properties that may be
assigned values. These values may be numeric or nonnumeric.
For better understanding of the concept, look at the example given below. The
following are the possible attributes and their corresponding values for an entity,
an employee of a given organization.
Attributes:
Values:
Name
XYZ
Age
34
Sex
M
Emp. No.
134
Entity Set : Entities with similar attributes (e.g., all the employees in an
organization) form an entity set.
Range of Values : Each attribute of an entity set has a range of values, the set
of all possible values that could be assigned to the particular attribute.
Information : The term information is sometimes used for data with given
attributes, or, in other words, meaningful or processed data.
The way data is organized into the hierarchy of fields, records and files reflect the
relationship between attributes, entitles and entity sets.
Field Single elementary unit of information representing an attribute of an
entity
Record - Collection of field values of a given entity
File Collection of records of the entities in a given entity set
Primary key : Each record in a file may contain many field items, but the value
in a certain field may uniquely determine the record in the file. Such a field K is
MSc. In Software
called a primary key, and the values K1, K2,. in such a field are called keys or
key values.
Consider these cases to understand Primary key better:
(1) Suppose an automobile dealership maintains a file where each record contains
the following data:
Serial Number
Type
Year
Price
Accessories
The Serial Number field can serve as a primary key for the file, since each
automobile has a unique serial number.
(2) Suppose an organization maintains a membership file where each record
contains the following data:
Name
Address
Telephone Number
Dues Owed
Although there are four data items, Name cannot be the primary key, since
more than one person can have the same name. Name and Address may be
group items and together can serve as a primary key. Note also that the
Address and Telephone Number fields may not serve as primary keys, since
some members may belong to the same family and have the same address
and telephone number. Dues Owed is out of the question because many
people can have the same value.
The above examples must have cleared your doubt about key words, which we
are going to use.
Records may also be classified according to length. A file can have fixed-length
records or variable-length records.
Fixed-length records All the records contain the same data items with the
same amount of space assigned to each data item.
Variable-length records - File records may contain different lengths.
The above can be explained as: Student records usually have variable lengths,
since different students take a varying number of courses. Usually, variablelength records have a minimum and maximum length.
MSc. In Software
The study of such data structures includes the following three steps:
(1)
(2)
(3)
Logical or mathematical description of the structure

Implementation of the structure on a computer
Quantitative analysis of the structure, which includes determining
the amount of memory needed to store the structure and the time
required to process the structure
DATA STRUCTURES
The logical or mathematical model of a particular organization of data is called a
data structure. Here we will discuss three types of data structures in detail. They
are: arrays, link list, and trees.
Arrays
The simplest type of data structure is a linear (or one-dimensional) array. By a
linear array, we mean a list of a finite number n of similar data elements referred
respectively by a set of n consecutive numbers, usually 1,2,3 n.
If we choose the name A for the array, then the elements of A are denoted by
subscript notation.
A1 , A2 , A3 ,.., An
Or by the parenthesis notation
A(1), A (2), A(3),, A(N)
Or by the bracket notation
A[1], A[2], A[3],, A[N]
Lets see an example so that we can understand it easily:

A linear array STUDENT consisting of the names of six students is pictured in this
figure
1
2
3
4
5
6
Name1
Name2
Name3
Name4
Name5
Name6
Fig 1-1
This is the simplest example of a single-dimensional array. Here STUDENT[1]

denotes Name1, STUDENT[2] denotes Name2, and so on.
A two-dimensional array is a collection of similar data elements where each
element is referred to by two subscripts. Such arrays are called matrices in
MSc. In Software
mathematics, and tables in business applications. Multidimensional arrays are

defined analogously.
Consider a block like this, having 3 rows and 4 columns. You can visualize a two
dimensional array just like this.
sw
The size of this array is denoted by 3 X 4 (read 3 by 4), since it contains 3 rows
(the horizontal lines of numbers) and 4 columns (the vertical lines of numbers).
If we denote the 1st array members as array[0][0], the following array members
will be denoted like this:
array[0][0]
array [0][1]
array [0][2]
array [0][3]
array [1][0]
array [1][1]
array [1][2]
array [1][3]
array [2][0]
array [2][1]
array [2][2]
array [2][3]
Fig 1-2
The position of the highlighted cell in the block in the array notation is
array[2][3].
Linked Lists
A linked list is the most important and difficult part of data structure. So if you
want to be an expert in data structures, emphasize more on linked lists,
understand the basic idea and try to solve as many examples as you can.
If we only give the theory of linked lists it will be difficult for you to understand.
Therefore, we will introduce it with an example. Consider a file where each record
contains a customer's name and his or her salesperson, and suppose the file
contains the data as appearing in the figure 1-3. Clearly the file could be stored in
the computer by such a table, i.e. by two columns of five names. However, this
may not be the most useful way to store the data.
MSc. In Software
Salesperson1
Salesperson2
Salesperson3
Salesperson4
Salesperson5
Salesperson6
Salesperson7
Salesperson8
1
2
3
4
5
6
7
8
Customer1
Customer2
Customer3
Customer4
Customer5
Salesperson1
Salesperson2,
Sales person3
Salesperson3,
Sales person8
Salesperson4
Salesperson5
Salesperson1
Fig. 1-3
Another way of storing data in the figure 1-3 is to have a separate array for the
salespeople and an entry (called a pointer) in the customer file, which gives the
location of each customer's salesperson. This is done in the figure 1-4, where
against every customer name we have written the number (pointer) of the
corresponding salesperson.
Practically speaking, in figure 1-3, in front of each customer we have specified his
salespersons name. Now in the above case the number of customers is very less
that is why we can afford to do this. But imagine a case where there are hundreds
of customers. In such a case, repeating the name of the sales person will
consume lot of space. Instead, we will give numbers to the sales persons and
mention that number in front of the customers name. Thus we can save a lot of
space.
Customer1
Customer2
Customer3
Customer4
Customer5
Customer1
Customer2
Customer3
Customer4
Customer5
a
b
c
d
e
1
2,3
3,8
4
5,1
Fig.1-4
Suppose the firm wants the list of customers for a given salesperson. Using the
data representation in this figure 1-4, the firm would have to search through the
entire customer file. One way to simplify such a search is to have a table
containing customer name and a number (pointer) corresponding to each
customer. Each salesperson would now have a set of numbers (pointer) giving the
position of his or her customers, as in this figure 1-5.
MSc. In Software
Salesperson1
Salesperson2
Salesperson3
Salesperson4
Salesperson5
Salesperson6
Salesperson7
Salesperson8
a,e
b
b,c
d
e
Fig. 1-5
Disadvantage:
The main disadvantage of this representation is that each salesperson may have
many pointers and the set of pointers will change as customers are added and
deleted.
Link List contd..

The most popular way to store such data is shown in figure 1-6. Here each
salesperson has one pointer which points to his or her first customer, whose
pointer in turn points to the second customer, and so on, with the salesperson's
last customer indicated by a 0. Consider figure 1-6 for the salesperson
Salesperson1.
1
Salesperson1
a
Customer1
1
b
Customer2
a
c
Customer3
b
0
Customer4
c
Fig. 1-6
Here 1 is the pointer of Salesperson1. This pointer points to Customer1. a is
the pointer of Customer1, which in turn points to Customer2. Similarly, b,
which is the pointer of Customer2 points to Customer3 and so on. Since
Customer4 is the last customer i.e. this customer is not further connected to any
other customer, its pointer has been assigned to 0. (In this picture we have
considered that Salesperson1 has got Customer1, Customer2, Customer3,
Customer4).
Trees
Data frequently contains a hierarchical relationship between various elements.
The data structure, which reflects a hierarchical relationship between various
elements, is called a rooted trees graph or, simply, a tree.
MSc. In Software
10
Trees will be defined and discussed in detail in later modules but here we indicate
some of their basic properties by means of two examples:
(a) An employees personnel record :
This may contain the following data items
i) Social Security Number
ii) Name
iii) Address
iv) Age
v) Salary
vi) Dependents
However, Name may be a group item with the subitems Last, First and MI (middle
initial). Also, address may be a group item with the subitems Street address and
Area address, where Area itself may be a group item having subitems City, State
and ZIP code. This hierarchical structure is explained in figure 1-7 (a).
Fig 1-7(a)
Another way of picturing such a tree structure is in terms of levels, as shown in
figure 1-7 (b).
01 Employee
02. Social Security Number
02. Name
03. Last
03 First
03 Middle Initial
02 Address
03 Street
03 Area
04 City
04 State
MSc. In Software
02 Age
02 Salary
02 Dependents
11
04 ZIP
Fig. 1-7(b)
(b) An algebraic expression in the tree structure format for calculating.
Let the expression be
(2x + y) (a - 7b)3
Now we want to represent the expression by the tree, so lets use a vertical arrow
() for exponential and an asterisk (*) for multiplication. Thus now we can show
the expression in terms of a tree diagram as shown in figure 1-8. Observe that
the order in which the operations will be performed is reflected in the diagram:
the exponentiation must take place after the subtraction, and the multiplication at
the top of the tree must be executed last.
Fig. 1-8
Some More data Structures

(a) Stack : A stack, also called a last-in first out (LIFO) system, is a linear list in
which insertions and deletions can take place only at one end, called the top.
This structure is similar in its operation to a stack of dishes on a spring. New
dishes are inserted only at the top of the stack and dishes can be deleted only
from the top of the stack.
(b) Queue : A queue, also called a first-in first-out (FIFO) system, is a linear list
in which deletions can take place only at the front of the list, and insertions
can take place only at the other end, the rear of the list. This structure
operates in much the same way as a line of people waiting at a bus stop. The
first person in line is the first person to board the bus. Another analogy is with
MSc. In Software
12
automobiles waiting to pass through an intersection--the first car in line is the

first car through.
(c) Graph : Data sometimes contain a relationship between pairs of elements,
which is not necessarily hierarchical in nature. For example, suppose an airline

flies only between the cities connected by lines as shown in the figure below.
The data structure, which reflects such a relationship, is called a graph.
City1
City2
City4
City3
City5
Fig. 1-9
DATA STRUCTURE OPERATIONS

Now we are going to see how data appearing in our data structures are processed
by certain operations. Remember that the particular data structure one chooses
for a given situation depends largely on the frequency with which specific
operations are performed. In this section we will introduce some of the frequently
used operations.
The following four operations play a major role in the text:
(1) Traversing: Accessing each record exactly once so that certain items in the
record may be processed. (The accessing and processing is sometimes called
"visiting" the record.)
(2) Searching: Finding the location of the record with a given key value, or
finding the locations of all records, which satisfy one or more conditions.
(3) Inserting: Adding a new record to the structure.
(4) Deleting: Removing a record from the structure.
Sometimes two or more of the operations may be used in a given situation; e.g.,
we may want to delete the record with a given key, which may mean we first
need to search for the location of the record.
The following two operations, which are used in special situations, will also be
considered.
(1) Sorting: Arranging the records in some logical order (e.g., alphabetically
according to some NAME key, or in numerical order according to some
NUMBER key, such as employee number or account number).
(2) Merging: Combining the records in two different sorted files into a single
sorted file.
MSc. In Software
13
A real example will make our idea clear about these concepts:
An organization contains a membership file in which each record contains the
following data for a given member.
Name
Address
Telephone Number
Age
Sex
(a) Suppose the organization wants to announce a meeting through a mailing
system. Then one would traverse the file to obtain Name and Address for each
member.
(b) Suppose one wants to find the names of all members living in a certain area.
Again one would traverse the file to obtain the data.
(c) Suppose one wants to obtain address for a given Name. Then one would
search the file for the record containing Name.
(d) Suppose a new person joins the organization. Then one would insert his or her
record into the file.
(e) Suppose a member dies. Then one would delete his or her record from the file.
(f) Suppose a member has moved and has a new address and telephone number.
Given the name of the member, one would first need to search for the record
in the file. Then one would perform the "update"--i.e., change items in the
record with the new data.
(g) Suppose one wants to find the number of members 65 or older. Again one
would traverse the file, counting such members.
ALGORITHMS: COMPLEXITY, TIME-SPACE TRADEOF
Consider time and space trade-offs in deciding on your algorithm.

Never be afraid to start over. Next time it may be both shorter and easier.
An algorithm is a well-defined list of steps for solving a particular problem. One

major purpose of this section is to develop efficient algorithms for the processing
of our data. The time and space it uses are two major measures of the efficiency
of an algorithm. If the storage space is available and otherwise unused, it is
preferable to use the algorithm requiring more space and less time. If not, then
time may have to be sacrificed.
Lets see these ideas with two examples:
Searching Algorithms
Consider a membership file in which each record contains, among other data, the
name and telephone number of its member. Suppose we are given the name of a
MSc. In Software
14
member and we want to find his or her telephone number. One way to do this is
to linearly search through the file, i.e., to apply the following algorithm:
Linear Search
Search each record of the file, one at a time, until the given Name and the
corresponding telephone number is found.
Consider that the time required to execute the algorithm is proportional to the
number of comparisons.
Second, assuming that each name in the file is equally likely to be picked, it is
intuitively clear that the average number of comparisons for a file with n records
is equal to n/2; that is, the complexity of the linear search algorithm is given by
C(n) = n/2.
Binary Search
Compare the given Name with the name in the middle of the list. This indicates
which half of the list contains Name. Then compare Name with the name in the
middle of the correct half to determine which quarter of the list contains Name.
Continue the process until Name is found in the list.
One can show that the complexity of the binary search algorithm is given by
C(n) = log2n.
Thus, for example, one will not require more than 6 comparisons to find a given
Name in a list containing 64 (=26) names.
Drawback
Although the binary search algorithm is a very efficient algorithm, it has some
major drawbacks. Specifically, the algorithm assumes that one has direct access
to the middle name in the list or a sublist. This means that the list must be stored
in some type of array. Unfortunately, inserting an element in an array requires
elements to be moved down the list, and deleting an element from an array
requires elements to be moved up the list.
An Example of Time-Space Tradeoff
Suppose a file of records contains names, employee numbers and much additional
information among its fields. For finding the record for a given name, sorting the
file alphabetically and using a binary search is a very efficient way. On the other
hand, suppose we are given only the employee number of the person. Then we
would have to do a linear search for the record, which is extremely timeconsuming for a very large number of records. How can we solve such a problem?
One way is to have another file, which is sorted numerically according to the
MSc. In Software
15
employee number. This, however, would double the space required for storing the
data. Another way, pictured in figure 1-10, is to have the main file sorted
numerically by employee number and to have an auxiliary array with only two
columns, the first column containing an alphabetized list of the names and the
second column containing pointers, which give the locations of the corresponding
records in the main file. This is one way of solving the problem that is done
frequently, since the additional space, containing only two columns, is minimal for
the amount of extra information it provides.
Employee. No.
Name
1-abc
2-xyz
3-pqr
4-mnp
5-lmn
Name1
Name2
Name3
Name4
Name5
Extra Data
XXXX
XXXX
XXXX
XXXX
XXXX
Pointer
1
2
3
4
5
Name
Name1
Name2
Name3
Pointer
1
2
3
Fig. 1-10
Quote of the chapter:
Act in haste and repent in leisure.
Program in haste and debug forever.
Summary
# An entity is something that has certain attributes or properties that may
be assigned values. These values may be numeric or nonnumeric.
# The logical or mathematical model of a particular organization of data is
called a data structure. Three most popular data Structures are arrays, link
list and trees.
# Some other data types are stacks, queues and graphs.
# We have discussed two searching algorithms, (i) linear search,
(ii) binary search.
# Sometimes two or more operations may be used in a given situation to get
optimum speed of a given situation.
MSc. In Software
Zee Interactive Learning Systems
16
MSc. In Software
2
COMPEXITY, RATE OF
Growth,
BIG O NOTATION
MAIN POINTS COVERED
INTRODUCTION
a) Floor and Ceiling Functions
b) Remainder function: Modular Arithmetic
c) Integer and Absolute Value Functions
d) Summation Symbol: Sums
e) Factorial Function
f) Permutations
g) Exponents and Logarithms
MSc. In Software
!
ALGORITHMIC NOTATION
CONTROL STRUCTURES
Sequence logic, or sequential flow

Selection logic, or conditional flow
1. Single alternative
2. Double alternative
3. Multiple alternatives
Iteration logic, or repetitive flow
COMPLEXITY OF ALGORITHMS
RATE OF GROWTH; BIG O NOTATION
SUMMARY
INTRODUCTION
This section gives various mathematical functions, which appear very often in
the analysis of algorithms and in computer science.
a) Floor and Ceiling Functions
Let x be any real number. Then x lies between two integers called the floor
and the ceiling of x. Specifically,
x , called the floor of x, denotes the greatest integer that does not exceed
x.
x , called the ceiling of x, denotes the least integer that is not less than x.
If x is itself an integer, then x = x ; otherwise x + 1 = x
3.14 =
5 =
-8.5 =
7
=
3
2
-9
7
3.14
5
-8.5
7
=4
=3
= -8
=7
b) Remainder function: Modular Arithmetic

Let k be any integer and let M be a positive integer. Then,
k (mod M)
MSc. In Software
(read k modulo M) will denote the integer remainder when k is divided by M.

More exactly, k (mod M) is the unique integer r such that
Where 0 r < M
k = Mq + r
When k is positive, simply divide k by M to obtain the remainder r.

Some examples:
25 (mod 7) = 4,
25(mod 5) = 0,
35 (mod 11) = 2,
3 (mod 8) = 3
[dividend
[dividend
[dividend
[dividend
=25, divider = 7, remainder is 4]

The term "mod" is also used for the mathematical congruence relation, which
is denoted and defined as:
a b (mod M)
if and only if
M divides (b a)
M is called the modulus, and a b (mod M) is read a is congruent to b

modulo M. The following aspects of the congruence relation are frequently
useful.
0 M (mod M)
and a M
a (mod M)
Arithmetic modulo M refers to the arithmetic operations of addition,

multiplication and subtraction where the arithmetic value is replaced by its
equivalent value in the set.
{ 0, 1, 2,., M -1}
or, in the set {1,2,3,., M}.
For example, in arithmetic modulo 12, sometimes called "clock" arithmetic,
6 + 9 3, 7 x 5 11, 1 - 5 8,
2 + 10 0
(The use of 0 or M depends on the application.)
12
c) Integer and Absolute Value Functions

Let x be any real number. The integer value of x, written INT(x), converts x
into an integer by deleting (truncating) the fractional part of the number.
Thus
INT(3.14) = 3,
=7
INT ( 5 ) = 2,
Observe that INT(x) = x

positive or negative.
or
INT (-8.5) = -8,
INT(x) = x
INT (7)
according to whether x is
MSc. In Software
The absolute value of the real number x, written ABS(x) or |x| , is defined as
the greater of x or -x. Hence ABS (0) = 0, and, for x 0, ABS (x) = x, if x is
positive, and -x, if x is negative. Thus,
| - 15| = 15 ,
|7| = 7,
4.44,
|-0.075| = 0.075
We note that
| x| =
|-x|
|-3.33|
and, for x 0,
= 3.33,
|x |
|4.44|
is positive.
d) Summation Symbol: Sums

Here we introduce the summation symbol (the Greek letter sigma).
Consider a sequence a1, a2, a3, .. then the sums
a1 + a2 + . an
and
am + am+1 + + an
will be denoted respectively, by

n
aj
j=1
and
n
aj
j=m
The expression a1b1 + a2b2 + + anbn is denoted as

n
ai bi
i=1
when n = 5, for a=b and I starting from 2, we have
5
j2
= 22 + 32 + 42 + 52 = 4 + 9 + 16 + 25 = 54
i=2
e) Factorial Function
The product of the positive integers from 1 to n, inclusive, is denoted by n!
(read "n factorial"). That is,
n! = 1 . 2 . 3 (n -2).(n - 1).n
It is also convenient to define 0! = 1.
For example,
(a)
(b)
2! = 1.2 = 2 ;
3! = 1.2.3 = 6 ;
MSc. In Software
(c)
(d)
(e)
4! = 1.2.3.4 = 24
5! = 5.4! = 5.24 = 120 ;
6! = 6. 5! = 6 . 120 = 720
f) Permutations
A permutation of a set of n elements is an arrangement of the elements in a
given order. For example the permutations of the set consisting of the
elements a , b , c are:
abc , acb, bac, bca, cab, cba
One can prove: There are n! permutations of a set of n elements.
Accordingly there are 4! = 24 permutations of a set of 4 elements, 5! = 120
permutations of a set with 5 elements, and so on.
g) Exponents and Logarithms
We consider first integer exponents (where m is a positive integer).
am = a . a . .. a ( m times )
a0 = 1,
a-m = 1 / am
Exponents are extended to include all rational numbers by defining, for any
rational number m / n
a
For example,
24 = 16 ,
m/n
= n am =
( a)
2-4 = 1 / 24 = 1/16 ,
1252/3 = 52 = 25
Logarithms are related to exponents as follows.

Let b be a positive number. The logarithm of any positive number x to the
base b written
log b x
represents the exponent to which b must be raised to obtain x. That is
y = log
and
by = x
are equivalent statements. Accordingly

log
8=3
since 23 = 8
MSc. In Software
log
log
log
since 10 2 = 100
since 2 6 = 64
since 10 3 = 0.001
100 = 2
2 64 = 6
10 0.001 = -3
10
Furthermore for any base b,

since b0 = 1
since b1 = b
log b 1 = 0
log b b = 1
The logarithm of a negative number and the logarithm of 0 are not defined.
Exponent function f(x) = bx
Logarithmic function
g(x) = log b x
For example log
40 = 3.6889.
Natural logarithms
Common logarithms
Binary logarithms
where e = 2.718281..
Logarithms to the base e

Logarithms to the base 10
Logarithms to the base 2
Notation: log x will mean log
x unless otherwise specified.
ALGORITHMIC NOTATION
An algorithm is a finite step-by-step list of well-defined instructions for
solving a particular problem. This section describes the format that is used to
present algorithms throughout the text. This algorithmic notation is best
described by means of examples.
An array DATA of numerical values is in memory. We want to find the
location LOC and the value MAX of the largest element of DATA. Given no
other information about DATA, one way to solve the problem is:
i) Initially we being with LOC = 1 and MAX = DATA [1].
ii) Then compare MAX with each successive element DATA [K] of DATA.
iii) If DATA [K] exceeds MAX, then update LOC and MAX
so that LOC = K and MAX = DATA [K].
The final values appearing in LOC and MAX give the location and value of the
largest element of DATA.
Algorithm (Largest Element in Array) A nonempty array DATA with N
numerical values is given. The algorithm finds the location LOC and the value
MAX of the largest element of DATA.
The variable K is used as a counter.
Step 1
[Initialize. ] Set K = 1, LOC = 1 and MAX = DATA [1].
MSc. In Software
Step 2.
Step 3.
Step 4.
Step 5.
[Increment counter.] Set K = K + 1.

