Design risk modelling and analysis: Case on Addis-Djibouti railway

construction project

1
Tesfaye Gashaw, 2Kassu Jilcha

1
African Railway Centre of Excellence, Institute of Technology, Addis Ababa University, Addis
Ababa, Ethiopia

2
School of Mechanical and Industrial Engineering, Institute of Technology, Addis Ababa
University, Addis Ababa, Ethiopia

1
tesgass@yahoo.com, 2jkassu@gmail.com

Abstract
The purpose of this paper is to develop a design risk analysis model and to analyze the effect of
these risks on railway construction project time and cost overruns. To assess the impact of design
related risks on the project objectives, questionnaires were distributed to twenty-two client experts
of railway construction projects. An integrated fuzzy-System dynamics approach was utilized to
assess the impact of the risks on project objectives. The fuzzy logic was used to handle the
uncertainty, vagueness of expert opinions and unavailability of data while system dynamics was
employed to address the feedback and dynamic effects of design related risks. The results of the
simulation showed that design delays, scope change by client and design changes were the
dominant design risks that have greater impact on cost and time. The proposed risk analysis model
can be used to handle uncertainty, vagueness of expert opinions, feedback and dynamic effects of
project risks. Moreover, the model helps managers understand the significant design risks and their
impact on time and cost overruns.

Keywords: Railway construction project; design risk; risk analysis; fuzzy logic; system dynamics

1
INTRODUCTION

Construction projects are one of the endeavors with numerous unique features such as long period,
complicated processes, abominable environment, financial intensity and dynamic organization
structures (Gupta & Thakkar, 2018; Taylan et al., 2014; Zou et al., 2007). Owing to their
complexity and dynamic nature as well as the involvement of various stakeholders, the construction
projects are exposed to effects of plentiful risk factors leading to cost and time overruns
(Nasirzadeh, Afshar, Khanzadi, et al., 2008).
PMI (2017) defines project risk as is an uncertain event or condition that, if it occurs, has a positive
or negative effect on one or more project objectives. This definition includes two key dimensions of
risk: uncertainty and effect on a project’s objectives. When assessing the importance of a project
risk, these two dimensions must both be considered. The uncertainty dimension may be described
using the term “probability” and the effect may be called “impact”. The definition of project risk
also encompasses uncertain events which could have a negative effect on a project’s objectives, as
well as those which could have a positive effect(PMI, 2017). Other definition given by
(ISO31000:2009) is similar to the above definition that comprises uncertain event and impact on
project objectives.
Since these risk factors have effect on project performance, it is then very important to assess the
risks for enhancing overall performance (Issa, 2013). Consequently, an effective risk management
approach is required to alleviate the impact of risks on project time, cost and related project
objectives. Hence, a suitable and effective risk assessment modeling techniques for construction
projects is needed for the success of construction projects (Elbarkouky et al., 2016; Tipili & Ilyasu,
2014).

For these reasons, different researchers adopted qualitative and quantitative risk analysis techniques
to analyze the impact of risks on construction project objectives. Different approaches have been
implemented to analyze the impact of construction project risks on project objectives. Probabilistic
approaches have been utilized to evaluate delay risk or cost risk by considering risk as an
estimation variance (Taroun, 2014). Researchers on construction project risk assessment adopted
the probabilistic approaches to analyze project risks using probability theory; Molenaar (2005) used
MCS for analyzing cost risk of highway megaprojects; probability of occurrence, and detection and

2
the level of criticality of risk events. In general, the probabilistic techniques are used to handle the
uncertainty of variables that arise from randomness (Thomas et al., 2006).
Since construction projects are risk prone and dependent on expert judgment and experience and
the uncertainty of construction projects arise due to fuzziness of parameters rather than randomness,
it is often difficult to quantify the subjective or qualitative information using probabilistic
approaches (Thomas et al., 2006;Islam & Nepal, 2016). Hence, different researchers include fuzzy
logic to deal with vague, fuzzy, and uncertain systems for assessing project risks and their impact
on project objectives(Thomas et al., 2006); Sharma et al., (2005) utilized a fuzzy FMEA to evaluate
the criticality or risk associated with item failure modes, and Abdelgawad& Fayek(2011) developed
a risk assessment methodology using fuzzy FTA to quantify the fuzzy probability of horizontal
directional drilling failure for pipeline construction projects.
To understand and rank risks in different hierarchy levels by considering multi-criteria and multi-
attributes, AHP has been utilized for risk assessment in different fields; highway construction
project risk assessment (Zayed et al., 2008); tunnel construction (Zhang et al., 2011); petroleum oil
refinery construction project (Dey, 2012). AHP is combined with FST to handle uncertainty, with
TOPSIS to determine variable weights (Islam et al., 2017). Hence, Taylan et al., (2014) developed
fuzzy AHP and TOPSIS to evaluate the construction projects and their overall risks under
incomplete and uncertain situations for projects selection and risk assessment.
Nonetheless, the complexity and dynamic nature of risks cannot be handled using Fuzzy AHP as it
cannot accommodate the interactions, dependencies, and feedback between higher and lower-level
elements. Therefore, ANP a technique has been introduced to address this limitation (Saaty ,1996)
to deal with interdependency of complex decision factors and helps decision makers determine
relationships among decision levels and their corresponding attributes (Ayaǧ & Özdemir,
2007;Yazdani et al., 2019). However, developing and constructing a pair wise matrix for large
number of risks is laborious and lengthy as FAHP (Islam et al., 2017). In addition, the ANP and
FANP approaches are unable to update when new information/data is available(Kabir et al., 2016).
Hence, Bayesian belief network (BBN) has been adopted to deal with the aforementioned
limitations of ANP and FANP for risk assessment (Kabir et al., 2016; Li et al., 2012). Fuzzy BBN
has also been utilized as an approach that can handle uncertainty, vagueness and imprecision of
data, deal with complex relationships, and updating new information to the system (Kabir et al.,

3
2016; Wang & Chen, 2017). However, fuzzy Bayesian belief network (FBBN) techniques require a
lot of time and effort and depend on domain experts(Zhang et al., 2016).
MCDM approaches, fuzzy ANP and fuzzy AHP approaches have been dominantly utilized for
various construction project risk analysis. However, these tools are incapable of updating the
emergence of new information/data. In addition, it is tiresome and lengthy to construct a pair wise
matrix for large number of risks. Hence, Bayesian belief network (BBN) has been adopted to deal
with the limitations of FANP or FAHP for risk assessment. However, BBN and FBBN cannot
handle the dynamic effect of project risks and their feedback effects over time.
Various methods for risk analysis have been proposed in the literature. All of them, however, have
one or more of the following limitations (Nasirzadeh, Afshar, & Khanzadi, 2008b)
 Traditional risk analysis methods do not consider the dynamic nature of risks throughout the
project life cycle and the feedback loop that affects the overall risk impact, and are
incapable of quantifying the full impact of various risks as they do not account for the
indirect effects of risks,
 Although potential risk events can have a significant impact on project cost, time, and
quality, traditional risk analysis techniques can only assess risk consequences based on their
impact on project cost or time.
 Most traditional risk management techniques deal with the risk analysis process from a
qualitative perspective.
The dynamics of project risk are difficult to understand and manage. The main challenge comes
from the subjective, dynamic and multi-factor nature of systemic risks. Not all techniques are
appropriate for dealing with and simulating problems of this nature. Managing the project risk
dynamics requires a different approach, based on a systemic and holistic perspective, capable of
capturing feedbacks and quantifying relevant subjective factors(Rodrigues, 2001).

