EXPLORING CREW BEHAVIOR DURING UNCERTAIN JOBSITE
CONDITIONS
By
Ankur Paresh Desai

A THESIS
Submitted to
Michigan State University
In partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Construction Management
2012

ABSTRACT

EXPLORING BEHAVIOR DURING UNCERTAIN JOBSITE
CONDITIONS
By
Ankur Paresh Desai
Construction projects are both time sensitive and dynamic in nature. Unexpected events pose a
great threat to construction projects and their successful completion. The traditional method of
tackling unknown unknowns in construction is based on a centrally controlling body - often the
superintendent of the project. This research explores the nature of these unknown events and the
crew behavior under these trying conditions. The study uses literature and an agent based model
to achieve these objectives. The study also explores the effect of having autonomous and selfmanaged crews in making decisions regarding equipment breakdowns on construction sites.
Modeling showed that the decision making by the self-managed team outperformed the team that
depended on external agents for making decisions. The delay caused by the same damage was on
average 40% lower for the crew making its own decisions compared to the crew that depended
on the superintendent. Considering that the model accounted for erroneous decisions made by
the crew, the prima facie result shows that allowing crews to be autonomous is an effective
strategy on the long run. As a result of the crews solving problems, the superintendent is also
expected to have more opportunities to concentrate on improving the coordination and planning
of work on site. It is also expected that conditions where crews are given autonomy to make
decision will result in a healthier and more satisfying experience for the workers and the
organization.

Dedicated to
My Father, Late Mr. Paresh D. Desai
My Mother, Mrs. Chanchal P. Desai and
My Sister, Ms. Anu P. Desai

iii

ACKNOWLEDGEMENTS
The last 2 years and 4 months at the Michigan State University have been the most enriching and
exciting years of my life yet. This is largely due to the people I have met and had the pleasure of
working and spending time with. I am extremely grateful to the Michigan State University, its
professors, my colleagues and friends for my time here and the opportunities I was presented
with. I hope I can come back in the future and contribute to the university in some form.

To my mentor, Dr. Tariq S. Abdelhamid, who provided me with unending support both in
academic issues and otherwise. His patience, knowledge and passion will always be an
inspiration for me. His expertise in the field of Lean Construction and love for teaching and
research never fails to amaze. I thank him for all the lessons in the classroom and also the lessons
in life. My association with Dr. Abdelhamid is one of the most important things I have gained
during my stay at the Michigan State University.

To my committee member, Prof. Timothy Mrozowski, I thank him for the continuous support and
encouragement during the thesis and the other projects I had the pleasure of working with you
on.

To my committee member, Dr. Mohamed El-Gafy, I thank him for the teachings in class and the
support on various issues on the thesis. Discussions with Dr. El-Gafy gave me confidence in the
direction I was taking on the research.

iv

To my committee member, Dr. Arika Ligmann-Zielinska, I thank you for the teachings in agent
based modeling. The class in spring of 2010 went a long way in encouraging me to pursue this
thesis topic.

To my family, Mrs. Chanchal P. Desai and Ms. Anu Desai, who have always trusted me in my
decisions and supported me in every step some more crazy than others. Having their support
allowed me to expand my horizons and achieve ever higher targets. My mother’s strength and
trust in me is the motivation I needed to complete this thesis.

To my dear friends, Shreyas Shetty, Josh Vertalka, Ramandeep Kaur, Andrew LeMarbe, TR
Ramkumar, Fedrico Steinvorth, Ryan Wang, Dalvir Singh Suri, and Naveen Nerwal. The
conversation I had with them about my research helped me refine the study and gave me a
different perspective on the same.

To my classmates and colleagues, Samarth Jain, Sagata Bhawani, Don Schafer, Stanley Samuels,
Eric, Wenda Nofera, Krupesh Kakkente, Raina, Pradnya Mahajan, Stephanie and Aditya Singh,
who I shared numerous learning moments with, I would like to thank them for sharing their
knowledge and their company.

In my time at Michigan State University, I was extremely fortunate to be presented with exciting
and rewarding employment opportunities.

To my first employers at the University, Jack Mumma, Sherie Shorman and Jannene Andrews, at
the Campus Planning and Administration (CPA) department of Michigan State University.
Working at CPA was a pleasure and often a blast! Conversations with Jack are still fresh in my
memory and I am grateful to him for showing me the right direction on many issues.
v

To my first employers in the construction industry in the US, Granger Construction Company,
who believed in me and gave me an opportunity to apply my learnings in the field. I am forever
grateful to Darryl Massa, Dennis Carignan, Ian Clutten, Mike Nowosad, Glenn Simon, Dave
McKelvey, Barry Roney, Glenn Granger, Al Granger, Hiedi Klum and David Olson. I am
grateful for the opportunities and sincerely thank them. Working at Granger was a steep learning
curve and it gave me immense confidence to be associated with the aforementioned persons.

My future employers, Enovio LLC., and in particular Ihab Ismail, who have stuck with me in
spite of my thesis taking longer than anticipated! Ihab is an inspiration to me in terms of
maintaining a work ethic and commitment to work. I hope to learn many more things in the
future at Enovio and grow along with the company.

There are many people who have played an important role in my stay at Michigan State
University and I would like to thank all my friends, collegues and mentors for making this time
unforgettable!

GO GREEN!

vi

TABLE OF CONTENTS
LIST OF TABLES .................................................................................................................... ix
LIST OF FIGURES ....................................................................................................................x
CHAPTER 1: INTRODUCTION ................................................................................................1
A) Motivation ...........................................................................................................................2
B) Problem Area .......................................................................................................................3
C) Goals and Objectives ...........................................................................................................5
D) Research Scope ....................................................................................................................6
E) Layout of the Thesis .............................................................................................................7
CHAPTER 2: LITERATURE REVIEW .....................................................................................9
A) Introduction ....................................................................................................................... 10
B) Levels of Uncertainty ......................................................................................................... 14
C) Introduction ....................................................................................................................... 18
D) Background on Unexpected Events .................................................................................... 19
E) Methods Suggested in Literature to Mitigate Unexpected Events ....................................... 21
F) Lean Construction .............................................................................................................. 36
H) Autonomy in an Organization ............................................................................................ 59
I)

Significance of the Research .............................................................................................. 63

J)

Agent Based Modeling ....................................................................................................... 64

CHAPTER 3: RESEARCH METHOD...................................................................................... 67
A) Introduction ....................................................................................................................... 68
B) Model Development ........................................................................................................... 72
C) Initial Model Setup ............................................................................................................ 74
D) Model Execution ................................................................................................................ 75
E) Model Output ..................................................................................................................... 76
F) Model Parameters .............................................................................................................. 77
CHAPTER IV: RESULTS ........................................................................................................ 79
A) Introduction ....................................................................................................................... 80
vii

B) Research Overview ............................................................................................................ 80
C) Research Goals and Objectives .......................................................................................... 82
D) Literature Review Results .................................................................................................. 82
E) Modeling Results ............................................................................................................... 83
F) Research Findings .............................................................................................................. 89
H) Conclusions ....................................................................................................................... 92
APPENDIX .............................................................................................................................. 95
A) Model Script ...................................................................................................................... 96
B) Excel Output from the model ........................................................................................... 107
C) Raw output from the model .............................................................................................. 110
REFERENCES ....................................................................................................................... 132

viii

LIST OF TABLES
Table 1: The TFV Theory of Production (Lauri Koskela 1992) ................................................. 39
Table 2: Evolution of models of Project Management (Laufer et al. 1996) ................................ 51
Table 3: Output from the AB model developed ....................................................................... 107

ix

LIST OF FIGURES
Figure 1: Certainty Matrix (Laufer et al. 1996) .......................................................................... 13
Figure 2: Varying Classes of Uncertainty (Hastings and McManus 2004) ................................. 17
Figure 3: The spectrum of dynamism and unknowns (Collyer and Warren 2007) ...................... 25
Figure 4: Control Methods (Collyer and Warren, 2007) ............................................................. 27
Figure 5: The three pillars to successful response to unexpected events (Geraldi et.al. 2009) ..... 30
Figure 6: John Boyd's OODA Loop........................................................................................... 35
Figure 7: Lean Project delivery System (Ballard 200)................................................................ 41
Figure 8: The Last Planner® System (Ballard 2000).................................................................. 45
Figure 9: Crew Delay VS Superintendent Delay (For interpretation of the references to color in
this and all other figures, the reader is referred to the electronic version of this thesis.).............. 84
Figure 10: The increase in knowledge of the crew ..................................................................... 86
Figure 11: Superintendent (Damage VS Delay) ......................................................................... 87
Figure 12: Construction crew taking decisions (Delay VS Damage) .......................................... 88

x

CHAPTER 1: INTRODUCTION

1

A) Motivation
Construction Operations are both dynamic and time sensitive in nature (Abdelhamid et. al,
2009). The dynamic nature and the interactive complexity make construction projects susceptible
to unexpected events. The unexpected events are often the source of project delays and budget
overruns. In many cases these unexpected events are unavoidable and the construction crew has
to be responsive in reacting to these situations. In a construction setting, change is seen more as a
rule than as an exception (Kim and Paulson 2003). These changes can occur due to changes in
design, supply chain issues, site conditions or unexpected operation delays (Sawhney and Walsh
2003). Even though the construction projects are governed by a master plan devised generally at
the start of a project, the onsite activities show more of an emergent behavior.

The unexpected situations are also bound to arise because of the tight coupling of activities on
the project master schedule. A small deviation from this master schedule can throw the project
off its path and create cost and schedule impacts. In construction, it does not take much for
deviations to occur given that activities are most often not conducted in a controlled
environment. In fact, the production phases of construction are highly susceptible to external
agents of variation such as the weather, the supply chain and lastly the economic environment.
Thus it is not a question of if but rather when these events will take place.

The literature provides modest insights into the construction crew’s behavior and decision
making under these uncertain situations that we have come to accept as inherent on construction
projects. The traditional way of dealing with these unexpected events is based on procedural

2

techniques such as executing a series of planned tasks to diffuse the situation (Pich et.al 2002).
Use of pre-planned contingencies is another method often used to deal with the unexpected
events. Most of the procedural methods are a means of reacting to the event and may or may not
be successful in diffusing the situation at hand. By the time project managers or superintendents
react to the situation, the project may have gone through considerable delays and cost overruns.
The delays are compounded because of the ripple effect that results from the linear nature of a
construction project.

Using an Agent-based Model (ABM), this research explores the behavior of construction
workers under unexpected events. The behavior of construction crews that possess the ability to
learn by making decisions (proactive) is contrasted with a construction crew that waits for the
superintendent (reactive) to address the situation.

B) Problem Area
The hard bid type procurement in the construction industry has often shown that contractors do
not completely understand the scope and/or the environment of the construction project in its
entirety before the price is fixed. A short time after taking ownership of the site, contractors
realize they knew less than they perceived about the project. Unexpected situations crop up
pushing the schedule and the budget beyond its set limits.

Unexpected events, specifically the known unknowns type and not the unknown unknowns one,
will cause delays in the project schedule because by definition that not foreseen and, hence, not
included in the project plan. These delays have a significant cost attached to them which rise
with the increase in delays. Strategies for combating unexpected events are developed in a

3

number of fields such as the military, IT and economy with noticeable results. The significance
of these delays is more prominent in the construction industry as shown by the track record of
construction projects being late and over budget frequently. A new strategy for combating known
unknowns is needed. We expect the delays due to unexpected situations to decrease as the
construction crews take responsibility and make decisions.

The research investigated the literature in the fields of mitigating uncertainties in a project
environment. Methods and principles such as Last Planner®, Integrated Project Delivery and
Lean Work Structuring show positive impact in combating unexpected situations or minimizing
them on construction operations at the project and production strategic planning levels but not at
the operational level of the crew. Through synthesis from literature and an agent based model,
the effectiveness of a responsive team is measured against the traditional command and control
environment. In conclusion, a learning method based on flexibility and responsiveness is
propagated for the goal of better mitigation of unexpected events in construction operations. This
method is only an early step to developing a strategy towards our goal.

The Florida Department of Transport (FDOT) did a review of 102 construction projects in the
state of Florida in the year 1996 (Turcotte 1996). The aim of the review was to investigate cost
overruns on the projects and establish accountability for these problems. The FDOT in 1996
maintained 12,000 centerline miles of highway in the state. (Turcotte 1996) Apart from
highways, they also maintain airports and rail services in the state. FDOT is thus a construction
savvy organization. In the review of the 102 construction projects, projects worth $302.7
Million, FDOT faced cost overruns of $28.6 Million or 9.5% of the total project value. The
study classified these into avoidable (reasonably foreseen) and unavoidable costs- 5.2% of the

4

costs were avoidable and the other 4.3% were classified as unavoidable. . Overruns due to
planning failures or design changes were deemed avoidable.

The FDOT study provided recommendations to remedy the avoidable costs but the unavoidable
cost overruns were not discussed. These latter costs are often neglected in the planning process
and there is no defined process to counter them once they do take place even though they could
comprise 5% or more of a project’s total budget. Kloppenborg and Opfer (2002, p. 21) found
that 29 percent of 3,554 project management articles deal with planning, 23 percent with control,
in comparison to only 1 percent which consider the implementation and execution phase of the
project. This underscored the need to investigate construction crew behavior during onsite
disruptions. This research focused on this need and is an attempt to better understand mitigation
of unexpected events. It is important that we accept that unknown events will take place and
build capability in the organization to tackle then instead of trying to stop them from taking place
altogether (Abdelhamid et. al, 2009).

C) Goals and Objectives
The primary goal of this research is to understand crew behavior during onsite disruptions. This
was accomplished through an in-depth literature review related to managing unknown events on
projects. The research explored the various existing management techniques for unexpected
events.

The literature review also compared the various techniques suggested in the literature to manage
unknown events. Most suggested techniques were based on the degree of dynamism of the
industry. The information technology industry, for example, is a highly dynamic industry in the

5

sense that the landscape of available techniques to accomplish tasks can change fundamentally
during the execution of a project. The construction industry is dynamic because it is subject to
constant execution whereas in other industries the tools and techniques that govern the execution
are more stable. A higher degree of innovation is often required to implement these tools for
effective and efficient execution of the tasks. This innovation often takes place at the site level
between the construction crews. Using an agent based model, this research examined this
dynamism in the construction industry by exploring construction crew behavior under
unexpected circumstances. The model instilled autonomy in the construction crews allowing
them to rectify events that come up without warning. The crews took decisions based on the
knowledge that they possessed.

The research thus tests the hypothesis that a self-managed autonomous crew will perform better
at mitigating unexpected events. This hypothesis is tested using an agent based model.

D) Research Scope
An owner survey conducted by Construction Management Association of America reported that
between 40 and 50% of construction projects are behind schedule and/or over budget (CMAA
2006). The FDOT study cited earlier also reported approximately 5% of construction delays
come from unavoidable situations. The literature has shown a distinct lack of management
principles to manage unknown unknowns in construction, specifically at the crew level.
Superintendents are generally charged with the management of these situations and are often
overloaded. These unknown situations when mismanaged can spiral into a crisis and cause a
disruption in the project. Using computer simulation, this thesis explored the methods to manage

6

unknown unknowns and tested a progressive approach by empowering construction workers to
take decisions and solve issues.

Unknown unknowns for construction projects are things we cannot begin to predict will happen.
These arise out of the blue i.e. we know they will happen, but we don’t know when they will
happen, and where they will take place. Neither do we know the magnitude of the disruption they
will cause to the project. In today’s world of increasing complexity and fast paced innovation,
we need to manage this complexity to take advantage of it.

The scope of this research is to explore the current construction workers’ behavior under these
complex and uncertain situations and compare it to a scenario where they are empowered to take
decisions.

An extensive literature review in relevant topics and an agent based model was conducted for
this research. The findings of the thesis are expected to afford a better understanding of how
construction workers should tackle these uncertain situations. Through literature and modeling,
tools and principles have been extracted and summarized to help manage unknown unknowns in
a more effective manner.

