A_Proposed_Genetic_Algorithm_Model_to_Im.pdf

A Proposed Genetic Algorithm Model to Improve
Effort Estimation in Agile Software Development
Hossam Abdel Rahman Mohamed Al-Ansary
Computer and Information Technology Dept, Cairo University
Email: Hossam_mm7@yahoo.com
Abstract— One of the main difficult tasks in software
development projects is an effort estimation accuracy. the
project's estimated cost, duration, and maintenance effort
early in the development life cycle is a big challenge to be
achieved for software projects. Agile software projects
require an innovative effort estimation model to help a
constructive cost accurate estimation. The main focus of
this paper is using the prediction model of genetic
algorithm to improve the effort estimation accuracy by
chromosomes that are made up of a gene pool that
includes user stories, friction factors, implementation
Level factor, and dynamic forces.
Index Terms—Genetic Algorithm, Agile Software
Development, Effort estimation.
1. Introduction
The management of the software development
project is the process of planning, organizing, staffing,
monitoring, controlling, and software project leading.
the life cycle of the project consists of four phases
Project initiation, project planning, project execution,
and project closure. the Project planning is a crucial
phase in the life cycle of the software project because it
includes many challenging activities that are necessary;
such as software estimation process that includes
estimating the size of the software product to be
produced, estimating the effort required, developing
preliminary project schedules, and finally, estimating
the overall cost of the project. [1]
Agile Software Development (ASD) is a group of
software development processes that are iterative,
incremental, self-organizing, and emergent that
continuously refined and prioritized following
principles of Just In Time (JIT) management. [2]
The main issue of determining effort estimation in
the agile methodology is to focus on the effort and
degree of difficulties of teamwork instead of an
individual. Agile Software Development uses different
techniques to deliver software products with maximum
customer satisfaction. Some of the techniques include
XP (Extreme Programming) as shown in fig 1, Scrum
Methodology, Adaptive Software Development (ASD).
Some of the existing work shows that during the use of
Agile Software Development, the team uses a story
point approach to calculate the effort with the help of
user story and project velocity as inputs. The main task
is to estimate the required effort keeping in mind the
technical complexity of the project. [3]
Fig 1: Agile Extreme Programing (XP) Lifecycle
The calculation and understanding of estimation
models based on user stories are difficult due to
inherent complex relationships between the related
friction factors. an implementation Level factor and
relationships used to estimate software development
effort could change over time and may differ for
different software development environments. to
address and overcome these problems, a new model
with accurate estimation will be desired. The paper
provides the solution that uses prediction model of a
genetic algorithm as an optimization algorithm that can
be used to improve the effort estimation accuracy in
agile software development methodology by
chromosomes that is made up of a gene pool that
include user stories, friction factors, implementation
Level factor, and dynamic forces. [4]
International Journal of Computer Science and Information Security (IJCSIS),
Vol. 22, No. 4, July-August 2024
https://google.academia.edu/JournalofComputerScience
https://zenodo.org/doi/10.5281/zenodo.12577978
1 https://sites.google.com/site/ijcsis/
ISSN 1947-5500

