Volume 10, Issue 2, February – 2025 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14930503

A Proposed Model for an Artificial Intelligence

Algorithm to Improve Pharmaceutical Industry
Hossam Abdelrahman Al-Ansary1
1
Information Technology and Computer Sciences -Cairo University, Égypt

Publication Date: 2025/02/26

Abstract: An Artificial Intelligence (AI) is revolutionizing the pharmaceutical industry by enhancing and streamlining
numerous processes, ranging from drug discovery to clinical trial optimization. AI technologies help accelerate processes,
reduce costs, and improve effectiveness, enabling pharmaceutical companies to achieve long-term competitive advantages.
This study aims to advance the pharmaceutical industry by designing a software solution that utilizes a genetic algorithm
based on artificial intelligence. In this research, drug components are represented as genes within a set of chromosomes,
allowing for the optimization of pharmaceutical molecule design by exploring vast spaces of possible chemical compounds.
The proposed algorithm identifies the most effective molecules while minimizing potential side effects, significantly
accelerating the drug discovery process by reducing the time and cost required to produce new or advanced formulations.
This innovative approach holds the potential to transform drug development and improve outcomes in the pharmaceutical
sector.

Keywords: Artificial Intelligence, Genetic Algorithm, Evolutionary Algorithm, Software Engineering, Knowledge Base.

How to Cite: Hossam Abdelrahman Al-Ansary. (2025). A Proposed Model for an Artificial Intelligence Algorithm to Improve
Pharmaceutical Industry. International Journal of Innovative Science and Research Technology,
10(2), 696-708. https://doi.org/10.5281/zenodo.14930503.

I. INTRODUCTION A Knowledge Base System is a cornerstone of modern

pharmaceutical operations, enabling innovation,
Artificial Intelligence (AI) is a field of computer compliance, and efficiency across the drug development
science aimed at developing systems capable of performing lifecycle. Its integration with advanced technologies like AI
tasks that typically require human intelligence. These tasks and big data analytics is transforming the industry, paving
include learning, reasoning, problem-solving, perception, the way for faster, safer, and more effective therapies. [4]
understanding natural language, and interacting with the
environment. .[1] The Research and Development (R&D) Department in
pharmaceutical industries is pivotal in driving innovation
The emergence of artificial intelligence (AI) has and ensuring the development of safe, effective therapeutics.
sparked a significant transformation in the pharmaceutical Its functions span the entire drug lifecycle, from discovery
sector, driving a paradigm shift across multiple areas such as to post-market monitoring.[5] The pharmaceutical
drug discovery, formulation development, manufacturing, manufacturing process is intricate and demands high
quality control, and post-market surveillance. This review precision and rigorous oversight to guarantee the quality,
provides a comprehensive analysis of the diverse impacts of efficacy, and safety of drugs. Below are the essential steps
AI-driven technologies on the entire pharmaceutical life involved in drug manufacturing:
cycle. It explores the use of genetic algorithms, data
analytics, and predictive modeling to streamline drug  Drug Discovery: Identification of chemical or biological
discovery, improve formulation optimization, and boost compounds with potential therapeutic effects for specific
manufacturing efficiency.[2] diseases.
 Preclinical Testing: Evaluation of compounds in
A Knowledge Base System (KBS) in the laboratory settings and animal models to assess efficacy
pharmaceutical industry is a centralized repository of and safety. [6]
structured and unstructured data, designed to store, organize,
and retrieve critical information. It serves as a vital tool for
decision-making, regulatory compliance, innovation, and
operational efficiency. [3]

IJISRT25FEB677 www.ijisrt.com 696

Volume 10, Issue 2, February – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14930503

