46

IEEE TRANSACTIONS ON ENGINEERING MANAGEMENT, VOL. 38, NO. I , FEBRUARY 1991

Project Risk Assessment Using the

Analytic Hierarchy Process
Mohammad A. Mustafa, Member, ZEEE, and Jamal F. Al-Bahar

Abstract-Often, construction projects fail to achieve their time,

budget, and quality goals. This is frequently due to the failure of the
contractor to analyze and assess unanticipated risks. The analytic hierarchy process (AHP) is a new approach that can be used to analyze and
assess project risks during the bidding stage of a construction project
and to overcome the limitations of the traditional approaches currently
used by contractors. The AHP presents a flexible, easily understood way
to assist the decision-maker in formulating his problem in a logical and
rational manner. The paper also includes a review of the AHP and its
application in the assessment of the riskiness of constructing the Jamuna
Multipurpose Bridge in Bangladesh.
Keywords-Risk management; project management; construction
management; risk assessment; the analytic hierarchy process; risk models; risk analysis; decision modeling.

I. INTRODUCTION

ONSTRUCTION, like many other industries in a free-enterprise system, has sizeable risk built into its profit structure [3]. From beginning to end, the construction process is
complex and characterized by a number of uncertainties. For
example, uncertainty about weather conditions, subcontractor
failure, and diffemt site conditions are typical risk variables that
exist in every construction project. As a result, many construction projects fail to achieve their time, cost and quality goals.
Most contractors, however, have developed a series of rules
of thumb to analyze and assess risk. These rules generally rely
on the contractors experience and intuition. Rarely do contractors quantify uncertainty and systematically assess the risks
involved in a project [l]. One reason might be the lack of a
rational, straightforward approach to combine all the facets of
risk systematically into a prioritized and manageable scheme.
However, as construction projects become more uncertain and
complex, intuition and tested rules of thumb often fail to anticipate and respond effectively to the extent of uncertainty and risk
in construction projects. Therefore, the need for a logical and
rational risk assessment procedure rather than tested rules of
thumb has increased [ 11.
Recently, a number of systematic models have been proposed
for use in the risk evaluation phase of the risk management
process. Kangari and Riggs [8] classify these methods into two
overall categories: 1) classical models (i.e., probabilistic analysis); and 2) conceptual models (i.e., fuzzy set analysis). Examples of classical models include Monte Carlo simulation [l] and
influence diagrams [4], [l]. An example of conceptual models is
fuzzy sets [8]. Kangari and Riggs [8] note that probabilistic
models suffer from two major limitations:
1) Some of these models require detailed quantitative inforManuscript received September 25, 1989; revised January 23, 1990. The
review of this paper was processed by Editor D. F. Kocaoglu.
The authors are with Tecno-Economics Division, Kuwait Institute for
Scientific Research, P.O.Box 24885, 13109 Safat, Kuwait.
IEEE Log Number 9037562.

mation which is not normally available in the real construction world.

2 ) The applicability of such models to real construction risk
analysis is limited. This is mainly due to the fact that many
of the contractors decision problems are imprecise, ill-defined, and vague in nature. Such characteristics are mostly
subjective in nature while classical models cannot handle
subjectivity.
The object of this paper is to introduce a new approach for
project risk assessment through the analytic hierarchy process
(AHP). The AHP provides a flexible and easily understood way
to analyze project risks. It is a multi-criteria decision analysis
methodology that allows subjective as well as objective factors
to be considered in the process which is precisely what is
needed. In this respect, the new approach is aimed at augmenting and enhancing intuition rather than replacing it. The AHP
gives managers a more rational basis on which to make decisions.
The AHP is used here to assist a contractor in the evaluation
of the riskiness of a project on which there is bidding. Thus, the
AHP will be used to provide a methodology for risk analysis and
assessment. As described before, the use of AHP allows the
management team to document and communicate an explicit,
common, and shared understanding of the degree of a projects
riskiness. In this way, the AHP becomes a living picture of the
managements understanding of the projects risks.
AHP has been applied in different fields [2], [ 5 ] , [9], [lo],
[ 121- [20]. In the field of project /construction management,
AHP has been applied in the evaluation of bidders (or biddings)
[13], and in the selection of the best crashing scheme when
factors other than time and cost are considered [14]. No prior
work has been conducted on the utilization of the AHP in
construction projects risk management. This paper attempts to
provide a basic application of the AHP in risk assessment. In the
next section we introduce the AHP. In Section 111, we give a
classification scheme of the risks encountered in a project. In
Section IV, we apply the AHP to assess the riskiness of constructing a bridge project in Bangladesh. In the last section, we
give a summary of the paper together with a discussion of the
efficient use of the AHP.