2. Genetic Algorithm
In artificial intelligence (AI), an evolutionary
algorithm (EA) is a subset of evolutionary computation,
a generic population-based metaheuristic optimization
algorithm. An evolutionary algorithm uses mechanisms
inspired by biological evolution, such as reproduction,
mutation, recombination, and selection. Candidate
solutions to the optimization problem play the role of
individuals in a population, and the fitness function
determines the quality of the solutions. The evolution
of the population then takes place after the repeated
application of the above operators. [5]
2.1 Genetic Algorithm Process
In Genetic evolutionary optimization Algorithms
techniques that have several alternative solutions to
transfer the problem from its real domain to the
domain of evolutionary algorithms. the result of the
solution is closer to the optimal, where the
evolutionary process starts with the number of random
solutions (population). [6]
These solutions (individuals)
are then encoded according to the current problem, and
the quality of each individual is evaluated through
computation a fitness of each individual in the initial
population, after which the current population changes
to a new population as showing fig 2 by applying three
basic operators:
Fig 2: Genetic Algorithm Flowchart
1) individuals Selection for reproduction using
some of the selection mechanisms
2) Create an offspring using crossover and
mutation operators. The probability of crossover and
mutation are selected based on the application.
3) Compute the new generation of the population.
This process will end either when the maximum
number of generations is reached or the optimal
solution is found.
2.2 Algorithm
1) Start: Create a random population of potential
solutions consisting of n individuals.
2) Fitness: Evaluate the fitness value f(x) of each
individual, x, in the population.
3) New population: Repeat the following steps to
create a new population until completion of the new
population.
4) Selection (Reproduction): The process of
choosing the "best" parents in the community for
mating, "best" is defined based on the current problem,
has an active role in solving premature convergence
resulting from a lack of diversity in the population,
therefore identifying the appropriate selection
technique is a critical step. [7]
2.3 Genetic Algorithm Parameters
The parameters of the Genetic Algorithm, including
the following:
1) The size of the Population: the selection of the
size of the population is a critical issue, if the size of
the population is very small, which means little search
space, so it is possible to reach the local optimum. if
the size of the population is too large, this will increase
the area of search and increase the mathematical load,
and therefore the process becomes slow, therefore, the
size of the population must be reasonable.
2) The rate of Crossover: This determines the
number of times a crossover occurs in the
chromosomes in one generation, which is between 0%-
100%; crossover rate is a delicate matter.
3) The rate of Mutation: Determines how many
genes mutate in one generation; that between 0%-
100%, and is also a delicate matter. decreases or
Increases in the rate of mutation and crossover can
have both a negative and positive effect. [8]
ISSN 1947-5500

4) The number of generations: Several cycles
before termination. In some cases, hundreds of loops
are sufficient and others are not, and this depends on
the complexity of the problem.
All these Genetic Algorithm parameters are important
because they determine the quality of the solution.
2.4 Biological Chromosomes in GA Algorithm
Can be surmised in the following points:
1) Genetic information is stored in the
chromosomes
2) Each chromosome is built of DNA
3) Genes are encoded in the chromosomes
4) Genes code for proteins
5) Every gene has a unique position on the
chromosome
The Chromosome in a genetic algorithm represents the
set of possible combinations within the search space. It
is commonly represented as a binary string of 0s and 1s
as showing in Fig 3. In this paper, the Chromosome
consists of the following parts: Story Point (SP),
Implementation Level Factors (ILF), Friction Factors
(FR), and Dynamic Forces. effectively constructing a
population of SP, ILF, FR, DF. For SP considered two
possibilities, represented by a binary string of 1 bit (19
= 2). For ILF, we considered three possible techniques,
also represented by a binary string of 3 bits (24 = 8).
For FR, DF, we considered twelve different
possibilities, which required 4 bits to represent (24 =
16). With this Chromosome representation, the goal of
the genetic algorithm is to find the Chromosome that
maximizes a fitness function. We only worked with
combinations within the proposed range. For example,
the range of the attribute selection is between 1 to 5.
This means that combinations with the attribute
selection set of 6 to 8 are not valid.
Fig 3: Binary String Example
3. Effort Estimation Techniques
To determine the process of the best-predicted
effort required to develop a software project is termed
as software effort estimation. This estimation can be
divided into three levels. The first level consists of size
estimation, the second level to involves effort
estimation and the third level is the cost estimation,
and the effort estimation is calculated in terms of
(Person-Months). [9]
Software effort estimation techniques can be
classified into algorithmic and non-algorithmic models.
The recent researches focused on an algorithmic model
such as:
1) Line of Code (LOC).
2) Functions point matrices.
3) COCOMO and COCOMO-II.
4) Object points.
5) Software life cycle management.
6) Use case estimation.
7) Story points.
Agile software development methodology plays an
important role in implementing communication goals
between customers in a planned manner. The main
issue of determining effort estimation in agile methods
is to focus on the effort and degree of difficulties of
teamwork rather than an individual. The main task is to
estimate the required effort keeping in mind the
technical complexity of the project. Agile software has
many advantages as it is iterative, modular,
incremental, customer-oriented and time-bound. [10]
This paper focused on the factors of the story point
to estimate the efforts in agile software development.
The story point is the most common estimation
technique for agile software development.
Story points refer to an estimate of the relative
scale of the work in terms of actual development effort.
Story Point estimation is done using relative sizing by
comparing one story with a sample set of big sized
stories. [11]
As Show in Teable1 Relative sizing across
stories tends to be much more accurate over a larger
sample, than trying to estimate each story for the effort
involved. Planning poker is one of the most useful
tools in agile software development.
Table 1：Story Point (SP) wights
Terms S Wight
Very Small Story VSS 1
Small Story SS 2
Medium Story MS 3
Large Story LS 5
Very Large Story VLS 8
Table 1 shows the weight of Story Point (SP) which
can have many iterations depending upon the system
requirements. The weights assigned to each customer
or user stories are predicted.
In Implementation Life Factors (ILF) includes the
terms Off the Shelf (OS), Full Experience Components
(FEC), Partial Experience Components (PEC), and
New Components (NC). as shows in Table 2 the
scaling factor for implementation level which
represents the level of understanding each user story.
Table 2：Implemntation Life Factors (ILF) Scale
Term Ser Scale
Off the Shelf OS 1
Full Experience Components FEC 2
Partial Experience Components PEC 3
New Components NC 4
The optimization process refers to study the
constraints of a project that should be completed before
ISSN 1947-5500

