This document provides an overview of the analytical hierarchy process (AHP), a multiple-criteria decision analysis technique for evaluating project alternatives. AHP involves structuring the decision problem into a hierarchy, measuring priorities through pairwise comparisons, and synthesizing the results. The document includes an example application of AHP and discusses its use in a case study evaluating design options for a grade separation on the California High-Speed Rail project.

1. Analytical Hierarchy Process
Decision making for
infrastructure alternatives
Authors
Sameer Deo
Duanne Gilmore

2. 32
Contacts
Sameer Deo
Senior Engineer
sameer.deo@arup.com
Duanne Gilmore
Associate
duanne.gilmore@arup.com
arup.com
Summary...........................................................................5
Introduction.......................................................................7
Analytical Hierarchy Process............................................9
Background....................................................................9
Structuring Complexity.................................................9
Measurement.................................................................9
Example.........................................................................11
Synthesis........................................................................11
Risks and Variables.......................................................14
Inconsistency Ratio.......................................................14
Sensitivity Analysis........................................................15
Rank Reversal................................................................15
Biases.............................................................................16
Case Study 1
California High-Speed Train Project/
Van Nuys Boulevard Grade-Separation Study.................21
Background....................................................................21
Methodology..................................................................22
Step 1–Structuring Complexity....................................22
Step 2–Measurement.....................................................22
Step 3– Synthesis............................................................24
Advanced Applications..................................................25
Grouptime......................................................................25
Case Study Conclusions................................................26
Findings.........................................................................27
Sources...............................................................................29
Contents

3. ©Arup
54
Engineers often make recommendations for project alternatives that
affect multiple stakeholders. Collaborative decision-making for project
alternatives can become challenging for engineers because of the many
opinions — the client, owner, public, architects, and other agencies
and stakeholders, all with different and often conflicting priorities.
The analytical hierarchy process (AHP), a multiple-criteria decision
analysis (MCDA) technique, offers a structured approach to decision
analysis in a multiple-criteria, multiple-stakeholder environment.
AHP begins with setting a goal, developing alternatives that may
be able to meet that goal, and then evaluating each alternative based
upon its ability to meet the stakeholders’ criteria. AHP is utilized in
many different fields, including economics, politics, finance, resource
allocation, conflict resolution, design, architecture, and transportation.
A case study on the California High-Speed Train project using AHP is
applied in this paper to determine the best possible project alternative.
Summary

4. A30 ©Anthony J. Branco
76
Not everything that counts can be counted
and not everything that can be counted, counts.
Albert Einstein
Infrastructure projects are large, complex, and expensive, and involve
collaborative decision-making among various entities, often with
incompatible ideas that lead to project delays and cost overruns. A
crucial problem engineers face when evaluating project options is
how to scientifically structure the decision-making process, quantify
qualitative human judgment, and then build consensus on a singular
engineering solution that would satisfy multiple stakeholders with
conflicting priorities. During our work on the California High-
Speed Train Project, Arup explored the use of an MCDA technique
as a possible solution for evaluating options for grade separations
(construction cost in excess of US$65b) along the high-speed rail line
in the Los Angeles metropolitan area. Through a case-study approach,
this paper is focused on the application of AHP in evaluating a single
grade separation with multiple design options against a set of criteria
and with multiple stakeholders’ input.
Introduction

5. 98
Analytical Hierarchy Process
has three primary functions: structure,
measurement, and synthesis.
Structuring Complexity
As illustrated in Figure 1, every element
of the hierarchy must ultimately be
related to an alternative. The relationship
must be direct or follow a path, but no
element should be isolated from the
alternatives. Careful consideration
must be given to develop performance
measures and priorities. Determining
which factors to include in the
hierarchical structure as “criteria” is the
most creative part of the AHP process
for the decision-maker.
Measurement
After structuring the problem, AHP
develops a set of priorities for the criteria
used to judge the alternatives using
the Fundamental Scale of Absolute
Numbers (see Table 1). The process of
prioritization resolves the problem of
dealing with different types of scales,
Figure 1: Typical AHP Structure
Goal
Criteria
Alternative #1 Alternative #2 Alternative #3
Criteria Criteria
Background
While teaching at the Wharton School,
Thomas Saaty was troubled by the
communication difficulties between
scientists and lawyers, noting their lack
of a practical and systematic approach
for priority-setting and decision-making.
He set out to find a way for ordinary
people to make complex decisions
(Gass). In his research, Saaty found that
humans often deal with complexity in a
hierarchical way and develop structures
through homogeneous clusters of
factors, with lower levels in the structure
corresponding to increased detail in
criteria.
In 1994, Saaty introduced AHP
as a pair-wise comparison tool for
multi-criteria decision-making. AHP
uses decision analysis mathematics
to determine priorities of various
alternatives using pairwise comparison
of different decision elements with
reference to a common criterion. AHP

