ARTICLE IN PRESS

Building and Environment 40 (2005) 12441254

www.elsevier.com/locate/buildenv

Judgmental risk analysis process development in

construction projects
Ahmet Oztas- , Onder Okmen
Civil Engineering Department, University of Gaziantep, Gaziantep, Turkey
Received 16 June 2003; accepted 22 October 2004

Abstract

Various risks and uncertainties exist in construction projects. These may not only prevent the projects to be completed within
budget and time limit, but also threaten the quality, safety and operational needs. In this context, risk analysis processes are the
systematic methods to analyze the potential project risks and develop risk response strategies in order to cope with risks and achieve
the desired objectives. This study proposes a new schedule risk analysis method named as judgmental risk analysis process (JRAP)
and offers a different project duration equation through JRAP. The process (JRAP) can be dened as a pessimistic risk analysis
methodology or a hypothesis based on Monte Carlo simulation that is effective in uncertain conditions due to its capability of
converting uncertainty to risk judgmentally in construction projects. A case study has also been developed to show how the
proposed process is applied on a construction project and to prove its validity.
r 2004 Published by Elsevier Ltd.

Keywords: Risk; Risk analysis; Monte Carlo simulation; Schedule analysis; Construction management

1. Introduction the previous data available. Since such a separation is

Each construction project has unique features that
differentiate it from even resembling projects. Construc-
tion techniques, design, contract types, liabilities, weath-
er, soil conditions, politic-economic environment and
many other aspects may be different for every new
commitment. This fuzzy atmosphere has been repre-
sented with the terms uncertainty or risk by construc-
tion managers and researchers, and they tried to control
this systematically through risk management and
analysis methods since the early 1990s [1]. Some
researchers like Flanagan et al. [2] and Pilcher [3] put
differentiation between these two terms. They have
mentioned that uncertainty represents the situations in
which there is no historical data; and risk, in contrast,
can be used for situations where success or failure is
determined in probabilistic quantities by beneting from
Corresponding author.
E-mail address: aoztas@gantep.edu.tr (A. Oztas-).
regarded as meaningless in the construction literature,
risk turns out to be the most consistent term to be used
for construction projects because some probability
values can be attached intuitively and judgmentally to
even the most uncertain events [2]. The uncertainty
represented quantitatively at some level is not the
uncertainty any more; rather it is the risk henceforth
and needs to be managed. In this context, this research
aims to prove this hypothesis by developing judgmental
risk analysis process (JRAP) and applying this process
in the assessment of schedule risks of a construction
project through a case study.
One important way of controlling risks in construc-
tion projects is to develop reliable project estimates and
schedules [4]. Probabilistic scheduling through conduc-
tion of simulations on prepared models provides more
powerful results fundamental to making decisions with
regard to the control of project risks. Barraza et al. [5]
has brought a stochastic approach to traditional S-curve
project performance control method by using simulation

0360-1323/$ - see front matter r 2004 Published by Elsevier Ltd.