the calibration to improve the stability of the velocity
calculation. This process includes two factors are
Friction Factor (FR) and Dynamic Forces (DF).
The friction includes a range of factors that may
affect the team velocity they are team composition, the
process, environmental factors, and dynamic of the
team. as shown in Table 3. the composition of the team
concerns the skills and attitudes of team members. The
process factor refers to the percentage of change in the
agile methods, release, building, and testing. The
environmental factors refer to the noise, poor
ventilation, poor lighting, uncomfortable seating and
desks, inadequate hardware, or software. Team
dynamics are patterns of interaction among team
members that determine the performance of the team.
Table 3： Friction Factors (FR)
Friction factors Stable volatile
Highly
volatile
Very highly
volatile
Team composition 1 .98 .95 .91
Process 1 .98 .94 .89
Environmental
factors
1 .99 .98 .96
Team dynamic 1 .98 .91 .85
As shown in Table 3 the Friction Factors (FR)
with a range of values. Each factor has been adjusted
according to their risk level (stable, volatile, highly
volatile, and very highly volatile).
The dynamic forces (DF) refers to the factors that
could lead to a loss of velocity. These factors are
unexpected or non-predictable. The dynamic forces
(DF) include some of the factors - team changes, new
tools, vendor defects, responsibilities out of the project,
personal issues, stakeholders, unclear requirements,
changing requirements, and reallocation. as shown in
Table 4, Those factors are ranked from (Normal, high,
very high, and extra high). [12]
Table 4： Dynamic Forces (DF) variables
Variable Factors
Normal
High
Very
High
Extra-
High
Expected to team change 1 .98 .95 .91
Introduction to a new tool 1 .99 .97 .96
Vendor’s Defect 1 .98 .94 .90
Responsibilities out of the project 1 .99 .98 .98
Personal Issues 1 .99 .99 .98
Expected Delay in Stakeholder
response
1 .99 .98 .96
Expected Ambiguity in Details 1 .98 .97 .95
Expected Changes in environment 1 .99 .98 .97
Expected Relocation 1 .99 .99 .98
when the project is complete, the effort will be
the sum of membership degree of Story Point (SP),
Implementation Life Factors (ILF), and population of
all individual user stories. The Eq. (2) shows the fuzzy
story size (FSS) by using the triangle membership
function μ.
3.1 Performance Criteria
There are some of the performance measures
criteria existing in the some of literature. The
performance of different models can be evaluated by
using the main criteria as the following:
1) Mean size of relative error- It is used to
calculate the percentage of the difference between the
actual and estimated effort and finally assess the
relative error averaged over a large number of values
for a set of the effort. [13]
actual i, is the actual effort required of i the data
test, predicted i, is the estimated effort from the
prediction of i data test. N represents the total count of
variables in the dataset. [14]
2) Mean squared error is used to calculate the
square of mean relative error by finding the difference
between estimated effort, actual and dividing by the
total number of datasets[15]
3) The third performance measure metric is the
percentage of prediction that represented as PRED(x),
given by
4) when complete the project, the effort will be
the sum of membership degree of SP, ILF, and Com
for all individual user stories. The below Equation
shows the fuzzy story size (FSS) by using the triangle
membership function μ. [16]
5) Fuzzy FR (FFR) is calculated as the product of
all four components of FR and the membership degree.
The below equation Shows the FFR calculation by
using the triangle membership function μ: [17]
ISSN 1947-5500