[Test counter.] if K > N, then
Write LOC, MAX, and Exit.
[Compare and update. ] if MAX < DATA[K], then
Set LOC = K and MAX = DATA [K]
[Repeat loop.] Go to Step 2.
The format for the formal presentation of an algorithm consists of two parts.
The first part identifies the variables, which occur in the algorithm and lists
the input data. The second part of the algorithm consists of the list of steps
that is to be executed.
CONTROL STRUCTURES
Algorithms and their equivalent computer programs are more easily
understood if they use self-contained modules and three types of logic or
flow of control.
(1)
(2)
(3)
Sequence logic, or sequential flow

Selection logic, or conditional flow
Iteration logic, or repetitive flow
These three types of logic are discussed below and in each case we show the
equivalent flowchart.
Sequence Logic (Sequential Flow)
In this case the modules are executed in the obvious sequence. The
sequence may be presented explicitly, by means of numbered steps, or
implicitly, by the order in which the modules are written.
Algorithm
.
.
.
Module A
Flow chart equivalent
Module A
Module B
Module B
Module C
.
.
.
Module C
Sequence logic.
MSc. In Software
Selection Logic (Conditional Flow)

Selection logic employs a number of conditions, which lead to a selection of
one out of several alternative modules. The structures, which implement this
logic are called conditional structures or if structures. For clarity, we will
frequently indicate the end of such a structure by the statement [End of if
Structure.]
These conditional structures fall into three types, which are discussed
separately.
Single alternative:
This structure has the form
if condition, then
[Module A]
[End of if Structure.]
The logic of this structure is pictured in Fig. 2-3(a). If the condition holds,
then Module A, which may consist of one or more statements, is executed;
otherwise Module A is skipped and control transfers to the next step of the
algorithm.
Double alternative:
This structure has the form:
if condition, then
[ Module A ]
else
[ Module B ]
[ End of if structure. ]
Multiple alternatives:
The structure has the form
if condition(1), then
[ Module A1 ]
else if condition(1), then
[ Module A2 ]
:
:
else if condition(M), then
MSc. In Software
else
[ Module AM ]
[ Module B]
[ End of if structure ]
The logic of the structure allows only one of the modules to be executed.
Example
The solution of the quadratic equation
ax2 + bx + c = 0
where a 0, is given by
-b ( b2 4ac)
=
2a
where D = b2 4ac is called the discriminant of the equation.

If D is negative then there are no real solutions.
If D = 0, then there is only one(double) real solution, x = -b/2a.
If D is positive, the formula gives two distinct solutions. The following
algorithm finds the solution of a quadratic equation.
Algorithm (Quadratic Equation) This algorithm inputs the coefficients A, B, C
of a quadratic equation and outputs the real solutions, if any.
Step 1.
Step 2.
Step 3.
Read A, B, C
Set D = B2 4AC
if D > 0, then
Set X1 = (-B + D)/2A and X2 = (-B - D)/2A
Write X1, X2
else if D = 0, then
Set X = -B / 2A
Write UNIQUE SOLUTION , X .
else
Write NO REAL SOLUTION .
[ End of If structure ]
Step 4.
Exit.
Iteration Logic (Repetitive Logic)
MSc. In Software
10
The third kind of logic refers to either of two types of structures involving
loops. Each type begins with a Repeat statement and is followed by a
module, called the body of the loop.
There are two types of such loops:
(1) repeat-for loop
Repeat for K = R to S step T
[ Module ]
[ End of Loop ]
(2) repeat-while loop
Repeat while condition
[ Module ]
[ End of loop ]
We have discussed an algorithm for finding the maximum element in an
array. Now we are going to discuss the same problem using a repeat-while
loop
Algorithm( Largest Element in Array ) Given a nonempty array DATA with N
numerical values this algorithm finds the location LOC and the value MAX of
the largest element of DATA.
1.
5.
6.
[ Initialize. ] Set K = 1 , LOC = 1 , and MAX = DATA[1].

2.
Repeat Steps 3 and 4 while K N
3. if MAX < DATA[K], then
Set LOC = K and MAX = DATA[K].
[ End of if structure. ]
4.Set K = K + 1.
[ End of Step 2 loop. ]
Write LOC, MAX.
Exit.
COMPLEXITY OF ALGORITHMS
In designing algorithms we need methods to separate bad algorithms from
good ones. This will enable us to choose the right one in a given situation.
The analysis of algorithms and comparisons of alternative methods constitute
an important part of software engineering. In order to compare algorithms,
we have to find out the efficiency of our algorithms. In this section we will
discuss how to find efficiency. Lets see this with an example.
Example 1
Food for thought:
MSc. In Software
11
What do you think are the possible criteria that measure the efficiency of an
algorithm?
(a) Time taken
(b) Length of the algorithm
(c) Memory space used
(d) Time required in writing the algorithm
(a) and (c) are the correct answers.
Suppose you are given an algorithm M, and the size of the input data is n.
Then the efficiency of the algorithm M depends on two main measures:
Time taken by the algorithm

Space used by the algorithm
Time: We can measure time by counting the number of key operations

performed during the execution of the algorithm.
Key operations - During the execution of an algorithm we have to perform
several operations. Among them, the operation that takes longer time as
compared to other operations is called key operation.
Space: The space is measured by counting the maximum memory needed by
the algorithm.
Then the complexity of an algorithm M is the function f(n), which gives the
running time and/or storage space requirement of the algorithm in terms of
size n of the input data. Frequently, the storage space required by an
algorithm is simply a multiple of the data size n.
Complexity basically refers to the running time of an algorithm.
Example 2
Suppose we have been given an English short story TEXT, and want to
search through TEXT for the first occurrence of a given 3-letter word, W. If
the 3-letter word is the, then it is likely that it will occurs near the
beginning of TEXT, so the complexity, f(n) will be small.
On the other hand, if W stands for the 3-letter word zee
Food for thought:
Where do you think the word zee will occur?
a) Beginning of the text
b) End of the text
MSc. In Software
12
c) Never occur
d) Not quite sure
Just for thought, any answer can be true.
The word zee is not a very common word so W may not appear at all, so
the complexity f(n) of the algorithm will be large.
The above discussion leads us to the question of finding the complexity
function f(n) for certain cases. The two cases one usually investigates in the
complexity theory are as follows:
(1) Worst case:
the maximum value of f(n) for any possible input
(2) Average case: the expected value of f(n)
(3) Best case:
sometimes we also consider the minimum possible value
of f(n), called the best case
Food for thought:
What is the best case while searching in an array for a specific element?
(a) Element occurs at the end of the array
(b) Element occurs at the beginning of the array
(c) Element occurs at the middle most position of the array
(d) Best case does not exist
(b)
is the correct choice, as only one element has to be compared

before the searching procedure can end.
Average case analysis of algorithms:

The analysis of the average case assumes a certain probabilistic distribution
for the input data. One such assumption might be that all possible
permutations of an input data set are equally likely. The average case also
uses the following concept in probability theory. Suppose the numbers n1, n2
, , nk occur with respective probabilities p1 , p2 , .. , pk. Then the
expectation or average value E is given by
E = n1p1 + n2p2 + . + nkpk
Example (Linear Search)
Suppose the linear array DATA contains n elements, and suppose a specific
ITEM of information is given. We want either to find the location LOC of ITEM
in the array DATA, or to send some message, such as LOC = 0, to indicate
that item does not appear in DATA. The linear search algorithm solves this
problem by comparing ITEM, one by one, with each element in DATA. That
MSc. In Software
13
is, we compare ITEM with DATA[1], then DATA[2], and so on, until we find
LOC such that
ITEM = DATA[LOC].
A formal representation of the algorithm is as follows:
Algorithm ( Linear Search ) A linear array DATA with N elements and a
specific ITEM of information are given. The algorithm finds the location LOC
of ITEM in the array DATA or sets LOC = 0
1. [ Initialize ] Set K = 1 and LOC = 0.
2. Repeat Steps 3 and 4 while LOC = 0 and K N.
3.
if ITEM = DATA[K], then Set LOC = K.
4.
Set K = K + 1. [ Increments counter. ]
[ End of Step 2 loop. ]
5. [ Successful? ]
if LOC = 0, then
Write ITEM is not in the array DATA
else
Write LOC is the location of ITEM.
[ End of If structure. ]
6. exit
We can find the complexity of the search algorithm by the number C of

comparisons between ITEM and DATA[K]. We seek C(n) for the worst case
and the average case.
Worst Case
Clearly the worst case occurs when ITEM is the last element in the array
DATA or is not there at all. In either case, we have
C(n) = n
Average Case
Here we assume that ITEM does appear in DATA, and it is equally likely to
occur at any position in the array. Accordingly, the number of comparisons
can be any of the numbers 1, 2, ,n, and each with probability p = 1/n.
Then
C(n) = 1 . 1/n + 2 . 1/n + + n . 1/n
= ( 1 + 2 + .. + n ) . 1/n
= n ( n + 1 ) / 2 . 1/ n
= (n+1)/2
MSc. In Software
14
This agrees with our intuitive feeling that the average number of
comparisons needed to find the location of ITEM is approximately equal to
half the number of elements in the DATA set.
Rate of Growth: Big O Notation or Big Oh Notation

If f(n) and g(n) are functions defined for positive integers, then to write
f(n) is O (g(n))
means that there exists a constant C such that |f(n)| < c |g(n)| for all
sufficiently large positive integers n. Under these conditions we also say that
f(n) has order at most g(n) or f(n) grows no more rapidly than g(n).
When we apply this notation, f(n) will normally be the operation count or
time for some algorithm, and we wish to choose the form of g(n) to be as
simple as possible. We thus write O(1) to mean computing time that is bound
by a constant, not dependent on n. O(n) means that the time is directly
proportional to n, and is called the linear time. We call O(n2 ) quadratic time,
O(n3) cubic, O(2n) exponential. These five orders, together with logarithmic
time O(log n) and O(n log n), are the once most commonly used in analyzing
algorithm.
Suppose M is an algorithm, and n is the size of input data in that algorithm.
Clearly the complexity f(n) of M increases as n increases. It is usually the
rate of increase of f(n) that we want to examine. This is usually done by
comparing f(n) with some standard function, such as
Logarithmic time : log 2n
Linear time :
n
Quadratic time :
n2
Cubic time :
n3
Linear cum logarithmic time : n log2n,
Food for thought:
What will be the name of the function whose functional form is 2n?
(a) Hyperbolic
(b) Parabolic
(c) Exponential
(d) Logarithmic
(c) is the correct choice
The rates of growth of these standard functions are given below in the
Fig 2-1.
MSc. In Software
15
Fig 2-1
The above table is arranged such that the rate of growth of the function
increases from left to right with log n having the lowest rate of growth and
2n having the largest rate of growth.
To indicate the convenience of this notation, we give the complexity of
certain well known searching and sorting algorithms:
(a)
(b)
(c)
(d)
Linear search
Binary search
Bubble sort
Merge-sort
:
:
:
:
O(n)
O(log n)
O(n2)
O(n log n)
Summary
# In order to compare algorithms, we have to find out its efficiency. This
will help us to employ the right one in order to solve problems
effectively.
# Suppose M is an algorithm, and n is the size of input data in that
algorithm. Clearly the complexity f(n) of M increases as n increases.
# Big O notation states that for a function f(n), there exists a positive
integer n0 and a positive number M such that, for all n > n0, we have
| f(n) | < M | g(n) |
Then we may write
f(n) = O ( g(n) )
This is called big O notation.
MSc. In Software
3
ARRAYS
MAIN POINTS COVERED
!
INTRODUCTION
LINEAR ARRAY
REPRESENTATION OF LINEAR ARRAY IN

MEMORY
TRAVERSING LINEAR ARRAYS
SORTING; BUBBLE SORT
SEARCHING; LINEAR SEARCH
MULTIDIMENSIONAL ARRAYS
REPRESENTATION OF MULTIDIMENSIONAL ARRAYS

IN MEMORY
SUMMARY
MSc. In Software
Introduction
ata-structures are classified into two broad categories linear and nonLinear.
The most elementary data-structure that we will introduce is array.
Advantages:
This has a linear structure
They are easy to traverse, search and sort
They are easy to implement
Disadvantages:
The length of the array cannot be changed once it is specified
Food for thought:
What do you think is the reason for the above disadvantage?
(a) Array size has to be fixed at the beginning
(b) Problem with memory allocation occurs if size is altered
(c) Once a fixed block of memory has been reserved for a particular
array it cannot be altered
(d) Variable length arrays are not required in real life
(a)
and (c) are the correct choices.
LINEAR ARRAY
A linear array is a list of a finite number n of homogeneous data elements
(i.e., data elements of the same type) such that:
(a) The elements of the array are referenced respectively by an index
set consisting of n consecutive numbers
(b) The elements of the array are stored respectively in successive
memory locations
Length or size of an array = The number of elements in the array
Length = UB LB + 1
Where UB is the largest index called the upper bound
And LB is the smallest index called the lower bound
Notation For Representing Arrays
Food for thought:
You can represent the elements of an array A by
(a)
A1, A2, A3, .., An
MSc. In Software
(b)
(c)
(d)
A (1), A (2), .., A (N)

A [1], A [2], A [3], .., A [N]
A{1}, A{2}, A{3}, .., A{N}
(a) , (b) and (c) are the correct choices

The subscript notation may denote the elements of an array A
A1, A2, A3, .., An
or by the parentheses notation (used in FORTRAN, PL/1 and BASIC)
A (1), A (2), .., A (N)
or by the bracket notation (used in C and Pascal)
A [1], A [2], A [3], .., A [N]
We will usually use the subscript notation or the bracket notation. Regardless
of the notation, the number K in A[K] is called a subscript or an index and
A[K] is called a subscripted variable. Note that a subscript allows any
element of A to be referenced by its relative position in A.
REPRESENTATION OF LINEAR ARRAYS IN MEMORY

Let LA be a linear array in the memory of the computer. The memory of the
computer is simply a sequence of addressed locations as shown in the fig 31.
Food for thought:
What is the nature of these memory locations?
a)
b)
c)
d)
Linear
Circular
Random
We cannot know about the memory locations
(a) Is the correct choice as seen from the diagram below
We use the notation

LOC (LA[K]) = address of the element LA [K] of the array LA
As we have previously noted, the elements of LA are stored in successive
memory cells. Accordingly, the computer does not need to keep track of the
address of every element of LA, but needs to keep track only of the address
of the first element of LA, denoted by
Base (LA)
and called the base address of LA.
Food for thought:
Why dont we keep track of all the array elements?
MSc. In Software
(a) Array elements except the first one are not required
(b) The first element contains information of all the other elements
(c) Knowing the first element and the position of the required element
we can traverse the array to reach that element
(c) is the correct choice as explained below
Using this address Base (LA), the computer calculates the address of any
element of LA by the following formula:
LOC(LA[K]) = Base (LA) + w(K - lower bound)
Where w is the number of words per memory cell for the array LA. Observe
that the time to calculate LOC(LA[K]) is essentially the same for any value of
K. Furthermore, given any subscript K, one can locate and access the content
of LA[K] without scanning any other element of LA.
.
100
101
102
103
104
.
Fig 3-1
TRAVERSING LINEAR ARRAYS

Let A be a collection of data elements stored in the memory of the computer.
Suppose we want to print the contents of each element of A or suppose we
want to count the number of elements of A with a given property. We can
accomplish this by traversing A, that is, by accessing and processing
(frequently called visiting) each element of A exactly once.
The following algorithm traverses a linear array LA. The simplicity of the
algorithm comes from the fact that LA is a linear structure.
MSc. In Software
Algorithm: (Traversing a Linear Array) Here LA is a linear array with lower

bound LB and upper bound UB. This algorithm traverses LA
applying an operation PROCESS to each element of LA.
1 [Initialize counter.] Set K = LB.
2 Repeat Steps 3 and 4 while K UB.
3
[Visit element.] Apply Process to LA [K].
4
[Increase counter.] Set K = K + 1.
[End of Step 2 loop.]
5. Exit.
Now we present the same algorithm using a different control structure.

Here is an alternative of the algorithm, which uses a repeat-for loop instead
of the repeat-while loop.
Algorithm: (Traversing a Linear Array) This algorithm traverses a
linear array LA with lower bound LB and upper bound UB.
1. Repeat for K = LB to UB
Apply PROCESS to LA[K]
[End of loop.]
2. Exit.
Note: The operation PROCESS in the traversal algorithm may use certain
variables, which must be initialized before PROCESS is applied to any of the
elements in the array. Accordingly, the algorithm may need to be preceded
by such an initialization step.
Inserting And Deleting
Let A be a collection of data elements in the memory of the computer.
"Inserting" refers to the operation of adding another element to the collection
A, and "deleting" refers to the operation of removing one of the elements of
A.
Food for thought:
In an array, from which end does insertion and deletion take place?
(a) Beginning of the array
(b) End of the array
(c) Middle of the array
(d) Beginning and end of the array
(c) and (d) are the correct choices, the choice of the end from which
insertion and deletion occurs solely depends on the programmer.
This section discusses inserting and deleting when A is a linear array.
MSc. In Software
We can easily insert an element at the "end" of a linear array provided the
memory space allocated for the array is large enough to accommodate the
additional element. On the other hand, if we need to insert an element in the
middle of the array. Then, on the average, half of the elements must be
moved downward to new locations to accommodate the new element and
keep the order of the other elements.
Similarly, deleting an element at the "end" of an array presents no
difficulties, but deleting an element somewhere in the middle of the array
would require each subsequent element to be moved one location upward in
order to "fill up" the array.
Example
Suppose TEST has been declared to be a 5-element array but data have been
recorded only for TEST[1], TEST[2] and TEST[3]. If X is the value of the next
test, then one simply assigns
TEST [4] = X
to add X to the list. Similarly, if Y is the value of the subsequent test, then
we simply assign
TEST[5] = Y
to add Y to the list, Now, however, we cannot add any new test scores to the
list.
SORTING; BUBBLE SORT
Let A be a list of n numbers. Sorting A refers to the operation of
rearranging the elements of A so they are in increasing order, i.e., so that
A [1] < A [2] < A [3] < < A [N]
For example, suppose A originally is the list
After sorting, A is the list
23,4,5,13,6,19,6
4,5,6,13,19,23
Sorting may seem to be an easy task. Actually, sorting efficiently may be

quite complicated. In fact, there are many, many different sorting
algorithms. Here we present and discuss a very simple sorting algorithm
known as the bubble sort.
MSc. In Software
Food for thought:

On which type of data can sorting take place?
(a) Numeric data
(b) Non-numeric data
(c) Symbolic data
(d) Binary data
(a) and (b) are the correct choices
Remark: The above definition of sorting refers to arranging numerical data in
increasing order. This restriction is only for notational convenience. Clearly,
sorting may also mean arranging numerical data in decreasing order or
arranging non-numerical data in alphabetical order. Actually, A is frequently
a file of records, and sorting A refers to rearranging the records of A so that
the values of a given key are ordered.
Bubble Sort
Suppose the list of numbers A[1]. A[2],., A[N] is in the memory. The
bubble sort algorithm works as follows:
Step 1. First we have to compare A[1], A[2] and arrange them in the desired
order, so that A[1] < A[2].
Similarly we can compare A[2] and A[3] and arrange them so that A[2] <
A[3]. Then compare A[3] and A[4] and arrange them so that A[3] < A[4].
We have to continue this process of comparison until we compare A[N - 1]
with A[N] and arrange them so that A[N-1] < A[N].
You can note that the Step 1 involves n-1 comparisons. (During Step 1, the
largest element is "bubbled up" to the nth position or "sinks" to the nth
position.) When Step 1 is completed, A[N] will contain the largest element.
Step 2. Repeat Step 1 with one less comparison; that is, now we stop after
we compare and possibility rearrange A[N-2] and A[N-1]. (Step 2 involves N2 comparisons and, when Step 2 is completed, the second largest element
will occupy A[N-1].)
Step3. Repeat Step 1 with two fewer comparisons; that is, we stop after
comparing and rearranging A[N-3] and A[N-2].
.......................................................
.......................................................
.......................................................
Step N-1. Compare A[1] with A[2] and arrange them so that A[1] < A[2}.
After n - 1 steps, the list will be stored in increasing order.
MSc. In Software
The process of sequentially traversing through all or part of a list is

frequently called a "pass". So each of the above steps is called a pass.
Accordingly, the bubble sort algorithm requires n-1 passes, where n is the
number of input items.
SEARCHING; LINEAR SEARCH

Consider a collection of data in the memory. Let DATA represent that
collection in the memory. Suppose we have been given a specific ITEM of
information to search. Searching refers to the operation of finding the
location LOC of ITEM in DATA, or printing some message that ITEM does not
appear there. The search is said to be successful if ITEM does appear in
DATA and unsuccessful otherwise.
Frequently, we may want to add the element ITEM to DATA after an
unsuccessful search for ITEM in DATA. One then uses a search and insertion
algorithm, rather than simply a search algorithm.
The complexity of searching algorithms is measured in terms of the number
f(n) of comparisons required in finding ITEM in DATA where DATA contains n
elements.
Food for thought:
What is the time complexity of the linear search algorithm?
(a) Linear
(b) Logarithmic
(c) Exponential
(d) Quadratic
(a) is the correct choice
We shall show that linear search is a linear time algorithm.
Linear Search
Suppose DATA is a linear array with n elements. We have not been given any
other information about DATA. The most intuitive way to search for a given
ITEM in DATA is to compare ITEM with each element of DATA one by one.
That is, first we test whether DATA[1] = ITEM, and then we test whether
DATA[2] = ITEM, and so on. This method, which traverses DATA sequentially
to locate ITEM, is called linear search or sequential search.
To simplify this, we first assign ITEM to DATA[N + 1], the position following
the last element of DATA. Then the outcome
MSc. In Software
LOC = N + 1
When LOC denotes the location where ITEM first occurs in DATA, it signifies
that the search is unsuccessful. The purpose of this initial assignment is to
avoid repeated testing to find out if we have reached the end of the array
DATA. This way, the search must eventually "succeed".
We have shown an algorithm for linear search.
Observe that Step 1 guarantees that the loop in Step 3 must terminate.
Without Step 1 the Repeat statement in Step 3 must be replaced by the
following statement, which involves two comparisons, not one:
Repeat while LOC N and DATA [LOC] = ITEM
On the other hand, in order to use Step 1, one must guarantee that there is
an unused memory location.
Algorithm : (Linear Search) LINEAR (DATA, N, ITEM,LOC) Here DATA is
a linear array with N elements and ITEM is a given item of
information. This algorithm finds the location LOC of ITEM in
DATA, or sets LOC = 0 if the search is unsuccessful.
1 [Insert ITEM at the end of DATA.] Set DATA [N + 1] = ITEM.
2 [Initialize counter,] Set LOC = 1.
3 [Search for ITEM.]
Repeat while DATA[LOC] ITEM
Set LOC = LOC + 1
[End of loop.]
4 [Successful ?] if LOC = n + 1, then Set LOC = 0.
Exit.
Complexity of the Linear Search Algorithm

Food for thought:
What is/are the factors on which the complexity of a given algorithm
depends?
(a) Number of steps
(b) Number of comparisons
(c) Number of arithmetic operations
(d) Memory space occupied by the algorithm
(a) and (c) are the correct choices
MSc. In Software
10
We have known that the complexity of our search algorithm is measured by

the number f (n) of comparisons required to find ITEM in DATA, where DATA
contains n elements. We have to consider two important cases as the
average case and the worst case.
Clearly the worst case occurs when have to search through the entire array
DATA, i.e., ITEM does not appear in DATA. In this case, the algorithm
requires f(n)=n+1 Comparisons. Thus, in the worst case, the running time is
proportional to n.
The running time of the average case uses the probabilistic notion of
expectation. Suppose Pk is the probability that ITEM appears in DATA[K], and
suppose q is the probability that ITEM does not appear in DATA. (Then P1 +
P2 + + Pn + q= 1.) Since the algorithm uses K comparisons when ITEM
appears in DATA[K], the average number of comparisons is given by
f(n) = 1. P1 + 2. P2 + + n. Pn + (n + 1). q
In particular, suppose q is very small and ITEM appears with equal
probability in each element of DATA. Then q = 0 and each P1 = 1/n.
Accordingly,
f(n) = 1.1 + 2. 1 + + n. 1 + (n + 1). 0 = (1 + 2 + +n). 1
n
n
n
n
= n(n + 1) . 1 = n + 1
2
n
2
That is in this special case, the average number of comparisons required to
find the location of ITEM is approximately equal to half the number of
elements in the array.
MULTIDIMENSIONAL ARRAYS
The linear arrays we have discussed so far are also called one-dimensional
arrays, since each element we represent in the array is a single subscript.
Most programming languages allow two-dimensional and three-dimensional
arrays, i.e., arrays where elements are referenced, respectively, by two and
three subscripts. In fact, some programming languages allow the number of
dimensions for an array to be as high as 7.
Food for thought:
Which of the following events have to be represented on the computer by
multidimensional arrays?
Chess board
Sales figures as per year of a certain firm
Co-ordinates in mathematics
-------------
yes
no
yes
MSc. In Software
11
Sales figures as per year and firm name