doi:10.1016/j.buildenv.2004.10.013
 ARTICLE IN PRESS
A. Oztas- , O. Okmen / Building and Environment 40 (2005) 12441254
techniques and probability distributions. Chehayeb et
al. [6] has developed a different simulation-based
scheduling method with continuous activity relation-
ships to enable effective use of systems simulation in
scheduling of construction projects. Similarly, Senior et
al. [7] has discussed a new scheduling system based on
statistical simulation. The reliability of probability
distribution functions used as input data in stochastic
scheduling systems has attracted the interest of research-
ers like Maio et al. [8] and Fente et al. [9].
Different methods have been used so far for measur-
ing the variations in activity durations and modelling
schedule risks of projects, e,g. Program Evaluation and
Review Technique (PERT) [10], Probabilistic Network
Evaluation Technique (PNET) [11], Narrow Reliability
Bounds (NRA) [12] and Monte Carlo simulation (MCS)
[2,10,13]. Activity durations in construction projects are
likely to vary because of the inherent risks. For this
reason, Dawood [13] and Ranashinghe [14] have also
developed equations for quantifying and modelling the
variation in activity durations and nding the overall
project duration.
2. Risk analysis and risk management
Each construction project has its own technical
characteristics that vary according to the construction
type, execution time and its environment. This leads to a
different risk atmosphere for each construction project.
In this context, risk management can be dened as a
systematic controlling procedure of risks that are
predicted to be faced in an investment or project [15].
However, it is neither an insurance system nor a magical
risk elimination method. It only aims to identify the
potential risks as early as possible and manage them for
preventing the harmful effects of the risks to the project
aims. The system utilizes the historical data, statistical
knowledge, computer modelling-running power, human
intuition, judgment, experience, gut feel, common sense
and willingness up to a full extent [2].
In the course of time, the management of risk has
become a key element for the completion of projects
within time schedule and planned budget. It is now a
common opinion that controllable and uncontrollable
risks can only be responded by utilizing risk manage-
ment process over the entire project, i.e. prior to the
tender and subsequently, by controlling and updating
the system periodically during the application of the pre-
determined plan.
Like every systematic procedure, risk management is
a stepwise phenomenon. These steps can be summarized
basically as risk identication, risk classication, risk
analysis, and risk response [2]. Risk identication and
classication comprise the qualitative investigation of
the risks. In the literature, a number of risk identica-
1245
tion and classication techniques are present such as
brainstorming, check lists, organizational charts and
mapping [1,16]. Risk analysis part is dominantly
quantitative and, therefore, it includes mathematical
and statistical operations [17]. Risk response is the last
risk management step at which the results of the
preceding steps are discussed and suitable risk mitiga-
tion actions are taken before facing risks.
3. JRAP methodology
The JRAP proposed in this research consists of a
number of managerial steps to be carried out and an
equation that offers the variation in each activitys
duration in the schedule network. The characteristics of
JRAP make this methodology effective in uncertain
conditions of which there is no or little previous data
and increases its applicability for converting high level
of uncertainty to risk judgmentally. Since, judgmental
decisions based on experience and intuition would insert
additional risks to the project network model, a
pessimistic way has been followed in analyzing the
overall project duration. In other words, the pessimistic
characteristic of JRAP decreases the effect of planning
engineers making inaccurate data estimation during
risk modelling. This pessimism is created through the
computing characteristics of the proposed equation that
is introduced within JRAP.
Another important point that should be mentioned
about JRAP is that it is not a whole risk management
system, rather it is considered as an analysis stage that is
performed during the conguration of a risk manage-
ment system. JRAP can be dened as a pessimistic risk
analysis methodology or a hypothesis that is effective in
schedule risk modelling of even the most uncertain
situations in construction projects.
Fig. 1 illustrates the logic and steps of JRAP and
claries the processs relationship with the other risk
management stages. The tasks labelled as D, E, F and G
represent the steps of JRAP. They make up the risk
management system together with the other tasks, which
are B, C, H and I. The task labelled as A, in contrast,
represents the classical approach, i.e. the arrangement of
project content by deterministic schedule planning.
JRAP methodology comprises the analysis part of
overall risk management system as seen in Fig. 1. In
other words, JRAP can only be applied just after
identifying and classifying the risks that would inuence
the activity durations of the shedule.
Deterministic way of schedule analysis prior to setting
up a risk management system (refer to Fig. 1) leads to a
more appropriate risk-modelling environment for any
project, even though deterministic approach avoids the
use of statistical data and assumes certain xed values.
The need for the deterministic model arises from its
 ARTICLE IN PRESS
1246 A. Oztas- , O. Okmen / Building and Environment 40 (2005) 12441254

Model
deterministic
time schedule
plan
Identify
of the project
risks
-A-
-B-

Classify
risks
-C-

RISK MANAGEMENT FRAMEWORK

Decide the critical risks Update the model
effective on activity durations
-D-

Assign probability
distributions
and max-min durations
-E- Control
the
progress
-I-

Establish activity-risk factor

matrix
-F-

Model stochastic time

schedule plan
of the project and Run Monte Respond
Carlo Simulation to risks
-G- -H-

JRAP METHODOLOGY

Fig. 1. The steps of judgmental risk analysis process (JRAP). Note: JRAP steps are labelled as D, E, F and G.