E) Layout of the Thesis
The thesis is divided into 5 chapters.

Chapter 1 gives an introduction to the issues faced by the construction industry when managing
unexpected events and establishes the need for a study to examine uncertainty given the rapidly

7

growing and fast changing construction world . The chapter is the motivation to conduct this
research.

Chapter 2 presents the literature review that was conducted in relevant areas to this thesis. A
study of the current method to manage uncertainty in construction and project management
practices is presented. A study into Lean Construction and the tools that can help us tackle
uncertainty on construction sites is presented. The chapter helped outline the research
methodology.

Chapter 3 presents the research methodology adopted. The research uses literature and an agent
based model to study the current and suggested practices to manage uncertainty. The chapter
goes on to explain the development of the components of the model and its outputs.

Chapter 4 outlines the hypothesis of the model. An explanation of the variety of uncertainties
faced in the field is presented before further specifying the approach.

Chapter 5 presents the findings of the model and its relevance to the construction industry. A
conclusion is drawn based on literature and the model outputs . An explanation is presented on
how we can better ready ourselves to manage unknown unknowns on construction sites.

8

CHAPTER 2: LITERATURE REVIEW

9

A) Introduction
“There are known knowns; there are things we know we know.
We also know there are known unknowns; that is to say we know there are some things we do
not know.
But there are also unknown unknowns – the ones we don't know we don't know.”

-

Donald Rumsfeld (Former United States Secretary of Defense)

Unknown unknowns are by definition not known. Some are hopeless to even contemplate
(Hastings and McManus 2004). However, as the secretary of state put it, we know they are out
there, so we need to apply some thought on how to mitigate it.

Craft skills are usually taught by an experienced practitioner through observation, imitation and
correction. Once sufficient skills have been attained by the learner, he or she has to continue
learning by practice and adapt the skills to suit their respective practice (Wood 2004). The
learner for some time may need support in the learning process but is largely left to adapt and
innovate. The knowledge to be transferred in this practice is largely tacit. The teacher or
facilitator of the learning process has to be comfortable with expressing his knowledge that is
largely tacit.

Dealing with unknown unknowns in construction is a lot like craft skills. Each unknown
unknown is a new problem and a unique problem. Subscribing to a process based approach to
deal with these events is difficult due to the fact that we do not know anything about them. We
will be much better prepared to deal with these events if we embrace that they will take place and
teach our crews how tackle them.

10

The foreman and construction crew collectively have a large amount of experience on
construction sites and dealing with issues. The centrally controlled environment in traditional
construction project organizations does not allow them to put this experience to use. Through
learning by doing, the construction foreman has immense tacit knowledge that can be utilized in
dealing with the inevitable unknown unknowns we counter on the construction site.

This research does not undermine the need for planning a construction project. The planning
stages in construction in fact need to be carried out more thoroughly and processes need to be
designed during the early stages of the project. Principles of systematic planning outlined by
Laufer et al. (1996) in fact play an important role in minimizing the occurrences of these
unexpected events. The planning of construction projects through the design and construction
phases bring order to the process.

The construction projects are managed to be as close to the planned sequence as possible and replanned when they deviate from that sequence. This process of planning and re-planning of work
is the base for management of construction projects in the production phases. The management
process serves us well in certain situations that are predictable. When faced with the unknown,
the sequence invariably fails the team. The centrally controlling entity (superintendent) works to
solve the problem to the best of his/her ability and the project is re-planned.

When faced with an unknown unknown, it is necessary to foster an environment where
innovation and creativity is encouraged. These issues are unique and it is difficult to create a
process to tackle them. Geraldi (2009) notes that we need a certain degree of chaos to foster
creativity and innovation. When there is overpowering reliance on order in an organization, it is
hard to deal with uncertainty.
11

Hyundai motor company artificially create crisis in their organization to foster creativity and to
keep their team ready for any unforeseen situations (Kim 1998). Literature has shown that good
project managers rely on creating an environment where the operation crew manages their work
and foster creatively to solve problems. The unexpected situations thus do not have to have
negative impacts, they can be utilized to innovate and create added value for the user.

In most cases, construction project organizations are often over reliant on order. As a result, the
knowledge and experience of the construction crews is underutilized in situations where
creativity is needed on the construction site. Through an approach of having self-managed and
autonomous construction crews, the construction project teams will be able to cope with the
unknown unknowns in an efficient manner.

In traditional construction, we have always tried to eliminate the uncertainty on projects. We are
better equipped to deal with unknown unknown when we embrace the fact that they are
inevitable and prepare the project team to deal with these situations. In a tightly coupled system
as construction, the effect of uncertainty is amplified. If left unattended, the event can spiral out
into a crisis and put the entire project at risk.

12

Figure 1: Certainty Matrix (Laufer et al. 1996)
Figure 9 from Laufer et al. (1996) shows us the relationship between the understanding of the
project and the stage at which the project is in the execution phase. In the traditional (old)
paradigm, when we get to the construction sites, the construction team assumes that they know
everything there is to know about the construction project. The project management team has
generally built contingencies into the project cost to cover any changes. Often, these
contingencies are based on past projects data and the expertise of the planning team.

Starting the project with the assumption that there is nothing uncertain about the project is
perhaps the most dangerous approach to managing uncertainty. There is surprise and panic when

13

things ‘go wrong’. The project plans, budgets and schedules are decided in this approach by a
project manager and pushed on to the construction team and subcontractors. Very little input is
sought from the subcontractors.

In the new paradigm that Laufer et al. (1996) show, the construction team accepts that there is
uncertainty about the project when starting the execution phase. This may be uncertainty in terms
of the performance of a subcontractor or in fact the weather that year. As the construction team
progresses with construction, they identify and resolve these issues thereby reducing the
prevailing uncertainty.

These are two very distinct approaches. Literature and the experiences in construction have
shown that it is a question of when and not if an unknown unknown will occur. Due to this, the
construction team is better equipped to handle the unknown unknown if they accept the existence
of the uncertainty. Uncertainty is exists on construction projects in various degrees. The
following section describes the different levels of uncertainty.

B) Levels of Uncertainty
Hastings and McManus (2004) studied the framework for understanding uncertainties and
techniques for mitigating it. They even propose taking advantage of uncertainties when they are
presented.

14

Many varying degrees of uncertainty exist on today’s complex projects. There are methods to
cope with some classes of uncertainty. For example, uncertainty is often tackled by building up
redundancies in the system. The excess resources help to cushion the blow of the uncertain
occurrence. The current environment of rapidly changing technologies, threats, needs and
budgets has created a need for better understanding of these classes of uncertainty.

The need for designing a robust, flexible and evolutionary systems and designs is foremost in
readying the organization to deal with these uncertainties.

McManus and Hastings (2004) note that uncertainties are things that are not known, or known
imprecisely. Uncertainties are factual and measurable; things are known or not known or known
to a degree.

The authors studied and classified the uncertainties based on the view point of a system architect
or designer. Even though they study a system from a system engineering point of view, these
classes apply very aptly to uncertainties existing in the construction industry.

Lack of Knowledge:
These comprise of facts that are not known or are only known imprecisely. These facts are
needed to complete the design in the intended way. The missing knowledge may need to be
collected or simply created based on customer needs and standards. The lack of knowledge will
exist in the early development stages of the project. Many of these uncertainties are reduced at
appropriate time as the designs progress.

Lack of Definition:

15

These again are specifics regarding designs that may not be known at the outset of the project.
These uncertainties are eliminated as we progress in design and define the scope of the project.
Care must be taken not to define too much of a system early in the project. This can lead to over
specifying elements of work and/or bad specifications.

Statistically characterized (random) Variables:
These are elements of work that cannot always be known precisely. Specifications are set for
these elements by setting standards within which the work is acceptable. The position of a
foundation is needed to be within required tolerances. This forms a limit within which the
foundation will be laid but the precise position is hard to attain till the work is executed. These
uncertainties are harder to handle. The industry is always looking for ways to reduce these
variations that act as sources of uncertainty.

Known Unknowns:

These are things that we know are unknown. These are best managed qualitatively. Change
orders in construction fall under this category. We know there will be changes, but cannot begin
to predict where they will arise from. Contingencies and redundancies are built into the project
plan to deal with these unknowns.

Often risk management plans are based on identification of known unknowns and dealing with
them ad hoc.

Unknown Unknowns:

16

These are by definition unknown. We cannot begin to understand where and when they will
occur. However, we know they will take place and therefore need to put some thought into how
we can best mitigate them when they occur. We build in high safety factors in to the structural
designs of bridges. This is done to cover the occurrence of an unknown unknown that can
potentially have devastating repercussions on the structure.

Lack of Knowledge

Lack of Definition

Statistically Characterized variables

Known Unknowns

Unknown Unknowns

Figure 2: Varying Classes of Uncertainty (Hastings and McManus 2004)

17

This research aims to explore the strategies to mitigate the unknown unknowns on construction
projects. An extensive literature review is conducted to study the existing strategies to mitigate
the unknown unknowns in construction and in project management. An agent based model is
developed to study the effects of self-managed and autonomous crews on mitigating the
occurrence of an unknown unknown event.

The current practices of risk management deal with the known knowns. The base of risk
management strategies lies in defining the known knowns and placing them into statistically
defined variables. Sometimes design options are assessed in this light (Hastings and McManus
2004).

Hastings and McManus note that a strategy based on building systems that are flexible, evolve
and have capability to interpolate are best suited to mitigating unknown unknowns.

C) Introduction
Chapter 1 laid out the motivation, needs and objectives of this research. One of the primary
objectives of the research is to study and understand the current management practices to manage
uncertainty in projects. An extensive literature review of the relevant topics in the industry was
conducted and outlined in this chapter. A study of the prevalent tools and techniques put forward
by Lean Construction are studied and presented as a means to manage uncertainty in the project
planning and execution phases.

18

D) Background on Unexpected Events
J. Reid (2000) noted that unexpected events are inevitable due to human involvement. Secondly,
unexpected events do not discriminate between the big and the small companies or the
specialized and general companies. Each will meet his/her demise at some point.

Projects are inherently uncertain; they are prone to unexpected events, i.e., events that are
predicted but are not expected to take place (Geraldi et.al. 2009). Geraldi (2009) uses the
example of the Titanic sinking as an example. The ship was considered to be a great engineering
success apart from being an extremely successful project. The event of the Titanic hitting an
iceberg could well have been predicted but nobody expected it to happen. Similarly, we can
predict that a backhoe will breakdown but we don’t expect it to happen. This element of surprise
can cause considerable delay to the project (as in case with the Titanic) before the issue is
resolved.

When the unexpected event of the backhoe breaking down does take place, we have
conventional means such as protocols, documents and processes to fall back on. Thus the
strategies are based on reacting to the situation. The other approach that is commonly followed
as a part of risk management is to try and avoid the unexpected situations all together (Geraldi
et.al. 2009). However such an approach can be dangerous as Loosemore (1998) argues that this
strategy can create over-reliance on the anticipation of events. The need to be resilient and
responsive to these within the organization is ignored causing long reaction times and subsequent
delays to follow. Due to the tight coupling of activities on a project and projects in an
organization, the unexpected event in the worst case scenario can spiral out into a crisis (Perrows
1999). Hallgren (2007) says that a crisis is anything that affects stability, security and vitality.

19

Thus in this tightly coupled system, the unexpected event can easily transform into a crisis if
mismanaged. The construction sector has the highest rate of business failures among all sectors
at 20.3% (Equifax Business Failure Report 2011).

There are numerous events that can disrupt a construction project. Hallgren and Wilson (2007)
conducted case studies on construction projects and their deviations from the project plan. The
authors concluded that not all events had negative impacts on the project progress. The events
that had the luxury of possessing float did not cause much delay to the overall project. The
impact only occurred when the deviations are large enough to make non-critical activities
critical. They considred the effect of the activities losing their float to be hidden from sight,
suggesting a loosely coupled system. The consequence of using a central planning practice and
procedures to deal with the unexpected introduces tight coupling in activities. Dealing with these
activities from a procedural stand point can often result in a short term solution.

Hwang and Lichtenthal (2000) identified two kinds of crisis in an organization. The first kind is
the abrupt crisis that can be associated with a particular event that took place and catches the
management off guard. The second kind of crisis is labeled a creeping crisis that develops and
accumulates stressors before eventually erupting. This can be related to construction in that if the
unexpected event delays the critical path, it is an abrupt crisis and would have a higher chance of
delaying the project. The creeping crisis can be compared to the event that disrupts a non-critical
path but can go unnoticed and develop into a full crisis if the situation is not mitigated.

The superintendent on a fairly large construction site is generally the central controlling entity.
He/she is responsible for the daily coordination and management of work, make ready planning,
preparing weekly work plans, maintain a high percent plan complete and problem resolutions.
20

There could be as many as twenty different crews working on a construction site at a particular
time. This puts pressure on the superintendent during unexpected events due to the use of a
central control paradigm. Traditionally, crews will call upon the superintendent to assist them in
resolving problems. At the same time, the crew is autonomous and self-managed, i.e., they will
also frequently attempt to make a decision to resolve the situation at hand and learn from it. This
research explores the behavior of construction crews during unexpected events using Agent
Based Modeling.

E) Methods Suggested in Literature to Mitigate Unexpected Events
There is a lot of literature that talks about general project management, risk management in the
various industries (Geraldi et. al. 2009). There are few studies conducted to analyze and develop
strategies to deal with these unexpected events. Literature in the construction industry provides
little insight in how the crew or the organization behaves during these events. The aim of this
study is to explore this gap through an agent based model where agents as construction crews are
exposed to these uncertain events. The effect of these crews making decisions in contrast with
the crews waiting on the superintendent is studied.

Related literature in construction is mainly that dealing with crisis management. Loosemore
(1998 a,b) found that important aspects such as mutual trust are missing during unexpected
events. Hallgren and Wilson (2008) studied the responses to these events as practiced by an
international construction company and had findings to the same effect. Suggestions for a
strategy based on sociology and philosophy are made to tackle these unforeseen events. They

21

treated a project as something people do rather than a structure of activities bound within an
organization.

There are various external factors that have bearing on the progress and outcome of a project.
These external factors can easily cause a case of an unexpected event. Hallgren and Wilson
considered 15 unexpected events faced by an international construction company that dealt
mainly with power generation. They handled over 100 projects per year. The company possessed
in house design, engineering and legal capabilities. The company maintained a rigorous planning
approach. One of the steps in the planning process was to identify and prevent potential risks to
the project. During the construction phase, construction related events were recorded in monthly
reports. The reports provided some learning for the project team from previous experiences.

Of the fifteen events recorded, 9 arose out of contract disputes, 3 out of miscellaneous causes, 1
each from a guerilla attack, fatality and transport problems. The events were evenly distributed
between creeping (8/15) and abrupt (7/15). The company utilized a dual structure to manage
projects, i.e., one site team and one corporate team. The line between the responsibilities of the
two teams was often fuzzy. Hallgren and Wilson identified responsiveness as one of the most
important criterion to prevent an unexpected event from spiraling into a catastrophe. The
conclusion drawn from the study, as stated by the authors was: “It was the task and not some
structure that kept the team together. Thus, it seems unlikely that a common approach to crisis
management is useful. There are situations that need immediate responses from all necessary and
available resources”

Pich et. al. 2002 found that there are instructionist, learning and selectionism methods to mitigate
complexity and the unknowns in a project. These methods are selected based on the state of the
22

world or the external aspects that affect a project’s progress and outcome. Pich notes “Where
levels of risk are considerably low, the project managers often deal with risk simply by including
slack or time and cost buffers in the projects.” Hence, projects which are well understood and
there are few unknowns are well managed by using a procedural approach to unforeseen events.
The authors suggest an instructionist or a procedural approach for the projects where the
environment is known and well understood. A learning approach is suggested for a project that
requires the project team to be flexible and be proactive to the situation at hand. The learning
approach is simply put, an extension of the instructionist approach where the team learns from
the past and incorporates in the present project to better the response philosophy. Unlike the
instructionist approach, the learning approach requires flexibility in terms of its personnel rather
than contingencies.