System dynamics (SD) approaches have been proposed for project risk assessment to assess the
dynamic effect of project risks and their feedback effects(Wan & Liu, 2014). These SD techniques
are also combined with fuzzy set theory to deal with the uncertainty and imprecision of input data
as well as subjectivity of domain experts for construction project risk assessment (Nasirzadeh et al.,
2014). However, these papers did not consider the likelihood and impact of risks on project
objectives. Most of the approaches that utilize SD as a risk assessment approach considers a single

4
risk factor to analyze their effect on project cost or time. Moreover, the implementation of fuzzy
system dynamics approach for risk analysis in railway construction projects is sparse. To bridge the
gap of implementing appropriate and effective risk assessment approach for design risks, a method
that can handle uncertainty, interdependency, dynamic and feedback effects of project risk factors is
proposed in this paper.

Research Methodology

Fuzzy sets theory and fuzzy logic

The fuzzy set concept was first introduced by Zadeh (1965) to handle vagueness in human

judgment. Fuzzy set theory (FST) can operate on inaccurate data to deal with approximate

reasoning. Since uncertainty in construction projects usually arose from partial information,

inherent imprecision or fuzziness of parameter estimates rather than randomness (Thomas et al.,

2006), probabilistic approaches cannot handle these uncertainties. Due to these facts, FST was

introduced to deal with approximate rather than precise reasoning. It is an appropriate technique to

handle uncertainty and vagueness that result from subjective judgment, unavailable and imprecise

information (Abdelgawad & Fayek, 2010).

Different types of fuzzy number membership functions (triangular, trapezoidal, and

Guassian) are used in decision making to interpret linguistic data into numerical format.

Membership functions with triangular shapes are frequently used to provide more precise results

(Zhang et al., 2016) and they are also very effective for solving multi-criteria decision making

(MCDM) problems when the available information is imprecise (Beltrão & Carvalho, 2019). Fuzzy

models using triangular fuzzy numbers are also employed successfully in construction risk

analysis(Nieto-Morote & Ruz-Vila, 2011).

Hence, triangular fuzzy numbers (TFNs) are used for representing likelihood and impact of

project risks on cost and time. TFNs can be represented as (l/m,m/u). The parameters l, m, and u

5
denote the smallest, most promising, and largest possible values that describe a fuzzy event,

respectively (Daǧdeviren & Yüksel, 2010). The membership function of TFN is defined as:

{
0 , x <l
x−l
, l≤ x ≤ m
~
μ ( x / M )= m−l (1)
u−x
, m≤ x ≤ u
u−m
0 , x >u

~ ~
Let fuzzy set A 1 be (l1, m1, u1) and A 2 be (l2,m2,u2), hence the operations on fuzzy numbers
(addition, subtraction, multiplication and division) are defined as (Zhang et al., 2016):

{
~
A1⊕ ~ A 2=(l 1 +l 2 , m1 +m2 ,u1 +u 2)
~
A1 ⊝ ~A 2=(l 1−l 2 ,l 1−l 2 ,u 1−u2)
~ ~ (2)
A 1 ⊗ A 2=(l 1 l 2 , m1 m2 , u1 u2 )
~ ~
A 1 Ø A 2=(l 1 /u 2 , , m1 /m2 , u1 /l 2)

where⊕ , ⊝,⊗∧Ø are the symbols of fuzzy number addition, subtraction, multiplication and
division respectively.

Figure 1. Membership functions of linguistics variables for risk likelihood

6
Experts prefer to use linguistic expression rather than numerical judgment to define the likelihood
of risk occurrence and its impact. The proposed approach defines five linguistic grades in order to
capture these linguistic expressions (Table xxx and XXX).
Table vvv: Linguistic scales of occurrence likelihood of risks
Linguistic scale Triangular fuzzy scale
Very unlikely (0,0,0.25)
Unlikely (0,0.25,0.5)
Medium (0.25,0.5,0.75)
Likely (0.5,0.75,1)
Very likely (0.75,1,1)
Note: the scale is taken from (Rezakhani, 2011)

Figure 2. Membership functions of linguistics variables for risk impact

Table 2: Linguistic scales of degree of risk impact
Linguistic scale Triangular fuzzy scale
Very Low (0,0,0.25)

Low (0,0.25,0.5)

Moderate (0.25,0.5,0.75)

High (0.5,0.75,1)

Very High (0.75,1,1)

Note: the scale is taken from (Rezakhani, 2011)

7
Each evaluator performed his or her assessments independently based on his or her experience,
intuition, and knowledge. The TFNs of m evaluators who explored a synthesized fuzzy degree of
impact value ~ w jfor each of the risk factors are then synthesized using an average score
computation, as shown in equation xxx.

[∑ ]
m
~ 1
❑~
i
w j= xj (3)
m i=1

where ~
w j = (𝑎1𝑗, 𝑎2𝑗, 𝑎3𝑗) represents the synthesized fuzzy degree of impact of the 𝑗th risk factor.
The synthesized results of the fuzzy risk assessment are still in fuzzy numbers format. Therefore, it
is necessary to further conduct defuzzification approach to transfer fuzzy numbers to crisp numbers.
Defuzzification
Defuzzification is the process of converting a fuzzy number into a crisp value. The crisp value is
useful for calculating the degree of design-related risks and evaluating the rank of project risks. A
variety of defuzzification methods are available, including the maximum membership principle, the
centroid method, the weighted average method, the mean max membership, the center of sums, the
center of largest area, and the first (or last) of maxima.
However, for this study, the crisp real number μ~a corresponding to the triangular fuzzy number μ~a =
(𝑙, 𝑚, 𝑢) is obtained from the following relation(Rahmani et al., 2016):

l+m+u
μ~a = (4)
3

Evaluating the degree of risk

After determining the degree of impact and occurrence likelihood of risk factors, the degree of
impact and occurrence likelihood of each risk factor are multiplied to compute the degree of risk
factors(R𝑗):
R𝑗 = I𝑗 × L𝑗
where I𝑗 denotes the normalized degree of impact of the 𝑗th risk factor and L𝑗 represents the
occurrence likelihood of the 𝑗th risk factors (Lu et al., 2014)

8
System dynamics
SD is the simulation modeling method used to represent the structure of complex systems and to
understand their behavior over time. It dynamically captures complex and nonlinear relationships
between components of a complex system(Marshall, Burgos-Liz, Ijzerman, Osgood, et al., 2015).
System Dynamics was developed at the end of 1950s by Jay W. Forrester as a method for modeling
large real world systems (Forrester, 1961). Forrester (1961) defined SD as: “the investigation of the
information-feedback characteristics of systems and the use of models for the design of improved
organizational form and guiding policy” , on the other hand, Coyle (1996) offered a detail definition
considering time and type of model as: “an approach that deals with the time-dependent behavior of
managed systems with the aim of describing the system and understanding, through qualitative and
quantitative models, how information feedback governs its behavior, and designing robust
information feedback structures and control policies through simulation and optimization”.

