Accelerat ing t he world's research.

A Study on Genetic Algorithm and its

Applications
IJCSE Editor

Related papers Download a PDF Pack of t he best relat ed papers 

Effect of Genet ic Algorit hm on Art ificial Neural Net work for Int rusion Det ect ion Syst em
IJCSE Edit or

A review: accuracy opt imizat ion in clust ering ensembles using genet ic algorit hms
Norwat i Must apha

Det ect ion Of Cyber At t ack Using Art ificial Int elligence Based Genet ic Algorit hm Wit h Feedback Ingest i…
Mr. Jay Parag Meht a, Digvijaysinh M Rat hod
 Review Paper Volume-4, Issue-10 E-ISSN: 2347-2693

A Study on Genetic Algorithm and its Applications

L. Haldurai1*, T. Madhubala2 and R. Rajalakshmi3
123*
Department of Computer Science (PG), Kongunadu Arts and Science College, Coimbatore, India

Available online at: www.ijcseonline.org

Received:19/Sep/2016 Revised: 28/Sep/2016 Accepted:21/Oct/2016 Published: 31/Oct/2016
Abstract— In order to obtain best solutions, we need a measure for differentiating best solutions from worst solutions. The measure
could be an objective one that is a statistical model or a simulation, or it can be a subjective one where we choose better solutions
over worst ones. Apart from this the fitness function determines a best solution for a given problem, which is subsequently used by
the GA to guide the evolution of best solutions. This paper shows how GA is combined with various other methods and technique to
derive optimal solution, increase the computation time of retrieval system the applications of genetic algorithms in various fields.
Keywords- Genetic Algorithm; Optimal Solution; Fitness function

I. INTRODUCTION
Genetic algorithms [1] are search and optimization
algorithms based on the principles of natural evolution, which
were first introduced by john Holland in 1970. Genetic
algorithms also implement the optimization strategies by
simulating evolution of species through natural selections.
Genetic algorithm is generally composed of two processes.
First process is selection of individual for the production of
next generation and second process is manipulation of the
selected individual to form the next generation by crossover
and mutation techniques [2]. The selection mechanism
determines which individual are chosen for reproduction and
how many offspring each selected individual produce. The
main principle of selection strategy is the better is an
individual; the higher is its chance of being parent.
Figure 1. Flowchart of GA System
II. GENETIC ALGORITHM
The GAs is computer program that simulate the heredity
Genetic algorithms (GA) are search algorithms and evolution of living organisms [3]. An optimum solution is
based on the principles of natural selection and genetics, possible even for multi modal objective functions utilizing
introduced by J Holland in the 1970’s and inspired by the GAs because they are multi-point search methods. Also, GAs
biological evolution of living beings. Genetic algorithms is applicable to discrete search space problems. Thus, GA is
abstract the problem space as a population of individuals, and not only very easy to use but also a very powerful
try to explore the fittest individual by producing generations optimization tool [4]. In GA, the search space consists of
iteratively. GA evolves a population of initial individuals to a strings, each of which representing a candidate solution to the
population of high quality individuals, where each individual problem and are termed as chromosomes. The objective
represents a solution of the problem to be solved. The quality function value of each chromosome is called its fitness value.
of each rule is measured by a fitness function as the Population is a set of chromosomes along with their
quantitative representation of each rule’s adaptation to a associated fitness. Generations are populations generated in
certain environment. The procedure starts from an initial an iteration of the GA [5]. Genetic algorithm to search a space
population of randomly generated individuals. of candidate solutions to identify the best one is as follows:
During each generation, three basic genetic operators are
sequentially applied to each individual with certain Procedure:
probabilities, i.e. selection, crossover and mutation [3]. {
1. [Start] Generate random population of n chromosomes
*
Corresponding Author: L. Haldurai (suitable solutions for the problem).
e-mail: [email protected]

© 2016, IJCSE All Rights Reserved 139

 International Journal of Computer Sciences and Engineering Vol.-4(10), Oct 2016, E-ISSN: 2347-2693

