Habitat International xxx (2016) 1e15

Contents lists available at ScienceDirect

Habitat International
journal homepage: www.elsevier.com/locate/habitatint

Alternative transit-oriented development evaluation in sustainable

built environment planning
Wann-Ming Wey a, *, Heng Zhang b, Yu-Jie Chang a
a
Department of Real Estate and Built Environment, National Taipei University, 151 University Road, San Shia District, New Taipei City, 23741, Taiwan, ROC
b
Department of Architecture, National Cheng Kung University, No.1, University Road, Tainan City 701, Taiwan, ROC

a r t i c l e i n f o a b s t r a c t

Article history: Following the increase in urban population density and house prices in the 1970s, urban residents began
Received 21 January 2016 searching for satisfactory living environments. With the construction of highways and advancements in
Received in revised form motor vehicle technology, urban development began extending to suburban regions, a phenomenon
23 February 2016
which is referred to as urban sprawl. The negative external effects caused by urban sprawl, including
Accepted 8 March 2016
Available online xxx
inefficient land use and pollution, highlight several problems such as the disjunction between urban land
use and transportation. To achieve sustainable transportation and resolve the aforementioned problems,
we established the transit-oriented development (TOD) evaluation criteria and formulated the TOD site
Keywords:
Sustainable transportation
selection standard, which fulfills the principle of sustainable transportation. Thus, a solution to problems
Transit-oriented development (TOD) related to urban sprawl is proposed, which may serve as a reference for planning units in devising
Fuzzy DELPHI method relevant processes.
Fuzzy analytic network process (FANP) The fuzzy Delphi method was adopted to select TOD evaluation criteria satisfying the principle of
Geographical information system (GIS) sustainable transportation, and the fuzzy analytic network process (FANP) was applied to determine the
Ankeng line of the New Taipei City mass weights of relevant planning criteria. The results show that all nine TOD evaluation criteria for sus-
rapid transit (MRT) system tainable transportation were supported by an expert consensus in the first phase of the study, indicating
that all of the selected criteria were critical in the TOD evaluation. In the second phase, according to the
weighted evaluation criteria, the environmental carrying capacity in high-density development was the
most crucial evaluation criterion, whereas the least critical criterion was equality in residence accessi-
bility. Finally, through the use of a geographical information system, the performance of each station on
the Ankeng Line of the New Taipei City mass rapid transit (MRT) system was evaluated according to the
TOD evaluation criteria. Combining the evaluation with the weighted sums of the FANP revealed that
Xinhe Elementary School MRT station was the optimal station for TOD. Accordingly, built environment
planning should be undertaken to develop the urban area surrounding this station. The findings of this
study may provide a reference for planning units in undertaking future TOD projects under the principle
of sustainable transportation.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction 1989). To solve these external problems, scholars have suggested

Following the increase in urban population and immigration,
overcrowding and urban sprawl began occurring because of urban
developments. The negative external effects of urban sprawl
include air pollution from motor vehicles, longer commuting times
because of traffic congestion, overuse of land resources caused by
low-density and leapfrog development, and driver safety (Peiser,
* Corresponding author.
E-mail addresses: wmwey@mail.ntpu.edu.tw (W.-M. Wey), changlin@mail.ncku.
edu.tw (H. Zhang).
applying the concept of sustainable development to resolve the
impact of urban development. In Our Common Future, the World
Commission on Environment and Development of the United
Nations (1987) defined sustainable development as “development
that meets the needs of the present without compromising the
ability of future generations to meet their own needs” (p. 10). This
effectively defines the goal of sustainable development as achieving
equality among generations in using resources. Practicing sustain-
able development requires simultaneously satisfying the economic,
social, and environmental dimensions of development. On applying
the concept of sustainable development in connecting urban
transportation and land use, scholars have advocated adopting the