----- yes
(Explanation: one axis represents the year and the other the
name of the firm)
Matrix
----- yes
(most obvious example)
Two-Dimensional Arrays
A two-dimensional (m, n) array A is a collection of m X n data elements
such that each element is specified by a pair of integers (such as j, k), called
subscripts, with the property that
ijm
and
ikn
The element of A with first subscript j and second subscript k will be denoted
by
Aj, k
or A [j, k]
We call two-dimensional arrays as matrices in mathematics and tables in

business applications. Hence, two-dimensional arrays are sometimes called
matrix arrays.
There is a standard way of drawing a two-dimensional m X n array A where
the elements of A form a rectangular array with m rows and n columns and
where the element A[j, k] appears in row j and column k (a row is a
horizontal list of elements, and a column is a vertical list of elements). Figure
3-2 shows the case where A has 3 rows and 4 columns. We emphasize that
each row contains those elements with the same first subscript, and each
column contains those elements with the same second subscript.
Columns
1
2
3
4
1 A[1, 1]
A [1, 2]
A[1,3]
A[1, 4]
2 A[2, 1]
A[2, 2]
A[2,3]
A[2, 4]
Rows
3 A[3, 1]
A [3, 2]
A[3, 3]
A[3, 4]
Fig.3-2-dimensional 3 x 4 array A.
Suppose A is a two-dimensional m X n array. The first dimension of A
contains the index set 1, . .. ., m, with lower bound 1 and upper bound m;
and the second dimension of A contains the index set 1, 2,., n, with lower
bound 1 and upper bound n. The length of a dimension is the number of
integers in its index set. The pair of length m X n (read "m by n") is called
the size of the array.
MSc. In Software
12
Some programming languages allow you to define multidimensional arrays in

which the lower bounds are not 1 (such arrays are sometimes called nonregular). However, the index set for each dimension still consists of
consecutive integers from the lower bound to the upper bound of the
dimension. The length of a given dimension (i.e., the number of integers in
its index set) can be obtained from the formula
Length = upper bound - lower bound + 1
Representation of Two-Dimensional Arrays in Memory

Let A be a two-dimensional m X n Array. Although we represent A as a
rectangular array of elements with m rows and n columns, the array will be
represented in the memory by a block of (m X n) sequential memory
locations. Specifically, the programming language will store the array A
either by
(1)
column by column, which is called column-major order, or
(2)
row by row, in row-major order.
Figure 3-4 shows these two ways when A is a two-dimensional 3 X 4 array.
We emphasize that the particular representation used depends upon the
programming language, not on the user.
(1,1)
(2,1)
(3,1)
(1,2)
(2,2)
(3,2)
(1,3)
(2,3)
(3,3)
(1,4)
(2,4)
(3,4)
(1,1)
(1,2)
(1,3)
(1,4)
(2,1)
(2,2)
(2,3)
(2,4)
(3,1)
(3,2)
(3,3)
(3,4)
Column1
Column2
Column3
Column4
Column major order
Row1
Row2
Row3
Row major order

Fig. 3-4
Recall that, for a linear array LA, the computer does not keep track of the
address LOCK (LA[K]) of every element LA [K] of LA, but it does keep track
of Base (LA), the address of the first element of LA. The computer uses the
formula
LOC(LA[K]) = Base (LA) + w(K - 1)
MSc. In Software
13
to find the address of LA[K] in time independent of K. (Here w is the number

of words per memory cell for the array LA, and 1 is the lower bound of the
index set of LA).
A similar situation also holds for any two-dimensional m x n array A. That is,
the computer keeps track of Base (A) which is the address of the first
element A[1,1] of array A and computes the address LOC (A[j, k]) of A[j, k]
using the formula
(Column-major order) LOC (A[j, k]) = Base (A) + w[M(k - 1) + (j - 1)]
Or the formula
(Row-major order) LOC (A[j, k]) = Base (A) + w [N(j-1) + (k - 1)]
Again, w denotes the number of words per memory location for the array A.
Note that the formula are linear in j and k, and that one can find that the
address LOC (A[j,k]) is time independent of j and k.
Example
Consider the 35 X 4 matrix array SCORE. Suppose Base (SCORE) = 200 and
there are w = 4 words per memory cell. Furthermore, suppose the
programming language stores two-dimensional arrays using row-major
order. Then the address of SCORE [12, 3], the third test of the twelfth
student is as follows:
LOC(SCORE[12, 3]) = 200 + 4 [4(12 - 1) + (3 - 1)] = 200 + 4[45] = 384
General Multidimensional Arrays

General multidimensional arrays are defined analogously. More specifically,
an n-dimensional m1 x m2 x x mn array B is a collection of m1, m2 .mn
data elements in which each element is specified by a list of n integers, such
as K1, K2,., Kn ,called subscripts, with the property that,
the elements of B with subscripts K1, K2,, Kn will be denoted by
B
K1, K2,, Kn
or
B [K1, K2,, Kn ]
The array will be stored in the memory in a sequence of memory locations.

Specifically, the programming language will store array B either in row-major
order or column-major order. By row-major order, we mean that the
elements are listed so that the subscripts vary like an automobile odometer,
i.e., the last subscript varies first (most rapidly), the next-to-last subscript
varies second (less rapidly), and so on. By column-major order, we mean
MSc. In Software
14
that the elements are listed such that the first subscript varies first (most
rapidly), the second subscript second (less rapidly), and so on.
Summary
In this module we studied,
# A linear array is a list of a finite number n of homogeneous data
elements of the same type.
# Let A be a collection of data elements stored in the memory of
the computer. Suppose we want to print the contents of each
element of A or suppose we want to count the number of
elements of A with a given property. We can accomplish this by
traversing A, that is, by accessing and processing (frequently
called visiting) each element of A exactly once.
# Let A be a collection of data elements in the memory of the
computer. "Inserting" refers to the operation of adding another
element to the collection A, and "deleting" refers to the operation
of removing one of the elements of A.
# Sorting means arranging numerical data in decreasing order or
arranging non-numerical data alphabetically.
# Searching refers to the operation of finding the location LOC of
ITEM in DATA, or printing some message that ITEM does not
appear there.
The linear arrays are called one-dimensional arrays, since each element we
represent in the array is a single subscript. Most programming languages
allow two-dimensional and three-dimensional arrays, i.e., arrays where
elements are referenced, respectively, by two and three subscripts. In fact,
some programming languages allow the number of dimensions for an array
to be as high as 7.
MSc. In Software
4
LINKED LISTS
MAIN POINTS COVERED
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
Introduction
Traversing a Linked List
Searching a Linked List
Memory allocation; garbage collection
Insertion into a linked list
Deletion from a linked list
Header linked lists
Two-ways lists
Binary trees
Representing binary trees in memory
Traversing binary trees
Traversal Algorithms using Stacks
Binary search trees
Searching and inserting in binary search trees
Deleting in binary search trees
Heap; Heapsort
Summary
MSc. In Software
Introduction
linked list is a linear collection of data elements, called nodes, where the linear
order is maintained by pointers. We divide each node into two parts.
# The information of the element

# The link field or nextpointer field, containing the address of the next node in the list
In figure 4-1, each node has two parts. The left part represents the information part of
the node, which may contain an entire record of data items (e.g., NAME, ADDRESS). The
right part represents the next pointer field of the node, and there is an arrow drawn from
it to the next node in the list. The pointer of the last node contains a special value, called
the null pointer, which is any invalid address.
Fig 4-1
We denote null pointer as X in the diagram, which signals the end of the list. The linked
list contains a list of pointer variables. One of them is START, which contains the address
of the first node in the list. We need only this address in START to trace through the list.
If the list contains no nodes it is called null list or empty list and is denoted by the null
pointer in the variable START.
TRAVERSING A LINKED LIST

Let LIST be a linked list. LIST requires two linear arrays e.g. INFO and LINK such that
INFO[k] and LINK[k] contains the information part and the nextpointer field of a node of
LIST respectively. Let the pointer START point to the first element and NULL indicate the
end of LIST. This section presents an algorithm that traverses LIST to process each node
once. You can use this algorithm in future applications.
Our traversing algorithm uses a pointer variable PTR that points to the node that is
currently being processed. Accordingly, PTR->LINK points to the next node to be
MSc. In Software
processed. Thus the assignment PTR = PTR->LINK moves the pointer to the next node
in the list, as pictured in Figure 4-2.
Fig 4-2
Here we have initialized PTR to START. Then processed INFO[PTR], the information at
the first node. Updated PTR by the assignment PTR = PTR->LINK, so that PTR points to
the second node. Then processed INFO[PTR], the information at the second node. Again
updated PTR by the assignment PTR = PTR->LINK, and then processed PTR[INFO],
the information at the third node. And so on. We continued until we reached PTR =
NULL, which signals the end of the list.
A formal presentation of the algorithm is as follows:
Algorithm: (Traversing a Linked List) Let LIST be a linked list in the memory. This
algorithm traverses LIST, applying an operation PROCESS to each element of
LIST. The variable PTR points to the node currently being processed.
1.
2.
3.
4.
Set PTR = START [Initializes pointer PTR]

Repeat Steps 3 and 4 while PTR NULL
Apply PROCESS to PTR->LINK
Set PTR = PTR->LINK [PTR now points to the next node]
[End of Step 2 loop]
5. Exit.
Example 1
The following procedure prints the information at each node of a linked list. Since the
procedure must traverse the list, it will be very similar to the Algorithm above.
Procedure: PRINT (INFO, LINK, START)
This procedure prints the information at each node of the list.
1.
Set PTR = START
2.
Repeat Steps 3 and 4 PTR NULL
3.
Write INFO [PTR]
4.
Set PTR = PTR->LINK [Updates pointer]
MSc. In Software

5.
Return
In other words, the procedure may be obtained by substituting the statement
Write INFO [PTR]
for the processing step in the above Algorithm.
Consider this procedure to find the number NUM of elements in a linked list.
Procedure: COUNT (INFO, LINK, START, NUM)
1.
2.
3.
4.
5.
6.
Set NUM = 0, [Initializes counter]

Set PTR = START [Initializes pointer]
Repeat Steps 4 and 5 while PTR NULL
Set NUM = NUM + 1 [Increases NUM by 1]
Set PTR = PTR->LINK [Updates Pointer]
Return
We can observe that the procedure traverses the linked list in order to count the number
of elements. Hence the procedure is very similar to the above traversing algorithm. Here,
however, we require an initialization step for the variable NUM before traversing the list.
SEARCHING A LINKED LIST

Let LIST be a linked list. Let us now try to search an ITEM in the LIST.
Case 1) Unsorted LIST

Suppose the data in LIST is not sorted, we can still search for ITEM in LIST by traversing
through the list using a pointer variable PTR and comparing ITEM with the contents
INFO[PTR] of each node, one by one, of LIST. Before we update the pointer PTR by PTR
= PTR->LINK we require two tests. First we have to check to see whether we have
reached the end of the list i.e. first we check to see whether
PTR == NULL
If not, we check to see whether
INFO[PTR] == ITEM
The two tests cannot be performed at the same time, since INFO[PTR] is not defined when
PTR == NULL. Accordingly, we use the first test to control the execution of a loop, and we
let the second test take place inside the loop. The algorithm is as follows:
Algorithm
SEARCH (INFO, LINK, START, ITEM, LOC)

List is a linked list in memory. This algorithm finds the location LOC of the
node where ITEM first appears in LIST, or sets LOC = NULL.
MSc. In Software
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Set PTR = START.

Repeat Step 3 while PTR NULL:
if ITEM == INFO[PTR], then;
Set LOC = PTR , and Exit.
else
Set PTR = PTR->LINK. [PTR now points to the next node.]
[End of If structure.]
[Search is unsuccessful.] Set LOC = NULL.
Exit.
Case 2) Sorted LIST

Suppose the data in LIST is sorted. Again we search for ITEM in LIST by traversing the list
using a pointer variable PTR and comparing ITEM with the contents INFO[PTR] of each
node, one by one. Now, however, we can stop once ITEM exceeds INFO[PTR].
The complexity of this algorithm is the same as that of other linear search algorithms. The
worst-case running time is proportional to the number n of elements in LIST, and the
average-case running time is approximately proportional to n/2.
Algorithm: SEARCH(INFO, LINK, START, ITEM, LOC)
LIST is a sorted list in memory. This algorithm finds the location LOC of
the node where ITEM first appears in LIST, or sets LOC = NULL
1. Set PTR = START
2. Repeat Step 3 while PTR NULL
3.
if ITEM < INFO [PTR], then
Set PTR = PTR->LINK [PTR now points to next node.]
else if ITEM == INFO [PTR] then
Set LOC = PTR, and Exit [Search is successful.]
else
Set LOC = NULL, and Exit [ITEM now exceeds INFO[PTR]]
[ End of If structure ]
4. Set LOC = NULL
5. Exit.
Drawback of link list as data structure
You know that with a sorted linear array we can apply a binary search whose running time
is proportional to log2n. On the other hand, a binary search algorithm cannot be applied to
a sorted linked list, since there is no way of indexing the middle element in the list.
Free pool
MSc. In Software
Together with the linked lists in memory, a special list is maintained which consists of
unused memory cells. This list, which has its own pointer, is called the list of available
space or the free-storage list or the free pool.
Suppose we implement linked lists by parallel arrays and insertions and deletions are to
be performed on two linked lists. Then the unused memory cells in the arrays will also be
linked together to form a linked list using AVAIL as its list pointer variable. Hence this
free-storage list will also be called the AVAIL list. Such a data structure will frequently be
denoted by writing
LIST (INFO, LINK, START, AVAIL)
In the pointer-type representation of linked lists, the language provides facilities for
returning memory that is no longer in use. In C the function free returns memory that
has been obtained by a call to malloc or calloc.
Note: More examples on memory allocation are given on the web. Please refer to the site
zeelearn.com
Syntax of malloc:
void *malloc( size_t size )
malloc returns a pointer to space for an object of size size_t, or NULL if the request cannot
be satisfied. The space is initialized to zero bytes.
Syntax of calloc:
void *calloc( size_t nobj , size_t size )
calloc returns a pointer to space for an array of nobj objects, each of size size_t, or NULL.
If the request cannot be satisfied the space is initialized to zero bytes.
Syntax of free:
void free( void *p )
free de-allocates the space pointed to by p. It does nothing if p is NULL. p must be a
pointer to space previously allocated by calloc or malloc.
Garbage Collection
The operating system of a computer may periodically collect all the deleted space onto the
free-storage list. Any technique, which does this collection, is called garbage collection.
Garbage collection usually takes place in two steps. First the computer runs through all
lists, tagging those cells which are currently in use, and then runs through the memory,
collecting all untagged space onto the free-storage list. The garbage collection may take
place when there is only some minimum amount of space or no space at all left in the
MSc. In Software
free-storage list, or when the CPU is idle and has time to do the collection. Generally
speaking, the garbage collection is invisible to the programmer.
Overflow and Underflow

Sometimes we insert new data into a data structure but there is no available space, i.e.,
the free-storage list is empty. This situation is usually called overflow. In such cases, we
need to modify the program for adding spaces to the underlying arrays. Notice that
overflow will occur with our linked lists when AVAIL= NULL and there is an insertion.
Similarly, the term underflow refers to the situation where we want to delete data from a
data structure that is empty. We can handle underflow by printing the message
UNDERFLOW. Observe that underflow will occur with our linked lists when START = NULL
and there is a deletion.
INSERTION INTO A LINKED LIST

Let LIST be a linked list with successive nodes A and B, as pictured in Fig.4-3(a). Suppose
we want to insert a node N into the list between nodes A and B
Fig. 4-3
We have shown insertion in Fig (b). That is, node A now points to the new node N, and
node N points to node B, to which A previously pointed.
Suppose our linked list is maintained in the memory in the form
LIST (INFO, LINK, START, AVAIL)
In the above discussion we did not consider the AVAIL list for the new node N. Let us
consider that the first node in the AVAIL list will be used for the new node N. Thus the
MSc. In Software
above figure looks like Figure 4-4. Observe that three pointer fields are changed as
follows:
(1) The nextpointer field of node A now points to the new node N, to which AVAIL
previously pointed
(2) AVAIL now points to the second node in the free pool, to which node N previously
pointed
(3) The nextpointer field of node N now points to node B, to which node A previously
pointed
Fig. 4-4
There are two special cases: If the new node N is the first node in the list, then START will
point to N; and if the new node N is the last node in the list, then N will contain the null
pointer.
Insertion Algorithms
Algorithms which insert nodes into linked lists come up in various situations. We will
discuss three of them here.
1) Inserting a node at the beginning of the list
2) Inserting a node after the node with a given location
3) Inserting a node into a sorted list
All our algorithms assume that the linked list is in the memory in the form LIST(INFO,
LINK, START,AVAIL) and the variable ITEM contains new information to be added to the
list.
Since our insertion algorithms will use a node in the AVAIL list, all the algorithms will
include the following steps:
MSc. In Software
(a)
Checking to see if space is available in the AVAIL list. If AVAIL is NULL, then the
algorithm will print the message OVERFLOW.
(b)
Removing the first node from the AVAIL list. Using the variable NEW to keep track of
the location of the new node, the steps can be implemented by the pair of
assignments (in this order).
NEW = AVAIL,
(c)
AVIAL = LINK->AVAIL
Copying new information into the new node. In other words,

INFO[NEW] = ITEM
The systematic diagram of the latter two steps is given in Fig. 4-5
Fig. 4-5
Inserting at the Beginning of a List

Suppose we have not sorted our linked list and there is no reason to insert a new node in
any special place in the list, the easiest place to insert the node is at the beginning of the
list. Following is such an algorithm:
Algorithm:
INSFIRST (INFO, LINK, START, AVAIL, ITEM)

This algorithm inserts ITEM as the first node in the list.
1. Set NEW = AVAIL
2. [OVERFLOW?] if AVAIL == NULL, then Write OVERFLOW, and Exit.
3. [Remove first node from AVAIL list.}
Set NEW = AVAIL and AVAIL = AVAIL->LINK
4. Set NEW[INFO] = ITEM. [Copies new data into new node.]
5. Set NEW->LINK = START. [New node now points to original first node.]
6. Set START = NEW. [Changes START so it points to the new node.]
7.
Exit.
MSc. In Software
10
Fig. 4-6
Inserting after a Given Node

Suppose we have been given a value of LOC where either LOC is the location of a node A
in a linked LIST or LOC = NULL. The following is an algorithm which inserts ITEM into LIST
so that ITEM follows node A or, in the first node when LOC = NULL.
Let N denote the new node (whose location is NEW). If LOC = NULL, then N is inserted as
the first node in the LIST as in algorithm INSFIRST. Otherwise, as pictured in Fig. 4-4 , we
let node N point to node B.
NEW->LINK = LOC->LINK
And we let node A point to the new node N by the assignment
LOC->LINK = NEW
A formal statement of the algorithm is as follows:
Algorithm:
1.
2.
3.
4.
5.
INSLOC (INFO, LINK, START, AVAIL, LOC, ITEM)

The algorithm inserts ITEM so that ITEM follows the node with location
LOC or inserts ITEM as the first node when LOC = NULL.
[OVERFLOW?] if AVAIL = NULL, then Write OVERFLOW, and Exit.
[Remove first node from AVAIL list.]
Set NEW = AVAIL and AVAIL = AVAIL->LINK.
Set NEW[INFO] = ITEM [Copies new data into new node.]
if LOC == NULL, then [Insert as first node.]
Set NEW->LINK = START and START = NEW.
else [Insert after node with location LOC.]
Exit
Inserting into a Sorted Linked List

Suppose we want to insert a node called ITEM into a sorted linked LIST. Then ITEM must
be inserted between nodes A and B so that
MSc. In Software
11
INFO(A) < ITEM < INFO(B)

The following is a procedure which finds the location LOC of node A, that is, which finds
the location LOC of the last node in LIST whose value is less than ITEM.
Traverse the list using a pointer variable PTR and comparing ITEM with PTR->INFO at
each node. While traversing, keep track of the location of the preceding node by using a
pointer variable SAVE, as pictured in Fig. 4-7. Thus SAVE and PTR are updated by the
assignment.
SAVE = PTR
and PTR = PTR->LINK
The traversing continues as long as PTR[INFO] > ITEM, or in other words, the traversing
stops as soon as ITEM < PTR[INFO]. The PTR points to node B, so SAVE will contain the
location of the node A.
The formal statement of our procedure is as follows. The cases where the list is empty or
where ITEM < START[INFO], so LOC= NULL, are treated separately, since they do not
involve the variable SAVE.
Procedure:
1.
2.
3.
4.
5.
6.
7.
8.
FINDA (INFO, LINK, START, ITEM LOC)

This procedure finds the location LOC of the last node in a sorted list such
that INFO[LOC] < ITEM, or sets LOC = NULL.
[List empty?] if START == NULL, then Set LOC = NULL, and Return.
[Special case?] if ITEM < START[INFO], then Set LOC = NULL, and
Return.
Set SAVE = START and PTR = START->LINK. [Initializes pointers.]
Repeat Steps 5 and 6 while PTR # NULL
if ITEM < PTR[INFO], then
Set LOC = SAVE, and Return.
Set SAVE = PTR and PTR = PTR->LINK. [Updates Pointers.]
Set LOC = SAVE
Return.
Fig. 4-7
MSc. In Software
12
Now we have all the components to present an algorithm, which inserts ITEM into a linked
list. The simplicity of the algorithm comes from using the previous two procedures.
Algorithm :
1.
2.
3.
INSSRT(INFO, LINK, START, AVAIL, ITEM)

This algorithm inserts ITEM into a sorted linked list.
[Use Procedure FINDA to find the location of the node preceding ITEM]
Call FINDA(START, ITEM, LOC).
[Use Algorithm INSLOC to insert ITEM after the node with location LOC.]
Call INSLOC(START, AVAIL, LOC, ITEM).
Exit.
DELETION FROM A LINKED LIST

Suppose N is a node between nodes A and B in linked list LIST, as pictured in Fig. 4-8.
Suppose node N is to be deleted from the linked list. The deletion occurs as soon as the
nextpointer field of node A is changed so that it points to node B. (Accordingly, when
performing deletions, one must keep track of the address of the node which immediately
precedes the node that is to be deleted.)
Suppose our linked list is maintained in the memory in the form
LIST (START, AVAIL).
Fig. 4-8
The above figure does not take into account the fact that, when a node N is deleted from
our list it will immediately return its memory space to the AVAIL list. Specifically, for
easier processing, it will be returned to the beginning of the AVAIL list. Thus a more exact
schematic diagram of such a deletion is the one in Fig. 4-9.
MSc. In Software
13
Fig. 4-9
Observe that three pointer fields are changed as follows:
(1)
(2)
(3)
The nextpointer field of node A now points to node B, where node N previously
pointed
The nextpointer field of N now points to the original first node in the free pool,
where AVAIL previously pointed
AVAIL now points to the deleted node N
There are two special cases: If the deleted node N is the first node in the list, then START
will point to node B; and if the deleted node N is the last node in the list, then node A will
contain the NULL pointer.
Deleting the Node Following a Given Node