power of highlighting the potential project risks. In
other words, it helps the planner especially during the
risk identication step.
3.1. Step 1deciding of main risks
The rst task undertaken during the conduction of
JRAP is to nd out the critical risks that may be
effective on activity durations. This is achieved through
examining the critical activities of the deterministic
schedule network and then selecting the risks which
inuence these critical activities from the risk list
predetermined in the risk identication step (step B in
Fig. 1) of risk management framework. For the sake of
simplication, the critical risks established in this way
are assumed to be effective on also non-critical activities.
3.2. Step 2assignment of probability distributions and
maximum minimum durations
The second step of JRAP is to attach probability
distributions to the identied risks in the preceding step.
This can be achieved through utilizing experience and
engineering judgment in case there is not enough past
 ARTICLE IN PRESS
A. Oztas- , O. Okmen / Building and Environment 40 (2005) 12441254
record to conduct detailed statistical analysis. This is
already the advantage and practicality of JRAP during
conditions when there is no previous data.
The random values between 0 and 1 (refer to Eq. (1))
will be selected by the utilized software according to the
character of the assigned probability distributions and
correlation coefcients during MCS. While distributions
are being assigned, maximum and minimum durations
of activities should also be determined. Eventually,
all of these values will be used in calculating activity
durations. Common distributions like beta, triangle and
lognormal can be used for representing the critical risks.
However, negative skewness is not offered in JRAP,
because it is less probable to face with a situation that
the real execution time of a construction activity would
be below the appraised most likely duration with a
greater chance than being above the most likely
duration. This is one of the characteristics of JRAP,
which makes it a pessimistic approach.
The corner values such as maximum, likely and
minimum durations of an activity can be determined
according to the type of key resources consumed for that
activity. The activities in construction projects can be
categorized as production-related, procurement-related
and administrative activities. The production-related
activities can either be labor-driven or equipment-
driven. According to this classication system, a variety
of methods like heuristics, statistical studies, cycle time
analysis, queuing theory and simulation are available to
calculate activity durations [10]. However, engineering
judgment, experience and intuition affect the consis-
tency of the results obtained by these techniques. It is
also possible to determine the corner duration values of
the activities judgmentally as it is applied in the case
study.
3.3. Step 3establishment of activity-risk factor matrix
In this step of the process, activity-risk factor matrix
is established. Activity-risk factor is the percentage
effect of each risk over each activity (refer to Table 2). In
activity-risk factor matrix, the varying effect of each risk
over each activity is quantied in percentages. However,
the total inuence of all risk factors should be 100% on
any given activity [13]. This assessment is performed
mainly in a judgmental way. In other words, if previous
data is unavailable and a fully uncertain situation
is encountered, a fully judgmental solution will be
produced.
3.4. Step 4modelling and simulation
The fourth JRAP step is modelling and simulating the
schedule network of the project. Deterministic schedule
model developed in project planning software, package
1247
programs performing spreadsheet modelling and MCS
are the main tools of this step.
At this step, MCS utilizes the activity-risk factor
matrix developed at the third step in order to calculate
the variation in activity durations. In order to enable
MCS to calculate the activity durations in a stochastic
manner, the following pessimistic equation (Eq. (1)) is
developed after some reformation on Dawoods [13]
equation (Eq. (2)).
Duration of activity MAXMin:Time
Risk Factor Affect;
Max:Time
 Risk Factor Effect;
Duration of activity MAXMin:Time
Max:Time  Min:Time
 RF1  Random1 RF2
 Random2 RFn
 Randomn ; Max:Time
 Max:Time  Min:Time
 RF1  Random1 RF2
 Random2 RFn
 Randomn ; 1
where Min.Time is the minimum duration that an
activity can be completed, Max.Time, the maximum
duration that an activity can be completed, RFn, the
percent effect of nth risk factor on an activity (taken
from activity-risk factor matrix), Randomn is a random
number, between 0 and 1, generated during MCS to
represent the nth risk factors probability distribution.
Duration of activity Min:Time Max:Time
 Min:Time  RF1
 Random1 RF2
 Random2 RFn
 Randomn : 2
Eq. (1) (activity duration equation), which is an
important fragment of the JRAP model, consists of two
main parts of which one of them is already Eq. (2) itself.
For this reason, the nal result of Eq. (1) is the
maximum of the results of these two parts. While Eq. (2)
calculates the duration of the activity by adding the risk
factor effect to the minimum activity time; Eq. (1), in
contrast, calculates also the subtracted value of the same
risk factor effect from the maximum duration and
eventually gives the maximum of the two results as the
activity duration. This approach provides to stand at the
worse side and creates a pessimistic view. This is one of
the cornerstones of the JRAP.
Now, as an example, assume that an activity of a
construction project has a maximum duration of 60 days
and a minimum duration of 30 days. Next, assume that
 ARTICLE IN PRESS
1248 A. Oztas- , O. Okmen / Building and Environment 40 (2005) 12441254

Table 1
Computations showing four manually conducted Monte Carlo iterations

Iteration number Data Computation

1 Random1 1.00 RF1 20%

Random2 0.00 RF2 20% Risk factor Effect 14.7
Random3 0.60 RF3 40% Duration of activity 45.3 days
Random4 0.25 RF4 20%
2 Random1 1.00 RF1 20%
Random2 1.00 RF2 20% Risk factor effect 16.2
Random3 0.00 RF3 40% Duration of activity 46.2 days
Random4 0.70 RF4 20% Max. duration 60 days