A selectionism approach is suggested for projects that undergo radical changes in its life cycle
rendering the previous learning ineffective. The nature of the unknown unknowns is radically
different from prior ones and will be different for the next. The project team thus puts forward a
reasonable policy and observes the outcome. The authors go on to conclude that given a fairly
complex project, the instructionist approach is more likely to fail. Learning and selectionism
strategies seem to outperform the intructionist approach to cope with complexity and ambiguity.

Strictly using selectionism is not suitable for all construction projects. Construction projects are
well grounded in terms of technology and methods used on construction sites. Selectionism is
more suitable for projects where there is a radical change that can occur in those aspects.
Learning as a tool can be used as a strategy to combat the unknown events. The agent based

23

model developed in this research tested the outcome of the learning ability of the construction
crew under unforeseen conditions.

Collyer and Warren (2008) mention that a dynamic project in terms of project management is
one which is influenced by changes in the environment in which it is conducted. The authors
mention that these projects that are conducted in highly uncertain environment are a key
unresolved issue in project management. Unknowns are an inherent character of a project. The
rate of resolving these unknowns has to be faster than the emergence of the events. This cycle is
especially critical in a linear operation such as construction (personal opinion). The authors go
on to present the following methods to tackle these projects and the uncertainty inherent in them:

1. Environment Manipulation
2. Planning approaches for dynamic environments
3. Scope control for dynamic environments
4. Controlled experimentation
5. Lifecycle strategies
6. Management controls
7. Culture and communications for dynamic environments
8.

Categorization

9. Leadership style
Most of the mentioned measures are aimed at and extremely useful in minimizing the
occurrences of these unknown events. However, it is well accepted that in spite of the most
stringent measures, the dynamism and the complexity of the projects are bound to cause events

24

that are hard to foresee. Moreover, no two projects are alike. This holds true in construction
where no two buildings are alike and will have the same purpose.

In complex projects, the managers may lack the experience required to establish accurate
controls (Collyer and Warren 2008). The establishment of strict controls may offend workers and
cause low morale and stifle creativity in the organization. Collyer and Warren also note that
burdening the workers with heavy processes and no incentives will discourage adaption to the
ever changing environment.

Figure 1 shows how the unknowns increase on a dynamic project as compared to operational
work. The dynamism in these projects is best tackled by giving the workers autonomy of action
rather than restricting them with procedures.

Figure 3: The spectrum of dynamism and unknowns (Collyer and Warren 2007)

25

Snell (1992) describes three types of project management controls; behavior, output and input. In
project management, a process or a plan is developed to carry out the tasks. The key is to adhere
to this process and any deviations from the same are monitored. This works well only if the
project and the processes are well understood and a stable process can be created (Collyer and
Warren 2008). In dynamic environments, strict behavior control measures can stifle creativity;
offend workers thereby creating low morale. A central controlling body, the manager, can setup
control measures that are counterproductive and result in unexpected behavior. Managers often
lack experience and knowledge to setup effective controls. If the process is flawed and the
employee can see it is flawed, it may be difficult to correct.

Snell (1992) also presents output or outcome control as a viable option for management of
complex changing environments. In this approach, clearly defined targets are set to provide
direction and discretion to the staff (Snell 1992). Snell advocated this approach for projects or
industries where behavior is difficult to define or expensive to monitor. Incentives are setup
based on reaching the defined targets.

The danger with output control however is deviations are harder to identify early. The mistake
may not be realized till the output is produced and measured.

Snell (1992) also presents an input control method for management of complex and changing
environments. This method is helpful in projects or industries where both behavior control and
measurement of outputs in difficult. The input in the systems can be controlled in terms of the
employees’ ‘knowledge, skills, abilities, values and motivation.’ This is achieved by holding
trainings, more involved staff selection and fostering interaction between employees. The
method relies on building capability within the team and them giving them freedom to achieve.
26

The input control method is the most suitable method for tackling unknown events on
construction projects. The process for mitigating these events is difficult to define and the
outcomes are not measurable. Building capability and then trusting the employees with
experience and knowledge to come up with a solution is an effective way to tackle unknown
events.

Ouchi (1979) explains that it may be viable to control the behavior of a warehouse picker, it
would require a very complex system to describe the behavior of a foreman’s job due to the
subtlety involved. He says a better approach will be to pick the foreman based on previously
demonstrated commitment to the organization’s objectives. The same can be applied to project
work. A team committed to the project and that has shown ability should be picked. Figure 2
shows the methods suggested by Snell (1992), Ouchi (1979) and Collyer (2007). Ouchi and
Collyer (2007) suggest that careful selection of the type of control can mean less control for
managers but it is better to optimize than focus on a ‘single unrealistic approach.’

Input Control

Process Control

Output Control

Eg. Recruitment,
training and
induction. Rely on
when process or goals
hard to define

Eg. Plans, procedures,
and checklists.
Requires predictable
environment

Eg. Rewards &
recognition. More
effective if
measurable

Figure 4: Control Methods (Collyer and Warren, 2007)

In line with Ouchi, Collyer suggests choosing a foreman who is experienced and has previously
demonstrated a high level of commitment to the organizations objectives and giving him/her

27

autonomy of action under those situations. Simons (1995) suggests that innovation should be
allowed between pre-defined boundaries. The boundaries can include codes of conduct and
safety measures. These are especially helpful in projects with high level of dynamism because
they afford flexibility to the team within reasonable boundaries.

Geraldi et al. 2009 studied unexpected events and associated responses in project management to
find out what makes a response successful during unexpected events. The authors concluded that
the successful responses were mainly based on three important pillars:

1. Responsiveness and functional structure: The authors note that quick and appropriate
responses to the unexpected events were a key ingredient in successful responses to these
events. The majority of the interviewed project managers mentioned quick decision
making and implementation of these decisions to be important to diffuse unforeseen
situations. Successful responses depended heavily on empowerment of the workforce
rather than micro management and excessive control by the higher levels of management.
Lack of autonomy caused slow response times and sometimes the team had to make
several responses before the resources were deployed. These severely affected the ability
to respond to the unexpected events successfully.
It is also important to note the difference between flexibility and responsiveness in the
organization. Flexibility of an organization in this case is its ability to change within its
existing pre-defined parameters. Responsiveness on the other hand refers to a team’s
ability to purposefully and quickly respond to a change or an event in the environment.
Agility is also a desirable quality in an organizational makeup. Agility refers to a team’s
ability to rapidly re-configure its self-according to new parameters (Geraldi et al. 2009).

28

2. Good interpersonal skills: Respondents in the study conducted by Geraldi et.al. (2009)
identified engagement with stakeholders as a key to successful responses to unexpected
events. Making sure that all sides involved were committed to enable enough resources to
solve the problem. Some participants complained of support system based on top down
management system. This sort of management is heavy on reporting on a daily basis. The
authors identify that this kind of management method often does not provide timely
response to the situation at hand on the contrary it uses up a lot of time and resources that
could be instrumental in solving the problem at hand.
3. Competent People: The mutual trust between the workforce and the judgment of the
management team emerged as an important aspect contributing to successful responses to
unexpected events. Events where people did not panic were better resolved.

29

Successful Response to Unexpected Events

Group Level
Good Interpersonal
Relationships

Organizational level
Responsive and Functioning
Structure





Engagement with
Stakeholders including ability
,
to negotiate solutions



Communication including
,
availability of information as
well as its communication

High Degree of Freedom
Pace to make and implement
decisions

Individual Level
Competent people


Competence of leader and
team



Behavior, including self
awareness and ability to deal
with stressful situation

Figure 5: The three pillars to successful response to unexpected events (Geraldi et.al. 2009)

30

Figure 5 shows the authors suggested three pillars approach to successful responses to
unexpected events.

The authors note that the ideas that emerge and the three pillars that contribute to the successful
responses to unexpected events point towards a post bureaucratic form of management system.
Post Bureaucratic form of management replaces the centralized, hierarchical form of governance
with a flat, de-centralized organization that emphasizes flexibility and responsiveness rather than
rule following.

Raelin (2010) investigated whether managerial control has outlived its due course. While this
might be true for some industries, the construction industry still has need for managers and
facilitators. The study however presents insights into Post Bureaucratic forms of management
that are very relevant to managing unexpected events in construction.

Raelin (2010) observes that our traditional and at first thought, rational view of management
relies on a degree of predictability of the outcome. This can work well for certain tasks but how
can we know whether a lawyer is going to win his/her case or whether a software engineer will
produce a better code? The post bureaucratic form of governance has been proposed as a method
to increase employee empowerment, autonomy and self-reliance.

There is a call for ‘soft control’ measures or informal control or culture control norms in project
management rather than through hard direction (Grey & Garsten, 2001). Raelin writes,
management by core cultural values to produce a highly motivated workforce has been a vital
tool for corporations. These soft control practices encourage team work, clan control rather than
central control and decrease the amount of surveillance required. The reduced workload on the

31

managers or the superintendent in the case of a construction site will afford him/her to
concentrate on activities such as weekly work planning and coordination of work.

William McKnight, chairman of the board of 3M Corporation from 1949 to 1966, is known for
his management theories that served as guiding principles at 3M (1948), he once said:

“Mistakes will be made. But if a person is essentially right, the mistakes he or she makes are not
as serious in the long run as the mistakes management will make if it undertakes to tell those in
authority [italics added] exactly how they must do their jobs.”

Geraldi (2009) presented a doctoral thesis aimed at exploring how multi-project companies
reconcile order (efficiency, control and clarity) and chaos (creativity, trust, uncertainty,
ambiguity). A lack of order in the organization can cause chaos and stifle efficiency of the
organization. Whereas, a lack of chaos can stifle the much needed creativity and trust needed to
innovate and solve problems.

Geraldi (2009) was employed at a large German construction company as a Research Assistant.
Her main goal during this time was to find how the company balanced chaos and order in the
organization. Geraldi notes that in order to deal with the heterogeneity, the company need to
reconcile order and chaos. More and more companies in search of greater productivity are
foregoing the chaos necessary in the organization. Geraldi notes that over 70 percent of project
managers complain of bureaucracy of processes (Geraldi 2011). This changes the function of the
project managers from managing of risk, process and creativity to managing of papers and forms.

Through this research, Geraldi found empirical data to confirm that organizations have and need
both order and chaos within them. The coexistence of order and chaos was at all levels from
32

project phases to higher levels of the enterprise. She found that companies tend to avoid
coexistence of order and chaos. They tilt either more on the order and compensate with low
chaos or vice versa. We opine that most construction companies tilt towards order. Most
processes are centrally controlled and a lack of chaos is guarded against yet it still exists in an
unbounded fashion.

One of the key findings of Geraldi’s work was that companies subject to external inflexibilities
need to build flexibility within the organization. The organizational flexibility helps absorb the
external inflexibilities that can be caused as a result of unexpected events or stakeholder
demands.

Geraldi concluded that even though order is deemed as fundamental to success of companies, her
findings pointed at excess of order does more harm than excess of chaos.

Abdelhamid et al. (2009) studied the management of unforeseen uncertainties with the help of
decision making cycles. The OODA (observe, orient, decide and act) loop of Col. John Boyd
was proposed as a suitable decision making cycle for tackling uncertainties. Figure 4 shows the
cycles in the OODA loop. The OODA loop is a continuous decision making cycle with four
stages: Observe, Orient, Decide and Act.

The observe phase of the OODA loop is when the person takes note of the issue at hand and
contrasts it to the surroundings. The orient phase is perhaps the most complex phase in the cycle.
The orientation of the observer to the surrounding is based on their perception of the same. A
person will see differently based on their culture, education and analytical capability. In other
words, although it appears as the second phase in the OODA loop, the orient phase affects the

33

observe phase and will affect the following decide phase as well. The orientation can trigger
further and different observation of the situation. This non-linearity or non-sequential nature of
the OODA loop was intentionally designed by Boyd to account for the complexity of the
decision making process. The decide phase is when the person takes a decision based on the
perceptions and understanding of the environment, and orientation. Based on the decision, the
person will act in the last phase of the loop. Again, the OODA loop is not a linear process. There
is feedback and feed forward amongst the different loop phases.

The OODA loop places great emphasis on the criticality of rapid decision making. Business
strategies based on the OODA loop consider quick decision making as an important method to
get ahead of the competition. Conceived for combat situations, the OODA loop has found use in
business transactions. The conclusion of the authors is that the OODA loop in conjunction with
the Last Planner® System is a viable conceptual framework to tackle unforeseen uncertainties.
These are in line with this research in the sense that the OODA loop emphasizes quick decision
making with the people involved and subsequently leaning from these decisions to make better
decisions

in

the

34

future.

Figure 6: John Boyd's OODA Loop

35

The aforementioned studies indicate that attempts at theoretical constructs for avoiding and
resolving unknown unknowns in project management are worthy endeavors. Insights into the
construction industry and the behavior of construction workers under unexpected events remain
under-researched. This research proposes to utilize the concepts of learning and the three pillars
to explore and understand the behavior of construction crews in uncertain conditions.

The research explores the impact of independent decision making; self-managed teams on the
construction site as compared to a conventional team dependent on the superintendent for
directives. Of the three pillars, we have to assume that the team has inherent qualities of
competent personnel and good interpersonal relationships. The responsiveness and its effects on
the project are studied through an agent based model. Learning within the construction crews is a
consequence of making decisions in the model. The tradition in the construction industry has
been that construction workers are hired from the ‘neck down’, i.e., there is no value that they
add to the job by the knowledge that they possess. This is considered a form of waste per Lean
principles. Crews who are on site each day will have vast hands-on knowledge about the work
they perform. They play an important role in predicting and mitigating unforeseen situations that
arise on the construction site. An involvement on these terms would also result in greater
satisfaction for the crew. The feeling that instead of carrying out a set of instructions, the team is
there for a larger contribution improves worker satisfaction and productivity (Liker 2003).

F) Lean Construction
A well accepted definition of construction management has been ‘the judicious allocation of
resources to complete a project at budget, on time and with desired quality.’ The theory of
construction management has in the past been based on the Transformation view. The traditional
36

construction practices view construction as a transformation of inputs into outputs (Koskela
1992). The theory of construction management is really a theory of task management. However
in a complex production system like construction, we cannot rely on a theory based solely on
transformation.

Relying on task management theories in construction management is extremely dangerous
because task management assumes certainty in the production system. In a complicated
production system such as construction, the idealization error associated to transformation can
become large (Koskela 1992). This is reflected in what happens in practice.

The transformation view of production tells us that production is a transformation of inputs to
outputs. There are a number of principles that govern the transformation theory of production.
For example the principles tell us that the production can be accomplished through decomposing
of the total transformation into tasks and executing these tasks. To decrease the cost or the time
required to complete the total transformation, we try to minimize the costs or time to execute
these tasks.

The flow view of production was firsts suggested by Gilberths in 1922. The flow view of
production formed the basis for what we know today as Lean production and Just-In-Time (JIT)
systems. The production system is viewed as a flow of material and information through the
system. The flow view of production advocates waste reduction from the flow processes
(Koskela 1992). The flow view takes into account principles cycle time, lead time reduction and
simplification.

37

The third view of production is the value generation view. The main aim of the production
system is to generate maximum value for the customer. The value generation view of production
is refined and used as foundation for quality movement in various industries.

To counter the problems caused by these assumptions, Koskela put forward the Transformation
Flow and Value (TFV) view of production in construction. The TFV theory is thus based on task
management, flow management and value management. Figure 5 shows a comparison of the
three production management theories.

