Professional Documents
Culture Documents
a r t i c l e i n f o a b s t r a c t
Article history: The growing number of studies to measure sustainability in transportation highlights the importance of
Received 7 June 2017 this research topic. Nonetheless, these studies do not address or even consider the sustainability of urban
Received in revised form freight transportation, such as the results of a systematic literature review conducted on the theme
10 January 2018
indicate. However, urban freight transportation has long-term impacts upon sustainability. Consequently,
Accepted 20 February 2018
it is important that freight transport companies operate in more sustainable manners. In this context, we
Available online 27 February 2018
address in this paper the challenge for transport operators in how to evaluate and select sustainable
configurations for urban freight transportation. We present a fuzzy multi-criteria decision making
Keywords:
Urban freight transportation
approach for selecting alternative configurations for the distribution chain in urban areas in terms of
Sustainability sustainability. Fuzzy multi-criteria is used to aggregate scores for sustainability assessment and selection
Indicators of the best alternative, considering thus not only different weights but also uncertainty. An application of
Indices the proposed framework to the assessment of alternative courier operations in Rio de Janeiro, Brazil, is
Fuzzy logic presented to demonstrate its applicability. A sensitive analysis is also carried out to determine how
Multi-criteria model different weights assigned to input parameters might impact the final priority ranking, and thus the final
decision. The framework presented is simple, flexible and wide useful, besides its practical applicability.
Companies and transportation operators could address to such tool to evaluate their distribution chain
and identify more sustainable configurations for their urban deliveries.
© 2018 Elsevier Ltd. All rights reserved.
https://doi.org/10.1016/j.jclepro.2018.02.234
0959-6526/© 2018 Elsevier Ltd. All rights reserved.
728 R.A.M. Bandeira et al. / Journal of Cleaner Production 184 (2018) 727e739
propulsion systems; information systems; and routing optimiza- Nevertheless, research has assessed their use mainly applied to
tion (Arvidsson et al., 2013; Oliveira and D'Agosto, 2017). passenger transportation rather than for UFT (Melo and Baptista,
From this list of potential initiatives, operators seek for those 2017), stressing another innovative character of this study.
able to reduce the impact of their activities, while bringing internal The paper is structured as follows: Section 2 presents a litera-
economic advantages (Browne and Allen, 2011). Arvidsson et al. ture review on STI. Section 3 proposes a set of sustainable indicators
(2013) stress that understanding these initiatives and their im- for UFT, and Section 4 presents the proposed methodology for
pacts is crucial to determine how to operate more efficiently. evaluating sustainable UFT. In Section 5, we apply the methodology
Therefore, from the perspective of transport operators, selecting to two alternative operations in Rio de Janeiro, Brazil, and analyze
initiatives to implement is a complex decision-making process. It the results. Section 6 provides the conclusions and steps for future
involves a diversity of quantitative and qualitative criteria to work.
measure the social, environmental and economic impacts of each
alternative, enabling thus a comparative analysis to determine the 2. Literature review on STI
initiative with the most potential in terms of sustainability.
Measures and concepts for more sustainable UFT have been Indicators are specific variables suited for quantification and
previously addressed in the literature, although most of them treat analysis (Litman, 2008). They must provide reliable information in
this subject in a broader approach (i.e. Munuzuri et al., 2005; Kin a simple and objective way, besides being easily understandable. In
et al., 2017), applied on one city or region (i.e. Dablanc, 2007). Be- general, indicators are useful for decision-making, measuring
sides, these studies focus on public policies (i.e. Munuzuri et al., progress and benchmarking (Egilmez et al., 2015).
2005; Russo and Comi, 2016; Kin et al., 2017), and do not address In this context, indicators can be a powerful tool to monitor
these initiatives in the managerial level, more specifically, to the transport sustainability (Hagshenas and Vaziri, 2012) and conse-
transport operator. quently to help operators to evaluate sustainable UFT. STI have
Beyond that, the perspective of operators is distinct. Public au- already been proposed in the past by different researches, but these
thorities focus mostly on social and environmental outcomes, while studies did not consider UFT. For confirming such assumption, we
transport operators mainly focus on economic outcomes; in other followed the methodology proposed by Thome et al. (2016) to
words, on the economic power for sustaining an activity or oper- develop a SLR to identify previous papers on sustainable indicators
ation (Mihaiu et al., 2010). These specificities make the decision- for UFT, and if they propose a composite index for this evaluation.
making process from the perspective of the operator unique and The research method adopts a SLR approach due to its purpose in
thus we aim to address the following research questions: (i) which developing a holistic conceptualization and synthesis of an
criteria should transport operators assess to select the most indi- emerging topic (herein sustainable indicators for UFT), providing
cated initiative to increase their level of sustainability?; (ii) is there results to reinterpret previous researches or a first step for further
a composite index that enables this analysis?; (iii) which are the research on the topic (Seuring and Gold, 2012).
steps to be followed to develop a composite index for evaluating The search was structured according to the following combi-
sustainable UFT from the perspective of transport operators? nation of keywords (sustainab* AND urban freight transport* AND
To address the first two research questions, we conducted a (index OR indicator)) restricted to the title, abstract and keywords
systematic literature review (SLR) to identify previous papers that in the Scopus database. It resulted in only seven papers, which were
propose sustainable transport indicators (STI) for UFT, and if they excluded because they did not directly propose sustainable in-
proposed a composite index for the evaluation. We address the dicators or an index for UFT, confirming thus the lack of studies on
third research question by proposing a framework that can be used this topic.