6. 1110
since measurements on different
scales cannot be directly evaluated
or combined. AHP uses pairwise
comparisons — directly comparing one
criterion to another, deciding between
the two which is more preferred. AHP
uses pairwise comparisons to determine
which criterion is more important by
comparing each criterion to every
other criterion, one at a time; the extent
preferred is given numerical values
1 through 9 (1 = same importance, 9
= one is extremely more important
than the other). Through these
pairwise comparisons, AHP turns a
multidimensional scaling problem into a
one-dimensional scaling solution.
Example
Suppose a woman is in the market for a
new car and has narrowed her options
to either a fuel-efficient sedan or a
sports car. The decision will be based on
the following metrics: (a) cost, (b) gas
mileage, and (c) style. To the decision-
maker, cost is a little more important
Fundamental Scale
Intensity of
Importance
Definition Explanation
1 equal importance
Two elements contribute equally to
the objective.
3 moderate importance
Experience and judgment
moderately favor one element over
another.
5 strong importance
Experience and judgment strongly
favor one element over another.
7 very strong importance
One element is favored very strongly
over another, its dominance is
demonstrated in practice.
9 extreme importance
The evidence favoring one element
over another is of the highest
possible order of affirmation.
2,4,6,8* values used as a compromise
1.1 - 1.9
when activities are very close a
decimal is added to 1 to show
their difference as appropriate
The option to use decimals can
indicate the relative importance of
similar criteria.
Reciprocals
of above
If Activity A has one of the above non-zero numbers assigned to it
when compared with Activity B, then B has the reciprocal value when
compared with A.
Cost
Gas
Mileage
Style
Cost 1 3 1/7
Gas
Mileage
1/3 1 3
Style 7 1/3 1
Table 2: Pairwise Comparisons
Table 1: Fundamental Scale of Absolute Numbers
*A value range of 1 through 9 makes it simple for the decision-maker to differentiate among
levels of importance. In this model, there are five levels of clear importance, so odd numbers are
normally used. However, even numbers can be used to provide more nuance — if a criterion is
more than “moderately” more important but not “strongly” more important, a value of 4 would
be used.
than gas mileage, so cost compared
to gas mileage receives a value of 3.
Reciprocating the opposite, gas mileage
compared to cost receives a value of 1/3.
Using the Fundamental Scale of Absolute
Numbers, the paired comparisons of the
criteria to the goal are summarized in
Table 2. As shown, style is much more
important than cost while gas mileage
is a little more important than style. The
“1” values diagonally across the table
merely imply that any criterion has equal
importance to itself. Once the criteria
have been prioritized, the alternatives
are then ranked according to how they
perform on each criterion.
It is possible for comparisons to be
inconsistent with each other — in the
example above, cost was assessed as
more important than gas mileage while
gas mileage is more important than style,
and yet style is more important than
cost. This inconsistency will not inhibit
any calculations in the AHP method, but
the inconsistency may indicate that the
scores are not the decision-maker’s true
preferences.
Synthesis
Once comparisons are made, the results
need to be synthesized. Each criterion
is given a numerical weight, determined
through the steps shown in Table 3.
Then each weight is multiplied by a
reciprocal root of the number of criteria
in the decision — for this example, to the
power of 1/3 because there are 3 criteria.
The weight is then determined by each
criterion’s power to the 1/n divided by
the total in that column.