38

Transformation view

Main Principle

Methods and
Practices

As a transformation of
inputs into outputs

As a flow of material,
composed of
transformation,
inspection, moving and
waiting

As a process where
value for the customer
is created through
fulfillment of his
requirements

Getting production
realized effectively
Workbreakdown
structure, MRP,
Organizational
responsibility charts

Elimincation of waste
(non-value adding
activities)
Continuous flow, pull
production control,
continuous
improvement

Elimination of value
loss (achieved value in
relation to best
possible value)
Methods for
requirement capture,
Quality finction
deployment

Taking care of that
what is unnecessary is
done as little as
possible

Conceptualization
of Production

Flow View

Value Generation
View

Taking care of that
customer requirements
are met in the best
possible manner

Taking care of what
Practical
has to be done
contribution
Suggested name
for practical
application of the
Task Management
Flow management
Value Management
view
Table 1: The TFV Theory of Production (Lauri Koskela 1992)

Lauri Koskela’s TFV view of construction management has played a major role in forming the
basis for what we know today as Lean Construction.

In the term Lean Construction, construction refers to the construction industry rather than the
actual activity of construction. Lean Construction is a new paradigm that presents a collaborative
model to project management. The project team including the owner, designers, construction
manager, and specialty subcontractors form an alliance with the success of the project as an
overarching aim.

39

Lean Construction is a complete delivery system, in the conceptual sense and not the contractual
one. Figure 6 illustrates the various phases and tasks that are carried out in the LPDS, which is
divided into the phases of Lean Programming, Lean Design, Lean Supply, Lean Assembly and
Lean Use. Through these phases, Lean Construction provides fundamental concepts and tools for
carrying out construction projects from inception through the design, construction and the use of
the structure.

Lean Construction presents various tools to achieve the aims of construction project
management. Many of these tools are very relevant to and help to prepare for unknown
unknowns

on

construction

40

sites.

Figure 7: Lean Project delivery System (Ballard 200)

41

LPDS presents various tools and techniques that help to make the project work more reliably and
with minimum of waste. The approach relies on involving the various project partners early in
the project and working together to achieve the best possible solutions to the problems. Bringing
the project partners together and solving issues results in a more certain and agreeable plan for
the project. Some of the tools are discussed further.

G) The Last Planner

LPS

®

®

®

System (LPS )

is a production management module developed by Howell and Ballard (1994). Production

Management can be divided into two functions; planning and control.

Planning activities establishes a goal and a desired sequence of activities to achieve this goal.
The control activities initiate re-planning of activities once a plan undergoes deviations and the
original sequence is not sufficient to reach the goals.

In a dynamic project environment such as construction, the planning activities cannot be outlined
with accuracy much before the work is to be performed on a construction site. As a result, the
next piece of work to be performed by a design team or a construction crew can hardly be
managed by a master schedule developed at the beginning of the project (Ballard 2000). In spite
of this reality, construction projects continue to be managed through a detailed CPM (Critical
Path Methods) schedule that is developed early on in the project development stage. This usually
works against successful production management. Howell and Ballard (1994) found that, on
average, only 50% of weekly planned production activities were executed as planned by

42

construction crews. They attributed this failure to the inability to predict workflow on the
construction site.

To solve this prevalent problem in the construction industry, Ballard and Howell (1994) began

®

introducing the Last Planner System. LPS . LPS

®

aims to make the construction operations

flow more predictable by utilizing a series of planning events that consider the best available
information from the construction site with respect to execution.

Constraints preventing

execution are pro-actively removed.

The LPS® uses a scheduling charrette, a gathering of all project players, with the following
outcomes; Master schedule, Phase schedule, MakeReady plan and the Weekly Workplan. The
work is planned based on different stages of readiness. Master and Phase schedule arrive at the
work that should be done. The makeready plan actively takes the work that should be done to
work that can be done and the weekly work plan is a listing of the the work that will be done (
work that is committed to by the crews).

Master Schedule: The master schedule is generally a critical path schedule that is based on
milestones and may be used to track the overall progress of the project. The schedule is based on
owner milestones and is developed early in the project planning process. The schedule in most
cases is a part of the contract between the owner and the contractor.

In most delivery methods for construction projects, the contractor is not brought into the design
process. This leads to the master schedule being full of guess work rather than intended work
plans by the contractors executing the work. The general contractor or construction manager may

43

not have enough knowledge about the design and the work to be carried out to accurately
estimate the time and resources required to execute the project. The same master schedule cannot
be used to manage production 6 months into the project. It is hard to gauge uncertainties like the
weather half way into the project.

In the LPS®, the master schedule is used as a milestone schedule. The master schedule is further
broken down into multiple phases.

Phase Schedule: A detailing of activities to take place between the milestones shown on the
master or the milestone schedule. The phase schedule is formed through a process called pull
planning rather than pushing dates on to the subcontractors. Phase Schedule (also called pull
planning session or pull scheduling session) brings together all trade partners into one room to
outline their plan for executing respective scopes of work. The durations, the resources and the
pre-requisite work required are derivatives of the pull planning sessions.

In pull planning, the work is pulled from the end of the project to the beginning. The project
team collaboratively fixes the activities and timelines for each phase. It is much easier to get buy
in from the entire project team because of their involvement in planning the schedule. Once the
phase schedule is fixed, the schedule is entered into scheduling software and communicated with
the project team.

44

Master Scheduling

SHOULD
Phase Scheduling

CAN

Make Ready Planning

WILL

Weekly Work Planning

DID

Learning

Figure 8: The Last Planner® System (Ballard 2000)
Make Ready Plan (or Look Ahead Plan): The master and phase schedules make up the
activities that should be performed. The Make Ready plan tells us what we can perform at any
given stage of the project.
The Make Ready plan or the look ahead plan is where we screen the activities on the phase
schedule for constraints that will stop it from being performed on site. The plans are generally
carried out using a 6-week window. The activities are screened for three fundamental constraints;
directives, pre-requisite work and resources.
Directives: These are information, instructions, product data, quality assurance
requirements or testing requirements that might be critical to the start of an activity. Lack

45

of these directives will result in the activity not starting on time or not being carried out
to required quality.
Pre-requisite work: The activities on the phase schedule require pre-requisite work to get
started. A steel column cannot be erected until the base plate is installed to requirements.
The pre-requisite work has to be of required quality and quantity to avoid rework later in
the project.
Resources: The activities on the phase schedule require resources such as material, labor
and equipment. It is imperative to ascertain that these resources are available before the
activities are put on the Makeready plan.
A lack of any of these will result in the failure to start the activity as planned. If the activity
screening results in the conclusion that it is ready to start, the activity is placed on the Makeready
plan as an activity that can be executed.
Weekly Work Plan (WWP): The Makeready plan is generally conducted on a 6 week look
ahead on the project. The Makeready plan is further broken down into tasks that the construction
crew can commit to be completed in the upcoming week. The WWP is what the construction
crew will accomplish within the next week on the construction site. These commitments are
made by the last planners on the projects, i.e. the foreman and crews that actually carry out the
tasks.
The commitments by each construction crew are logged and tracked during the week. The work
is committed to only a week in advance and therefore the judgment of the amount of work that
will be done is more accurate than that on the phase schedule. The weekly work plan is accurate
also because that the crew is committing to the work to be done and not a project manager from
46

the office. The fact that the work is planned and committed to by the last planners and only a
week in advance increases the reliability of the commitment.
Percent Plan Complete (PPC): Once each crew commits to complete a number of tasks in the
upcoming week, the project team updates the weekly work plan on a daily basis with the actual
number of tasks completed. For the tasks to be counted as complete, they must be finished to
satisfy all quality requirements.
The percentage of tasks complete at the end of the week against the tasks committed by the crew
at the start of the week is the PPC of the week. The PPC is calculated for each crew. PPC is an
indicator for the reliability of construction operations on a weekly basis. PPC forms a baseline
each week from which the construction crew tries to better its performance.
During tracking of the work done, the project team cites reasons for any tasks that are not
completed in the given week. These are often termed as variances due to which the tasks were
not completed. An analysis of the reported variances leads to identification of problems on the
construction site that the team needs to resolve.
The aim of the exercise is always to get as close to 100% PPC as we can. By removing the
variances reported in the field, the team creates a more reliable workflow and better results. As a
result, the project will fare better compared to the estimate and schedule.
The PPC is often made public on construction projects. This is an attempt to foster competition
between the crews. The subcontractors that are performing below par are easily found out and
are under pressure to better their performance so that the entire project is not held back.

47

As noted earlier, we cannot completely eliminate variations in the project plan. . The LPS

®

is

production management tool that is aimed at minimizing variation on the construction site. It
does this by combating the sources of variation, mainly those arising from known unknowns
The variations that remain in the system are those arising from unknown unknowns.
In traditional construction, the progress of the project may be updated on a monthly basis. Hence,
the identification of any deviations from the planned schedule is at the least a month from the
occurrence of the deviation. The measures that are taken to mitigate the deviations are therefore
reactive. The source of the deviation itself is harder to identify. It is a greater possibility in this
case that the deviation could cause a cascading event and cause delays on the projects.
Work Structuring
Work Structuring in Lean Construction is defined as ‘the development of operations and process
design in alignment with the product design, the design of supply chains, the allocation of
resources and design for assembly efforts.’ The goal of work structuring is to make workflow
more reliable and quick while delivering value to the customer (Ballard 2000).
Work structuring in construction answers the following questions (Ballard 1999):
1. What units of work will be allocated to what workers?
2. How will the work be sequenced?
3. How will the work be released from one sequence of workers to the next?
4. Will consecutive groups of workers execute the work in a continuous manner or will the work be
decoupled?
5. Where are the decoupling buffers needed and how should they be sized?
6. When will the units of work be done?

48

Work Structuring thus is a process that should be re-evaluated during the course of the project
(Tsao 2004). Work structuring is by no means a new concept. Construction teams have been
structuring work for as long as we have been constructing. However, the formal process of work
structuring forms an integral part of designing a resilient project. Work structuring can play an
important role in combating unforeseen situations as there is attention paid to the design of the
operation processes. Situations that can cause an event that affects the project in operations are
often unearthed and solutions are found to perform the operation better.

For example, it may be found that due to high ceilings, there is a safety risk in installing
electrical conduits on the ceiling. The contractor may point out to the designers that this causes
low efficiency, excessive need for equipment use and safety issues. The team may decide to
lower the ceilings or design the process to eliminate as many of the issues as they can. As a
result, some of the sources that can create an unexpected event are eliminated. The project gains
as a whole.

Tsao et al. (2004) note the process of work structuring is fundamentally different from the
traditional work breakdown structure in the construction industry. We understand from the TFV
theory presented by Koskela that construction management is not merely a transformation from
inputs to outputs. The work breakdown structure used in construction is a process of breaking
down the project into tasks. These tasks are assigned costs on estimates and time on the schedule.
Work breakdown structures are thus tasks management tools. Through work breakdown
structures we cannot understand the complexity caused by the interaction and overlap in
processes of construction operations (Tsao et al. 2004). While the design of each separate part
may appear to be sensible, the production process as a whole may be far from optimal. The

49

breakdown structure approach may amplify the errors made upstream in the production process
and are only realized later in the downstream processes.

Tsao et al (2004) conclud that the solution is to design the production process in construction in
line with, and in a fashion similar to, the product design. This eliminates much rework,
unexpected situations and conflicts in design. One of the findings they refer to is is that having
more detailed designs at proposal stage is not the solution. A higher level of design early by the
architect is often a constrictive practice and inhibits innovation and creativity. The solution is to
have the contractors come into the design process early and contribute by identifying the issues
with executing the design in the field.

Laufer et al. (1996) investigated how managers of capital projects successfully complete their
job in today’s uncertain and complex construction environment. They noted that the successful
management practices they found in their studies deviated greatly from traditional management
methods.

To understand the new management philosophies, we have to understand the evolution of
management concepts over the years and decades. Laufer et al (1996) notes 4 distinct
generations of management principles, each incorporating the traits of the preceding generation
(as Figure 8 illustrates).

In the 1960s, the first generation can be traced back to CPM and PERT techniques. These
techniques emerged in large defense and space contracts. The construction industry was quick to
adapt this approach to project management as a method in its own. Construction as an industry
dedicated substantial amount of financial and professional resources to further develop this

50

technique. This approach concentrates on the coordination of sequential and parallel activities
and their control. The technique was very much like flight scheduling problems in major airlines.
CPM and PERT proved to be effective tools for managing simple projects with low levels of
uncertainty (Laufer et al. 1996). This approach was termed as scheduling by Laufer et al. (1996).

Central
Concept

Era of
Model

Dominant
Project
Characteristics

Main Thrust

Metaphor

1960s

Scheduling
(control)

Teamwork
(integration)
Reducing
uncertainty
(Flexibility)

Simple, Certain

Coordinating

Schedule
regional flights
in an airline

1970s

Complex,
uncertain

Cooperation
between
participants

Conducting a
symphony
orchestra

1980s

Simultaneity

1990s

Complex,
uncertain

Complex,
Uncertain,
quick

Making stable
decisions

Orchestrating
contending
demands

Exploring an
unknown
country
Directing a
three-ring circus
continuously
switching acts
based on the
crowd's
response

Means
Information
technology,
planning
specialists
Process
facilitation,
definition of
roles
Search for
information,
selective
redundancy

Experience
responsiveness
and
adaptability

Table 2: Evolution of models of Project Management (Laufer et al. 1996)
In the 1970s, a teamwork approach was adapted for project management. There was a realization
of the need for managing complex projects, projects consisting of dissimilar and highly
interdependent components requiring many specialty skills. The challenge here was to interlace
these components and integrate the work process across the teams. The manager in this case took
charge of facilitating the process and defining roles (Laufer et al. 1996)

51

The two approaches based on scheduling and teamwork fit well into an environment with low
uncertainty. It was soon realized that projects rarely have static goals. Ever changing goals,
perpetual changes and unexpected events constantly disrupted the project schedules.

In the 1980s, the more resourceful project managers aimed at reducing uncertainty on projects.
The main test in this technique was to make stable and long standing decisions that are relevant
to the project. The managers collected as much information and experience as they could and
made educated decisions. Flexibility was introduced by introducing selective buffers to ensure
that the project stays in track.

The 1990’s brought a need for faster execution of projects (Laufer et al. 1996). The project
managers were managing more projects and time to market was the driving force behind most
projects. The successful project managers managed tasks by integrating them and keeping them
separate at the same time (Laufer et al. 1996). They resorted to a myriad of informal and formal
procedures. The goals and methods were decided on simultaneously and interactively rather than
separately and sequentially. This is termed as Simultaneity by Laufer et al. (1996).

Today, as the projects get more complex, the managers on projects face issues of large issues and
small details. It is easy for managers to get stuck in one domain (Laufer et al. 1996). The more
successful project managers swing back and forth in sync with rising and falling need – termed
Simultaneous Management. As the projects get more and more complex, there needs to be a
different concept to execute project management. Laufer et al. (2009) outline 9 principles to
achieve better simultaneous management techniques to tackle the complexity in projects. The
authors’ nine principles are divided into three groups: Planning, leadership and integration, and
systems.
52

Planning:

1. Systematic and integrative planning:

Laufer et al. argue ‘The need for speed should not be a reason to spare time and effort on
early, integrative, and systematic planning.’ Systematic planning is a multi-phase design
process. The phases in systematic planning include setting up objectives, deciding on the
means and strategies for implementing and determining the resources needed to execute the
project. The typical areas dealt with during the planning stage include project definition,
budget, schedule, logistics, human resources and procurement of construction. The
documents created during the planning exercise guide the project team through execution of
the project, authorization activities, communications and control of the project.

Systematic planning is a pillar for successful project management. However, to translate the
principles effectively into reality under complex and uncertain conditions, we need to have
timely decision making (Laufer et al. 1996).

2. Timely Decision making:

Under conditions of uncertainty, managers often resort to stable decisions; they cannot look
for the optimal decision. It is easier for managers to take a complete look at information and
inputs in the early planning phases of the project. They can therefore afford to pay much
greater attention to quality and completeness in the planning phases. The emphasis here is
not only to react to information, it is to proactively decide on matters and identify missing
information.

53

In early stages of planning, the importance of decision making assumes great importance.
Successful project managers do not see their role as making the best decision. It is their duty
to ensure that the best decision is made (Laufer et al. 1996). Managers should facilitate the
decision making process. The early stages of planning are often seen as a production of
decisions. It is the project managers duty to ensure the right decisions are made and to the
right amount of detail.