3 Random1 0.50 RF1 20% Min. duration 30 days

Random2 0.35 RF2 20% Risk factor effect 15.3
Random3 0.40 RF3 40% Duration of activity 45.3 days
Random4 0.90 RF4 20%
4 Random1 0.00 RF1 20%
Random2 0.10 RF2 20% Risk factor EFFECT 9.0
Random3 0.30 RF3 40% Duration of activity 51.0 days
Random4 0.80 RF4 20%

the random numbers produced from the distributions
assigned to four risks are (1, 0, 0.6, 0.25), (1, 1, 0, 0.7),
(0.5, 0.35, 0.4, 0.9) and (0, 0.1, 0.3, 0.8) to be used during
four manual iterations of MCS. Besides, from activity-
risk factor matrix, the weather has 20%, design changes
has 20%, soil conditions has 40% and labor productiv-
ity has 20% effect on the duration of the assumed
activity. All of this information is enough to calculate
the duration of the activity using Eq. (1). The computa-
tions are summarized in Table 1.
Nevertheless, manual application of MCS is incon-
sistent while computer programs are capable of doing
the same job faster, more precisely and with much more
running preferences. Furthermore, such programs pro-
vide an easy and effective environment for data entrance
and worksheet modelling.
The assessment of correlation between risk factors is
as important as the identication of risks and is more
important than deciding on probability distributions of
risk elements. For this reason, any risk model should
consist of the correlation coefcients, if correlation
exists. Otherwise, the simulation results would not be
realistic.
4. Case study
4.1. Purpose
The purpose of this case study is to give an example to
the application of JRAP, to show how it works and to
prove its validity. For achieving these goals and for the
sake of comprehensiveness, a small and real completed
construction project at which one of the authors
involved throughout the whole process has been chosen.
4.2. Project definition
The project is the decoration of a bank ofce, which is
located at a rural area of the South-Eastern part of
Turkey. The contract used is an admeasurement type
with xed unit-price including specic clauses offered by
the promoter. The promoter is the States biggest bank.
The project start date was 30th November 1999 and the
completion date was 25th January 2000 in the contract.
In other words, the contractual project duration was 57
days. However, the project has been completed in 122
days, i.e. 65 days after the project deadline negotiated
within the contract documents.
4.3. Scope and major project risks
Since the project is the decoration of a bank ofce, the
usage of some special decorative materials is unavoid-
able. These special materials might be related to
electrical or mechanical services, or related to internal
or external nishes according to the content of the
architectural design. This is a risk for the contractor
from both cost and duration point of views because the
location of the construction site is far from the main
manufacturers, which exist in the western industrialized
regions.
The second important risk creating characteristic of
the project is that the possible material changes
requested by the contractor after the approval of the
contract is not permitted. This is due to the image
sensitivity of the architects of the bank, i.e. a standard
image style has been determined for the bank and all
ofces have to be constructed according to this standard
all over the country. This situation creates an inexible
environment for the contractor.
 ARTICLE IN PRESS
A. Oztas- , O. Okmen / Building and Environment 40 (2005) 12441254
Next, the contract is an admeasurement type with
xed unit price, which can be considered as a risky
contract strategy from the contractors point of view.
The main reason of this is that the payment is at the end
of each month after the measurement of the completed
parts but the promoter does not compensate cost
escalation, although there is high ination in the country
and interest rates are very high. Furthermore, transport-
ing cost is not paid to the contractor. The states bill of
quantities and the banks own unit prices for special
activities are used in measuring payment certicate
values.
In case of late completion of the project, 0.2% of
tender price is cut off as a liquidated damage for each
day and this issue has been mentioned strictly in
contract documents. All of these oblige the contractor
to complete the project as early as possible before the
project deadline, which is committed within the con-
tract, because of the time value of money and the
reputation of the company.
Fortunately, weather conditions are not very effec-
tive on the construction activities because most of them
will be carried out in a closed environment; at the
ground and base oors of a multi-storey building.
Besides, soil condition is not a risk factor for the same
reason.
Although the construction site is at a region that has a
high human-resource potential, working quality is not
high. The abundance of this resource means that there
are no human-resource allocation and levelling pro-
blems during scheduling. In other words, the activities
are mainly material driven, not labor driven.
4.4. Schedule characteristics
The project start date was 30th November 1999 and
the completion date was 25th January 2000 in the
contract. The completion date has been determined by
negotiation between the promoter and the contractor
without preparing any deterministic schedule!
In this context, rst, deterministic schedule of the
project is developed using MS Project planning soft-
ware. Likely durations of the activities (refer to Table 5)
are determined whether judgmentally or by examining
previous similar projects. Experience has also made it
easier to construct the network logic and activity
interrelationships. Finally, the deterministic plan has
shown that the likely project completion date was 12th
February 2000, which was 17 days after the contractual
completion date. This shows that even the deadline
determined without risk analysis is a later date than the
contractual completion date. It should be noted that
JRAP will also provide the opportunity of comparing
this deterministic deadline with the other stochastic
deadline values and show whether this deterministic
value is optimum or not.
1249
4.5. JRAP application
The rst task undertaken in application of JRAP on
this project is determining the critical risks of the project
because the rst step of JRAP is the identication of
risks that are effective on critical activities explored
during the critical path analysis (refer to Fig. 1). The
critical risks of the project are determined as: defective
design, design changes, subcontractors default, uctua-
tion in labor productivity, delay in resolving disputes,
difculties/delays in availability of materials, equipment
and labor, inadequate quality of work and need for
correction, promoter delays (unable to get approvals,
lack of payment, etc.), and changes in quantity/scope of
work [13,18].
Three different computer package programs namely
MS Project, MS Excel and Crystal Ball have been used
during this JRAP application. Before producing the
spreadsheet risk model, activity-risk factor matrix has
been formed. Table 2 presents activity-risk factor matrix
of the bank decoration project. Afterwards, the remain-
ing steps of JRAP, which are assigning probability
distributions and maximumminimum durations, and
modelling and simulating have been carried out
subsequently. The statistical distributions assigned to
risks and the correlation coefcients between them are
given in Tables 3 and 4, respectively. Experience and
engineering judgment have been utilized for setting up
this data.
All the data gathered and processed up to here have
been utilized in setting the spreadsheet risk model so
that it has become possible to perform MCS on the
produced model by the risk analysis software, Crystal
Ball. Table 5 shows the main spreadsheet model
developed by the spreadsheet program (MS Excel) for
the current case study. Tables 24 contain and provide
the supportive data to Table 5. For instance, the cells
just near the means row in Table 3 are the assumption
cells as it is called in the program and here they
represent the probability distributions assigned to the
critical risks. The cells at the last column and the cell at
the bottom of the late nish time column of Table 5 are
called forecast cells and represent the resultant values
that vary with respect to the relevant assumption cells.
After running preferences and correlation coefcients
had been introduced into the risk analysis model,
simulation was conducted. The report produced by
Crystal Ball at the end of MCS mainly consists of
information about the assumption cells and forecast
cells. Indeed, desired and necessary answers mostly exist
among the parts related to forecast cells in the
simulation report.
The spreadsheet model in Table 5 has two kinds of
forecast cells: total project duration and critical activity
query cells. The cell at the bottom of the late nish time
column gives the total project duration and it is related
 1250
Table 2
Activity-risk factor matrix