7. 1312
Now that the weights have been
established for each criterion, the
alternatives must compared.
Tables 4–6 show the calculations that
determine each car’s performance in
relation to the weighted criteria. Using
the same calculation as the criteria
weighting process above, these matrices
A B C D E F G
Cost
Gas
Mileage
Style
Product
B x C x D
Power
1/n E^(1/3)
Weight
F/Ftotal
Cost 1 3 1/7 0.429 0.754 0.245
Gas
Mileage
1/3 1 3 1.000 1.000 0.325
Style 7 1/3 1 2.333 1.326 0.431
Total 3.080 1.000
Sedan Sports car Product
Power or
1/n
Weight of
Performance
Sedan 1 3 3.000 1.732 0.750
Sports car 1/3 1 0.333 0.577 0.250
Total 2.309
Sedan Sports car Product
Power or
1/n
Weight of
Performance
Sedan 1 3 3.000 1.732 0.750
Sports car 1/3 1 0.333 0.577 0.250
Total 2.309
Sedan Sports car Product
Power or
1/n
Weight of
Performance
Sedan 1 1/2 0.500 0.707 0.333
Sports car 2 1 2.000 1.414 0.667
Total 2.121
Table 4: Alternatives’ Cost Performance Comparison
Table 5: Alternatives’ Gas Mileage Performance Comparison
Table 6: Alternatives’ Style Performance Comparison
Table 3: Synthesis of Criteria Weighting Process
Figure 2: Car Hierarchy
Figure 2: Hierarchy with Weights
demonstrate mathematically that the
involved parties believe that the sedan is
three times as cost-effective as the sports
car, has three times the gas mileage, but
only half the style.
AHP objectively advises the decision-
maker to buy the sedan instead of the
sports car.
• sedan (0.750)
• sports car (0.250)
Sedan performance = (0.245 × 0.750) + (0.325 × 0.750) + (0.431 × 0.333) = 0.571
Sports car performance = (0.245 × 0.250) + (0.325 × 0.250) + (0.431 × 0.667) = 0.430
• sedan (0.750)
• sports car (0.250)
• sedan (0.333)
• sports car (0.667)
Goal:
Buy a new car
Criteria:
Gas Mileage (0.325)
Alternative #1:
Sedan
Alternative #2:
Sports car
Criteria:
Cost (0.245)
Criteria:
Style (0.431)
Goal:
Buy a new car
Criteria:
Gas Mileage (0.325)
Criteria:
Cost (0.245)
Criteria:
Style (0.431)

8. 1514
Risks and Variables
Inconsistency Ratio
The inconsistency ratio of a pairwise
comparison matrix identifies the
uniformity of the decision-maker’s
judgments. If A>B and B>C, then
logically, A>C. However, in the example
above, cost was assessed as more
important than gas mileage and gas
mileage as more important than style,
but style as more important than cost.
The inconsistency ratio is a way of
quantifying how applicable this logic is
throughout the pairwise comparison.
AHP assumes that inconsistency is part
of the psychology of human decision-
makers. Extreme judgments, lack
of information, and the structuring
of the problem can all lead to large
inconsistencies. Inconsistency in the
way questions are answered often means
the “intuitiveness” of a result, a key
ingredient for consensus, is lost.
Sensitivity Analysis
Once an AHP problem has been
structured, judgments made, and
a preferred alternative identified,
sensitivity analysis can be performed.
A sensitivity analysis identifies how
sensitive a set of preferences, typically
the ranking of alternatives, are to altering
the criteria upon which they are based.
If the relative importance of a criterion
changes, then it follows that the preferred
alternative may also change. However,
if changing the relative importance of a
criterion has no impact on the preferred
alternative, then it can be reasonably
argued that that particular criterion can
be excluded and consideration given to
more important criteria. Analogous to a
construction schedule where activities on
the critical path warrant more attention
than those with “float,” criteria upon
which the alternatives are dependent
warrant more scrutiny than those that
have little impact on the end result.
Sensitivity analysis can also be used
to assess the impact of a particular
stakeholder or decision-maker on the
preferred alternative. When considering
the judgments from many stakeholders,
it is useful to determine how sensitive
a preferred alternative is to a particular
stakeholder, or to weight the relative
importance of stakeholder’s judgments.
In the context of infrastructure
engineering, sensitivity analysis is a
powerful technique for building consensus.
By mathematically demonstrating that
a particular criterion or decision-maker
has no impact on the ranking of the
alternatives, stakeholders can agree to
disregard that criterion or perspective in
the context of the decision at hand.
The structure of an AHP problem can
lead to inconsistency if it requires
judgment on criteria that are more than
one order of magnitude apart in terms
of their relative importance. The most
extreme judgment that can typically be
made is a nine, or within one magnitude
of difference.
Inconsistency up to 10% is acceptable.
Inconsistency ratios higher than 10%
may be indicative of poor structure,
ambiguity in the questioning, or
disingenuous responses from the
decision-makers. High inconsistency
should be identified, reviewed, and
resolved by improving the problem
structure, clarifying the question, and
revisiting the judgments. Once the
results have been checked for their
inconsistency ratio, a sensitivity analysis
can be performed.
Rank Reversal
If three alternatives are ranked A>B>C,
then adding a much lesser alternative
D should not impact the ranking of
alternatives A, B, and C. However, this
is not always true — the introduction of
a new alternative can change the ranking
of the original alternatives in unintuitive
ways. Rank reversal is a phenomenon of
AHP (and all other MCDA techniques)
where the relative ranking of a set of
alternatives will change whenever
another alternative is added or deleted.
Rank reversal can occur due to the
irrationality of the decision-maker or,
more problematically, as a function of
the mathematics behind AHP. Research
into the causes and potential mitigation
of rank reversal is inconclusive and
thus problematic and often ignored in
all but the most advanced applications
of AHP. Infrastructure engineers often
rearrange priorities and adding and
subtracting alternatives, due to changes
in scope, new information, and design
development. Thus it is important to
be aware of rank reversal as a potential
issue with AHP.
One way to avoid rank reversal is to
measure the alternatives on an absolute
scale rather than a relative scale. In other
words, an alternative is rated against
a benchmark or standard rather than
relative to other alternatives. While
absolute scales preserve rank regardless
of how many alternatives are added
or subtracted, they require a known
benchmark to derive scale. This limits
their functionality when the criteria to be
ranked are intangible, as is often the case
in engineering.
Gerald Desmond Bridge, Long Beach, CA ©Arup