http://dx.doi.org/10.1016/j.habitatint.2016.03.003
0197-3975/© 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Wey, W.-M., et al., Alternative transit-oriented development evaluation in sustainable built environment
planning, Habitat International (2016), http://dx.doi.org/10.1016/j.habitatint.2016.03.003
2 W.-M. Wey et al. / Habitat International xxx (2016) 1e15
principle of sustainable transportation, which requires promoting
transportation behaviors and involves the following subgoals: (a)
Economic efficiency: Transportation resources must be used effi-
ciently to maximize their effects and facilitate economic develop-
ment in other areas. (b) Environmental sustainability: Decisions
regarding transportation behaviors must consider the external ef-
fects of transportation such as air pollution, noise, and traffic ac-
cidents. Additionally, transportation infrastructure developments
must not exceed the environmental carrying capacity in order to
prevent irreparable damage to the natural environment. (c) Social
equity: Promoting transportation behaviors must involve fair
consideration of the benefits to all people and cater to the needs of
vulnerable groups. Transit-oriented development (TOD) empha-
sizes integrating urban land use with public transportation. In other
words, urban residents must be able to accomplish their daily living
activities entirely through using public transportation and without
relying on privately owned motor vehicles (Wey & Chiu, 2013).
Thus, the use of motor vehicles would be reduced, effectively pre-
venting urban sprawl and lowering air pollution. In addition, the
traffic volume of public transportation systems would increase,
thereby promoting the economic efficiency of such systems.
However, the concept of diversity, such as diversified land use and
levels of habitability, must be stressed for such developments to be
successful (Blowers, 1993; Calthorpe, 1993). Accordingly, the
objective of this study was to fulfill sustainable development
through coordinating transportation and land use behaviors.
Because the developmental concepts of TOD are related to the goals
of sustainable transportation, TOD was considered a suitable
planning approach for promoting sustainable transportation.
This study was conducted to establish TOD evaluation criteria
based on the principles of sustainable transportation to enable
effectively selecting mass rapid transit (MRT) stations in Taipei for
future TOD projects. To combine theories with practice, the Ankeng
line of the New Taipei City MRT system was selected as the research
target. The most suitable station on Ankeng Line for TOD was
determined according to the principles of sustainable trans-
portation, and built environment plans were formulated for
developing the urban area around that station.
To establish the research model, the fuzzy Delphi method was
first applied to select the criteria through administering question-
naires to a panel of experts. Subsequently, the criteria deemed
suitable for establishing the evaluation model were determined.
Second, the fuzzy analytic network process (FANF) was adopted to
assign weights to the criteria derived from the expert question-
naires. Finally, a geographic information system (GIS) was used as
an analytical tool to designate an appropriate index for each cri-
terion. An index map of the studied region was constructed to form
the basis for scoring each station according to the evaluation
criteria. The scores of each criterion were then multiplied by the
weighted values to identify the optimal station according to the
highest score on the TOD evaluation criteria. Finally, a built envi-
ronment plan was formulated for developing the urban area near
the selected station.
2. Literature review
Transportation planning and strategies affect not only the
landscape style and land use of a city, but also the relationship
between urban sprawl and population density (Peiser, 1989). In
summary, urban sprawl in Taiwan is due to inappropriate urban
development planning and management. Specifically, an excess of
urban development projects has caused rapid urban expansion, as
well as the inadequate and inefficient use of public facilities. In the
urban housing market, vacancy rates and house prices in urban
centers have been exorbitant. The proliferation of private motor
Please cite this article in press as: Wey, W.-M., et al., Alternative transit-oriented development evaluation in sustainable built environment
planning, Habitat International (2016), http://dx.doi.org/10.1016/j.habitatint.2016.03.003
vehicles also contributed to exacerbating the urban sprawl.
Collectively, the aforementioned problems originated from
inappropriate transportation and land use. To solve these problems,
scholars have suggested that philosophy of urban development
must shift toward one of sustainable development, and urban and
transportation infrastructure planning must follow the compre-
hensive consideration of environmental sustainability, social eq-
uity, and economic efficiency. Unrestrained use of natural resources
for economic development has created a shortage of resources and
damaged the natural environment, thus prompting the emergence
of sustainable development concepts. Sustainable development is
aimed at preventing shortages of usable natural resources for
future generations while pursuing economic development. There-
fore, when using resources, efficiency must be emphasized,
wastage must be avoided, and renewable resources must be uti-
lized extensively (World Commission on Environment and
Development, 1987; Blowers, 1993; Meijer, Adriaens, Linden, &
Schik, 2011).
Sustainable transportation can be regarded as an operations
approach to the sustainable development of transportation sys-
tems. Sustainable transportation requires balancing the environ-
mental, social, and economic dimensions in transportation
behaviors. On the environmental dimension, transportation au-
thorities must consider the externality of transportation behaviors
in its decision-making, such as air pollution, noise, and traffic ac-
cidents. Moreover, the development of transportation infrastruc-
ture must not exceed the carrying capacity of the natural
environment. On the social dimension, promoting transportation
behaviors must involve fair consideration of the benefits to all
people and cater to the needs of vulnerable groups such as elderly
adults, people with disabilities, and disadvantaged groups. Finally,
on the economic dimension, transportation resources must be used
and maintained efficiently by minimizing resource use in order to
maximize efficiency (Zhao, 2010).
Sustainable transportation requires the development of sus-
tainable transportation behaviors in all three of these dimensions.
Most of the aforementioned green transportation strategies have
focused on environmental sustainability. Therefore, to promote
sustainable transportation, social equity and economic efficiency
must also be considered to balance the three dimensions of sus-
tainable transportation behaviors.
The concept of TOD is derived from urban planning theories that
emerged in response to urban sprawl including smart growth, new
urbanism, and sustainable development. Smart growth involves
controlling urban growth through flexible management ap-
proaches. The principles of smart growth include high-density
development, mixed land use, redeveloping the old regions in cit-
ies, constructing accessible pedestrian spaces, and diversifying the
selection of transportation vehicles. Grant and Tsenkova (2012)
stressed that using encouragement management approaches (e.g.,
floor area incentives and flexible partitions), in contrast to growth
management approaches, to achieve the goals of smart growth
provides more flexibility and incentives to decelerate urban sprawl.
New urbanism emphasizes diversity in designing cities, including
diversifying traffic, which involves the coexistence of pedestrians,
public transportation, and cars. Additionally, this concept accen-
tuates localizing communities and focusing on specialization in
order to project a city image of closeness and diversity, thereby
reducing the use of private transportation vehicles and orienting
future cities toward sustainable development.
Calthorpe (1993) introduced the concept of TOD, defining it as
“moderate and high-density housing, along with complementary
public uses, jobs, retail and services, concentrated in mixed-use
developments at strategic points along the regional transit sys-
tems.” In particular, Calthorpe addressed the importance of
 W.-M. Wey et al. / Habitat International xxx (2016) 1e15
creating pedestrian-friendly spaces. Concentrating various facilities
for activities of daily living in a TOD region can improve pedestrian
accessibility. Furthermore, buildings should satisfy public needs
and constructed on a human scale, and previous development
strategies that are solely car-oriented must be avoided.
La Greca, Barbarossa, Ignaccolo, Inturri, and Martinico (2011)
verified that concentrating urban development near trans-
portation facilities improves pedestrian accessibility to such sta-
tions within walking distance, thereby further promoting the
financial sustainability of public transportation systems. On envi-
ronmental sustainability, Tiwari, Cervero, and Schipper (2011)
emphasized that reduced carbon emissions is a benefit of TOD;
through combining the three dimensions (3D; i.e., density, di-
versity, and design) of TOD with the ASIF2 paradigm (avoid, shift,
improve and finance) (Dalkmann & Brannigan, 2007) which de-
termines emissions as a function of total transport activity, modal
shares according to mode of transport, fuel intensity according to
mode of transport and type of fuel, and the carbon content of fuels
according to mode of transport. Furthermore, they formulated
numerous low-carbon urban planning and transportation strate-
gies, which facilitated a reduction of carbon emissions by up to
52.5%.
TOD designs can be examined according to the scale definition
of TOD proposed by Calthorpe (1993), in which a TOD project is
characterized as a mixed-use community where the area scale of
the community is within approximately 2000 pedestrian steps to a
central commercial station (walking time: approximately 10 min,
walking distance: approximately 600 m). The TOD region and
density of the city of Austin in the United States was designed to
have an accessible walking distance of 800 m to any public trans-
portation station. The radial design is divided from the center
outward into a gateway zone, midway zone, and transition zone.
The highest density of use is in the midway zone, and this density
decreases gradually at further distances from the center. Regarding
usage patterns, the midway zone features the highest number of
commercial areas, and residential areas are more prominent in the
regions outside of this zone. The concept of design diversity em-
phasizes diversity in the choice of residence (i.e., variety in the
types of housing that are available). Therefore, exorbitant house
prices must be avoided to provide equal opportunity for people to
live in the region.
Cervero and Kockelman (1997) proposed the design criteria for
the 3D principle of TOD. The density criteria are population density,
employment density, and accessibility to jobs. The diversity criteria
include the proportion of dissimilar land uses among hectare grid
cells within a tract, vertical mixture, per developed acre intensities
of land uses, and proximities to commercial-retail uses. Finally, the
urban design criteria include street designs, pedestrian and cycling
provisions, and site designs. Cervero, Sarmiento, Jacoby, Gomez,
and Neiman (2009) subsequently added two more criteria (i.e.,
destination accessibility and distance to transit), thereby forming
the 5D principle for TOD projects.
When Calthorpe (1993) proposed the concept of TOD, the
concept was divided into two typesdurban TOD and neighborhood
TODdaccording to the types of public transportation systems. The
types of public transportation systems for urban TOD are light and
heavy rail systems as well as express buses (similar to bus rapid
transit systems). Urban TOD areas feature high-density businesses
and work opportunities as well as moderateehigh density of resi-
dential areas. The primary types of public transportation systems in
neighborhood TOD areas are local transportation systems and bus
extensions. The travel time to a primary transportation system
should not exceed 10 min. Neighborhood TOD areas have a mod-
erate residential density, and the primary commercial behavior is
retail businesses. The principal goal of developing this type of
Please cite this article in press as: Wey, W.-M., et al., Alternative transit-oriented development evaluation in sustainable built environment
planning, Habitat International (2016), http://dx.doi.org/10.1016/j.habitatint.2016.03.003
3
region is to create a positive living environment with abundant
public open spaces and recreational zones. Regarding the land-use
distribution ratios, urban (neighborhood) TOD areas feature a
public space usage ratio of 5%e15% (10%e15%), a central commer-
cial/office zone ratio of 30%e70% (10%e40%), and a residential ratio
of 20%e60% (50%e80%).
Sustainable transportation involves encouraging transportation
behaviors that promote the goals of sustainable development,
which can be divided into the following three subgoals: (a) Eco-
nomic efficiency: Transportation resources must be used efficiently
to maximize their effects while facilitating economic development
in nearby regions. (b) Environmental sustainability: Decisions
regarding transportation behaviors must consider the external ef-
fects of transportation such as air pollution, noise, and traffic ac-
cidents. Additionally, transportation infrastructure developments
must not exceed the environmental carrying capacity in order to
prevent irreversible damage to the natural environment. (c) Social
equity: Promoting transportation behaviors must involve fair
consideration of the benefits to all people and account for the needs
of vulnerable groups. To enable integrating transportation with
land use and thereby realize sustainable transportation, we adop-
ted TOD as the primary planning model in order to shift current
urban transportation system toward sustainable transportation
systems. However, because of the restriction of government bud-
gets and natural environmental conditions near stations, not all
stations are suitable for TOD of the surrounding area. Therefore, the
station closest to the target of sustainable transportation must be
selected for TOD planning following a thorough evaluation. Thus,
establishing a set of TOD evaluation criteria is crucial.
3. Research design
We reviewed the literature to understand the implications of
sustainable transportation and TODda land use planning approach
to achieving a sustainable transportation system. Sustainable
transportation was established as the principal development goal,
with economic efficiency, environmental sustainability, and social
equity in connecting transportation and land use defined as sub-
goals. Accordingly, we then established the criteria corresponding
to the subgoals in the design concept of TOD. Subsequently, we
verified the necessity of including the criteria in the evaluation in
order to ensure an objective assessment of the planning resulting
from our subjective judgment. Therefore, given this necessity, the
fuzzy Delphi method was adopted for selecting the criteria ac-
cording to the level of emphasis that a panel of experts placed on
each criterion. After the evaluation criteria were selected, the
weights of the criteria were calculated on the basis of the impor-
tance assigned to each criterion by using the fuzzy analytic network
process (FANP). The FANP enabled comparing each pair of criteria
on their importance, thus ensuring that the opinions of the scholars
were organized scientifically and objectively as well as simplifying
the calculation of the weights. Furthermore, the dependency
among the criteria was considered in the FANP, thereby avoiding
the weight distortion caused by interdependence among the
criteria. However, not all of the experts understood each criterion
clearly. Therefore, a geographical information system (GIS) was
adopted to evaluate the performance of each of the stations. When
a GIS is used to evaluate a project, the index corresponding to each
evaluation criterion must be verified; therefore, a means for visu-
alizing the index data must be established. The value of each index
was stratified to establish a basis for scoring the stations. The
weights obtained through the FANP were summed and multiplied
by the stratified index scores. Finally, the scores in each criterion
were multiplied by the weights and then summed through a matrix
operation. The project with the highest score represented the most
4 W.-M. Wey et al. / Habitat International xxx (2016) 1e15

suitable station for TOD.