by operators to analyze and compare, in terms of sustainability, Therefore, we broadened the scope of the literature review,
alternative configurations for their distribution chain in urban considering any publication on STI. We excluded the restriction on
areas. For that purpose, we develop a fuzzy multi-criteria model for papers that only considered UFT, and extended it to publications on
evaluation and selection of sustainable UFT, considering not un- sustainable indicators for any type of transportation, aiming to
certainty but also different weights for various indicators, identify a list of STI that could be adapted to assess sustainability in
addressing thus a literature gap previously pointed out by Reisei UFT. This result would support the sustainable indicators proposed
et al. (2014) and Santos and Ribeiro (2013). We opted to use fuzzy for UFT.
logic due to its capacity of converting human knowledge and its Earlier studies on STI gave priority to the environmental
decision-making ability into rule-based frameworks and mathe- component of transportation sustainability, and only recent studies
matical formula (Rajak et al., 2016), besides the fact that some input present more balance among environmental, economic and social
parameters could not be determined analytically, requiring expert indicator (Santos and Ribeiro, 2013). Haghshenas and Vaziri (2012)
knowledge (Awasthi et al., 2011). analyzed 17 studies, from 1999 to 2010, on STI, grouping them in
Moreover, we present an application of this methodology to three categories: economic, environmental and social. However,
assess alternative strategies to the distribution process of a postal none of them was related to UFT.
company in Rio de Janeiro, Brazil. The current distribution process Recent studies have not considered UFT in their proposal of STI
is considered the base-line, and an alternative scenario with the either. Awasthi and Chauhan (2011) presented an integrated
adoption of electric tricycles is proposed for a comparative decision-making approach based on Analytical Hierarchy Process
assessment. The relevance of this particular case study is justified (AHP) for analyzing how transportation measures, such as mode
due to the importance of the last-mile of parcel deliveries. The cost sharing or multi-modal transport solutions, impact city sustain-
of global parcel deliveries, which are a key step for postal systems, ability, using a index composed by nine criteria. However, none of
express mail, private courier companies and less-than-truckload them concerns UFT and the paper do not evaluate transport oper-
shipping carriers (Dennis, 2017), amounts to 70 billion euros, and ations per se, but transport measures. Aggregating uncertainty to
the last mile's share in this total reaches 50% (Joerss et al., 2016). this decision-making process, Awasthi et al. (2011) used fuzzy
Furthermore, e-bikes and e-tricycles are sustainable alternatives to TOPSIS for assessing from the perspective of transportation plan-
urban mobility due to its potential of reducing environmental ex- ners the sustainability of urban transportation systems, such as
ternalities and traffic disturbance, outperforming conventional carsharing and park-and-ride. Nevertheless, the criteria and alter-
modes on almost all environmental metrics (Cherry et al., 2009). natives considered did not include UFT either.
R.A.M. Bandeira et al. / Journal of Cleaner Production 184 (2018) 727e739 729
Shiau and Liu (2013) proposed a set of indicators for helping UFT environmental performance indicators: UFT contributes
local governments to measure transportation sustainability at city to 40% of air pollution in European cities (Lindhome, 2010). The
level. They initially categorized 21 indicators into four aspects air pollutants of local impact, and thereby with greater impacts
(economy, environment, society, and energy), but only two of them in urban areas, more common in the transportation sector are
were related to UFT, that is ‘truck loading’ and ‘effect of public depot CO, NOx and PM. Vehicles that run on diesel predominantly
on freight transshipment’. AHP was applied to select ten key in- contribute to the emission of NOx and PM, while cars and mo-
dicators to measure sustainable transportation strategies to be torcycles, that run on gasoline or ethanol, are mainly responsible
implemented in Taipei, Taiwan, but this final list did not include for the emission of CO (D'Agosto and Ribeiro, 2009). Therefore,
indicators or strategies related to UFT. the indicator to measure such impact is ‘emission intensity of air
Reisei et al. (2014) used principle component analysis/factor pollutants (CO, NOx, and PM)’ (in g/t.km). Since there is a high
analysis (PCA/FA) for developing a composite sustainability index to correlation between fossil energy consumption and the emis-
assess transport sustainability for statistical local areas, allowing a sions of greenhouse gases, we opted to use only the indicator
comparative assessment that would help transport planning in ‘fossil energy consumption’ (in TEP/t.1000 km) to measure the
Melbourne, Australia. The proposed index is based on nine in- impact on natural resources consumption. It is important to
dicators, but none of them was related to UFT. The same occurred in reinforce that UFT is more polluting than long distance freight
the study of Haghshenas et al. (2015) when developing an urban transportation due to short trips and stops required in urban
dynamics model to evaluate different transportation development distribution. Besides, UFT also accounts for a significant part of
scenarios. ambient noise in cities, causing discomfort to people. The indi-
Rajak et al. (2016) presented a fuzzy based approach to evaluate cator to measure such impact is noise from the transportation
urban transportation sustainability performance, based on 60 at- system (in dBA). The indicator ‘products transported per m2’
tributes. However, only one out of these attributes is related to UFT indicates the impact of the operation in the public space, a
(freight ton-miles). Ignaccolo et al. (2016) proposed a methodology system that can transport a higher volume (or weight) in less
to evaluate an indicator of “Transportation Energy Dependence” of space occupies less public space and therefore tends to have less
an urban area, which consists of geographically-based socio-eco- impact. However, in UFT, loads tend to be more voluminous than
nomic data and demand flows according to the travel purpose at heavy, so the load capacity in volume is usually used as a per-
the traffic zone scale, and an energy model, to characterize the formance indicator in addition to the use of loading capacity in
vehicle categories. But UFT was not considered either. weight (tons) (Schoemaker et al., 2006), so the indicator is
Table 1 summarizes STI proposed in previous researches that measured in volume (m3) per square meter.
