You are on page 1of 6

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/314175061

A Brief Study and Analysis of Different Searching Algorithms

Conference Paper · February 2017

CITATION READS

1 2,135

4 authors, including:

Najma Sultana Sourabh Chandra


Calcutta Institute Of Technology ,Kolkata,India Calcutta Institute of Technology, Kolkata
4 PUBLICATIONS   6 CITATIONS    39 PUBLICATIONS   107 CITATIONS   

SEE PROFILE SEE PROFILE

Sk Safikul Alam
Jubilant Life Sciences
37 PUBLICATIONS   107 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

[Elsevier] Intelligent Environmental Data Analysis and Pollution Management View project

SENTIMENT ANALYSIS FOR PRODUCT REVIEW View project

All content following this page was uploaded by Sourabh Chandra on 02 March 2017.

The user has requested enhancement of the downloaded file.


2017 SECOND IEEE INTERNATIONAL CONFERENCE ON E LECTRICAL , COMPUTER AND COMMUNICATION T ECHNOLOGIES

A Brief Study and Analysis of Different


Searching Algorithms
Najma Sultana Smita Paira
Dept. of Computer Science and Engineering Dept. of Computer Science and Technology
Calcutta Institute of Technology, Uluberia IIEST, Shibpur
Howrah, India Howrah, India
najma.cse@gmail.com smtpaira@gmail.com

Sourabh Chandra Sk Safikul Alam


Dept. of Computer Science and Engineering Dept. of Computer Science and Engineering
Calcutta Institute of Technology, Uluberia Calcutta Institute of Technology, Uluberia
Howrah, India Howrah, India
sourabh.chandra@gmail.com mail2safikul@gmail.com

Abstract— This paper presents the review of certain types of search algorithms used to search an element from
important and well discussed traditional as well as a set of elements. Search algorithm includes both
proposed search algorithms with respect to their time traditional search algorithm i.e. which are used very
complexity, space Complexity , merits and demerits frequently and proposed search algorithms i.e. the search
with the help of their realize applications. This paper algorithms which are proposed by the researcher's
also highlights their working principles. As it is well recently .Traditional search algorithms are like linear
known that every sorted or unsorted list of elements search[2], binary search[1] , interpolation search[16] and
requires searching technique so many search jump search[12] . There are some proposed search
algorithms have been invented. Among traditional algorithms too, as network localization using tree search
search algorithms, a comparison table has been made algorithm, a fast string searching algorithm etc[14].
in order to check and establish their benefits and Though there are many search algorithms, we have
drawbacks. Among some of the well known proposed discussed and compared some of the useful, efficient
search algorithms like fast string searching algorithm, algorithms in the following sections. Time complexity
multiple solution vector approach and bi linear search analysis has been made in order to compare them
etc, a comparison has also been made. perfectly.

Keywords—Network localization; tree search; BPCF; II. THEORITICAL REVIEW ON


harmony search; SOS; Best first search; LR-LD TRADITIONAL SEARCH ALGORITHMS
I. INTRODUCTION In this section , we have made an comparison study on the
four traditional search algorithms namely linear search ,
Searching is one of the basic operations of Data Structure, binary search ,interpolation and jump search .The
which is used to find a particular data from a list of data. comparison study on these algorithms are based on the
Searching can operates both on sorted and unsorted list of features, time complexity analysis, merit and demerit
elements. Depending on the type of data structure, a respectively. The Table I below briefs about the study.
suitable search algorithm is chosen .There is so many

Table I. Comparison Table of Traditional Search Algorithms


Algorithm Linear Search Binary Search Interpolation Search Jump Search
Features Sequentially checks the Also known as half Improved variant of Also known as Block
target element in the listInterval search or Binary search. Works search, Where step
until and unless it is logarithmic search on the probing number is calculated
found or all the elements Which follows divide position of required from the list length
are checked [3]. and Reduce method. value to search a and searching is done
particular data from a in some interval of
list. Works only on blocks.
sorted elements.
Time Complexity Best case=O(1) Average Best case =O(1) Best case =O(1)
Analysis case=O(n) Worst Average Case =O(log Average Case

944 978-1-5090-3239-6/17/$31.00 © 2017 IEEE


2017 SECOND IEEE INTERNATIONAL CONFERENCE ON E LECTRICAL , COMPUTER AND COMMUNICATION T ECHNOLOGIES

Algorithm Linear Search Binary Search Interpolation Search Jump Search


case=O(n) n) Worst Case =O(log =O(log(log n)) Worst
n) case=O(n)
Merit Easy to understand, Average case and Execution time for It is very useful When
implement and Resource worst case order are Average case is Much jumping back is
efficient. Works on both better than that of lower than linear and significantly lower
sorted and Unsorted linear search. Worst binary search. than jumping
elements. case order is also forward. More
better than interpolate efficient than linear
on search. search.
Demerit Is not very efficient than Works only on sorted Works only on Sorted Works only on Sorted
binary and interpolation n elements. elements and worst elements.
search as it requires lot of case is n iterations Implementation of
Comparison s to find a This approach is
particular data. Considered to be
More difficult than
Binary search.