11. THEANALYTIC
HIERARCHY
PROCESS
The AHP developed by Saaty [16] is a robust and flexible
multi-criteria decision analysis methodology. Formulating the
decision problem in a hierarchical structure is the first and
probably the most important step. The hierarchy should be
constructed so that elements at the same level are of the same
order of magnitude and must be capable of being related to some
or all elements in the next higher level. In a typical hierarchy,
the top level reflects the overall objective (focus) of the decision
problem. The elements affecting the decision are represented in

0018-9391/91/0200-06$01.~ @ 1991 IEEE

MUSTAFA

intermediate levels. The lowest level comprises the decision

hierarchy provides a clear and simple

The General Progenitorsof Risk

options. This type of

illustration of all the factors affecting the decision and their

relationships. Once the hierarchy has been constructed, the
decision maker begins the prioritization procedure to determine
the relative importance of the element in each level of the
hierarchy. Elements in each level are pairwise compared with
respect to their importance in making the decision under consideration. The comparison takes this form: How important is
element 1 when compared to 2 with respect to a specific element
in the immediately higher level? For each level, starting at the
top of the hierarchy and working down, a number of square
matrices are formed from the results of comparing the elements
of that level with respect to an element in the level immediately
above. The elements are arranged into homogeneous groupings
on clusters.
The decision maker can express his preferences between
every two elements verbally as: equally preferred (or important,
or likely), moderately preferred, strongly preferred, very
strongly preferred or extremely preferred. These descriptive
preferences would then be translated into absolute numbers [I],
131, [5], 171 and 191, respectively, with [2], [4],161, and [8] as
intermediate values for compromise between two successive
qualitative judgments. The verbal scale used in AHP enables the
decision maker to incorporate subjectivity, experience and
knowledge in an intuitive and natural way.
After forming the comparison matrices, the process moves to
the phase of deriving relative weights for the various elements.
The relative weights of the elements of each level with respect to
an element in the adjacent upper level are computed as the
components of the normalized eigenvector associated with the
largest eigenvalue of their comparison matrix. The composite
weights of the decision alternatives are then determined by
aggregating the weights through the hierarchy. This is done by
following a path from the top of the hierarchy to each alternative
at the lowest level and multiplying the weights along each
segment of the path. The outcome of this aggregation is a
normalized vector of the overall weights of the options. The
mathematical basis for determining the weights has been established [16]. Questions raised about efficient use of AHP are
addressed in the last section.
The use of the AHP to model and analyze real world problems
can be made much easier using a microcomputer implementation
of the method such as Expert Choice 161, [ 151. It makes structuring and modifying the hierarchy simple and quick. It eliminates
tedious calculations.

111. IDENTIFICATION
AND CLASSIFICATION
OF PROJECT
RISKS

We now classify the various potential sources of risk in

construction projects. Fig. 1 summarizes the general categories
of risk. By considering these categories separately one can
develop a detailed classification scheme of a projects risks.
Fig. 2 is a proposed classification scheme that shows the
different sources of risk in construction projects. The proposed
classification scheme is composed of six risk categories:
acts of God risks;
physical risks;
financial and economical risks;
political and environmental risks;
design risks;
job site-related risks.