can be adapted to UFT, as well as their frequency of use. UFT economic performance indicators: Although UFT consti-
Eventually, it is important to stress the limited number of tutes a small proportion in the total freight transportation
studies that aggregate individual indicators in a single composite length, it invokes a high proportion of the transportation costs,
index, such indicated by Reisei et al. (2014) due to the difficulty of accounting for 28% of total transportation costs, mainly due to
measuring sustainable transportation directly. In fact, only four access restrictions (delivery windows and vehicle restrictions)
papers from Table 1 proposed a transportation sustainability index (Schoemaker et al., 2006). Transportation costs are standardized
(Zito and Salvo, 2011; Awasthi et al., 2011; Haghshenas and Vaziri, per the momentum of transport (t.km) to be compared through
2012; Reisei et al., 2014). Nevertheless, Zito and Salvo (2011) and the indicator ‘transportation cost per t.km’. The ‘load factor’
Haghshenas and Vaziri (2012) attributed equal importance for the (occupancy rate) is also an economic indicator that reflects the
different indicators, which may not be the case. Reisei et al. (2014) efficiency of transportation operations and hence is associated
overcame this limitation using PCA/FA. However, we cannot to their competitiveness. The ‘average operational speed’ (in
compare the weights of different indicators calculated by this km/h) is also considered an economic indicator that is related to
method since they do not measure indicators’ theoretical impor- traffic congestion, in which UFT plays a key role, decreasing
tance. Awasthi et al. (2011) used fuzzy TOPSIS to overcome this operator productivity, besides causing major disturbance to
limitation and to aggregate uncertainty to the decision-making urban population.
process. Nonetheless, the index proposed in these researches UFT social performance indicators: Accidents in UFT impacts
focus on the perspective of local authorities, aiming to help in the traffic flow, both by the difficulties in removing the vehicle and
process of urban and transport planning, but do not consider the loads, increasing congestion. The indicator ‘fatality and
criteria or alternatives related to UFT. Moreover, none of these re- injured of traffic accidents per vehicles’ is used to measure the
searches focused on analyzing alternative UFT from the perspective system's safety, and thus is related to the impact on health. This
of the transport operator, as the scope of this paper. indicator is calculated through the ratio between the number of
From the SLR, we could not identify in the literature indicators injured and the total of vehicles of the fleet in this operation. To
or a composite index that transport operators could assess to measure the impact of the operation to the lives of the local
evaluate the sustainability of alternative UFT. Therefore, based on population, we use the indicators ‘Income generated per t.km’
this analysis on STI, we propose on Section 3 a set of sustainable and ‘tax generated per t.km’ (in US$ per t.km), which captures
indicators for UFT. the total amount of labor and money reverted to the local
Campos et al. (2009) listed the main impacts that can be asso- population by the distribution operation.
ciated with urban transportation for the three dimensions of sus-
tainability: (i) in the environmental aspect: air pollution, noise and Table 2 presents a list of sustainable indicators for UFT, which
natural resources consumption; (ii) in the social aspect: health, were selected from previous researches on STI or were developed
equality and justness of opportunities; (iii) in the economic aspect: according to criteria identified in the SLR. We recommend the use
urban economy, transports costs, competitiveness and subsides. of such indicators, but adjustments may be required according to
Based on such impacts and on the list of possible STI that can be the circumstances of the case assessed. These indicators can be
adapted to UFT (Table 1), we selected ten sustainable indicators that aggregated in a social, an economic and an environmental index, as
can be used to assess UFT alternatives and classified them according presented in section 4.
to the triple bottom line of sustainability.
730
Table 1
Sustainable transportation indicators from previous studies that can be adapted to UFT.
Table 2
UFT sustainable indicators.
3. Developing the sustainable urban freight transportation their type and linguistic terms, as presented in Table 3.
composite index The input parameters are converted into linguistic variables, in
the fuzzification process, using fuzzy sets (Filippo et al., 2007). For
This section presents the steps for the development of a com- each indicator (Table 3), a fuzzy set must be determined from
posite index for evaluating and selecting sustainable UFT. This in- specialist opinions and then modelled as trapezoidal membership
dex measures multi-dimensional aspects of sustainability that functions.
could not be completely addressed by individual indicators alone.
Nonetheless, a single index cannot address every question, and thus 3.3. Development of the fuzzy network architecture
we use multiple indicators. We propose an environmental, a social
and an economic performance index for UFT, based on the ten Fig. 1 summarizes the variables and presents the fuzzy network
sustainable indicators proposed in Section 3. These indices are then architecture used for calculating the sustainability index for UFT,
aggregated in a composite index. For indicator quantifications, we which has 16 linguistic variables and six inference blocks (IB1 to IB6
propose a fuzzy multicriteria model, considering different weights in Fig. 1). The sustainability indicators inference blocks (at the
and uncertainty. lowest tier) are aggregated into intermediate tiers that provide the
Fuzzy logic enables converting human knowledge and its deci- impact performance indices, which range in a scale from 0 to 10.
sion making ability into a mathematical formula, which is hard to Clustering these indices gives the sustainable UFT composite index
achieve by conventional analytical techniques (Rajak et al., 2016). in a scale from 0 (low level of sustainability) to 10 (high level of
Hence, we opted to use fuzzy logic to develop a sustainable index sustainability).
for UFT since several input parameters require expert judgments to
be determined. Along these lines, we proposed a five-step frame-
work for evaluating sustainable UFT. 3.4. Inference procedure and defuzzyfication
Table 3
Fuzzy variables.