4. Proposed Algorithm
The proposed design to is calculating the effort
estimation in agile software development by creating a
Multi-Population of Genetic Algorithm as showing in
Fig 4. the effort estimation is calculated as offspring
generated by inbreeding and crossbreeding in each
phase from software project implementation phases,
and the candidate for each population of the next
generation, each population include two chromosomes.
[18] [19]
The Proposed algorithm is described as follows:
1) Start- Initializing Population, by creating a two
population of potential solutions consisting of n
individuals.
Population (A) - include two chromosomes, the first
one called (Story Point) SP Chromosome as shown in
Table 5, which contains some of the gens. each gene is
representing a wight of Story Point factors, and the
second chromosome called Implementation Life
Factors (ILF), which also contains some of the gens.
each gene is representing a scale of ILF scales.
Population (B) - include two chromosomes, the first
one called Friction Factor (FR) Chromosome, That
Contains some of the gens. each gene is representing a
risk factor of FR scales. and the second chromosome
called Dynamic Forces (DF), which also contains some
of the gens. each gene is representing a variable of DF
factors. [20]
Fig 3: Multi Population of Genetic Algorithm
2) Fitness - Evaluate the fitness value f(x) of each
individual, x, in the population (A), then Evaluate the
fitness value in the population (B)
Story Point (SP) Chromosome:
Table 5 shows SP Chromosome which can have
many iterations depending upon the system
requirements. The weights assigned for each user
stories are predicted.
Table 5：Story Point (SP) Chromosome
0001 0010 0011 0101 1000
1 2 3 5 8
VSS SS MS LS VLS
Very
Small
Story
Small
Story
Medium
Story
Large
Story
Very
Large
Story
Implementation life factors (ILF) Chromosome
Table 6 shows (ILF) Chromosome scaling factor for
implementation level which represents the level of
understanding each user story, that include terms Off
the Shelf (OS), Full Experience Components (FEC),
Partial Experience Components (PEC), and New
Components (NC).
Table 6：(ILF) Chromosome
0001 0010 0011 0100
1 2 3 4
OS FEC PEC NC
Off the Shelf
Full
Experience
Components
Partial
Experience
Components
New
Components
Friction Factors (FR) Chromosome
Table 7 shows the (FR) Chromosome, with a range
of values. Each factor has been adjusted according to
their risk level (stable, volatile, highly volatile, and
very highly volatile).
Table 7：Friction Factor (FR) Chromosome
Stable Volatile
Highly
volatile
Very
highly
volatile
4 3 2 1
Team composition 0100 0011 0010 0001
Process 0100 0011 0010 0001
Environmental factors 0100 0011 0010 0001
Team dynamic 0100 0011 0010 0001
Team composition 0100 0011 0010 0001
Friction Factors (FR) Chromosome
Table 8 shows the (DF) Chromosomes refers to the
factors that could lead to loss of velocity. These factors
are unpredictable and unexpected.
Table 8：Dynamic Forces (DF) Chromosome
Normal
High
Very
High
Extra-
High
Expected to team change 0001 0010 0011 0100
Introduction to a new tool 0001 0010 0011 0100
Vendor’s Defect 0001 0010 0011 0100
Responsibilities out of the project 0001 0010 0011 0100
Personal Issues 0001 0010 0011 0100
Expected Ambiguity in Details 0001 0010 0011 0100
Expected Changes in the
environment
0001
0010 0011 0100
Expected Relocation 0001 0010 0011 0100
ISSN 1947-5500