In the following, we first outline the extent analysis method on
FANP and then the method is applied to an alternative transit-
oriented development evaluation in sustainable built environ-
ment planning problem. Let X ¼ (x1, x2, …, xn) be an object set, and
U ¼ (u1, u2, …, um) be a goal set. According to the method of Chang's
(1992) extent analysis, each object is taken and extent analysis for
each goal, gi; is performed, respectively. Therefore, m extent anal-
ysis values for each object can be obtained, with the following signs
(Kahraman, Cebeci, & Ruan, 2004):

1 2 m
Mgi ; Mgi ; …; Mgi ; … i ¼ 1; 2; …; n; Fig. 1. The interaction between M1 and M2.

Step 3: The degree possibility for a convex fuzzy number to be greater than k
j
where all the Mgi (j ¼ 1, 2, …, m) are TFNs. convex fuzzy numbers Mi ði ¼ 1; 2; …; kÞ can be defined by
Steps of the Chang's extent analysis can be given as in the
following: VðM  M1 ; M2 ; …; Mk Þ
¼ V½ðM  M1 Þ and ðM  M2 Þ and / and ðM  Mk Þ
Step 1: The value of fuzzy synthetic extent with respect to the ith
¼ minVðM  MiÞ; i ¼ 1; 2; 3; …; k: (7)
object is defined as
2 31
X
m X
n X
m
Mgi 54 Mgi 5
j j
Si ¼ (1)
j¼1 j¼1 j¼1
P j Assume that
To obtain m j¼1 Mgi , perform the fuzzy addition operation of m
extent analysis values for a particular matrix such that
d0 ðAi Þ ¼ min VðSi  Sk Þ (8)
0 1
X
m X
m X
m X
m
for k ¼ 1; 2; …; n; ksi . Then the weight vector is given by
Mgj i ¼ @ lj ; mj ; uj A (2)
j¼i j¼1 j¼1 j¼1 T
W 0 ¼ ðd0 ðA1 Þ; d0 ðA2 Þ; …; d0 ðAn ÞÞ (9)
Pn Pm j 1
and to obtain ½ i¼1 j¼1 Mgi  ,
perform the fuzzy addition oper-
j where Ai ði ¼ 1; 2; …; nÞ are n elements.
ation ofMgi ðj ¼ 1; 2; :::; mÞ values such that
! Step 4: Via normalization, the normalized weight vectors are
X
n X
m
j
X
n X
n X
n
Mgi ¼ li ; mi ; ui (3)
i¼1 j¼i i¼1 i¼1 i¼1 W ¼ ðdðA1 Þ; dðA2 Þ; …; dðAn ÞÞT (10)

and then compute the inverse of the vector in Equation (3) such where W is a non-fuzzy number.
that There are considerable advantages in incorporating de-
mographic and environmental modelling directly into the TOD
2 31
X
n X
m   planning processes to enable the assessment of the environmental
4 1 1 1
Mgi 5
j effects of different transport proposals and scenarios simulta-
¼ Pn ; Pn ; Pn (4)
j¼1 j¼1 i¼1 ui i¼1 mi i¼1 li neously with the assessment of their traffic effects. This modelling
system in our proposed sustainable built environment planning
was developed using the Symbology function of Arc GIS 10.0 and is
intended for use by urban planners as an add-on module to existing
Step 2: The degree of possibility of transport planning models. The development of the system relies
M2 ¼ ðl2 ; m2 ; u2 Þ  M1 ¼ ðl1 ; m1 ; u1 Þ is defined as on the capabilities of GIS and uses, as its prime input, the output
   data from prediction models or other necessary data, overlaid with
VðM2  M1 Þ ¼ sup min mM1 ðxÞ; mM2 ðyÞ (5) land use information in the immediate vicinity of the modelled
yx road traffic network. The system is designed to provide rapid in-
formation to the planner on the environmental effects of any TOD
and can be equivalently expressed as follows: proposal being considered and thereby aids in the selection of a
preferred TOD scenario based on environmental, not just any one
VðM2  M1 Þ ¼ hgtðM1 ∩M2 Þ ¼ mM2 ðdÞ
8 consideration related outcomes.
>
> 1; if m2  m1 Models tend not to be used if they interfere too much with the
>
>
< way planners work. There are three main methods available for
0; if l1  u 2
¼ integrating GIS and models: embedding GIS functionalities
>
> l1  u 2
>
> entirely within the model (the model dominant approach);
: ; otherwise
ðm2  u2 Þ  ðm1  l1 Þ implementing the models within the GIS (the GIS dominant
(6) approach); and the interface approach. Details of these ap-
proaches are outlines in Affum and Taylor (1998), Lewis (1990)
where d is the ordinate of the highest intersection point D between and Trinidad and Marquez (1994). Of these methods, the inter-
mM1 and mM2 .(see Fig. 1). To compare M1 and M2, we need both the face approach to the integration of GIS and models, tends to
values of VðM1  M2 Þ and VðM2  M1 Þ. provide the least interference and has been adopted for the

