The Automation of Critical Path Method using Machine Learning: A Conceptual Study

International Journal of Engineering and Management Research e-ISSN: 2250-0758 | p-ISSN: 2394-6962
Volume-11, Issue-3 (June 2021)
www.ijemr.net https://doi.org/10.31033/ijemr.11.3.38
237 This Work is under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
The Automation of Critical Path Method using Machine Learning: A
Conceptual Study
Othman Aljumaili
Student, Department of Engineering Management, Istanbul Gedik University, TURKEY
Corresponding Author: othman.adnan.ibrahim@gmail.com
ABSTRACT
This research aims to shed light on the use of
machine learning in improving, developing and automating
the critical path method, solving its problems, studying this
effect and its dimensions, and discussing that from many
aspects.
The research is divided into two theoretical and
practical parts. The theoretical part is concerned with
studying the critical path method and its advantages,
problems and challenges, as well as studying machine
learning and artificial intelligence and its dimensions,
reviewing materials and sources related to this, and then
presenting suggestions and future solutions based on this
study. As for the practical section, it is a questionnaire that
targeted a segment of engineers, in particular, and others
who have sufficient experience in both the critical path
method and machine learning, and seeking their opinions on
both topics.
The result of the theoretical research was 14
theories or proposals that were presented based on the
foregoing study. As for the practical questionnaire, a sample
of 127 was taken. Through statistical analysis, the results
were analyzed and discussed separately, and then a
conclusion was drawn regarding them.
Keywords— Critical Path Method, Machine Learning,
Artificial Intelligence
I. INTRODUCTION
Before starting to discuss the Critical path
method, it is important to understand the way project
management process, like planning, scheduling,
monitoring, and controlling along with corrective actions.
The main goals of any project are to finish a pre-known
amount of work under a fixed duration at a pre-estimated
cost to the desired quality. To attain these goals, project
planning, scheduling, monitoring, and controlling are
important. The three stages: planning, scheduling,
monitoring, and controlling are a cycle process [1]. There
are different project management methods and tools, that
can be used for every one of these stages according to the
type and the size of the project. For the planning stage,
there is a significant number of methods such as Critical
path Method (CPM), Bar Charts, Program Evaluation and
Review Techniques (PERT), Linear Scheduling Method
(LSM), Line of Balance (LOB), Work Study technique and
operation techniques [1]. The Critical Path Method is a
powerful tool for planning and management of all sorts of
projects. It offers an accurate mathematical policy for
planning, scheduling and control and permits evaluation
and comparison of substitute work programs, construction
methods and kinds of equipment through changing activity
durations individually, resources or bonds between
activities [1]. CPM is a quite important and effective tool
in the project management and hence, should be included
in the project management software packages. There are
different project management techniques based on critical
path method. They are needed at different stages of a
project. Despite the fact that machine learning is a branch
from computer science, it varies from classical
computational methods. In classical computing, algorithms
are collection of frankly codded directions employed by
computers for calculating or solving a problem. Machine
learning algorithms on the other hand, let computers train
using data inputs then analyze them statistically in order to
extract output values that are falling in a particular range.
Because of this, machine learning facilitates computers in
building models from sample data in order to automate
decision-making processes based on data inputs. Machine
learning is a continuously developing field. Because of
this, there are some considerations to be kept in mind as
working with machine learning methodologies, or analyze
the impact of machine learning processes [2]. Investing in
this field to solve different problems and challenges is the
new trend for researchers and scientists around the world.
II. METHODOLOGY
The goal of this research is to make the first step
and the first research that focuses the lights on the
investment in the capabilities and enhancements that ML
can contribute and make in the improvement journey of
CPM, and this was done by dividing the research into two
parts:
 The first part is about collecting information
about the CPM through previous articles and
researches and also collect information about ML
and its capabilities and branches and try to make a