Consider the LIST again. Suppose we have been given the location LOC of a node N in
LIST. Furthermore, we are given the location LOCP of the node preceding N. When N is
the first node, LOCP = NULL. The following algorithm deletes N from the list.
Algorithm :
DEL(INFO, LINK, START, AVAIL, LOCK, LOCP)

This algorithm deletes the node N with location LOC, LOCP is the
location of the node which precedes N or, when N is the first node,
LOCP = NULL.
1.
if LOCP == NULL, then
Set START = START->LINK. [Deletes first node.]
else
Set LOCP->LINK = LOC->LINK . [Deletes node N.]
2.
[Return deleted node to the AVAIL list.]
Set LOC->LINK = AVAIL and AVAIL = LOC.
3.
Exit.
MSc. In Software
14
START = START->LINK is the statement, which effectively deletes the first node from
the list. This covers the case when N is the first node.
Figure 4-10 is the schematic diagram of the assignment START = START->LINK
Fig. 4-10
Figure 4-11 is the schematic diagram of the assignment LOCP->LINK = LOC->LINK
which effectively deletes node N when N is not the first node.
The simplicity of the algorithm comes from the fact that we are already given the location
LOCP of the node, which precedes node N. In many applications, we must first find LOCP.
Fig. 4-11
Deleting the Node for a Given ITEM of Information

Let LIST be a linked list in memory. Suppose we have been given an ITEM of information
and we want to delete from the LIST the first node N that contains ITEM. (If ITEM is a key
value, then only one node can contain ITEM.) Recall that before we delete N from the list,
we have to know the location of the node preceding N. Accordingly, first we will give a
procedure which finds the location LOC of the node N containing ITEM and the location
LOCP of the node preceding node N. If N is the first node, we set LOCP = NULL, and if
ITEM does not appear in LIST, we set LOC = NULL.
Traverse the list, using a pointer variable PTR and then we will compare ITEM with
INFO[PTR] at each node. While traversing, keep track of the location of the preceding
node by using a pointer variable SAVE, as pictured in Figure 4-7.
Thus SAVE and PTR are updated by the assignments.
MSc. In Software
SAVE = PTR
AND
15
PTR = PTR->LINK
We will continue with the traversing as long as PTR->INFO ITEM, or in other words, the
traversing stops as soon as ITEM = PTR->INFO. Then
PTR contains the location LOC of node N and
SAVE contains the location LOCP of the node preceding N
The formal statement of our procedure is as follows: The cases where the list is empty or
where START->INFO = ITEM (i.e., where node N is the first node) are treated separately,
since they do not involve the variable SAVE.
Procedure : FINDB(INFO, LINK, START, ITEM, LOC, LOCP)
This procedure finds the location LOC of the first node N which
contains ITEM and the location LOCP of the node preceding N.
If ITEM does not appear in the list, then the procedure
sets LOC = NULL; and if ITEM appears in the first
Node, then it sets LOCP = NULL.
1. [List empty?] if START = NULL, then
Set LOC = NULL and LOCP = NULL, and Return
[End of if Structure.]
2. [ITEM in first node] if START->INFO = ITEM, then
Set LOC = START and LOCP = NULL, and Return.
[End of if Structure,]
3. Set SAVE = START and PTR = START->LINK. [Initializes pointers.]
4. Repeat Steps 5 and 6 while PTR NULL.
5. if INFO[PTR] = ITEM, then
Set LOC = PTR and LOCP = SAVE, and Return.
[End of it Structure]
6. Set SAVE = PTR and PTR = PTR->LINK. [Updates pointers.]
7. Set LOC = NULL. [Search unsuccessful.]
8. Return.
Now we can easily present an algorithm to delete the first node N from a linked list, which
contains a given ITEM of information. The simplicity of the algorithm comes from the fact
that the task of finding the location of N and the location of its preceding node has already
been done in the above procedure.
Algorithm :
DELETE(INFO, LINK, START, AVAIL, ITEM)
MSc. In Software
16
This algorithm deletes from a linked list the first node N which contains
the given item of information.
1. [Use Procedure above to find the location of N and its preceding
node.]
Call FINDB(START, ITEM, LOC, LOCP]
2. if LOC = NULL, then write ITEM node in list, and Exit.
3. [Delete node.]
if LOCP = NULL, then
Set START= START->LINK. [Deletes first node.]
else
Set LOCP->LINK = LOC->LINK.
[End of If structure.][Return deleted node to the AVAIL list.]
Set LOC->LINK = AVAIL and AVAIL = LOC.
4. Exit.
HEADER LINKED LISTS

A header-linked list is a linked list, which always contains a special node, called the
header node, at the beginning of the list. The following are two kinds of widely used
header lists.
(1)
A grounded header list is a header list where the last node contains the null pointer.
(The term "grounded" comes from the fact that in many cases the electrical ground
symbol is used to indicate the null pointer.)
(2)
A circular header list is a header list where the last node points back to the header
node.
Figure 4-12 contains schematic diagrams of these header lists. Unless otherwise we have
stated or implied, our header lists will always be circular. Accordingly, in such a case, the
header node also acts as a sentinel indicating the end of the list.
We can observe that the list pointer START always points to the header node. Accordingly,
LINK [START] = NULL indicates that a grounded header list is empty, and
LINK[START] = START indicates that a circular header list is empty.
MSc. In Software
17
Fig. 4-12
Although header lists in the memory may maintain our data, the AVAIL list will always be
maintained as an ordinary linked list.
We frequently use Circular header lists instead of ordinary linked lists because many
operations are much easier to state and implement, using header lists. This comes from
the following two properties of circular header lists,
(1)
The null pointer is not used and hence all pointers contain valid addresses
(2)
Every (ordinary) node has a predecessor, so the first node may not require a
special case
The next example illustrates the usefulness of these properties.

Algorithm : (Traversing a Circular Header List) Let LIST be a circular header list in the
memory. This algorithm traverses LIST, applying an operation PROCESS to
each node of LIST.
1. Set PTR = START->LINK. [Initializes the pointer PTR.]
2. Repeat Steps 3 and 4 while PTR START.
3. Apply PROCESS to PTR->INFO.
4. Set PTR = PTR->LINK. [PTR now points to the next node.]
5. Exit.
Example
Suppose LIST is a linked list in the memory, and a specific ITEM of information is given.
MSc. In Software
18
The algorithm below finds the location LOC of the first node in LIST, which contains ITEM
when LIST is an ordinary linked list. The following is such an algorithm when LIST is a
circular header list.
Algorithm : SRCHHL (INFO, LINK, START, ITEM, LOC)
LIST is a circular header list in memory. This algorithm finds the location
LOC of the node where ITEM first appears in LIST or sets LOC = NULL.
1. Set PTR = START->LINK.
2. Repeat while PTR.INFO ITEM and PTR START
Set PTR = PTR->LINK. [PTR now points to the next node.]
[End of loop]
3. if PTR->INFO = ITEM, then
Set LOC = NULL.
4. Exit.
The two tests which control the searching loop (step 2) were not performed at the same
time in the algorithm for ordinary linked list because for ordinary link list PTR->INFO is
not defined when PTR = NULL.
Enough with link list. Take a break and then continue again.
TWO-WAY LISTS
Each list we have discussed above is called a one-way list, since there is only one way we
can traverse the list.
We have introduced a new list structure, called a two-way list, which can be traversed in
two directions: in the usual forward direction from the beginning of the list to the end, or
in the backward direction from the end of the list to the beginning. Furthermore, given the
location LOC of a node N in the list, you now have immediate access to both the next
node and the preceding node in the list. This means, in particular, that you are able to
delete N from the list without traversing any part of the list.
A two-way list is a linear collection of data elements, called nodes, where each node N is
divided into three parts.
(1) An information field INFO which contains the data of N
(2) A pointer field FORW that contains the location of the next node in the list
(3) A pointer field BACK, which contains the location of the preceding node in the list
The list also requires two more pointer variables: FIRST, which points to the first node in
the list, and LAST, which points to the last node in the list. Figure 4-13 shows such a list.
MSc. In Software
19
Observe that the null pointer appears in the FORW field of the last node in the list and
also in the BACK field of the first node in the list.
Fig. 4-13
We can observe that, using the variable FIRST and the pointer field FORW, we can
traverse a two-way list in the forward direction as before. On the other hand, using the
variable LAST and the pointer field BACK, we can also traverse the list in the backward
direction.
Suppose LOCA and LOCB are the locations of nodes A and B, respectively, in a two-way
list. Then the way the pointers FORW and BACK are defined gives us the following:
Pointer property: FORW [LOCA] = LOCB if and only if
BACK [LOCB] = LOCA
In other words, the statement that node B follows node A is equivalent to the statement
that node A precedes node B.
We can maintain two-way lists in memory by means of linear arrays in the same way as
one-way lists except that now we require two pointer arrays, FORW and BACK, instead of
one pointer array LINK. We also require two list pointer variables, FIRST and LAST,
instead of one list pointer variable START. On the other hand, the list AVAIL of available
space in the arrays will still be maintained as a one-way list--using FORW as the pointer
field--since we delete and insert nodes only at the beginning of the AVAIL list.
Two-Way Header Lists

The advantages of a two-way list and a circular header list may be combined into a twoway circular header list and it is pictured in Figure 4-14. The list is circular because the
two end nodes point back to the header node. Observe that such a two-way list requires
only one list pointer variable START, which points to the header node. This is because the
two pointers in the header node point to the two ends of the list.
MSc. In Software
20
Fig. 4-14
TREES
So far, we were concentrating on linear types of data structures: strings, arrays, lists, and
queues. Here we define a nonlinear data structure called a tree. This structure is mainly
used to represent data containing a hierarchical relationship between elements, e.g.
records, family and tables of contents.
First we investigate a special kind of tree, called a binary tree, which can be easily
maintained in the computer. Although such a tree may seem to be very restrictive, we will
see later in the chapter that more general trees may be viewed as binary trees.
BINARY TREES
A binary tree T is defined as a finite set of elements, called nodes, such that:
(a)
(b)
T is empty (called the null tree or empty tree) or

T contains a distinguished node R, called the root of T and the remaining nodes of
T form an ordered pair of disjoint binary trees T1 and T2.
If T does contain a root R, then the two trees T1 and T2 are called the left and right
subtrees of R respectively. If T1 is nonempty then its root is called the left successor of R.
Similarly, if T2 is nonempty, then its root is called the right successor of R. We frequently
represent a binary tree T by means of a diagram. Specifically, the diagram in figure 4-15
represents a binary tree T as follows
MSc. In Software
21
Fig. 4-15
(i) T consists of 11 nodes, represented by the letters A through L excluding I
(ii) The root of T is the node A at the top of the diagram
(iii) A left-downward slanted line from a node N indicates a left successor of N and a rightdownward slanted line from N indicates a right successor of N
Observe that
(a)
(b)
B is left successor and C is a successor of the node A

The left subtree of the root A consists of the nodes B, D, E and the right subtree
consists of the nodes C, E, H, J, K and I
Any node N in a binary tree T has either 0,1 or 2 successors. The nodes A, B, C and H
have successors. The nodes E and J have only one successor and the nodes D, F, G, L and
K have no successors. The nodes with no successors are called terminal nodes.
The above definition of the binary tree T is recursive since T is defined in terms of binary
subtrees T1 and T2. This means, in particular, that every node N of T contains a left and a
right subtree. Moreover, if N is a terminal node then both its left and right subtrees are
empty.
Binary trees T and T` are said to be similar if they have the same structure or in other
words if they have the same shape. The trees are said to be copies if they are similar and
if they have the same contents at corresponding nodes.
Food for thought:
Consider these four binary trees. Which is the right option for similar trees?
MSc. In Software
(a)
(b)
(c)
(d)
22
(a) and (b)

(b) and (c)
(c)and (d)
(a) and (b) and (d)
(d) is the right answer.

The three trees (a), (c) and (d) are similar. In particular, the trees (a) and (c) are copies
since they also have the same data at corresponding nodes. The tree (b) is neither similar
nor a copy of the tree (d) because, in a binary tree we distinguish between a left
successor and a right successor even when there is only one successor.
Terminology
We frequently use terminology to describe family relationships between the nodes of a
tree T. Specifically, suppose N is a node in T with left successor S1 and right successor S2,
then N is called the parent (or father) of S1 and S2. Analogously, S1 is called the left child
(or son) of N, and S2 is called the right child (or son) of N. Furthermore, S1and S2 are said
to be siblings (or brothers). Every node N in a binary tree T, except that the root has a
unique parent, called the predecessor of N.
The terms descendant and ancestor have their usual meaning. That is, a node L is called a
descendant of a node N (and N is called an ancestor of L) if there is a succession of
children from N to L. In particular, L is called a left or right descendant of N according to
whether L belongs to the left or right subtree of N.
Terminology from graph theory and horticulture is also used with a binary tree T.
Specifically, the line drawn from a node N of T to a successor is called an edge, and a
sequence of consecutive edges is called a path. A terminal node is called a leaf, and a
path ending in a leaf is called a branch.
Each node in a binary tree has got a level number. First, we have assigned the root R of
the tree T the level number 0, then for every other node we have assigned a level
number, which is 1 more than the level number of its parent. Furthermore, those nodes
with the same level number are said to be of the same generation.
The depth (or height) of a tree T is the maximum number of nodes in a branch of T. This
turns out to be 1 more than the largest level number of T. Binary trees T and T` are said
to be similar if they have the same structure or, in other words, if they have the same
shape. The trees are said to be copies if they are similar and if they have the same
contents at corresponding nodes.
Extended Binary Trees: 2-Trees

A binary tree T is said to be a 2-tree or an extended binary tree if each node N has either
0 or 2 children. In such a case, the nodes with 2 children are called internal nodes, and
the nodes with 0 children are called external nodes. Sometimes, we distinguish the nodes
in diagrams by using circles for internal nodes and squares for external nodes.
MSc. In Software
23
Fig. 4-16
We get the term extended binary tree from the following operation. Consider any binary
tree T, such as the tree in figure 4-16. Then we may convert T into a 2-tree by replacing
each empty subtree by a new node, as pictured in the figure 4-16(b). Observe that the
tree is, indeed, a 2-tree. Furthermore, the nodes in the original tree T are now the
internal nodes in the extended tree, and the new nodes are the external nodes in the
extended tree.
REPRESENTING BINARY TREES IN MEMORY

Let T be a binary tree. We will discuss two ways of representing T in memory in this
section. The first and usual way is called the link representation of T and is analogous to
the way linked lists are represented in the memory. The second way, which uses a single
array, is called the sequential representation of T. The main requirement of any
representation of T is that one should have direct access to the root R of T and, given any
node N of T, one should have direct access to the children of N.
Representation of a binary tree node in C

struct tree_node{
char (or int) info ;
struct tree_node *left ;
struct tree_node *right ;
}
Where left and right are pointers to the left son and right son of that node.
The root of the tree will be denoted as:
struct
tree_node *root ;
i.e. a pointer to the root node of the tree.
Sequential Representation of Binary Trees
MSc. In Software
24
Suppose T is a binary tree that is complete or nearly complete. Then there is an efficient
way of maintaining T in memory called the sequential representation of T. This
representation uses only a single linear array TREE as follows:
(a) The root R of T is stored in TREE[1].
(b) If a node N occupies TREE[K], then its left child is stored in TREE[2*K] and its
right child is stored in TREE[2*K+1]
Again, NULL is used to indicate an empty subtree. In particular,
TREE[1] = NULL indicates that the tree is empty.
Figure 4-16(b) is the sequential representation of the binary tree T shown in figure 416(a). Observe that we require 14 locations in the array TREE even though T has only 9
nodes. In fact, if we include null entries for the successors of the terminal nodes, then we
would actually require TREE[29] for the right successor of TREE[14]. Generally speaking
the sequential representation of a tree with depth d will require an array with
approximately 2d+1 elements. Accordingly this sequential representation is usually
inefficient unless, as stated above, the binary tree T is complete or nearly complete. For
example, the tree T in Fig. 4-15 has 11 nodes and depth 5, which means it would require
an array with approximately 26 = 64 elements.
Fig. 4-16
TRAVERSING BINARY TREES

There are three standard ways in which we can traverse a binary tree T with root R. These
three algorithms, called preorder, inorder and postorder, are as follows:
Preorder:
(1)
Process the root R.
MSc. In Software
Inorder:
(2)
(3)
Traverse the left subtree of R in preorder.

Traverse the right subtree of R in preorder.
(1)
(2)
(3)
Traverse the left subtree of R in inorder.

Process the root R.
Traverse the right subtree of R in inorder.
Postorder: (1)
(2)
(3)
25
Traverse the left subtree of R in postorder.

Traverse the right subtree of R in postorder.
Process the root R.
We can observe that each algorithm contains the same three steps, and that the left
subtree of R is always traversed before the right subtree. The difference between the
algorithms is the time at which the root R is processed. Specifically, in the "pre"
algorithm, the root R is processed before the subtrees are traversed; in the "in" algorithm,
the root R is processed between the traversals of the subtrees; and in the "post"
algorithm, the root R is processed after the subtrees are traversed.
The three algorithms are sometimes called, respectively, the node-left-right (NLR)
traversal, the left-node-right (LNR) traversal and the left-right-node (LRN) traversal.
Observe that each of the above traversal algorithms is recursively defined, since the
algorithm involves traversing subtrees in the given order. Accordingly, we will expect that
a stack be used when the algorithms are implemented on the computer.
Note: More examples on binary tree are given on the web. Please refer to the site
zeelearn.com
TRAVERSAL ALGORITHMS USING STACKS

Suppose a binary tree T is maintained in memory by some linked representation
TREE (INFO, LEFT, RIGHT, ROOT)
We will discuss the implementation of the three standard traversals of T, which were
defined recursively in the last section, by means of non-recursive procedures using stacks.
We will discuss the three traversals separately.
Preorder Traversal
The preorder traversal algorithm uses a variable PTR (pointer), which will contain the
location of the node N currently being scanned. This is pictured in this figure 4-17, where
L(N) denotes the left child of node N and R(N) denotes the right child. The algorithm also
uses an array STACK, which will hold the addresses of nodes for future processing.
MSc. In Software
26
Fig. 4-17
Algorithm: Initially push NULL onto STACK and then set PTR = ROOT. Then repeat
the following steps until PTR = NULL or, equivalently, while PTR NULL
(a)
(b)
Proceed down the left-most path rooted at PTR, processing each node N on the
path and pushing each right child R(N), if any, onto STACK. The traversing ends
after a node N with no left child L(N) is processed. (Thus PTR is updated using
the assignment PTR = LEFT [PTR], and the traversing stops when LEFT [PTR]
== NULL.)
[Backtracking.] Pop and assign to PTR the top element on STACK.
If PTR NULL, then return to Step (a); otherwise Exit.
(Note that the initial element NULL on STACK is used as a sentinel.)

We simulate the algorithm in the next example. Although the example works with the
nodes themselves, in actual practice the locations of the nodes are assigned to PTR and
are pushed onto the STACK.
Algorithm : PREORD (INFO, LEFT, RIGHT, ROOT)
A binary tree T is in memory. The algorithm does a
preorder traversal of T, applying an operation PROCESS to each of its
nodes. An array STACK is used to temporarily hold the addresses of
nodes.
1. [Initially push NULL onto STACK, and initialize PTR.]
set TOP = 1, STACK{1} = NULL and PTR = ROOT
2. Repeat Steps 3 to 5 while PTR = NULL
3. Apply PROCESS to INFO [PTR].
4. [Right child?]
if RIGHT[PTR] = NULL, then [Push on STACK.]
Set TOP = TOP + 1, and STACK[TOP] = RIGHT [PTR].
[End of if structure.]
5. [Let child?]
if LEFT [PTR] = NULL, then
Set PTR = STACK [TOP] and TOP = TOP - 1.
6. Exit.
MSc. In Software
27
Inorder Traversal
The inorder traversal algorithm also uses a variable pointer PTR, which will contain the
location of the node N currently being scanned, and an array STACK, which will hold the
addresses of nodes for future processing. In fact, with this algorithm, a node is processed
only when it is popped from STACK.
Algorithm: Initially push NULL onto STACK (for a sentinel) and then set PTR = ROOT.
Then repeat the following steps until NULL is popped from STACK.
(a) Proceed down the left-most path rooted at PTR, pushing each node N
onto STACK and stopping when a node N with no left child is pushed onto
STACK.
(b)[Backtracking.] Pop and process the nodes on STACK. If NULL is popped,
then Exit. If a node N with a right child R (N) is processed, set PTR = R (N)
(by assigning PTR = RIGHT [PTR] and return to Step (a).
We emphasize that a node N is processed only when it is popped from STACK.
Note: Examples on Traversal algorithm are given on the web. Please refer to the site
zeelearn.com
Postorder Traversal
The postorder traversal algorithm is more complicated than the proceeding two
algorithms, because here we may have to save a node N in two different situations. We
distinguish between the two cases by pushing either N or its negative, -N, onto STACK.
(In actual practice, the location of N is pushed onto STACK, so -N has the obvious
meaning.) Again, a variable PTR (pointer) is used which contains the location of the node
N that is currently being scanned, as shown in this figure 4-17.
Algorithm: Initially push NULL into STACK (as a sentinel) and then set PTR = ROOT.
Then repeat the following steps until NULL is popped from STACK.
(a) Proceed down the left-most path rooted at PTR. At each node N of the path,
push N onto STACK and, if N has a right child R (N), push -R(N) onto STACK.
(b)[Backtracking.] Pop and process positive nodes on STACK. If NULL is popped,
then Exit. If a negative node is popped, that is, if PTR = - N for some node N, set
PTR= N (by assigning PTR: = -PTR) and return to Step (a).
We emphasize that a node N is processed only when it is popped from STACK and it is
positive.
BINARY SEARCH TREES

We will discuss one of the most important data structures in computer science, a binary
search tree. This structure enables you to search for and find an element with an average
running time f(n) = O(log2n). It also enables us to easily insert and delete elements.
This structure contrasts with the following structures.
MSc. In Software
28
(a)
Sorted linear array. Here you can search for and find an element with a
running time f(n) = O(log2 n), but it is expensive to insert and delete
elements.
(b)
Linked list. Here you can easily insert and delete elements, but it is
expensive to search for and find an element, since you must use a linear
search with running time f(n) = O(n).
Although each node in a binary search tree may contain an entire record of data, the
definition of the binary tree depends on a given field whose values is distinct and may be
ordered.
Suppose T is a binary tree. Then we call T a binary search tree (or binary sorted tree) if
each node N of T has the following property: The value at N is greater than every value in
the left subtree of N and is less than every value in the right subtree of N. (It is not
difficult to see that this property guarantees that the inorder traversal of T will yield a
sorted listing of the elements of T.)
Note: Examples on binary search tree are given on the web. Please refer to the site
zeelearn.com
SEARCHING AND INSERTING IN BINARY SEARCH TREES

