Machine learning programming

e-ISSN: 2582-5208
International Research Journal of Modernization in Engineering Technology and Science

( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:05/Issue:01/January-2023 Impact Factor- 6.752 www.irjmets.com
A STUDY ON GENETIC ALGORITHM AND ITS APPLICATIONS

Namira Begum Sayyed*1
*1Department Of IT (Information Technology), S.I.W.S College, Mumbai, India.
DOI : https://www.doi.org/10.56726/IRJMETS32980
ABSTRACT
We need a way to differentiate the best and worse solutions in order to get the best outcomes. A statistical
model or simulation could be the metric, or it could be a subjective one where we favor better results over
worse ones. In addition, the fitness function chooses a top solution for a particular issue, which is then used by
the GA to direct the evolution of other top solutions. The use of genetic algorithms in many fields is
demonstrated in this study along with how they may be integrated with a number of other approaches to arrive
at the best possible solution and speed up information retrieval systems.
I. INTRODUCTION
John Holland proposed genetic algorithms [1], which are search and optimization techniques based on the ideas
of natural evolution, in 1970. By replicating the evolution of species through natural selection, genetic
algorithms also carry out the optimization procedures. Typically, a genetic algorithm consists of two steps. The
first step is choosing an individual to produce the next generation, and the second is modifying the chosen
individual to create the next generation using crossover and mutation processes [2]. Which individuals are
chosen for reproduction and how many children each selected person produces are determined by the
selection mechanism. The main principle of selection strategy is that a person's likelihood of having children
increases with their level of competence.
II. GENETIC ALGORITHM
Figure 1. Flowchart of GA System

In the 1970s, John Holland invented Genetic Algorithms (GA), which are search algorithms based on the
principles of natural selection and genetics and were inspired by the biological evolution of living organisms. By
producing generations iteratively, genetic algorithms abstract the problem space as a population of individuals
and attempt to identify the fittest individual. A population of beginning individuals is transformed by GA into a
population of high-quality individuals, where each individual offers a potential solution to the issue at hand. A
fitness function, which represents each rule's adaptation to a certain environment quantitatively, evaluates
each rule's quality. Starting with a population that was initially created randomly.
www.irjmets.com International Research Journal of Modernization in Engineering, Technology and Science
[696]
e-ISSN: 2582-5208
Three fundamental genetic processes selection, crossover, and mutation are successively applied to each
person with varying probabilities over each generation [3].
The GAs is a computer software that simulates how genes and living things evolve [3]. Due to the fact that GAs
are multi-point search techniques, an optimal solution is still feasible for objective functions that use many
modes. Additionally, discrete search space issues can be solved with GAs. Thus, GA is a very effective
optimization tool that is also quite user-friendly [4]. In GA, the search space is made up of strings known as
chromosomes, each of which represents a potential solution to the issue. Each chromosome's fitness value is its
objective function value. A population is a collection of chromosomes with the corresponding fitness. In one
iteration of the GA, generations are populations created [5]. The following describes a genetic algorithm for
searching a set of potential solutions to find the best one[6]:
Procedure:
1. [Start] Create a population of n chromosomes at random (suitable solutions for the problem).
2. [Fitness] Analyze each chromosome x in the population to determine its fitness, f(x).
3. [New Population] Repeat the steps below until the new population is finished to create a new population
a. [Selection] Choose two parent chromosomes from a population according to how fit they are (the more fit,
the greater the possibility of selection).
b. [Crossover] Crossover the parents to create a new offspring with a crossover probability (children). If no
crossover occurred, the offspring is a perfect replica of the parents.
c. [Mutation] Create new children with a mutation probability at each locus (position in chromosome).
d. [Accepting] Place new offspring (children) in a new population.
4. [Replace] Use the newly formed population to run the algorithm again.
5. [Test] Stop and return the most effective solution available to the present population if the final condition is
met.
6. [Loop] Go to Step 2.
III. GENETIC OPERATORS
By using genetic operations [6] including selection, crossover, and mutation, genetic algorithms (GA) searches
for optimum solutions.
1. Selection Operation
The goal of the selection process is to choose elitist people from the current population who can produce
offspring as parents. Fitness measures are used as a standard of measurement to determine whether someone
is elitist. The best chromosomes can be chosen using a variety of techniques, including the Boltzman selection
method, tournament selection, rank selection, steady-state selection, elitism selection, and others. Some of
them will be described below.
a. Roulette Wheel Selection
Parents are chosen based on their fitness. The greater the chromosomes the chances of selection are more.
Imagine a roulette wheel with all of the chromosomes in the population placed on it. Each chromosome is
assigned a size according to its fitness function, and the chromosome with the highest fitness will be chosen
more frequently.
b. Rank Selection
When the fitness levels are substantially different, the preceding selection procedure will have issues. The other
chromosomes will have very little probability of being chosen, for instance, if the best chromosomal fitness
makes up 90% of the total roulette wheel. With rank selection, each chromosome is given fitness after the
population is first sorted according to fitness. The worst will have fitness 1, the second-worst will have fitness
2, etc., while the best will have fitness N (number of chromosomes in population). The remaining chromosomes
then have a chance to be chosen. Therefore, rather than fitness, the likelihood that a chromosome will be
chosen is inversely correlated with its rank in this sorted list. However, because the best chromosomes do not
differ greatly from other chromosomes, this strategy may result in delayed convergence.