System dynamics approach represents the feedback processes along with stock and flow structures,
time delays and nonlinearities that determine the dynamics of the system (Sterman, 2000), hence,
the approach is worthwhile for managing and simulation of complex processes that involve changes
over time and allow feedback system. System dynamics modeling is useful for managing and
simulation of processes with two major characteristics: (1) they involve changes over time and (2)
they allow feedback-the transmission and receipt of information(Nasirzadeh & Nojedehi, 2013).

Feedback, accumulations (stocks), rates (flows) and time delays are key elements of SD. Stocks are
an accumulation or aggregation of something. Flows are rates- feed in and out of stock and have the
same units of stock per unit of time. Nonlinearity is an important concept in SD that is linked to the
existence of feedback processes. It means that the effect is rarely proportional to the cause of the
event. One of the key assumptions in the SD is that the behavior of the system is due to its structure
and not to external forces or factors. In general terms, SD can produce patterns and trends, as well
as mean values and outputs of the model. Modelers and stakeholders can analyze patterns and
trends resulting from simulation experiments with different policies or strategies ('what-if'
questions) to inform decision-making (Marshall, Burgos-Liz, Ijzerman, Osgood, et al., 2015).

Feedback and Causal Loop Diagrams

9
A causal diagram consists of variables connected by arrows denoting the causal influences among
the variables. The important feedback loops are also identified in the diagram. Each causal link is
assigned a polarity, either positive (+) or negative (-) to indicate how the dependent variable
changes when the independent variable changes(Sterman, 2000).

Feedback systems are closed loop systems and the inputs are changed on the basis of output. The
feedback system has a closed loop structure that brings back the results of past action to control
future actions. In a closed system, the problem is perceived, action is taken, and the result affects
further action. The distinguishing feature of a closed loop system is therefore the feedback path of
information, decision and action linking output to input (Bala et al., 2017).

Feedback system may be classified as either positive feedback systems or negative feedback
systems. Positive feedback systems generate growth, and negative feedback systems are goal
seeking. A feedback loop is a closed sequence of causes and effects, that is, a closed path of action
and information (Richardson & Pugh, 1981). A feedback loop consists of two or more causal links
between elements that are connected in such a way that if one follows the causality starting at any
element in the loop, one eventually returns to the first element. Feedback loops give rise to
nonlinear behavior, even if all constitutive causal relationships are linear.
Stock and flow Diagrams

SD simulation models are constructed using Stock-Flow Diagrams (SFDs). These diagrams consist
of stock variables, flow variables, auxiliary variables, parameters and constants, causal links
between variables and causal links with delay signs.

A stock variable (level or a state variable) accumulates, i.e. integrates flows over time. During
simulation, a stock variable can only be changed by ingoing and outgoing flow variables. Stocks
accumulate or integrate their flows; the net flow into the stock is the rate of change of the stock.

The state of the system is described by the stock variables. Stocks are accumulations or
aggregations of something. Stocks (also known as state variables) are accumulations of inflows and
out flows over a period of time. When the system is stopped for an instant, stock will have a value
that determines the state of the system at that instant. The flow variables (also known as rates of
change) change the accumulations of the stocks and control the rates of flow. Flows (rates) feed in

10
and out of stocks and have the same units of stocks per time unit(Marshall, Burgos-Liz, Ijzerman,
Crown, et al., 2015).
Hence the integral equation:

Stock (t) =∫ [Inflow ( s)−Outflow (s)]ds + Stock(t 0 )

t0

,where Inflow(s) represents the value of the inflow at any time s between the initial time to and the
current time t. Equivalently, the net rate of change of any stock, its derivative, is the inflow less the
outflow, defining the differential equation

d(Stock)/dt = Inflow(t) - Outflow(t).

System dynamics for construction project risk analysis

Risk is a dynamic event. Overruns, slippage, and other problems are rarely traceable to a single
discrete event in time. In a project, risk occurs in a complex network composed of many
interrelated causes and effects, and these networks generate closed-loop feedback. Risk dynamics
are generated by various feedback loops that occur within the project system(Rodrigues, 2001).
SD modeling can be applied adequately to construction domain problems. The reason is that
Construction projects(Sterman, 1992)
• Are extremely complex, consisting of multiple interdependent components
• Are highly dynamic
• Involve multiple feedback processes
• Involve nonlinear relationships
• Involve both hard (quantitative) and soft (qualitative) data.
SD is suited to handle these situations more than any other modeling process (Nasirzadeh, Afshar,
& Khanzadi, 2008b)
Owing to these advantages, SD has been proposed for safety and risk management that require a
systemic approach to capture feedback and quantify the subjective factors (Rodrigues, 2001). The
method has also been used a potential solution to handle the complexity of projects in various
application areas ; construction project risk analysis (Nasirzadeh, Afshar, & Khanzadi, 2008a,
2008b), hospital systems(Koelling & Schwandt, 2005), aircraft maintenance (Mohaghegh, 2010).

11
Han et al., (2013) used system dynamics to develop an SD model in order to capture the dynamics
of design errors and assess their negative impacts on schedule delays for university building
construction projects; Wang& Yuan (Wang & Yuan, 2016)proposed an SD model to analyze the
effect of dynamic risk interactions on schedule delay in infrastructure projects in China; Xu et al.,
(2018) employed system dynamics to develop a dynamic approach for investigating the effect of
risks on infrastructure project schedule performance; Nasirzadeh et al.,(2007) used system
dynamics approach to evaluate the consequences of alternative response scenarios and to analyze
the impact of project risks on cost, quality and delay; Li et al., (2014) utilized SD approach to
develop a cost risk assessment model for analyzing the interactions of risks and quantifying their
impact on the construction project cost.

However, SD models cannot handle the vague and uncertain information of risk factors. In order to
handle the vagueness and uncertainty of information for inputs of risk assessment, fuzzy logic has
been integrated with SD to deal with vagueness and uncertainty. Hence, a hybridized fuzzy SD
model for risk assessment has been proposed by different researchers; Nasirzadeh, Afshar,
Khanzadi, et al.,(2008) employed fuzzy SD approach to conduct construction project risk analysis
by which interrelated risks are modeled, simulated and quantified in terms of their impact on
project time and cost for bridge construction project.