Suppose T is a binary search tree. We will discuss the basic operations of searching and
inserting with respect to T. In fact, a single search and insertion algorithm will give the
searching and inserting.
Suppose we give an ITEM of information. The following algorithm finds the location of
ITEM in the binary search tree T, or inserts ITEM as a new node in its appropriate place in
the tree.
(a) Compare ITEM with the root node N of the tree.
(i) If ITEM < N, proceed to the left child of N.
(ii) If ITEM > N, proceed to the right child of N.
(b) Repeat Step (a) until one of the following occurs:
(i) We meet a node N such that ITEM = N. In this case the search is successful.
(ii) We meet an empty subtree, which indicates that the search is unsuccessful, and
we insert ITEM in place of the empty subtree.
In other words, proceed from the root R down through the three T until finding ITEM in T
or inserting ITEM as a terminal node in T.
Example
Suppose the following six numbers are inserted in order into an empty binary search tree:
43, 63, 54, 39, 60, 16
MSc. In Software
29
This figure 4-18 shows the six stages of the tree. We emphasize that if the six numbers
were given in a different order, then the tree might be different and we might have a
different depth.
Fig. 4-18
The formal presentation of our search and insertion algorithm will use the following
procedure, which finds the locations of a given ITEM and its parent. The procedure
traverses down the tree using the pointer PTR and the pointer SAVE for the parent node.
This procedure will also be used in the next section, on deletion.
Observe that, in Step 4, there are three possibilities: (1) the tree is empty, (2) ITEM is
added as a left child and (3) ITEM is added as a right child.
Procedure : FIND (INFO, LEFT, RIGHT,ROOT, ITEM, LOC, PAR)

A binary search tree T is in memory and an ITEM of information is given. This
procedure finds the location LOC of ITEM in T and also the location PAR of the
parent of ITEM. There are three special cases:
(i) LOC = NULL and PAR = NULL will indicate that the tree is empty.
(ii) LOC NULL and PAR=NULL will indicate that ITEM is
the root of T.
(iii) LOC = NULL and PAR NULL will indicate that ITEM is not in
T and can be added to T as a child of the node N with location PAR.
1. [Tree empty?]
if ROOT = NULL, then Set LOC = NULL and PAR = NULL, and Return.
2. [ITEM at root?]
MSc. In Software
30
if ITEM = ROOT->INFO, then Set LOC = ROOT and PAR = NULL, and
Return.
3. [Initialize pointers PTR and SAVE.]
if ITEM < ROOT->INFO, then
Set PTR = ROOT->RIGHT and SAVE = ROOT.
else.
Set PTR = ROOT->RIGHT and SAVE = ROOT.
[End of If Structure.]
4. Repeat Steps 5 and 6 while PTR NULL
5.
[ITEM found?]
if ITEM = PTR->INFO, then Set LOC = PTR and
PAR = SAVE, and Return.
6.
if ITEM < PTR->INFO, then
Set SAVE = PTR and PTR = PTR->LEFT.
else
Set SAVE = PTR and PTR = PTR->RIGHT.
[End of If Structure.]
7. [Search unsuccessful.] Set LOC = NULL and PAR = SAVE.
8. Exit.
Notice that, in step 6, we move to the left child or the right child according to whether
ITEM < PTR->INFO or ITEM > PTR->INFO.
Algorithm: INSBST(INFO, LEFT, RIGHT, ROOT, AVAIL, ITEM, LOC)
A binary search tree T is in memory and an ITEM of information is given. This
algorithm finds the location LOC of ITEM in T or adds ITEM as a new node in
T at location LOC.
1. Call FIND(INFO, LEFT, RIGHT, ROOT, ITEM, LOC, PAR).
2. if LOC NULL, then Exit.
3. [Copy ITEM into new node in AVAIL list.]
(a) if AVAIL = NULL, then Write OVERFLOW, and Exit.
(c) Set LOC = NEW, NEW->LEFT = NULL and
NEW->RIGHT = NULL
4. [Add ITEM to tree.]
if PAR = NULL, then
Set ROOT = NEW
else if ITEM < PAR->INFO = NEW
Set PAR->LEFT = NEW
else
Set PAR-> RIGHT = NEW
5. Exit.
MSc. In Software
Observe
i)
ii)
iii)
31
that, in step 4, there are three possibilities:

The tree is empty.
ITEM is added as left child.
ITEM is added as right child.
Complexity of the Searching Algorithm

Suppose we are searching for an item of information in a binary search tree T. We have to
observe the depth of the tree for the number of comparisons. This comes from the fact
that we proceed down a single path of the tree. Accordingly, the running time of the
search will be proportional to the depth of the tree.
Suppose we have been given n data item, A1, A2,....An, and suppose we insert the items in
order into a binary search tree T. It can be shown that the average depth of the n trees is
approximately c log2 n, where c = 1.4. Accordingly, the average time f(n) to search for an
item in a binary tree T with n elements is proportional to log2 n, that is f(n) = O(log2 n).
Application of Binary Search Trees

Consider a collection of n data items, A1, A2,....An. Suppose we want to find and delete all
duplicates in the collection. One straightforward way to do this is as follows:
Algorithm A: Scan the elements from A1 to An (that is, from left to right).
(a) For each element AK, compare AK with A1, A2,.....,AK-1, that is,
AK with those elements which precede AK.
(b) If AK does occur among A1, A2,...., Ak-1, then delete AK.
compare
After all the elements have been scanned, there will be no duplicates.
Example
Suppose Algorithm a is applied to the following list of 15 numbers:
14, 10, 17, 12,10, 11, 20, 12, 18, 25, 20, 8 22, 11, 23
Observe that the first four numbers (14, 10, 18 and 12) are not deleted. However,
A5
A8
A11
A14
=
=
=
=
10 is
12 is
20 is
11 is
deleted, since A5 = A2
When Algorithm A is finished running, the 11 numbers

14, 10, 17, 12, 11, 20, 18, 25, 8 22, 23
MSc. In Software
32
which are all distinct, will remain.

Consider now the time complexity of algorithm A, which is determined by the number of
comparisons. First of all, we assume that the number d of duplicates is very small
compared with the number n of data items. Observe that the step involving AK will require
approximately k -1 comparisons, since we compare AK with items A1, A2,...AK-1.
Accordingly, the number f(n) of comparisons required by Algorithm A is approximately
0 + 1 + 2 + 3 + ....+ (n-2) + (n-1) = (n-1)n/2 = O(n2)
For example, for n = 1000 items, Algorithm A will require approximately 500,000
comparisons. In other words, the running time of Algorithm A is proportional to n2.
Using a binary search tree, we can give another algorithm to find the duplicates in the set
A1, A2,..., An of n data items.
Algorithm B: Build a binary search tree T using the elements A1, A2,...., An. In building
the tree, delete AK from the list whenever the value of AK already appears
in the tree.
The main advantage of Algorithm B is that each element AK is compared only with the
elements in a single branch of the tree. It can be shown that the average length of such a
branch is approximately clog 2 k, where c = 1.4. Accordingly, the total number f(n) of
comparisons required by Algorithm B is approximately nlog2n, that is, f(n) = O(nlog2n).
For example, for n = 1000, Algorithm B will require approximately 10,000 comparisons
rather than the 500 000 comparisons of Algorithm A. (We note that, for the worst case,
the number of comparisons for Algorithm B is the same as for Algorithm A.)
Note: Explanation of Deletion from binary search and examples on deletion from binary
search tree are given on the web. Please refer to the site zeelearn.com
HEAP; HEAPSORT
In this section we will discuss another tree structure, called a heap. We will use heap in an
elegant sorting algorithm called heapsort. Suppose H is a complete binary tree with n
elements. (Unless otherwise stated, we assume that H is maintained in memory by a
linear array TREE using the sequential representation of H, not a linked representation.)
Then H is called a heap, or a maxheap, if each node N of H has the following property:
The value at N is greater than or equal to the value at each of the children of N.
Accordingly, the value at N is greater than or equal to the value at any of the descendants
of N.
A minheap is defined analogously: The value at N is less than or equal to the value at any
of the children of N.
Example
MSc. In Software
33
Consider the complete tree H in this figure 4-19. Observe that H is a heap. This means, in
particular, that the largest element in H appears at the "top" of the heap, that is, at the
root of the tree. This figure 4-19(b) shows the sequential representation of H by the array
TREE. That is, TREE [1] is the root of the tree H, and the left and right children of node
TREE [K] are, respectively, TREE[2K] and TREE [2K + 1]. This means, in particular, that
the parent of any nonroot node TREE[J] is the node TREE [J / 2] (where J / 2 means
integer division). Observe that the nodes of H on the same level appear one after the
other in the array TREE.
101 92 97 67 57 97 50 67 37 50 57 66 67 26 39 19 41 31 27 25
1 2
3 4
5 6
7 8
9 10 11 12 13 14 15 16 17 18 19 20
(b) Sequential representation
Fig. 4-19
Inserting into a Heap
Suppose H is a heap with N elements, and suppose an ITEM of information is given. We
insert ITEM into the heap H as follows:
(1) First adjoin ITEM at the end of H so that H is still a complete tree, but not necessarily
a heap.
(2) Then let ITEM rise to its "appropriate place" in H so that H is finally a heap.
We will illustrate the way this procedure works before stating the procedure formally.
Example
Consider the heap H in this figure 4-20. Suppose we want to add ITEM = 71 to H. First we
adjoin 71 as the next element in the complete tree; that is, we set TREET [21] = 71. Then
MSc. In Software
34
71 is the right child of TREE [10] = 50. The path from 71 to the root of H is pictured in
this figure 4-20(a). We now find the appropriate place of 71 in the heap as follows:
(a)
Compare 71 with its parent, 50. Since 71 is greater than 50, interchange 71 and
50; the path will now look like this 4-20(b).
(b)
Compare 71 with its new parent, 57. Since 71 is greater than 57, interchange 71
and 57; the path will now look like this 4-20(c).
(c)
Compare 71 with its new parent, 92. Since 71 does not exceed 92, ITEM = 71 has
risen to is appropriate place in H.
This figure 4-23(d) shows the final tree. A dotted line indicates that an exchange has
taken place.
Fig. 4-23
Note: More examples on Inserting into heapsort are given on the web. Please refer to the
site zeelearn.com
Procedure:
INSHEAP (TREE, N, ITEM)
MSc. In Software
35
A heap H with N elements is stored in the array TREE, and an ITEM of

information is given. This procedure inserts ITEM as a new element of H.
PTR gives the location of ITEM as it rises in the tree, and PAR denotes the
location of the parent of ITEM
1. [Add new node to H and initialize PTR.]
Set N = N + 1 and PTR = N
2. [Find location to insert ITEM.]
Repeat Steps 3 to 6 while PTR < 1.
3.
Set PAR = [PTR/2]. [Location of parent node.]
4.
If ITEM < TREE [PAR], then:
Set TREE [PTR] = ITEM, and Return.
5.
Set TREE [PTR] =TREE[PAR]. [Moves node down.]
6.
Set PTR = PAR. [Updates PTR.]
7. [Assign ITEM as the root of H.]
Set TREE[1] = ITEM
8. Return.
Observe that ITEM is not assigned to an element of the array TREE until the appropriate
place for ITEM is found. Step 7 takes care of the special case that ITEM rises to the root
TREE [1].
Suppose an array A with N elements is given. By repeatedly applying the above Procedure
to A, that is, by executing
Call INSHEAP(A, J, A[J + 1])
For j = 1, 2,.......N - 1, we can build a heap H out of the array A.
Deleting the root of a Heap
Suppose H is a heap, with N elements, and suppose we want to delete the root R of H.
This is accomplished as follows:
(1) Assign the root R to some variable ITEM
(2) Replace the deleted node R by the last node L of H so that H is still a complete
tree, but not necessarily a heap.
(3) (Reheap) let L sink to its appropriate place in H so that H is finally a heap.
Again we illustrate the way the procedure works before stating the procedure formally.
Example
MSc. In Software
36
Consider the heap H in this figure 4-24(a), where R = 95 is the root and L=22 is the last
node of the tree. Step 1 of the above procedure deletes R = 95 by L=22. This gives the
complete tree in this figure 4-24(b), which is not a heap. Observe, however, that both the
right and left subtrees of 22 are still heaps. Applying Step 3, we find the appropriate place
of 22 in the heap as follows:
(a)
Compare 22 with its two children, 85 and 70. Since 22 is less than the larger
child 85, interchange 22 and 85 so the tree now looks like this 4-24(c).
(b)
Compare 22 with its two new children, 55 and 33. Since 22 is less than the
larger child, 55, interchange 22 and 55 so the tree now looks like this 4-24(d).
(c)
Compare 22 with its new children, 15 and 20. Since 22 is greater than both
children, node 22 has dropped to its appropriate place in H.
Thus Fig. 4-24(d) is the required heap H without its original root R.
Fig. 4-24
Remark: As we are inserting an element into a heap, we must verify that the above
procedure does always yield a heap as a final tree. Again we leave this verification to the
reader. We also note that Step 3 of the procedure may not end until the node L reaches
the bottom of the tree. i.e. until L has no children.
The formal statement of our procedure is as follows.
MSc. In Software
37
Procedure: DELHEAP(TREE, N, ITEM)

A heap H with N elements is stored in the array TREE. This procedure assigns
the root TREE[1] of H to the variable ITEM and then reheaps the remaining
elements. The variable LAST saves the value of the original last node of H.
The pointers PTR, LEFT and RIGHT gives the locations of LAST and its left
and right children as LAST sinks in the tree.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Set ITEM = TREE[1]. [Remove root of H]

Set LAST = TREE[N] and N= N-1. [Removes last node of H.]
Set PTR =1,LEFT=2 and RIGHT = 3. [Initializes pointers]
Repeat Steps 5 to 7 while RIGHT N
if LAST TREE[LEFT] and LAST TREE[RIGHT], then
Set TREE[PTR] =TREE[LEFT] and Return.
if TREE[RIGHT] TREE[LEFT], then
Set TREE[PTR] = TREE[LEFT] and PTR= LEFT.
else
Set TREE[PTR]= TREE[RIGHT] and PTR= RIGHT.
Set LEFT =2*PTR and RIGHT= LEFT+1
if LEFT=N and if LAST<TREE[LEFT], then Set PTR = LEFT.
Set TREE[PTR] = LAST.
Return.
The step 4 loop repeats as long as LAST has a right child. Step 8 takes care of the special
case in which LAST does not have a right child but does have a left child (which has to be
the last node in H). The reason for the two if statement in Step 8 is that TREE[LEFT]
may not defined when LEFT > N.
Application to Sorting
Suppose an array A with N elements is given. The heapsort algorithm to sort A consists of
the two following phases:
Phase A: Build a heap H out of the elements of A.
Phase B: Repeatedly delete the root element of H.
Since the root of H always contains the largest node in H, Phase B deletes the elements of
A in decreasing order. A formal statement of the algorithm is as follows.
Algorithm: HEAPSORT(A, N)
An array A with N elements is given. This algorithm sorts the elements of A.
1. [Build a heap H, using Procedure 7.9.]
Repeat for j=1 to N-1
MSc. In Software
38
Call INSHEAP(A, J, A[J=1]).

[End of loop.]
2. [Sort A by repeatedly deleting the root of H]
Repeat while N > 1
(a) Call DELHEAP(A,N,ITEM).
(b) Set A[N+1] =ITEM.
[End of Loop.]
3. Exit
The purpose of step2(b) is to save space. That is, one could use another array B to hold
the sorted elements of A and replace Step 2(b) by
Set B[N+1]= ITEM
However, the reader can verify that the given Step2(b) does not interfere with the
algorithm, since A[N+1] does not belong to the heap H.
Complexity of Heapsort
Suppose the heapsort algorithm is applied to an array A with n elements. The algorithm
has two phases, and we analyze the complexity of each phase separately.
Phase A. Suppose H is a heap. Observe that the number of comparisons to find the
appropriate place of a new element ITEM in H cannot exceed the depth of H. Since H is a
complete tree, its depth is bounded by log2 m where m is the number of elements in H.
Accordingly the total number g(n) of comparisons to insert the n elements of A into H is
bounded as follows:
g(n) n log2n
Consequently the running time of Phase A of heapsort is proportional to n log2 n.
Phase B. Suppose H is a complete tree with m elements, and suppose the left and right
subtrees of H are heaps and L is the root of H. Observe that reheaping uses 4
comparisons to move the node L one step down the tree H. Since the depth of H does not
exceed log2 m, reheaping uses at most 4 log2 m comparisons to find the appropriate place
of L in the tree. This means that the total number h(n) of comparisons to delete the
elements of A from H, which requires reheaping n times is bounded as follows:
H(n) 4n log2 n
Accordingly, the running time of Phase B of heapsort is also proportional to n log2n.
Since each phase requires time proportional to n log2n, the running time to sort the nelement array A using heapsort is proportional to n log2n, that is, f(n)=O(n log2,n).
Observe that this gives a worst-case complexity of the heapsort algorithm.
MSc. In Software
39
Summary
# A linked list is a linear collection of data elements, called nodes, where the linear
order (or linearity) is given by means of pointers. We divide each node into two
parts. The first part contains the information of the element, and the second part,
called the link field or next pointer field, contains the address of the next node in
the list.
#
Languages like C that support dynamic memory allocation with structures and
pointers use the following technique. The linked list node is defined as a structure.
# If ITEM is actually a key value and we are searching through a file for the record
containing ITEM then ITEM can appear only once in LIST.
# The list, which has its own pointer, is called the list of available space or the freestorage list or the free pool.
#
Sometimes we insert new data into a data structure but there is no available space,
i.e., the free-storage list is empty. This situation is usually called overflow. The
term underflow refers to the situation where we want to delete data from a data
structure that is empty.
Suppose we have not sorted our linked list and there is no reason to insert a new
node in any special place in the list. Then the easiest place to insert the node is at
the beginning of the list.
# A header-linked list is a linked list, which always contains a special node, called the
header node, at the beginning of the list.
# We frequently use Circular header lists instead of ordinary linked lists because
many operations are much easier to state and implement-using header lists.
# We have introduced a new list structure, called a two-way list, which can be
traversed in two directions: in the usual forward direction from the beginning of the
list to the end, or in the backward direction from the end of the list to the
beginning.
# Binary tree is a nonlinear data structure called a tree. This structure is mainly used
to represent data containing a hierarchical relationship between elements, e.g.
records, family and tables of contents.
# We frequently use terminology describing family relationships to describe
relationships between the nodes of a tree T. Specifically, suppose N is a node in T
with left successor S1 and right successor S2. Then N is called the parent (or father)
of S1 and S2. Analogously, S1 is called the left child (or son) of N, and S2 is called
the right child (or son) of N. Furthermore, S1and S2 are said to be siblings (or
brothers). Every node N in a binary tree T, except the root, has a unique parent,
called the predecessor of N.
MSc. In Software
40
MSc. In Software
5
STACKS, QUEUES,
RECURSION
Main Points Covered
!
!
!
!
!
!
!
!
!
!
!
!
!
Introduction
Stacks
Array Representation Of Stacks
Arithmetic Expressions; Polish Notation
Consider this C-Program on Stack
Quicksort
Recursion
Factorial Function
Fibonacci Sequence
Divide-and-Conquer Algorithms
Towers Of Hanoi
Queues
Representation of Queues
MSc. In Software
Introduction
e have already learned linear lists and linear arrays, which allows us to
insert and delete elements at any place in the list i.e. at the beginning, at
the end, or in the middle. There are certain frequent situations in computer
science when we want to restrict insertions and deletions so that they can take
place only at the beginning or the end of the list, not in the middle. Two of the data
structures that are useful in such situations are stacks and queues.
Stack : A stack is a linear structure in which items may be added or removed only
at one end. An example of such a structure is, a stack of dishes. Observe that an
item may be added or removed only from the top of any of the stacks. This means,
in particular, that the last item to be added to a stack is the first item to be
removed. Accordingly, stacks are also called last-in first-out (LIFO) lists. Other
names used for stacks are "piles" and "push-down lists." Although the stack may
seem to be a very restricted type of data structure, it has many important
applications in computer science.
Queue: A queue is a linear list in which items may be added only at one end and
items may be removed only at the other end. The name "queue" likely comes from
the everyday use of the term. Observe a queue at the bus stop. Each new person
who comes takes his or her place at the end of the line, and when the bus comes,
the people at the front of the line board first. That is, the person who comes first,
boards first and who comes last, boards last. Thus queues are also called first-in
first-out (FIFO) lists. Another example of a queue is a batch of jobs waiting to be
processed, assuming no job has higher priority than the others.
STACKS : Special terminology is used for two basic operations associated with
stacks.
(a) "Push" is the term used to insert an element into a stack.
(b) "Pop" is the term used to delete an element from a stack.
Remember that these terms are used only with stacks, not with other data
structures.
Suppose the following 6 elements are pushed, in order, onto an empty stack:
A, B, C, D, E, F
The figure 5-1 shows three ways of picturing such a stack. For notational
convenience, we will frequently designate the stack by writing.
STACK: A , B, C, D, E, F
The implication is that the right-most element is the top element. We emphasize
that, regardless of the way a stack is described, its underlying property is that
MSc. In Software
insertions and deletions can occur only at the top of the stack. This means E cannot
be deleted before F is deleted, D cannot be deleted before E and F are deleted, and
so on. Consequently, the elements may be popped from the stack only in the
reverse order of that in which they were pushed onto the stack.
top
(b)
F
E
D
C
B
A
Fig 5-1 Diagrams of stacks
Fig 5-1
Postponed Decisions
We use stacks frequently to indicate the order of the processing of data when
certain steps of the processing must be postponed until other conditions are
fulfilled. We have illustrated it below.
Suppose we are processing some project A, and desire to move on to project B,
then we need to complete project B before we return to project A. We place the
folder containing the data of A onto a stack, as pictured in the figure 5-2(a) and
begin to process B. However, suppose that while processing B we are led to project
C, for the same reason. Then we place B on the stack above A, as pictured in figure
5-2(b) and begin to process C. Furthermore, suppose that while processing C we
are likewise led to project D. Then we place C on the stack above B, as pictured in
the figure 5-2(c) and begin to process D.
(a)
(b)
(c)
(d)
(e)
(f)
MSc. In Software
Fig. 5-2
On the other hand, suppose we are able to complete the processing of project D.
Then the only project we may continue to process is project C, which is on top of
the stack. Hence we remove folder C from the stack, leaving the stack as pictured
in figure 5-2(d) and continue to process C. Similarly, after completing the
processing of C, we remove folder B from the stack, leaving the stack as pictured in
figure 5-2(e) and continue to process B. Finally, after completing the processing of
B, we remove the last folder A, from the stack, leaving the empty stack pictured in
figure 5-2(f) and continue the processing of our original project A.
Notice that, at each stage of the above processing, the stack automatically
maintains the order that is required to complete the processing.
ARRAY REPRESENTATION OF STACKS

We can represent stacks in the computer in various ways. Usually we denote it by
means of a linear array. Unless otherwise stated or implied, each of our stacks will
be maintained by a linear array STACK; a pointer the stack can hold variable TOP,
which contains the location of the top element of the stack; and a variable MAXSTK,
which gives the maximum number of elements that the stack can hold. The
condition
TOP = 0
or
TOP = NULL
will indicate that the stack is empty.
Figure 5-3 pictures such an array representation of a stack. Since TOP = 3, the
stack has three elements, XXX, YYY and ZZZ; and since MAXSTK = 8, there is room
for 5 more items in the stack.
Fig 5-3
Overflow:
MSc. In Software
The procedure for adding (pushing) an element is called PUSH and removing (pop)
an item is called POP. In executing the procedure PUSH, we have to test whether
there is room in the stack for the new item. If not, then we have the condition
known as overflow.
Underflow:
The case is same in executing the procedure POP. We must first test whether there
is an element in the stack to be deleted. If not, then we have the condition known
as underflow.
Procedure:
Procedure:
PUSH(STACK, TOP, MAXSTK, ITEM)

This procedure pushes an ITEM onto a stack.
1. [Stack already filled?]
if TOP = MAXSTK, then print OVERFLOW, and return.
2. Set TOP = TOP + 1. [Increases TOP by 1.]
3. Set STACK [TOP] = ITEM.
[Inserts ITEM in new TOP Position]
4. Return
POP(STACK, TOP, ITEM)

This procedure deletes the top element of STACK and assigns
it to the variable ITEM.
1. [Stack has an item to be removed]
if TOP = 0 then Print UNDERFLOW, and return.
2. Set ITEM = STACK [TOP]. Assigns TOP element to ITEM.]
3. Set TOP = TOP - 1. [Decreases TOP by 1.]
4. Return
It is observed that frequently, TOP and MAXSTK are global variables; hence the
procedures may be called using only
PUSH (STACK, ITEM)
and
POP (STACK, ITEM)
respectively. We note that the value to TOP is changed before the insertion in
PUSH, but the value of TOP is changed after the deletion in POP.
Lets see few examples on the above algorithm:
(a)
Consider the stack in figure 5-3. We simulate the operation PUSH (STACK,
WWW):
1.
Since TOP = 3, control is transferred to Step 2.
2.
TOP = 3 + 1 = 4
MSc. In Software
3.
STACK [TOP] = STACK [4] = WWW.
4.
Return
Note that W is now the top element in the stack.