combination of the two and summarize that into
practical and solid ideas and thoughts that can
possible be implemented in further researches and
discuss each of them separately.
 The second part is a survey that is made for a
wide slice of engineers collecting information
about several aspects:
 Engineers’ awareness of ML and its abilities.
 The aspects that need to be reconsidered in
developing CPM.
 Automation of CPM and what are the aspects
that are most needed to be automated than
others.
 Is the automation in CPM really needed?
 What other improvements that ML can
contribute in developing CPM?
The results after that would be analyzed and
discussed one by one.
III. CRITICAL PATH METHOD
The Critical Path Method (CPM) which was
introduced in 1956 has proven that it is an effective useful
tool for planning and controlling construction projects.
critical path method enables project managers to evaluate
the early and late starting and ending time of each activity,
and then being able to calculate the float (slack) time of the
activities if available, to define critical activities, and to
evaluate the impact of changing the durations and the
logical relations on the overall duration. the use of the
critical path method (CPM) in all industries including
construction in the last three decades has dramatically
increased because of its benefits and the huge
advancement that has been made in computer hardware
and scheduling software [3]. in construction projects, CPM
is a very important method because it helps the contractor
to determine and calculate when and how many resources
needed, also enables the vendors to estimate when they can
deliver materials, and the subcontractors to determine
when exactly they can start their work. though, the critical
path method has a lot of serious limitations that haven't yet
been overcome. The computing efficiency and the
analytical capabilities of the critical path method also need
to be enhanced in order to meet the changing requirements
of the construction industry [4]. Construction involves
unique environments, challenges, and project management
needs, which are not found in other industries. although the
industry includes many large companies, statistics indicate
that more than two-thirds of construction companies have
less than five employees [5]. The majority of these small
companies are specialist subcontractors working with a
general contractor. This type of companies experiences the
highest number of failures, as reported in a survey [6]. The
survey shown the factors that contribute in failure, such as
insufficient cash flow, underbidding, lack of experience in
estimating and monitoring costs, and external difficulties.
These factors, indicate a lack of efficient project
management, which is in part due to the drawbacks
associated with critical path method, particularly the lack
of direct mathematical formulation for satisfying project
constraints such as resource limits and deadlines. Despite
the many practical insights provided by commercial
software and professional organizations, for many
construction professionals, particularly small contractors
and trades, the use of critical path method and project
management tools does not go beyond creating a schedule
with a neat appearance in order to satisfy contract
requirements [7].
The critical path method (CPM) was developed in
the late Fifties by researchers at the E. I. Du Pont de
Nemours Company. When firstly developed, the
traditional form of Critical path method networks was
named as AOA or "activity on arrow" diagram, which
allows only Finish-to-Start relationship among the
activities. This means that activities cannot overlap and
that all preceding activities must be finished before a
current activity can start. With the start of the Precedence
Diagram Method (PDM), more flexibility in activity
relationships has been added while the schedule
calculations still use CPM analysis. In precedence
networks, an activity can be connected from either its start
or its finish, which besides the traditional Finish-to-Start
relationship, allows the use of three additional
relationships between project activities: Start-to-Start,
Finish-to-Finish, and Start-to-Finish. Another
characteristic of Precedence Diagram Method is that the
periods of time that can be assigned between the start
and(or) finish of one activity and the start and (or) finish of
a succeeding activity. These periods of time between the
activities are called leads and lags. A lead is the amount of
time by which an activity precedes the start of its
successor(s), while a lag is the amount of delay between
the completion of one task and the start of its successor.
Most of commercial software, like Primavera, Project
Planner and Microsoft Project permit the use of non-
traditional relationships with lags. Several surveys showed
that over the years, CPM use has been increasing in the
construction industry. [8] analyzed data from three surveys
done in 1974, 1990, and 2003 which investigated whether
Engineering News Record’s (ENR) top 400 contractors are
using CPM. The study revealed an increasing CPM use
that reached 98% of the respondents in 2003. By surveying
a mix of both large and small contractors, a recent survey
by [9] shown that it is not only the large ENR 400 firms
who use CPM to manage their projects, but also small and
mid-size construction companies. All the respondents
Saied that they used CPM scheduling some of the time at
least, with 45% reported that they used it all of the time