 Clinical Trials: Human testing conducted in three  Safe Disposal

phases:
 Waste Management: Proper disposal of chemical and
 Phase I: Focuses on safety and determining the biological waste in line with environmental regulations.
appropriate dosage.
 Phase II: Assesses efficacy and identifies potential side  Continuous Improvement
effects.
 Phase III: Confirms efficacy and safety in a larger patient  Process Optimization: Ongoing refinement of
population. manufacturing processes to enhance efficiency and
reduce costs.
 Formulation and Design  New Product Development: Exploration of new
therapeutic applications for existing active ingredients.
 Formulation Development: Determination of the drug's
final form (e.g., tablets, capsules, injections) and the Pharmaceutical manufacturing relies on the
inclusion of excipients such as preservatives and coloring collaboration of chemists, pharmacists, engineers, and
agents. [7] quality assurance experts to ensure the production of safe
 Optimization: Ensures the formulation's chemical and and effective medications for patients. This research focuses
biological stability. on the R&D Department for Process Optimization, which
involves the ongoing refinement of manufacturing processes
 Manufacturing to enhance efficiency and reduce costs by using genetic
algorithm. Where The research and development team
 Active Pharmaceutical Ingredient (API) conduct numerous attempts to refine the formulation, adjust
Production: Synthesis of the active ingredient through the proportions and values of the preparation, and carry out
chemical reactions or biotechnological methods. experiments until the best possible outcome is achieved. [9]
 Primary Manufacturing: Combining the active ingredient
with excipients to produce the final drug product. II. EVOLUTIONARY ALGORITHMS
 Packaging: Packaging the drug in appropriate containers
(e.g., bottles, blister packs) with accurate labeling. [8] An evolutionary algorithm is a type of artificial
intelligence-based computational tool designed to solve
 Quality Control complex problems by simulating processes inspired by
biological evolution. It mimics natural behaviors such as
 Chemical and Physical Testing: Ensures each batch reproduction, mutation, and recombination to iteratively
meets specified standards. improve solutions over time. Evolutionary algorithms
operate through a Darwinian-like process of natural
 Biological Testing: Verifies efficacy and safety.
selection, where weaker solutions are eliminated, and
 Documentation: Maintains detailed records of all
stronger, more viable options are retained and refined in
processes and tests to ensure traceability and regulatory
subsequent iterations. The ultimate goal is to arrive at
compliance.
optimal or near-optimal solutions that achieve the desired
outcomes. [10]
 Regulatory Approval
 Benefits of Evolutionary Algorithms
 Submission of Applications: Applications are filed with
Evolutionary algorithms offer several advantages,
regulatory bodies (e.g., FDA, EMA) for marketing
making them highly effective for solving a wide range of
approval. problems:
 Inspections: Facilities are audited to ensure adherence to
regulatory standards.  Increased Flexibility:
Evolutionary algorithms can be adapted and modified
 Distribution and Marketing to address highly complex problems across various domains.
Their flexible framework allows them to meet specific
 Distribution: The drug is supplied to pharmacies, objectives and tackle challenges that traditional methods
hospitals, and other healthcare providers. may struggle with. [11]
 Marketing: Promotional activities are conducted in
compliance with advertising regulations.  Better Optimization:
These algorithms consider a vast "population" of
 Post-Marketing Surveillance potential solutions, enabling them to explore a wide search
space. Unlike traditional methods that may be limited to a
 Pharmacovigilance: Ongoing monitoring of side effects narrow set of solutions, evolutionary algorithms are not
after the drug is released to the market. restricted to a single approach, increasing the likelihood of
 Updates: Medical information is revised based on new finding optimal or near-optimal solutions. [12]
findings.

IJISRT25FEB677 www.ijisrt.com 697

Volume 10, Issue 2, February – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14930503

 Unlimited Solutions:  Selection:

Unlike classical optimization techniques that often
focus on maintaining a single best solution, evolutionary  Choosing the fittest chromosomes to produce offspring
algorithms can generate and present multiple potential (new generations).
solutions to a problem. This diversity allows decision-  Usually done using methods like "Roulette Wheel
makers to choose from a range of viable options based on Selection" or "Rank Selection."
specific criteria or constraints. [13]
 Crossover:
III. GENETIC ALGORITHMS AND
EVOLUTIONARY ALGORITHMS  A process where two chromosomes (parents) are
combined to produce a new chromosome (offspring).
Genetic Algorithms (GAs) are among the most  One or more points in the chromosome are selected, and
prominent and widely used algorithms within the broader genes are exchanged between the parents.
category of Evolutionary Algorithms (EAs). Inspired by the
principles of biological evolution, GAs utilize mechanisms  Mutation:
such as selection, crossover, and mutation to continuously
improve and optimize solutions. They are particularly  A random change in one or more genes of the
effective for problems involving continuous improvement, chromosome.
optimization, and search in complex, multidimensional  Helps maintain genetic diversity and avoid getting stuck
spaces. [14] in local optima.
Evolutionary algorithms, including genetic algorithms,  Replacement:
provide a powerful and flexible approach to problem-solving
by leveraging the principles of natural selection and  Replacing the least fit individuals in the population with
biological evolution. Their ability to explore vast solution new offspring.
spaces, adapt to complex problems, and generate multiple
 Replacement can be complete or partial.
potential solutions makes them invaluable tools in fields
ranging from engineering and logistics to artificial
 Mechanism of Genetic Algorithms:
intelligence and data science. [15]
 Initialization: Create an initial population of solutions
IV. GENETIC ALGORITHMS
randomly.
 Evaluation: Calculate the fitness value for each
Genetic Algorithms (GAs) are optimization techniques
individual in the population using the fitness function.
inspired by natural evolutionary processes, such as natural
selection, mutation, and crossover (reproduction). These  Selection: Select the fittest individuals to produce
algorithms are used to solve complex or nonlinear offspring.
optimization problems, where traditional methods are either  Crossover: Apply the crossover process to create new
inefficient or difficult to apply. [16] individuals.
 Mutation: Apply mutations to the new individuals to
A. Components of Genetic Algorithms: increase diversity.
 Replacement: Replace old individuals with new ones.
 Chromosome:  Iteration: Repeat steps 2 to 6 until a stopping condition
is met (e.g., a specific number of generations or reaching
 Represents a potential solution to the problem. a satisfactory solution).
 Typically represented as a string of genes (numbers,
Genetic algorithms are used in various fields such as
characters, or other data).
design optimization, machine learning, robotics, and many
other applications that require finding optimal solutions in
 Population:
large and complex search spaces. [17]
 A group of potential solutions (chromosomes). B. Genetic Algorithm Process
 Initially generated randomly. In genetic evolutionary optimization algorithms,
techniques are used to transform a problem from its real
 Fitness Function: domain into the domain of evolutionary algorithms by
generating several alternative solutions. The goal is to
 Evaluates the quality of each chromosome (solution). achieve a result that is closer to the optimal solution. The
 Depends on the specific problem being solved. evolutionary process begins with a set of random solutions,
known as the population. [6] These solutions, referred to as
individuals, are encoded based on the specific problem at
hand. The quality of each individual is then evaluated by
computing its fitness value within the initial population.