(b)
Consider again the stack in figure 5-3. This time we simulate the operation
POP (STACK, ITEM):
1.
Since TOP = 3, control is transferred to Step 2.
2.
ITEM = ZZZ
3.
Top = 3 - 1 = 2
4.
Return.
Observe that STACK [TOP] = STACK [2] = YYY is now the top element in the stack.
Minimizing Overflow
There is an essential difference between underflow and overflow in dealing with
stacks. Underflow depends exclusively on the given algorithm and the given input
data, and hence there is no direct control by the programmer. Overflow, on the
other hand, depends on the arbitrary choice of the programmer for the amount of
memory space reserved for each stack, and this choice does influence the number
of times overflow may occur.
Generally, the number of elements in a stack fluctuates as elements are added to
or removed from a stack. Accordingly, the particular choice of the amount of
memory for a given stack involves a time-space tradeoff. Initially reserving a great
deal of space for each stack will decrease the number of time overflow may occur.
However, this may be an expensive use of space if most of the space is seldom
used. On the other hand, reserving a small amount of space for each stack may
increase the number of times overflow occurs. The time required for resolving an
overflow, such as by adding space to the stack, may be more expensive than the
space saved.
Various techniques have been developed which modify the array representation of
stacks so that the amount of space reserved for more than one stack may be more
efficiently used. Most of these techniques lie beyond the scope of this text. One of
these techniques is shown in fig 5.4.
Suppose we have been given an algorithm, which requires two stacks, A and B. We
can define an array STACKA with n1 elements for stack A and an array STACKB with
n2 elements for stack B. Overflow will occur when either STACKA contains more
than n1 elements or STACKB contains more than n2 elements.
MSc. In Software
Suppose we define a single array STACK with n = n1 + n2 elements for stacks A and
B together. As pictured in the figure below, we define STACK [1] as the bottom of
stack A and let A grow to the right, and we define STACK [n] as the bottom of
stack B and let B 'grow' to the left. In this case, overflow will occur only when A and
B together have more than n = n1 + n2 elements. This technique will usually
decrease the number of times overflow occurs even though we have not increased
the total amount of space reserved for the two stacks. In using this data structure,
the operations of PUSH and POP need to be modified.
Fig. 5-4
ARITHMETIC EXPRESSIONS; POLISH NOTATION

Let Q be an arithmetic expression involving constants and operations. In this
section we have shown an algorithm, which finds the value of Q by using reverse
Polish (postfix) notation. We will see that the stack is an essential tool in this
algorithm.
Recall that the binary operations in Q may have different levels of precedence.
Specifically, we assume the following three levels of precedence for the usual five
binary operations:
Highest
: Exponentiation (^ )
Next highest: Multiplication (*) and division (/)
Lowest
: Addition (+) and subtraction (-)
Food for thought
Suppose we want to evaluate the following parenthesis free arithmetic expression:
4 ^ 2 + 9 * 4 ^ 3 - 10 / 2 .
The right answer will be
(a)
(b)
(c)
587
299990
999995
(a) is the right choice.

First we evaluate the exponentiations to obtain
16 + 9 * 64 5
MSc. In Software
Then we evaluate the multiplication and division to obtain 16 + 576 - 5. Last, we

evaluate the addition and subtraction to obtain the final result, 587. Observe that
the expression is traversed three times, each time corresponding to a level of
precedence of the operation.
Polish Notation (prefix notation)

For most common arithmetic operations, we place the operator symbol between its
two operands.
For example,
A+B,
C- D,
E*F ,
G/H
This is called infix notation. With this notation, we must distinguish between
(A + B) * C
and A + (B * C)
by using either parentheses or some operator-precedence convention such as the

usual precedence levels discussed above. Accordingly, the order of the operators
and operands in an arithmetic expression does not uniquely determine the order in
which the operations are to be performed.
A Polish mathematician Jan Lukasiewicz, introduced a post notation methodology
in which the operator symbol is placed before its two operands. For example,
+AB
-CD
*EF
/GH
We translate, step by step, the following infix expressions into Polish notation using
brackets [ ] to indicate a partial translation:
(A + B) * C
= [+ AB]*C
= *+ ABC
A + (B * C)
= A + [*BC]
= + A * BC
(A + B)/(C - D) = [+ AB]/[ - CD] = / + AB CD
The fundamental property of Polish notation is that the order in which the
operations are to be performed is completely determined by the positions of the
operators and operands in the expression. Accordingly, one never needs
parentheses when writing expressions in Polish notation.
Reverse Polish notation (postfix notation) refers to the analogous notation in
which the operator symbol is placed after its two operands:
AB+
CD-
EF*
GH/
Again, we do not need parentheses to determine the order of the operations in any
arithmetic expression written in reverse Polish notation.
MSc. In Software
The computer usually evaluates an arithmetic expression written in infix notation

into postfix notation, and then it evaluates the postfix expression. In each step, the
stack is the main tool that is used to accomplish the given task.
We will illustrate first how stacks are used to evaluate postfix expressions, and then
we show how stacks are used to transform infix expressions into postfix
expressions.
Evaluation of a Postfix Expression
Suppose P is an arithmetic expression written in postfix notation. The following
algorithm, which uses a STACK to hold operands, evaluates P.
Algorithm: This algorithm finds the VALUE of an arithmetic expression P written in
postfix notation.
1.
2.
3.
4.
5.
6.
Add a right parenthesis ")" at the end of P. [This acts as a

sentine1.]
Scan P from left to right and repeat Steps 3 and 4 for each element
of P until the
Sentinel ")" is encountered.
If an operand is encountered, put it on STACK.
If an operator (x) is encountered, then
(a) Remove the two top elements of STACK, where A is
the top
Element and B is the next-to-top element.
(b) Evaluate B (x) A.
(c) Place the result of (b) back on STACK
Set VALUE equal to the top element on STACK.
Exit.
We note that, when Step 5 is executed, there should be only one number on
STACK.
Consider an arithmetic expression: 5 * ( 6 + 2 ) 12 / 4.
In postfix notation it will become: 5, 6, 2, +, *, 12, 4, /, (Commas are used to separate the elements of P so that 5, 6, 2 is not interpreted
as the number 562).
Symbol Scanned
(1)
5
(2)
6
(3)
2
(4)
+
(5)
*
STACK
5
5,6
5,6,2
5,8
40
MSc. In Software
(6)
(7)
(8)
(9)
(10)
12
4
/
)
10
40,12
40,12,4
40,3
37
How the STACK works:

STACK first takes 5, and then it takes 6 and then 2 one after another. After that it
takes +. Then it executes this + on the last two numbers that it takes away from
the STACK, giving us 8. Next it takes * from the STACK. It executes multiplication
on the two numbers that it has got in its hand, giving us 40. Next it takes away 12
and 4 one after another. When it meets a division sign, it executes a division on the
last two numbers that it has taken away from the STACK, giving us 3. Remember it
still has got 40 in its hand. Now it takes - sign from the STACK and applies this on
the two numbers that it has got in its hand. ) sign is the sign for the end of the
procedure.
Transforming Infix Expression into Postfix Expressions

We can transform infix expression into postfix expression. Let Q be an arithmetic
expression written in infix notation. Besides operands and operators, Q may also
contain left and right parentheses. We assume that the operators in Q consist only
of exponentiations (^), multiplications (*), divisions (/), additions (+) and
subtractions (-), and those have the usual three levels of precedence as given
above. We also assume that operators on the same level, including
exponentiations, are performed from left to right unless otherwise indicated by
parentheses. (This is not standard, since expressions may contain unary operators
and some languages perform the exponentiations from right to left. However, these
assumptions simplify our algorithm.)
We have given the algorithm, which transforms the infix expression Q into its
equivalent postfix expression P. The algorithm uses a stack to temporarily hold
operators and left parentheses. The postfix expression P will be constructed from
left to right using the operands from Q and the operators, which are removed from
STACK. We begin by pushing a left parenthesis onto STACK and adding a right
parenthesis at the end of Q. The algorithm is completed when STACK is empty.
Algorithm:
POLISH (Q, P)
Suppose Q is an arithmetic expression written in infix notation.
This algorithm finds the equivalent postfix expression p.
1. Push "(" onto STACK, and add ")" to the end of Q.
2. Scan Q from left to right and repeat steps 3 to 6 for each
element of Q until the STACK is empty
3. if an operand is encountered, add it to P.
4. if a left parenthesis is encountered, push it onto STACK.
5. if an operator (x) is encountered, then
MSc. In Software
11
a) Repeatedly pop from STACK and add to P each operator

(on the top of STACK), which has the same precedence as
or higher precedence than (x).
b) Add (x) to STACK.
6. if a right parenthesis is encountered, then
a. Repeatedly pop from STACK and add to P each operator
(on the top of STACK) until a left parenthesis is
encountered.
b. Remove the left parenthesis. [Do not add the left
parenthesis.]
(End of Step 2 loop.]
7. Exit.
Consider this C-Program on Stack

Let us suppose that we wish to make a function that will read a line of input and
then write it backward. We can accomplish this task by putting each character onto
a stack as it is read. When the line is finished, we then pop the characters off the
stack, and they will come off in the reverse order. Consider this function first.
Void Reverse Read(void)
{
item_type item;
item_type *item_ptr = &item;
stack_type stack;
stack_type *stack_ptr = &stack;
initialize(stack_ptr);
while(!full(stack_ptr) && (item =
push(item,stack_ptr);
while( !empty(stack_ptr))
{
pop(item_ptr, stack_ptr);
}
putchat(\n);
/*initialize the stack to be empty */

getchar()) ! = \n)
/* push each item onto the stack*/
/*pop an item from the stack
This is a typical program for stack. Difficult to understand? Read on.

In the above function we are calling five other functions.
1) push()
2) pop()
3) initialize()
4) empty()
5) full()
*/
MSc. In Software
12
Our rule is one step at a time. Let us now discuss the prototypes of all these
functions. Let MAXSTACK be a symbolic constant giving the maximum size allowed
for stacks and item_type be a type describing the data that will be put into the
stack.
So, first few lines of your program will be like this.
#define MAXSTACK 10
typedef char item_type;
typedef struct struct_tag {
int top;
item_type entry[MAXSTACK];
} stack_type;
Boolean_type empty(stack_type *);
Boolean_type full(stack_type*);
Void initialize(stack_type*);
Void push(item_type, stack_type *);
Void pop(item_type *, stack_type *);
/* Push : push an item onto the stack. */
void push (Item_type item, Stack_type * stack_ptr)
{
if (stack_ptr->top >= MAXSTACK)
Error ("Stack is full");
else
stack_ptr->entry [ stack_ptr->top++] = item;
}
/* Pop: Pop an item from the stack.*/
void pop (item_type *item_ptr, Stack_type * stack_ptr)
{
if (stack_ptr->top < = 0)
Error ("Stack is empty");
else
*item_ptr = stack_ptr->entry [- - stack_ptr->top];
}
Error is a function that receives a pointer to a character string, prints the string,
and then invokes the standard function exit to terminate the execution of the
program.
/* Error : print error message and terminate the program.*/
void Error (char *s)
{
fprintf (stderr, "%s\n", s);
exit (1);
}
MSc. In Software
13
In normal circumstances Error is not executed because, as in ReverseRead, we

check whether the stack is full before we push an item and we check if the stack is
empty before we pop an item.
OTHER OPERATIONS
/* Empty : returns non-zero if the stack is empty.*/
Boolean_type Empty (Stack_type *stack_ptr)
{
return stack_ptr->top < = 0;
}
/* Full : returns non-zero if the stack is full.*/
Boolean_type Full (Stack_type *stack_ptr)
{
return stack_ptr->top > = MAXSTACK;
}
The next function initializes a stack to be empty before it is first used in a program:
/* Initialize : initialize the stack to be empty.*/
void initialize (Stack_type *stack_ptr)
{
stack_ptr->top = 0;
}
QUICKSORT, AN APPLICATION OF STACKS
Let A be a list of n data items. We can consider A as numerical or alphabetical.
"Sorting A" refers to the operation of rearranging the elements of A so that they
are in some logical order, such as numerically ordered when A contains numerical
data, or alphabetically ordered when A contains character data. In this section we
will discuss only one sorting algorithm, called quicksort, in order to illustrate an
application of stacks.
Quicksort is an algorithm of the divide-and-conquer type. That is, the problem of
sorting a set is reduced to the problem of sorting two smaller sets. We illustrate this
"reduction step" by means of a specific example.
Suppose A is a list of following numbers:
(49),
37,
13,
57,
81,
93,
43,
64,
102,
25,
91,
(69)
The reduction step of the quicksort algorithm finds the final position of one of the
numbers i.e. 49. Then we consider the last number, 69, scanning the list from right
to left, comparing each number with 49 and stopping at the first number less than
49. The number is 25. Interchange 49 and 25 to obtain the list.
MSc. In Software
(25),
37,
13,
57,
81,
93,
43,
64,
14
102, (49),
91,
69
(Observe that the numbers 91 and 69 to the right of 49 are each greater than 49.)
Beginning with 25, next scan the list in the opposite direction, from left to right,
comparing each number with 49 and stopping at the first number greater than 49.
The number is 57. Interchange 49 and 57 to obtain the list.
25, 37,
13,
(49),
81,
93,
43,
64,
102, (57),
91,
69
(Observe that the numbers 25, 37 and 13 to the left of 49 are each less than 49).
Beginning this time with 57, now scan the list in the original direction, from right to
left, until you meet the first number less than 49. It is 43. Interchange 49 and 43
to obtain the list.
25,
37,
13,
(43),
81,
93,
49,
64,
102,
57,
91,
69
(Again, the numbers to the right of 49 are each greater than 49.) Beginning with
43, scan the list from left to right. The first number greater than 49 is 81.
Interchange 49 and 81 to obtain the list.
25,
37,
13,
43,
(49),
93,
(81),
64,
102,
57,
91,
69
(Again, the numbers to the left of 49 are each less than 49.) Beginning with 81,
scan the list from right to left seeking a number less than 49. We do not meet such
a number before meeting 49. This means all numbers have been scanned and
compared with 49. Furthermore, all numbers less than 49 now form the sub-list of
numbers to the left of 49, and all numbers greater than 49 now form the sub-list of
numbers to the right of 49, as shown below:
25,
37,
13,
43,
(49),
93,
(81),
64,
102,
57,
91,
69
First sub-list
Second sub-list
Thus 49 is correctly placed in its final position, and the task of sorting the original
list A has now been reduced to the task of sorting each of the above sub-lists.
We have to repeat the above reduction step with each sub-list containing 2 or more
elements. Since we can process only one sub-list at a time, we must be able to
keep track of some sub-lists for future processing. We can accomplish this by using
two stacks, called LOWER and UPPER, to temporarily "hold" such sub-lists. That is,
the addresses of the first and last elements of each sub-list, called its boundary
values, are pushed onto the stacks LOWER and UPPER, respectively. The reduction
step is applied to a sub-list only after its boundary values are removed from the
stacks. The following example illustrates the way the stacks LOWER and UPPER are
used.
MSc. In Software
Procedure:
15
QUICK (A, N, BEG, END, LOC)

Here A is an array with N elements. Parameters BEG and END
contain the boundary values of the sublist of A to which this
procedure applies, LOC keeps track of the position of the first
element A(BEG) of the sublist during the procedure. The local
variables LEFT and RIGHT will contain the boundary values of
the list of elements that have not been scanned.
1. [Initialize] Set LEFT = BEG, RIGHT = END and LOC = BEG.
2 [Scan from right to left.]
(a)
Repeat while A [LOC] < [RIGHT] and LOC RIGHT
[End of loop.]
(b)
if LOC = RIGHT, then Return
(c)
if A[LOC]> A[RIGHT], then.
(i)
[Interchange A[LOC] and A [RIGHT]
TEMP = A [LOC], A[LOC] = A [RIGHT]
A [RIGHT] = TEMP.
(ii) Set LOC = RIGHT
(iii) GO to Step 3.
3 [Scan from left to right]
(a)
Repeat while A[LEFT] < A [LOC] and LEFT LOC.
LEFT = LEFT + 1
[End of loop.]
if LOC = LEFT, then RETURN>
(b)
(c)
if A[LEFT]> A [LOC], then

(i)
[Interchange A[LEFT] and A [LOC]
TEMP = A [LOC], A [LOC] = A[LEFT],
A [LEFT] = TEMP
(ii) Set LOC = LEFT
(iii) Go to Step 2.
Algorithm : (Quicksort) This algorithm sorts an array A with N elements.

1.
[Initialize.] TOP = NULL
2.
[Push boundary values of A onto stacks when A has 2 or more
elements.]
If N>1, then TOP = TOP + 1, LOWER [1] = 1, UPPER[1] = N
3.
Repeat Steps 4 to 7 while TOP NULL.
4.
{pop sublist from stacks.]
Set BEG = LOWER [TOP], END = UPPER [TOP]
TOP = TOP - 1.
5.
Call QUICK (A, N, BEG, END, LOC).
6.
[Push left sublist onto stacks when it has 2 or more elements.]
if BEG < LOC - 1, then
TOP = TOP + 1, LOWER [TOP] = BEG,
UPPER [TOP] = LOC 1
MSc. In Software
7.
8.
16
[Push right sublist onto stacks when it has 2 or more

elements.]
if LOC + 1 < END, then
TOP = TOP + 1, LOWER [TOP] = LOC + 1,
UPPER [TOP] = END
Exit
Let us now discuss a little C-function on this logic.

/* This function is responsible to sort a contiguous list */
List_type *quick_sort(list_type *lp)
{
sort(lp,0,lp->count-1);
return lp;
}
The sort() will look like this.
/* This function is responsible to sort the contiguous list between low and high */
/* We have used the recursion function here. In our C program we have already
told you about a recursion function. We will discuss more about it a little later. */
Void sort(list_type *lp, int low, in high)
{
int pivotloc;
if(low < high)
{
pivotloc = partition(lp,low,high);
sort(lp,low,pivotloc 1);
sort(lp,pivotloc +1, high);
}
}
/* This function is responsible to return the pivot location for a contiguous list */
int partition (list_type *lp, int low, int high)
{
int i,pivotloc;
key_type pivotkey;
swap(low, (low +high)/2,lp);
/* swap pivot into first position */
pivotkey = lp ->entry[low].key;
pivot = low;
/* remember its location
*/
for(i=low+1;i<=high;i++)
if(LT(lp->entry[i].key,pivotkey))
swap(++pivotloc,1,lp);
MSc. In Software
17
swap(low,pivotloc,lp);
return pivotloc;
}
Read this function very carefully. We are leaving it now. We will discuss more about
it in our program section.
Complexity of the Quicksort Algorithm
We can measure the running time of a sorting algorithm by the number f(n) of
comparisons required to sort n elements. Generally speaking, the algorithm has a
worst-case running time of order n2/2, but an average-case running time of order
nlog n. The reason for this is indicated below.
The worst case occurs when we sort the list, the first element requires n
comparison to get recognized if it remains in the first position. Furthermore, if the
first sublist is empty, the second sublist will have n - 1 elements. Accordingly, we
have to do n-1 comparison to recognize that it remains in the second position. And
so on. Consequently, there will be a total of
f(n) = n + ( n 1 ) + . + 2 + 1 = n ( n + 1 ) / 2
= n2/2 + O(n) = O( n2 )
comparisons. Observe that this is equal to the complexity of the bubble sort
algorithm.
The complexity f(n) = O(n log n) of the average case comes from the fact that, on
the average, each reduction step of the algorithm produces two sublists.
(1)
(2)
(3)
(4)
Reducing
Reducing
Reducing
Reducing
the
the
the
the
initial list places 1 element and produces two sublists.

two sublists places 2 elements and produces four sublists.
four sublists places 4 elements and produces eight sublists.
eight sublists places 8 elements and produces sixteen sublists.
And so on. Observe that the reduction step in the kth level finds the location of 2k-1
elements; hence there will be approximately log2 n levels of reductions steps.
Furthermore, each level uses at most n comparisons, so f(n) = O (n log n). In fact,
mathematical analysis and empirical evidence have both shown that,
f(n) = 1.4 [n log n]
is the expected number of comparisons for the quicksort algorithm.
RECURSION
Recursion is an important concept in computer science. There are many algorithms,
which we can best describe in terms of recursion. In this section we have
introduced this powerful tool.
MSc. In Software
18
Consider a procedure P containing either a Call statement to itself or a Call

statement to a second procedure that may eventually result in a Call statement
back to the original procedure P. Then P is called a recursive procedure. So that the
program will not continue to run indefinitely, a recursive procedure must have the
following two properties.
(1) There must be certain criteria, called base criteria, for which the procedure
does not call itself.
(2) Each time the procedure does call itself (directly or indirectly), it must be
closer to the base criteria.
A recursive procedure with these two properties is said to be well defined. Similarly,
a function is said to be recursively defined if the function refers to itself. Again, in
order to avoid the definition being circular, it must have the following two
properties.
(1)
(2)
There must be certain arguments, called base values, for which the function
does not refer to itself.
Each time the function does refer to itself, the argument of the function must
be closer to a base value.
A recursive function with these two properties is also said to be well defined.
The following examples should help clarify these ideas.
Factorial Function
We can explain the recursion with the example of factorial function. The product of
the positive integers from 1 to n, inclusive, is called "n factorial" and is usually
denoted by n!:
n! = 1. 2 . 3(n - 2) (n - 1) n
It is also convenient to define 0! = 1, so that the function is defined for all
nonnegative integers.
Thus we have
0! = 1
1! = 1 2! = 1.2 = 2
5! = 1.2.3.4.5 = 120
and so on. Observe that