-7

47

AND AL-EAHAR: PROJECT RISK ASSESSMENT USING ANALYTIC HIERARCHY PROCESS

Njture
naM,

n
1

Environment

Society

Fig. 1. The general progenitors of risk.

While the list of potential risks in every category is neither

complete nor exhaustive, it represents most of the typical risks
associated with a project. It would be impractical to describe
every possible risk. Instead, we focus our attention on the details
of general categories of risk.

Acts of God Risks

The risk of acts of God,or sometimes referred to as force
majeure is a term appropriated, largely by the insurance industry, to describe events that are unpredictable and simply beyond
anyones direct control. These events occur as a result of nature
and are often referred to as natural phenomena. The common
risks under this category are those related to physical damages
and personal injuries due to earthquakes, floods, fires, landslides, etc. force majeure is a term commonly used in contracts
and normally covers war and civil commotion rather than embargoes and sabotage, etc.

Physical Risks
The typical risks that fall under this category are associated
with damage of a property or asset that the contractor owns or
has under his possession. Such risks include: damage to structure or property; damage to equipment and material; labor
injuries and death; and forcible prevention of work or access to
work.

Financial and Economic Risks

Most risks that evolve in construction projects are financially
related. Project funding is obviously a potential economic risk
for contractors. Inadequate sources of project funds by an owner
or funding agent may create time delays and financing problems
which to many contractors are unbearable. The owner must have
enough funds to complete the work, and must make these funds
available to the contractor in a suitable manner and time that
enables the contractor to proceed with the work. Another potential risk, which is not frequently mentioned, is the risk of
financial failure by one of the sub-contractors. While infrequent,
the order of magnitude of the consequences of such failure needs
to be considered.

Political and Environmental Risks

These risks may arise from the interactions between the
contractor, the host government and the surrounding environment or society. Typically, political risks are salient in foreign
operations or international projects. The main risks in this area
include: expropriation of the contractors equipment by the host
government; customs and export/import restrictions on imported
materials; and local laws and environmental control regulations.

48

IEEE TRANSACTIONS ON ENGINEERING MANAGEMENT, VOL. 38, NO. I , FEBRUARY 1991

23

I1
Acts of God

Financial B
Economic

Physical

Flood
Earthquake
Landslide
Fire
Wind Damage
. Lighting
- Weather

Damage to
Structure
Damageto
Equipment
Labor Injuries
Material and
Equipment Fire
and Theft

Risk Categories In a
Construction Project

Inflation
Availability of
Funds From
Client
Exchange
Rate
Fluctuation
Financial
Default of
Subcontractor
NonConvertibility

,
. Changesin
l a w s and
Regulations
War and
Civil Disorder
- Requirement
for Permits
and Their
Approval
Pollution and
Safety Rules
. Expropriation
- Embargoes

Incomplete
Design Scope
Defective
Design
- Errorsand
Omissions
- Inadequate
Specifications
- Design
Changes

Labor Dispute
and Strike
Labor
Productivity
Different Site
Conditions
Defective
Work
Equipment
Failure

Fig. 2 . The proposed risk classification scheme.

Design Risks
Risks normally incurred by the design professional include
defective design, ambiguous specifications and plans, errors and
omissions in design, inaccurate geological and geotechnical exploration, and interaction of design with methods of construction.

Job Site-Related Risks

Each sector of the construction industry has its own job
site-related risks. Different site conditions, for example, are a
major determinant of work progress in tunneling and earth
moving contracts. On the other hand, building contractors may
find labor productivity and equipment breakdown as a major
determinant in schedule delays [2]. Hence, job site-related risks
are numerous owing to the different operations in construction
projects. The following are potential risks that are considered
typical in every construction project:
availability and productivity of labor;
soil, site and other changed conditions;
material shortages and quality.
Not every project involves all these kinds of risk. This
classification can be used as a guide to identify the potential risks
in a project.

IV . APPLICATION
We now give an application of the AHP to assess the riskiness
of an international construction project: The Jamuna Multipurpose Bridge in Bangladesh [I].