IJISRT25FEB677 www.ijisrt.com 698

Volume 10, Issue 2, February – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14930503

Following this evaluation, the current population evolves continues until one of the following stopping conditions is
into a new population through the application of three met:
fundamental operators, as illustrated in Fig 1. [18]
 The maximum number of generations is reached.
 An optimal or satisfactory solution is found.

The algorithm terminates at this point, providing the

best solution discovered during the evolutionary process.

C. GA Algorithm

Fig 1: Genetic Algorithm Flowchart Fig 2: Genetic Algorithm

 Selection of Individuals for Reproduction:  Start:

Individuals are chosen for reproduction using specific Begin by creating an initial population of potential
selection mechanisms, such as roulette wheel solutions. This population consists of n randomly generated
selection, tournament selection, or rank-based selection. individuals, each representing a possible solution to the
These mechanisms prioritize individuals with higher fitness problem. [21]
values, increasing their chances of contributing to the next
generation. [19]  Fitness Evaluation:
Evaluate the fitness value f(x) of each individual x in
 Creation of Offspring: the population. The fitness function measures how well each
New offspring are generated by solution performs relative to the problem's objectives.
applying crossover and mutation operators. Individuals with higher fitness values are considered better
solutions.
 Crossover: Combines genetic information from two
parents to produce one or more offspring. The probability  New Population Creation:
of crossover is determined based on the application. Repeat the following steps to generate a new
 Mutation: Introduces random changes to the offspring's population until the desired population size is achieved:
genetic information to maintain diversity and explore
new solutions. The probability of mutation is also  Selection (Reproduction):
selected based on the application. Select the "best" individuals from the current
population to serve as parents for the next generation. The
 Computation of the New Generation: definition of "best" depends on the specific problem and is
The new generation of the population is formed by typically based on fitness values. Selection plays a crucial
replacing less fit individuals with the newly created role in maintaining diversity within the population and
offspring. This ensures that the population evolves toward preventing premature convergence, where the algorithm
better solutions over time [20]. This iterative process settles on a suboptimal solution. Choosing an appropriate

IJISRT25FEB677 www.ijisrt.com 699

Volume 10, Issue 2, February – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14930503

selection technique is a critical step in ensuring the  Number of Generations:

algorithm's effectiveness. [7] This parameter defines the number of cycles the
algorithm will execute before termination. In some cases, a
 Crossover (Recombination): few hundred generations may suffice to find a satisfactory
Combine the genetic information of selected parents to solution, while more complex problems may require
produce offspring. This process mimics biological significantly more iterations. The optimal number of
reproduction and introduces new combinations of traits into generations depends on the problem's complexity and the
the population. [22] desired solution quality.