9. 1716
©Arup

10. 1918
Relative measurement, or paired
comparisons, are most widely used
because many scales are intangible and
can only be relatively compared, not
absolutely.
Biases
When using pairwise comparisons
considerations of how we ask the
decision-maker to rank the criteria and
the effect it may have on the results.
First, the framing of the question affects
the answer. For example, within the
boundaries of a pairwise comparison, we
can ask of the decision-maker, “What is
more important: cost or environment?”
Another way to frame the question would
be to ask, “What is more important:
an increase in cost of 2% or obtaining
LEED Platinum status for the project’s
efforts to be sustainable?” The latter is a
more specific, less ambiguous, and well-
defined question, likely to get a different
answer than the former. Ambiguity
can also be used strategically — it can
Descriptive scale Numerical scale Value
None 0 0
A Tad 1 1/9 or 0.1111
Moderate 2 2/9 or 0.2222
Moderate to Good 3 3/9 or 0.3333
Good 4 4/9 or 0.4444
Good to Very Good 5 5/9 or 0.5556
Very Good 6 6/9 or 0.6667
Very Good to Excellent 7 7/9 or 0.7778
Excellent 8 8/9 or 0.8889
Outstanding 9 9/9 or 1.0000
Table 7: Descriptive versus Numerical Scale
promote engagement and discussion
among a group of decision-makers.
Secondly, the type of rating scale
we ask decision-makers to use can
affect results — a numerical scale will
provide outcomes very different from
a descriptive scale. While the decision-
maker’s ratings will inherently be
subjective, descriptive scales are likely
to be more inconsistent than numerical
or graphical scales. For example, a
numerical rating scale of 0 to 9 may
translate to a descriptive scale of “None”
to “Outstanding” with eight gradations
between.
In order to calculate a result, we must
assign a numerical difference between
items on the descriptive scale, for
example, between “Moderate” and
“Moderate to Good.” In Table 6 we
say that the numerical difference is
0.5556 – 0.4444 or 11%. Thus we are
saying “Moderate to Good” is 11%
better than “Moderate,” but how do we
know the decision-makers will assign
the same value to the descriptive scale?
Most software packages also allow
user-defined gradations, so rather than
assuming equal-increment jumps, we
can widen the gap between “Moderate”
and “Moderate to Good” to 18%,
assuming the sum of all gradations
equals 1, and the AHP methodology will
be unaffected. What will change is the
result.
All measurement in AHP is
fundamentally reliant on the scale, so
when structuring an AHP problem,
it is critically important to consider
its effect and the decision-maker’s
interpretation of it on the outcome.
Numerical or graphical scales will
provide the most transparency, but
descriptive scales can be more engaging
and can be intelligently applied to suit
the requirements of a problem, or can
be applied to lean towards a particular
alternative.
©Arup