Please cite this article in press as: Wey, W.-M., et al., Alternative transit-oriented development evaluation in sustainable built environment
planning, Habitat International (2016), http://dx.doi.org/10.1016/j.habitatint.2016.03.003
 W.-M. Wey et al. / Habitat International xxx (2016) 1e15
planning model we proposed. Under this approach, the compo-
nent environmental prediction sub-models were developed as
routines that behave like modules resident within the GIS and
operate via a user interface. The entire process runs transparently
to the user via the user interface.
The procedure for establishing the research model was divided
into the following five main steps, which were further divided into
substeps:
 Step 1: Verify the subgoals of each goal and the corresponding
criteria.
 Step 1-1: Verify the subgoals or targets within each research
goal
 Step 1-2: Establish the evaluation criterion suitable for each
subgoal or target
 Step 2: Use the FDM to select the criteria appropriate for this
study.
 Step 2-1: Form a group of experts and administer question-
naires for them to evaluate the level of importance of each
criterion by using a 3-point estimate (optimistic values,
optimal values, and conservative values) and identify the
influencing relationships among the criteria.
 Step 2-2: Calculate the mean optimistic and conservative
values of each criterion. Eliminate the optimistic and conser-
vative values that are more than two standard deviations from
the mean. Recalculate the mean values with the remaining
scores.
 Step 2-3: Establish a triangular fuzzy number based on the
optimistic and conservative values for each criterion.
 Step 2-4: Use the grey zone verification method to confirm
whether the values assigned by the experts indicate that
consensus has been reached. If not, then use the arithmetic
means of the optimistic and conservative values as the final
consensus value. If consensus is achieved, then obtain the
fuzzy sets from the fuzzy intersection of the two triangular
fuzzy numbers, and use the stratified scores with the largest
membership values in the fuzzy sets as the consensus values.
If a grey zone (i.e., complete overlap between the values) ex-
ists, then the questionnaire survey on the criterion must be
repeated.
 Step 2-5: Rank the expert consensus values for each criterion.
If there is a large gap between the values of any two criteria,
then the criterion above the gap can be considered more
critical to a given subgoal and designated as the selection
threshold. If the gap between the ranked consensus values is
not large, then the criteria of the minimal consensus values
should be assessed to determine whether they meet a certain
standard. If the value is relatively high and the experts
consider the criterion vital, then the minimal consensus
values can be included as the selection threshold.
 Step 3: Calculate the weight of each criterion using the FANP.
 Step 3-1: Use the expert questionnaires to determine the
comparison values for each pair of evaluation criteria ac-
cording to their importance. This step involves applying the
aforementioned fuzzy analytic hierarchy process (FAHP) to
calculate the weight of each criterion without considering the
interdependency among the criteria according to the com-
parison values.
 Step 3-2: Conduct a consistency test on the comparison
values. Logical inconsistency in comparing the criterion
importance values assigned by the experts (e.g., A > B, B > C,
and C > A) must be avoided. Therefore, Cronbach's alpha co-
efficient must be used to further verify the FAHP expert
questionnaires. Proceed to the next step only when the veri-
fication is satisfactory.
Please cite this article in press as: Wey, W.-M., et al., Alternative transit-oriented development evaluation in sustainable built environment
planning, Habitat International (2016), http://dx.doi.org/10.1016/j.habitatint.2016.03.003
5
 Step 3-3: Examine the interdependency among the criteria. To
understand the mutual relationships among the criteria, for
each dependent criterion, compare the level of importance
between each pair of interdependent criteria to obtain the
actual mutual relationship between the criteria.
 Step 3-4: Verify the consistency in the questionnaire tables.
Each paired comparison table must pass the consistency test
before its weight can be calculated.
 Step 3-5: Multiply the two matrices obtained from the pre-
vious two steps with each other to acquire the FANP weight-
ings, including the level of importance between each pair of
criteria and the products of the interdependency between
each pair of criteria.
 Step 4: Evaluate the objective performance of each station ac-
cording to the nine criteria by using the GIS.
 Step 4-1: Establish the objective evaluation indices applicable
for the GIS under each evaluation criterion.
 Step 4-2: Confirm the region of study.
 Step 4-3: Survey and collect index data.
 Step 4-4: Convert the performance of each station to a score
according to the rated standard.
 Step 5: Combine the FANP weightings with the performance
P
scores for each station, and use the matrix algorithm Wi Xi to
determine the optimal station for TOD.
4. Empirical study and analysis
After the research model was established, the effectiveness of
the model was empirically verified. The main considerations of this
study are the criteria derived from the FDT expert questionnaires,
the criteria weights derived from the FANP questionnaires, the in-
dex scores for the criteria determined using the GIS, and the total
scores and selected stations. The Ankeng line of the New Taipei City
MRT system was the target of this study. The remainder of this
chapter describes the empirical analysis of the study region, the
selection of the TOD evaluation criteria, the establishment of the
criterion weightings, and the evaluation and selection of the sta-
tions through using the GIS.
4.1. Study region
The Ankeng line is located in the Ankeng region of Xindian
District, New Taipei City, approximating the intersection between
Xindian and Zhonghe (Fig. 2). The region is positioned in the
western outskirts of Xindian District. Geographically, the region is
mostly composed of mountainsides and hills, the slopes of which
are less steep in regions approaching the Xindian River. The
developed regions are mainly concentrated on both sides of Ankang
Road, and numerous large residential communities are located on
the mountainside. Because the traffic in this region relies primarily
on a 15-m-wide four-lane road (i.e., Ankang Road), severe traffic
congestion is common during peak traffic. To address this problem,
the Department of Transportation constructed Anyi Road.
(http://www.xindian.ris.ntpc.gov.tw/web/AB?
command¼showDetail&postId¼276334) to relieve the traffic
inflow at Xindian, and extended the New Taipei Riverside
Expressway to relieve traffic inflow at Zhonghe and Banqiao. In
addition, a 12-m reserved route is retained on Anyi Road for
building the MRT Ankeng Line.
The Ankeng line of the MRT is a light-rail transit system with a
lowemoderate traffic volume. From the perspective of the TOD
scale, the optimal range of development for the Ankeng line is
approximately 400e600 m from each MRT station on that line.
Therefore, the median value of 500 m from each station was
6 W.-M. Wey et al. / Habitat International xxx (2016) 1e15
Fig. 2. Location of the Ankeng region.
selected as the evaluation region for built environment develop-
ment. The administrative zones in the region include 12 villages in
Xindian and six villages in Zhonghe (Fig. 3). This study targeted
these 18 villages to collect spatial and attribute data for each index.
Regarding the maps we will focus mainly on the case of the
Ankeng Line of the New Taipei City mass rapid transit (MRT) sys-
tem, noting that the general results have been fairly similar in
Fig. 3.1
4.2. Selecting the TOD evaluation criteria
4.2.1. TOD evaluation criteria
Using sustainable transportation as the core value and aiming
for the goals of economic efficiency, environmental sustainability,
and social equity in transportation behaviors, we organized the
TOD evaluation criteria according to the aforementioned goals.
After examining the three major goals of urban sustainability and
integrating the planning concepts of TOD as an evaluation standard,
we reviewed the literature and generalized the resulting evaluation
factors as the criteria base for the empirical study stated as Table 1.
4.2.2. Calculating the criterion weightings
The selected FANP questionnaire respondents were experts in
transportation and land use and considered a representative in
their fields. Because further investigation was required in the FANP
questionnaire survey, additional time was needed to complete this
phase. Considering the time cost and the quality and validity con-
cerns in administering the questionnaire, only one expert from
industry, government, and academia was recruited as the ques-
tionnaire respondents. The industry expert was an urban planning
technician from a consultancy responsible for the built environ-
ment planning of the areas near the MRT lines in Taipei City; the
government expert was an officer from the Urban and Rural
1
For visualisation of the changes on a graphical scale we refer the reader to kml
file in the journal homepage which could be explored with Google-Earth software
[http://dx.doi.org/10.1016/j.habitatint.2016.03.003].
Please cite this article in press as: Wey, W.-M., et al., Alternative transit-oriented development evaluation in sustainable built environment
planning, Habitat International (2016), http://dx.doi.org/10.1016/j.habitatint.2016.03.003
Development Department in New Taipei City; and the academic
expert was a professor of urban planning and designs.
The remainder of this section presents the FAHP results and
consistency analysis, as well as the FANP results. The criterion
weightings assigned by the three experts are then averaged and
analyzed.
4.3. FANP calculation process and result by experts
Fuzzy ANP is a general theory of relative measurement used to
derive composite-priority-ratio scales from individual-ratio scales
that represent relative influence of factors that interact with
respect to control criteria (Niemiraa & Saaty, 2004). Through the
“supermatrix”, which is composed of matrices of column priorities,
the FANP framework catches the consequence of dependence/
feedback within and among components.
To generate global priorities in a system with interdependent
influences, the obtained local priority vectors and matrices are
entered in a matrix to form a “supermatrix” as follows:
2 Goal Criteria Alternatives 3
Goal 0 0 0
W¼
Criteria 4 W21 W22 0 5
Alternatives 0 W32 I
where “I” is the identity matrix, and entries of zeros correspond to
those elements that have no influence. After forming the super-
matrix, a weighted supermatrix is derived by transforming all
columns sum to unity, i.e. like the concept of Markov chain for
ensuring column stochastic (Huang, Tzeng, & Ong, 2005). The detail
of mathematical processes of the ANP approach can refer to Saaty
(1996). And then, if the supermatrix formed covers the whole
network, the priority weights of alternatives can be found in the
column of alternatives-to-goal in the limit supermatrix (Sarkis,
2003).
The process for solving interdependent TOD evaluation criteria
prioritization is summarized as follows: In order to consider
interdependence, the first step is to identify the multiple criteria
 W.-M. Wey et al. / Habitat International xxx (2016) 1e15 7

Fig. 3. Development region of Ankeng MRT Line (K stands for station No.).

Table 1
TOD evaluation criteria based on the principle of sustainable transportation.

Dimension Criterion

A. Economic Efficiency A-1 Population density in the development region

A-2 Spatial density of commercial and retail facilities
A-3 Design of pedestrian spaces with up-scalable traffic volume
B. Environmental Sustainability B-1 Environmental carrying capacity for high-density development
B-2 Externality resulting from mixed land use
B-3 Density of open spaces for improving quality of life
C. Social Equity C-1 Inhibition capacity of floor area incentives against house prices
C-2 Equality in residence accessibility for all people
C-3 Daily living safety of the development region

that merit consideration and then draw a relationship which shows
the degree of interdependence among the criteria. Next is to
determine the degree of impact or influence between the criteria.
When comparing the criteria for each criterion, the decision maker
will respond to questions such as: “In comparing criteria A and B,
on the basis of A. Economic Efficiency, which criterion is
preferred?” The responses are presented numerically, scaled on the
basis of Saaty's 1e9 scale (Saaty, 1996) with reciprocals, in a project
comparison matrix. The final step is to determine the overall pri-
oritization of these TOD evaluation criteria.
The AHP technique introduced by Saaty is applied to a problem
without considering the interdependence properties among
criteria. However, there exist interdependence relationships among
the identified criteria in TOD evaluation criteria prioritization
problem. Generally, if we promote a A-1 Population density in the
development region, the performance of A-2 Spatial density of
commercial and retail facilities will be increased in hereafter op-
erations. Similarly, in order to increase the B-1 Environmental
carrying capacity for high-density development, we will have to
increase B-2 Externality resulting from mixed land use. Accord-
ingly, there exist interdependence relationships among these
criteria, that is, the A-1 criterion influences A-2, B-1, and B-3; the A-
2 criterion influences A-1, B-2, B-3 and C-3 criteria (see Fig. 4).
Where all the CI and CR values considered in this computation
process, show the consistence for the pairwise comparison results.
In order to check network structure or relationship in consid-
ered criteria, we need to have group discussion because the type of
network or relationship depends on the decision makers'
judgment.

Please cite this article in press as: Wey, W.-M., et al., Alternative transit-oriented development evaluation in sustainable built environment
planning, Habitat International (2016), http://dx.doi.org/10.1016/j.habitatint.2016.03.003
8 W.-M. Wey et al. / Habitat International xxx (2016) 1e15

A-3 Design of
pedestrian spaces
A-2 Spa al density
with up-scalable
of commercial and
traffic volume
retail facili es
B-1
Environmental
carrying capacity
for high-density
A-1 Popula on
density in the
development region

B-2 Externality
resul ng from
mixed land use
C-3 Daily living
safety of the
development
region
B-3 Density of
open spaces for
C-2 Equality in improving quality
residence C-1 Inhibi on of life
accessibility for all capacity of floor
people area incen ves
against house

Fig. 4. Interdependences among nine criteria.