 Mutation: All these parameters are vital because they collectively

Introduce random changes to the offspring's genetic determine the efficiency, accuracy, and overall success of the
information. Mutation helps maintain genetic diversity and Genetic Algorithm in finding high-quality solutions.
enables the algorithm to explore new areas of the search Properly tuning these parameters ensures a balance between
space that might not be reachable through crossover alone. exploration (searching new areas) and exploitation (refining
existing solutions), ultimately leading to better results. [27]
 Replacement:
Replace the old population with the new population of E. Biological Chromosomes in GA Algorithm
offspring, ensuring that the population size remains constant.
 The key points can be summarized as follows:
This iterative process continues until a termination
condition is met, such as reaching a maximum number of  Chromosomes serve as the storage units for genetic
generations or achieving a satisfactory solution. [23] information.
 Each chromosome is composed of DNA.
D. Genetic Algorithm Parameters  Genes, which are embedded within the chromosomes,
The parameters of the Genetic Algorithm play a crucial carry specific instructions.
role in determining its performance and the quality of the  These genes are responsible for coding proteins.
solutions it produces. These parameters include the  Every gene occupies a distinct and unique location on
following: the chromosome.
 Population Size: The Chromosome in a genetic algorithm represents the
Selecting an appropriate population size is a critical set of possible combinations within the search space. It is
decision. If the population size is too small, the search space commonly represented as a binary string of 0s and 1s as
becomes limited, increasing the risk of converging to a local showing in Fig 3. In this paper, the Chromosome consists of
optimum rather than the global one. Conversely, if the the following parts of the drug, Active Ingredient (API),
population size is too large, the search area expands Inactive Ingredients (IAI), Fillers (FI) Increase the volume
significantly, leading to increased computational load and of the drug, Binders (BI) Ensure cohesion of the ingredients,
slower processing. Therefore, it is essential to choose a (Colorants) Provide color, Preservatives (PR) Prevent
reasonable population size that balances exploration and bacterial contamination, Flavors (FL) Improve taste and
efficiency. [24] Lubricants (LU) Facilitate the manufacturing process. [28]
 Crossover Rate: Effectively constructing a population of API, IAI, FI,
The crossover rate determines the frequency at which BI PR, FL and LU. For API considered two possibilities,
crossover operations occur between chromosomes within a represented by a binary string of 1 bit (19 = 2). For IAI, we
single generation. This rate typically ranges between 0% and considered three possible techniques, also represented by a
100%. Setting the crossover rate is a delicate task, as it binary string of 3 bits (24 = 8). For FI, BI, PR, Fl and UL we
directly influences the algorithm's ability to explore new considered twelve different possibilities, which required 4
solutions. An inappropriate crossover rate can either limit bits to represent (24 = 16). With this Chromosome
diversity or introduce excessive randomness, both of which representation, the goal of the genetic algorithm is to find the
can negatively impact the algorithm's performance. [25] Chromosome that maximizes a fitness function. We only
worked with combinations within the proposed range. For
 Mutation Rate: example, the range of the attribute selection is between 1 to
The mutation rate specifies the proportion of genes that 5. This means that combinations with the attribute selection
undergo mutation in a generation, also ranging between 0% set of 6 to 8 are not valid.
and 100%. Like the crossover rate, the mutation rate requires
careful tuning. While a higher mutation rate can introduce
diversity and prevent premature convergence, an excessively
high rate can disrupt good solutions. Conversely, a very low
mutation rate may limit the algorithm's ability to escape local
optima. Both increases and decreases in the mutation and Fig 3: Binary String Example
crossover rates can have either positive or negative effects
on the algorithm's outcomes. [26]

IJISRT25FEB677 www.ijisrt.com 700

Volume 10, Issue 2, February – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14930503

V. PERFORMANCE CRITERIA

In the pharmaceutical industry, equations for

calculating the values of drug components are used to
determine the precise quantities of active ingredients and
excipients (such as solvents, fillers, binders, etc.) required
for drug manufacturing. These calculations depend on
several factors, including the required dosage, the dosage  Calculating Loading Capacity in Capsules
form (tablets, capsules, syrups, etc.), and the properties of When preparing capsules, the quantity of the active
the materials used. Below are some basic equations and ingredient and excipients is calculated based on the caps
concepts used in calculating drug components: evaluate the
average relative error across a large dataset of effort values,
providing a measure of estimation accuracy. [29]

 Calculating the Quantity of Active Pharmaceutical

Ingredient (API)
The active ingredient is the main component of the drug
that provides the therapeutic effect. The required quantity is  Calculating Overall Process Efficiency
calculated based on the specified dosage and the number of In pharmaceutical manufacturing, the overall process
units (e.g., number of tablets or capsules). efficiency is calculated to assess losses or waste.

𝐐𝐮𝐚𝐧𝐭𝐢𝐭𝐲 𝐨𝐟 𝐀𝐏𝐈 = Dosage per unit × Number of unit

 Calculating the Quantity of Excipients

Excipients are inactive components added to improve
the properties of the drug (e.g., stability, solubility, shape,
etc.). They are calculated based on the percentage or total These equations provide a foundation for accurately
weight of the drug. calculating drug components, taking into account the dosage,
pharmaceutical form, and the physical and chemical
properties of the materials used.

VI. PROPOSED DESIGN

 Calculating Solution Concentration The proposed design to calculating the values of drug
For liquid medications (e.g., syrups or injections), the components is used to determine the precise quantities and
concentration of the active ingredient in the solution is accelerate the drug discovery process by reducing the time
calculated. and cost required for produce a new or advanced formulation
by creating a Multi-Population of Genetic Algorithm as
showing in Fig 4. The drug is determined based on offspring
produced through inbreeding and crossbreeding during each
phase of the drug composition process. In every phase,
candidates for the next generation's population are selected,
 Calculating Dilution Factor with each population consisting of two chromosomes. [30]
When diluting a concentrated solution, the dilution [31] The proposed algorithm can be outlined as follows:
equation is used:
 Start Initialization of Population: Begin by generating
C1×V1=C2×V2 two populations of potential solutions, each
comprising n individuals.
Where:  Population (A): This population consists of three
chromosomes. The first chromosome is termed
 C1C1: Initial concentration. the Active Pharmaceutical Ingredient (API)
 V1V1: Initial volume. Chromosome, as illustrated in Table 5. It comprises
several genes, with each gene representing the weight of
 C2C2: Final concentration.
API factors. The second chromosome is referred to as
 V2V2: Final volume.
the Inactive Ingredients (IAI) Factors Chromosome,
which also contains multiple genes. Each gene in this
 Calculating Molar Quantity
chromosome represents a scale of IAI factors. Similarly,
Sometimes, the concentration of substances is
the third chromosome is designated as the Fillers (FI)
expressed using molarity (number of moles per liter).
Chromosome, structured in the same manner as the IAI
chromosome. [32]

