4-IJCI Vol. 3 No. 3-March 2024-Paper3-Dr. Obaid

Monte Carlo Simulation for Risk Management
in ِِِAgile Software Development

Ali Hamzah Obaid
Al-Furat Al-Awsat Technical University, Babylon Technical Institute
inb.ali210@atu.edu.iq
Khansaa Azeez Obayes Al-Husseini

Al-Furat Al-Awsat Technical University, Babylon Technical Institute
inb.khanssa@atu.edu.iq
Abstract
Agile development methodologies are becoming increasingly popular in agile
development projects due to their iterative nature and adaptability. The Monte Carlo
method is distinguished by its statistical technique of using random samples to obtain
many satisfactory results for solving uncertainty. It can be used in agile software
development by building different scenarios and measuring their impact on the
project results through statistical operations. Its ability to meet significant challenges
in managing risk and uncertainty mitigation.
In this study, we propose a Monte Carlo simulation-based approach to measure and
analyze risks in the agile development process. It provides a probabilistic framework
for risk assessment by integrating the Monte Carlo simulation methodology to model
the various development process variables. It was addressed by identifying the main
risk factors within agile projects, such as tasks, availability of resources, and external
dependencies. The proposed approach contributes to mitigating potential risks in
agile software development using Monte Carlo theory, which provides a systematic
framework that enables work teams and management to overcome risk and
63
International Journal of Computers and Informatics, London Vol (3), No (3), 2024
https://doi.org/10.59992/IJCI.2024.v3n3p3 E-ISSN 2976-9361
uncertainty associated with the agile development process in dynamic project
environments.
Keywords: Monte Carlo, Agile, Software Development, Risk, Simulation.
1. Introduction
Recently, automation of the decision-making process has played an essential role in
the development of intelligent systems. As software development companies face
particular possibilities of risks, the need arises to predict the risks that these companies
face during the development process. In the agile development process, there are
many uncertainties and challenges that stakeholders and project managers may face,
the most important of which are the changing requirements, especially by the end
users of the program. These requirements significantly affect the planned timetable
for completing the project and the material costs allocated for the process.
Development, which in turn negatively affects the quality of the final product [1].
Therefore, we see many development teams tending to reduce risks by using several
methods, and one of these methods is the use of Monte Carlo simulation because of
its remarkable ability and effectiveness in managing risks and uncertainty in software
development projects. This method identifies ways and opportunities to achieve the
plans and goals set for modelling and evaluating alternative scenarios for developing
volatile events and changing requirements during the development period [2].
Monte Carlo simulation is based on statistical calculations that take a set of repeated
random samples to obtain numerical results and apply them to estimate the risks of an
agile development project by building different scenarios and analyzing their potential
effects on the development process results. The Monte Carlo method can be used in
agile development projects to assess risks through:
1- Identifying potential risks due to changing requirements and uncertainties
related to costs or work teams.
64
2- Identifying the main requirements and variables during the development
process that could affect, in one way or another, the quality of the final product.
3- Identify and set frequency distributions for the selected variables.
4- Building a simulation model that includes the introduction of variables and
their probability distributions, which in turn represents the dynamic model of
the project.
5- Create a generator to generate random samples, as the Monte Carlo method
allows the generation of many random samples from the probability
distributions of variables.
6- Evaluating and analyzing the results of the simulation process.
7- Iterate and improve the model and plan to mitigate potential risks.
Monte Carlo simulation is a powerful tool for improving the efficiency of project
management due to its approach based on risk assessment by generating many series
of mathematical statistics and predicting potential risks and its effectiveness in
mitigating those risks because it is of great importance in dealing with uncertainty,
which allows decision makers to have confidence in Make bolder decisions by dealing
with the potential for risk and uncertainty. [3]
Purpose: The purpose of this work is to study the theoretical aspects of using risk
simulation in agile software development using the Monte Carlo method.
Objectives: Identify risk factors, integrate probability distributions, and develop a
Monte Carlo simulation model to calculate risks in agile software development.
2. Risk of Agile Software Development
Risk is an uncertain event that hurts the project's success [4]. Agile development is
considered one of the methodologies used for software management and
development, which focuses on building the program in multiple stages and within
65
specific periods. Each stage produces a program distinguished from the previous one
by several additional tasks, and each stage is considered an actual program [5].
What distinguishes agile software is that it is an iterative software development
process, and each iteration has a specific scope and period. These periods may cause
an increase in the possibility of risk due to unequal priorities for delivering the project
in a short time. Adapting to changing requirements is a feature of agile software, and
this feature makes uncertainties affect the project results, which increases the
probability of risk. Hence, risk management in agile software can include many risk
factors that must be considered, including [6].
1- Teamwork: This leads to the inability to communicate with work teams, which
causes a delay in delivering the project on time.
2- Participation: Decision-makers' participation in the project life cycle is
necessary to ensure their requirements are met. Understanding the requirements
and meeting their needs accurately leads to the risk of the product being
delivered and working efficiently, as the involvement of stakeholders
throughout the project period is essential to meet their needs and feedback
consistent with the project objectives.
3- Final delivery: The deadlines for delivering the product are among the most
essential and significant risks facing the agile software development process
due to the large number of variables.
Software development based on agile methodologies is subject to many potential
risks due to fluctuations in requirements, technical complexities, the possibility of
restrictions on the resources used, and the dynamics of the teams' work. All of these
factors contribute to uncertainty in project implementation management.
From the above, development process management teams must prioritize teamwork
practices and effective communication [7]. This effectiveness and risk reduction can
66
be achieved by holding meetings between development teams and stakeholders.
Participation strategies can help stakeholders mitigate risks and ensure that The
development process is moving toward good delivery. [8]
3. Risk factors within Agile projects
Many software organizations tended to adopt agile methods for software
development, but they noticed failures in the development process that were reported