Linear search is a sequential search which checks each execution value for best case but for worst case it takes
and every element of the entire list to search a particular much more time to execute the program.
data from the list. It is an easy, resource efficient
algorithm which can work on both sorted and unsorted list Bi Linear search a new session of searching was proposed
of elements. Binary search or logarithmic search is a more by researchers Smita Paira et al. in the year 2014.The
efficient search algorithm to search a particular data from algorithm works on both sorted and unsorted list of
a sorted list of data. Determines the middle position of the elements. As the name suggest that is Bi Linear, searching
list and search accordingly either in left or right of the propagates at both the ends of the list for a search item in
middle. Worst case and average case orders are much the list thus its worst case scenario is far better than other
appreciable than linear search. search algorithms. The researchers Phisan Kaewprapha et
al. proposed a search algorithm named Network
Interpolation search or improved variant binary search localization using tree search in the year 2016, the
works by calculating the probing position in the list. algorithm is a heuristic process which works under the
Average case execution time is very much noticeable O presence of anchor nodes (nodes, whose locations are well
(log n).Thus is a helpful algorithm when the search value known) to find the locations of desired unknown nodes
is of average one. Jump search algorithm refers to a search [6]. The addition of a new node in the existing list is done
in sorted list of elements. It is also known as block search by traversing and checking it’s compatibility for
as it uses a minimum gapping value or length to search a neighboring nodes and non neighboring nodes. Apart
particular data. It overcomes the demerits of sequential from solving a problem faster this method do not promises
search for worst case scenario. to be optimal.

III. REVIEW ON SOME RECENTLY Researchers Sankar Peddapati et al. proposed an


PROPOSED SEARCH ALGORITHMS algorithm on multiple solutions vector approach in the
year 2016.The term multiple solution is an area multiple
In this section some of the well discussed proposed search criteria decision making. Vector approach of multiple
algorithms have been described in short along with the variable problems can be solved comparing a large
table of comparison. These are as follows: number of generated solutions with the position of best
solution vector. If the random number so generated is less
A fast string searching algorithm was proposed by than the best position consideration factor (BPCF) the
researchers Robert S Boyer et al. in the year 1977, which new solution are evaluated from the best individual
the foresights the string being searched for. The basic solution. Along with the merit of solving a multiple
knowledge behind the working of this algorithm is that variable problem easily it has a demerit of having high
more information is gathered by matching the pattern value of execution time.
from right than from the left of the string .If the last
character of the searched character name is matched then An iterative way of searching named unsupervised feature
it will examine the second last of the character name and selection technique based on harmony search was
so on until we get a positive result or terminating the proposed in the year 2016 by researchers Laith
process getting a negative output thus work as a Muhammad Abualigah et al. It improves the text
determinative process [4]. This method has a low clustering technique by means of random search
technique. It works on its corresponding parameters