IJISRT25FEB677 www.ijisrt.com 701

Volume 10, Issue 2, February – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14930503

 Population (B): This population comprises four A. Active Pharmaceutical Ingredient (API) Chromosome:
chromosomes. The first chromosome is termed The Active Pharmaceutical Ingredient (API)
the Binders (BI) Chromosome, which contains several Chromosome is illustrated in Table 5. This chromosome can
genes, with each gene representing a risk factor undergo multiple iterations based on the specific
associated with BI scales. The second chromosome is requirements of the system. The weights assigned to each
referred to as the Preservatives (PR) Chromosome, API are predicted values.
which also includes multiple genes, each representing a
variable of PR factors. Additionally, there are two other Table 1: API Chromosome
chromosomes: Flavors (FL) and Lubricants (LU), both kilogram Gram Pound Ounce Milligrams
structured similarly to the BI chromosome, with genes VSS SS MS LS VLS
representing their respective factors. [33]. 1 2 3 5 8
0001 0010 0011 0101 1000

B. Inactive Ingredients (IAI) Chromosome

Table 6 shows (IAI) Chromosome scaling factor for
which represents the level of understanding each binder, that
include percentage of (VSS), percentage of Components
(SS), percentage of Components (MS), and New
Components (LS).

Table 2: (IAI) Chromosome

Anti-caking
Stabilizers Vehicles Retardants
Agents
% % % %
1 2 3 4
0001 0010 0011 0100

Table 7 shows the ((FI) Chromosome, which also

includes multiple genes, each representing a variable of PR
factors.

Table 3: (FI) Chromosome

VSS SS MS LS
% % % %
2 3 4 8
0010 0011 0100 1000
Fig 4: Multi Population of Genetic Algorithm
C. Binders (BI) Chromosome
 Fitness Evaluation: Calculate the fitness Table 8 shows the (BI) Chromosomes refers to the
value f(x)f(x) for each individual xx in Population (A). fac-tors that could lead to such as hardness, disintegration
This involves assessing the performance or suitability of time, and drug release rate. These factors are unpredictable
each solution based on predefined criteria. Once and unexpected.
completed, repeat the same process to evaluate the fitness
value for each individual in Population (B). This step
ensures that the quality of potential solutions in both
populations is quantified and compared effectively.

Table 4: Binders (BI) Chromosome

Normal High Very High Extra-High
Adhesion 0001 0010 0011 0100
Solubility 0001 0010 0011 0100
Chemical Stability 0001 0010 0011 0100
Biocompatibility 0001 0010 0011 0100
Flowability 0001 0010 0011 0100
Compressibility 0001 0010 0011 0100
Disintegration Time 0001 0010 0011 0100

IJISRT25FEB677 www.ijisrt.com 702

Volume 10, Issue 2, February – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14930503

The other chromosomes (PR, Fl, and LU) follow a Assuming we have a new product project divided into
process identical to that of the BI chromosome. sub-scenarios, we assign random values to the Active
Pharmaceutical Ingredient (API) in Population (A), as
After the transformation, all chromosomes (API, IAI, shown in Table 9. These random values represent initial
FI, BI, PR, FL, and LU) will be represented as genes within potential solutions for the API factors within the project.
their respective chromosomes. These genes will then These values will serve as the starting point for the genetic
undergo standard genetic algorithm operations— algorithm operations—Selection, Crossover,
Selection, Crossover, and Mutation—independently for and Mutation—applied to Population (A) to optimize the
each population (A and B). API configuration for the new product.

Table 5: Population A
Gen0 Gen1 Gen2 Gen3 Gen4 Gen5
API1 API 2 API 3 API 4 API 5 API 6 API Project 1
2.4 0.7 8 -2 5 1.1 44.1
SS SS VLS SS VLS SS

As illustrated in Table 10, the six APIs, which these APIs will be represented as a gene in the chromosome
collectively account for 44.1% of the total values in the first of the genetic algorithm.
project, have the potential to improve the results. Each of
.
Table 6: Project 1 Chromosome
APIs Chromosomes) – Population A API’ 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

The API’ = API1 Gen0 + API Gen1 + API 3 Gen2 API

Gen3 + API 5 Gen4 + API 6 Gen5

The objective is to identify the optimal set of

parameters (API1:API6 ) that accurately maps the given API’ = 4 API1 + 2 API2 + 7 API3 + 5 API4 + 11 API5 + API6
input to its corresponding output. This will be achieved by API’ = 4 x 2.4 – 2 x 0.7 + 7 x 8 + 5 x -2 +11 x 5 + 1.1
leveraging the genetic algorithm to evolve the chromosome API’ = 110.3
and refine the parameter values for the best possible
outcome.

API’ = 4 API1 + 2 API2 + 7 API3 + 5 API4 + 11 API5 + API6

To calculate the fitness function using the following

equation in the regression model: Then calculate the fitness function for all chromosomes
in the population (A) with the same previous steps as shown
in Table 11.