and another 40% reported that they used CPM most of the
time. The primary uses of CPM were reported as planning
(before construction), control (during construction), and
claim analysis. The disadvantages of CPM were also
reported as implementation requiring excessive work, logic
abuse, too much specialist dependent, and lack of
responsiveness to the needs of field personnel. These
findings agree with the results of a survey [10]: CPM did
not gain the trust of the construction industry as a project
control tool. This is true because in spite of the reports by
contractors that they used CPM for project control, as in
[8] survey. They may find it useful for analyzing progress
status and updating activity data but not as beneficial
method in supporting other important aspects, such as
recovering execution problems and corrective actions.
IV. MACHINE LEARNING
Since their creation, human beings have been
using many different types of tools to accomplish various
tasks. The creativity of human’s brain led to the invention
of different machines. Those machines made the human
life much easier by enabling people to meet different life
needs, including constructions, travelling, computing and
industries. in spite of the rapid developments in the
machine industry, intelligence has remained the essential
difference between humans and machines in
accomplishing their tasks. human use his senses to collect
information from the surrounding environment; the human
brain works on analyzing that information and takes
suitable decisions based on that information. Machines, on
the other hand, are not intelligent by nature. A machine
does not have the ability for analyzing data and taking
decisions. For example, a machine is not expected to
understand the story of Sinbad, fall in love, or contact with
other machines through a common language. The era of
intelligent machines began in the mid-twentieth century
when Alan Turing thought about the possibility thinking
ability for machine. Since then, the artificial intelligence
(AI) branch of computer science has developed fast.
Humans have had the aspirations for creating machines
that have the same level of intelligence as humans. Many
science-fiction movies have expressed these dreams, such
as Artificial Intelligence; Her; I, Robot; The Terminator;
and Star Wars. The history of Artificial intelligence started
in 1943 when Waren McCulloch and Walter Pitts invented
the first neural network model [11]. in 1950, Alan Turing
introduced the next noticeable work in the development of
the AI when he asked his famous question [12]: ―can
machines think?‖ He introduced the B-type neural
networks and the concept of test of intelligence. In 1955,
Oliver Selfridge suggested the use of computers for pattern
recognition [13]. In 1956, Marvin Minsky, John McCarthy,
Nathan Rochester of IBM, with Claude Shannon organized
the first summer AI conference at Dartmouth College,
United States [14]. In the second conference, the term
―artificial intelligence‖ was used for the first time.
Cognitive science term was originated in 1956, during a
symposium about information science at the MIT [15]. in
1957, Rosenblatt invented the first perceptron. After that in
1959, John McCarthy invented the LISP programming
language [16]. In 1962, David Hubel and Torsten Wiesel
proposed using neural networks for the computer vision.
Joseph Weizenbaum developed ―Eliza‖ the first expert
system that could diagnose a disease through its symptoms
[17]. The National Research Council (NRC) of USA
created the Automatic Language Processing Advisory
Committee (ALPAC) in 1964 to expand the research in the
natural language processing. But after many years, the
research was terminated because of the low progress and
high expenses. The book Perceptrons written by Marvin
Minsky and Seymour Papert was published in 1969 [18],
which was talking about the limitations of neural networks
Resulted in stopping organizations’ funding for research
on neural networks. The period from 1969 to 1979
witnessed an increase in the research of knowledge-based
systems. The programs Dendral and Mycin are examples
of this research. In 1979, Paul J. Werbos suggested the first
effective model for neural network with backpropagation
[19]. In 1986 Geoffrey Hinton, Ronald Williams, and
David Rumelhart discovered a method that allows a
network to learn how to distinguish between nonlinear
separable classes, and they gave it the name
backpropagation [20]. In 1986, Terrence Sejnowski and
Charles Rosenberg developed a speech recognition
artificial neural network called ―NETTalk‖. In 1987,
Arthur W. Burks and John H. Holland created an adapted
computing system that has the learning capability. In 1989,
Dean Pomerleau proposed a three-layered neural network
named ALVINN (autonomous land vehicle in a neural
network), designed for road following. In the year 1997,
Garry Kasparov the world chess champion was defeated
by the Deep Blue chess machine which was designed by
IBM. In 2011, a computer developed by IBM named
Watson defeated Ken Jennings and Brad Rutter, the
champions of the television game show Jeopardy!
The period from 1997 till nowadays witnessed and
witnessing rapid developments in natural language
processing, reinforcement learning, computer vision,
computer hearing, emotional understanding, image
processing, pattern recognition, cognitive computing,
knowledge representation, and so on. These trends aim to
provide machines that have the ability of gathering data
through senses similar to the human senses and then
processing them using machine learning methods and the
computational intelligence tools to make predictions and
decisions at the same level as humans. The term machine
learning means enabling machines to learn without