2. [Fitness] Evaluate the fitness f(x) of each chromosome x in 2) Rank Selection

the population. The previous selection method will have problems
3. [New population] Create a new population by repeating when the fitness’s differ very much. For example, if the best
following steps until the new population is complete chromosome fitness is 90% of the entire roulette wheel, then
a. [Selection] Select two parent chromosomes from a the other chromosomes will have very few chances to be
population according to their fitness (the better fitness, the selected. Rank selection first sorts the population by fitness
bigger chance to be selected). and then every chromosome receives fitness from this ranking.
b. [Crossover] With a crossover probability cross over the The worst will have fitness 1, second worst 2 etc. and the best
parents to form a new offspring (children). If no crossover will have fitness N (number of chromosomes in population).
was performed, offspring is an exact copy of parents.
After this, all the chromosomes have a chance to be selected.
c. [Mutation] With a mutation probability mutate new
The probability that a chromosome will be selected is then
offspring at each locus (position in chromosome).
proportional to its rank in this sorted list, rather than its fitness.
d. [Accepting] Place new offspring in a new population.
4. [Replace] Use new generated population for a further run But this method can lead to slower convergence, because the
of algorithm. best chromosomes do not differ so much from other ones.
5. [Test] If the end condition is satisfied, stop, and return the
best solution in current population. 3) Elitism Selection
6. [Loop] Go to step 2. When creating new population by crossover and
} mutation; we have a big chance, that we will lose the best
chromosome. Elitism is name of method, which first copies
III. GENETIC OPERATORS the best chromosome (or a few best chromosomes) to new
population. The rest is done in classical way. Elitism can very
GA searches for better solutions by genetic operations, rapidly increase performance of GA, because it prevents
including selection operation, crossover operation and losing the best found solution.
mutation operation.
B. Crossover Operations
A. Selection Operation
The generation of successors in a GA is determined by a
Selection operation is to select elitist individuals as parents in set of operators that recombine and mutate selected members
current population, which can generate offspring. Fitness of the current population. The two most common operators
values are used as criteria to judge whether individuals are are crossover and mutation. The crossover operator produces
elitist. There are many methods how to select the best two new offspring from two parent strings, by copying
chromosomes, for example roulette wheel selection, Boltzman selected bits from each parent. The bit at position i in each
selection, tournament selection, rank selection, steady-state offspring is copied from the bit at position i in one of the two
selection, elitism selection and some others. Some of them parents. The choice of which parent contributes the bit for
will be described shortly. position i is determined by an additional string called the
crossover mask. Figure 3.10 below illustrates crossover
1) Roulette Wheel Selection operator briefly. There are three types of crossover operators,
Parents are selected according to their fitness. The better the namely as single-point, two-point and uniform crossover.
chromosomes are, the more chances to be selected they have.
Imagine a roulette wheel where are placed all chromosomes 1) Single-Point Crossover
in the population, every has its place big accordingly to its In the single-point crossover, the crossover mask is
fitness function like on the Figure 2. Chromosome with always constructed so that it begins with a string containing n
bigger fitness will be selected more times. contiguous 1s, followed by the necessary number of 0s to
complete the string. This results in offspring in which the first
n bits are contributed by one parent and the remaining bits by
the second parent. Each time the single-point crossover
operator is applied, the crossover point n is chosen at random,
and the crossover mask is then created and applied. To
illustrate, consider the single-point crossover operator at the
top of the figure and consider the topmost of the two offspring
in this case. This offspring takes its first five bits from the
first parent and its remaining six bits from the second parent,
because the crossover mask 11111000000 specifies these
choices for each of the bit positions. The second offspring
uses the same crossover mask, but switches the roles of the
Figure 2. Roulette Wheel Selection

© 2016, IJCSE All Rights Reserved 140

 International Journal of Computer Sciences and Engineering Vol.-4(10), Oct 2016, E-ISSN: 2347-2693