Proposed Integrated Risk Modeling and Analysis

System dynamics modeling processes

SD modeling process is not strictly defined in a literature; however different authors propose their
own modeling steps. Forrester (1961) developed four steps to create a system dynamics model. The
first step is the articulation of the problem: defining the purpose of modeling and identifying the
entities, interactions and behaviors to highlight. The second step is to describe the causal
relationships between these entities, by building the causal (or influence) diagram. The third step
corresponds to the introduction of stock variables and flow in the system by building a stock flow
diagram. The fourth step is to formulate simulation models.
Coyle (1996) proposed the structure of system dynamics approach. The steps are: problem
recognition, problem understanding and system description (Influence diagrams), Qualitative

12
analysis, Simulation modeling and policy testing and design. Albin (1997) proposed a four-stage
modeling process, i.e, conceptualization, formulation, testing and implementation.
Major steps used in a typical system dynamics study: problem identification, model
conceptualization (construction of a conceptual model), model formulation (construction of a
formal model), model analysis and validation, policy analysis and design, and
implementation(Barlas, 1996).
Table 1. Different Approaches to the system dynamics modeling process—Stages of system
dynamics modeling process (adapted from Martinez-Moyano & Richardson, 2013)
(Randers, 1980) (Richardson & Pugh, Sterman(2000) (Martinez-Moyano &
1981) Richardson, 2013)
Conceptualization Problem Problem articulation Problem identification and
identification and (Boundary Selection) definition
definition
System Formulation of System conceptualization
conceptualization Dynamic Hypothesis
Formulation Model Formulation Formulation of a Model Formulation
Simulation Model
Testing Analysis of model Testing Model Testing and
behavior Evaluation
Model evaluation
Policy Design and
Implementation Policy analysis Evaluation Model use, implementation
& dissemination
Model use or Design of learning
implementation strategy/Infrastructure

For this research, the process proposed by Randers (1980) is adopted to model and analyze the
impact of project risks of railway construction projects on time and cost.

Conceptualization

13
Railway construction projects are risk prone projects that are affected by various uncertain factors
that have impact on project cost and time. Since risks in the railway construction projects are complex
and dynamic in nature, they need in-depth investigation for assessment . These uncertain factors (risks)
need to be identified and analyzed in order to manage the projects effectively and to contribute the
success of the project accordingly.
Hence, the main factors that causes project cost and time overruns are identified and the purpose of
the model is clearly defined in this modeling stage. To recognize the problem, the project risk
factors are identified based on available reports and studies and expert opinions. The project risks
have been analyzed and identify the important variables generating the observed dynamic behavior
of the system. The purpose of this paper is to develop a risk analysis model that addresses the
interdependence, feedback and dynamic effects of risks for railway construction projects. Causal loop
diagrams in the model were used to describe the conceptual model structure derived from a
modeler’s understanding of system and show the dynamic of variables involved in the system. The
model causal loop diagrams indicate how the variables are related with each other in the system.
Stock and flow diagrams were drawn from the causal loop diagrams. The time horizon of the model
is determined considering the average time taken to complete the railway construction projects in
Ethiopia. For this case, a time of seventy months is taken for this system dynamics modeling.

Causal loop diagrams

Causal loop diagrams (CLDs) are a useful tool for representing the feedback structure of systems. A
causal diagram is made up of variables that are connected by arrows that represent the causal
influences between the variables. Variables are linked together by causal links, which are depicted
by arrows. To indicate how the dependent variable changes when the independent variable changes,
each causal link is assigned a polarity, either positive (+) or negative (-).
A loop identifier highlights the important loops, indicating whether they are positive (reinforcing)
or negative (balancing) feedback loops. A positive link indicates that when the cause increases, the
effect increases above what it would have been otherwise, and when the cause decreases, the effect
decreases below what it would have been otherwise. A negative link indicates that as the cause
increases, the effect decreases below what it would have been otherwise, and as the cause
decreases, the effect increases above what it would have been otherwise (Sterman, 2000).

14
The causal loop diagram of design related risk analysis model illustrates the interaction of the
potential project risks that are encountered during design processes. The interrelationships of these
factors are shown using polarity of the links.

This causal loop diagram helps visualize a set of feedback loops. Positive feedback systems and
negative feedback systems are two types of feedback systems. Positive (reinforcing) feedback
systems promote growth, whereas negative (balancing) feedback systems promote goal-seeking. A
feedback loop is a closed sequence of causes and effects, or a closed path of action and information
(Richardson & Pugh, 1981). The feedback loop consists of causal relationships between two or
more elements that are connected in such a way that if one follows the causal relationship starting
from any element in the loop, it will eventually return to the first element. Feedback loops can lead
to non-linear behavior, even if all constitutive causality is linear.

The main elements to be considered in the design-related risk analysis system are scope changes,
design changes, design errors and omissions, schedule pressure, rework (redesign) and delayed
design. These risks need to be properly managed to reduce the impact of design-related risks on
project objectives. Determining how the various factors are related to each other is critical to
understanding the performance of the project system(Love et al., 2010). To determine the root
cause of the project delay and cost overrun, it is necessary to examine the relationship between the
project variables.

Changes in scope are caused by additions, deletions, omissions, or changes in the nature of the
work to be done. The majority of change orders are at the client's request and are typically in the
form of design changes(Love et al., 2010). The client's change in scope of work has an impact on
the design change. Changes in design necessitate rework, which increases the risk of design-related
risks impacting cost and time. Design errors are also major factors that have a significant impact on
project success. These factors can occur as a result of a misunderstanding of the client's
requirements, a tight project schedule, inadequate site information, or poor coordination among
relevant parties.

Increased project duration risks put pressure on schedule, and this pressure also triggers the need to
complete the design on time. As a result, designers can work longer; however, it causes fatigue and
decreased productivity, which in turn increases the risk of scheduling pressure.

15
The degree of the risk mitigation rate depends on the level of risk and its impact on cost and time.
The control system reduces the impact of design-related risks on project objectives, thereby
creating a negative loop to balance the adversely enhanced impact of design-related risks on project
success. The remaining factors demonstrate the causal interdependence of design-related risk
factors. Figure 3 depicts the model's overall causal loop diagram.

<Redesign>
+ +
+ Change in project
Design delay duration
-
R
+
RWCP duration
Lack of -
Change in Design Related B +
coordination mitigation rate
+ design risk + Risk -
+ -
- + +
<Consultant's Incomplete RWCP cost
lack of design <RWCP
duration +
experience> <Lack of R >
coordination>

<RWCP cost>
hire employees
+ +
Design change Schedule pressure +
+ + +
Change in scope of -
+ + Change in
work by client Design error + overtime Project Cost
+ + B
+ + +
Redesign + +
Political +
instability
fatigue
labor
productivity -
+ Redesign Cost
Cost
Change in government Inadequate site underestimation +
investigation +
regulation and law <Lack of
misunderstanding of + + coordination>
<Redesign>
client's requirement
+

Consultant's lack of
experience Unsafe project
location

Figure 3. Design related risk analysis model causal loop diagram

Formulation and Simulation

Simulation model is another version of mental model or casual loop diagram, but written in
equations and computer code (Coyle 1996). In the process of model behavior, computer simulation
was used to determine how all the variables within the system behave over time. Following the
development of causal loop diagrams that represent feedback loops, interdependence, and the effect

16
of project risk factors on project objectives, a model with socks and flows is created to assess their
dynamic effect over time. After the model structure was defined, the basic equations were
formulated to create the simulation model. The parameter values of each model are entered in the
equation of the stock flow diagram of the risk analysis model.
The simulation models are formulated following the creation of the SFDs. The variables in the
model take into account the values of the variables that influence it and from which it receives
information. The dynamic behavior of the system is generated by equations that simulate the
system's behavior over time using initial values for stock variables. A system dynamics model can
thus be used to investigate the long-term behavior of complex systems (Bouloiz et al., 2013).