Questionnaire Calculation Process and Results (Expert A)

wf ahp w21 w22 w23 w24 w25 w26 w27 w28 w29 Wf anp

A1 0.0778  0.392 0.104 0.000 0.201 0.000 0.128 0.000 0.000 0.000 ¼ 0.1193
A2 0.0452 0.095 0.501 0.000 0.000 0.258 0.089 0.000 0.000 0.078 0.0661
A3 0.1726 0.000 0.045 0.637 0.000 0.000 0.238 0.000 0.000 0.000 0.1505
B1 0.3155 0.199 0.104 0.000 0.520 0.000 0.044 0.000 0.000 0.078 0.2035
B2 0.0379 0.000 0.246 0.000 0.078 0.637 0.000 0.000 0.000 0.193 0.0891
B3 0.1619 0.199 0.000 0.258 0.201 0.105 0.456 0.000 0.000 0.036 0.2067
C1 0.0187 0.046 0.000 0.000 0.000 0.000 0.000 0.500 0.500 0.000 0.0222
C2 0.0187 0.023 0.000 0.000 0.000 0.000 0.000 0.500 0.500 0.193 0.0499
C3 0.1518 0.046 0.000 0.105 0.000 0.000 0.044 0.000 0.000 0.420 0.0926
C.R. 0.05 0.04 0.03 0.04 0.02 0.04 0.05 0.00 0.00 0.02

Questionnaire Calculation Process and Results (Expert B)

wf ahp w21 w22 w23 w24 w25 w26 w27 w28 w29 Wf anp

A1 0.2703  0.3903 0.2351 0.0000 0.2350 0.0000 0.0826 0.0000 0.0000 0.0000 ¼ 0.1982
A2 0.2659 0.0947 0.4748 0.0000 0.0000 0.1530 0.0431 0.0000 0.0000 0.0609 0.1593
A3 0.1541 0.0000 0.1291 0.6370 0.0000 0.0000 0.1502 0.0000 0.0000 0.0000 0.1363
B1 0.1197 0.1700 0.1191 0.0000 0.6248 0.0000 0.2521 0.0000 0.0000 0.2367 0.1622
B2 0.0356 0.0000 0.0418 0.0000 0.0458 0.7766 0.0000 0.0000 0.0000 0.0559 0.0451
B3 0.0256 0.0568 0.0000 0.2583 0.0944 0.0704 0.4466 0.0000 0.0000 0.0240 0.0807
C1 0.0688 0.2180 0.0000 0.0000 0.0000 0.0000 0.0000 0.8571 0.2500 0.0000 0.1293
C2 0.0457 0.0276 0.0000 0.0000 0.0000 0.0000 0.0000 0.1429 0.7500 0.1452 0.0537
C3 0.0144 0.0426 0.0000 0.1047 0.0000 0.0000 0.0254 0.0000 0.0000 0.4773 0.0352
C.R. 0.08 0.06 0.04 0.04 0.10 0.10 0.07 0.00 0.00 0.08

Questionnaire Calculation Process and Results (Expert C)

wf ahp w21 w22 w23 w24 w25 w26 w27 w28 w29 Wf anp

A1 0.0780  0.3520 0.2747 0.0000 0.2854 0.0000 0.2407 0.0000 0.0000 0.0000 ¼ 0.1685
A2 0.1174 0.1803 0.4263 0.0000 0.0000 0.3196 0.1230 0.0000 0.0000 0.1197 0.1536
A3 0.0849 0.0000 0.0952 0.5584 0.0000 0.0000 0.0385 0.0000 0.0000 0.0000 0.0635
B1 0.2736 0.1851 0.1598 0.0000 0.4723 0.0000 0.2044 0.0000 0.0000 0.2442 0.2012
B2 0.2115 0.0000 0.0439 0.0000 0.1697 0.5584 0.0000 0.0000 0.0000 0.1938 0.1798
B3 0.1275 0.1375 0.0000 0.3196 0.0725 0.1220 0.3449 0.0000 0.0000 0.0660 0.1309
C1 0.0226 0.0447 0.0000 0.0000 0.0000 0.0000 0.0000 0.6667 0.3333 0.0000 0.0294
C2 0.0324 0.0309 0.0000 0.0000 0.0000 0.0000 0.0000 0.3333 0.6667 0.0346 0.0334
C3 0.0521 0.0694 0.0000 0.1220 0.0000 0.0000 0.0485 0.0000 0.0000 0.3417 0.0397
C.R. 0.03 0.10 0.04 0.02 0.02 0.02 0.06 0.00 0.00 0.03

Please cite this article in press as: Wey, W.-M., et al., Alternative transit-oriented development evaluation in sustainable built environment
planning, Habitat International (2016), http://dx.doi.org/10.1016/j.habitatint.2016.03.003
 W.-M. Wey et al. / Habitat International xxx (2016) 1e15 9

4.4. Averaged FANP weightings
Three FANP questionnaires were distributed. The final weights
were obtained by summing and averaging the FANP weightings,
and the arithmetic mean was adopted as the mean of the weighting
scores of the three experts. Each expert represented the viewpoints
of the industry, the government, and academia. Because the experts
invited in this study are independent entities from distinct do-
mains, their views are also independent from others. If the samples
are independent from one another, then using the arithmetic mean
is a suitable calculation approach; if the samples are interrelated,
then the geometric mean would apply. Therefore, the arithmetic
mean was adopted for calculating the weighted mean. The FANP
mean calculation process and result is depicted as follows:
The criterion weightings from the FANP questionnaires were
obtained and then averaged. According to the weighted scores, the
criteria were ranked in the order of B-1, environmental carrying
capacity for high-density development (0.1890); A-1, population
density in the development region (0.1620); B-3, density of open
spaces for improving quality of life (0.1394); A-2, spatial density of
commercial and retail facilities (0.1263); A-3, design of pedestrian
spaces with up-scalable traffic volume (0.1168); B-2, externality
By utilizing the application and capacities of GIS, we are able to
be noticed their evaluation index via GIS for each of the specific
criteria. For example, in order to determine the area of the com-
mercial and retail facilities within 500 m of each station, the den-
sity of these facilities in each village was multiplied by the actual
area within 500 m from the station in the village. The calculated
area was used the score for the density of commercial and retail
facilities. Furthermore, regarding the design of pedestrian spaces
with up-scalable traffic volume (Criterion A-3), to determine the
status of the sidewalks in the development area near the station,
graphical data were obtained from the sidewalk information sys-
tem in the Public Works Department of New Taipei City Govern-
ment, from which the sidewalk locations in the region were
selected. These locations were overlaid with the region within
500 m from each station to calculate the sidewalk area within the
development regions, which was used as the index score.
And then, to evaluate the factors associated with earthquakes
and fault lines, a geological sensitivity map obtained from the
regional planning GIS in the Construction and Planning Agency
(CPA) of the Ministry of the Interior was used as the overlay data
source. The map was overlaid with the regional map of the Ankeng
line. Fig. 5 shows the geological sensitivity map of the Ankeng line
from the score obtained from GIS for geologically sensitive land
considered.

8 2 3 2 3 2 39
>
> 0:1193 0:1982 0:1685 > >
>
> 6 0:0661 7 6 0:1593 7 6 0:1536 7>>
>
> 6 7 6 7 6 7>>
>
> 6 0:1505 7 6 0:1363 7 6 0:0635 7>>
>
> 6 7 6 7 6 7 >
>
>
> 6 0:2035 7 6 0:1622 7 6 0:2012 7>>
< 6 7 6 7 6 7=
Expert A6
6 0:0891 7
7 þ Expert B6 6 0:0451 77 þ Expert C66 0:1798 7
7 =3
>
> 6 0:2067 7 6 0:0807 7 6 0:1309 7>>
>
> 6 7 6 7 6 7>>
>
> 6 0:0222 7 6 0:1293 7 6 0:0294 7>>
>
> 6 7 6 7 6 7>>
>
> 4 0:0499 5 4 0:0537 5 4 0:0334 5>>
>
> >
>
: ;
0:0926 0:0352 0:0397
2 3 2 3
A1:Population density in the development region 0:1620
6 A2:Spatial density of commercial and retail facilities 7 6 0:1263 7
6 7 6 7
6 A3:Design of pedestrian spaces with up  scalable traffic volume 7 6 0:1168 7
6 7 6 7
6 B1:Environmental carrying capacity for high  density development 7 6 0:1890 7
6 7 6 7
¼6
6 B2:Externality resulting from mixed land use 7 ¼ 6 0:1046 7
7 6 7
6 B3:Density of open spaces for improving quality of life 7 6 0:1394 7
6 7 6 7
6 C1:Inhibition capacity of floor area incentives against house prices 7 6 0:0603 7
6 7 6 7
4 C2:Equality in residence accessibility for all people 5 4 0:0456 5
C3:Daily living safety of the development region 0:0558