Activity Risk no. 1 2 3 4 5 6 7 8 9 Total (%)

number
Activity name Defective Design Subcontractors Fluctuation in Delay in Promoter delays Difculties/ Inadequate Changes in
design (%) changes (%) default (%) labour resolving disputes (unable to get delays in quality of work quantity/scope of
productivity (%) (%) approvals, lack of availability of and need for work (%)
payment, etc.) materials, correction (%)
(%) equipment and
labour (%)

1 Tender
2 Sanction 70 30 100
3 Site delivery 70 30 100

A. Oztas- , O. Okmen / Building and Environment 40 (2005) 12441254

4 Order materials 20 80 100
5 Procure materials 20 80 100
6 Remove existing old materials 70 30 100
7 Detail measurement; by aluminum 40 40 20 100
subcontractor
8 Order and procure aluminum 80 20 100
9 Manufacture aluminum joineries 10 10 20 10 10 20 20 100
10 Approval of weight records of 80 20 100

ARTICLE IN PRESS
aluminum joineries
11 Detail measurement; by glazier 40 40 20 100
12 Order and procure joinery glasses 80 20 100
13 Install electrical tting 20 10 10 30 20 10 100
14 Install suspended ceiling 10 10 20 10 20 20 10 100
15 Install HVAC, sanitary and 20 10 10 30 20 10 100
mechanical tting
16 Level slab cover sub-layer 60 40 100
17 Cover slab 10 30 40 10 10 100
18 Paint interior walls and columns 10 40 20 20 10 100
19 Cover WC walls and slab 10 40 30 10 10 100
20 Install aluminum joineries 20 30 20 30 100
21 Install joinery glasses 30 20 20 30 100
22 Install security alarm 30 20 20 30 100
23 Reinforcement and framework; 30 30 40 100
cash room walls
24 Approval and carriage of cash 20% 20 30 30 100
room door
25 Approval of reinforcement and 20 80 100
framework
26 Pour concrete 60 40 100
27 Remove frameworks 60 40 100
28 Cure concrete 60 40 100
29 Install cash room door 40 20 30 10 100
30 Paint walls and cover slab of cash 40 60 100
room
31 Delivery of money machine 100 100
32 Install money machine 40 60 100
33 Finish exterior works 10 10 30 40 10 100
34 Cover mechanical and electrical 30 30 30 10 100
units
35 Temporary commissioning 40 60 100
36 Approval of extra project duration 40 60 100
37 Approval of temporary 40 60 100
commissioning
Table 3
Probability distributions assigned to the risks