Stock and flow diagram of design related risk analysis model

The qualitative structure of the system is described by causal loop diagrams of each subsystem of
the construction project risk analysis model. They show how various system variables interact with
one another. They also show whether the relationships are positive or negative. However, they do
not reveal any quantitative relationships between the variables. The quantitative aspects of the
railway construction project risk analysis system are captured by stock and flow structures. Stock
and flow diagrams (SFDs) are created by combining levels and rates variables from causal loop
diagrams. They provide information about the values of various variables as well as the rate at
which those values are associated with them.
The design related risk analysis model stock flow diagram depicts the interdependence of stocks
(design risk, project duration, and project cost), rates (change in design risks, mitigation rate, rate of
change in project duration, and change in project cost), auxiliary and exogenous variables
The SFD of the model is developed based on the causal loop diagrams of the respective system.
These diagrams clearly show the relationships between the various variables of stock, flow and
auxiliary.

17
 Incomplete <Design error>
design
<Design change>
<RWCP
Lack of duration>
coordination
Design related risk
Change in
design risk mitigtaion rate
Consultant's lack of
experience

Design delay

Change in Change in
Change in government RWCP duration
project time Project Cost
regulation and law

Design change
RWCP Cost

Change in scope Redesign redesign cost

of work labor cost
cost underestimation

unsafe project
<RWCP Design error Inadequate site location Hire employees
Cost> <Lack of investigation
coordination>
Political
Schedule instability <Schedule
pressure pressure>

Misunderstanding of
Poor labor
client's requirement
<Lack of productivity <Consultant's
lack of
coordination> experience>
<RWCP
duration>
fatigue overtime

Figure 4. Stock and flow diagram for design related risk analysis model

Results and discussion

The interaction of project risks can be quantified using system dynamics simulation models. The
ability of the system dynamics simulation model to simulate scenarios in which multiple risks
interact is its most notable feature. The SD simulation model's feedback nature ensures that the
cross-impacts of different risks are captured and that direct and indirect risks are quantified.

The design-related risk analysis model is simulated in order to assess the impact of each risk factor
on the project objectives (time and cost). The figures below depict the impact of design-related risk
factors on project duration and cost overruns.

Simulation and results

To begin, three case studies on the effect of design-related risk factors were conducted in order to
understand the impact of these design-related risk variables on time and cost overruns.

18
 .6 Dmnl
1
.4 Dmnl
1
1 3 3 3 3 3 3
3
3
3
1
.3 Dmnl 1 3
.2 Dmnl
3
1 1
3
1
0 Dmnl 3
1
0 Dmnl 3
2 2 2 2 2 2 2 2 1 2 1
2
1
2 2 2
2 1 1 1
0 16 31 47 62
Time (Month)
Design risk 1 1 1 1 1 1 1 1 1 1 1 1 Dmnl
Project cost 2 2 2 2 2 2 2 2 2 2 2 2 2 Dmnl
project duration 3 3 3 3 3 3 3 3 3 3 3 Dmnl

Figure 5. Simulation patterns of design related risk, project time and cost overruns

Figure 6. simulation patterns of design risk impact on project cost

According to the graph above, design risk increases exponentially from 0.187 to 0.542 after 14
months and then decreases to around 0 after 60 months of project execution. In a 62-month railway

19
project construction period, the impact of project risks on project cost ranges from 0 to 0.0308, and
the risk of project duration ranges from 0 to 0.313.
Figure 5 depicts the dynamic simulation patterns of the stock variables, indicating the behavior
trends for design risk, project cost risks, and project duration, while Figure 6 specifically depicts the
pattern of the impact of design risks on project cost overrun. Table 2 also shows the dynamic
simulation outputs for all system variables in the design risk analysis model.
Table 2 Simulation results of design related risks

Maximum Minimum Average Standard

Design risk factors deviation
Change in government regulation and law 0.02010 0.02010 0.02010 0.00000
Change in scope of work 0.06764 0.03684 0.05898 0.01074
Client's lack of experience 0.01674 0.01674 0.01674 0.00000
Contractor's experience 0.01100 0.01100 0.01100 0.00000
Cost underestimation 0.00056 0.00056 0.00056 0.00000
Design change 0.06764 0.03684 0.05898 0.01074
Design delay 0.13478 0.10398 0.12612 0.01074
Design error 0.02729 0.02683 0.02699 0.00014
Fatigue 0.03162 0.01890 0.02330 0.00395
Hire employees 0.03162 0.01890 0.02330 0.00395
Inadequate site investigation 0.02873 0.02873 0.02873 0.00000
Incomplete design 0.06262 0.06262 0.06262 0.00000
Labor cost 0.03162 0.01890 0.02330 0.00395
Lack of coordination 0.03614 0.03614 0.03614 0.00000
Misunderstanding of client's requirement 0.00025 0.00025 0.00025 0.00000
Overtime 0.03162 0.01890 0.02330 0.00395
Poor labor productivity 0.00114 0.00068 0.00084 0.00014
Redesign 0.09493 0.06367 0.08597 0.01086
Redesign cost 0.09493 0.06367 0.08597 0.01086
Project cost 0.03080 0.00000 0.02214 0.01074
Project duration 0.31345 0.00000 0.22979 0.10703
Schedule pressure 0.03162 0.01890 0.02330 0.00395
Unsafe project location 0.02993 0.02993 0.02993 0.00000

Effect of exogenous variables on project delay and cost overrun

The dynamic simulation was carried out to investigate the effects of exogenous parameters on
design risks, project cost risks, and project duration risks. Exogenous variables refer to variables
whose causes are outside the model and whose role is to explain other variables or results in the
model. These variables are identified in the modeling process and boundary determination. The

20
exogenous variables for this design related risk analysis model are changes in government
regulations and laws, unsafe project location, client’s and contractor’s experience on similar
projects. Duration 0.282 to 0.344, cost 0.0248 to 0.0374.

RWCP duration
.4
2 2 2 2 2 2 2
2
2 3 3 3 3 3 3 3
.3 2 3
3
1 1 1 1 1
1 1
3 1
2
Dmnl

1
.2 3
1
2
3
1
.1
2 3 1

2 3 1
0 3 1
1 2
0 16 31 47 62
Time (Month)
RWCP duration : Exogenous risks[-20%] 1 1 1 1 1 1 1 1
RWCP duration : Exogenous risks[+20%] 2 2 2 2 2 2 2 2
RWCP duration : Exogenous risks baseline 3 3 3 3 3 3 3

Figure 7. Effect of exogenous variables on project duration

21
 RWCP Cost
.04 2 2 2 2 2 2
2 2
2
2 3 3 3 3 3 3 3
.03 3
3
2 3 1 1 1 1 1 1
1
Dmnl

1
.02 3 1
2
1
3
.01 1
2 3
1
2 3 1
0 1
1 2 3
0 16 31 47 62
Time (Month)
RWCP Cost : Exogenous risks[-20%] 1 1 1 1 1 1 1 1
RWCP Cost : Exogenous risks[+20%] 2 2 2 2 2 2 2 2 2
RWCP Cost : Exogenous risks baseline 3 3 3 3 3 3 3 3

Figure 8. Effect of exogenous variables on project cost

Effect of design change

Design change in construction projects is one of the main risks that occurs frequently. These risks
mainly arise from change in scope of work by clients. The majority of change orders are at the
client's request and are typically in the form of design changes (Love et al., 2010). Changes in
design necessitate rework, which increases the risk of design-related risks impacting cost and time.
Figure xx and xx depict the effect of design changes on railway construction project duration and
cost. The figures show the level of impact of design change within the range of 20% from the
current risk level. The simulation results show that changing the design has a significant impact on
the project objectives. Within this 20 percent risk range of design change, the risk of delay ranges
from 0.280 to 0.353, while the risk of project overrun ranges from 0.024 to 0.040 within the project
timeline.