resulting from mixed land use (0.1046); C-1, inhibition capacity of
floor area incentives against house prices (0.0603); C-3, daily living
safety of the development region (0.0558); and C-2, equality in
residence accessibility for all people (0.0456).
4.5. Evaluating and selecting the station project using the GIS
In order to apply the TOD evaluation criteria in assessing the
Ankeng line, an index for each criterion was developed for using the
GIS to evaluate the performance of each station to facilitate the
selection process (Table 2). The remainder of this section describes
the processes involved in using the GIS to evaluate the performance
of each station based on the evaluation indices. Subsequently, the
performance results for each of the stations are presented, and the
optimal station for TOD is selected according to the weighted
results.
In this map, the geological sensitivity is divided into four levels
of disaster potential: zero potential, mild potential, moderate po-
tential, and severe potential. Scores were allocated according to
each level (zero potential ¼ 3; mild potential ¼ 2; moderate
potential ¼ 1; and severe potential ¼ 0). The index scores were
obtained by taking the grid characteristic scores within 500 m of
the stations and multiplying it by the area. The station with the
highest geological sensitivity score (235,573) was K9 station. For
the 10-level standardized operation, the station attained a score of
10. Accordingly, the score of 235,573 was divided into 10 segments,
and the scores within the standardized operation was allocated to
the stations according to their corresponding geological sensitivity
scores, yielding the final results and scores shown in Fig. 6.
For the Evaluation Index via GIS of Criterion B-2, i.e., Mixed Land
Use Status in the Development Region, comparison of the mixed
land use status in the development region. Although TOD

Please cite this article in press as: Wey, W.-M., et al., Alternative transit-oriented development evaluation in sustainable built environment
planning, Habitat International (2016), http://dx.doi.org/10.1016/j.habitatint.2016.03.003
10 W.-M. Wey et al. / Habitat International xxx (2016) 1e15

Table 2
Evaluation index via GIS for each criteria.

Criterion Evaluation Index via GIS

A-1 Population density in the development region Population density in the administrative zone near the station
A-2 Spatial density of commercial and retail facilities Floor density of commercial and retail buildings in the administrative zone near the station
A-3 Design of pedestrian spaces with up-scalable traffic volume Area of existing pedestrian sidewalks in the development region
B-1 Environmental carrying capacity for high-density development Environmentally sensitive land
B-2 Externality resulting from mixed land use Mixed land use status in the development region
B-3 Density of open spaces for improving quality of life Area of parks, green lands, and schools in the development region
C-1 Inhibition capacity of floor area incentives against house prices Extent of floor area incentives offered for urban planning in the region
C-2 Equality in residence accessibility for all people Mean house price in the development region
C-3 Daily living safety of the development region Minimal accessible distance to public safety facilities in the development region

Botanical Garden MRT station (K1) was identified as featuring the

largest total area of open spaces (221,589 m2) in its development
region. According to the 10-level standardized operation, the sta-
tion attained a score of 10. The area of 221,589 m2 was divided into
10 segments, and the standardized operation score was allocated
according to the levels to which the open-space score of each sta-
tion belonged.
Currently, the floor area incentive measures in Taiwan are
specified in the guideline for Land Use Zoning, which is a part of the
detailed urban plans. Therefore, the floor area incentive regulations
vary with each urban planning zone and must be determined ac-
cording to the urban planning zone each region is situated in. In this
study, the floor area incentive implemented in the Xindian urban
planning zone received the highest appraisal, and the station in this
zone attained an index score of 10. Although the floor area incen-
tive in the Ankeng urban planning zone was not as high as that of
Xindian, it displayed a considerable incentive standard, attaining an
index score of 8. The stations located in the Xindian zone include

Fig. 5. Geological sensitivity map of Ankeng

emphasizes mixed land use, an incompatible mix of land uses can

have negative external effects. This study used the 2011 CPA land
use partition map as an overlay for the nine MRT stations in the
study region. The mixed use status of the partition of each station
was investigated using the search command for the location of each
station. The record indicated that K8 (Xinhe Elementary School)
MRT station attained the highest index score.
For the Evaluation Index via GIS of Criterion B-3, i.e., Area of the
Parks, Green Lands, and Schools in the Development Region, we can
elucidate the performance of each station for Criterion B-3 (density
of open spaces for improving quality of life), the types of open-
space land use selected in this study included parks, green lands,
and schools. The 2011 CPA land use partition map was used to
overlay the partition graphical data of parks, green lands, and
schools in the areas with the development region of the Ankeng
line, revealing the distribution status of the open spaces in Ankeng.
The index scores were calculated according to the area of open
Fig. 6. Standardized operation allocated to the stations according to their corre-
spaces in the development region of each MRT station. Erpazhi sponding geological sensitivity scores.

Please cite this article in press as: Wey, W.-M., et al., Alternative transit-oriented development evaluation in sustainable built environment
planning, Habitat International (2016), http://dx.doi.org/10.1016/j.habitatint.2016.03.003
 W.-M. Wey et al. / Habitat International xxx (2016) 1e15 11

Shisihzhang (K9), Xinhe Elementary School (K8), Taiwan Malt 4.5.1. Calculating the optimal station
Factory (K7), the AnheeAnkang Intersection (K6), AnzhongeAnyi For each station, the index scores of the nine criteria were
Road Intersection (K5), and Automobile Anyi Road Intersection organized into a matrix. The score of the alternative project was
(K4); those situated in the Ankeng zone include the Ciaoxine Anyi calculated using the FANP matrix algorithm, in which the FANP
Road Intersection (K3), Meiguie Anyi Road Intersection (K2), and weighting vectors were multiplied by the scores in the matrix. The
Erpazhi Botanical Garden station (K1). results show that Xinhe Elementary School station (K8) attained
According to Evaluation Index of Criterion C-2 (equality in the highest score (7.4624). Table 3 lists the calculation process and
residence accessibility for all people), i.e., Mean House Price in the results, and Fig. 9 shows the ranking of the scores for the stations.
Development Region, lower house prices in the development re-
gion allow low-income families to live in these regions, thus ful-
filling the diversity of residents criterion for the development
region. The index for this criterion was the mean house price in the
development region. House price data were obtained from the Real
Estate Actual Transaction Price Inquiry Service web site (http://lvr.
land.moi.gov.tw/). The transaction prices listed on the site were
obtained from the Department of Land Administration of the
Ministry of the Interior and were thus considered reliable. Addi-
tionally, the site features a function for averaging the house prices
of a selected region. As shown in Fig. 7, numerous transaction data
were scaled to automatically calculate the mean house prices in the
region, thus simplifying the calculation. In addition, the mean
house prices were standardized into score levels, which were used
to allocate scores for each station.
Finally, the scoring of each development region according to the
Evaluation Index via GIS of Criterion C-3 (daily living safety of the
development region), i.e., Minimal Accessible Distance to Public
Safety Facilities in the Development Region was assessed according
to the minimal accessible distance to fire and police stations. A
score of 5 was allocated to MRT stations within 200 m of a police or
fire station; 4 for those within 200e400 m; 3 for those within
400e600 m; 2 for those within 600e800 m; 1 for those within
800e1000 m; and 0 for those further than 1000 m. Thus, whether
the areas are safe for activities of daily living could be determined.
The fire stations in the development region of the Ankeng line
include the AnkangeAnhe Division of the Xindian Squadron as well
as the Ankang Division; police stations within the development
region include Ankang Station, Anhe Station, and Shuangcheng
Station. Through point data conversion, the locations of the five
public facilities were input into Arc GIS 10.0 for individual buffer
analysis. Subsequently, the location overlaying function was used to
examine the buffer zones in which the stations are located. The
stations were then allocated daily living safety scores (Fig. 8). Fig. 8. Scoring for Criterion C-3 (daily living safety of the development region).

Fig. 7. Sketch map of the mean sectional house prices. Source: Show Taiwan 3D Real Estate Actual Transaction Price Inquiry Service (2013).