Low M/L M M/H High Steps 1 to 4 result in a final score for assessing and ranking
[( )]
alternative UFT in terms of sustainability. The final ranking in-
11 12 13 14 15
(3) dicates the highest score alternative as the best sustainable UFT,
= and thus recommends its implementation.
n: number of linguistic terms used for the outputs (n ¼ 5); Cargo bikes or tricycles can only cover specific types of business,
since it can only cope with small weight freight, such as parcels, to
Different weights (pij) can be assigned for each intermediate be delivered within short distances, being thus considered viable
variable within each performance index through vector pi1xm (Eq. solutions for postal service deliveries (Melo and Baptista, 2017).
(4)), which is applied to the membership values of the linguistic Under such context, this paper applies the proposed sustainable
term k to obtain a row vector Ai1x5 (Eq. (5)) for each indicator i. UFT composite index to assess the use of electric tricycles in urban
These calculated values of aik are then aggregated in Matrix A (Eq. areas by a postal company. The current distribution process is the
(6)) and used in the center of-maximum method (Eq. (7)) to obtain baseline, and the use of e-tricycles is considered an alternative
the scores for each impact performance index Si. scenario, based on a trial experiment in the districts of Copacabana
and Leme, in Rio de Janeiro, Brazil. These districts present high
pi1xm ¼ ½pi1 …pim (4) population density (36,000 inhabitant/km2) and are located in a
small plain area (4.1 km2), squeezed among a mountain, a lagoon
1 and the beach. These factors provide a high complexity scenario for
Ai1x5 ¼ Pm $pi $Mi ¼ ½ai1 ai2 ai3 ai4 ai5 (5) urban distribution, and thus this is the reason for choosing this area
j¼1 pij
for the experimental test-rig.
=
[ 11
21
31
12
22
32
13
23
33
14
24
34
15
25
35
] (6) In the Traditional Intermodal Distribution (DTI), the postman
walks from the DC to the bus stop and then takes a bus to the
service zone, where he/she performs the deliveries walking and
carrying its own bag. Due to the legal limit of 10 kg for the weight of
ðai1* 0Þ þ ðai2* 2:5Þ þ ðai3* 5Þ þ ðai4* 7:5Þ þ ðai5* 10Þ the bag, the postman must collect the exceeding cargo weight in
Si ¼ P5 (7) predefined support points (PAD), where a light duty vehicle (LDV)
k¼1 aik
brings the additional load to be delivered. At the end of the de-
A vector U (Eq. (8)) is determined to assign the weight u i liveries, the postman returns to the DC taking a bus, such as illus-
attributed each impact factor I (i.e., social, economic and environ- trated in Fig. 2.
mental). Then row vector B1x5 is calculated (Eqs. (9)e(11)). Each In the Electric Tricycle Supported Distribution (DTE), the post-
element bk of B1x5 is the membership's value corresponding to the man drives an electric tricycle from the DC to the first delivery
linguistic term k. point. At the delivery zone, he/she parks the vehicle and carries out
R.A.M. Bandeira et al. / Journal of Cleaner Production 184 (2018) 727e739 733
Neo ¼ noise emission by bus the costs per km of the tricycle and LDV, as indicated in NTC (2014).
do ¼ distance travelled by the postman by bus (km) The total transportation cost per km for each operation is US$
df ¼ distance travelled by LDV (km) US$1.05 for DTI and US$0.74 for DTE.
avb ¼ average bus occupancy (passengers per vehicle)
np ¼ number of postmen attended by LDV ¼ 10 in DTI Average operational speed: For DTI, the postman travels
Fossil Energy Consumption per km: DTE uses e-tricycles with 13.32 km in 2h02min per route, and thus the average opera-
very low energy consumption per km (varying from 0.13 to 0.28 tional speed is 6.5 km/h (Eq. (18)). In DTE, the total distance
kwh/km), besides most of the electric energy consumed in Rio travelled per route (in the tricycle and walking) is 10.74 km in
de Janeiro is of hydroelectric origin, so we will not consider fossil 1h35min, and thus the average speed is 6.8 km/h.
energy consumption in this operation. Nonetheless, the per-
formance of the bus used in DTI, according to the model, fuel V ¼ (tdt)/(tdt) (18)
and years, is 2.3 km/l (MMA, 2013). The performance of the flex-
gasoline LDV used in DTI is 4.56 km/l, according to data provided Wherein, V ¼ average operational speed; tdt ¼ Total distance trav-
by the postal company. Based on this data and on the travelled elled per trip (km); tct ¼ Total round trip (h).
distances, the average fossil energy consumption per km in DTI
is 0.03 TOE/1000 km, as calculated by Eq. (15). Fatality and injured of traffic accidents per t.km: Although
21% of the casualties in traffic accidents in Rio de Janeiro in-
1
*fg *df 1
*fd *do volves bicycles, tricycles or motorcycles, there was no traffic
pf
np þ po
ðavbÞ accident during the DTE trial. Therefore, we could not consider
En ¼ (15) this indicator on this analysis.