After the change, all estimation effort techniques
(SP, ILF, FR, DF) to genes in chromosomes will
perform some operations of genetic algorithm
(Selection, Crossover, Mutation) Severally for each
population.
Assuming we have an agile software, project divided
into sub-projects and we put random values of a story
point in the population (A) shows Table 9- as the
following:
Table 9：Population A
SP1 SP2 SP3 SP4 SP5 SP6
Project 1 2.4 0.7 8 -2 5 1.1
Project 2 -0.4 2.7 5 -0.1 7 0.1
Project 2 -0.1 2 2 -3 2 0.9
Project 3 4 7 12 6.1 1.4 -4
Project 4 3.1 4 0 2.4 4.8 0
Project 5 -2 3 -7 6 3 3
As showing in Table 10 – the 6 stories point (SP)
with total estimation effort (44.1%) in the first project.
that can enhance the result for estimation effort and
each story point (SP) will represent a gen in the genetic
algorithm chromosome.
Table 10：project 1 Chromosome
Gen0 Gen1 Gen2 Gen3 Gen4 Gen5
Chromosome 1
SP1 SP2 SP3 SP4 SP5 SP6 SP Project 1
2.4 0.7 8 -2 5 1.1 44.1
SS SS VLS SS VLS SS
The SP = SP1 Gen0 + SP2 Gen1 + SP3 Gen2 + SP4 Gen3
+ SP5 Gen4 + SP6 Gen5
The goal is to find the set of parameters (SP1:SP6)
that map the following input to its output
SP’ = 4 SP1 + 2 SP2 + 7 SP3 + 5 SP4 + 11 SP5 + SP6
To calculate the fitness function using the following
equation in the regression model:
SP’ = 4 SP1 + 2 SP2 + 7 SP3 + 5 SP4 + 11 SP5 + SP6
SP’ = 4 x 2.4 – 2 x 0.7 + 7 x 8 + 5 x -2 +11 x 5 + 1.1
SP’ = 110.3
Then calculate the fitness function for all
chromosomes in the population (A) with the same
previous steps as shown in Table 11.
Table 11：Fitness Function for all Chromosomes
Story Points (SP Chromosomes) – Population A SP’ F(C)
2.4 0.7 8 -2 5 1.1 110.3 0.015
-0.4 2.7 5 -0.1 7 0.1 100.1 0.018
-0.1 2 2 -3 2 0.9 13.9 0.033
4 7 12 6.1 1.4 -4 127.9 0.012
3.1 4 0 2.4 4.8 0 69.2 0.04
-2 3 -7 6 3 3 3 0.024
3) Selection – The individuals in this type are
selected according to their rank, but first we need to
sort the chromosomes based on their fitness value. this
process will be choosing the best parents or the best
chromosomes from table 12 Which carries the best
fitness function.
As shown in Table 12 the best parents from
population A
Table 12：Best Chromosomes
Story Points (SP Chromosomes) – Population A SP’ F(C)
P1 -0.1 2 2 -3 2 0.9 13.9 0.033
P2 3.1 4 0 2.4 4.8 0 69.2 0.04
P3 -2 3 -7 6 3 3 3 0.024
4) Crossover - in this operation will take two
parents (chromosomes) to generate a new offspring by
switching segments of the parent genes. It is more
likely that the new offspring (children) as shown in fig
4 will contain a good part of their parents, and
consequently perform better as compared to their
ancestors.
Fig 4: Crossover operation
The first crossover between two parents (P1, P2) as
showing in fig 5 to generate new offspring.
Fig 5: New Generate of Offspring
Then makes a crossover between other parents (P1,
P2) with the same previous steps.
5) Mutation – The mutation is one of the most
important phases of the genetic algorithm, by using
random of values of the story point factors (VSS, SS,
MS, LS, VLS), where it is changed or is switched
between specific genes within a single chromosome to
create chromosomes, that provide new enhancement
solutions for the generation of estimation effort.
Fig 6: New Generate of Offspring
As shown in Fig 6 the describe the mutation of the
chromosome by replacing the gene (4.8) with the gene
(2.4) as a random value from story point factors.
Then makes the mutation for all chromosomes that
generated form the crossover phase, with the same
previous steps.
ISSN 1947-5500