11. 2120
Case Study 1
California High-Speed Train
Project/Van Nuys Boulevard
Grade-Separation Study
Background
The potential for a structured MCDA methodology such as AHP in
engineering is demonstrated in large infrastructure projects. The
California High-Speed Train Project is one example of a project where
AHP could be implemented to test its applicability and benefits in
alternatives evaluation on infrastructure projects.
Consider the 20-plus grade-separation schemes — in this hypothetical
example, each scheme has 5 distinct locations dependent on preferred
alignment and each location has 8 possible treatments. Thus there are
4,000 possible grade-separation schemes to evaluate. Assume each one
is ranked against 10 key criteria and by 8 executive stakeholders, all of
whom have a preference based on qualitative rather than quantitative
criteria. AHP provides a methodology to structure a problem of this
scale and complexity.
For this case study, a single grade separation, Van Nuys Boulevard, on
a preferred alignment is considered.
California High-Speed Rail ©NC3D Media California High-Speed Rail ©NC3D Media

12. 2322
Methodology
Step 1–Structuring Complexity
AHP begins with an objective, criteria,
and alternatives. The objective is to
find the best alternative for the grade
separation required to cross the high-
speed rail tracks at Van Nuys Boulevard,
in the San Fernando Valley, California.
The key criteria were established as
traffic circulation, right-of-way, utilities
and drainage, constructability, transit,
design compliance, and pedestrian
provision. Lastly, three preferred options
worthy of further consideration were
identified from the initial eight options
shown in Figure 4 — options 1, 2, and 6.
Step 2– Measurement
Once an objective, alternatives, and
criteria are defined, four decision-makers
individually made their judgments
regarding the relative importance of
the criteria and preferences among the
alternatives. The decision-makers made
the pairwise comparisons using the using
an absolute scale (1 to 9). These pairwise
comparisons result in a matrix for each
decision-maker, as shown in Figure 6.
Figure 5: Case Study Hierarchical Structure
Optimum Grade
Separation
VA1 VA2 VA6
Traffic
circulation
Right of
Way
Utilities &
Drainage
Constructability Transit
Compliant
design
Pedestrians
Figure 4: Grade-Separation Treatment Menu
PEDESTRIAN
TRAFFIC
DESIGN
TRANSIT
RIGHT-OF-WAY
CONSTRUCTABILITY
UTILITIES
PEDESTRIAN 1 4 3 1/2 1/4 ... ...
TRAFFIC 1/4 1 6 5 ... ... ...
DESIGN 1/3 1/6 1 2 ... ... ...
TRANSIT 2 1/5 1/2 1 ... ... ...
RIGHT-OF-WAY 4 ... ... ... 1 ... ...
CONSTRUCTABILITY ... ... ... ... ... 1 ...
UTILITIES ... ... ... ... ... ... 1
Figure 6: Pairwise comparison matrix