ðdf þ doÞ
Wherein: Taxes per t.km: As the same freight is delivered in both oper-
ations, the same amount of taxes is paid and thus this indicator
En ¼ fossil energy consumption per km in DTI will not be assessed in this analysis.
pf ¼ LDV's performance (km/l) Total income per km: is calculated for DTE and DTI according to
po ¼ bus' performance (km/l) Eq. (19). Considering the total distance travelled on each route,
fg ¼ factor of conversion from m3 of gasoline to TOE (tonne of oil 26 days of operation per month and the monthly income
equivalent) ¼ 0.86 TOE generated (US$ 220.07 of wages for postman, driver and main-
fd ¼ factor of conversion from m3 of diesel to TOE (tonne of oil tenance staff plus taxes for DTI, and US$ 148.20 for DTE), the
equivalent) ¼ 0.98 TOE total income generated per km is US$ 0.81 for DTI and US$0.53
do ¼ distance travelled by the postman by bus (km) for DTE. Nonetheless, due to difficulties presented by the spe-
df ¼ distance travelled by LDV (km) cialists to attribute the lower and upper limits to this indicator in
avb ¼ average bus occupancy (passengers per vehicle) the fuzzyfication process, we opted to use the ratio between the
np ¼ number of postmen attended by the LDV ¼ 10 in the DTI value of the total income generated per km by the assessed
Products transported per m2: Most of the DTI is carried out on operation and the average value to the city of Rio de Janeiro as
bus and a large part of both operations is carried out on foot. this indicator. The specialists felt more comfortable on estab-
Therefore, neither operation occupies significant amount of lishing limits to the indicator as a comparison ratio.
public space on the streets and parking areas to perform their
deliveries. Consequently, we considered that this indicator is not tw
In ¼ (19)
entirely representative in this analysis and it was not tdt$nd
considered. Wherein:
Load factor: Each postman deliveries on average 40 kg per day
in e-tricycles of 50 kg of capacity, so in DTE the load factor of the In ¼ total income per km
vehicle is 80%. In DTI, a LDV delivers on average 40 kg of extra tw ¼ total wages for postman, drivers and maintenance staff
loads for each postman, and, in Copacabana and Leme, each LDV plus taxes
with load capacity of 800 kg attends 10 postmen with an tdt ¼ Total distance travelled per trip (km)
amount of 400 kg. So, the load factor for DTI is 50%. nd ¼ number of days of operation per month
Total transportation cost per km: was calculated according to
Eqs 16 and 17 developed by the National Association of Freight The average income generated per km in UFT in Rio de Janeiro
Transport and Logistics (NTC, 2014). was calculated considering the minimum wage of US$410.65 for
truck drivers (SCVRTTCGP, 2016) and that 0.73% of Brazilian
TCTID ¼ (cm/h * tct) þ (bf * n) þ (cv/h *vct) þ (cv/km *tdv) (16) economically active population (EAP) works in road freight or
distribution activities (IBGE, 2014). Since the EAP of the city of Rio
Where, TCTID ¼ Total cost of TID; cm/h ¼ Cost of postman per hour; de Janeiro is 3,083,000 inhabitants, it is estimated that 22,505
tct ¼ Total round trip time; bf ¼ Bus fare (US$ 1.00); n ¼ Number of employees work on UFT, generating a monthly income of US$
bus trips; cv/h ¼ Cost of LDV driver per hour; vct ¼ LDV round trip 9,536,297.88. Considering that the total number of internal-
time; cv/km ¼ Cost of the LDV per km; tdv ¼ Total distance travelled external and external-internal displacements for UFT in the city is
by the LDV (km). equivalent to 16,647 trips/day (PDTU, 2014) and that these move-
ments occurred only in business days (22 per month), an amount of
TCDET ¼ (cm/h * tct) þ (ct/km * tdt) (17) US$ 26.03 of income is generated per trip. However, there is no
survey on the average distance travelled by routes for UFT in Rio de
Wherein, TCDET ¼ Total cost of DET; cm/h ¼ Cost of postman per Janeiro. Jaller et al. (2016) measured an average distance of 48.1 km
hour; tct ¼ Total round trip time; ct/km ¼ Cost of the tricycle per per daily route to UFT in the metropolitan area of Mexico City
km; tdt ¼ Total distance travelled by the tricycle (km). (1485 km2), thus, given that the area of the metropolitan region of
Costs related to vehicle depreciation, maintenance, taxes, Rio de Janeiro is 1255 km2 and that both are megacities of devel-
compulsory insurance and, fuel were considered when calculating oping countries, this value was adopted to Rio de Janeiro, obtaining
R.A.M. Bandeira et al. / Journal of Cleaner Production 184 (2018) 727e739 735
Table 5
Values of the input parameters.
thus the value of US$0.54 of income generated per kilometer (I1, I2 and I3, respectively) and the sustainability composite index
travelled. Aggregating 96% of taxes applied to wages in Brazil, we (SIUFT) for each alternative UFT were calculated considering the
have a total income of US$ 1.06/km, so the ratio values for DTI and same weights for each input and intermediate variable in the
DTE are, respectively, 0.76 and 0.5. model, such as follows:
Table 5 summarizes the values of the input parameters for DTI
and DTE. It is important to reinforce that even though we initially
4.5. Evaluation and selection of best alternative
considered the ten UFT sustainable indicators proposed in Section
2, only seven of them apply to the scenario analyzed.