After complete, the mutation phase will become the
first generation of the Population (A) as show the
generation 1 in Table 13.
Table 13: Generation 1 from Population A
Generation 0 – Population A
P1 -0.1 2 2 -3 2 0.9
P2 3.1 4 0 2.4 4.8 0
P3 -2 3 -7 6 3 3
Generation 1 – Population A
P1 -0.1 2 2 2.4 4.8 0
P2 3.1 4 0 6 3 3
P3 -2 3 -7 -3 2 0.9
When the new generation generated from the
population (A), all the phases of the genetic algorithm
will be re-applied on this generation (Fitness value,
Selection, Crossover, Mutation). The algorithm will be
stopped when achieves the goal, after completing the
mating pool of all chromosomes and completing the
mutation phase for all genes in the population, that use
random values of the story points factors, To ensure
that all point stories are addressed to extract effort
estimation for only story points Technique. can change
the value of the gene of story point (SP) chromosomes
to binary sting if the gene contains a set of attributes.
In implementation Life Factors (ILF) technique, that
the second chromosomes of the population (A) and the
Population (B), that include two chromosomes
(Friction Factor FR, Dynamic Forces DF), will be
appley the proposed genetic algorithm phases (Fitness
value, Selection, Crossover, Mutation), with the same
previous steps in all phases. The purpose of using the
Multi-Genetic Algorithm to accelerate the algorithm,
by dividing each of the two chromosomes into one
population.
5. Results
The Genetic Algorithm parameters that were
selected in the first test of the population (A), Which
underwent a set of proposed algorithm operations
showed the following results for effort estimation for
story points chromosomes:
Fig 7: Effort Estimation of SP
as shown in Fig 7 the proposed algorithm producing
an accurate estimation of effort for story point.
In implementation Life Factors (ILF) Chromosomes,
the proposed genetic algorithm extracted accurate
results, which represents the level of understanding
each user story.
Fig 8: Effort Estimation of ILF
Fig 9 shows the proposed genetic algorithm extract
the effort estimation for dynamic forces assigned
values, which refers to the factors that could lead to
loss of velocity. The factors are ranked from (Normal,
high, very high, and extra high)
Fig 9: Effort Estimation of Dynamic Forces
Fig 10 shows the Effort Estimation for friction factor
FR with a range of values, also put each factor
according to their risk level (stable, volatile, highly
volatile, and very highly volatile).
Fig 10: Effort Estimation - Friction Factor
ISSN 1947-5500