13. 2524
Step 3–Synthesis
Synthesis is the stage between making
pairwise comparisons and arriving
at a preferred alternative, essentially
the mathematical theory behind AHP.
Squaring the pairwise comparison
matrix and summing and normalizing
the rows results in a vector with 7 rows
(1 for each criterion), which represents
the relative weighting of the criteria.
This vector will always be 1 column by
n rows, where n equals the number of
criteria.
Once the pairwise comparisons are
synthesized for the criteria, resulting in
criteria vector C, we repeat the process
on the alternatives.
The alternatives’ performances
concerning each criterion are pairwise
compared to each other. The comparison
values are then synthesized to obtain
an n x m matrix, where n is the number
of criteria and m is the number of
alternatives. Thus we get a 3 x 7
alternatives matrix A, as shown in
Figure 8.
Multiplying the criteria vector C by the
alternatives matrix A, and normalizing
the result, produces a 3 x 1 vector R,
which represents the ranked alternatives.
R is the preference vector, which in
this case translates to 33.31% for VA1,
31.07% for VA2, and 35.62% VA6 —
thus VA6 is the preferred alternative.
The percentages demonstrate the relative
C =
0.2096 PEDESTRIAN
0.2053 TRAFFIC
0.1863 DESIGN
0.1803 TRANSIT
0.1408 RIGHT-OF-WAY
0.0447 CONSTRUCTABILITY
0.0329 UTILITIES
Figure 7: Criteria vector showing relative weighting of the criteria
A =
0.293
VA1
PEDS 0.113
VA1
TRAF 0.293
VA1
DESN 0.113
VA1
TRAN 0.396
VA1
ROW 0.220
VA1
CONS 0.703
VA1
UTIL
0.113
VA2
PEDS 0.040
VA2
TRAF 0.362
VA2
DESN 0.259
VA2
TRAN 0.431
VA2
ROW 0.362
VA2
CONS 0.255
VA2
UTIL
0.113
VA6
PEDS 0.293
VA6
TRAF 0.362
VA6
DESN 0.425
VA6
TRAN 0.113
VA6
ROW 0.293
VA6
CONS 0.255
VA6
UTIL
Figure 8: Alternatives matrix showing to what extent each alternative satisfies the criteria
C x A = R =
0.331 VA1
0.3107 VA2
0.3562 VA6
Figure 9: Rankings vector showing extent to which each alternative satisfies the objective
strength of the alternative in terms of
satisfying the objective, in this instance
the optimum grade separation. The
difference between the percentages of
the alternatives gives an indication of
how preferred they are. In this instance,
all alternatives are relatively close, a
reflection of the similarity of the impacts
of each alternative.
Advanced Applications
The results can be interrogated once
synthesis has been performed in order
to build consensus among the decision-
makers and apply sensitivity analysis,
altering our judgments and changing our
assumptions to evaluate the impact on
our preferred alternative.
In our case study, there were multiple
decision-makers. During synthesis this
is accounted for by using a mean average
of their individual judgments. Another
option at this stage would be to weight
the decision-makers themselves, making
one person more influential than another,
a situation inherent to any decision-
making process.
Sensitivity analysis can also be
performed to determine how sensitive
our ranked alternatives are to the
pairwise comparisons, for example,
by selectively excluding particular
criteria or decision-makers’ judgments.
Sensitivity analysis can explore a
wide range of methods, in which case
software proves critical to avoiding
laborious calculations. In this case study,
hand calculations have been employed
in parallel to running the software
to confirm our methodology, but for
problems with larger numbers of criteria
or alternatives, or to perform robust
sensitivity analyses, hand calculations
and spreadsheets become impractical.
Grouptime
During the synthesis, a function of the
software called grouptime was used.
This was a platform for discussing the
weighting of the criteria and ranking
of the alternatives. Until this stage the
four decision-makers had made the
judgments independently based on their
own experience and knowledge of the
project. Using grouptime, these decision-
makers were brought together and could
argue the case for their judgments and
see the sensitivity of the rankings to
their weighted criteria. For example, one
criterion could be ignored altogether
to determine the extent of impact,
if any, on the ranking of the results.
Changing decision-maker’s judgments
or ignoring them entirely is also an
option. The possibilities for analysis and
interrogation were numerous and their
value immeasurable.