3. Isolation and Absorption:

This principle is designed to produce the most robust plan capable of tackling the unexpected
situations during the execution phase. This principle however does not deal with missing
information or guidelines. The emphasis here is on the multiple interdependent and dissimilar
elements of the project.

Laufer et al. (2009) note that under situations with high certainty, high interdependence of
processes result in optimal utilization of resources. However, under environments with
uncertainty, this is not a good practice. A few uncertain tasks can quickly result in a
cascading effect in downstream tasks. Laufer et al. explain it is the same as one bad apple
spoiling all the good apples in a basket. The independent and dissimilar tasks therefore need
to be decoupled by employing redundant resources.

The probability of having a successful project diminishes by having a large number of
uncertain activities. Therefore, it is a good practice to decouple the uncertain activities from
the rest by using these redundant resources. This is an effective way to shield the project

54

from sources of uncertainty. Where there is high uncertainty on projects, the authors suggest
using set based concepts i.e. using two parallel alternatives for development.

Leadership and Integration

4. Inward and Outward Leadership:

To have a successful project, the project manager should be brought in early in the design
process and should remain a part of the project till the customer takes it over. The project
managers have a large influence on the cost, schedule and the operations on the project. The
main duty of the project manager however lies in integrating the various project partners. For
a construction project, that may be the architect, the owner and the various subcontractors.

Leadership is an important characteristic for the project manager in turbulent times. The
project manager thus need to exhibit leadership both within the organization by motivating
and aligning the teams work towards the goals of the project. The complexity of the projects
today makes them susceptible to influence from outside the project. The project management
needs to be a leader in terms of tackling the supply chain for example.

5. Multi-Phase Integration

Multiple phase integration principles consist of important organizational and time related
aspects to achieve speed and quality of decisions (Laufer et al. 1996). The construction
project goes through various phases with different organizations taking responsibility for
each. A typical construction project may comprise of the following phases: feasibility
studies, conceptual planning, project definition, engineering design, procurement,
construction and commissioning/startup. Traditionally, the first three phases are often more
55

design and owner oriented while the contractor takes responsibility of the later three. There is
a need to bring the leaders of the later phases into the project at the latest by the feasibility
studies stage. Similarly, the architects and the designers need to be included in the
construction phases downstream.

Using the principles of multiple phase integration has three advantages. Firstly, due to the
participation of the downstream members in the upstream discussions, the decisions are made
with more experience and in a more educated manner. There are fewer surprises for the
contractors downstream as they are involved in the decision making processes upstream.
Secondly, the production teams downstream understand better the objectives and the intent of
the designers. The construction teams will find it easier to align their efforts better with the
efforts of the project. Thirdly, the communication between the upstream and the downstream
participants will improve. The collaborative nature of the multi-phase integration produces a
transparent, bi-directional and continuous communication process on project (Laufer et al.
1996). As a result, there will be less need for change later in the construction phases and a
much better level of decision making productivity.

6. Multi-Disciplinary Teams

Laufer et al. (1996) explain that projects are executed by teams and not by a single project
manager. Today’s complex construction projects require an inter-disciplinary team to make
decisions and execute work. The team is committed to aligning their interests with the
interests of the project. The individual on the team should depend on one another on
important decisions.

56

The project team as a whole work to create a sense of belonging on the project and pride in
their respective specializations (Laufer et al. 1996). High performance project teams however
realize that individual successes are inseparable form team successes.

The multi-disciplinary teams are generally tied together with loose guidelines and room for
creativity and innovation. As the project progresses, there is a transition from unknowns to
known and from creative discovery to planned activity (Laufer et al. 1996). These multidisciplinary teams help to tackle the unexpected situations when they occur. The high level
of creativity and varied knowledge is an asset to finding innovative solutions.

Systems

7. Intensive Communication:

Intensive communications are perhaps the central factor in making a project successful. This
hold true for most projects much more so for today’s complex construction projects. Laufer
et al. (1996) note that the best remedy to reduce uncertainty in an ambiguous, clouded,
changing and volatile world is in fact the creation and flow of information. When a project is
undergoing fast developments, the information flow must keep up. There is a significant
amount of information that is circulated during the duration of the project. The team is hard
pressed for time to receive, scrutinize and impart this information to others.

Laufer et al. compare the information flow on projects to traffic in a metropolitan city. There
is a large volume of it, it comes from many places and goes to many different places.

8. Simple Procedures:

57

Master project managers often resort to the simplest forms of processes. Although complex
and indifferent in nature, the processes can be applied to today’s construction projects in a
similar way. A network of processes is the backbone of project management. Some
companies however, have a tendency to use too many processes.

Laufer et al. note that in successful companies, the formal processes are often used as a
means for horizontal communications. The processes are much less used for vertical and intra
organization control. While designing the processes, it should always be kept as close to the
ultimate value to the customer as possible. The processes should therefore be developed with
constant input from the field staff. The procedures should show in detail the product to be
attained rather than the way to get there. On the ‘how’ side, it should be much more like a
general outline.

The result of this is a simple and flexible process that requires only minimal amounts of
paperwork and approvals.

9. Systematic Monitoring

The traditional means of monitoring or ‘control’ is to make certain that we compel the events
to conform to the plan. The project managers in simultaneous management measure the
utilization of resources, the achievements of objectives and the validity of the objectives
itself.

The principles of simultaneous management present concepts that are instrumental in eliminating
the sources of uncertainty in the design phases and mitigating them in the production phase of
construction.

58

H) Autonomy in an Organization

The study of project management literature in the construction industry and other industries to
manage unexpected and dynamic events points us to maintaining flexibility and responsiveness
of action in the organization. The flexibility can be achieved through autonomy in the
organization.
From the studied literature, we have also learned that the unexpected events are best embraced
and not avoided as they will inevitably happen. Since we cannot avoid them, we have to prepare
the organization and the project teams to tackle them when they occur. In an autonomous and
responsive team, these unexpected events often prove to be opportunities for organizational
learning. Hyundai, one of the worlds’ leading automotive companies creates internal crises in its
organization at various levels to foster innovation and organizational learning (Kim 1998).
Polanyi (1966) identifies that there are two distinct dimensions to knowledge in an organization;
explicit and tacit. Explicit knowledge is knowledge that can be codified and transmitted in form
of texts, specifications and designs. Tacit knowledge on the other hand is embedded deep within
human minds. Tacit knowledge is only expressed through action, commitment and involvement
in a particular context. Tacit knowledge is only acquired through experience such as observation,
action and practice.
Kim (1998) notes that the tacit knowledge in an organization is nothing but the core of its prior
knowledge base. The organization may have some explicit knowledge in the form of manuals,
instructions and procedures. The explicit knowledge however is useful only when tacit

59

knowledge enables its members to utilize them. Thus the performance of the organization is
often a function of its tacit knowledge embodied within its members.
The transfer of this tacit and explicit knowledge within the organization is instrumental in
sustaining learning in the organization (Kim 1998). The transfer of knowledge takes place in the
following ways:
Tacit to Tacit (socialization): Tacit knowledge is transferred from an individual to the
organization through training and workshops within the organization.
Explicit to Explicit (combination): Explicit knowledge is transferred when an individual takes
discrete pieces of explicit information and combines it to form new explicit information.
Tacit to Explicit (externalization): Tacit knowledge is converted to explicit knowledge when an
individual identifies the foundations of his/her tacit knowledge.
Explicit to Tacit (Internalization): The transfer from explicit to tacit knowledge takes place when
new explicit knowledge is shared throughout the organization and members start using,
expanding and reframing the knowledge.
All organizations are learning systems, they learn as they develop and deliver products (Kim
1998). The organizational learning and the intensity required to transfer knowledge depends on
factors such as autonomy, flexibility and leadership among others. The autonomy and flexibility
in the organization facilitate more efficient sharing of knowledge. More specifically, Kim (1998)
notes that autonomy in the organization creates an environment that enables teams to function
creatively. The creativity is an important trait to possess for mitigating unexpected situations.

60

Hallgren and Olsson (2008) conducted a study to find out how unexpected events are managed in
practice. The authors resonate that unexpected events will occur on projects and they will
inevitably cause delays on projects that substantially increase project costs. The authors note that
the deviations on projects are managed by tools and methods that are reactive. These tools and
methods do not describe what actually happens when a deviation occurs, i.e., they do not tell us
actions, they only tell us what is considered to be the right procedure.
One of the main characteristics of a project is that it is unique (Hallgren and Olson 2008). This
trait of the project in itself causes a degree of uncertainty on the project. The projects are planned
assuming linear rationality. The work is broken down using a work breakdown structure (WBS)
into tasks and assigned costs, time and responsibility. The changes to this project plan are
managed by tools such as earned value, information systems, expert judgment, and formal and
informal project management. The access to these different tools and techniques creates an
illusion of the project being planned and executed in a controlled manner (Hallgren and Olsson
2008).
There will be unexpected events that occur even after the use of tools and methods to control the
project. These unexpected situations are problematic and require a lot of effort on behalf of the
project manager. A major part of the uncertainty is because of not knowing where the deviation
will arise from. These events will occur sooner or later due to the lack of independence between
the different components of work (Hallgren and Olsson 2008). The common answer to these
unexpected situations has been to add contingencies and redundancies but neither is enough even
in an extremely highly controlled environment (Perrow 1999).

61

Instead of looking at management of unexpected situations by tools and methods, we should look
to understand the management of these unexpected situations from an organizing point of view
(Hallgren and Olsson 2008). In an organizing point of view, the nonlinear form of the project can
be understood. When things start falling apart on a project, the participants act not according to
the planned procedure but rather based on improvisational pattern. Moreover, the management of
unexpected events has proved to be a trial and error endeavor with some basic structures
(Hallgren and Olsson 2008).
Hallgren and Olsson (2008) present a case study of a power plant construction project where
they studied the occurrence and management of an unexpected event. They noted that most of
the literature on the subject of unexpected events concentrates on what should be done rather
than what is done. The literature describes procedures to follow when everything falls into place.
Despite the efforts of project teams to eliminate the unexpected events, the events are persistent
and they remain very expensive to projects. The study concludes that the project team handled
the unexpected event by an informal process. The authors suggest participative observations as
an effective way for the team to understand these unexpected situations and learn how to act
when they occur.
The author’s case study showed that the unexpected event of a piece of equipment not reaching
on time did not receive much attention. Thus bringing out the emphasis on what is done rather
than drawing up plans for what should be done.
Hallgren and Olsson also identify the occurrences of the unexpected events as an opportunity for
learning by reflection. The project teams must hold individual and collective reflection. Through
these sessions, the team members’ prior knowledge is made explicit through participation

62

whereby new ideas and knowledge can be created. The knowledge created in the reflection
cycles are instrumental in building capability of the project team in mitigating the next
unexpected event that occurs.
Engwall and Svensson (2004) suggest forming cheetah teams to combat unexpected events. The
cheetah teams are ad hoc teams of skilled members that are separate from the team carrying out
the work. Cheetah teams are called up to solve unexpected events that threaten holding up the
project and causing budget overruns.

Construction is a management intensive industry (Betts and Ofori 1992). A shift from the
management intensive nature is key to mitigating unexpected events. It is essential to give the
foreman and crew autonomy of action in order to act quickly to mitigate the event from spiraling
into a crisis.

I) Significance of the Research
There are numerous principles and theories for dealing with these uncertainty and complexity
(Pich et.al. 2002). Perrow (1994) argues that it is not a question of ‘if’ but ‘when’ the unexpected
event will occur. Perrow states ‘no matter how hard we try there will be serious accidents
because of interactive complexity (which allows the inevitable error to interact in unexpected
ways and defeat safety systems) and tight coupling (in which small errors propagate into major
ones) of most risky systems.’

Unknown unknowns thus pose a complicated problem in management. We know they will
happen but cannot begin to predict where they will come from and when. However, the fact that

63

we know they will happen and will pose significant threats to the project forces us to study and
develop an approach to tackle them.

This research considers contrasting self-managed autonomous methods against the traditional
centrally controlled method to explore the behavior of construction crews under unexpected
events. The effects on a construction project are the results of an agent based model. A better
understanding of the current practices and the autonomous approach can provide insights to
strategies to address unexpected events.

The construction industry is time sensitive in nature. Unexpected events can cause considerable
delays on construction projects and subsequently cause cost overruns and delays. A better
understanding of the behavior of construction workers facing unexpected events is a step in the
right direction towards formulating a strategy to help embrace and tackle these situations on
construction projects.

J) Agent Based Modeling
Agent Based Modeling (ABM) is considered as a system in which a simulation experiment is
created around a set of autonomous agents that interact with each other and the surrounding
environment to imitate the real world proceedings (Sanchez and Lucas 2002). ABM is used in
various fields of social sciences, ecology, politics, economics, marketing and sales. In each of
these fields, the agents are designed such that they sense and respond stochastically to the
conditions in the local environment to replicate complex large scale system behavior. Each agent
is developed as an individual decision making entity.

64

Though limited in numbers, ABM simulations have been used in the construction industry to
simulate various facets of the construction process. Construction has been widely viewed in
history as a centrally controlled process. One plan is created and it is assumed that the
construction will be carried out according to the plan and the entities will have little impact on
the process (Sawhney and Walsh 2003). Howell (1999) suggested that the dynamics of a
construction industry could be better explained through a decentralized and autonomy based
concept. At a micro level, the onsite construction activities show more “organic control” rather
than central control system (Howell 1999). These behavior types are more easily modeled using
ABM. Moreover, a construction project is subject to constant alterations due to changes in
design, unexpected delays and interruptions, field conditions, etc.

Sawhney and Walsh (2003) cited that most current approaches to create a safe and congenial
work environment on the construction site is focused on past indicators and not enough study of
organizational, managerial and human factors have been incorporated. Much of this is due to the
lack of availability of adequate modeling techniques and analysis tools. It is extremely difficult
to model aspects of trust and support within a construction crew, safety rules and procedures and
amount of time available to carry out the work, etc.

Another aspect that makes the construction industry unique is that some of the workers are more
risk tolerant than others and readily take risks that others would not. Accidents can occur to both
the risk averse and the risk taking workers but in different proportions. The work environment
could also become more unsafe due to the actions taken by the workers and their interactions
with the environment. These observations cited by Sawhey and Walsh were specific to the safety
in construction but also can be applied to other aspects.

65

ABM simulations have been adapted to reduce the inefficiencies of the construction claim
negotiation process, to simulate the processes in production home building, the response of
crews to company safety cultures and to simulate dynamic interactions in the supply chain
(Palaniappan et. al. 2007) among others. “Agent Based Modeling Simulations is a promising area
that has the potential of providing answers to many questions that have remained unanswered in
the construction discipline” Sawhney and Walsh (2003). The approach can be effectively used
to study the emergent nature of the construction projects and the proactive application of human
factors in construction. The complex nature of the construction industry makes ABM simulations
a suitable tool for its analysis.

66

CHAPTER 3: RESEARCH METHOD

67

A) Introduction
Construction projects by their dynamic nature are susceptible to change and uncertainties. The
conclusion from studies and literature is that uncertainties in construction should be embraced
and accepted as a challenge rather than artificially ruling them out (Abdelhamid et al. 2009). The
construction teams would improve their workflow reliability through meaningful discussions
innovations. The probability of the occurrence of these uncertainties is low and is generally
managed by a project contingency incorporated in the budget.

This research aims at unknown events not covered by this project contingency budget. Known
unknowns in construction can be represented by owner scope changes, change orders, etc. We
know that these exist and that they are bound to occur. Planning for these is often done through
the introduction of time and cost buffers in the estimates. Unknown unknowns however cannot
be anticipated and can be of varying nature.

The traditional planning and procedural methods to tackle these issues are often found wanting.
The site superintendent is often put under enormous pressures when these events transpire.
Hallgren and Wilson studied unknown unknowns for a construction company. Out of the fifteen
events documented, eight events occurred abruptly and seven were of the creeping nature. Nine
events arose out of contract disputes, one each out of fatality, guerilla attack and transport issues.
The remaining three were identified as miscellaneous causes. This research will explore the
behavior of construction workers and compares the performance of self-managed autonomous
crew to a centrally controlled crew. As mentioned in literature, the dynamic nature of the

68

construction project should point to a learning strategy in contrast to a rule based hierarchy for
better performance in managing uncertainty.