6. Conclusion
This paper aimed to propose to improve the effort
estimation in agile software development methodology
by using Multi - genetic algorithm. the proposed model
has been put a population of candidate solutions for
(Story Points, Implementation Life Factors, Fraction
Factors, and Dynamic forces) to optimization the effort
estimation in agile software development methodology.
Each candidate solution has a set of properties (Parents,
chromosomes) which can be mutated and altered;
traditionally, solutions fit represented in decimal or
binary as strings of 0s and 1s, and other encodings are
also possible.
References
[1] Srivastava, P. R., & Kim, T. H¸ Application of
genetic algorithm in software testing. International
Journal of software Engineering and its Applications,
2009, 3 (4), 87-96.
[2] Raja, P. V., & Bhaskaran, V. M, Improving the
Performance of Genetic Algorithm by reducing the
population size, International Journal of Emerging
Technology and Advanced Engineering ,2013, 86-91.
[3] Nagy Ramadan Darwish, “Improving the Quality of
Applying eXtreme Programming (XP) Approach",
International Journal of Computer Science and Information
Security (IJCSIS), November 2011, ISSN 1947-5500, Vol. 9
No. 11.
[4] Patil, S., & Bhende, M, Comparison and Analysis of
Different Mutation Strategies to improve the Performance of
Genetic Algorithm, (IJCSIT) International Journal of
Computer Science and Information Technologies,2014,5 (3),
4669-4673.
[5] Paulinas, M, & Ušinskas A, A survey of genetic
algorithms applications for image enhancement and
segmentation, Information Technology and control,2015,36.
[6] Sharma, P., & Wadhwa, A, Analysis of Selection
Schemes for Solving an Optimization Problem in Genetic
Algorithm. International Journal of Computer Applications,
2014, 93 (11).
[7] Kumar, A, Encoding schemes in genetic algorithm,
International Journal of Advanced Research in IT and
Engineering, 2013, 2 (3), 1-7.
[8] Soni, N., & Kumar, T, Study of Various Mutation
Operators in Genetic, (IJCSIT) International Journal of
Computer Science and Information Technologies ,
2014. 5 (3), 4519-4521.
[9] Dong, M., & Wu, Y., Dynamic Crossover and Mutation
Genetic Algorithm Based on Expansion Sampling. Artificial
Intelligence and Computational Intelligence, 2009, 141-149.
[10] Ziauddin Zia,Shahid Kamal Tipu, and Shahrukh Zia
“An Effort Estimation Model for Agile Software
Development”, Advances in Computer Science and its
Applications (ACSA), Vol.2, 2012.
[11] Eldos, T. ,A New Migration Model For Distributed
Genetic Algorithms, Proceedings of the International
Conference on Scientific Computing ,2006, (CSC06), (pp.
128-134).
[12] Korejo, I., Yang, S., Brohi, K., & Khuhro, Z. U, Multi-
Population Methods with Adaptive Mutation for Multi-
Modal Optimization Problems. International Journal on Soft
Computing, Artificial Intelligence and Applications
(IJSCAI), 2013 ,2 (2).
[13] Sandeep Kad, Vinay Chopra, Fuzzy Logic based
framework for Software Development Effort Estimation, An
International Journal of Engineering Sciences, 2011, Vol. 1.
[14] Lee, W. C., Wang, J. Y., & Lee, L. Y, A hybrid genetic
algorithm for an identical parallel-machine problem with
maintenance activity, Journal of the Operational Research
Society, 2015.
[15] Evita Coelho,Anirban Basu, “Effort Estimation in Agile
Software Development using Story Points,” International
Journal of Applied Information Systems (IJAIS),Volume 3–
No.7, August 2012.
[16] Sandeep Kad, Vinay Chopra, “Fuzzy Logic based
framework for Software Development Effort Estimation”, An
International Journal of Engineering Sciences (IJES), Vol. 1,
2011.
[17] Vishal Sharma and Harsh Kumar Verma, “Optimized
Fuzzy Logic Based Framework for Effort Estimation in
Software Development”, International Journal of Computer
Science Issues (IJCSI), Vol. 7, Issue 2, No 2, 2010.
[18] Osaba, E., Onieva, E., Carballedo, R., Diaz, F., &
Perallos, A, An adaptive multi-crossover population
algorithm for solving routing problems. In Nature Inspired
Cooperative Strategies for Optimization, .2014, (NICSO
2013) (pp. 113-124). Springer International Publishing.
[19] H. Pourvaziri and B. Naderi, “A hybrid multi-
population genetic algorithm for the dynamic facility layout
problem,” Applied Soft Computing Journal, 2014, vol. 24,
pp. 457–469.
[20] Patil, S., & Bhende, M, Comparison and Analysis
of Different Mutation Strategies to improve the
Performance of Genetic Algorithm. (IJCSIT)
International Journal of Computer Science and
Information Technologies, 2014 5 (3), 4669-4673.
Hossam Al-Ansary: Doctor degree for computer science
in Cairo University, Computer and Information Technology
Dept.
ISSN 1947-5500

A_Proposed_Genetic_Algorithm_Model_to_Im.pdf

More Related Content

Similar to A_Proposed_Genetic_Algorithm_Model_to_Im.pdf (20)

More from Hossam Al-Ansary (7)

Recently uploaded (20)

A_Proposed_Genetic_Algorithm_Model_to_Im.pdf