5! = 1.2.3.4.5 = (4!).5 = 120
3! = 1.2.3 = 6
4! = 1.2.3.4.= 24
6!= 1.2.3.4.5.6. = (5!).6 = 720
This is true for every positive integer n,; that is,

n! = n. ( n - 1)!
Accordingly, the factorial function may also be defined as follows:
MSc. In Software
Definition:
19
(Factorial Function)
(a)
If n = 0, then n! = 1
(b)
If n > 0, then n! = n . (n - 1) !
Observe that this definition of n! is recursive, since it refers to itself when it uses (n
- 1) ! However, (a) the value of n! is explicitly given when n = 1 (thus) is the base
value); and (b) the value of n! for arbitrary n is defined in terms of a smaller value
of n which is closer to the base value 0. Accordingly, the definition is not circular, or
in other words, the procedure is well-defined.
The following are two procedures that calculate n factorial.
Procedure A:
FACTORIAL (FACT, N)
This procedure calculates N! and returns the value in the
variable FACT.
1. if N = 0, then Set FACT = 1, and return.
2. Set Fact = 1, [Initializes FACT for loop.]
3. Repeat for K = 1 to N.
Set FACT = K * FACT.
[End of loop.]
4. Return.
Procedure B:
FACTORIAL (FACT, N)
This procedure calculates N! and returns the value in the variable
FACT.
1. if N = 0, then, Set FACT = 1, and Return.
2. Call FACTORIAL (FACT, N - 1).
3. Set FACT = N * FACT.
4. Return.
We can observe that the first procedure evaluates N!, using an iterative loop
process. The second procedure, on the other hand, is a recursive procedure, since it
contains a call to itself.
Suppose P is a recursive procedure. During the running of an algorithm or a
program, which contains P, we associate a level number with each given execution
of procedure P as follows. The original execution of procedure P is assigned level 1;
and each time procedure P is executed because of a recursive call; its level is 1
more than the level of the execution that has made the recursive call.
The depth of recursion of a recursive procedure P with a given set of arguments
refers to the maximum level number of P during its execution.
MSc. In Software
20
Fibonacci Sequence
The celebrated Fibonacci sequence (usually denoted by F0, F1, F2, ) is as follows.
0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55,
That is, F0, = 0 and F1 = 1 and each succeeding term is the sum of the two
preceding terms. For example, the next two terms of the sequence are
34 + 55 = 89
and 55 + 89 = 144
A formal definition of this function is as follows:

Definition: (Fibonacci Sequence)
(a)
If n = 0 or n = 1, then Fn = n
(b)
If n > 1, then Fn = Fn-2 + Fn-1.
A procedure for finding the nth term Fn of the Fibonacci sequence is given below
Procedure: FIBONACCI (FIB, N)
This procedure calculates FN and returns the value in the first
parameter FIB.
1.
if N = 0 or N = 1, then: Set FIB = N, and return.
2.
Call FIBONACCI (FIBA, N -2)
3.
Call FIBONACCI (FIBB, N-1)
4.
Set FIB = FIBA + FIBB
5.
Return.
Divide-and-Conquer Algorithms
We consider a problem P associated with a set S. Suppose A is an algorithm which
partitions S into smaller sets such that the solution of the problem P for S is
reduced to the solution of P for one or more of the smaller sets. Then A is called a
divide-and-conquer algorithm.
We can use the quicksort algorithm to find the location of a single element and to
reduce the problem of sorting the entire set to the problem of sorting smaller sets.
We can use the binary search algorithm to divide the given sorted set into two
halves so that the problem of searching for an item in the entire set is reduced to
the problem of searching for the item in one of the two halves.
We can view a divide-and-conquer algorithm as a recursive procedure. The reason
for this is that the algorithm A may be viewed as calling itself when it is applied to
the smaller sets. The base criteria for these algorithms are usually the one-element
sets. For example, with a sorting algorithm, a one-element set is automatically
sorted; and with a searching algorithm, a once-element set requires only a single
comparison.
MSc. In Software
21
Ackermann Function
The Ackermann function is a function with two arguments each of which can be
assigned any nonnegative integer: 0, 1, 2, This function is defined as follows.
Definition: (Ackermann Function)
(a)
If m = 0, then A(m, n) = n + 1
(b)
If m 0 but n = 0, then a (m, n) = A(m-1, 1).
(c)
If m 0 and n 0, then A (m, n) = A (m - 1, A (m, n -1))
Once more, we have a recursive definition, since the definition refers to itself in
parts (b) and (c). Observe that A (m, n) is explicitly given only when m = 0. The
base criteria are the pairs.
(0, 0), (0, 1), (0, 2), (0, 3),.(0, n).
Although it is not obvious from the definition, the value of any A (m, n) may
eventually be expressed in terms of the value of the function on one or more of the
base pairs.
The Ackermann function is too complex to evaluate on any example. Its importance
comes from its use in mathematical logic. The function is stated here in order to
give another example of a classical recursive function and to show that the
recursion part of a definition may be complicated.
TOWERS OF HANOI
In the preceding section we have given some examples of recursive definition and
procedures. In this section we show how recursion may be used as a tool in
developing an algorithm to solve a particular problem. The problem we pick is
known as the Towers of Hanoi problem.
Suppose we have been given three pegs, labeled A, B and C, and suppose on peg A
there are placed a finite number n of disks with decreasing size. This is pictured in
figure 5-5 for the case n = 6. The object of the game is to move the disks from peg
A to peg C using peg B as an auxiliary. The rules of the game are as follows:
(a)
(b)
We can move only one disk at a time. Specifically, only the top disk on any
peg may be moved to any other peg.
At no time can a larger disk be placed on a smaller disk.
MSc. In Software
22
Fig. 5-5
Sometimes we will write X
Y to denote the instruction "Move top disk from peg X
to peg Y, " where X and Y may be any of the three pegs.
The solution to the Towers of Hanoi problem for n = 3 appears in fig. 5-6. Observe
that it consists of the following seven moves:
n=3
Move
Move
Move
Move
Move
Move
Move
top
top
top
top
top
top
top
disk
disk
disk
disk
disk
disk
disk
from
from
from
from
from
from
from
peg
peg
peg
peg
peg
peg
peg
A
A
C
A
B
B
A
to
to
to
to
to
to
to
peg
peg
peg
peg
peg
peg
peg
C.
B.
B.
C.
A.
C.
C.
MSc. In Software
23
Fig 5-6
In other words,
For completeness, we also give the solution to the Towers of Hanoi problem for n =
1 and n = 2
MSc. In Software
n = 1:
n = 2:
A
A
C
B,
C,
24
Note that n = 1 uses only one move and that n = 2 uses three moves.
Rather than finding a separate solution for each n, we use the technique of
recursion to develop a general solution. First we observe that the solution to the
Towers of Hanoi problem for n > 1 disks may be reduced to the following subproblems.
(1) Move the top n - 1 disks from peg A to peg B.
(2) Move the top disk from peg A to peg C: A
C
(3) Move the top n - 1 disks from peg B to peg C.
The reduction is illustrated in figure 5-7 for n = 6. That is, first we move the top
five disks from peg A to peg B, then we move the large disk from peg A to peg C,
and then we move the top five disks from peg B to peg C.
MSc. In Software
25
Fig. 5-7
Let us now introduce the general notation
TOWER (N, BEG, AUX, END)
To denote a procedure, which moves the top n, disks from the initial peg BEG to the
final Peg END using the peg AUX as an auxiliary. When n = 1, we have the following
obvious solution:
TOWER (1, BEG, AUX, END) consists of the single instruction
BEG
END
Furthermore, as discussed above, when n > 1, the solution may be reduced to the
solution of the following three sub-problems.
(1) TOWER (N - 1, BEG, END AUC)
(2) TOWER (1, BEG, AUX, END) or BEG ->
(3) TOWER (N - 1, AUX, BEG, END)
END
We can solve each of these three sub-problems directly or is essentially the same
as the original problem using fewer disks. Accordingly, this reduction process does
yield a recursive solution to the Towers of Hanoi problem.
Observe that the recursive solution for n = 4 disks consists of the following 15
moves:
MSc. In Software
26
In general, this recursive solution requires f (n) = 2n - 1 moves for n disks.

Let us now discuss a small C-program on this.
#define DISK 64(say)
main()
{
move(DISK,BEG,END,AUX);
}
void move(int n, int a, int b, int c)
{
if(n>0)
{
move(n-1, a,c,b);
printf(Move a disk from %d to %d.\n,a,b);
move(n-1,c,b,a);
}
}
Procedure: TOWER (N, BEG, AUX, END)
This procedure gives a recursive solution to the Towers of
Hanoi problem for N disks.
1. if N = 1, then
(a) Write, BEG -> END
(b) Return
2. [Move N - 1 disks from peg BEG to peg AUX.]
Call TOWER (N-1, BEG, END, AUX).
3. Write BEG -> END
3. [Move n - disk from peg AUX to peg END. ]
Call TOWER (N-1, AUX, BEG, END)
4. Return.
One can view this solution as a divide-and-conquer algorithm, since the solution for
n disks is reduced to a solution for n - 1 disks and a solution for n = 1 disk.
The Tower of Hanoi problem illustrates the power of recursion in the solution of
various algorithmic problems.
QUEUES
MSc. In Software
27
A queue is a linear list of elements in which deletion can take place only at one end,
called the front, and insertions can take place only at the other end, called the rear.
We use the terms "front" and "rear" to describe a linear list only when it is
implemented as a queue.
We also call Queues as first-in first-out (FIFO) lists, since the first element in a
queue will be the first element out of the queue. In other words, the order in which
elements enter a queue is the order in which they leave. This contrasts with stacks,
which are last-in first-out (LIFO) lists.
Queues are important in everyday life. People waiting in line at a bank form a
queue, where the first person in line is the first person to be waited on. The
automobiles waiting to pass through an intersection form a queue, in which the first
car in line is the first car through. An important example of a queue in computer
science occurs in a timesharing system, in which programs with the same priority
form a queue while waiting to be executed.
Example
We have shown figure 5-8(a) is a schematic diagram of a queue with 4 elements,
where AAA is the front element and DDD is the rear element. You can observe that
the front and rear elements of the queue are also, respectively, the first and last
elements of the list. Suppose we delete an element from the queue. Then it must
be AAA. This yields the queue in figure (b), where we get BBB as the front element.
Next, suppose EEE is added to the queue and then FFF is added to the queue. Then
they must be added at the rear of the queue, as pictured in figure (c). Note that
FFF is now the rear element. Now suppose we delete another element from the
queue; then it must be BBB, to yield the queue in figure (d). And so on. Observe
that in such a data structure, EEE will be deleted before FFF because it has been
placed in the queue before FFF. However, EEE will have to wait until CCC and DDD
are deleted.
Fig. 5-8
MSc. In Software
28
Representation of Queues
We can represent queues in the computer in various ways, usually by means of
one-way lists or linear arrays. Unless otherwise stated or implied, each of our
queues will be maintained by a linear array QUEUE and two pointer variables
FRONT, containing the location of the front element of the queue, and REAR,
containing the location of the rear element of the queue. The condition FRONT =
NULL will indicate that the queue is empty.
Figure (5-9) shows the way the array will be stored in memory using an array
QUEUE with N elements. This figure also indicates the way elements will be deleted
from the queue and the way new elements will be added to the queue. Observe
that whenever an element is deleted from the queue, the value of FRONT is
increased by 1; this can be implemented by the assignment
FRONT = FRONT + 1
Similarly, whenever an element is added to the queue, the value of REAR is
increased by 1; this can be implemented by the assignment
REAR = REAR + 1
This means that after N insertions, the rear element of the queue will occupy
QUEUE [N] or, in other words, eventually the queue will occupy the last part of the
array. This occurs even though the queue itself may not contain many elements.
Suppose we want to insert an element ITEM into a queue at the time the queue
does occupy the last part of the array, i.e. when REAR = N. One way to do this is to
simply move the entire queue to the beginning of the array, changing FRONT and
REAR accordingly, and then inserting ITEM as above. This procedure may be very
expensive. The procedure we adopt is to assume that the array
MSc. In Software
29
Fig. 5-9
QUEUE is circular, that is, QUEUE[1] comes after QUEUE[N] in the array. With this
assumption, we insert ITEM into the queue by assigning ITEM to QUEUE [1].
Specifically, instead of increasing REAR to N + 1, we reset REAR = 1 and then
assign.
QUEUE [REAR] = ITEM
Similarly, if FRONT = N and an element of QUEUE is deleted, we reset FRONT = 1
instead of increasing FRONT to N + 1.
Suppose that our queue contains only one element, i.e.,
FRONT = REAR NULL
And suppose that the element is deleted. Then we assign
FRONT = NULL
and REAR = NULL
to indicate that the queue is empty.
MSc. In Software
30
Comparative definition between Stack & Queue

Initialise the stack to be empty
Determine if the stack is empty or not
Determine if the stack is full or not
If the stack is not full, insert a node at
the top end of the stack
If the stack is not empty, retrieve the
top node
If it is not empty, delete the node at the
top
Initialise the queue to be empty

Determine if the queue is empty or not
Determine if the queue is full or not
If the queue is not full, insert a node at
the end of the queue
If the queue is not empty, retrieve the
first node
If it is not empty, delete the first node
Summary
#
A stack is a linear structure in which items may be added or removed only at

one end. Stacks are also called last-in first-out (LIFO) lists. Other names
used for stacks are "piles" and "push-down lists."
"Pop" is the term used to delete an element from a stack. "Push" is the
term used to insert an element into a stack.
Quicksort is an algorithm of the divide-and-conquer type. That is, the

problem of sorting a set is reduced to the problem of sorting two smaller
sets.
Consider a procedure P containing either a Call statement to itself or a Call

statement to a second procedure that may eventually result in a Call
statement back to the original procedure P. Then P is called a recursive
procedure.
We consider a problem P associated with a set S. Suppose A is an algorithm

which partitions S into smaller sets such that the solution of the problem P
for S is reduced to the solution of P for one or more of the smaller sets. Then
A is called a divide-and-conquer algorithm.
A queue is a linear list of elements in which deletion can take place only at
one end, called the front, and insertions can take place only at the other end,
called the rear.
MSc. In Software
6
SORTING AND SEARCHING
TECHNIQUES
MAIN POINTS COVERED
!
Introduction
Sorting
Insertion Sort
Selection Sort
Merging
Merge-sort
Summary
MSc. In Software
INTRODUCTION
orting and Searching are important operations in computer science. Sorting

refers to the operation of arranging data in some given order, such as
increasing or decreasing, with numerical data, or alphabetically, with
character data. Searching refers to the operation of finding the location of a given
item in a collection of items.
We have already discussed some of the sorting and searching algorithm such as,
linear and binary search but there are many more sorting and searching
algorithms. The particular algorithm one chooses depends on
!
!
the properties of the data

the operations that one may perform on the data
We will discuss the complexity of each algorithm; that is, the running time f(n) of
each algorithm and the space requirements of our algorithms.
Sorting and searching apply to a file of records, and here are some standard
terminologies of that field. Each record in a file F can contain many fields, but
there may be one particular field whose values uniquely determine the records in
the file. Such a field K is called a primary key, and the values k1, k2 .... in such a
field are called keys or key values. Sorting the file F usually refers to sorting F
with respect to a particular primary key, and searching in F refers to searching for
the record with a given key value.
SORTING
Let A be a list of n elements A1, A2,.An in memory. Sorting A refers to the
operation of rearranging the contents of A so that they are increasing in order
(numerically or lexicographically), so that
A1 A2
A3
An.
Since A has n elements, there are n! ways that the contents can appear in A. Each
sorting algorithm must take care of this n! possibilities.
Example
Suppose an array named DATA contains 8 elements as follows:
DATA:
87, 44, 21, 2, 55, 20, 76, 100
After sorting, DATA must appear in memory as follows:

DATA: 2, 20, 21, 44, 55, 76, 87, 100
MSc. In Software
Since DATA consists of 8 elements there are 8! = 40320 ways that the numbers
2, 20,.100 can appear in DATA.
Complexity of Sorting Algorithms
We measure the complexity of a sorting algorithm of the running time as a
function of the numbers n of items to be sorted. We note that each sorting
algorithm S will be made up of the following operations, where A1, A2. An contain
the items to be sorted and B is an auxiliary location (used for temporary storage):
(a)
(b)
(c)
Comparisons, which test whether Ai < Aj or test whether Ai < B

Interchanges, which switch the contents of Ai and Aj or of Ai and B
Assignments, which set B = Ai and then set Aj = B or Aj = Ai
Normally, we use the complexity function to measures the number of

comparisons, since the number of other operations is at most a constant factor of
the number of comparisons.
There are two main cases whose complexity we will consider the worst case and
the average case.
Food for thought
In average case analysis what is the usual assumption that one has to make?
(a) The elements are sorted in ascending order.
(b) The elements are sorted in descending order.
(c) The probabilistic assumption that all the n! permutations are equally likely.
(d) No assumptions are required.
(c) is the correct choice
In studying the average case, we make the probabilistic assumption that all the n!
permutations of the given n items are equally likely.
Just to give you a feel of how things work, we give you the approximate number
of comparisons and the order of complexity of some algorithms that we have
already discussed in the previous modules.
Algorithm
Bubble Sort
Quicksort
Heapsort
Worst case
n(n-1)/2 = O(n2)
n(n+3)/2 = O(n2)
3nlogn = O(n logn)
Average Case
n(n-1)/2 = O(n2)
( 1.4)nlogn = O(n logn)
3nlogn = O(n logn)
Remark - Note first that the bubble sort is a very slow way of sorting. Its main
advantage is the simplicity of the algorithm. Observe that the average-case
complexity (n log n) of heapsort is the same as that of quicksort, but its worstcase complexity (n log n) seems quicker than quicksort (n2). However, empirical
MSc. In Software
evidence seems to indicate that quicksort is superior to heapsort on rare

occasions.
Sorting Files; Sorting Pointers
Suppose a file F of records R1, R2,, Rn is stored in memory. Sorting F refers to
sorting F with respect to some field K with corresponding values k1, k2 kn. That
is, the records are ordered so that
kn
k1 k2
The field K is called the sort key. (Recall we call K as a primary key if its values
uniquely determine the records in F.) Sorting the file with respect to another key
will order the records in another way.
Example
Suppose the personnel file of a company contains the following data on each of its
employees:
Name
Employee Number
Sex
Salary
If we sort the file with respect to the Name key will yield a different order of the
records than sorting the file with respect to the Employee Number key. The
company may want to sort the file according to the Salary field even though the
field may not uniquely determine the employees. Sorting the file with respect to
the Sex key will likely be useless; it simply separates the employees into two
subfiles, one with the male employees and one with the female employees.
Sorting a file F by reordering the records in memory may be very expensive when
the records are very long. Moreover, the records may be in secondary memory,
where it is even more time-consuming to move records into different locations.
Accordingly, we may prefer to form an auxiliary array POINT containing pointers
to the records in memory and then sort the array POINT with respect to a field
KEY rather than sorting the records themselves. That is, we sort POINT so that
KEY[POINT[1]] KEY[POINT[2]] KEY[POINT[N]]
Note that choosing a different field KEY will yield a different order of the array
POINT.
INSERTION SORT
Suppose an array A with n element A[1], A[2], A[N] is in memory. The selection
sort algorithm scans A from A[1] to A[n], inserting each element A[K] into its
proper position in the previously sorted subarray A[1], A[2], A[K-1]. That is:
Pass 1.
A[1] by itself is trivially sorted.
MSc. In Software
Pass 2. A[2] is inserted either before or after A[1] so that : A[1], A[2] is sorted.
Pass 3. A[3] is inserted either into its proper place in A[1],A[2] that before A[1],
between A[1] and A[2] or after A[2] so that : A[1],A[2],A[3] is sorted
.
Pass 4. A[4] is inserted into its proper place in A[1], A[2], A[3] so that:
A[1], A[2], A[3], A[4] is sorted.
.
Pass N. A[N] is inserted into its proper place in A[1], A[2],.., A[N-1] so that
A[1], A[2],.A[N] is sorted.
Real life instances where insertion sort is used:

This sorting algorithm is frequently used when n is small.
Example
There remains only the problem for us of deciding how to insert A[k] in its proper
place in the sorted subarray A[1], A[2],A[k-1]. We can accomplish this by
comparing A[k], with A[k-1] comparing A[k] with A[k-2], comparing A[k] with
A[k-3], and so on, until first meeting an element A[j] such that A[j]A[k]. Then
each of the elements A[k-1],A[k-2,A[j+1] is moved forward one location, and
A[k] is then inserted in the j+1st position in array.
We simplify the algorithm if there always is an element A[j] such that A[j] A[k];
otherwise we must constantly check to see if we are comparing A[k] with A[1] .
This condition can be accomplished by introducing a sentinel element A[0] = -
(or a very small number).
Example
Suppose an array A contains 8 elements as follows:
7, 3, 4, 1, 8, 2, 6, 5
This figure above illustrates the insertion sort algorithm. The circled element
indicates the A[k] in each pass of the algorithm, and the arrow indicates the
proper place for inserting A[k].
The formal statement of our insertion sort algorithm is as follows:
Algorithm: (Insertion Sort) INSETION (A, N).
This algorithm sorts the array A with N elements.
(1)
Set A[0] = - [Initializes sentinel element.]
(2)
Repeat Steps 3 to 5 for K = 2, 3,., N
(3)
Set TEMP = A[K] and PTR = K-1.
MSc. In Software
(4)
Repeat while TEMP < A[PTR]

(a) Set A[PTR+1]=A[PTR]. [Moves element forward]
(b) Set PTR=PTR-1.
[End of loop.]
Set A[PTR+1]=TEMP. [Inserts element in proper place]
(5) Return.
Observe that there is an inner loop, which is essentially controlled by the variable
PTR, and there is an outer loop, which uses k as an index,
Complexity of Insertion Sort

We can easily compute the number of comparisons, f(n) in the insertion sort
algorithm. First of all, we get the worst case when the array A is a reverse order
and the inner loop must use the maximum number K-1 of comparisons. Hence
n(n-1)
f(n) = 1 + 2+.+ n(n-1) = --------- = O(n2)
2
Furthermore, we can show that, on the average, there will be approximately (k1)/2 comparisons in the inner loop. Accordingly, for the average case,
1 2
n-1
n(n-1)
f(n) = --+---++ ----- = --------- = O(n 2)
2 2
2
4
Thus the insertion sort algorithm is a very slow algorithm when n is very large.
The above results are summarized in the following table.
Algorithm
Insertion Sort
Worst Case
n(n 1)/2 = O(n2)
Average Case
n( n 1)/4 = O(n2)
Time may be saved by performing a binary search, rather than a linear search, to
find the location in which to insert A[K] in the subarray A[1],A[2],A[K-1]. This
requires, on the average, log K comparisons rather than (k-1)/2 comparisons.
However, one needs to move (k-1)/2 elements forward. Thus the order of
complexity is not changed. Furthermore, insertion sort is usually used only when
n is small, and in such a case, the linear search is about as efficient as the binary
search.
SELECTION SORT
Suppose an array A with n element A[1] A[2],A[n] is in memory. The selection
sort algorithm for sorting A works as follows. First we have to find the smallest
element in the list and put it in the first position then find the second smallest
element in the list and put it in the second position. And so on.
MSc. In Software
The stepwise procedure is as follows:

Pass 1.
Find the location LOC of the smallest in the list of n elements

A[1],A[2],..A[n] and then interchange A[LOC] and A[1]. Then
A[1] is sorted.
Pass 2.
Find the location LOC of the smallest in the sublist of n-1 element
A[2],A[3],..A[n] and then interchange A[LOC] and A[2] Then
A[1],A[2] is sorted, since A[1] A[2].
Pass 3.
Find the location LOC of the smallest in the sublist of n-2 elements
A[3],A[4],.A[n], and then interchange A[LOC] and A[3] . Then
A[1],A[2],..,A[3] is sorted since A[2] A[3].
Pass n-1
Find the location LOC of the smaller of the elements A[n-1], A[n],
and then interchange A[LOC] AND A[n-1]. Then
A[1],A[2],.A[n] is sorted , since A[n-1] A[n].
Thus A is sorted after n-1 passes.