Risks: 1 2 3 4 5 6 7 8 9

Defective Design Subcontractors Fluctuation in Delay in Promoter delays Difculties/ Inadequate Changes in
design changes default labour resolving (unable to get delays in quality of work quantity/scope
productivity disputes approvals, lack availability of and need for of work
of payment, etc.) materials, correction

A. Oztas- , O. Okmen / Building and Environment 40 (2005) 12441254

equipment and
labour

Distribution Type Triangle Lognormal Triangle Triangle Triangle Lognormal Lognormal Triangle Triangle
Mean 0.46 0.39 0.37 0.40 0.37 0.44 0.44 0.42 0.41

ARTICLE IN PRESS
Table 4
Correlation coefcients assigned to the risks

Risks 1 2 3 4 5 6 7 8 9

Defective Design Subcontractors Fluctuation in Delay in resolving Promoter delays Difculties/delays Inadequate Changes in
design changes default labour disputes (unable to get in availability of quality of work quantity/scope of
productivity approvals, lack of materials, and need for work
payment, etc.) equipment and correction
labour

1 1.0
2 0.6 1.0
3 0.0 0.0 1.0
4 0.0 0.0 0.0 1.0
5 0.3 0.3 0.0 0.0 1.0
6 0.1 0.2 0.0 0.0 0.1 1.0
7 0.0 0.3 0.0 0.0 0.0 0.0 1.0
8 0.0 0.0 0.3 0.0 0.1 0.0 0.0 1.0
9 0.2 0.1 0.0 0.0 0.0 0.0 0.1 0.0 1.0

1251
 1252
Table 5
Spreadsheet risk model of the project

Time schedule risk analysis

Activity Activity name Predecessor and Lag Minimum Likely duration Maximum Simulated Early start Early nish Late start Late nish Total oat 0 critical
number relationship (days) duration (days) (days) duration (days) duration (days) time time time time (days) activity
1 non-critical
activity

1 Tender 0 0 0 0 0 0 0 0 0
2 Sanction 1 FS 2 2 4 3.3 0 3.3 0 3.3 0 0
3 Site delivery 2 FS 2 1 1 2 1.6 5.3 6.9 5.3 6.9 0 0
4 Order materialsa 2 FS 1 1 1 1 3.3 4.3 44.9 45.9 41.7 1

A. Oztas- , O. Okmen / Building and Environment 40 (2005) 12441254

5 Procure materials 4 FS 10 15 40 27.8 4.3 32.1 47.2 75 42.9 1
6 Remove existing old materials 3 FS 2 3 4 3.1 6.9 10 6.9 10 0 0
7 Detail measurement; by aluminum 6 FS 1 1 2 1.7 10 11.7 10 11.7 0 0
subcontractor
8 Order and procure aluminum 7 FS 15 20 45 32.7 11.7 44.3 11.7 44.3 0 0
9 Manufacture aluminum joineries 8 FS 15 20 30 24.2 44.3 68.6 44.3 68.6 0 0
10 Approval of weight records of 9 FS 1 1 1 1 68.6 69.6 68.6 69.6 0 0