using the agile approach [9]. Due to the iterative nature of the agile approach, a lack
of requirements and unclear scope are considered to be among its main risk factors.
Therefore, these institutions sought to mitigate the risks associated with agile
development and management of the development process, as it was noted that there
are multiple risks, for example, risks related to the time of the development process,
the time of the delivery process, the financial resources allocated to the development
process, the interaction of work teams, and the dynamics of communication between
them or with Stakeholders. Here, we explain a group of essential factors that may
affect the development process of agile methods.
1- Risks related to stakeholders: Through the cooperation of work teams with project
owners or end users in conducting interviews or workshops and listening to the
project requirements.
2- Risks associated with time management: One of the most critical areas in the
development process is the process of delivering the product on time and
evaluating work levels and their impact on the various performance measures of
the work teams by running the simulation process with different requirements
scenarios and time limits. Stakeholders can obtain A clear vision of the potential
risks in time management. This allows them to make better decisions in
determining the time required to complete the task by the work teams to balance
the material costs and the period required to complete the task.
67
3- Risks associated with changing requirements: By considering the uncertainty of
requirements for stakeholders or end users, Monte Carlo simulation is used to
improve the process of predicting changing requirements during the development
process, where stakeholders can evaluate the likely results with work teams and
develop more robust and realistic forecasting models. This allows for better
forecasting of requirements and improves planning for managing the
development process and allocating the necessary resources.
By considering these factors that may affect the development process, agile
development teams can build a clear and comprehensive understanding of the
potential risks and predict them, which enables them to address these risks and
challenges that they may face early and proactively and mitigate them during the
development process period, [10]as The iterative nature of the agile approach allows
for continuous adaptation and improvement of its vision and strategy in challenging
potential risks.
4. Risk calculation simulation model
The Monte Carlo method for any project's success depends on the input data's
accuracy and the simulation model's rigor, precision, and realism. To create any
model that simulates risks, we take into account cases of uncertainty and variation in
project requirements, and this is what the Monte Carlo method provides us with to
calculate software development risks. Figure (1) shows a guide to applying the
Monte Carlo method to analyze software development risks.
Building a simulation model first requires identifying the variable requirements in
the development process that contribute to the success of any project and that may
be subject to uncertainty, such as cost variables, lack of resources, changing teams,
etc.). [11] Probability distributions are then determined to estimate the probability of
different outcomes for each variable stage. To build a model that represents the flow
68
of project tasks for the development process using Monte Carlo theory, Triple
statistical data are used based on previously recorded data or expert judgment. This
model simulates the development process, which depends mainly on the specified
variables and their probability distributions. It is characterized by generating random
samples of probability distributions from the variables specified during the
simulation and repeating them. For each simulation process, the results are evaluated
based on random samples, which may include evaluating the project period, cost, or
scope... When looking at the distributions of project results, the results are analyzed
from variables that are characterized by uncertainty, such as standard deviation
calculations and percentages; through the results obtained in the analysis process,
management efforts focus on reducing risks and uncertainty, and high-risk areas are
identified through variation in results. Depending on the results obtained, improving
the simulation model by adjusting the probability distributions or incorporating new
requirements to reach the best results is possible. Communication and iteration in
agile software development allow project teams to participate with stakeholders in
discussing the results obtained from the simulation model built by the Monte Carlo
method. It facilitates reducing risks through the project's progress and repeating the
model to make the best decisions [12].
To enhance the accuracy of risk assessment and reduce the possibility of uncertainty,
the simulation model is constantly updated with each development in terms of
requirements or variables in the life of the development process.
69
Figure [1]: Monte Carlo method to analyze software development risks
Risk Distribution according to the Monte Carlo Model
To distribute risk calculations using Monte Carlo theory in agile software
development, a hypothetical table that includes multiple iterations and focuses on
task duration variables using the triangular distribution can be built. See table (1)
Since the task duration includes (optimistic, most likely, and pessimistic), these
estimates include the minimum, maximum, and most likely time each task may take.
Random value: It is calculated by a random number generator.
Project results: represents the sum of all task durations for each iteration.
After applying the simulation for sufficient iterations, the results can be analyzed to
understand the distribution of potential project outcomes and calculate statistical
metrics.
70
Table (1): Calculation of risks in agile software development according to Monte Carlo theory
5. Uncertainty Model
In any Monte Carlo simulation process, it is necessary to develop a model of
uncertainty that is related to new requirements or variables that arise during the
period of the development process that contribute to uncertainty, such as the
availability of resources, project scope variables, or any other process related to the
development process. [13]
Through Monte Carlo simulation, random samples are selected from the probability
distribution within a specific range of possible, most likely, and pessimistic
estimates. To calculate the project results, these samples are taken, which may
include collecting periods of individual or team tasks, and these results are recorded
to complete the process of analyzing the results.
After completing the simulation.
1- Mean (Average)
71
Sum of all outcomes
Mean =
Number of iteration
2- Standard Deviation
∑(Squared differences from the mean
Standard Deviation = √
Number of iterations
3- Percentiles
Examine percentiles (eg,10th,50th ...) to understand the range of potential outcomes.
The various project variables (task durations, availability of resources, speed of
work) or any other factors related to the development process that contribute to
uncertainty are calculated, and the probability distributions for each specific variable
are determined according to mathematical equations to model uncertainty solutions
according to the following:
- Triangular Distribution
Where a,b, and c are the minimum, maximum, and most likely values .
- Normal Distribution
(×−𝜇)2
−
P(×)= 𝜎√2𝜋𝑒 1
2𝜎2
Where µ is the mean and 𝜎 is the standard deviation