Under the traditional system, when an unforeseen situation arises (let us take the example of a
scaffold breaking down for the purpose of illustration) the crew would stop work and wait for
further directives. The superintendent/foreman is the central controlling entity. She/he is
generally called upon to find a solution to the problem at hand. She/he might take a decision
based on the perceived damage that he/she interprets. This diverts the superintendent from his
primary objective on site that is looking over the make ready plans and coordinating work. The
delay caused on the construction site due to this stoppage in work can cause a ripple effect
throughout the project if it is at a critical stage.

This research proposes a decision making model where the last planners (the foreman) who
perform the work take a decision based on their knowledge and foresight.

By giving the last planners the freedom to take decisions, it is possible that the situation which
could have been resolved or prevented quickly goes unseen. It is easy for a crew to bypass a
situation if they think they do not need to deal with it. The ‘not my job’ thinking is thus avoided
on the job. Moreover, an event that is unattended could easily spiral out into a full blown crisis
and cause catastrophic effects to the organization. Responsiveness has been identified as one of
the three pillars for success in response to unexpected events (Gerali et.al. 2009)

In this research, an environment simulating a construction site setting and agents representing the
construction workers on that site will be created. The environment is affected by various
uncontrollable constraints like size, location, type of project and weather conditions and also

69

subject to other internal factors such as trenches, equipment, worker diversity, safety culture, etc.
The environment is thus a complex combination of these factors. The agents on the other hand
are subject to movements in the environment. The agents move progressively in the
environment. The movement represents the completion of work. Thus when the agent starts in
the environment, the work done is zero and when it exits, the work is deemed complete. The
agents will have properties of age, culture, responsibilities and other personal attributes but more
importantly, they will be given the ability to learn from making decisions in case they encounter
an unforeseen event. The agents in the system will interact with each other and the environment
and learn from each other through these interactions. Various decision making cycles can be
tried out using this methodology. Decision cycles such as the OODA loop and PDCA are
suggested in literature to encounter these events (Abdelhamid et.al. 2009). The crews can be
trained in these decision making cycles to raise their respective individual competencies.

Geraldi 2009 mentions the three pillars for successful responses to unexpected situations. The
three pillars are responsiveness and functioning structure at the organizational level, good
interpersonal relations at the group level and finally competent people at the individual level.
The three pillars are in line with the literature that was studied in this research.

The individual level requirement for competent people has to come from the human resources
management. The employees should be hired and trained to execute their jobs more efficiently.
Continuing education opportunities to hone their skills can be helpful to improve the individual
competency.

The interpersonal relationships are an important part of any organization. Literature about
simultaneous management (Laufer et al. 1996) also brings forward the importance of
70

communication within the organization. The communication assumes an extremely important
role in project organizations. These organizations bring together teams skilled at certain tasks
and a general contractor who manages these teams. The teams are together for a short time i.e.
the duration of the construction project. Managing the communication between these teams
therefore is a difficult task and needs skillful construction managers to execute. The
superintendent is responsible for these teams and coordinating the work on construction sites.
The responsibility for solving unforeseen issues on construction sites falls on the superintendent
in traditional construction practices.

At the organizational level, Geraldi et al. mention the need for a responsive and flexible
structure. The current organization structures on construction projects do not afford the team this
flexibility and responsiveness needed to cope with unexpected situations. The superintendent
takes charge of coordination of work and also managing the production activities. The
superintendents are often overloaded with work. The management of these unexpected situations
take away time from important activities like look ahead planning and weekly work planning.
This in turn causes more unexpected events. The superintendent thus is the centrally controlling
body on the construction site.

The agent based model compares the centrally controlled construction site with a construction
site where there are autonomous and self-managed crews. The self-managed crews perform work
and take decisions to mitigate any unforeseen events that may take place on the construction site.
The superintendent will thus have more opportunities to manage the communication and
coordination of the work on site.

71

The model utilizes the learning strategies as outlined in Pich et al. (2002). The crew actively
makes decisions and this enhances their knowledge of the environment. As a result, they would
be better contributors in the next event that transpires. The model also aims to incorporate the
three pillars to success in meeting unexpected events as outlined by Geraldi et.al. (2009). The
modeling of competent people in the organization and good interpersonal relations are assumed
to be inherent. The responsiveness is tested by the model. The aim of including the
responsiveness element is incorporated in the construction crews to study the effect of
autonomous self-managed teams in tackling unexpected events. Thus the crew is autonomous
and operates on a learning strategy to tackle unexpected events.

The interaction and decision making in the crews fosters an environment where the tacit
knowledge is transferred within and between construction crews. The cumulative knowledge
within a system increases and the crew will be better equipped to face the next unforeseen
situation.

B) Model Development
The model is developed to analyze the workers’ and the superintendent’s behavior under
uncertain situations and how they deal with them. The model is developed using python version
2.7. The model has the following agents

1. Construction Site:
The construction site is a two dimensional grid of ‘0’ that represent the construction site. The
grid acts as a platform over which all of the active agents iterate. The site changes depending

72

upon the situation of the other agents in question. When initialized, the site comprises of all ‘0’s
and one unknown occurrence randomly generated to represent the unforeseen situation.

2. Construction Crew:
The construction crews are perhaps the most active agents in the model. They iterate over the
construction on each progressive step. As they move forward, they represent work performed on
the site. The crews have attributes of name, knowledge and size. The name is initialized as ‘1’.
Knowledge possessed by the crew is a value between 0 to 100 that determines the errors that
they produce while taking decisions.

The crew proceeds on the construction site in an orderly fashion until they face an unforeseen
situation. Each simulation step simulates the base case scenario, in which a traditional approach
is retained, i.e., the superintendent is the central controlling entity. The construction crew stops
work and call on the superintendent to solve the situation. A counter is started to keep track of
the amount of time elapsed before the work can proceed again. Each simulation step also
simulates the autonomous decision making crews’ response to the event. The crew takes a
decision without waiting for the superintendent to intervene. The superintendent is also called
however to see if the decision taken based on the crew’s knowledge is correct. In the second
case, there is no wait time associated with the decision. The delay is caused solely by the amount
of damage that has occurred.

3. Superintendent:

The superintendent is initialized as a dormant agent. It is called upon by the crew when they
encounter an unforeseen situation. Once called, the superintendent iterates over to the location on

73

the site where the issue is and takes a decision to solve the problem. The decision taken by the
superintendent in all cases is assumed to be the right one, i.e., the superintendent has 100%
knowledge. In the case where the crew takes active measures, the crews’ decisions are compared
to the superintendent’s decision to judge whether the crew made the right decision.

The Damage:

The damage is an unforeseen situation that occurs on a random location on the site. Damage has
a random attribute between 5 and 100 that tells us the extent of the damage. The higher the
number, the greater the damage. The damage is created when the site is initialized. The damage
triggers a decision from the construction crew in the model.

Assumptions

1. It is assumed that the construction crews are all using the same equipment and there are
equal chances for the equipment to malfunction for each crew
2. The cost of repairing and replacement is not considered in the model with the assumption
that in comparison between the two scenarios, the costs cancel out.
The Superintendent’s decision is absolutely correct

C) Initial Model Setup
The model uses an environment through which the construction crews (agents) iterate. The
default site is a ten by ten grid of the character ‘0’. The size of the site is customizable. A ten by
ten grid however is sufficient to obtain sizeable results and also manageable in terms of
computer processing power required. The value ‘0’ denotes a normal work condition in the
74

environment, i.e., the work proceeds smoothly without any negative impacts. The environment is
an abstract representation of a crew’s process on a construction project. It represents the
procurement, set up, carrying out of the work, and demobilizing of equipment. If we consider a
drywall crew, the first 10% of the iterations can be considered procurement and mobilization.
The following 80% can represent the installing of drywall in a sequential manner. The remaining
10% can be seen as inspections and demobilizing activities. Most construction activities are
linear in nature, i.e., they either progress in a horizontal or a vertical manner.

The agent objects are representations of the construction crew that would iterate over the
environment. The name is the character that would be placed in the environment and helps to
identify the agent during simulation.

When the program is run, the site is initialized with all values as ‘0’ to denote normal conditions.
Once the site is initialized, an unexpected event occurs at a random location with a random
magnitude that represents damage. The crew starts moving from the first position on site and
progressively forward until it encounters the unexpected event.

D) Model Execution
The agents or the construction crew iterates over the environment or the site as if going through
the construction process right from mobilization to demobilization. The values ‘1’ in the
environment point out the current location of the agents. In each time step, the agents move to
the next location if the next location has a normal working condition or a value of ‘0’. During the
iterations, if the agent hits a cell with value of over ‘5’ which denotes the location of the event,

75

the method superintendent is called in the base case to take a decision. In the autonomous crew’s
case, the agent takes a decision on its own based on the knowledge it possesses.

This method decides whether the agent will choose to repair the equipment or completely replace
it. The method uses two comparisons for deciding which action to take. A randomly generated
damage coefficient is checked to be less than 40. It would be impractical to replace the
equipment if the damage was less than 40%. The other comparison is between the agents risk
coefficient entered by the user and a randomly generated number. If both the conditions are
satisfied, the agent decides to repair the equipment, otherwise the equipment is replaced.

If the repair method is selected based on conditions, the method calculates the number of days
the project is likely to be delayed during repair. The method identifies the number of workers in
the crew and the damage occurred as the factors to calculate delay. The two values are multiplied
and delay is calculated.

The replace method has no factors affecting it since there is no cost considered in this model.
The assumption is that it would take 3 days for the equipment to be replaced. Thus, the delay in
replacing the equipment is calculated by multiplying the number of construction workers by 3.

The crews would each be delayed and the total delay of the project is calculated and printed.

E) Model Output
The entire process is repeated 100 times. The model exports results to an excel spreadsheet. The
outputs are as follows:

76

1. The current iteration number (tick)
2. The amount of delay caused by the crew taking the decision
3. The amount of delay caused by the superintendent taking the decision
4. The magnitude of damage that occurred
5. Whether the decision taken by the crew was right or wrong
6. The learning in the iteration
The model can be altered to run any number of iterations. Running the model 100 times gives us
a sufficient indicator of the results. The results stay consistent over greater amounts of iterations.

F) Model Parameters
The model has the following parameters:
1. The size of the grid
The size of the grid for this model is 10 X 10. This can be changed to test the hypothesis on
various sizes of the construction environment. The increase in size causes the delay caused due
to disruption to increase proportionately. A sensitivity analysis with the size of the grid found
that the results stay consistent.

However, on a very small construction site, the improvement due to decentralization is
negligible.

2. The decision parameter
The construction workers take a decision based on the magnitude of the damage. The parameter
in this study is kept at forty (40). The parameter controls the criteria for the workers’ decision. If

77

the parameter is kept at 40, the correct decision is to replace the equipment when damage is over
40%.

The parameter can have whole number values between 10 and 100.
3. Error rate of the workers
The error rate effectively controls the starting knowledge of the construction worker. In this
model, the error rate at the start is kept at 4. The error rate is a number between 0 and 10. The
error rate of 4 means that 4 out of every 10 decisions that the construction crew will take will
be wrong. The parameter is dynamic and decreases every time the construction worker takes
a wrong decision and ‘learns’ from it.

4. Learning factor for the workers
The learning factor controls the increase in the knowledge of the worker (or the decrease in the
error rate) each time the construction crew takes an incorrect decision. The value can be a
number between 0 and 0.99.

Each time the construction worker takes a wrong decision, the learning factor is multiplied by the
error rate to decrease the error rate an effectively increase the knowledge of the agents.

78

CHAPTER IV: RESULTS

79

A) Introduction
Chapter 5 provides an overview of the research conducted, and the major findings. The relevance
of the results to the construction industry is discussed. This chapter also discusses the limitations
of the research.

B) Research Overview
This research focused on understanding the crew behavior under uncertain situations. A thorough
literature review was conducted in the fields of construction and project management to
investigate the current methods suggested to minimize and mitigate uncertainty on a construction
site. An exploration into the crew behavior during these uncertain times provided a better
understanding of how we can effectively tackle these events in the future.

Chapter 1 presented the need and the significance of the research. Studies have been conducted
that have shown that the unexpected events indeed have a large impact on the construction
operations. Chapter 1 also outlined the goals and the objectives of the research.

Chapter 2 provided a literature review of the current practices in construction and project
management to tackle unexpected events. The chapter outlined literature in the fields of
management of uncertainty, project management, Lean Construction and Agent based modeling.
The literature review in these topics provided the needed background to carry out the research.

80

Chapter 3 described the research methodology for this research. The chapter explained the
relevance of the studied literature in construction management and managing the unexpected in
construction. An agent based model was developed to simulate the occurrence of an unexpected
event on a construction site. The model helped understand the behavior of the construction
worker under uncertain conditions. A parallel scenario was simulated where the construction
teams were self-managed and autonomous, in which the construction teams decided to take
decisions and rectify the issues they are faced with. As a result, the crews gained in knowledge
as they took decisions. Through this agent based model and the literature reviewed, the behavior
at the crew level was better understood when faced with an unexpected event. Chapter 4
presented the findings of the research and the relevance of the findings to the construction
industry. This chapter will put forward the learning from the model and literature and state its
relevance in today’s complex construction industry.

81

C) Research Goals and Objectives
The overall goal of this research was to understand the behavior of construction crews under
unexpected events. One of the primary goals of the thesis was also to document the current
approaches to unknown unknowns in construction and the strategy to deal with them on the
construction site.

To study the current practices dealing with unknown unknowns in construction, a literature
review was conducted in the fields of construction management, Lean Construction, managing
uncertainty in project management and Agent Based Modeling.

Agent Based Modeling was chosen as the research method to develop a model that allowed
investigating the crew behavior on a construction site. Agent based modeling is a versatile tool
that has been used in construction to study the safety and performance of supply chain among
other things. This modeling approach allowed creating construction crews that interacted and
took decisions. As a result of the interaction, there was learning that took place in the crew and
they were better equipped to make the next decision.

D) Literature Review Results
The literature review brought forward the fundamental strategies to tackle unexpected situations.
Studies propagate planning in greater details, planning of the production process, selecting a
capable and responsive team among other measures. In construction, we are often faced with
extremely dynamic and uncertain projects. To tackle these uncertainties, we need to embrace the
fact that they will happen rather than trying to avoid them altogether.

82

The literature also showed the need for a responsive, flexible and autonomous production team
to handle uncertainty. Geraldi (2009) explains the three pillars of successful response to
unknown unknowns. On the construction sites, these pillars translates into providing the last
planners with a degree of freedom to innovate and creatively solve problems. The construction
crews should also be autonomous and be able to quickly decide and implement actions. At the
group level, good communications need to be maintained. At the individual level, we need to
develop competencies in the employees. The Human resources department at the companies has
to draw guidelines which must be followed when recruiting new employees. Leaders should be
developed from within the organizations who lead the team to deal with stressful situations.

These learning points from the literature were applied in the model to test the behavior of
construction crews under traditional project organizations. These results were compared to the
results of an autonomous and self-managed crew faced with an unexpected event. The results are
presented in the next section of this chapter.

E) Modeling Results
The results from the model based on 100 runs were analyzed. The first measure of interest was
the amount of delay caused while the crews made the decisions and the corresponding delay
when the superintendent took the decision on how to proceed. The equipment damage ( the
assumed uncertain condition encountered) was constant for both simulated instances where the
superintendent takes the decision and where the crews take the decision.

83

250

Crew VS Superintendent Delay

200

Delay

150

100

50

0
0

20

40

60
Damage

80

100

120

Figure 9: Crew Delay VS Superintendent Delay (For interpretation of the references to color in this
and all other figures, the reader is referred to the electronic version of this thesis.)