Table 7: Fitness Function for all Chromosomes

APIs Chromosomes) – Population A API’ 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

 Selection – First, all chromosomes are sorted in selected as parents for the next generation. These
descending order based on their fitness values. This chromosomes are chosen from Table 12, which contains
means the chromosome with the highest fitness value the fitness values for all individuals.
(closest to 1) will be ranked first, and the one with the
lowest fitness value will be ranked last. As shown in Table 12 the best parents from population
 Rank-Based - Selection After sorting, the top-performing A.
chromosomes (those with the highest fitness values) are

IJISRT25FEB677 www.ijisrt.com 703

Volume 10, Issue 2, February – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14930503

Table 8: Best Chromosomes

API1 API2 API3 API4 API5 API6
Scenario 1 2.4 0.7 8 -2 5 1.1
Scenario 2 -0.4 2.7 5 -0.1 7 0.1
Scenario 3 -0.1 2 2 -3 2 0.9
Scenario 4 4 7 12 6.1 1.4 -4
Scenario 5 3.1 4 0 2.4 4.8 0
Scenario 6 -2 3 -7 6 3 3

 Crossover - in this operation will take two parents change is applied randomly to ensure diversity in the
(chromosomes) to generate a new offspring by switching population.
segments of the parent genes. It is more likely that the
new offspring (children) as shown in fig 5 will contain a  Creating Enhanced Solutions:
good part of their parents, and consequently perform The mutation process generates new chromosomes that
better as compared to their ancestors. may provide improved or alternative solutions for the effort
estimation problem. By introducing random changes,
mutation helps the genetic algorithm escape local optima and
explore a broader search space.

 Controlled Mutation Rate:

The mutation rate (probability of mutation) is typically
kept low to avoid disrupting good solutions while still
allowing for exploration. For example, a mutation rate of 1-
Fig 5: Crossover Operation 5% is common.

 The first crossover between two parents (P1, P2) as  Importance of Mutation:
showing in fig 6 to generate new offspring.
 Diversity: Mutation ensures that the population remains
diverse, preventing the algorithm from converging too
quickly to suboptimal solutions.
 Exploration: It allows the algorithm to explore new
regions of the search space that might contain better
solutions.
 Innovation: By introducing random changes, mutation
can lead to innovative solutions that might not emerge
Fig 6: New Generate of Offspring through selection and crossover alone.

Then makes a crossover between other parents (P1, P2)

with the same previous steps.

Mutation is a critical phase in the genetic algorithm, as

it introduces diversity into the population and helps explore
new solutions that might not be reached through selection
and crossover alone. Here's how the mutation process works
in your context:
Fig 7: New Generate of Offspring
 Random Selection of API Factors:
The mutation process involves randomly selecting As shown in Fig 7 the describe the mutation of the
values from the predefined API factors: VSS (Very Small chromosome by replacing the gene (4.8) with the gene (2.4)
Small), SS (Small Small), MS (Medium Small), LS (Large as a random value from story point factors.
Small), and VLS (Very Large Small). These values are used
to modify or switch specific genes within a single Then makes the mutation for all chromosomes that
chromosome. generated form the crossover phase, with the same previous
steps.
 Gene Modification:
During mutation, one or more genes (representing API After complete, the mutation phase will become the
parameters) in a chromosome are altered. For example, a first generation of the Population (A) as show the generation
gene representing API1 might be change from VSS to LS or 1 in Table 13.
any other random value from the available factors. This

IJISRT25FEB677 www.ijisrt.com 704

Volume 10, Issue 2, February – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14930503

Table 9: 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 VII. PROPOSED SOFTWARE
population (A), all the phases of the genetic algorithm will
be re-applied on this generation (Fitness value, Selection, This program was designed using the Agile software
Crossover, Mutation). The algorithm will be stopped when engineering methodology, with the collaboration of the
achieves the goal, after completing the mating pool of all Research and Development department, the program was
chromosomes and completing the mutation phase for all designed by Python code to simulate a generic drug recipe,
genes in the population, that use random values of the story specifying ingredients, quantities, and steps. This code
points factors, To ensure that all point stories are addressed allows you to create or improve a drug recipe, add
to extract effort estimation for only story points Technique. ingredients with quantities, specify preparation steps, and
can change the value of the gene of story point (API) then display the entire recipe.
chromosomes to binary sting if the gene contains a set of
attributes.  Python Software Code
This program was designed using the Agile software
In implementation Life Factors (IAI) technique, that engineering methodology, with the collaboration of the
the second chromosomes of the population (A) and the Research and Develop-ment department, the program was
Population (B), that include two chromosomes (PR, UL and designed by Python code to simulate a generic drug recipe,
FI), will be appley the proposed genetic algorithm phases specifying ingredients, quantities, and steps. This code
(Fitness value, Selection, Crossover, Mutation), with the allows you to create or improve a drug recipe, add
same previous steps in all phases. The purpose of using the ingredients with quantities, specify preparation steps, and
Multi-Genetic Algorithm to accelerate the algorithm, by then display the entire recipe, the Python code used to
dividing each of the two chromosomes into one population. optimize drug formulations based on constraints such as
efficacy, toxicity, and maximum weight. The code relies on
the PuLP library to solve the optimization problem.