22
 RWCP duration
.4
2 2 2 2 2 2 2
2
2 3 3 3 3 3
3 3
.3 2 3
3
1 1 1 1
1 1
1
3 1
2 1
Dmnl

.2 3
1
2 3
1
.1
2 3 1

2 3 1
0 3 1
1 2
0 16 31 47 62
Time (Month)
RWCP duration : Design Risk [design change -20%] 1 1 1 1 1 1 1
RWCP duration : Design Risk [design change +20%] 2 2 2 2 2 2 2
RWCP duration : Design Risk 3 3 3 3 3 3 3 3 3 3 3

Figure 9. Effect of design change on project duration

RWCP Cost
.05

2 2 2 2 2 2
2
.0375 2
2
3 3 3 3 3 3 3
3
Dmnl

.025 2 3
1 1 1 1 1 1
1 1
3 1
2 1
.0125 3
1
2 3
1
2 3 1
0 1
1 2 3
0 16 31 47 62
Time (Month)
RWCP Cost : Design Risk [design change -20%] 1 1 1 1 1 1 1 1
RWCP Cost : Design Risk [design change +20%] 2 2 2 2 2 2 2 2
RWCP Cost : Design Risk 3 3 3 3 3 3 3 3 3 3 3

Figure 10. Effect of design change on project cost

Effect of design errors

Errors have become an inextricable part of the design process in construction and engineering
projects (Love et al., 2012). In the construction industry, design errors are unavoidable, and they
have a negative impact on project management efficiency and effectiveness and can adversely
23
affect cost, schedule and safety (Love et al., 2012; Peansupap & Ly, 2015). Figures 11 and 12 show
how varying the risk level of design errors affect the duration and cost of railway construction
project. According to the simulation results, design errors have a significant impact on the project
objectives. Within the 20% design error risk range, the delay risk ranges from 0.298 to 0.329, while
the project cost overrun risk ranges from 0.0272 to 0.0347 within the project timeline.

RWCP duration
.4

3 3 3 3 3 3 3
3 1 1 1 1 1 1
.3 3 1 2 2 2 2 2 2
1 2
3 1 2
2
3 1 2
Dmnl

.2
3 1 2

.1 3 1
2

2
3 1
2
0 3 1
1 2
0 16 31 47 62
Time (Month)
RWCP duration : Design Error base 1 1 1 1 1 1 1 1 1
RWCP duration : Design Error[-20%] 2 2 2 2 2 2 2 2 2
RWCP duration : Design Error[+20%] 3 3 3 3 3 3 3 3

Figure 11. Effect of design errors on project duration

RWCP Cost
.04
3 3 3 3 3 3 3
3
3 1 1 1 1
1 1
.03 3 1
1
2 2 2 2 2 2
1 2
2
3 2
Dmnl

1
.02 2
3 1
2
.01 3 1 2

3 1 2
0 2
3 1
1 2
0 16 31 47 62
Time (Month)
RWCP Cost : Design Error base 1 1 1 1 1 1 1 1 1
RWCP Cost : Design Error[-20%] 2 2 2 2 2 2 2 2 2
RWCP Cost : Design Error[+20%] 3 3 3 3 3 3 3 3 3

Figure 12. Effect of design errors on project cost

24
Model Validation

Validation of dynamic simulation models is one of the most vital phases in the process of modeling
real systems. The tests used to validate the models were boundary adequacy, structure assessment,
parameter assessment, extreme conditions test, integration error test, and sensitivity analysis. The
test for robustness used to check the models for realistic behavior when stressed to extreme
conditions. The testing proved whether the models are consistent with the system behavior with
respect to their purposes. Where tests failed, iterations were carried out in the modeling process to
develop a valid model. The validation methods employed for this design risk analysis model are
direct structure tests that do not use simulation and compare model with knowledge about the real
system; information from the system (empirical) and from generalized knowledge (theoretical),
Structure-oriented behavior tests that use simulation in order to study the behavioral results as a
mean to indirectly find out structural flaws and behavior pattern tests by which result in a measure
of model accuracy in reproducing the behavior patterns found in the real system (Barlas, 1996).
Implementation

After validating the models with relevant testing methods, they were experimented for various
practical consequences. The aim of the study was to identify the weakness in the existing risk
management procedure and recommend a better system to assess risks at the early stage of railway
construction project development before construction commenced.

25
 Conclusions

This paper aimed to analyze the impact of design-related risks on railway construction project
delays and cost overruns. It used an integrated system dynamics approach to deal with feedback and
the dynamic effect of design-related risks, as well as fuzzy logic to deal with the uncertainty and
vagueness of experts' judgment on risk likelihood and impact rating. Whereas other risk
management tools are ineffective in analyzing these systemic behaviors of interrelationships
between risks and feedback loops, the SD approach provides a different analysis in which project
risks and their impact on project objectives are explicitly considered and quantified. Based on the
opinions of experts involved in railway construction projects, the interrelationship between these
risks affecting project objectives of railway construction projects is developed using a causal loop
diagram of design related risk analysis model. The quantitative dynamic risk analysis model is then
developed from the qualitative model by incorporating the mathematical relationships between the
design risk factors. To analyze the impact of design-related risks on project cost and time, the
fuzzy-SD model of design risk analysis is simulated. During the simulation period, the level of
design-related risks, as well as their impact on project cost and schedule, are determined. The
patterns and results of dynamic simulation show that design delays, design changes, and scope
changes by clients tend to dominate and cause cost and time overruns.
The performance and validity of the proposed fuzzy SD approach are evaluated using various
verification and validation tests for the design risk analysis model. The validation test results
demonstrate the effectiveness of the proposed model for project risk analysis in determining the risk
impact on project cost and time.
The application of this approach in the context of a railway construction project risk analysis
system provides a dynamic analysis of design-related risks and their impact on project objectives.
This model can be generalized and applied to other project risks, with some modifications, to
analyze the impact of risks on project cost and time. Since the data for this study were gathered
from experts at the Ethiopian Railway Corporation, the simulation result may be influenced because
other stakeholders' opinions were not taken into account.