14. 2726
California High-Speed Rail Grade Separation Schematic California High-Speed Rail Grade Separation Schematic
Case Study Conclusions
In this case study, utilizing AHP as a
decision-making methodology allowed
the project team to arrive at a preferred
alternative in an efficient manner with
input from each stakeholder. AHP is not
only tried and tested, but structured,
transparent, and justifiable to the client
and wider stakeholders. For the purposes
of this paper, a single grade separation on
Van Nuys Boulevard was investigated.
Historically, across the California
High-Speed Train Project, as many as 50
grade separations have been considered,
each with eight possible alternatives.
For a large number of these alternatives
the weighting of the criteria, and thus
the first half of the AHP synthesis,
may be identical. It would be unusual
to have different criteria and different
weighting of those criteria for each of
the 50 schemes. Pairwise comparisons
of the criteria rarely vary greatly across
similar locations. For example, within
a city we may consistently judge utility
diversions as less important than right-
of-way land take, or traffic circulation
as more important than constructability.
Assuming our criteria weighting for
Van Nuys is applicable at other locations
within the San Fernando Valley, this
case study provides a foundation to
implement an AHP analysis to multiple
scenarios without replicating the entire
process. These economies of scale would
further add value and efficiency to the
decision-making process of large-scale
infrastructure projects. Opportunities to
develop generalized criteria weighting
within a company or for a particular
client open up exciting possibilities to
harness the power of AHP analysis.
With a complex infrastructure project
such as the California High-Speed Train
Project, where many stakeholders hold
differing, often strong, opinions, AHP
is ideally suited not only to structure
the complexity of the problem but
also to build consensus among the
stakeholders in regards to an alternative.
Lack of consensus and a lack of belief
in a solution often stall or inhibit
funding and forward movement of
large schemes. Stakeholders often find
themselves entrenched in unswaying
beliefs and unwillingness to buy into and
support a compromise for the greater
good of a project. AHP could provide
all stakeholders space to share their
preferences and critical transparency
to interrogate their relative importance.
Using a robust and measurable decision-
making process can help ensure that
the project team chooses the correct
alternative that everyone can believe in
and support moving forward.
Findings
AHP methodology brings a powerful,
transparent, and structured approach to
decisions on infrastructure projects. Too
often significant decisions in traditional
option analysis are determined by
subjective, biased, or even random
factors — the opinion of “the loudest
voice in the room” or simply the last
issue identified, as it is fresh on the
minds of the decision-makers. Older
issues tend to be underestimated as we
become desensitized to them through
their numerous tablings at progress
meetings. Often, the extrovert’s issues
are always critical, while the introvert’s
are forgotten or underestimated. The
decision process is plagued by a lack of
structure, a bias toward the loudest voice,
and a general fatigue in considering
historical issues. AHP provides a
documentable and replicable way of
structuring problems involving multiple
criteria. The vast body of academic study
and accompanying literature, as well as
commercially available software, provide
confidence in AHP as a decision-making
tool. AHP provides a platform for group
decision-making, allowing a spectrum of
stakeholders to have their voices heard
and critically build consensus through
early involvement and collaboration.
This research has established that the
AHP process is a relevant and useful
technique in the decision-making
process for infrastructure projects. The
study has shown that even when a result
has been determined, a professional and
pragmatic view is needed to draw an
actionable conclusion.

15. 2928
Sources
1. Forman, Ernest H. and Saul I. Gass. “The Analytic Hierarchy Process – An
Exposition.” Operations Research, 49.4 (2001): 469-486. Print.
2. Piantanakulchai, Mongkut and Nattapon Saengkhao. “Evaluation of Alternatives
in Transportation Planning Using Multi-Stakeholders Multi-Objectives AHP
Modeling.” Proceedings of the Eastern Asia Society of Transportation Studies,
Vol. 4 (2003): 1613-1628. Print.
3. Saaty, Thomas L. and Luis G. Vargas. Models, Methods, Concepts & Applications
of the Analytic Hierarchy Process. Boston: Kluwer Academics, 2001. Print.
4. Sadasivuni, Raviraj, Charles G. O’Hara, Rodrigo Nobrega, and Jeramiah Dumas.
“A Transportation Corridor Case Study for Multi-Criteria Decision Analysis.”
ASPRS 2009 Annual Conference. 9 Mar. 2009, Baltimore. Starkville: Mississippi
State University, n.d. Print.
5. Shin, Yong B., Seungho Lee, Sun Gi Chun, and Dalsang Chung. “A Critical
Review of Popular Multi-Criteria Decision Making Methodologies.” Issues in
Information Systems, 14.1 (2013): 358-365. Print.
6. Triantaphyllou, Evangelos and Stuart H. Mann. “Using the Analytic Hierarchy
Process for Decision Making in Engineering Applications: Some Challenges.”
International Journal of Industrial Engineering: Applications and Practice, 2.1
(1995): 35-44. Print.
©Arup

16. Gerald Desmond Bridge, Long Beach, CA ©Arup
3130
About Arup
Arup is the creative force at the heart of many of the world’s most
prominent projects in the built environment and across industry.
We offer a broad range of professional services that combine to make a
real difference to our clients and the communities in which we work.
We are truly global. From 92 offices in 40 countries, our 12,000
planners, designers, engineers, consultants and technical specialists
deliver innovative projects across the world with creativity and passion.
Founded in 1946 with an enduring set of values, our unique
trust ownership fosters a distinctive culture and an intellectual
independence that encourages collaborative working. This is reflected
in everything we do, allowing us to develop meaningful ideas,
help shape agendas and deliver results that frequently surpass the
expectations of our clients.
We shape a better world.
Contributors
Josh Bird
Engineer
Josue Enriquez
Engineer
Mackenzie Van Thof
Graduate Engineer
Daleen Saah
GIS Consultant
Jesse Vernon
Technical Editor
Precila Disman
Josephine Donde
Business Development
Doreen Patron
Graphic Design
Acknowledgements
Tony Marshall
Tim Corcoran
Richard Prust