945
2017 SECOND IEEE INTERNATIONAL CONFERENCE ON E LECTRICAL , COMPUTER AND COMMUNICATION T ECHNOLOGIES

following the storage of solutions in the harmony memory algorithm i.e. an application of best first search algorithm
matrix. Invention of a new solution takes place following to demand control in the year 2016 to minimize the
certain rules, if the new solution is far better than the deficiency of electricity in Zombia. The algorithm
previous one it will replace that otherwise not. When searches for the best combination of household appliances
harmony search reaches its maximum value of iterations according to their power rate and ranking. System
the process will terminate and the HMCR and PAR continuously checks for the power of PV system and
parameters will help in searching global and local matches whether the said combination of appliances input
solutions. This algorithm beats other text clustering demand matches with the supply or not. Algorithm runs
techniques through improved performances of fast through each and every appliances and alerts the user
information retrieval. The foremost drawback of this which appliance should be switched on or off as per their
algorithm is that it performs well only in a limited area. priorities. The priority (represented by a number) is given
by the user. The system is prepared in such a way that it
Bid based economic dispatch using symbiotic organisms tells when the consumer is connected to the electricity
search algorithm was proposed by researchers Archana grid or not. The main advantage of Best first search
Tiwari et al. In the year 2016, which says that the algorithm over Breadth first search algorithm is that as it
economic load dispatch problem that is the maximization uses heuristic functions it is applicable for large number
criteria of social profit while satisfying different of appliances and produces result much faster[20]. But
constraints like power balance, generator bid and sometimes the result may be disvalued if the heuristic
customer bid quantities can be done using symbiosis function so produced contains some errors.
technique [15] .The word symbiosis refers to the
relationship between two different species .SOS a A nearest neighbor search algorithm for LR-LD on high
metaheuristic process, works on the principle of survival SNR was proposed by researchers Thae Thae Yu Khine et
of the fittest. The fittest species will eradicate the other al. in the year 2016, this algorithm includes two parts that
and occupy its position. The algorithm works by is derivation of unimodular matrix and feasible detection
calculating mutualism vector, commensalism function or of symbols. But mainly focuses on the second drawback
parasitism vector and benefit factor(BF).SOS attempts to of lattice reduction (LR) technique (feasible set of the
move a population, called ecosystem of possible solutions detected symbol cannot be found without huge
to promising areas of the search space during the search operation)[18].It produces feasible set of detected symbols
for optimal solution for which social profit respecting the without lesser amount of procedures and calculations .The
generation cost and customer demand will be higher. Has algorithm work same as of previous LR algorithm to
no algorithm specific parameters thus easy to understand derive the unimodular(matrix having perpendicular
and compute [19].May produce result with little bit error vectors) matrix[11].Its operation is easy and execution
(as it is a meta-heuristic process). time is much lower. The table for various proposed search
algorithms is described in Table II below:-
Researchers Kashale Chimmanga et al. proposed a search

Table II. Comparison of Proposed Search Algorithms


Method Feature Merit Demerit
Fast string Searches for the pattern in the text Execution time for best case If the pattern is not present
searching from backward direction i.e. first is much lower than other in the text i.e. worst case
algorithm searches for the right positioned similar type of string execution time is much
character of the pattern and searching algorithms. higher than other string
continues to search for the initial searching algorithms.
characters. [13] Unable to search for a
numeric value in the text.
It does not work efficiently
for small alphabets like
DNA.
Bi-linear search Modified form of linear search. Works on both sorted and -
algorithm Search bidirectionally i.e. unsorted list of elements [5].
searching starts from both the ends. Worst case run time is of
Along with best case order worst O(1) which is far better than
case order is also of 1. the worst case run time of
linear, binary, interpolation
and jump search algorithms.

946
2017 SECOND IEEE INTERNATIONAL CONFERENCE ON E LECTRICAL , COMPUTER AND COMMUNICATION T ECHNOLOGIES

Method Feature Merit Demerit


Network Heuristic process of determination. Produces faster result Output is not much
Localization n Works in presence of anchor nodes, .Efficient in terms of memory optimal as it is a heuristic
using Tree search uses compatibility factor to localize usage [23]. process of searching.
algorithm certain unknown nodes. Promises
to be faster enough
Multiple solution Generates N numbers of random Solves confined optimization Execution time is higher as
vector Approach solutions. Identifies best solution problem easily where it requires many
and replaces it by the previous one. multiple variables are present. calculations.
[7]
Unsupervised Uses random search technique Beats other text clustering Has limited scope, only
feature Selection [21]. Invention of new solutions Techniques through Improved being accessible within a
technique based on takes place following certain rules. performance of fast certain portion.
Harmony search The process terminates when it information retrieval [8].
algorithm reaches its maximum iteration.
Bid based Meta-heuristic in nature. The It is simple, easy to Do not produce accurate or
economic dispatch method follow some formulations understand. Has less optimal result every time.
Using symbiotic like objective functions and computational time than
organisms Search constraints (power balance, Other algorithms like
algorithm generator bid quantities and simulated annealing and
customer bid quantities constraints) genetic algorithm [9]. It has
[22]. no algorithm specific
parameters.
Based on the symbiotic relationship
a particular algorithm
(commensalism, mutualism or
parasitism) is chosen
Application of best Maximizes use of energy from For a number of appliances it Greedy in nature, i.e. may
first search solar source than from the is found to be more efficient lead to a solution very
algorithm to economic grid. Follows some in Searching optimal quickly with some errors if
demand control heuristic information to select the appliances Combinations than heuristic functions are not
better appliances combinations in breadth first search algorithm. optimal.
order to minimize power demand
[10]. It also helps the user to select
own appliance combinations and
do switch on/off according to the
power generate.
A nearest neighbor Derivation of unimodular matrix is Operation is simpler and
Search algorithm same as in case of other LR Time complexity is minimal
for LRLD on high methods. Evaluation of optimal set than the previous LR-LD
SNR of detected symbol from unknown algorithm.
transformed symbols is done
quickly.