 Recipe Code

IJISRT25FEB677 www.ijisrt.com 705

Volume 10, Issue 2, February – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14930503

 Genetic Algorithm Code

VIII. RESULT the program is run, the result is displayed, which includes the
drug prescription, the improved drug components, as well as
The Genetic Algorithm parameters that were selected the preparation steps for the drug listed in the knowledge
in the first test of the population (A), Which underwent a set base as show in Fig-7 and Fig 8
of proposed algorithm operations showed the results when

Fig 8: The Result of Recipe Code

IJISRT25FEB677 www.ijisrt.com 706

Volume 10, Issue 2, February – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14930503

Fig 9: The Result of Genetic Algorithm Code

 Validation and Verification Methodology [2]. S. Bodade, N. Bajad Mam, S. Deshmukh, S,

Mangesh, Artificial Intelligence in Pharmaceutical
 Check Feasibility of Constraints- The constraints should Industry, International Journal of Advanced Research
not conflict, ensuring a feasible solution exists. in Science, Communication and Technology
 Manually verify- The total weight does not exceed (IJARSCT),Volume 3, Issue 3, December 2023.
50.Toxicity remains within limits. Minimum beneficial [3]. P. Sajja and R. Akerkar, Knowledge-Based Systems
compounds are met. for Development, Advanced Knowledge Based
 Run Multiple Test Cases Modify the constraints slightly Systems: Model Applications Research, vol. 1, pp. 1-
and observe whether the model still finds an optimal 11, 2010.
solution. Example: Change up Bound values for [4]. W. Lo and J. Choobineh, Knowledge-Based Systems
ingredients to check edge cases. as Database Design Tools: A Comparative
 Compare with Manual Calculation - Solve the objective Study, Intelligent Support Systems Technology:
function manually for small values and compare results. Knowledge Management, 2002.
 Check Solver Status Ensure LpStatus[status] == [5]. Mak, K.-K.; Pichika, M.R. Artificial Intelligence in
"Optimal", indicating a successful optimization. Drug Development: Present Status and Future
Prospects. Drug Discov. Today, 24, 773–780, 2019.
IX. CONCLUSION [6]. Scannell, J.W.; Blanckley, A.; Boldon, H.;
Warrington, B. Diagnosing the Decline in
This paper aimed to propose to enhance the Pharmaceutical R&D Efficiency. Nat. Rev. Drug
pharmaceutical industry by leveraging artificial intelligence, Discover., 11, 191–200, 2012.
specifically through a multi-gene genetic algorithm, to [7]. Paul, D., Sanap, G., Shenoy, S., Kalyane, D., Kalia,
optimize drug formulation effectiveness. The proposed K. and Tekade, R.K., . Artificial intelligence in drug
model considers critical constraints such as weight, toxicity, discovery and development. Drug discovery today,
and beneficial components. The approach involves 26(1), p.80,2021.
generating a set of candidate solutions, each characterized by [8]. In Bioinformatics Tools for Pharmaceutical Drug
modifiable properties (e.g., parents, chromosomes), these Product Development; John Wiley & Sons, Ltd.:
solutions are encoded in binary or decimal systems as strings Hoboken, NJ, USA,pp. 1–18, 2023.
of 0s and 1s, though other encoding methods are also [9]. Lalitkumar K. Vora , Amol D. Gholap , Keshava
feasible. and implemented this, a Python-based program has Jetha , Raghu Raj Singh Thakur ,Hetvi K. Solanki
been developed to simulate the genetic algorithm, aiming to and Vivek P. Chavda,Artificial Intelligence in
refine drug formulations while adhering to the specified Pharmaceutical Technology and Drug, School of
constraints. Pharmacy, Queen’s University Belfast, 97 Lisburn
Road, Belfast BT9 7BL, UK© 2023 by the
REFERENCES authors.Licensee MDPI.
[10]. Mafarja, M., Mirjalili, S., Aljarah, I., & Al-Zoubi, A.
[1]. K Huanbutta a, K Burapapadh , P Kraisit , P M. Hybridizing Genetic Algorithm with Advanced
Sriamornsak , K Suwanpitak ,Artificial intelligence- Metaheuristics for Optimization Problems. Swarm
driven pharmaceutical industry: A paradigm shift in and Evolutionary Computation, 77, 101273. [DOI:
drug discovery, formulation development, 10.1016/j,2023.
manufacturing, quality control, and post-market [11]. Nguyen, T. T., Nguyen, M. T., & Vo, B. A Review
surveillance,European Journal of Pharmaceutical on Genetic Algorithms and Their Applications in
Sciences, 69:262–277, 2024 Optimization. Artificial Intelligence Review, 56(2),

IJISRT25FEB677 www.ijisrt.com 707

Volume 10, Issue 2, February – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.5281/zenodo.14930503