Project Background
The project involves constructing a 4 760-m long span across
the Jamuna river. The bridge will have multiple purposes: 1)

carrying road and rail traffic, 2) carrying high voltage electrical

cables and communication lines across the river. Designs are
being developed in steel and concrete for both the substructure
and superstructures of the main bridge. River training works
consist of guide bunds (dykes) at both river banks erected
perpendicular to the bridge axis, and having a length of approximately 2 800 m each. Major dredging works (approximately 30
million m3) are required to construct the slope protection of the
guide bunds to 30 m below the present flood plain. Floating
equipment and other modem off-shore construction techniques
with their associated hazards and risk will be used to build the
bridge.

The AHP Risk Assessment Model

In this section, the steps followed in applying the AHP to
assess the riskiness of the Jamuna Multipurpose Bridge project
are discussed.
Step I : Structuring the elements of the problem into a
hierarchy.
For this purpose, various factors that affect the expected level
of risk of the Jamuna Bridge Project are first identified. Because
of space limitation, the significant risk factors of only the three
most relevant risk categories have been selected for consideration in this demonstration. The following is a summary of these
factors.

Financial and Economic Risks

Because of the project's complexity, the large number of
contracts and the complications associated with involving international know how and equipment, it is critical to identify the
financial risks that could be faced in this project. Those considered here are:
Subcontractors Financial Default: Subcontractors financial
default may result in serious financial problems and time delays

MUSTAFA AND AL-BAHAR: PROJECT RISK ASSESSMENT USING ANALYTIC HIERARCHY PROCESS

49

sand, or erode the river embankments before they are completed.

Soil Subsidence and Collapse: Areas considered vulnerable
to subsidence and collapse are the approach roads, access routes,
and flood embankments. It is known that the soils at the bridge
location consist of loosely consolidated silty micaceous sand
deposits, which extend to great depths. During the lengthy wet
season, these soils will be relatively unstable, especially because
of the proximity of the river, which is a cause for concern
particularly for deep excavations. While the extent of the physical damage may be relatively small, the consequential delay
costs could be extensive.
These factors are incorporated in levels 2 and 3 of the
hierarchy as shown in Fig. 3. Level 1 represents the construction
firms overall goal, namely: The most likely risk in the bridge
project. Level 4 contains the three possible levels or intensities
of the total risk of the project. Although not in Figure 3 an
additional level of outcomes (severe; moderate; strong; fair;
weak; favorable; unfavorable) could be inserted and considered
between the subfactors level and the risks level.
Step 2: Develop the relative weights of the various elements.
The importance of the factors and subfactors and the likelihood of the levels of risk are determined next. For this purpose,
judgments are elicited from the management of the construction
firm and matrices of judgments are formulated. For example, in
order to determine the relative importance of the three factors of
the second level (financial, political and acts of God), a 3 x 3
matrix is formed as shown in Table I.
Political Risks
This matrix shows that Financial risks are judged to be weakly
The country of Bangladesh is not at war with any neighboring more important than political risks on the AHP verbal ratio scale
country, and the chances of a war breaking out in the region are presented earlier (equivalent to three times on the numerical
very small. Therefore, we only consider risk changes in local scale). The main reason for this is the managements concern
laws: civil disturbance, sabotage, changes in government poli- about the consequences of finance-related risks (especially the
cies regarding customs and export/import restrictions, embar- unavailability of funds from the government). The government
of Bangladesh is keen to see the bridge built and is not expected
goes and expropriation of contractors equipment and plant.
to change laws or regulations in a way that would affect the
success of the project. On the other hand, money is a very
Acts of God Risks
Equipment, materials used and the resulting structure are scarce resource in that country and the government might find it
subject to loss or damage during their transportation to the work necessary and opportune to shift money allocated to such a
site and during construction because of the following types of project to another activity.
The matrix also shows that when financial risks and acts of
risks:
God
risk are compared, the management was not sure whether to
Earthquakes: The selected bridge location at the town of
Sirajganj is situated in a moderately active seismic region. The judge it as strongly more important (five on the numerical scale)
consultants are aware of the possible impact of an earthquake on or very strongly more important (seven on the numerical scale).
the bridge, its foundations, the surrounding river training works The judgment in between was therefore chosen. The reason for
and embankments. During the construction phase, however, the low importance of acts of God risks is that the bridge will be
designed taking into consideration these kinds of factors. For
there are other exposures to earthquake risks that include:
similar reasons, political risks are judged to be five times more
the partially completed permanent works and how they can important than acts of God.
This matrix needed only three (in general n(n - 1)/2) judgwithstand all stresses within the design magnitude;
ments to be made which are enough to fill the triangle above the
the contractors temporary works;
diagonal. The diagonal elements of the matrix are each equal to
the sensitivity of soil to liquefication effects.
one (elements are compared with themselves), and lower trianWater Damage and Floods: The site of the project is subject gle elements of the matrix are reciprocals of upper triangle
to annual flooding when the rivers and their tributaries are elements.
A close examination of the judgments made to determine the
swollen by melting snow from the mountains and from monsoon
rains. Therefore, protection of camp sites located in the flood relative importance of the three factors considered shows that
plains and of river works are of paramount importance. Another management has not been fully consistent. Inconsistency is
source of concern is the possibility of the river overflowing to permissible in AHP so long as it does not exceed a ratio of about
outflank the construction and subsequently causing damage to 0.10 justified in the theory as a cut-off point. In the case of the
structures vulnerable to the force of water. An early or late flood matrix comparing the factors, the inconsistency ratio was 0.081.
Similar procedures are followed to elicit judgments on the
would break through into the excavated basin and fill it up with