IV. CONCLUSION demerits in order to have optimal search algorithms for


accurate and fast search operations in near future.
Searching is one of the important operation of data
structure. Different searching algorithms enable us to look REFERENCES
for a particular data from the entire list. Traditional search
algorithms are well known for their basic characteristics [1]. Data structure using C by Reema Thareja(first published
like searching a data from array index i.e. either from 2011 by Oxford University Press) in INDIA Ram Book
Binding House,New Delhi.
sorted list(like binary search, interpolation search and [2]. Data structures using c and c++ by Yedidyah
jump search)or from unsorted list of elements like Linear langsam,Aaron M. Tenenbaum second Indian printing
search algorithm [17]. Different proposed search (Prentice Hall of India private limited) New Delhi-
algorithms have been also discussed whose working are 110001.
unique and condition specific i.e. they have been meant [3]. Data Structures by Seymour Lipschutz and G A
Vijayalakshmi Pai (Tata McGraw Hill companies).
for a particular job. The merits of each and every
[4]. R S Boyer and J Srother Moore, “ A fast string searching
algorithm are highly appreciable but we cannot ignore algorithm”,Communication of the association for
their drawbacks. Researchers should focus on those

947
2017 SECOND IEEE INTERNATIONAL CONFERENCE ON E LECTRICAL , COMPUTER AND COMMUNICATION T ECHNOLOGIES

computing machinery Inc.20(10),1977, vol-20,pp. 762- demand control”,2016 IEEE PES Power Africa
772. Conference,pp. 51-55,IEEE 2016.
[5]. Smita Paira,Sourabh Chandra,Sk Safikul Alam and [11]. Thae Thae Yu Khine,Koji Araki, Daisuke Mitsunaga and
Subhendu Sekhar Patra, “Bi Linear search a new session Hua-An Zhao” A nearest neighbor search algorithm for
of searching”, IJARCSSE, Vol. 4, Issue 3, pp. 459-463, LR-LD on high SNR” IEEE 2016.
March 2014. [12]. en.wikipedia.org/wiki/Jump search
[6]. Phisan Kaewprapha,Thaewa TansarnandNattakan [13]. www.cs.utexas.edu/users/moore/publications/fstrpos.pdf
Puttarak,”Network localization using tree search
algorithm”IEEE 2016. [14]. ieeexplore.ieee.org/Xplore/home.jsp
[7]. Sankar Peddapati and K.K. Phanisri Kruthiventi,”A New [15]. ieeexplore.ieee.org/document/7571840/
Random Search Algorithm: Multiple Solution Vector [16]. en.wikipedia.org/wiki/Interpolation search_algorithm
Approach”, 2016 6th International Advanced Computing [17]. www.studytonight.com/data-structures/search-algorithms
Conference,pp.187-190, IEEE 2016. [18]. www.cs.auckland.ac.nz/~jmor159/PLDS210/niemann/s_
[8]. Laith Muhammad Abualigah, Ahamad Tajudin man.pdf
Khaderand Mohammed Azmi AI-Betar ,“Unsupervised [19]. www.researchgate.net/publication/262227023_Symbiotic
feature selection technique based on harmony search”, _Organisms_S
2016 7th International Conference on Computer Science earch_A_new_metaheuristic_optimization_algorithm
and Information Technology (CSIT),IEEE 2016.
[20]. en.wikipedia.org/wiki/Best-first_search
[9]. Archana Tiwari and Manjaree Pandit ,”Bid based
economic dispatch using symbiotic organisms search [21]. en.wikipedia.org/wiki/Harmony_search
algorithm”, 2nd IEEE International Conference on [22]. www.iaeng.org/publication/WCECS2010/WCECS2010_
Engineering and Technology (ICETECH), 17th& pp460-465.pdf
18thMarch 2016, Coimbatore, TN, India,IEEE 2016. [23]. en.wikipedia.org/wiki/tree search
[10]. Kashale Chimmanga, Josephat Kalezhi and Phillimon
Mumba,“Application of best first search algorithm to

948

View publication stats

You might also like