[697]
e-ISSN: 2582-5208
c. Elitism Selection
There is a strong probability that the best chromosome will be lost when a new population is created through
crossover and mutation. The process known as elitism involves first copying the best chromosome (or a small
number of best chromosomes) to a new population. The remaining is completed using traditional methods.
Elitism can vastly improve GA performance since it keeps the best solution from being lost.
2. Crossover Operations
A group of operators that recombine and mutate a subset of the current population produce the generation of
successors in a GA. Mutation and crossover are the two most popular operators. By replicating a few selected
bits from each parent string, the crossover operator creates two new children from two parent strings. Each
offspring's bit at position I copies the bit at that same location in one of the two parents. An additional string
known as the crossover mask determines which parent contributes the bit for position i.
3. Mutation Operations
A second kind of operator, in addition to recombination operators that create children by fusing elements of
two parents, creates offspring from a single parent. Specifically, the mutation operator modifies the value of a
single bit at random to cause minor random changes in the bit string. Frequently, mutation appears after
crossover.
IV. OUTLINE OF GENETIC ALGORITHM
The genetic algorithm's operation [7] is outlined given below:
1. An initial population created at random is the first step of the method.
2. After that, the algorithm generates a series of fresh populations. The algorithm builds the subsequent
population at each stage using members of the current generation. The algorithm follows these steps to
produce the new population:
a. Computes the fitness value of each member of the current population to determine their score. The raw
fitness scores refer to these numbers.
b. Translates the raw fitness ratings into a more useful set of numbers by scaling them. Expectation values are
the name given to these scaled values.
c. Chooses members, referred to as parents, according to their expectations.
d. The elite are chosen from among some of the less fit members of the current population. The subsequent
population inherits these outstanding individuals.
e. generates offspring from the parents. Crossover is the process of mixing the vector entries of two parents,
while mutation is the process of producing random changes to a single parent.
f. The children are used to create the next generation, replacing the current population.
3. When one of the halting requirements is satisfied, the algorithm terminates.
4. To account for linear and integer limitations, the technique uses modified steps.
5. The algorithm is then changed to accommodate nonlinear restrictions.
V. APPLICATION OF GENETIC ALGORITHM
The optimization of issues and solutions[8] is one of the fundamental uses for genetic algorithms, which have
many other uses as well. We employ optimization to locate the best answer to every issue. There are various
advantages to undertaking optimization using genetic algorithms, which can be referred to as genetic
optimization.
One of the most prevalent real-world combinatorial optimization issues that can be resolved by genetic
optimization is this one. The primary goal of this problem is to determine the best path for the salesman to take
within the confines of a given map that shows routes and distances between places. We don't find the answer
once when using genetic algorithms to determine the ideal route construction. We can produce child solutions
that can carry over the characteristics of parent solutions after each cycle. Applications for TSP include
manufacturing, logistics, and planning.

[698]
e-ISSN: 2582-5208
Algorithm [9]:
1. Create a random population starting point.
2. Try to determine the chromosome's fitness.
3. Repeat until finished:
a. Choose parents.
b. Crossover and mutation are performed.
c. Determine the new population's fitness.
d. Add it to the gene pool.
A salesman is given a list of towns; his task is to determine the shortest path between each city and the
beginning city while visiting each place precisely once. Cities are employed as genes, and a string created using
these characters is referred to as a chromosome. A population is targeted using a fitness score that is equal to
the combined path length of all the cities listed.
VI. CONCLUSION
Genetic Algorithm as we know is employed in repeatedly modifying a population of individual solutions, as this
is a method for solving such constrained and unconstrained problems to achieve optimization. The whole idea
as we know it is based on natural selection that drives biological evolution. This paper is presented keeping in
mind genetic optimization that again employs the concepts of genetic algorithms. Hence it is supported by
genetic optimization applications applied on a popular Travelling Salesman Problem. genetic algorithm
concepts applied in travelling salesman problem A salesman is given a list of towns; his task is to determine the
shortest path between each city and the beginning city while visiting each place precisely once.
VII. REFERENCES
[1] J. Holland, “Adaptation in natural and artificial systems”, University of Michigan press, Ann Arbor,
1975.
[2] Noraini Mohd Razali, John Geraghty “A genetic algorithm performance with different selection
strategies”, Proceedings of the World Congress on Engineering Vol II, 2011.
[3] D. E. Goldberg, “Genetic Algorithm in Search, Optimization and Machine Learning, Reading, MA:
Addison-Wesley, 1989.
[4] Gaidhani, Chaitali R., Vedashree M. Deshpande, and Vrushali N. Bora. "Image Steganography for
Message Hiding Using Genetic Algorithm." (2014).
[5] Radwan A., Latef B., Ali A., and Sadek O., “Using Genetic Algorithm to Improve Information Retrieval
Systems”, World Academy of Science and Engineering Technology, 17, Issue-2, Page No (6-13), 2006.
[6] L. Haldurai, T. Madhubala and R. Rajalakshmi “A Study on Genetic Algorithm and its Applications”,
Kongunadu Arts and Science College, Coimbatore, India, Volume-4, Issue- 10, E-ISSN:2347-2693
[7] https://in.mathworks.com/help/gads/how-the-genetic-algorithm-works.html
[8] https://analyticsindiamag.com/10-real-life-applications-of-genetic-optimization/
[9] https://www.geeksforgeeks.org/traveling-salesman-problem-using-genetic-algorithm/

[699]

Machine learning programming

Uploaded by

Machine learning programming

Uploaded by

e-ISSN: 2582-5208

International Research Journal of Modernization in Engineering Technology and Science

A STUDY ON GENETIC ALGORITHM AND ITS APPLICATIONS

Figure 1. Flowchart of GA System

www.irjmets.com International Research Journal of Modernization in Engineering, Technology and Science

www.irjmets.com International Research Journal of Modernization in Engineering, Technology and Science

www.irjmets.com International Research Journal of Modernization in Engineering, Technology and Science

You might also like