for the prime contractor. The prime contractor should review

carefully the qualifications of the potential subcontractors. Negligence by subcontractors such as failing to complete their job or
to pay their bills may result in substantial delays, and consequently higher costs.
Unavailability of Funds: Unavailability of funds from the
government and untimely payments for the contractors are risk
factors that should be considered by a contractor. The project
will be financed by the World Bank with co-financing from other
international development financing agencies. However, there
are a number of conditions and restrictions on the Government
of Bangladesh regarding the use of funds allocated to this
project.
Inflation and Price Escalation: These risks are generally
speculative in nature. It is not possible to insure against cost
increases which arise purely as a result of inflation. For example, the quotes price for tabular steel piles and fabricated steel
box girders and other major items may fluctuate. The contractor
may, or may not, incur additional costs due to price changes in
such major construction items. In addition, charter rates for
floating derricks, dredging equipment, piling barges and associated operations also vary. Any increase in charter rates for those
heavy offshore equipment may result in substantial costs for the
contractor. In addition, the contractor should assess the loss of
use cost of such expensive equipment. Published price indices of
various construction materials and the general trend of worldwide inflation are valuable sources for predicting inflation and
assessing the potential for price escalation.

IEEE TRANSACTIONS ON ENGINEERING MANAGEMENT, VOL. 38, NO. 1, FEBRUARY 1991

50

Assess the R
I
Constructing a Bridge
Project

GOAL

FACTORS

LEVEL OF
RISK

Total Risk

Legend
F1 : Financialand Economic Risks

F11 : Subcontractors Failure

F12 : UnavailabilityOf Funds
F13 : Inflation

F3 : Acts of God Risks

F2 : Political Risks
F21 : Embargoes / expropriation
~ 2 :2Changes in Local ~ W S

F31:Eanhquakes
F32 : Water Damage and Floads
F33 : Soil Subsidence and
Collapse

Fig. 3. The AHP risk assessment model.