- Uniform Distribution
1
P(x)=
𝑏−𝑎
72
6. Monte Carlo Simulation Model for Risk Calculation in Agile
Software Development
The Monte Carlo method has gained significant traction in evaluating software
development risks because of its importance in managing and mitigating challenges
effectively. There are several steps to follow a Monte Carlo simulation loop to
evaluate risks in agile software development. In the beginning, variables and their
probability distributions, representing aspects of uncertainty, must be identified.
Certainty, random samples are generated for each variable based on probability
distributions. The results are calculated based on the random samples entered into
the simulation process and the generation and calculation process is repeated several
times to reach the best results.
By generating random samples, we assume that we have variables (X1,
PO= X1+X2+X3+……... +Xn.
Designing accurate and comprehensive mathematical equations to calculate risks in
agile software development is difficult due to the dynamic and iterative nature that
characterizes agile development methodologies. However, we provide a
representation of certain aspects of the process in the form of mathematical
equations:
73
Triangular distribution of task duration
Triangular distribution of task duration:
Where: a is the optimistic estimate, b is the most likely estimate, and c is the
pessimistic estimate.
Expected Task Duration:
𝑎+4𝑏+𝑐
Expected Duration=
6
𝑐−𝑎
Standard Deviations =
6
Project Outcomei= ∑𝑛𝑗=1 𝑇𝑟𝑖𝑎𝑛𝑔𝑢𝑙𝑎𝑟𝐷𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛(𝑎𝑗 , 𝑏𝑗 , 𝑐𝑗