26
 References

Abdelgawad, M., & Fayek, A. R. (2010). Risk management in the construction industry using
combined fuzzy FMEA and fuzzy AHP. Journal of Construction Engineering and
Management, 136(9), 1028–1036. https://doi.org/10.1061/(ASCE)CO.1943-7862.0000210
Abdelgawad, M., & Fayek, A. R. (2011). Fuzzy reliability analyzer: Quantitative assessment of risk
events in the construction industry using fuzzy fault-tree analysis. Journal of Construction
Engineering and Management, 137(4), 294–302. https://doi.org/10.1061/(ASCE)CO.1943-
7862.0000285
Ayaǧ, Z., & Özdemir, R. G. (2007). An intelligent approach to ERP software selection through
fuzzy ANP. International Journal of Production Research, 45(10), 2169–2194.
https://doi.org/10.1080/00207540600724849
Bala, B. K., Arshad, F. M., & Noh, K. M. (2017). System Dynamics Modelling and Simulation. In
System Dynamics Review. Springer Science+Business Media.
Barlas, Y. (1996). Formal aspects of model validity and validation in system dynamics. System
Dynamics Review, 12(3), 183–210. https://doi.org/10.1002/(sici)1099-
1727(199623)12:3<183::aid-sdr103>3.0.co;2-4
Beltrão, L. M. P., & Carvalho, M. T. M. (2019). Prioritizing Construction Risks Using Fuzzy AHP
in Brazilian Public Enterprises. Journal of Construction Engineering and Management,
145(2), 1–11. https://doi.org/10.1061/(ASCE)CO.1943-7862.0001606
Bouloiz, H., Garbolino, E., Tkiouat, M., & Guarnieri, F. (2013). A system dynamics model for
behavioral analysis of safety conditions in a chemical storage unit. Safety Science, 58(1), 32–
40. https://doi.org/https://doi.org/10.1016/j.ssci.2013.02.013
Coyle, R. G. (1996). System Dynamics Modelling: A Practical Approach. Chapman & Hall.
Daǧdeviren, M., & Yüksel, I. (2010). A fuzzy analytic network process (ANP) model for
measurement of the sectoral competititon level (SCL). Expert Systems with Applications,
37(2), 1005–1014. https://doi.org/10.1016/j.eswa.2009.05.074
Dey, P. K. (2012). Project risk management using multiple criteria decision-making technique and
decision tree analysis: A case study of Indian oil refinery. Production Planning and Control,
23(12), 903–921. https://doi.org/10.1080/09537287.2011.586379
Elbarkouky, M. M. G., Fayek, A. R., Siraj, N. B., & Sadeghi, N. (2016). Fuzzy arithmetic risk
analysis approach to determine construction project contingency. Journal of Construction
Engineering and Management, 142(12), 1–11. https://doi.org/10.1061/(ASCE)CO.1943-
7862.0001191
Forrester, J. W. (1961). Industrial dynamics. MIT Press.
Gupta, V. K., & Thakkar, J. J. (2018). A quantitative risk assessment methodology for construction
project. Sadhana - Academy Proceedings in Engineering Sciences, 1–16.
https://doi.org/10.1007/s12046-018-0846-6
Han, S., Love, P., & Peña-Mora, F. (2013). A system dynamics model for assessing the impacts of
design errors in construction projects. Mathematical and Computer Modelling, 57(9), 2044–
2053. https://doi.org/10.1016/j.mcm.2011.06.039
Islam, M. S., & Nepal, M. (2016). A Fuzzy-Bayesian Model for Risk Assessment in Power Plant
Projects. Procedia Computer Science, 100, 963–970.
https://doi.org/10.1016/j.procs.2016.09.259

27
Islam, M. S., Nepal, M. P., Skitmore, M., & Attarzadeh, M. (2017). Current research trends and
application areas of fuzzy and hybrid methods to the risk assessment of construction projects.
Advanced Engineering Informatics, 33(1), 112–131. https://doi.org/10.1016/j.aei.2017.06.001
Issa, U. H. (2013). Implementation of lean construction techniques for minimizing the risks effect
on project construction time. Alexandria Engineering Journal, 52(4), 697–704.
https://doi.org/10.1016/j.aej.2013.07.003
Kabir, G., Sadiq, R., & Tesfamariam, S. (2016). A fuzzy Bayesian belief network for safety
assessment of oil and gas pipelines. Structure and Infrastructure Engineering, 12(8), 874–889.
https://doi.org/10.1080/15732479.2015.1053093
Koelling, P., & Schwandt, M. J. (2005). Health systems: A dynamic system - Benefits from system
dynamics. Proceedings - Winter Simulation Conference.
https://doi.org/10.1109/WSC.2005.1574393
Li, P., Chen, G., Dai, L., & Zhang, L. (2012). A fuzzy Bayesian network approach to improve the
quantification of organizational influences in HRA frameworks. Safety Science, 50(7), 1569–
1583. https://doi.org/https://doi.org/10.1016/j.ssci.2012.03.017
Li, Wang, J., & Yuan, H. (2014). Analysis of the Interactions Between Risks Affecting
Construction Project Cost. In D. Yang & Y. Qian (Eds.), Proceedings of the 18th International
Symposium on Advancement of Construction Management and Real Estate (pp. 423–431).
Springer Berlin Heidelberg.
Love, P. E. D., Davis, P. R., Ellis, J. M., & Cheung, S. O. (2010). A systemic view of dispute
causation. International Journal of Managing Projects in Business, 3(4), 661–680.
https://doi.org/10.1108/17538371011076109
Love, P. E. D., Lopez, R., Edwards, D. J., & Goh, Y. M. (2012). Error begat error: Design error
analysis and prevention in social infrastructure projects. Accident Analysis & Prevention,
48(1), 100–110. https://doi.org/https://doi.org/10.1016/j.aap.2011.02.027
Lu, S.-T., Yu, S.-H., & Chang, D.-S. (2014). Using Fuzzy Multiple Criteria Decision-Making
Approach for Assessing the Risk of Railway Reconstruction Project in Taiwan.
TheScientificWorldJournal, 2014(1), 1–8. https://doi.org/10.1155/2014/239793
Marshall, D. A., Burgos-Liz, L., Ijzerman, M. J., Crown, W., Padula, W. V., Wong, P. K.,
Pasupathy, K. S., Higashi, M. K., & Osgood, N. D. (2015). Selecting a dynamic simulation
modeling method for health care delivery research - Part 2: Report of the ISPOR dynamic
simulation modeling emerging good practices task force. Value in Health, 18(2), 147–160.
https://doi.org/10.1016/j.jval.2015.01.006
Marshall, D. A., Burgos-Liz, L., Ijzerman, M. J., Osgood, N. D., Padula, W. V., Higashi, M. K.,
Wong, P. K., Pasupathy, K. S., & Crown, W. (2015). Applying dynamic simulation modeling
methods in health care delivery research - The SIMULATE checklist: Report of the ISPOR
simulation modeling emerging good practices task force. Value in Health, 18(1), 5–16.
https://doi.org/10.1016/j.jval.2014.12.001
Martinez-Moyano, I. J., & Richardson, G. P. (2013). Best practices in system dynamics modeling.
System Dynamics Review, 29(2), 102–123. https://doi.org/10.1002/sdr.1495
Mohaghegh, Z. (2010). Combining system dynamics and Bayesian belief networks for socio-
technical risk analysis. International Conference on Intelligence and Security Informatics:
Public Safety and Security. https://doi.org/10.1109/ISI.2010.5484736
Molenaar, K. R. (2005). Programmatic cost risk analysis for highway megaprojects. Journal of
Construction Engineering and Management, 131(3), 343–353.
https://doi.org/10.1061/(ASCE)0733-9364(2005)131:3(343)