Please cite this article in press as: Wey, W.-M., et al., Alternative transit-oriented development evaluation in sustainable built environment
planning, Habitat International (2016), http://dx.doi.org/10.1016/j.habitatint.2016.03.003
12 W.-M. Wey et al. / Habitat International xxx (2016) 1e15

Table 3
Calculation table for the optimal station.

wf anp Station A1 A2 A3 B1 B2 B3 C1 C2 C3 Score

A1 0.1620  K9 6 9 7 10 1 1 10 6 0 ¼ 5.9379
A2 0.1263 K8 10 10 5 10 5 2 10 8 6 7.4624
A3 0.1168 K7 2 5 6 10 2 1 10 7 3 4.9855
B1 0.1890 K6 2 5 6 10 2 1 10 8 4 5.0869
B2 0.1046 K5 3 7 4 6 1 3 10 8 0 4.4629
B3 0.1394 K4 2 5 10 5 1 3 10 10 0 4.6513
C1 0.0603 K3 2 4 7 6 1 3 8 10 2 4.3544
C2 0.0456 K2 2 5 8 6 3 2 8 9 5 4.7891
C3 0.0558 K1 2 2 0 4 3 10 8 10 1 4.0354

With these results, the area surrounding Xinhe Elementary School
station was selected as the target for planning a built environment
design.
5. Sustainable built environment planning for the alternative
TOD station
In this study, nine TOD evaluation criteria were selected, the
weightings of which were calculated using the FANP. The priorities
for built environment planning were arranged according to these
weightings. Thus, the priorities in resource allocation were clearly
defined, thereby preventing poor resource management and
wastage.
According to index scores, the station with the highest score
(7.4624) was Xinhe Elementary School station (K8). Therefore, it
was selected as the target for built environment planning. The
planning was formulated according to a combination of the design
principles of the nine criteria and the light-rail transit-oriented
development (LRTOD) planning scale. Suggestions were proposed
for the built environment planning regarding the current status of
the development region of K8 station. The following describes the
particulars of the designs for the criteria:
5.1. B-1 environmental carrying capacity for high-density
development
Regions within 200 m of the station can be used for high-density
development. Geologically sensitive land areas with mild disaster
potential located on the southwest side and northern areas can be
developed into environmentally friendly neighborhood parks. In
addition, the regions near these land areas can be developed into
low-density residential zones. Potential hillside disasters in exist-
ing collective housing zones must be considered. Retaining walls
requiring repairs in sloped areas must be reinforced. Future
development around the geologically sensitive lands must be
evaluated more cautiously.
5.2. A-1 population density in the development region near the
station
High-density mixed residentialecommercial development is
recommended for the predetermined areas near the MRT station
and vacant land owned by the Water Resource Office. This devel-
opment not only increases the population density near the station,
thereby improving the efficiency of the public transportation sys-
tem, but also promotes commercial uses at the station, raising the
convenience in daily living. In the LRTOD scale, the development

Fig. 9. Line chart showing the summed weightings for each station.

Please cite this article in press as: Wey, W.-M., et al., Alternative transit-oriented development evaluation in sustainable built environment
planning, Habitat International (2016), http://dx.doi.org/10.1016/j.habitatint.2016.03.003
 W.-M. Wey et al. / Habitat International xxx (2016) 1e15
density of a region correlates positively with the distance to an MRT
station. To improve the population density in the development
region, we recommend that a renewal plan first be promoted for
the old low-density apartments near the station to increase their
usage density. To promote the quality of daily living in the resi-
dential zones at the boundary of the development region (i.e.,
>200 m from the station), the development of low-density resi-
dential zones in high open-space communities with buildings
approximately 8 stories high should be prioritized.
5.3. B-3 density of open spaces for improving quality of life
Currently, the total area of the open spaces in the development
region constitutes only 14% of the total area of that region, which is
below the open space ratio of 20% expected for the LRTOD. Con-
structing neighborhood parks similar to Tianshan Park is recom-
mended for geologically sensitive land areas with mild disaster
potential, which would provide recreational areas and supplement
the 6% shortfall. The geologically sensitive lands on the southwest
side can be developed into neighborhood parks, which can then
form a circular-shaped park with Tianshan Park located at the
northwest side and Hsintien Creek at the north and east sides. Thus,
open spaces for recreation can be provided, forming a green belt in
the area.
5.4. A-2 spatial density of commercial and retail facilities
The development region contains several institutional sites. The
vacant green area near the Water Resource Office can be developed
for high-density mixed residentialecommercial use, and the in-
stitutions within the region can be moved into the buildings. Thus,
the commercialeretail density around the station can be increased,
providing greater convenience for people living in the area. Addi-
tionally, the released institutional sites could be used to establish
welfare facilities such as social housing.
5.5. A-3 design of pedestrian spaces with up-scalable traffic volume
Adding sidewalks on the connecting roads on both sides of Anhe
Road is recommended to guide pedestrians to the public trans-
portation stations. Pedestrian comfort in the sidewalk areas must
be improved by adding weather coverings or street trees to existing
sidewalks. The tile layouts must be upgraded to prevent floods
during heavy rain. In addition, barrier-free sidewalks must be
constructed to provide disabled accessibility. Arcades must be
regulated strictly and cannot be used arbitrarily as parking spaces
for bicycles or motorcycles, which affects the rights of pedestrians.
5.6. B-2 externality resulting from mixed land use
Currently, only the mixture of “schools in commercial zones”
easily results in the generation of negative external influences;
however, according to the current situations of these schools, the
commercial behaviors surrounding the schools have not caused
excessive disturbance to the schools. The institutional site beside
Xinhe Elementary School is a green area reserved for the MRT
station. According to our recommendation, the site can be used to
develop mixed residentialecommercial or purely commercial
buildings. However, this might have negative external effects on
the school. Circulation elements and access openings must be
established to ensure that commercial behaviors do not affect
student safety or the tranquility of the school.
Please cite this article in press as: Wey, W.-M., et al., Alternative transit-oriented development evaluation in sustainable built environment
planning, Habitat International (2016), http://dx.doi.org/10.1016/j.habitatint.2016.03.003
13
5.7. C-1 inhibition capacity of floor area incentives against house
prices
Current legal regulations contain numerous floor area incentive
measures, with the incentive restrictions being the loosest in re-
gions near MRT stations. Therefore, measures for floor area in-
centives can be used to expand the supply of residential properties
and reduce house prices. Private property developers can provide
additional social housing in the development region and admin-
ister higher floor area incentives. This would not only reduce house
prices, but also increase the supply of social housing. Accelerate
urban renewal plans for old regions near the station, which would
not only improve daily living safety, but also promote the supply of
housing in the regions.
5.8. C-3 daily living safety of the development region near the
station
Xinhe Elementary School Station is approximately 500 m from
both the fire and police stations. Although the two facilities are on
the boundary of the development region, it is nonetheless a reliable
distance. No additional safety facilities require establishment.
Tighten controls over illegal parking on access roads and narrow
alleys to facilitate emergency relief.
5.9. C-2 equality in residence accessibility for all people
Increasing the housing supply in the regions with low house
prices on the right of Anhe Road and adjacent to the school would
reduce house prices in the area. Applicable approaches include
floor area incentives for urban renewal. Mixed use of public facility
sites can be utilized to release additional public land for the gov-
ernment to provide a certain amount of social housing to low-
income earners and other disadvantaged groups. Moreover, floor
area incentives can be offered to encourage developers to provide
social housing to solve the housing problem for these groups,
thereby improving social diversity and equality in residence
accessibility for all people.
6. Study analysis and concluding remarks
We specified sustainable transportation as our goal and applied
TOD as the land use planning approach to achieve this goal in an
effort to solve the urban sprawl caused by the lack of coordination
between land use and transportation infrastructure developments.
According to the results of this study, the analysis and conclusion as
well as suggestion for further studies are presented as follows:
Three subgoals in sustainable transportation (economic effi-
ciency in transportation behaviors, environmental sustainability in
transportation infrastructure, and social equity in transportation
accessibility) were applied as the goals corresponding to the TOD
evaluation criteria. In reference to literature on sustainable trans-
portation and TOD, nine TOD evaluation criteria were formulated.
The fuzzy Delphi method was employed to ensure that the pro-
posed criteria selection approach was objective. The results indi-
cated that all of the selected criteria attained high consensus value,
suggesting that these criteria can be incorporated in future studies
for further analysis. Accordingly, all nine selected TOD evaluation
criteria corresponding to the principle of sustainable transportation
were regarded as critical to the goal of sustainable transportation;
thus, all of them must be considered when evaluating whether a
region is suitable for TOD in order to realize sustainable
transportation.
The FANP results in the second phase of this study revealed that
the environmental carrying capacity for high-density development
14 W.-M. Wey et al. / Habitat International xxx (2016) 1e15
was the highest-scored criterion with a weighted value of 0.1890.
This result indicates that if the environmental carrying capacity of
the region near a station is satisfactory, then the station is a strong
candidate for TOD. A possible reason for this is that after consid-
ering the relationship between TOD and sustainable transportation,
the experts maintained that through high-density, mixed-use, and
favorable urban designs, TOD is aimed at encouraging people who
live near stations to use the public transportation system. However,
such a rigid development approach is not applicable to the envi-
ronment near all stations; hence, the carrying capacity in in each
region must be considered on a case-by-case basis. To protect the
life and property of people living in the development region, this
approach must not be applied in disaster-prone regions. Therefore,
this criterion attained a considerably high weight.
The three lowest-weighted criteria were the inhibition capacity
of floor area incentives against house prices (C-1), the equality in
residence accessibility for all people (C-2), and the daily living
safety of the development region (C-3). This can be ascribed to the
experts indicating that the convenience of living near MRT stations
inevitably inflates house prices in the area. Therefore, to enhance
social equity, promoting social welfare and strengthening care fa-
cilities for disadvantaged groups are considerably more effective
than increasing the number of residential buildings to lower house
prices.
In previous studies that have applied the analytic hierarchy
process or analytic network process, project criteria have typically
been evaluated by experts and scholars. However, experts and
scholars are not fully knowledgeable about the details of the project
and typically allocate scores based on a brief description of the
project. To address this evaluation problem, this study applied the
GIS to evaluate the project. The GIS was used to identify regions
with the highest index scores for each criterion and to determine
the characteristics of these areas. Through a standardized opera-
tion, the objective data were then divided into 10 levels; the highest
level was given a score of 10, and the lowest level was given a score
of 1. Finally, the FANP weightings were summed for the FANP
matrix computation to obtain the final scores for the project. Thus,
the results were relatively objective, which reduced the uncertainty
in the experts' subjective assessment and the rate of error from the
experts’ lack of understanding of the project.
On built environment planning, by re-evaluating the TOD
evaluation criteria that satisfied the goals of sustainable trans-
portation and by considering the appropriate LRTOD scale, we
formulated recommendations for future built environment plan-
ning. The priorities of the stations in the planning were arranged
according to the ranking of the criteria weightings. In the actual
planning of the development region, for each criterion, the devel-
opment status of the region within 500 m of Xinhe Elementary
School station was examined. Subsequently, the statuses were
compared with the suggestions derived from the TOD sustainable
transportation evaluation criteria and the LRTOD scale. Problems
indicated by these statuses were improved. For example, if the ratio
of the public open spaces was insufficient, then more public open
spaces were recommended. If the suggestion according to a crite-
rion was already satisfied, then the planning for this criterion was
unnecessary. Finally, we offered recommendations for built envi-
ronment planning according to the nine criteria.
This paper proposes an integrated approach to identify the most
important criteria that can signify alternative transit-oriented
development evaluation in sustainable built environment plan-
ning. The proposed approach, which systematically integrates both
FDM and FANP techniques into a GIS planning, is used to judge the
sustainable built environment planning and evaluate the overall
impact of MRT stations. The proposed approach also assesses the
impact of identified criteria on the TOD's stations in Taipei city,
Please cite this article in press as: Wey, W.-M., et al., Alternative transit-oriented development evaluation in sustainable built environment
planning, Habitat International (2016), http://dx.doi.org/10.1016/j.habitatint.2016.03.003
Taiwan. The sustainable built environment planning in Taipei has
come to a turning point, as the traffic and population growing has
become a serious concern and put a tremendous pressure on both
environment and people in Taipei. It is therefore important to find
ways or new lifestyles such as TOD formulations that are more
flexible, inclusive, and sustainable.
The approach applies the Fuzzy and ANP method to model the
inner and outer dependences between the criteria in the matrix.
Based on the prioritization of the weights, the relative importance
of principles is determined. Finally, given these relative importance
degrees, the GIS application is utilized to consider both environ-
mental and sustainable planning requirements. These technical
requirements are viewed as feasible and important criteria for
designing urban MRT station environment. The proposed approach
not only identifies the TOD's physical components that contribute
to (and do not contribute to) the urban transportation environment
design when taking end-user's needs into consideration, but also
finds hidden causal relationships between physical components
and user's needs. The paper illustrates the proposed approach in an
empirical study in Taipei for showing the proposed approach is a
valuable tool to planners. The approach enables planners to quickly
identify the features of environment constructs and to output some
important information when testing the present structure of urban
transit-oriented development evaluation in sustainable built envi-
ronment planning. In doing so, planners will be able to select the
best plan option or to make an effective decision. In other words, in
our empirical study, planners will be able to consider end-use
needs when inputting necessary planning elements to improve
urban built environment in Taipei City and increase user satisfac-
tion by utilizing output information generated from the modeling
process.
Acknowledgements
The authors are grateful to the anonymous referee and the Ed-
itor of the journal for their constructive comments that significantly
improve the paper. The authors also wish to express their appre-
ciations to the Ministry of Science and Technology, R.O.C. for the
support of the project MOST 100-2410-H-305-061-MY2. All opin-
ions, errors, omissions and such are solely the responsibility of the
authors.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.habitatint.2016.03.003.
References
Affum, J. K., & Taylor, M. A. P. (1998). Integration of geographic information systems
and models for transport planning, land use and environmental analysis. In
Proceedings of the 1st Asia Pacific conference on transportation and the environ-
ment (vol. 2, pp. 624e632). Singapore: Center for Transportation Research,
National University of Singapore.
Blowers, A. (1993). Planning for a sustainable environment. London: Town and
Country Planning Association.
Calthorpe, P. (1993). The Next American Metropolis. New York: Princeton Architec-
tural Press.
Cervero, R., & Kockelman, K. (1997). Travel demand and the 3Ds: density, diversity,
and design. Transpn Res.-D, 2(3), 199e219.
Cervero, R., Sarmiento, O. L., Jacoby, E., Gomez, L. F., & Neiman, A. (2009). Influences
of built environments on walking and cycling: lessons from Bogota . Interna-
tional Journal of Sustainable Transportation, 3(4), 203e226.
Chang, D. Y. (1992). Extent analysis and synthetic decision, optimization techniques
and applications. World Scientific, 1(1), 352.
Dalkmann, H., & Brannigan, C. (2007). Transport and climate change, module 5e,
sustainable transport, a sourcebook for policy-makers in developing cities, GTZ
global. Federal Eschborn, Germany: Ministry for Economic Cooperation and
Development Accessed 14.03.11 www2.gtz.de/urbanet/library/detail1.asp?
number¼6399.
 W.-M. Wey et al. / Habitat International xxx (2016) 1e15 15