TABLE I1

TABLE I
JUM~MENTS MATRIX
AND WEIGHTS
OF FACTORS
With
Respect
to Goal

F1

F2

F3

Relative
Importance

F1
F2
F3

1
113
116

3
1
115

6
5
1

0.635
0.287
0.078

relative importance of the subfactors and the relative likelihood

of the levels of risk (high, medium and low total risk). Three
matrices of judgments on subfactors and eight matrices on the
three levels of risks are formed.
The relative importance of the various factors are then computed as the components of the normalized eigenvectors of the
matrices. For example, using the matrix shown in Table I, the
relative importance of the factors of risk are computed. Financial risks has the highest relative importance (0.635), followed
by political risks (0.287) and finally acts of God with a relative
weight of 0.078. The relative importance of the subfactors of
Level 3 with respect to the overall goal are shown in column 2
of Table II. The likelihoods of the various levels of risk given
the outcome of the subfactors with respect to the overall goal are
shown in columns 3-5 of Table II. The results show that
unavailability of funds is the most influential subfactor in determining the level of risk of this project with a relative importance
of 0.433, followed by changes in local laws subfactors 0.205.
Inflation and soil subsidence and collapse are seen by the management as the least important subfactors with priority 0.007.
Step 3: Synthesize and determine likelihoods of levels of
risks. In this step, the likelihoods of high, medium, and low total
risk are determined by aggregating the relative weights through
the hierarchy.
The results show that when all factors are considered with the
judgments made by the management of the firm, the project is
characterized as low risk (with a likelihood of 0.401) as shown
at the bottom of Table 11.
Sensitivity Analysis: The outcome of the analysis presented
above is highly dependent on the hierarchy established by the
management and the relative judgments made about the various
elements of the problem. Changes in the hierarchy or the

PRIORITIES
OF FACTORS.
SUB-FACTORS.
AND LEVELS
OF RISK
Factors
F1

F2

F3

Sub-factors

High

Levels of Risk
Medium

Low

0.099

0.031

0.020

0.041

0.121

0.271

0.005

0.017

0.030

0.008

0.019

0.055

0.040

0.147

0.018

0.005

0.015

0.002

0.035

0.010

0.004

0.002

0.005

0.001

0.235

0.364

0.401

0.635
F11

0.150

F12

0.433

F13

0.052

0.287
F2 1

0.082

F22

0.205

0.078
F3 1

0.022

F32

0.049

F33

0.007

Likelihood of level of risk

judgments may lead to a change in the outcome. Using Expert

Choice the sensitivity of the outcome to different changes can be
tested. For example, Fig. 4 shows the sensitivity of the outcome
to changes in the relative importance of the Financial risks
factor. With its current weight of 0.635, the project is judged to
be of low total risk. If the relative importance of this factor were
judged differently and its weight decreased to 0.55 or less, the
project would be described as a medium total risk project.
Sensitivity of the decision to other factors can be tested in a
similar manner.

V. SUMMARY
AND DISCUSSION
In this paper, we have investigated the subject of risk assessment in construction projects. A scheme of classifying the
variuos sources of risk has been developed. The AHP has been
introduced and applied in assessing the riskiness of constructing
The Jamuna Multipurpose Bridge. Risk factors significant to this
project are identified and incorporated in this assessment. Results show that the project is an overall low risk project.

51

MUSTAFA AND AL-BAHAR: PROJECT RISK ASSESSMENT USING ANALYTIC HIERARCHY PROCESS

1.oo

0.90
0.80

0.10

0.001

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8 0.9
Financial Risks

Fig. 4. Sensitivity of the levels of total risk to financial risks.