28
Nasirzadeh, F., Afshar, A., & Khanzadi, M. (2007). System Dynamics Approach to Optimum
Response Selection in Construction Project Risks. International Project Management
Conference.
Nasirzadeh, F., Afshar, A., & Khanzadi, M. (2008a). Dynamic risk analysis in construction
projects. Canadian Journal of Civil Engineering, 35(8), 820–831. https://doi.org/10.1139/L08-
035
Nasirzadeh, F., Afshar, A., & Khanzadi, M. (2008b). System dynamics approach for construction
risk analysis. International Journal of Civil Engineering, 6(2), 120–131.
Nasirzadeh, F., Afshar, A., Khanzadi, M., & Howick, S. (2008). Integrating system dynamics and
fuzzy logic modelling for construction risk management. Construction Management and
Economics, 26(11), 1197–1212. https://doi.org/10.1080/01446190802459924
Nasirzadeh, F., Khanzadi, M., & Rezaie, M. (2014). Dynamic modeling of the quantitative risk
allocation in construction projects. International Journal of Project Management, 32(3), 442–
451. https://doi.org/10.1016/j.ijproman.2013.06.002
Nasirzadeh, F., & Nojedehi, P. (2013). Dynamic modeling of labor productivity in construction
projects. International Journal of Project Management, 31(6), 903–911.
https://doi.org/10.1016/j.ijproman.2012.11.003
Nieto-Morote, A., & Ruz-Vila, F. (2011). A fuzzy approach to construction project risk assessment.
International Journal of Project Management, 29(2), 220–231.
https://doi.org/10.1016/j.ijproman.2010.02.002
Peansupap, V., & Ly, R. (2015). Evaluating the Impact Level of Design Errors in Structural and
Other Building Components in Building Construction Projects in Cambodia. Procedia
Engineering, 123, 370–378. https://doi.org/https://doi.org/10.1016/j.proeng.2015.10.049
PMI. (2017). A guide to the project management body of knowledge (PMBOK® guide) (6th ed.).
Project Management Institute, Inc.
Rahmani, A., Hosseinzadeh Lotfi, F., Rostamy-Malkhalifeh, M., & Allahviranloo, T. (2016). A
New Method for Defuzzification and Ranking of Fuzzy Numbers Based on the Statistical Beta
Distribution. Advances in Fuzzy Systems, 2016(1), 1–8. https://doi.org/10.1155/2016/6945184
Randers, J. (1980). Elements of the system dynamics method. Pegasus Communications.
Rezakhani, P. (2011). Fuzzy Risk Analysis Model for Construction Projects. International Journal
of Civil and Structural Engineering, 2(2), 516–531.
https://doi.org/10.6088/ijcser.00202010130
Richardson, G. P., & Pugh, A. L. (1981). Introduction to system dynamics modeling with dynamo.
Wright Allen Pr.
Rodrigues, A. G. (2001). Managing and Modelling Project Risk Dynamics A System Dynamics-
based Framework. Fourth European Project Management Conference: PMI Europe.
Saaty. (1996). Decision Making with Dependence and Feedback: The Analytic Network Process.
RWS Publications.
Sharma, R. K., Kumar, D., & Kumar, P. (2005). Systematic failure mode effect analysis (FMEA)
using fuzzy linguistic modelling. International Journal of Quality and Reliability
Management, 22(9), 986–1004. https://doi.org/10.1108/02656710510625248
Sterman, J. D. (1992). System Dynamics Modeling for Project Management. In Unpublished
manuscript.
Sterman, J. D. (2000). Business dynamic:systems thinking and modeling for a complex world. The
McGraw-Hill Companies, Inc.
Taroun, A. (2014). Towards a better modelling and assessment of construction risk: Insights from a

29
 literature review. International Journal of Project Management, 32(1), 101–115.
https://doi.org/10.1016/j.ijproman.2013.03.004
Taylan, O., Bafail, A. O., Abdulaal, R. M. S., & Kabli, M. R. (2014). Construction projects
selection and risk assessment by fuzzy AHP and fuzzy TOPSIS methodologies. Applied Soft
Computing, 17(1), 105–116. https://doi.org/10.1016/j.asoc.2014.01.003
Thomas, A. V., Kalidindi, S. N., & Ganesh, L. S. (2006). Modelling and assessment of critical risks
in BOT road projects. Construction Management and Economics, 24(4), 407–424.
https://doi.org/10.1080/01446190500435275
Tipili, L. G., & Ilyasu, M. S. (2014). Evaluating the impact of risk factors on construction projects
cost in Nigeria. The International Journal Of Engineering And Science.
Wan, J., & Liu, Y. (2014). A System Dynamics Model for Risk Analysis during Project
Construction Process. Open Journal of Social Sciences, 2(6), 451–454.
https://doi.org/10.4236/jss.2014.26052
Wang, & Chen, C. (2017). Fuzzy comprehensive Bayesian network-based safety risk assessment
for metro construction projects. Tunnelling and Underground Space Technology, 70(1), 330–
342. https://doi.org/10.1016/j.tust.2017.09.012
Wang, J., & Yuan, H. (2016). System Dynamics Approach for Investigating the Risk Effects on
Schedule Delay in Infrastructure Projects. Journal of Management in Engineering, 33,
4016029. https://doi.org/10.1061/(ASCE)ME.1943-5479.0000472
Xu, X., Wang, J., Li, C. Z., Huang, W., & Xia, N. (2018). Schedule risk analysis of infrastructure
projects: A hybrid dynamic approach. Automation in Construction, 95(1), 20–34.
https://doi.org/https://doi.org/10.1016/j.autcon.2018.07.026
Yazdani, M., Abdi, M. R., Kumar, N., Keshavarz-Ghorabaee, M., & Chan, F. T. S. (2019).
Improved Decision Model for Evaluating Risks in Construction Projects. Journal of
Construction Engineering and Management, 145(5), 1–13.
https://doi.org/10.1061/(ASCE)CO.1943-7862.0001640
Zadeh, L. A. (1965). Fuzzy Sets. Information and Control, 8(3), 338–353.
https://doi.org/doi:10.1016/s0019-9958(65)90241-x
Zayed, T., Amer, M., & Pan, J. (2008). Assessing risk and uncertainty inherent in Chinese highway
projects using AHP. International Journal of Project Management, 26(4), 408–419.
https://doi.org/10.1016/j.ijproman.2007.05.012
Zhang, Cai, M., Li, X., & Mu, Z. (2011). Prioritizing highway tunnel risk factors with AHP
method. 2011 International Conference on Information Science and Technology, ICIST 2011.
https://doi.org/10.1109/ICIST.2011.5765186
Zhang, L., Wu, X., Qin, Y., Skibniewski, M. J., & Liu, W. (2016). Towards a Fuzzy Bayesian
Network Based Approach for Safety Risk Analysis of Tunnel-Induced Pipeline Damage. Risk
Analysis, 36(2), 278–301. https://doi.org/10.1111/risa.12448
Zou, Zhang, G., & Wang, J. (2007). Understanding the key risks in construction projects in China.
International Journal of Project Management, 25(6), 601–614.
https://doi.org/10.1016/j.ijproman.2007.03.001

30