Where: i is the iteration index, n is the number of project variables, and aj,bj,cj are
the parameters of the triangular distribution for the j-th project variable.
Mean project Outcome= 𝑁1 ∑𝑁
𝑖=1 𝑃𝑟𝑜𝑗𝑒𝑐𝑡 𝑂𝑢𝑡𝑐𝑜𝑚𝑒𝑖
Standard Deviations of Project Outcome =

Where:
N= It is the number of simulation iterations.
74
These equations provide a simplified representation of some of the nuances in the
overall approach to risk calculation. In practice, performing Monte Carlo simulations
for agile software development often involves detailed modeling, coding, and
simulation techniques tailored to the specific context of a software development
project, which may require context-specific conditions such as probability
distributions and different modeling techniques.
7. Proposed Framework
Software development faces many risks, which result from unexpected events or
circumstances that may positively or negatively impact the development process.
Therefore, when starting any project, it is necessary to develop scenarios to manage
potential risks [14]. In the form of (2), we put together a proposal for managing a set
of scenarios for facing risks in agile software development. They are distributed
according to specific strategies to mitigate risks, including strategies for adapting to
changing requirements and regular review of all requirements and modifications
required by stakeholders through communication and agreement to develop plans. A
well-studied scenario to avoid the risk. A scenario for developing a prioritization of
risks based on the size of the impact of these risks and the probability of their
occurrence, recording the potential risks, and classifying them according to their
priority based on the degree of their impact. A scenario for knowing the risks based
on determining the type of potential risk, studying the history of these risks, and
determining their levels of importance. And the possibility of managing to avoid
these risks. Risk analysis scenarios are used to evaluate situations of risk and
uncertainty to reduce their potential impact on the development process and analyze
the dependency of these risks. Feedback scenarios through an organized mechanism
to review the opinions of users and stakeholders, monitor them continuously, and
adapt them to the requirements of agile methodologies. Documentation and
75
preparation of reports on potential risks and recording them. Training and awareness
of work teams on the importance of mitigating risks in the development process and
verifying them. Improvement process based on user feedback perspectives, and
changeable requirements.
Figer [2]: Proposed framework for risk assessment