Figure 11 illustrates how the data compares when the delay caused by the damage is plotted
against time. The blue plots represent the delay caused by the crew agent taking the decision and
the red plots are the delay caused when the superintendent taking the decision.

The figures indicate that, as modeled in this research, the decision making by the self-managed
team outperformed the team that depended on external agents for making decisions. The delay
caused by the same damage was on average 40% lower for the crew making its own decisions
compared to the crew that depended on the superintendent. Considering that the model

84

accounted for erroneous decisions, whether made by the superintendent or the crew itself, the
prima facie result shows that allowing crews to be autonomous is an effective strategy on the
long run.

The model assumed that the crew started off with a knowledge coefficient of seventy (70) on a
scale of 1 to 100. Learning took place every time the crew made a decision. The knowledge of
the crew increased each time they made decisions. This is based on the observations of Kim
(1998). Kim observed that there are two kinds of knowledge in any organization.

The first kind of knowledge is the explicit knowledge possessed by the company -the manuals,
training programs, and processes. The more subtle and perhaps more important kind of
knowledge with regard to tackling known or unknown unknowns is the tacit knowledge owned
by the organization. Tacit knowledge is much less accessible. Tacit knowledge is gained through
experience, observation and practice. The construction team in the model learns through these
principles. Tacit knowledge in the team increases every time they take a decision and solve a
problem.

When a problem is solved incorrectly, the crew gains in knowledge. In the case of making a right
decision, the crew gains knowledge only applying the tacit knowledge that they already possess.
There is learning in this case among the team members but to a lesser degree. Figure 12 shows
the increase in knowledge of the crews as they made decisions throughout the simulations. The
reduction in the error rate over time (number of runs) matches real-life behavior, further giving
support the model was reasonably realistic.

85

Error Rate in Workers' Decisions
40

Error Rate (%)

35
30
25
20
15
10

5
1
5
9
13
17
21
25
29
33
37
41
45
49
53
57
61
65
69
73
77
81
85
89
93
97

0

Iterations

Error Rate

Figure 10: The increase in knowledge of the crew

A number of the lean methods and principles are aimed at reducing variation on the construction
site. This helps make the project more predictable and easier to manage. Reducing variation is a
fundamental concept in production planning (The Last Planner® System). Figure 13 shows the
delay caused against the damage occurred for the superintendent taking the decisions. The data is
very varied. For example, for a damage of magnitude 60, the delay caused can be anywhere
between the range of 60 and 160 units which is an extremely large variation. The variation is
mostly caused due to the location of the unforeseen situation with respect to the location of the
superintendent. The superintendent takes time to iterate over to the location of the damage and
subsequently solves the problem. This is a representation of reality to some extent where the

86

superintendent might be busy with other responsibilities. The unforeseen situation thus might not
be top priority for the superintendent. The negative by product is also in that the superintendent
who should be focusing on coordination of work and planning of work is sidetracked due to
these situations arising.

Damage VS Delay (Superintendent)
120
100

Damage

80
60

40
20
0

0

50

100

150
Delay

Figure 11: Superintendent (Damage VS Delay)

87

200

250

Damage VS Delay (Crew)
120
100

Damage

80

60
40
20
0
0

20

40

60
Delay

80

100

120

Figure 12: Construction crew taking decisions (Delay VS Damage)
Figure 14 shows the delay generated against damage occurred when the self-managed crew takes
a decision. The plot is almost a straight line as derived from the model results. This is because
there is no wait time associated with the decision making. The agents encounter the problem and
instantaneously take a decision to rectify it. The decision has a possibility of being wrong
causing a negative impact. The straight line is a modeling result and will not likely be observed
in reality. It is however fair to conclude that the variation in this case is a great deal smaller than
in the previous case. When the agent encounters the event, the foreman perceives damage and
takes active measure to correct it. The decision adds to the agent’s knowledge through learning
and he/she is better equipped to tackle the next occurrence.

88

Thus we observe that based on model results, there is less varation in the decision making
process amongst the self-managed crew and there is savings in terms of delay caused due to
damage occurring. The error rate that exists is expected to subside dramatically at first as there
are wrong decisions made more frequently and a lot of learning occurs. As time passes and more
exposure is given to the crew, the knowledge in the agents will build up and error rate will fall.

F) Research Findings
A thorough literature review was conducted to understand the current practices used in handling
unexpected events in construction. The findings from the literature review were as follows:

1. The occurrences of unexpected events is inevitable and we must prepare to combat them
The literature review has shown that unexpected events are inevitable due to the increasing
complexity on construction projects today and the highly interactive nature of our construction
projects. It is thus imperative that we build capabilities in our construction project teams to
combat these uncertainties. These unexpected events when mismanaged can lead to far reaching
damages to the project plan. The events also have the potential to cascade into a full blown crisis.

2. Lean Construction offers various tools that help to minimize the occurrences and impacts
of unexpected events
Lean Construction offers various tools and techniques to ensure that the issues that could arise on
construction projects are analyzed and eliminated early in the project. Elimination of these issues
will help reduce unexpected events on site. Techniques such the Last Planner® System help
improve the communication and organization on construction sites thereby exposing any issues

89

in real time and can be eliminated before they cause any disruptions. Lean Construction also
propagates the use of the workers’ knowledge and experience which is in line with what this
thesis proposes.

The findings from the literature review were also used in the agent based model developed to
test the two cases of interest in this research. Based on the assumptions of the model and after
analysis of the model results, we find that:

1. A decentralized system appears to help the project to recover from unknown unknowns
more effectively
A decentralized system where the crew and the construction workers are empowered to take
decisions and help to mitigate encountered disruptions, as modeled in this research, appears to be
an effective method to combat unexpected events. The modeling showed that the unexpected
events are found and dealt with earlier in the decentralized system. The shorter time taken,
compared to waiting for the superintendent to help, to resolve disruptions had a positive effect on
the project completion time. The delay caused by the disruptions was found, in the model, to be
approximately 40% lower when the system of decision making was decentralized.

2. A responsible and flexible organization is helpful in recovery from unknown unknowns
The model results indicate that an organization that is responsive in reacting to disruptions was
more effective in its mitigation. Prior research in Lean Construction agrees that the flexibility of
construction site operations can often be increased by using multi-functional crews that are
capable and knowledgeable in various trades. The use of cheetah teams have also been found to
be useful in mitigating unexpected events.

90

3. Superintendent was under much less pressure in a system where the crews were
autonomous and self-managed
In the traditional approach modeled in this research, the superintendent was responsible for
solving all issues that arose on site. In the decentralized system, every worker and crew was
actively working to keep the operations on the construction site on track. It can be logically
inferred here that this should reduce the workload on the site superintendent. In turn, this allows
spending time on coordination of the construction activities and planning activities that are part
of the Last Planner® System.

Again, based on the prima facie results of the agent based model implemented in this research,
using a system where workers take responsibilities and decisions to help manage the project
appears to the utilize workers’ knowledge and experience better. This may give the workers a
sense of more involvement in the project as well as possibly aligning the crews’ interest with that
of the project.

G) Research Limitations
The research shows us through the model that disruptions on construction sites can in fact be
better managed through maintaining a self-managed and autonomous construction team.
However, the research is subject to the following limitations:

1. Learning in the construction teams is non-linear in nature. The construction workers gain
knowledge each time they make a wrong move. The learning in the model is a direct
factor of the knowledge gain by the crew, which was assumed to be linear.

91

2. The research tested the behavior of a single construction crew at a time. The model tested
the results for a single superintendent and a single autonomous construction crew. The
model may be expanded to examine behavior emergence in case of multiple crews and
superintendents.
3. The model does not consider other aspects that affect the crew behavior such as culture,
traditions, etc. These are eliminated from the scope of the model to maintain parsimony
of results and to analyze the effect of the unit of analysis (the crew making decision
versus the superintendent making them).

H) Conclusions
Significant amounts of time and energy on construction projects is spent on managing
uncertainty. Due to the tight coupling of activities in construction, the present means of planning
and management do not help this cause. Projects are planned in advance using unrealistic
schedules that are pushed on to the subcontractors. The Construction Management Association
of America conducted an owner survey that notes that according to owners, between 40 – 50 %
of construction projects are behind schedule and / or over budget (6th CMAA Owners Survey).
This indicates that the current planning and management projects for construction projects is
lacking.

A study conducted in Florida notes that 4.8% of the construction cost was overruns due to
unavoidable conditions. There is little consideration given to these situations in construction
management. Identifying and eliminating them will result in considerably more reliable projects.
The burden of solving these issues is often placed on the site superintendent. This study focused
on exploring how these events can be addressed.
92

This thesis advocated a progressive approach where the construction crews are trusted to make
decisions and take responsibility for solving these problems as and when they occur. Through an
approach of having a responsive and flexible structure, trust in the competencies of the workers
and good leaders, we can tackle these situations effectively.

The strategy illustrated here is not to avoid changes and uncertainty but to embrace it and prepare
a project team ready to tackle it. Through literature and an agent based model, a strategy of
maintaining self-managed, learning and autonomous construction crews is advocated to tackle an
unexpected event. The dynamic nature of construction makes it suitable for such measures rather
than the traditional rule based strategy.

Empowerment of the workers to make decisions has been successfully used in various industries
for quick and accurate executions, typically giving advantage over competitors or an adverse
situation. Hyundai Motor Company fabricates crises in the organization as a means to build
resilience against these unforeseen events. Through the literature review and the agent based
model, we find that uncertainty was better tackled through an approach of autonomy and
responsive decision making.

The model used in this research incorporated learning characteristics in the construction crews to
analyze how they make decisions and how it affected delays under an unforeseen situation. The
results show that when the construction crews make decisions based on their knowledge and
within reasonable boundaries, the delay caused can decrease over time and there is more
reliability in the process.

93

As a result of the crews solving problems, the superintendent is also expected to have more
opportunities to concentrate on improving the coordination and planning of work on site. In
closing, it is expected that conditions where crews are given autonomy to make decision will
result in a healthier and more satisfying experience for the workers and the organization.

94

APPENDIX

95

A) Model Script
The following script is written in Python Programming Language version 2.6. Python is an open
source programming language.

##### ANKUR DESAI, MICHIGAN STATE UNIVERSITY ### ###
##### MS CONSTRUCTION MANAGEMENT #########
##### MODEL TO EXPLORE CREW BEHAVIOR ######
##### DEEMBER 9TH 2011 #####################

import numpy,random,copy,time,csv
totaloutput = []
tick = 0
knowledge = 4
wrongcount = 0

while tick <100:

tick = tick + 1
print "#############################################", tick
### Setup Construction Site ###
size = 10

96

row = 10
end = 0

site = numpy.zeros ((size,size), int)

totaldely = 0
string = []
Delay = 0

class agents:

def __init__(self, name, k, crew, knowledge):
self.name = name
self.k = k
self.damage = 0
self.dec = ""
self.know = knowledge

def __str__(self):
grid = site
return grid

97

def move_agents (self,agent):
print agent.name
tradestring = agent.name

for i in range(size):
for j in range(size):
if site [i,j] == 0:
site [i,j] = agent.name
##

print site, "\n\n\n"
##

time.sleep(0.75)
site[i,j] = 0

else:

##

site[i,j] = agent.name
agents.met_damage(agent,i,j, site)
damage = site.max()
agents.decision (agent, Delay, damage)

def met_damage(agent, row, colum,site):

98

site[0,0] = super1.name
print site, "\n\n\n"
site [0,0] = 0
Delay = 0
for i in range(size):
for j in range(size):
if site[i,j] < 4:
site [i,j] = super1.name
##

print site, "\n\n\n"
##

time.sleep(0.75)
site [i,j] = 0
Delay = Delay +1

##

print Delay
end = 1
else:
##

agents.decision(Delay, agent, e1)
damage = site.max()
agents.superdecision(agent, Delay, damage)

##

def know (agent):

##

knowledge = 4

##

newknowledge = newknowledge*0.9

99

##

##

print 'new knowledge =', knowledge

knowledge = agen1.knowledge
def decision(agent, Delay, workerdamage):

wrongcount = 0
if workerdamage > 40:
agentdecision = "Replace"
if random.randint(0,10)> agen1.know:
dec = "Right"

else:
dec = "Wrong"
agen1.know = agen1.know*0.95
print "Construction Worker will REPLACE the damage"
print "\nThe amount of delay caused is 0 but the decision is:", dec
print "\nThe cost of the delay is: ", workerdamage
agen1.damage = workerdamage
agen1.dec = dec

else:

100

agentdecision = "Repair"
if random.randint(0,10)> agen1.know:
dec = "Right"

else:
dec = "Wrong"
agen1.know = agen1.know*0.95

print "Construction Worker will Repair the damage"
print "\nThe amount of delay already occured is 0 but the decision is: ", dec
print "\nThe cost of the delay is: ", workerdamage
agen1.damage = workerdamage
agen1.dec = dec

def superdecision (agent,Delay, damage):
if damage >50:
superdecision = "Replace"
print "The superintendent decided to REPLACE the damage"
print "\nThe amont of delay already occured is ", Delay
print "The cost of the delay is: ", damage+Delay

101

super1.damage = damage+Delay

else:
superdecision = "Repair"
print "The Superintendent decided to Repair the damage"
print "\n The amount of delay already occured is ", Delay
print "The cost of the delay is: ", damage+Delay
super1.damage = damage+Delay

##
##
##

while i>row:
for j in range(size):
if site[i,j] == 0 or site [i,j] == "1":

##
##

site[i,j] = super1.name

##

print site, "\n\n\n"

##

time.sleep(0.75)

##

site [i,j] = 0

##
##

for j in range(colum):
if site[i,j] == 0 or site[i,j] == "1":

##

site[i,j] = super1.name

##

print site, "\n\n\n"

##

time.sleep(0.75)

102

##

site [i,j] = 0

##

def calculatedamage(self):
pass

class superintendents:

def __init__(self, name, k):
self.name = name
self.k = k
self.damage = 0

## def move_superintendent(superintendent):###PLACE SUPERINTENDENT AND
MOVE SUPERINTENDENT####
##
##
##
##

for i in range(size):
for j in range(size):

##

random_row = random.randint(0,size-1)

##

random_colum = random.randint(0,size-1)

103

##

site[random_row,random_colum] = super1.name

##

print site

##

site[random_row,random_colum] = 0

def call_superintendent(row, colum, k):
##random_row = random.randint(0,size-1)
##random_colum = random.randint(o,size-1)
##site[random_row,random_colum] = superintendentstring
site [0,0] = super1.name
print site, "\n\n\nSUPERINTENDENT IS NOW MOVING TOWARDS THE
DAMAGE\n\n\n"

for i in range(size):
while i<row:
for j in range(size):
if site[i,j] == 0 or site[i,j] == "1":
site [i,j] = superintendent.name
##

print site, "\n\n"
time.sleep(.75)
site[i,j] = 0

104

class unknown_event():

def __init__(self):
self.damage = random.randint(5,100)
print "Damage Magnitude = ", self.damage
damagerow = random.randint(0,size-1)
damagecolum = random.randint(0,size-1)
site[damagerow, damagecolum] = self.damage
print "\n\n\n\n\n\n\n","DAMAGE HAS BEEN PLACED\n", site,"\n\n\n\n\n"
agents.move_agents(agen1,agen1)
self.damagerow = damagerow
self.damagecolum = damagecolum

class Results:

def __init__(self):
pass

105

###################################
###############################

INITIALIZE

agen1 = agents("1", 0.75, 60, knowledge)
super1 = superintendents("2", 1)
e1 = unknown_event()

output = (tick, e1.damage, super1.damage, agen1.damage, agen1.dec, agen1.know)
totaloutput.append(output)
knowledge = agen1.know

###################################
##############################

WRITE RESULTS

w = csv.writer(open('output.csv', 'w'))
w.writerows(totaloutput)

106

B) Excel Output from the model
The model when run, exports the results into a Microsoft Excel worksheet. The table below
shows the results for a run of the model for 100 such events. The starting error rate is kept at
40%.