two parents. Therefore, it contains the bits that were not used has a component that scores the classification accuracy of the
by the first offspring. rule over a set of provided training examples. Often other
criteria may be included as well, such as the complexity or
2) Two-Point Crossover generality of the rule. More generally, when the bit-string
In two-point crossover, offspring are created by hypothesis is interpreted as a complex procedure (e.g., when
substituting intermediate segments of one parent into the the bit string represents a collection of if-then rules that will
middle of the second parent string. Put another way, the be chained together to control a robotic device), the fitness
crossover mask is a string beginning with n0 zeros, followed function may measure the overall performance of the
by a contiguous string of nl ones, followed by the necessary resulting procedure rather than performance of individual
number of zeros to complete the string. Each time the two- rules.
point crossover operator is applied, a mask is generated by
randomly choosing the integers n0 and nl. For instance, in the V. REVIEW OF LITERATURE
example shown in Figure 3.10 the offspring are created using
a mask for which n0= 2 and nl = 5. Again, the two offspring A novel hybrid genetic k-means algorithm (GKA)
[6] to find a globally optimal partition of a given data into a
are created by switching the roles played by the two parents.
specified number of clusters. The proposed GA circumvent
expensive crossover operator used to generate valid child
3) Uniform Crossover
chromosomes from parent chromosomes. It hybridized the
Uniform crossover combines bits sampled uniformly
GA by using a classical gradient descent algorithm used in
from the two parents, as illustrated in Figure 3. In this case the clustering viz., K-means algorithm. In genetic K-means
crossover mask is generated as a random bit string with each algorithm (GKA), K-means operator was defined and used as
bit chosen at random and independent of the others. a search operator instead of crossover. It defined a biased
mutation operator specific to clustering called distance-based-
mutation. The authors used finite Markov chain theory to
prove that the proposed GKA converges to the global
optimum. It was also observed that GKA searches faster than
some of the other evolutionary algorithms used for clustering.
An improved version of GKA known as Fast
Genetic K-means Algorithm (FGKA) was proposed in [7].
The proposed GA featured several improvements over GKA.
It was evident from experiments in [6][7] that K-means
algorithm might converge to a local optimum, both FGKA
and GKA always converge to the global optimum. FGKA
initializes the population to P0 and obtains the next
population by applying selection, crossover and mutation
operators and it keeps on evolving until some termination
condition is met. Illegal strings are permitted in FGKA during
initialization phase, but were considered as the most
Figure 3. Crossover and Mutation Operations
undesirable solutions by defining their total within cluster
variation (TWCVs) as infinity (+∞). By allowing illegal
C. Mutation Operations strings the overhead of illegal string in the evolution process
In addition to recombination operators that produce was avoided and thus improved the time performance of the
offspring by combining parts of two parents, a second type of algorithm as compared to GKA.
operator produces offspring from a single parent. In
particular, the mutation operator produces small random Incremental Genetic K-means Algorithm (IGKA)
changes to the bit string by choosing a single bit at random, proposed in [8] was an extension to previously proposed
then changing its value. Mutation is often performed after clustering algorithm, the Fast Genetic K- means Algorithm
crossover as in Figure 3. (FGKA). The performance of IGKA was found to be better
when the mutation probability was small. IGKA was based
calculating the Total Within-Cluster Variation (TWCV) and
IV. FITNESS FUNCTION to cluster centroids incrementally whenever the mutation
The fitness function defines the criterion for ranking probability was small for the clustering task. Like FGKA,
potential hypotheses and for probabilistically selecting them IGKA also always converges to the global optimum.
for inclusion in the next generation population. If the task is to
learn classification rules, then the fitness function typically

© 2016, IJCSE All Rights Reserved 141

 International Journal of Computer Sciences and Engineering Vol.-4(10), Oct 2016, E-ISSN: 2347-2693