8. Conclusion
In the agile development process, Monte Carlo simulation provides a systematic
approach to assessing and managing risks in managing the chain of changing
requirements during the development process by considering uncertainties in
requirements management and predicting variables during the agile development
cycle, allowing stakeholders to make crucial decisions and develop their outlook and
its strategy to improve the efficiency of the development process. The Monte Carlo
76
method and its use in developing agile software projects also provide its efficiency
in managing changing and recurring requirements, assessing the risks associated with
the physical aspects, distributing the tasks of work teams, and predicting the
requirements and goals set for the development process to reach the best final
product.
Using the Monte Carlo method in the agile development approach helps better
understand the uncertainties associated with the development process and mitigate
risks very effectively if new data and information are available while updating
ongoing requirements as the work progresses.
References
1- Tawunwoot, W., & Chomngern, T. (2023, November). Prototype Structure Design Process
Using Information Flow Diagram (IFD) for Mobile Application Learning System. In 2023 7th
International Conference on Information Technology (InCIT) (pp. 492-497). IEEE.
2- Magennis, T. (2012). Managing software development risk using modeling and Monte Carlo
simulation. In Lean Software and Systems Consortium Conference, Boston.
3- Mahadevan, S. (1997). Monte Carlo simulation. Mechanical Engineering-New York and Basel-
Marcel Dekker-, 123-146.
4- Marle, F. (2020). An assistance to project risk management based on complex systems theory
and agile project management. Complexity, 2020, 1-20.
5- Williams, L. (2010). Agile software development methodologies and practices. In Advances in
computers (Vol. 80, pp. 1-44). Elsevier.
6- Al-Husseini, K. A. O., & Obaid, A. H. (2020). Interaction between project tasks and risk
management tasks in software development. Periodicals of Engineering and Natural Sciences, 8
(4), 2300-2308.
7- Linderman, K., Schroeder, R. G., & Sanders, J. (2010). A knowledge framework underlying
process management. Decision Sciences, 41(4), 689-719.
8- Al-Husseini, K. A., & Obaid, A. H. (2019). Analysis and risk management in software
development using the logical-algebraic model. In CEUR Workshop Proceedings (Vol. 2475,
pp. 241-248).
77
9- Shrivastava, S. V., & Rathod, U. (2017). A risk management framework for distributed agile
projects. Information and software technology, 85, 1-15.
10- Tam, C., da Costa Moura, E. J., Oliveira, T., & Varajão, J. (2020). The factors influencing the
success of on-going agile software development projects. International Journal of Project
Management, 38(3), 165-176.
11- Jonkers, R. K., & Shahroudi, K. E. (2020). A design change, knowledge, and project
management flight simulator for product and project success. IEEE Systems Journal, 15(1),
1130-1139.
12- Mun, J. (2006). Modeling risk: Applying Monte Carlo simulation, real options analysis,
forecasting, and optimization techniques (Vol. 347). John Wiley & Sons.
13- Arunraj, N. S., Mandal, S., & Maiti, J. (2013). Modeling uncertainty in risk assessment: An
integrated approach with fuzzy set theory and Monte Carlo simulation. Accident Analysis &
Prevention, 55, 242-255.
14- Al-Husseini, K. A. O., & Obaid, A. H. (2018). Development of Risk Management Tools in
Question-Answering Based Software Design Environment. International Journal of Computer
Science and Mobile Computing-IJCSMC, 7(6), 165-174.
78

4-IJCI Vol. 3 No. 3-March 2024-Paper3-Dr. Obaid

Uploaded by

Copyright:

Available Formats

4-IJCI Vol. 3 No. 3-March 2024-Paper3-Dr. Obaid

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

4-IJCI Vol. 3 No. 3-March 2024-Paper3-Dr. Obaid

Uploaded by

Copyright:

Available Formats

Monte Carlo Simulation for Risk Management

in ِِِAgile Software Development

Khansaa Azeez Obayes Al-Husseini

Where µ is the mean and 𝜎 is the standard deviation

Project Outcomei= ∑𝑛𝑗=1 𝑇𝑟𝑖𝑎𝑛𝑔𝑢𝑙𝑎𝑟𝐷𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛(𝑎𝑗 , 𝑏𝑗 , 𝑐𝑗

Standard Deviations of Project Outcome =

Figer [2]: Proposed framework for risk assessment

You might also like