programming them exactly. Generally, there are six
machine learning methods:
 Supervised learning.
 Unsupervised learning.
 Semi-supervised learning.
 Reinforcement learning.
 Deep learning.
 Deep reinforced learning.
V. CONCLUSION
After a detailed study of both the critical path
method and the machine learning technique, and a review
of their sources, previous studies and expert opinions,
some theoretical proposals that could be achieved in the
future were made regarding the use of machine learning to
improve and develop the critical path method based on the
capabilities that this technology possesses, which are as
follows:
1. Using machine learning for the auto-estimation of
projects period.
2. Using machine learning for the auto-suggestion of any
substitutes.
3. Using machine learning for the auto-detection of the
critical path.
4. Using machine learning for the auto-suggestion for
different scenarios.
5. Using machine learning to auto-suggest different and
best crashes or fast tracking have to be made.
6. Solve resources limitation by processing both
resources and project duration and tasks at the same
time.
7. Use the clustering technique for the auto-sorting of
tasks, after that decision tree can be helpful for the
arrangement of the tasks based on their features.
8. Feeding deep learning with many different
modifications suggested for the improvement of CPM
and try to conclude the best model of them:
theoretically this can be efficient method.
9. Using machine learning for the auto weather-
prediction and making pre-alarms.
10. Auto-check and alarm in case any problem is made
while programing the CPM.
11. Logical abuse solving.
12. Reduce specialist’s dependency.
13. Simplifying implementing CPM.
14. Restricting projects deadlines.
REFERENCES
[1] Jayawardena A. K. W. (2012). Mitigating delays in
donor funded projects in Sri Lanka. Engineer-Journal of
Institution of Engineers Sri Lanka, XXXV(01), 65-75.
[2] Lisa, T. (2017). An introduction to machine learning.
Digital Ocean Community, 1-3.
[3] Liberatore, M. J., Pollack-Johnson, B., & Smith, C. A.
(2001). Project management in construction: Software use
and research directions. Journal of Construction
Engineering and Management, ASCE, 127, 101–107.
[4] Ahuja, V. & Thiruvengadam, V. (2004). Project
scheduling and monitoring: current research status.
Construction Innovation, 4, 19-31.
[5] Halphin, D.W. & Woodhead, R.W. (1998).
Construction management. (2nd
ed.). New York: John
Wiley & Sons.
[6] Russell, J. S. & Radtke, M. W. (1991). Subcontractor
failure. Case history. AACE International Transactions,
E.02.1-E.02.6.
[7] Baweja, S. S. (2006). CPM schedules -Why and how.
AACE International Transactions, PS.22.1-PS.22.5.
[8] Kelleher, A. (2004). An investigation of the expanding
role of the critical path method by ENR’s top 400
contractors. Master’s thesis, Faculty of Virginia
Polytechnic Institute and State Univ., Blacksburg, VA.
[9] Hawkins, C. V. (2007). Assessing CPM scheduling
software for the small to mid-size construction firm.
Master’s thesis, Faculty of the Graduate School of the
University of Maryland, College Park, Maryland.
[10] Galloway, P. D. (2006). Survey of the construction
industry relative to the use of CPM scheduling for
construction projects. Journal of Construction Engineering
and Management, ASCE, 132, 697–711.
[11] Warren S. McCulloch & Walter Pitts. (1943). A
logical calculus of the ideas immanent in nervous activity.
Bulletin of mathematical biophysics, 5, 115–133.
[12] A. M. Turing. (1950). I.—Computing machinery and
intelligence. Mind, 59(236), 433–460.
[13] Oliver S. & Marvin M. (1955). Self-Pattern
recognition and modern computers. In: Western Joint
Computer Conference, pp. 91–93.
[14] Marvin M., John McCarthy, Nathan R., & Claude S.
(1955). A proposal for the dartmouth summer research
project on artificial intelligence.
[15] Marvin Minsky & Seymour Papert. (1972).
Perceptrons: An introduction to computational geometry.
(2nd
ed.). Cambridge MA: The MIT Press,
[16] McCarthy, John. (1959). Letter to the editor.
Communications of the ACM, 2(8).
[17] Weizenbaum, Joseph. (1976). Computer power and
human reason: from judgment to calculation. W. H.
Freeman and Company.
[18] Paul J. Werbos. (1979). Neural networks for control.
[19] David E. Rumelhart, Geoffrey E. Hinton, & Ronald J.
Williams. (1986). Learning representations by back-
propagating errors. Available at:
https://www.nature.com/articles/323533a0.

The Automation of Critical Path Method using Machine Learning: A Conceptual Study

More Related Content

Similar to The Automation of Critical Path Method using Machine Learning: A Conceptual Study (20)

More from Dr. Amarjeet Singh (20)

Recently uploaded (20)

The Automation of Critical Path Method using Machine Learning: A Conceptual Study