A GA-based unsupervised clustering technique was generated from the created model are fed to the GA which is
proposed in [9], which selects cluster centers directly from the used to find the corresponding inputs so that automation of
data set, thus speeding up the fitness evaluation process by test cases generation from output domain is completed.
constructing a look-up table in advance and saving the
In [16] introduces a new algorithm based on the
distances between all pairs of data points. Binary
traditional genetic algorithm, for the traditional GA algorithm
representation rather than string representation is used to
the new algorithm has done some improvements: By
encode a variable number of cluster centers and more
introducing genetic selection strategy, decreased the
effective operators for selection, crossover, and mutation were
possibility of being trapped into a local optimum. Compared
introduced.
the traditional genetic algorithm, the new algorithm enlarges
A novel clustering algorithm for mixed data was the searching space and the complexity is not high. By
proposed in [10].Most of the existing clustering algorithms analyzing the testing results of benchmarks functions
were only efficient for the numeric data rather than the mixed optimization, it is concluded that in the optimization precision,
data set but the proposed GA worked efficiently for datasets the new algorithm is efficiency than the traditional genetic
with mixed values by modifying the common cost function. algorithm. We also use this new algorithm for data
classification and the experiment results shown that our
A hybrid genetic based clustering algorithm, called
proposed algorithm outperforms the KNN with greater
HGA-clustering was proposed in [11] to explore the proper
accuracy.
clustering of data sets. This algorithm, with the cooperation of
tabulist and aspiration criteria, has achieved harmony between Managing groundwater supplies has found AI and
population diversity and convergence speed. GAs useful. [17] Proposed used GAs to fit parameters of a
model to optimize pumping locations and schedules for
A genetic algorithm was proposed in [12] which
groundwater treatment. They then combined the GA with a
designed a dissimilarity measure, termed as Genetic Distance
neural network (NN) to model the complex response
Measure (GDM) to improve the performance of the K-modes
functions within the GA [18]. Then [19] combined Simulated
algorithm which is an extension of k- means.
Annealing (SA) and GAs to maximize efficiency and well use
A novel approach of [13] parallel indexing the color the easily applied parallel nature of the GA. Most recently,
and feature extraction of images and genetic algorithm has [20] together with Peralta used a Pareto GA to sort optimal
been implemented. Its main functionality is image-to-image solutions for managing surface and groundwater supplies,
matching and its intended use is for still-image retrieval. The together with a fuzzy-penalty function while using an
evaluation criteria are provided by the GA and have been Artificial Neural Network (ANN) to model the complex
successfully employed as a measure to evaluate the efficacy aquifer systems in the groundwater system responses.
of content-based image retrieval process.
Evolutionary methods have also found their way into
In [14], P.R. Srivastava and Tai have presented a oceanographic experimental design. In [21] showed that a
method for optimizing software testing efficiency by genetic algorithm is faster than simulated annealing and more
identifying the most critical path clusters in a program. The accurate than a problem specific method for optimizing the
SUT is converted into a CFG. Weights are assigned to the design of an oceanographic experiment. [22] found that an
edges of the CFG by applying 80-20 rule. 80 percentage of evolutionary programming strategy was more robust than
weight of incoming credit is given to loops and branches and traditional methods for locating an array of sensors in the
the remaining 20 percentage of incoming credit is given to the ocean after they have drifted from their initial deployment
edges in sequential path. The summation of weights along the location.
edges comprising a path determines criticality of path. Higher
VI. CONCLUSION
the summation more critical is path and therefore must be
tested before other paths. In this way by identifying most Genetic Algorithms proved to be better in finding areas of
critical paths that must be tested first, testing efficiency is complex and real world problems. Genetic Algorithms are
increased. adaptive to their environments, as this type of method is a
In [15], Zhao used the neural network and GA for platform appearing in the changing environment. In Present
the functional testing. Neural network is used to create a these algorithms are more applicable. Several improvements
model that can be taken as a function substitute for the SUT. must be made in order that GAs could be more generally
The emphasis is given on the outputs which exhibit the applicable.
important features of SUT than inputs. In that case, test cases REFERENCES
should be generated from the output domain rather than input
domain. The feed forward neural network and back [1] J. Holland, “Adaptation in natural and artificial systems”,
propagation training algorithm is used for creating a model. University of Michigan press, Ann Arbor, 1975.
Neural network is trained by simulating the SUT. The outputs

© 2016, IJCSE All Rights Reserved 142

 International Journal of Computer Sciences and Engineering Vol.-4(10), Oct 2016, E-ISSN: 2347-2693