Fig. 4 presents the results for DTI and DTE and shows that DTE
Table 6 presents the fuzzy sets for the input parameters, ob-
presents slightly better results for the sustainable composite index,
tained from interviews with seven Brazilian specialists on UFT. The
as well as a better economic performance than the DTI. Both op-
specialists are academics with PhD on transport engineering and
erations present similar environmental performances, but DTE also
with experience in the area varying from 5 to 30 years.
presents better results regarding this category. Nonetheless, DTI
presents better social impacts than the DTE, with a higher I3.
4.3. Development of the fuzzy network architecture For the results presented in Fig. 4, we have considered that the
environmental, economic and social impacts have equal impor-
The fuzzy network architecture used to obtain the sustainability tance and thus we used the same weight for each performance
index for ranking DTI and DTE is similar to that presented in Fig. 1 indicator when calculating the sustainable composite index
with the required adaptation due to the modifications in the set of (SIUFT). Nonetheless, this may not be the case and thus it becomes
indicators, such as presented in Table 5. necessary to verify the importance of the weights attributed to the
input criteria in the selection of the best alternative for UFT. For that
4.4. Inference procedure and defuzzyfication reason, we conducted a sensitivity analysis with the purpose of
identifying how sensitive SIUFT is to changes in the weights
The environmental, economic and social performance indices attributed to each impact performance indicator. Consequently, we
Matrix A:
DTI DTE
L M H L M H
[
0 0 0 0.1 0.9
] [
0 0 0 0 1
= 0 0.5
0.3 0
0 0
0.7 0
0
0
0 0 0.2 0.8 0
0.8 0 0.2 0 0 ]
Performance index Ii:
DTI DTE
o Environmental index: I1=9.75 I1=10
o Economic index: I2=2.5 I2=7.0
o Social index: I3=3.5 I3=1
Vector B:
DTI DTE
o B= [0.1 0.167 0.233 0.033 0.3] [0.267 0 0.133 0.267 0.333]
Table 6
Fuzzy sets for the input parameters.
Low [c1, a1, b1, d1] Medium [c2, a2, b2, d2] High [c3, a3, b3, d3]
10
9
8
7
6
5 DTI
4
3 DTE
2
1
0
Environmental Economic Social SIUFT
performance performance performance
vary the weights attributed for each impact performance indicator right weight for each criterion in the decision process and also
(I1, I2 and I3) individually from 1 to 10, and observe the effects on highlights the relevance of conducting a sensitive analysis in situ-
SIUFT, such as illustrated in Figs. 5e7. ations in which there is uncertainty regarding the definition of the
Figs. 5 and 6 show that changes in the weights of the environ- relevance of different criteria.
mental and economic performance indicators, when analyzed At long last, the case study shows that the use of electric tri-
individually, do not invert SIUFT scores for DTI and DTE. However, cycles as an alternative to the last mile postal distribution in the city
variations in the weight of the social indicator, when analyzed of Rio de Janeiro was able to promote economic and environmental
individually, do change SIUFT scores to the point that DTI presents benefits, corroborating to the results presented by Cherry et al.
better results, that is, if a higher importance is given to the social (2009) that indicates that electric bikes and tricycles outperform
impact of the operation, DTI may be considered more sustainable other modes on almost all environmental metrics.
than DTE. This result reinforces the importance of choosing the Moreover, the case study confirms the consistency and
8.80
8.30
7.80
7.30
6.80
6.30
SIUFT
5.80
0 2 4 6 8 10
DTI DTE Weight
Fig. 5. Sensitivity analysis for the SIUFT varying the weigh for the environmental performance index.
R.A.M. Bandeira et al. / Journal of Cleaner Production 184 (2018) 727e739 737
8.00 5. Conclusion
7.00
SIUFT
7.00
6.00
5.00
4.00
3.00
2.00
1.00
0.00
1 2 3 4 5 6 7 8 9 10
DTI DTE
Fig. 7. Sensitivity analysis for the SIUFT varying the weigh for the social performance index.
738 R.A.M. Bandeira et al. / Journal of Cleaner Production 184 (2018) 727e739
we chose the presented case study, the proposed methodology is global comparison. Ecol. Indicat. 15 (2012), 115e121.
Haghshenas, H., Vaziri, M., Gholamialam, A., 2015. Evaluation of sustinable policy in
not specific to postal operation. The case study confirms the con-
urban transportaion using system dynamics and world cities data : a case study
sistency and applicability of the framework, so it can be applied to in Isfahan. Cities 45 (2015), 104e115.
assess, in terms of sustainability, alternative configurations of any IBGE, 2014. Pesquisa Anual de Serviços 2014 e Se rie Atualizada de 2012-2013.
other distribution chain in urban areas. As a matter of fact, we can Instituto Brasileiro de Geografia e Estatística, Rio de Janeiro, RJ.
Ignaccolo, M., Inturri, G., Le Pira, M., Capri, S., Mancuso, V., 2016. Evaluating the role
apply the proposed methodology in a comparative assessment for of land use and transport policies in reducing the transport energy dependence
almost any set of alternatives of UFT operations, such as deliveries of a city. Res. Transport. Econ. 55 (2016), 60e66.
from centers of distribution to stores, home-deliveries from e- Jaller, M., S anchez, S., Green, J., AndFandin ~ o, M., 2016. Quantifying the impacts of
sustainable city logistics measures in the Mexico City Metropolitan Area.
commerce or even to urban waste collection. The application of Trans.Res. Proc. 12, 613e626. https://doi.org/10.1016/j.trpro.2016.02.015.
such methodology to assess different case studies could be the Joerss, M., Schro €der, J., Neuhaus, F., Klink, C., Mann, F., 2016. McKinsey & Company
subject of research for future studies. Parcel Delivery: the Future of Last Mile, pp. 1e32.