ARTICLE IN PRESS
aluminum joineries
11 Detail measurement; by glazier 10 FS 1 1 2 1.7 69.6 71.3 69.6 71.3 0 0
12 Order and procure joinery glasses 11 FS 10 15 30 21.8 71.3 93.1 71.3 93.1 0 0
13 Install electrical tting 6 FS 2 3 5 3.7 10 13.7 61.9 65.6 52 1
14 Install suspended ceiling 13 FS 3 4 10 7.3 13.7 21 65.6 73 52 1
15 Install HVAC, sanitary and 6 FS 4 5 8 6.3 10 16.3 66.7 73 56.7 1
mechanical tting
16 Level slab cover sub-layer 14,15 FS 1 2 3 2 21 23 73 75 52 1
17 Cover slab 5,16 FS 1 2 4 2.6 32.1 34.7 75 77.6 42.9 1
18 Paint interior walls and columns 17 FS 3 3 4 10 6.6 37.7 44.3 80.6 87.2 42.9 1
19 Cover WC walls and slab 5,15 FS 1 2 3 2.1 32.1 34.1 85.2 87.2 53.1 1
20 Install aluminum joineries 10,18,19 FS 3 4 8 5.9 69.6 75.5 87.2 93.1 17.6 1
21 Install joinery glasses 12,20 FS 2 3 6 4.7 93.1 97.7 93.1 97.7 0 0
22 Install security alarm 21 FS 1 2 3 2.2 97.7 99.9 97.7 99.9 0 0
23 Reinforcement and framework; 6 FS 2 2 4 3.1 10.0 13.1 75.2 78.3 65.3 1
cash room walls
24 Approval and carriage of cash room 4 FS 2 2 4 3.3 4.3 7.5 92.1 95.3 87.8 1
door
25 Approval of reinforcement and 23 FS 1 1 1 1.0 13.1 14.1 78.3 79.3 65.3 1
framework
26 Pour concrete 25 FS 1 1 1 1.0 14.1 15.1 79.3 80.3 65.3 1
27 Remove frameworks 26 FS 3 1 1 1 1.0 18.1 19.1 83.3 84.3 65.3 1
28 Cure concrete 27 FS 7 4 4 4 4.0 26.1 30.1 91.3 95.3 65.3 1
29 Install cash room door 24,28 FS 1 2 3 2.4 30.1 32.5 95.3 97.7 65.3 1
30 Paint walls and cover slab of cash 29 FS 1 2 3 2.2 32.5 34.7 97.7 99.9 65.3 1
room
31 Delivery of money machine 4 FS 20 30 60 43.5 4.3 47.8 45.9 89.4 41.7 1
32 Install money machine 31 FS 3 3 7 5.3 47.8 53.1 89.4 94.7 41.7 1
33 Finish exterior works 32,5 FS 3 4 7 5.2 53.1 58.3 94.7 99.9 41.7 1
34 Cover mechanical and electrical 15,5 FS 3 4 6 4.8 32.1 36.9 95.1 99.9 63.1 1
units
35 Temporary commissioning 22,33,34,30 FS 1 1 2 1.7 99.9 101.6 99.9 101.6 0.0 0
36 Approval of extra project duration 35 FS 1 1 2 1.5 101.6 103.1 101.6 103.1 0.0 0
37 Approval of temporary 36 FS 1 1 2 1.5 103.1 104.7 103.1 104.7 0.0 0
commissioning
 ARTICLE IN PRESS
A. Oztas- , O. Okmen / Building and Environment 40 (2005) 12441254 1253