Grant, G., & Tsenkova, S. (2012). New urbanism and smart growth movements. In
International encyclopedia on housing & home. Oxford: Elsevier.
Huang, J. J., Tzeng, G. H., & Ong, C. S. (2005). Multidimensional data in multidi-
mensional scaling using the analytic network process. Pattern Recognition Let-
ters, 26, 755e767.
Kahraman, C., Cebeci, U., & Ruan, D. (2004). Multi-attribute comparison of catering
service companies using fuzzy AHP: the case of Turkey. International Journal of
Production Economics, 87, 171e184.
La Greca, P., Barbarossa, L., Ignaccolo, M., Inturri, G., & Martinico, F. (2011). The
density dilemma. A proposal for introducing smart growth principles in a
sprawling settlement within Catania Metropolitan Area. Cities, 28(6), 527e535.
Lewis, S. (1990). Use of GIS in transport modeling. ITE journal, 60(March), 34e38.
Meijer, M., Adriaens, F., Linden, O., & Schik, W. (2011). A next step for sustainable
urban design in the Netherlands,. Cities, 28, 536e544.
Niemiraa, M. P., & Saaty, T. L. (2004). An analytic network process model for
financial-crisis forecasting. International Journal of Forecasting, 20, 573e587.
Peiser, R. B. (1989). Density and urban sprawl. Land Economics, 65(3), 193e204.
Saaty, T. L. (1996). The analytic network process. New York: RWS Publications.
Sarkis, J. (2003). A strategic decision framework for green supply chain manage-
ment. Journal of Cleaner Production, 11(4), 397e409.
Tiwari, R., Cervero, R., & Schipper, L. (2011). Driving CO2 reduction by integrating
transport and urban design strategies. Cities, 28(5), 394e405.
Trinidad, G. S., & Marquez, L. O. (1994). Interfacing GIS with models for urban
planning and analysis. In CSIRO international seminar, 10e11 August, Higehett
Victoria.
Wey, W. M., & Chiu, Y. H. (2013). Assessing the walkability of pedestrian environ-
ment under the transit-oriented development. Habitat International, 38,
106e118.
World Commission on Environment and Development. (1987). Our common future.
Oxford: Oxford University Press.
Zhao, P. (2010). Sustainable urban expansion and transportation in a growing
megacity-consequences of urban sprawl for mobility on the urban fringe of
Beijing. Habitat International, 34, 236e243.

Please cite this article in press as: Wey, W.-M., et al., Alternative transit-oriented development evaluation in sustainable built environment
planning, Habitat International (2016), http://dx.doi.org/10.1016/j.habitatint.2016.03.003