Joumard, R., Gudmundsson, H., Folkeson, L., 2011. Framework for assessing in-
Beyond that, the case study indicated that, although the pro- dicators of environmental impacts in the transport sector. Transport. Res. Rec.: J
posed sustainable UFT indicators can be used as input variables to Transport Res Board 2011 (2242), 55e63.
the model, the procedure enables adaptations if the characteristics Kim, G., Han, S., 2011. Comparative analysis of transportation sustainability in OECD
countries. J. East Asia Soc. Transp. Stud. 9, 82e97.
of the assessed operation and data availability require them. For Kin, B., Verlinde, S., Mommens, K., Macharis, C., 2017. A stakeholder-based meth-
instance, due to the lack of data on UFT in Rio de Janeiro, changes odology to enhance the success of urban freight transport measures in a multi-
were necessary in the units of measurement of several indicators level governance context. Res. Transport. Econ. 65, 10e23.
Lindhome, M., 2010. A sustainable perspective on urban freight transport: factors
for this application. This does not invalidate the proposed meth-
affecting local authorities in the planning procedures. Proc. Soc. Behav. Sci. 2,
odology and neither the presented application, but it reinforces the 6205e6216.
importance of measuring data and statistics on UFT. Lindhome, M., Behrends, S., 2012. Challenges in urban freight transport planning e
Finally, the framework presented is simple, flexible and wide a review in the Baltic Sea Region. J. Transport Geogr. 22 (2012), 129e136.
Litman, T., 2008. Sustainable Transportation Indicators, Sustainable Transportation
useful, so it can be applied for evaluating the sustainability of other Indicators Subcommittee of the Transportation Research Board.
UFT operations, as long as the required data is available. The pro- Macharis, C., Woxenius, J., Lier, T., 2014. The 4 As of Sustainable Logistics. Emerald).
posed composite index helps transport operators to choose from McKinnon, A.C., Browne, M., Whiteing, A.E., 2012. Green logistics. In: Improving the
Environmental Sustainability of Logistics. Kogan Page.
potential alternatives for distribution configuration of UFT opera- Mello, S., Baptista, P., 2017. Evaluating the impacts of using cargo cycles on urban
tions. Moreover, it can be used for benchmarking among different logistics: integrating traffic, environmental and operational boundaries. Eur.
urban freight delivery companies as well, and thus clients could use Transp. Res. Ver 9 (30), 29e30.
Mihaiu, D., Opreana, A., Cristescu, M.P., 2010. Efficiency, effectiveness and perfor-
this index in their decision-making process regarding the selection mance of the public sector. Rom. J. Econ. Forecast 13, 132e147.
of transport companies. Furthermore, it can be applied as the basis MMA, 2000. Resoluça ~o CONAMA nº 272 de 2000 estabelece o limite m aximo de
for the development of a sustainability standard and certification ruído emitido por veículos de carga e passageiro em aceleraça ~o. Ministe
rio do
Meio Ambiente, Brasília, DF.
program for UFT. Such program would recognize participants that MMA, 2013. Inventa rio Nacional de Emisso ~ es Atmosfe ricas por Veículos Automo-
achieve a minimum level of performance. Based on the proposed tores Rodovia rios 2013: Ano-base 2012. Ministe rio do Meio Ambiente, Brasília,
STI and composite index, specific targets would be set for different DF.
Munuzuri, J., Larraneta, J., Onieva, L., Corte s, P., 2005. Solutions applicable by local
levels of achievement.
administrations for urban logistics improvement. Cities 22, 15e28.
NTC, 2014. Manual de Ca lculo de Custos e Formaça ~o de Preços 2014. NTC&Logística,
Braslia, DF. Available at. http://www.portalntc.org.br/media/images/
References publicacoes/manual-de-calculo-e-formacao-de-precos-rodoviario-2014/.
Accessed in March 17, 2017.
Appleton, B., Davies, M., 2008. Smart transportation ranking report. Green Apple OECD, October 1999. Indicators for the Integration of Environmental Concerns into
Canada.Arvindsson, 2013. Transport Policies. Environment Policy Committee. Working Group on the State
Arvidsson, N., Woxenius, J., Lammga, C., 2013. Review of road Hauliers' measures for of the Environment.
increasing transport efficiency and sustainability in urban freight distribution. Oliveira, C., D'Agosto, M., 2017. Reference Guide on Sustainability. Instituto Brasi-
Transport Rev. 33 (1), 107e127. leiro de Transporte Sustenta vel, Rio de Janeiro.
Awasthi, A., Chauhan, S., 2011. A hybrid approach integrating Affinity Diagram, AHP Oliveira, L., D'Agosto, M., Fernandes, V., Oliveira, C., 2014. A financial and environ-
and fuzzy TOPSIS for sustainable city logistics planning. Appl. Math. Model. 36, mental evaluation for the introduction of diesel-hydraulic hybrid-drive system
573e584. https://doi.org/10.1016/j.envsoft.2010.11.010. in urban waste collection. Transport. Res. D 31, 100e109.