Example
Suppose an array A contains 8 elements as follows:
7, 3, 4, 1, 8, 2, 6, 5
The problem is, finding, during the kth pass, the location LOC of the smallest
among the elements A[k], A[k+1],.A[n] . This may be accomplished by using a
variable MIN to hold the current smallest value while scanning the subarray from
A[K] to A[n]. Specifically, first set MIN = A[k] and LOC=K, and then traverse the
list, comparing MIN with other element A[j] as follows:
(a)
(b)
If MIN A[j], then simply move to the next element.

If MIN > A[j] then update MIN and LOC by setting MIN = A[j] and LOC =j.
After comparing MIN with the last element A[N],MIN will contain the smallest
among the elements A[k], A[k+1],A[n] and LOC will contain its location.
The above process will be stated separately as a procedure.
Procedure: MIN (A, k, n, LOC)
An array A is in memory. This procedure finds the location LOC of the
smallest element among A[k], A[k+1],A[n].
(1)
(2)
Set MIN = A[k] and LOC = k [Initializes pointers]

Repeat for j = k+1, k+2, n
if MIN > A[j] , then Set MIN = A[j] and LOC= A[j] and LOC =
j.
MSc. In Software
(3)
[End of loop]
Return
The selection sort algorithm can now be easily stated.

Algorithm: (Section Sort) SELECTION (A, n)
This algorithm sorts the array A with n elements.
(1) Repeat Steps 2 and 3 for k=1, 2, .., n-1
(2) Call MIN (A, k, n, LOC)
(3) [Interchange A[k] and A[LOC].]
Set TEMP = A[k],A[k]=A[LOC] and A[LOC]= TEMP.
(4) Exit
Complexity of the Selection Sort Algorithm

First we have to note that the number f(n) of comparisons in the selection sort
algorithm is independent of the original order of the elements. Observe that
MIN(A,k,n,LOC) requires n-k comparisons. That is there are n-1 comparisons
during Pass 1 to find the smallest element, there are n-2 comparisons during Pass
2 to find the second smallest element, and so on. Accordingly,
n(n-1)
f(n)=(n-1)+(n-2) +.+2+1 = ------- = O(n2)
2
The above result is summarized in the following table:
Algorithm
Selection Sort
Worst Case
n(n 1)/2 = O(n2)
Average Case
n( n 1)/4 = O(n2)
Remark: The number of interchange and assignments depends on the original

order of the elements in the array. A, but the sum of these operations does not
exceed a factor of n.
MERGING
Suppose A is a sorted list with r elements and B is a sorted list with s elements.
The operation that combines the elements of A and B into a single sorted list C
with n = r + s elements is called merging. One simple way to merge is to place
the elements of B after the elements of A and then use some sorting algorithm on
the entire list. We cannot take advantage of the fact that A and B are individually
sorted. Given below is an efficient algorithm. First, however, we indicate the
general idea of the algorithm by means of two examples.
MSc. In Software
Description of the formal algorithm

We will translate the above discussion into a formal algorithm which mergers a
sorted r-element array A and a sorted s-element array B into a sorted array C,
with n = r + s element. First, of all, we must always keep track of the locations of
the smallest element of A and the smallest element of B, which have not yet been
placed in C. Let NA and NB denote these locations, respectively. Also, let PTR
denote the location in C to be filled. Thus, initially we set NA = 1, NB = 1 and PTR
= 1. At each step of the algorithm, we compare
A[NA]
and B[NB]
and assign the smaller element to C[PTR]. Then we increment PTR by setting
PTR = PTR+1 and we either increment NA by setting NA = NA+1, or increment NB
by setting NB = NB+1, according to whether the new element in C has come from
A or from B. Furthermore, if NA > r then the remaining elements of B are
assigned to C; or if NB > s, then the remaining elements of A are assigned to C.
The formal statement of the algorithm is as follows:
Algorithm: MERGING (A, R, B, S, C)
Let A and B be sorted array with R and S element, respectively. This
algorithm merges A and B into an array C with N=R+S elements.
(1)
(2)
3.
4.
[Initialize] Set NA = 1, NB= 1 and PTR = 1.

[Compare.] Repeat while NA R and NB S;
if A[NA] < B[NB], then
(a)[Assign element from A to C] Set C[PTR] = A[NA].
(b)[Update pointer.] Set PTR =PTR+1 and
NA =NA+1.
else
(a) [Assign element from B to C] Set C[PTR] =B[NB]
(b) [Update pointer] Set PTR = PTR+1 and NB = NB+1.
[End of if structure]
[End of loop.]
[Assign remaining elements to C]
if NA > R, then
Repeat for K = 0, 1, 2, ., R-NA
Set C[PTR+K] = A[NA+K].
[End of loop.]
else
Repeat for K = 0, 1, 2, .. , R NA
Set C[PTR + K] = A[NA + K].
[End of loop]
Exit
MSc. In Software
10
Complexity of the Merging Algorithm

The input consists of the total number n = r + s of elements in A and B. Each
comparison assigns an element to the array C, which eventually has n elements.
Accordingly the number f(n) of comparisons cannot exceed n
f(n) n = O(n)
In other words, the merging algorithm can be run in linear time,
Example
Suppose A has 5 elements and B has 100 elements. Then if we merge A and B by
the above algorithm it will perform approximately 100 comparisons. On the other
hand, only approximately log 100 = 7 comparisons are needed to find the proper
place to insert an element of A and B using the binary search and insertion
algorithm.
The binary search and insertion algorithm does not take into account the fact that
A is sorted. Accordingly, the algorithm may be improved in two ways as follows:
(Here we assume that A has 5 elements and B has 100 elements)
(1)
Reducing the target set: Suppose after the first search we find that A[1]
is to be inserted after B[16]. Then we need to use only a binary search on
B[17] ,.B[100] to find the proper location to insert A[2], and so on.
(2)
Tabbing: The expected location for inserting A[1] in B is near B[20] (that
is, B[s/r]), not near b[50]. Hence we first use a linear search on B[20].
B[40], B[80] and B[100] to find B[K] such that A[1] B[K] , and then we
use a binary search on B[K-20], B[k-19],B[k]. (This is analogous to using
the tabs in a dictionary, which indicate the location of all the words with the
same first letter.)
MERGE-SORT
Suppose an array A with n elements A[1], A[2],.A[n] is in memory. The mergesort algorithm which sorts A will first be described by means of a specific
example.
Suppose the array A contains 14 elements as follows:
72, 37, 43, 25, 59, 91, 64, 13, 84, 22, 54, 47, 80, 33
Each pass of the merge-sort algorithm will start at the beginning of the array A
and merge pairs of sorted subarrays as follows:
Pass 1. Merge each pair of elements to obtain the following list of sorted pairs:
MSc. In Software
37
72
25
43
59, 91
13
64 22
84 47
11
54 33
80
Pass 2. Merge each pair of pairs to obtain the following list of sorted quadruplets
25, 37, 43, 72
13, 59, 64, 91
22, 47, 54, 84
33,80
Pass 3. Merge each pair of sorted quadruplets to obtain the following two sorted
subarrays:
13, 25, 37, 43, 59, 64, 72, 91
22, 33, 47, 54, 80, 84
Pass 4. Merge the two-sorted subarrays to obtain the single sorted array
13, 22, 25, 33, 37, 43, 47, 54, 59, 64, 72, 80, 84, 91
The original array A is now sorted.
Description
The above merge-sort algorithm for sorting an array A has the following important
property. After Pass k, we will partition the array A into sorted subarrays where
each subarray, except possibly the last, will contain exactly L = 2k elements.
Hence the algorithm requires at the most log n passes to sort an n-elements array
A.
We will translate the above informal description of merge-sort into a formal
algorithm, which will be divided into two parts. The first part will be a procedure
MERGEPASS, which uses the procedure discussed above to execute a single pass
of the algorithm, and the second part will repeatedly apply MERGEPASS until A is
sorted.
We apply the MERGEPASS procedure to an n-element array, A which consists of a
sequence of sorted subarrays. Moreover, each subarray consists of L elements
except that the last subarray may have fewer than L elements. Dividing n by 2 *
L, we obtain the quotient Q which tells the number of pairs of L-elements sorted
subarrays; that is
Q = INT(N/(2*L))
(We use INT(X) to denote the integer value of X.) Setting S = 2*L*Q, we get the
total numbers S of elements in the Q pairs of subarray. Hence R = N-S denotes
the number of the remaining element. The procedure first merges the initial Q
pairs of L-element subarray. Then the procedure takes care of the case where
there is an odd number of subarray (when R L) or where the last subarray has
fewer than L elements.
MSc. In Software
12
The formal statement of MERGEPASS and the merge-sort algorithm is as follows:
Procedure: MERGEPASS(A,N, L, B)
The N-element array A is composed of sorted subarrays where each
subarray has L elements except possibly the last subarray, which
may have fewer than L elements. The procedure merges the pairs of
subarrays of A and assigns them to the array B.
1. Set Q = INT(N/(2*L)), S= 2*L*Q and R= N S .
2. [Use Procedure discussed above to merge the Q pairs of
subarrays.]
Repeat for J = 1, 2, ., Q
(a) Set LB = 1+(2*J-2)*L [find lower bound of first array]
(b) Call MERGE (A, L, LB, A, L, LB+L, B, LB)
[End of loop.]
3 [Only one subarray left?]
if R L then
Repeat for J = 1,2R
Set B( S + J ) = A( S + J )
[End of loop]
else
Call MERGE (A, L, S+1, A, R, L+S+1, B, S+1)
[End of if structure]
4 Return.
Algorithm:
MERGESORT(A,N)
This algorithm sorts the N-element array A using an auxiliary array
B.
1. Set L = 1. [Initialize the number of elements in the subarray]
2 Repeat Steps 3 to 6 while L < N
3.
Call MERGEPASS(A, N, L, B).
4.
Call MERGEPASS(B, N, 2*L, A).
5.
Set L = 4 * L.
6. Exit
Since we want the sorted array to finally appear in the original array A, we must
execute the procedure MERGEPASS an even number of times.
Complexity of the Merge-Sort Algorithm
We use f(n) to denote the number of comparisons needed to sort an n-element
array A using the merge-sort algorithm. Recall that the algorithm requires at
MSc. In Software
13
most log n passes. Moreover each pass merges a total of n elements, and by the
discussion on the complexity of merging. Each pass will require at most n
comparisons. Accordingly for both the worst case and average case,
f(n) n log n
Observe that this algorithm has the same order as heapsort and the same
average order as quicksort. The main drawback of merge-sort is that it requires
an auxiliary array with n element. Each of the other sorting algorithms we have
studied is that it requires only a finite number of extra locations, which is
independent of n.
The above results are summarized in the following table:
Algorithm
Merge-Sort
Worst Case
n log n = O(n log n)
Average Case
n log n = O(n log n)
Extra
O(n)
Summary
# Sorting and Searching are important operations in computer science.
Sorting refers to the operation of arranging data in some given order, such
as increasing or decreasing, with numerical data, or alphabetically, with
character data. Searching refers to the operation of finding the location of a
given item in a collection of items.
# We measure the complexity of a sorting algorithm of the running time as a
function of the numbers n of items to be sorted.
# We learned about Insertion sort, selection sort, Merge Sort.
MSc. In Software
7
HASHING TECHNIQUES
MAIN POINTS COVERED
Hashing
Hash Functions
Collision Resolution
Open Addressing
(Linear probing and modification)
Chaining
Summary
MSc. In Software
HASHING
he search time for each algorithm we discussed so far depends on the

number of elements of data. Hashing is a searching technique, which is
essentially independent of the number of elements.
The terminology, which we will be using in our presentation of hashing, will

be oriented towards file management. First of all we assume that there is a
file F of n records with a set k of keys that uniquely determine the records in
F. Secondly we assume that F is maintained in memory by a table T of m
memory locations and that L is the set of memory addresses of the locations
in T. For notational convenience, we assume that the keys in k and the
addresses in L are (decimal) integers.
We will introduce the subject of hashing by the following example.
Example
Suppose a company with n employees assigns an employee number to each
employee. We can in fact, use the employee number as the address of the
record in memory. The search will require no comparisons at all but a lot of
space will be wasted.
The idea of using the key to determine the address of a record is an excellent
idea, but it must be modified so that a great deal of space is not wasted. This
modification takes the form of a function H from the set K of keys into the
set L of memory addresses. Such a function,
H: K # L
is called a hash function or hashing function. Unfortunately, such a
function H may not yield distinct values. It is possible that two different keys
k1 and k2 will yield the same hash address. This situation is called collision.
And some method must be used to resolve it.
Accordingly the topic of hashing is divided into two parts:
(1) Hash function and
(2) Collision resolutions
We will discuss each of these two parts with you separately.
Hash Functions
The two principal criteria in selecting a hash function H: K # L are as follows:
First of all, the function H should be very easy and quick to compute. Second,
the function H should, as far as possible uniformly distribute the hash
addresses throughout set L so that there are minimum number of collisions.
MSc. In Software
We cannot guarantee that the second condition can be completely fulfilled

without actually knowing before hand the keys and addresses. However,
certain general techniques do help. One technique is to chop a key k into
pieces and combine them in some way to form the hash address H(k).
We will now illustrate some popular hash functions. We emphasize that the
computer can easily and quickly evaluate each of these hash functions.
(a)
Division method: Choose a number m larger than the number n of
keys in K (We usually chose the number m to be a prime number or a
number without small divisors. Since this frequently minimizes the number of
collisions.) The hash function H is defined by
or
H (k) = k (mod m)
H(k) = k(mod m) + 1
Here k(mod m) denotes the remainder when k is divided by m. The second

formula is used when we want the hash addresses to range from 1 to m
rather than from 0 to m 1.
(b)
Midsquare method: Here, the key k is squared and the hash function
H is defined by
H(k) = 1
where 1 is obtained by deleting digits from both ends of k2. We emphasize
that the positions of k2 must be the same for all the keys.
(c)
Folding method: We have to partition the key k into a number of
parts k1, , kr where each part, except possibly the last, has the same number
of digits as the required address. Then the parts are added together, ignoring
the last carry, that is.
H(k) = k1 + k2 + .+ kr
where the leading-digit carries, if any, are ignored. Sometimes for extra
milling the even-numbered parts, k2, k4, .are each reversed before the
addition.
Example
We can consider the company in the previous example each of whose 68
employees is assigned a unique 4-digit employee number. Suppose L
consists of 100 two-digit addresses: 00, 01, 02,,99. We apply the above
hash functions to each of the following employee numbers.
3205,
7148,
2345
MSc. In Software
(a)
Division method. Choose a prime number m close to 99, such as m =
97. Then
H(3205) = 4,
H(7148) = 67, H(2345)=17
That is, dividing 3205 by 97 gives a remainder of 4, dividing 7148 by 97
gives reminder of 67, and dividing 2345 by 97 gives a remainder of 17. In
the case that the memory addresses begin with 01 rather than 00, we
choose that the function H(k) = k(mod m) + 1 to obtain:
H(3205) = 4+1=5,
(b)
H(7148) = 67 + 1 = 68,
H(2345) = 17+1=18
Midsquare method. The following calculations are performed:
k :
3205
k2 :
10 272 025
H(k):
72
7148
2345
51 093 904 5 499 025
93
99
Observe that the fourth and fifth digits, counting from the right, are chosen
for the hash address.
(c)
Folding method. Chopping the key k into two parts and adding yields
the following hash addresses:
H(3205)= 32+05 = 37
H(7148)=71+48=19 H(2345)=23+45=68
Observe that the leading digit 1 in H(7148) is ignored . Alternatively, we may

want to reverse the second part before adding, thus producing the following
hash addresses:
H(3205)=32+50=82, H(7148)=71+84=55
H(2345)=23+54=77
Collision Resolution
Suppose we want to add a new record R with key k to our file F, and the
memory location address H(k) is already occupied. This situation is called
collision. In this subsection we discuss two general ways of resolving
collisions. The particular procedure that we have chosen depends on many
factors. One important factor is the ratio of the number n of keys in K (which
is the number of records in F) to the number m of hash addresses in L. This
ratio, = n/m, is called the load factor.
First we will show that collisions are almost impossible to avoid. Specifically,
suppose a student class has 24 students and suppose the table has space for
365 records. One random hash function is to choose the students birthday
as the hash address. Although the load factor = 24/365 7% is very
small. We can show that there is a better than fifty-fifty chance that two of
the students have the same birthday.
MSc. In Software
We can measure the efficiency of a hash function with a collision resolution

procedure by the average number of probes (key comparisons) needed to
find the location of the record with a given key k.
The efficiency depends mainly on the load factor . Specifically, we are
interested in the following two efficiencies that depend mainly on the load
factor . Specifically, we are interested in the following two quantities.
S(
) = average number of probes for a successful search
U(
) = average number of probes for an unsuccessful search
Open Addressing: Linear Probing and Modifications

Suppose we add a new record R with key k to the memory table T, but that
the memory location with hash address H(k) = h is already filled. One natural
way to resolve the collision is to assign R to the first available location
following T[h]. ( We assume that the table T with m locations is circular, so
that T[1] comes after T(m) ). Accordingly, with such a collision procedure,
we will search for the record R in the table T by linearly searching the
locations T[h], T[h+1], T[h+2], until finding R or meeting an empty location,
which indicates an unsuccessful search.
The above collision resolution is called linear probing. The average number
of probes for a successful search and for an unsuccessful search are known
to be the following respective quantities.
S(
) =
1
1
---- (1 + ----)
2
1-
and
U(
) =
1
1
---- (1+ ------2
(1-
)2
(Here = n/m is the load factor.)
Chaining
Chaining involves maintaining two tables in memory. First of all, as we used
before, there is a table T in memory, which contains the records in F, except
that T now has an additional field LINK that is used so that all records in T
with the same hash address h may be linked together to form a linked list.
Second, there is a hash address table LIST that contains pointers to the
linked lists in T.
Suppose a new record R with key k is added to the file F. We place R in the
first available location in the table T and then add R to the linked list with
pointer LIST[H(k)]. If the linked lists of records are not sorted, then R is
simply inserted at the beginning of its linked list. Searching for a record or
deleting a record is nothing more than searching for a node or deleting a
node from a linked list.
MSc. In Software
The average number of probes, using chaining, for a successful search and
for an unsuccessful search are known to be the following approximate
values:
1
S(
) 1 + --
and U(
) e- +
2
Here the load factor = n/m may be greater than 1, since the number m of
hash addresses in L (not the number of locations in T) may be less than the
number n of records in F.
Example
Lets consider again the data in the previous example where the 8 records,
have the following hash addresses:
Record:
H(k):
A
4
B
8
C
2
D
11
E
4
X
11
Y
5
Z
1
Using chaining, the records will appear in memory as pictured in figure 7-1
Observe that the location of a record R in Table T is related to its hash
address. A record is simply put in the first node in the AVAIL list of table T.
In fact, table T need not have the same number of elements as the hash
address table.
Fig. 7-1
The main disadvantage of chaining is that we need 3m memory cells for the
data. Specifically there are m cells for the information field INFO, there are m
MSc. In Software
cells for the link field LINK, and there are m cells for the pointer array LIST.
Suppose each record requires only 1 word for its information field. Then it
may be more useful to use open addressing with a table with 3m locations,
which has the load factor 1/3, then to use chaining to resolve collisions.
Summary
$ The terminology, which we will be using in our presentation of
hashing, will be oriented towards file management.
$ We have used the two principal criteria in selecting a hash function H:
K # L are as follows. First of all, the function H should be very easy

and quick to compute. Second the function H should, as far as possible
uniformly distribute the hash addresses throughout the set L so that
there are minimum number of collisions.
$ Suppose we want to add a new record R with key k to our file F, but
suppose the memory location address H(k) is already occupied. This
situation is called collision.
$ Chaining involves maintaining two tables in memory. First of all, as we
used before, there is a table T in memory, which contains the records
in F, except that T now has an additional field LINK that is used so that
all records in T with the same hash address may be linked together to
form a linked list.

Data Structures

Uploaded by

Data Structures

Uploaded by

Introduction to Data Structures

Complexity, Rate of Growth,

The Need to learn Data structures

Coding, Testing And Refinement

Basic Terminology; Elementary Data

Data Structures Operations

Complexity, Time-Space Tradeof

THE NEED TO LEARN DATA STRUCTURES

ome of the rules for good programming are:

CODING, TESTING AND REFINEMENT

4. Keep your functions short and simple.

BASIC TERMINOLOGY; ELEMENTARY DATA

Logical or mathematical description of the structure

A(1), A (2), A(3),, A(N)

Or by the bracket notation

A[1], A[2], A[3],, A[N]

Lets see an example so that we can understand it easily:

This is the simplest example of a single-dimensional array. Here STUDENT[1]

mathematics, and tables in business applications. Multidimensional arrays are

Link List contd..

Some More data Structures

automobiles waiting to pass through an intersection--the first car in line is the

which is not necessarily hierarchical in nature. For example, suppose an airline

DATA STRUCTURE OPERATIONS

ALGORITHMS: COMPLEXITY, TIME-SPACE TRADEOF

Consider time and space trade-offs in deciding on your algorithm.

An algorithm is a well-defined list of steps for solving a particular problem. One

Sequence logic, or sequential flow

RATE OF GROWTH; BIG O NOTATION

b) Remainder function: Modular Arithmetic

(read k modulo M) will denote the integer remainder when k is divided by M.

When k is positive, simply divide k by M to obtain the remainder r.

=25, divider = 7, remainder is 4]

M is called the modulus, and a b (mod M) is read a is congruent to b

Arithmetic modulo M refers to the arithmetic operations of addition,

c) Integer and Absolute Value Functions

Observe that INT(x) = x

INT (-8.5) = -8,

d) Summation Symbol: Sums

will be denoted respectively, by

The expression a1b1 + a2b2 + + anbn is denoted as

Logarithms are related to exponents as follows.

are equivalent statements. Accordingly

Furthermore for any base b,

Logarithms to the base e

Notation: log x will mean log

x unless otherwise specified.

[Initialize. ] Set K = 1, LOC = 1 and MAX = DATA [1].

[Increment counter.] Set K = K + 1.

Sequence logic, or sequential flow

Flow chart equivalent

Selection Logic (Conditional Flow)

where D = b2 4ac is called the discriminant of the equation.

Iteration Logic (Repetitive Logic)

[ Initialize. ] Set K = 1 , LOC = 1 , and MAX = DATA[1].

Time taken by the algorithm

Time: We can measure time by counting the number of key operations

is the correct choice, as only one element has to be compared

Average case analysis of algorithms:

We can find the complexity of the search algorithm by the number C of

Rate of Growth: Big O Notation or Big Oh Notation

REPRESENTATION OF LINEAR ARRAY IN

TRAVERSING LINEAR ARRAYS

SORTING; BUBBLE SORT

SEARCHING; LINEAR SEARCH

REPRESENTATION OF MULTIDIMENSIONAL ARRAYS

and (c) are the correct choices.

A1, A2, A3, .., An

A (1), A (2), .., A (N)

(a) , (b) and (c) are the correct choices

REPRESENTATION OF LINEAR ARRAYS IN MEMORY

We use the notation

TRAVERSING LINEAR ARRAYS

Algorithm: (Traversing a Linear Array) Here LA is a linear array with lower

Now we present the same algorithm using a different control structure.

Sorting may seem to be an easy task. Actually, sorting efficiently may be