[2] Noraini Mohd Razali, John Geraghty “A genetic algorithm Computer Sciences and Engineering, Volume-04, Issue-10,
performance with different selection strategies”, Proceedings Page No (10-18), Oct -2016
of the World Congress on Engineering Vol II, 2011. [19] Shieh, H-J. and R.C. Peralta,: Optimal system design of in-situ
[3] D. E. Goldberg, “Genetic Algorithm in Search, Optimization bioremediation using genetic annealing algorithm. In Ground
and Machine Learning, Reading, MA: Addison-Wesley, 1989. Water: An Endangered Resource, Proceedings of Theme C,
[4] Gaidhani, Chaitali R., Vedashree M. Deshpande, and Vrushali Water for a changing global community, 27th Annual
N. Bora. "Image Steganography for Message Hiding Using Congress of the International Association of Hydrologic
Genetic Algorithm." (2014). Research, pp 95-100, 1997.
[5] Radwan A., Latef B., Ali A., and Sadek O., “Using Genetic [20] Fayad, H, Application of neural networks and genetic
Algorithm to Improve Information Retrieval Systems”, World algorithms for solving conjunctive water use problems, Ph.D.
Academy of Science and Engineering Technology, 17, Issue-2, Dissertation, Utah State University, 152 pp, 2001.
Page No (6-13), 2006. [21] Barth, H, Oceanographic Experiment Design II: Genetic
[6] K. Krishna and M. N. Murty, “Genetic K-Means Algorithm”, Algorithms, Journal of Oceanic and Atmospheric Technology,
IEEE Transaction on Systems, Man, and Cybernetics—Part B: 9, 1992, pp. 434-443, 1992.
CYBERNETICS, Vol. 29, No. 3, June 1999. [22] Porto, V.W., D.B. Fogel, and L.J. Fogel, Alternative neural
[7] Shrivastava, Animesh, and Singh Rajawat. "An network training methods, IEEE Expert Syst, June 1995.
implementation of hybrid genetic algorithm for clustering
based data for web recommendation system." Int J Comput Sci Authors Profile
Eng 2.4 (2014): 6-11.
[8] Yi Lu1, Shiyong Lu1, Farshad Fotouhi1, Youping Deng, d.
Susan, J. Brown,” an Incremental genetic K- means algorithm Mr. L. Haldurai done his Bachelor of Science,
and its application in gene expression data analysis”, BMC Master of Computer Applications and Master of
Bioinformatics 2004. Philosophy from Bharathiar University. He
currently working as Assistant Professor in
[9] Sangari, Fardin Esmaeeli, and Mehrdad Nabahat. "Efficacy of Department of Computer Science (PG),
Different Strategies in Graph Coloring with Parallel Genetic Kongunadu Arts and Science College.His main
Algorithms." (2014): 138-141. research work focuses on Data Mining and Data
[10] LI Jie, G. Xinbo, “A GA-Based Clustering Algorithm Communications. He has 8 years of Teaching experience and 2
forLarge Data Sets With Mixed Numeric and Categorical years of Research experience.
Values”,IEEE, Proceedings of the Fifth International
Conference onComputational Intelligence and Multimedia
Applications(ICCIMA’03) 0-7695-1957-1/03, 2003. Ms. T. Madhubala done Bachelor of Science and
[11] Y. Liu, Kefe and X. Liz,” A Hybrid Genetic Based Clustering Master of Science from Bharathiar University.
Algorithm”, Proceedings of the Third InternationalConference She currently working as Assistant Professor in
on Machine Leaming and Cybernetics, Shanghai,26-29 August Department of Computer Science, Kongunadu
2004. Arts and Science College. Her main research
[12] S. Chiang, S. C. Chu, Y. C. Hsin and M. H. Wang, ”Genetic work focuses on Data Mining. She has 3 years of
Distance Measure for K-Modes Algorithm”,International Teaching experience.
Journal of Innovative Computing, Informationand Control
ICIC,ISSN 1349- 4198, Volume 2, Number 1, pp.33-40
February 2006. Ms. R. Rajalakshmi done Master of Computer
[13] L. Haldurai and V. Vinodhini, "Parallel Indexing on Color and Applications and Master of Philosophy from
Texture Feature Extraction using R-Tree for Content Based Bharathiar University. She currently works as
Image Retrieval", International Journal of Computer Sciences Assistant Professor in Department of Computer
and Engineering, Volume-03, Issue-11, Page No (11-15), Nov Science (PG), Kongunadu Arts and Science
-2015, E-ISSN: 2347-2693 College. Her main research work focuses on Data
[14] Praveen Ranjan Srivastava and Tai-hoon Kim, “Application of Mining. She has 8 years of Teaching experience and 2 years of
genetic algorithm in software testing” International Journal of Research experience.
software Engineering and its Applications, 3(4), pp.87- 96,
2009.
[15] Ruilian zhao, shanshan lv, “Neural network based test cases
generation using genetic algorithm” 13th IEEE international
symposium on Pacific Rim dependable computing. IEEE,
pp.97 – 100, 2007.
[16] Sreedharamurthy S K and H.R.Sudarshana Reddy, "Feature
Subset Selection Using Genetic Algorithms for Handwritten
Kannada Alphabets Recognition", International Journal of
Computer Sciences and Engineering, Volume-03, Issue-06,
Page No (94-99), Jun -2015.
[17] Aly, A.H. and R.C. Peralta,: Comparison of a genetic
algorithm and mathematical programming to the design of
groundwater cleanup systems, Water Resources Research,
35(8), pp. 2415-2425, 1999.
[18] Amin Dastanpour, Suhaimi Ibrahim and RezaMashinchi,
"Effect of Genetic Algorithm on Artificial Neural Network for
Intrusion Detection System", International Journal of

© 2016, IJCSE All Rights Reserved 143