Table 6 considered as a different approach to the analysis stage

Probable durations in percentiles and critical activities with 100% of risk management systems. Judgmental risk analysis
certainty
process is comprised of a number of steps as shown in
Forecast: total project duration Critical activities (with Fig. 1. The methodology is mainly based on MCS and
100% certainty) an equation that is proposed in the preceding sections.
This equation (Eq. (1)) enables the planner to observe
Percentile (%) Days Activity no.
the variation in activity durations during MCS. These
0 97.5 2 simulated values regarding the activity durations are
10 101.1 3 then utilized in the spreadsheet risk model through
20 102.6 6 arithmetic formulas that are produced according to the
30 104.0 7
logical relationships between the activities of the time
40 105.5 8
50 106.9 9 schedule network.
60 108.4 10 The reason for the proposed risk analysis method to
70 110.1 11 be qualied as judgmental is that the lack of historical
80 112.1 12 data related to the activities of the time schedule
90 114.6 21
network does not interfere with performing risk analysis
100 124.8 22
35 on the network, i.e. engineering judgment, experience
36 and intuition of the planner have all been utilized
37 through JRAP. Obviously, such an analysis strategy
would create additional risks due to the subjective data
used. Nevertheless, the pessimistic character of the
to the other varying cells with the formula within the proposed activity duration equation (Eq. (1)) and
frame of the activity network logic. Moreover, Eq. (1) is skewness in probability distributions assigned to critical
utilized in the simulated duration column. On the other risks compensate this undesired situation.
hand, the remaining forecast cells (the cells at the last The other objective of the present research was to
column) show how often each activity becomes critical illustrate the application of JRAP using a construction
so that it can be decided on which activities managerial project data. For this, JRAP was applied on a bank
effort should be focused and necessary precautions decoration project using the computer package pro-
should be taken in advance. grams, which are MS Project, MS Excel and Crystal
Ball. The results of the case study have shown that
4.6. Discussion of the results JRAP has been capable of answering the questions
such as:
Table 6 shows the summary of the results of the
simulation performed with 10,000 trials. It consists of
In how many days and with how much probability it
the project duration frequency values and the critical is possible to complete the project?
activities with 100% certainty. For instance, the project
Which activities are the most critical activities?
can be completed in 110.1 days with 70% probability
Which risk elements are more effective on project
according to these results. If at least this extra duration?
information had existed prior to the sanction, project
overrun would not have occurred. In other words, the The answers to such questions are obviously useful to
risk percentiles of the project duration would have make sound decisions prior to the contracting. As in the
helped the parties agree on a more suitable and realistic case study, the project completion duration specied in
contractual deadline. Unfortunately, not carrying out the contract, which is 57 days, came out to be only a big
deterministic and stochastic network analysis of the mistake because the results of JRAP has shown that the
project prior to the sanction resulted in late completion total project duration is 114.6 days with 90% and 124.8
of the current project. That is, the project was completed days with 100% probability. The commissioning docu-
in 122 days, i.e. 65 days after the project deadline written ments of the project have already supported this result
in the contract documents. because the construction has been completed in exactly
122 days, i.e. 65 days after the negotiated contract
deadline.
5. Conclusion As a general concluding remark, JRAP stands as a
practical schedule risk analysis methodology in the
The aim of this paper was to propose a time schedule overall risk management framework and it also con-
risk analysis methodology that is applicable on con- tributes to the dynamism of the risk management system
struction projects. This methodology, which is named as through its stochastic characteristics. However, it is
Judgmental Risk Analysis Process (JRAP), can be open to development through its application over more
 ARTICLE IN PRESS
1254 A. Oztas- , O. Okmen / Building and Environment 40 (2005) 12441254
complex construction projects of longer durations from
different categories. Furthermore, the methodology can
be modied and converted into a more data (past
record)-dependent characteristic by adaptation of mod-
ern techniques like genetic algorithms, articial neural
networks and expert systems.
References
[1] Edwards L. Practical risk management in the construction
industry. London: Thomas Telford; 1995.
[2] Flanagan R, Norman G. Risk management and construction.
Cambridge: Backwell Scientic; 1993.
[3] Pilcher R. Project cost control in construction. London: Collins;
1985.
[4] Isidore LJ, Back WE. Multiple analysis for probabilistic cost and
schedule integration. ASCE Journal of Construction Engineering
and Management 2002;128(3):2119.
[5] Barraza GA, Back WE, Mata F. Probabilistic monitoring of
project performance using SS-curves. ASCE Journal of Con-
struction Engineering and Management 2000;126(2):1428.
[6] Chehayeb NN, AbouRizk SM. Simulation-based scheduling with
continuous activity relationships. ASCE Journal of Construction
Engineering and Management 1998;124(2):10715.
[7] Senior BA, Halpin DW. Simplied simulation system for
construction projects. ASCE Journal of Construction Engineering
and Management 1998;124(1):7281.
[8] Maio C, Schexnayder C, Knutson K, Weber S. Probablity
distribution functions for construction simulation. ASCE Journal
of Construction Engineering and Management 2000;126(4):
28592.
[9] Fente J, Schexnayder C, Knutson K. Dening probablity
distribution function for construction simulation. ASCE Journal
of Construction Engineering and Management 2000;126(3):
23441.
[10] Grifs FH, Farr JV. Construction planning for engineers.
Singapore: McGraw-Hill; 2000.
[11] Illumoka AA. A tolerance analysis approach to network
scheduling for engineering project management. International
Journal of Production Research 1987;25(4):53147.
[12] Chapman CB. A risk engineering approach to project risk
management. International Journal of Project Management
1990;8(1).
[13] Dawood N. Estimating project and activity duration: a risk
management approach using network analysis. Construction
Management and Economics 1998;16:418.
[14] Ranasinghe M. Qualication and management of uncertainty in
activity duration networks. Construction Managemant & Eco-
nomics 1994;12(1):1529.
[15] Okmen O. Risk analysis and management of construction projects
tendered under design-build (turnkey) contract system. M.Sc.
Thesis. Gaziantep: University of Gaziantep, 2002.
[16] Kartam NA. Risk and its management in the Kuwaiti construc-
tion industry: a contractors perspective. International Journal of
Project Management 2001;19(4):32535.
[17] Oztas- A, Okmen O. Risk analysis in xed-price design-build
construction projects. Building and Environment 2004;39:22937.
[18] Gursel AP. Risk perception and trends of Turkish construction
companies. M.S. Thesis. Ankara: Middle East Technical Uni-
versity, 1998.