Table 3: Output from the AB model developed

Iteration Damage
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29

13
67
36
22
37
97
67
17
68
53
68
70
26
76
97
22
47
71
97
54
90
88
41
42
35
41
61
72
25

Super’s Workers Workers
Decision Decision Decision
65
13 Wrong
124
67 Wrong
103
36 Right
23
22 Wrong
62
37 Right
127
97 Right
152
67 Right
26
17 Right
153
68 Right
56
53 Wrong
121
68 Wrong
108
70 Right
99
26 Right
105
76 Right
101
97 Right
74
22 Right
87
47 Right
91
71 Wrong
141
97 Right
143
54 Right
186
90 Right
151
88 Right
81
41 Right
128
42 Right
113
35 Right
114
41 Right
134
61 Wrong
144
72 Right
72
25 Right

107

Error
Rate
3.8
3.61
3.61
3.4295
3.4295
3.4295
3.4295
3.4295
3.4295
3.258025
3.095124
3.095124
3.095124
3.095124
3.095124
3.095124
3.095124
2.940368
2.940368
2.940368
2.940368
2.940368
2.940368
2.940368
2.940368
2.940368
2.793349
2.793349
2.793349

Error
Rate %
38
36.1
36.1
34.295
34.295
34.295
34.295
34.295
34.295
32.58025
30.95124
30.95124
30.95124
30.95124
30.95124
30.95124
30.95124
29.40368
29.40368
29.40368
29.40368
29.40368
29.40368
29.40368
29.40368
29.40368
27.93349
27.93349
27.93349

Table 3 continued
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72

95
56
36
73
63
51
74
10
70
60
5
71
28
47
38
13
24
5
50
45
82
41
68
94
33
9
10
86
97
76
13
26
30
71
45
99
53
44
75
89
11
58
41

164
123
43
117
153
108
120
60
152
95
63
165
90
114
56
99
65
59
114
100
122
62
77
102
131
87
69
128
160
86
50
101
104
144
81
182
144
56
111
112
45
85
73

95
56
36
73
63
51
74
10
70
60
5
71
28
47
38
13
24
5
50
45
82
41
68
94
33
9
10
86
97
76
13
26
30
71
45
99
53
44
75
89
11
58
41

108

Wrong
Right
Wrong
Right
Wrong
Right
Right
Right
Wrong
Right
Wrong
Wrong
Right
Wrong
Wrong
Right
Right
Right
Right
Right
Right
Right
Wrong
Right
Right
Right
Right
Right
Right
Right
Wrong
Right
Right
Right
Right
Right
Right
Right
Right
Right
Right
Right
Right

2.653682
2.653682
2.520998
2.520998
2.394948
2.394948
2.394948
2.394948
2.2752
2.2752
2.16144
2.053368
2.053368
1.9507
1.853165
1.853165
1.853165
1.853165
1.853165
1.853165
1.853165
1.853165
1.760507
1.760507
1.760507
1.760507
1.760507
1.760507
1.760507
1.760507
1.672481
1.672481
1.672481
1.672481
1.672481
1.672481
1.672481
1.672481
1.672481
1.672481
1.672481
1.672481
1.672481

26.53682
26.53682
25.20998
25.20998
23.94948
23.94948
23.94948
23.94948
22.752
22.752
21.6144
20.53368
20.53368
19.507
18.53165
18.53165
18.53165
18.53165
18.53165
18.53165
18.53165
18.53165
17.60507
17.60507
17.60507
17.60507
17.60507
17.60507
17.60507
17.60507
16.72481
16.72481
16.72481
16.72481
16.72481
16.72481
16.72481
16.72481
16.72481
16.72481
16.72481
16.72481
16.72481

Table 3 continued
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100

81
18
10
85
61
93
51
16
10
59
91
19
47
97
74
45
16
47
11
88
78
17
31
82
22
52
22
64

158
57
14
121
84
139
90
110
63
94
145
0
131
0
162
134
67
71
62
183
144
48
124
102
26
149
77
132

81
18
10
85
61
93
51
16
10
59
91
0
47
0
74
45
16
47
11
88
78
17
31
82
22
52
22
64

109

Right
Right
Right
Right
Right
Right
Right
Right
Right
Right
Wrong
Right
Right
Wrong
Right
Wrong
Right
Right
Right
Right
Right
Right
Right
Right
Right
Right
Right
Right

1.672481
1.672481
1.672481
1.672481
1.672481
1.672481
1.672481
1.672481
1.672481
1.672481
1.588857
1.588857
1.588857
1.509414
1.509414
1.433944
1.433944
1.433944
1.433944
1.433944
1.433944
1.433944
1.433944
1.433944
1.433944
1.433944
1.433944
1.433944

16.72481
16.72481
16.72481
16.72481
16.72481
16.72481
16.72481
16.72481
16.72481
16.72481
15.88857
15.88857
15.88857
15.09414
15.09414
14.33944
14.33944
14.33944
14.33944
14.33944
14.33944
14.33944
14.33944
14.33944
14.33944
14.33944
14.33944
14.33944

C) Raw output from the model
The following shows the results from the model running in the Python console. The results show
the first 10 iterations of a run of the model.

############################################# 1
Damage Magnitude = 9

DAMAGE HAS BEEN PLACED
[[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 0 0 0 0]

110

[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 9 0 0 0]]

1
[[2 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 0 0 0 0]
[0 0 0 0 0 0 9 0 0 0]]

The Superintendent decided to Repair the damage

111

The amount of delay already occured is 96
The cost of the delay is: 105
Construction Worker will Repair the damage

The amount of delay already occured is 0 but the decision is: Wrong

The cost of the delay is: 9
############################################# 2
Damage Magnitude = 21

DAMAGE HAS BEEN PLACED
[[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]

112

[ 0 0 0 0 0 0 0 0 0 0]
[21 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

1
[[ 2 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[21 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

113

The Superintendent decided to Repair the damage

The amount of delay already occured is 50
The cost of the delay is: 71
Construction Worker will Repair the damage

The amount of delay already occured is 0 but the decision is: Right

The cost of the delay is: 21
############################################# 3
Damage Magnitude = 48

DAMAGE HAS BEEN PLACED

114

[[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[48 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

1
[[ 2 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[48 0 0 0 0 0 0 0 0 0]

115

[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

The Superintendent decided to Repair the damage

The amount of delay already occured is 50
The cost of the delay is: 98
Construction Worker will REPLACE the damage

The amount of delay caused is 0 but the decision is: Wrong

The cost of the delay is: 48
############################################# 4
Damage Magnitude = 87

116

DAMAGE HAS BEEN PLACED
[[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 87 0 0 0 0 0 0 0 0]]

1
[[ 2 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]

117

[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 87 0 0 0 0 0 0 0 0]]

The superintendent decided to REPLACE the damage

The amont of delay already occured is 91
The cost of the delay is: 178
Construction Worker will REPLACE the damage

The amount of delay caused is 0 but the decision is: Right

The cost of the delay is: 87
############################################# 5
Damage Magnitude = 16

118

DAMAGE HAS BEEN PLACED
[[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 16 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

119

1
[[ 2 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 16 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

The Superintendent decided to Repair the damage

The amount of delay already occured is 44
The cost of the delay is: 60
Construction Worker will Repair the damage

The amount of delay already occured is 0 but the decision is: Right

120

The cost of the delay is: 16
############################################# 6
Damage Magnitude = 38

DAMAGE HAS BEEN PLACED
[[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 38 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

121

1
[[ 2 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 38 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

The Superintendent decided to Repair the damage

The amount of delay already occured is 15

122

The cost of the delay is: 53
Construction Worker will Repair the damage

The amount of delay already occured is 0 but the decision is: Wrong

The cost of the delay is: 38
############################################# 7
Damage Magnitude = 23

DAMAGE HAS BEEN PLACED
[[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 23 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]

123

[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

1
[[ 2 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 23 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

124

The Superintendent decided to Repair the damage

The amount of delay already occured is 35
The cost of the delay is: 58
Construction Worker will Repair the damage

The amount of delay already occured is 0 but the decision is: Right

The cost of the delay is: 23
############################################# 8
Damage Magnitude = 65

DAMAGE HAS BEEN PLACED
[[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]

125

[ 0 0 0 0 0 0 0 0 0 0]
[65 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

1
[[ 2 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[65 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]

126

[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

The superintendent decided to REPLACE the damage

The amont of delay already occured is 30
The cost of the delay is: 95
Construction Worker will REPLACE the damage

The amount of delay caused is 0 but the decision is: Right

The cost of the delay is: 65
############################################# 9
Damage Magnitude = 91

127

DAMAGE HAS BEEN PLACED
[[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 91 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

1
[[ 2 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]

128

[ 0 0 0 91 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

The superintendent decided to REPLACE the damage

The amont of delay already occured is 43
The cost of the delay is: 134
Construction Worker will REPLACE the damage

The amount of delay caused is 0 but the decision is: Wrong

The cost of the delay is: 91
############################################# 10
Damage Magnitude = 43

129

DAMAGE HAS BEEN PLACED
[[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[43 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

1

130

[[ 2 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]
[43 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0]]

The Superintendent decided to Repair the damage

The amount of delay already occured is 80
The cost of the delay is: 123
Construction Worker will REPLACE the damage

The amount of delay caused is 0 but the decision is: Wrong

The cost of the delay is: 43

131

REFERENCES

132

REFERENCES
Abdelhamid T.S., Schafer Don, Mrozowski Tim, Jayaram V., Howell G., Mohamed El-Gafy
(2009) “Working Through Unforeseen Uncertainties through OODA Loop” Proceedings
of the 17th Annunal conference for the International Group of Lean Construction.
Betts, Martin, and George Ofori. "Strategic planning for competitive advantage in
construction." Construction Management & Economics10, no. 6 (November 1992):
511. Business Source Complete, EBSCOhost (accessed October 17, 2011).
Charles Perrow (1999) “Organizing to Reduce the Vulnerabilities of Complexity”Journal of
Contingencies and Crisis Management. Volume 7, Issue 3 pages 150-155. September
1999
Collyer Simon,Warren
Clive M.J., “Project management approaches for dynamic
environments”, International Journal of Project Management, Volume 27, Issue 4, May
2009, Pages 355-364,
Cynthia C. Y. Tsao, Iris D. Tommelein, Eric S. Swanlund and Gregory A. Howell, “Work
structuring to Achieve Integrated Product-Process Design.” Journal of Construction
Engineering and Management November/December 2004 DOI: 10.1061/(ASCE)07339364(2004)130:6(780)
Engwall Mats and Svensson Charlotta ‘Cheetah Teams.’ Harvard Business Review May 2009.
Equifax
Business
Failure
Report
2011
(http://www.equifax.co.uk/Aboutus/Press_releases/2011/EQUIFAX_2011_BUSINESS_FAILURES_REPORT.html)
accessed on 15th October 16, 2011
FMI/CMAA
Sixth
Annual
Survey
(http://cmaanet.org/user_images/owners_survey_6.pdf)

of

Owners

Geraldi Joanna G. “Reconciling Order and Chaos in Multi-project Firms.”, Internation Journal of
Managing Projects in Business. Vol. 2 No. 1 2009, pp. 149-158
Grey Chris and Garsten Christina, “Trust, Control and Post Bureaucracy”,Organization Studies,
Vol. 22, No. 2 March 2001.
Herman Glenn Ballard, “The Last Planner System of Production Control.” Doctoral Thesis
submitted to the Faculty of Engineering, The University of Birmingham.
Howell G. (1999) “White Paper for Berkley/Stanford CE & M Research Workshop”

133

Hwang, Peter, and J. David Lichtenthal. "Anatomy of Organizational Crises." Journal of
Contingencies & Crisis Management 8, no. 3 (September 2000): 129. Business Source
Complete, EBSCOhost (accessed October 20, 2011).
Joana G. Geraldi, Liz Lee-Kelley, Elmar Kutsch “The Titanic sunk, so what? Project manager
response to unexpected events”, International Journal of Project Management, Volume
28, Issue 6, August 2010, Pages 547-558
Keesoo Kim, Boyd C. Paulson 2003 “Agent-Based Compensatory Negotiation Methodology to
Facilitate Distributed Coordination of Project Schedule Changes”, Journal of Computing
in Civil Engineering. Vol.17, No. 10, (2003).
Kloppenborg, Timothy J., and Warren A. Opfer. "The Current State of Project Management
Research: Trends, Interpretations, and Predictions." Project Management Journal 33, no.
2 (June 2002): 5. Business Source Complete, EBSCOhost (accessed October 17, 2011)
Koskela Lauri. “We need a theory for construction.” (1992)
Laufer Alexander, Denker Gordon R., Shenhar Aaron J., “Simultaneous Management: The key
to excellence in capital projects.” International Journal of Project Management Vol. 14,
No. 4, pp. 189-199, 1996
Liker J.K. “The Toyota Way” McGraw-Hill; 1 edition (December 17, 2003)
Markus Hällgren, Eva Maaninen-Olsson, (2009) "Deviations and the breakdown of project
management principles", International Journal of Managing Projects in Business, Vol. 2
Iss: 1, pp.53 – 69
Markus Hällgren, Timothy L. Wilson, (2007) "Mini-muddling: learning from project plan
deviations", Journal of Workplace Learning, Vol. 19 Iss: 2, pp.92 – 107
Markus Hällgren, Timothy L. Wilson, “The nature and management of crises in construction
projects: Projects-as-practice observations”, International Journal of Project
Management, Volume 26, Issue 8, November 2008, Pages 830-838
Mcmanus, H.; Hastings, D.; , "A framework for understanding uncertainty and its mitigation and
exploitation in complex systems" Engineering Management Review, IEEE , vol.34, no.3,
pp.81,
Third
Quarter
2006
doi: 10.1109/EMR.2006.261384
Palaniappan, A. Sawhney, M.A. Janssen, K.D. Walsh 2007 “Modeling Construction Safety as an
Agent Based Emergent Phenomenon”, 24th International Symposium on Automation and
Robotics in Construction (ISARC 2007). Construction automation group I.I.T., Madras.

134

Pich Micheal, Loch Christoph and De Meyer Arnoud “On Uncertainty, Ambiguity and
Complexity in Project Management”, Management Science, Vol. 48, No. 8, August 2008.
Pages 1008-1023
Polanyi M. “The tacit dimension.’ Library of Congress Catalog Card number Ó6-21015
©Micheal Polanyi
Raelin Joseph A. “The End of Managerial Control?” Group & Organization Management April
2011 36: 135-160, first published on December 31, 2010
Reid Janine L. “Crisis Management: Planning and media relations for the Design and
Construction Industry”, Wiley; 1 edition (March 1, 2000)
Richard J. Curts and Douglas E. Campbell (2001)“Avoiding Information Overload through the
Understanding of OODA loops, A Cognitive Hierarchy and Object-Oriented Analysis
and Design” ©Curts and Campbell 2001
Sawhney, A. Bashford, H. Walsh, K. Mulky, A. R. “Agent Based Modeling and Simulations in
Construction” Winter Simulation Conference 2003. Volume 2 pages 1541-1547
Simons Robert (1995) “Control in the Age of Empowerment” Harvard Business Review (1995)
Snell Scott A. “Control Theory in Strategic Human Resource Management: The Mediating
Effect of Administrative Information.” The Academy of Management Journal Vol. 35,
No. 2 (Jun., 1992), pp. 292-327
Susan M. Sanchez and Thomas W. Lucas. 2002. “Exploring the world of agent-based
simulations: simple models, complex analyses.” In Proceedings of the 34th conference on
Winter simulation: exploring new frontiers (WSC '02). Winter Simulation Conference
116-126.
Turcotte John W., Director of Florida Department of Transport “Review of the Florida
Department of Transport’s Performance in Controlling Construction Cost Overruns and
Establishing Accountability for These Problems.” November 27, 1996.
Wood, N. (2004). “Unknown unknowns: knowledge elicitation for multimedia in craft learning.”
In: Challenging Craft: International Conference, Gray's School of Art, Aberdeen,
Scotland, 8-10 September 2004

135