Awasthi, A., Chauhan, S., Omrani, H., 2011. Application of fuzzy TOPSIS in evaluating PDTU, 2014. Plano Diretor de Transporte Urbano da Regi~ ao Metropolitana do Rio de
sustainable transporation systems. Expert Syst. Appl. 12270e12280, v. 38. Janeiro. Secretaria de Estado de Transporte. Governo do Estado do Rio de
Browne, M., Allen, A., 2011. Enhancing the sustainability of urban freight transport Janeiro.
and logistics. Transport Commun. Bull. Asia Pac 80, 1e19. Pitfield, D.E., Castillo, H., 2010. ELASTICda methodological framework for identi-
Campos, V., Ramos, R., Correia, D., 2009. Multi-criteria analysis procedure for sus- fying and selecting sustainable transport indicators. Transp. Res. D 15, 179e188.
tainable mobility evaluation in urban areas. J. Adv. Transport. 43 (4), 371e390. Quack, H., 2012. Improving urban freight transport sustainability by carriers - best
Cherry, C.R., Weinert, J.X., Xinmiao, Y., 2009. Comparative environmental impacts of practices from The Netherlands and the EU project CityLog. Proc. Soc.. Behav.
electric bikes in China. Transp. Res. Part D Transp. Environ 14, 281e290. https:// Sci. 39, 158e171.
doi.org/10.1016/j.trd.2008.11.003. Rai, H., Lier, T., Meers, D., Macharis, C., 2017. Improving urban freight transport
Choubassi, C., Seedah, D., Jiang, N., Walton, C., 2016. Economic analysis of cargo sustainability: policy assessment framework and case study. Res. Transport.
cycles for urban mail delivery. Transport. Res. Rec. 2547, 102e110. Econ. 64, 26e35.
D'Agosto, M., Ribeiro, S., 2009. Assessing total and renewable energy in Brazilian Rajak, S., Parthiban, R., Dhanalakshmi, R., 2016. Sustainable transportation systems
automotive fuels. A life cycle inventory (LCI) approach. Renew. Sustain.Energy perfromance evaluation using fuzzy logic. Ecol. Indicat. 71, 503e513.
Rev. 13 (6e7), 1326e1337. Reisei, M., Aye, L., Rajabifard, A., Ngo, T., 2014. Transport sustainability index:
Dablanc, L., 2007. Goods transport in large European cities: difficult to organize, Melbourne case study. Ecol. Indicat. 43 (2014), 288e296. https://doi.org/10.
difficult to modernize. Transport. Res. Part A. Davis 41, 280e285. 1016/j.ecolind.2016.07.031.
Dennis, W. Parcel and small package delivery industry paperback. Available online: Russo, F., Comi, A., 2016. Urban freight transport planning towards Green goals:
https://www.amazon.com/Parcel-Small-Package-Delivery-Industry/dp/ synthetic environmental evidence from tested results. Sustainability 8, 381.
1461021545 Accessed in August 04, 2017. https://doi.org/10.3390/su8040381.
ECOSTART, 2014. Bicicletas e Triciclos Ele tricos. Available at: http://ecostart.com.br/ Santos, A., Ribeiro, S., 2013. The use of sustainability indicators in urban passenger
information/index/sobre-a-ecostart. Acessed 6 November 2016. transport during the decision-making process : the case of Rio de Janeiro, Brazil.
Egilmez, G., Gumus, S., Kucukvar, M., 2015. Environmental sustainability bench- Curr. Opin. Environ. Sustain. 5, 251e260. https://doi.org/10.1016/j.cosust.2013.
marking of the U.S. and Canada metropoles: an expert judgment-based multi- 04.010.
criteria decision making approach. Cities 42, 31e41. Schoemaker, J., Allen, J., Huschebeck, M.E., Monigl, J., 2006. Quantification of Urban
Fillipo, S., Ribeiro, P.C.M., Ribeiro, S.K., 2007. A Fuzzy Multi-Criteria Model applied to Freight Transport Effects I. Best Urban Freight Solutions II.
the management of the environmental restoration of paved highways. Trans- SCVRTTCGP, 2016. Circular Informativa nº 001/Agosto 2016. Sindicato dos Con-
port. Res. Part D 12 (6), 423e436. dutores de Veículos Rodovi arios e Trabalhadores em Transporte de Carga em
Haghshenas, H., Vaziri, M., 2012. Urban sustainable transportation indicators for Geral e Passageiros no Município do Rio de Janeiro, Rio de Janeiro-RJ. Available
R.A.M. Bandeira et al. / Journal of Cleaner Production 184 (2018) 727e739 739
at. http://www.rodoviariosrio.com.br/circular-cct-carga-2016.htm. Accessed in review in operations management. Prod. Plann. Contr. 10.1080/09537287.2015.
November 2016. 1129464.
Seuring, S., Gold, S., 2012. Conducting content-analysis based literature reviews in Toth-Szabo, Z., Varhelyi, A., Koglin, T., Angjelevska, B., 2011. Measuring Sustain-
supply chain management. Supply Chain Manag.: Int. J. 7 (5), 544e555. ability of Transport in the City d Development of an Indicator-set. Department
Shiau, T., Liu, J., 2013. Developing an indicator system for local governments to of Technology and Society, Lund University, Bulletin 261. Traffic & Roads. Lund.
evaluate transport sustainability strategies. Ecol. Indicat. 34, 361e371. https:// Zito, P., Salvo, G., 2011. Toward an urban transport sustainability index: an European
doi.org/10.1016/j.ecolind.2013.06.001. comparison. Eur. Transport Res. Rev. 3, 1e17.
Thome , A.M., Scavarda, L.F., Scavarda, A.J., 2016. Conducting systematic literature