1157-1190. [DOI: 10.1007/s10462-023-10412- [23]. Verma and Archana, "A Survey on Image Contrast
5],2023. Enhancement Using Genetic Algorithm", International
[12]. Hossam Alansary,” A Proposed Genetic Algorithm Journal of Scientific and Research Publications, vol. 2, no.
Model to Improve Effort Estimation in Agile 7, 1, pp. 2, July 2012.
Software Development” International Journal of [24]. Arkhipov DI, Wu D, Wu T, Regan AC, A parallel genetic
Computer Science and Information Security algorithm framework for transportation planning and
(IJCSIS), Vol. 22, No.4,2024 logistics management, IEEE Access 8:106506–106515,
[13]. Ochoa, G., & Veerapen, N. Genetic Algorithms: 2020.
Advances, Trends, and Future Directions. [25]. Cheng H, Yang S, Multi-population genetic algorithms
Evolutionary Computation, 31(1), 1-27. [DOI: with immigrants scheme for dynamic shortest path routing
10.1162/evco_a_00364] 2023. problems in mobile ad hoc networks. Applications of
[14]. Cai, Z., Sun, Y., & Chen, J. Multi-Objective Genetic evolutionary computation, Springer, In, pp 562–571, 2010.
Algorithms: A Comprehensive Survey. IEEE [26]. Das AK, Pratihar DK, A direction-based exponential
Transactions on Evolutionary Computation, 27(5), mutation operator for real-coded genetic algorithm. IEEE
1054-1075. [DOI: 10.1109/TEVC..3273856], 2023. International Conference on Emerging Applications of
[15]. Abd Elaziz, M., Oliva, D., & Hinojosa, S. Genetic Information Technology, 2018.
Algorithms in Machine Learning: Applications and [27]. Dash SR, Dehuri S, Rayaguru S,Discovering interesting
Challenges. Expert Systems with Applications, 235, rules from biological data using parallel genetic algorithm,
120312. [DOI: 10.1016/j.eswa.120312], 2024. 3rd IEEE International Advance Computing Conference
[16]. Kar, A. K. A Review of Evolutionary Algorithms for (IACC), Ghaziabad,, pp. 631–636, 2013.
Optimization Problems: Trends and Future [28]. de Ocampo ALP, Dadios EP , Energy cost optimization in
Directions." Applied Soft Computing, 139, 110931. irrigation system of smart farm by using genetic algorithm,
[DOI:10.1016/j.asoc..110931], 2023. 2017IEEE 9th international conference on humanoid.
[17]. Mirjalili, S., & Heidari, A. A. Advances in Nanotechnology, Information Technology,
Evolutionary Computation: State-of-the-Art and Communication and Control, Environment and
Emerging Trends. Swarm and Evolutionary Management (HNICEM), Manila, pp 1–7, 2017.
Computation, 80, 101290. [29]. Deb K, Pratap A, Agarwal S, Meyarivan T , A fast and
[DOI:10.1016/j.swevo.101290], 2023. elitist multiobjective genetic algorithm: NSGA-II. IEEE
[18]. Wu, G., & Luo, Q. ,A Survey on Hybrid Evolutionary Trans Evol Comput 6(2):182–197. 2002
Algorithms for Large-Scale Optimization." [30]. Di Fatta G, Hoffmann F, Lo Re G, Urso A , A genetic
Information Sciences, 646, 119204. algorithm for the design of a fuzzy controller for active
[DOI:10.1016/j.ins.2023.119204], 2024. queue management. IEEE Trans Syst Man Cybern Part C
[19]. Hossam Alansary,” A Proposed Genetic Algorithm Appl Rev 33(3):313–324, 2003.
Model to Improve Effort Estimation in Agile [31]. D-Manríquez, R Alvarado, B Zambrano, G Melendez, El
Software Development” International Journal of Leal JC , An automatic document classifier system based
Computer Science and Information Security on genetic algorithm and taxonomy. IEEE Access
(IJCSIS), Vol. 22, No.4,2024 6:21552–21559, 2018.
[20]. Eldos, T. ,A, New Migration Model For Distributed [32]. EkbataniFard GH, Monsefi R, Akbarzadeh-T M-R,
Genetic Algorithms , Proceedings of the International Yaghmaee M et al. A multi-objective genetic algorithm
Conference on Scientific Computing , , (CSC06), (pp. 128- based approach for energy efficient qos-routing in two-
134), 2006. tiered wireless sensor net-works. In: wireless pervasive
[21]. W Z Guo, L J Hao and W J. Di, Application and computing (ISWPC), 2010 5th IEEE international
Development of Artificial Intelligence Technology for the symposium on. IEEE, pp 80–85. 2010.
Data Management and Analysis in Forestry[J], Artificial [33]. Shukla AK, Singh P, Vardhan M , A new hybrid feature
Intelligence Computational Intelligence. Aici. International subset selection framework based on binary genetic
Conferenc, vol. 4, no. 4, pp. 438-441, 2009. algorithm and information theory. International Journal of
[22]. M H Huang and R T. Rust, Artificial Intelligence in Computational Intelligence and Applications
Service, Journal of Service Research, vol. 21, no. 2, pp. 18(3):1950020(1–10), 2019.
155-172, 2018.

AUTHORS’ PROFILES

Hossam Abdelrahman Al-Ansary.Ph.D. degree in an information system with Cairo University. His currently position is IT
Director at Alandalous Pharmaceutical.

IJISRT25FEB677 www.ijisrt.com 708