The AHP provides a valuable support for contractors in the

decision making process. It is a comprehensive framework for
thinking through the decision problem. The numerical outcomes
of the method are less important than the systematic thinking
environment it offers. While this paper presents the use of AHP
in the risk analysis and assessment stage of the projects risk
management process, it can also be used in evaluating alternative
responses to risk. The project manager might find the AHP
useful to support decision making in bid evaluation, equipment
selection, staff selection, corporate stability and competitiveness,
and other areas.
Some questions regarding the efficient use of the AHP have
been raised. The first concern is related to building the hierarchy
and the recommendation of (7 k 2) elements under any node in
order to preserve consistency [111, [ 161. This raises the question
of how to handle cases where the number of elements is greater
than 9. When the number of elements is large, they can be
grouped in clusters and then comparable clusters are merged.
One can use a first pass to arrange elements in ascending or
descending order according to the property at hand. This total
ordering is then decomposed into clusters beginning with the
largest element (for example) and including only few elements
(7 k 2) in each cluster. Then, the elements in the first cluster
are compared. To ensure that elements in two adjacent clusters
be comparable, the smallest element of the largest cluster is
included as the largest element of the adjacent cluster. The
derived values of the elements in the adjacent cluster are then
divided by the priority of the common element, and then all are
multiplied by the weight of that element in the larger cluster. In
this way, the priorities of the elements of both clusters become
comparable and the two clusters are then merged. The process is
then repeated by using the smallest element of the adjacent
cluster as the largest element of the third cluster. This sorting
and comparison must be carried out separately for each property. Hypothetical elements may have to be considered to make
possible the smooth transition from one cluster to another.
The second source of concern is the number of judgments
required to derive relative priorities. From traditional measurement of tangibles, it seems reasonable that AHP needs no more
than ( n - 1) judgments to relate one element to the remaining
( n - 1) from which one can automatically construct all other
comparisons by forming ratios. However, when we are dealing
with intangible criteria for which we have no measurement, we

are no longer certain of the precise correspondence of the

strength of a judgment to a numerical value, which represents
that judgment, nor are we certain of the judgment itself. Therefore, we should use redundant information in more than ( n - 1)
judgments to improve both the quality of the judgment and the
validity of the derived numerical scale. The latter validity of the
derived scale can be seen to depend on the amount of information represented in the fundamental scale. Thus we need a way
to estimate the cost of information and detemine how valuable it
is to produce a valid outcome. Harker [7] has developed such a
way. It would seem that a complex problem with which one has
had to contend for a long period of time, perhaps measured in
years, may require longer discussion and a larger number of
judgments and redundancy for its solution. Generating these
judgments may be a somewhat more protracted process than a
simple problem of choice. Therefore, the number of judgments
and their quality is an important concern in a complex decision.
The participants may have to meet more than once, or their
judgment may be canvassed through a carefully designed questionnaire. Eliciting relevant information is a scientific concern
not to be abrogated or aborted through slick rationalization.
To the practicing scientist and engineer, the use of personal
judgment in the AHP which involves interpretation of stimuli in
the form of comparisons may be construed as a limitation. Here
the individual uses his mind directly to create scales as outlined
in the AHP without the intervention of instruments of measurement. However, every one would agree that whatever its numerical outcome may be, whether in yards or meters, tons or
bushels, fahrenheit or Celsius, measurement itself must be interpreted by experts. On the other hand, most of the elements in the
AHP model are intangibles measured through judgment as there
is no other way for their calibration. A fundamental way to
check the soundness of such judgments elicited from a diversity
of experts is to calculate their consistency. By taking redundant
judgments from experts, cooperating together to align their
thinking, we can ensure a modicum of validity. The AHP has
also been extended to deal with situations where there are
conflicts in the judgments of experts [18].

REFERENCES
[l] J. F. AI-Bahar, Risk management in construction projects: A
systematic analytical approach for contractors, Ph.D. dissertation, Department of Civil Engineering, University of California,
Berkeley, 1988.

IEEE TRANSACTIONS ON ENGINEERING MANAGEMENT, VOL. 38, NO. 1 , FEBRUARY 1991

A. Arbel and A. Seidman, Performance evaluation of flexible

manufacturing systems, IEEE Trans. Syst., Man. Cybern.,
vol. SMC-14, no. 4, 1984.
D. B. Ashley, Construction project risk-sharing, Tech. Report
No. 220, Department of Civil Engineering, Stanford University,
Stanford, CA, 1977.
D. B. Ashley, Influence diagramming for analysis of project
risks, Project Management Journa1,vol. XV, no. 1, pp.
56-62, 1984.
J. F. Bord and A. Feinberg, A two-phase method for technology selection and system design, IEEE Trans. Eng. Manag.,
vol. 36, no. 1, p. 28, 1989.
E. Forman, T. L. Saaty, M. A. Selly, and R. Waldron, Expert
Choice. Pittsburgh, PA: Expert Choice, Inc., 1983.
P. T. Harker, The incomplete pairwise comparisons in the
analytic hierarchy process, Mathematical Modelling, vol. 9,
pp. 837-848, 1987.
R. Kangary and L. S . Riggs, Construction risk assessment by
linguistics, IEEE Trans. Eng. Manag., vol. 36, no. 2, pp.
126-131, May 1989.
R. Khorramshahgol, H. Azani, and Y . Gousty, IEEE Trans.
Eng. Manag., vol. 35, no. 4, pp. 265-270, 1988.
M. Liberatore, An Extension of the analytic hierarchy process
for industrial R&D project selection and resource allocation,
IEEE Trans. Eng. Manag., vol. EM-34, no. 1, pp. 12-18,
1987.
G. A. Miller, The magical number seven, plus or minus two:
Some limits on our capacity for information processing, Psychological Review, vol. 63, no. 2, pp. 81-97, 1956.
M. A. Mustafa, Microcomputer-based expert support system
for engineering project selection and resource allocation, in
Proc. Annual Metting of ASEM, St. Louis, MO, 1987.
M. A. Mustafa arid T. C. Ryan, Decision support for bid
evaluation, paper presented at ORSA 1TIMS Meeting, Vancouver, BC, Canada, May 1989.
M. A. Mustafa and E. L. Murphree, Decision support approach
for project compression, Project Management Journal, vol.
XX, no. 2, pp. 29- 34, 1989.
M. A. Mustafa and H. H. El-Agha, A spreadsheet implementation for a two-phase approach of the analytic hierarchy process,
unpublished manuscript, 1989.
T. L. Saaty, The Analytic Hierarchy Process. New York:
McGraw-Hill, 1980.
T. L. Saaty, Priority setting in complex problems, IEEE
Trans. Eng. Manag., vol. EM-30, no. 3, pp. 140-55, 1983.
T. L. Saaty and J. M. Alexander, Conflict Resolution, The
Analytic Hierarch,y Approach. New York: Praeger, 1989.

1191 W. C. Wedley, Consistency tests for incomplete AHP Matrices

-A comparison of two methods, in Proc. Administrative
Sciences Association of Canada-Management Science Division, June 1989.
[20] F. Zahedi, The analytic hierarchy process: A survey of the
method and its applications, Interfaces, vol. 16, pp. 96-108,
1986.

Mohammad A. Mustafa was born in Jordan in

1954. He received the B.Sc. degree in electrical
engineering from the University of Baghdad,
Baghdad, Iraq, in 1976, and the Masters and
Doctorate of Science degrees in engineering
management from George Washington University, Washington, DC.
Subsequently, he worked for seven years in
the construction industry in different capacities.
He then worked in the areas of benefit-cost
analysis, risk analysis, decision support systems, and decision analysis, as an Assistant Professor and a Consultant
for a number of international organizations, including the World Bank.
At present, he is an Associate Research Scientist at the Techno-Economics Division of the Kuwait Institute for Scientific Research.
His current research includes project evaluation, decision analysis and
policy planning especially in the areas of construction and telecommunication.

Jamal Ai-Bahar was born in Kuwait in 1960.

He received the B A . , M.S., and Ph.D. degrees in civil engineering, from Kuwait University, the University of Michigan, Ann Arbor,
and the University of California, Berkeley, respectively.
He was an Assistant Cost and Planning Engineer for two years, and at present he is engaged
in teaching and research at Kuwait University.
His current research interests are risk management, decision analysis, and application of
construction managem.ent principles in construction projects. He is a
member of ASEC and PMI.