Journal of Cleaner Production xxx (2016) 1e13

Contents lists available at ScienceDirect

Journal of Cleaner Production

journal homepage: www.elsevier.com/locate/jclepro

Life Cycle Assessment and Multi-criteria Decision Analysis: Selection

of a strategy for domestic food waste management in Rio de Janeiro
Ana Carolina Maia Angelo a, *, Anna Bernstad Saraiva a, Joa
~o Carlos Namorado Clímaco b,
Carlos Eduardo Infante c, Rogerio Valle a
a
Production Engineering Program, Federal University of Rio de Janeiro, COPPE/UFRJ, Rio de Janeiro, Brazil
b
Universidade de Coimbra, INESC-C, Coimbra, Portugal
c ~o Joa
Federal University of Sa ~o del-Rei, UFSJ, Minas Gerais, Brazil

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

Article history: Results from Life Cycle Assessment (LCA) of solid waste management are, in many cases, difficult to
Received 3 March 2016 interpret because there is no entirely satisfying alternative. Therefore, most of the studies focus on a
Received in revised form small number of environmental impact categories. While this might facilitate the interpretation of the
8 December 2016
results, it also creates a risk of excluding relevant aspects from the assessment. For instance, the limited
Accepted 11 December 2016
Available online xxx
impact-coverage of Carbon footprint calculations where exclusively climate change related greenhouse
gas emissions are considered. On the other hand, the inclusion of a larger number of impact categories in
the study creates a risk of conflicting results. Multi-criteria Decision Analysis (MCDA) is a powerful
Keywords:
Life Cycle Assessment
approach for decision aiding, where highly diversified indicators can be analysed in the same framework
Food waste helping to organise the available information and identify pros and cons in the decision process as well as
Multi-criteria analysis its aggregation. This paper aims at improving decision making in solid waste management by combining
VIP-Analysis LCA and MCDA techniques harmonically, using management of food waste from households in the city of
Rio de Janeiro as a case study. Oriented to support decisions in choice problems, the software VIP-
Analysis was used to choose the best food waste treatment option among the results of the LCA con-
ducted. The VIP-Analysis is perfectly applicable in aggregating the LCA results due to its ability to handle
problems with impreciseinformation, improving the decision making process.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction priority on managing solid waste: non-generation, reduction,

The Brazilian National Solid Waste Policy (PNRS - Política
Nacional de Resíduos So lidos) was implemented in 2010 after 20
years of congress debates. It has established three main goals:
eradication of dumps by 2014, preparation and implementation of
Municipal Solid Waste Plans at state and municipal level, and
implementation of selective collection and composting systems for
treating the organic fractions of municipal solid waste (BRASIL, Law
No. 12,305, 2010). Nevertheless, few municipalities have really
eliminated their dumps e 45.95% of total solid waste generated in
Brazil in 2014 was still disposed of in open dump sites and 14.65% in
controlled landfills (IPEA, 2014).
Similar to the waste hierarchy suggested by the European
Commission (Directive, 2008/98/EC), PNRS considers the following
* Corresponding author.
E-mail address: ana.maia@sage.coppe.ufrj.br (A.C.M. Angelo).
reutilisation, recycling, solid waste treatment and environmentally
adequate disposal. In general, applying this waste hierarchy should
deliver the best overall environmental option. However, in cases
where specific circumstances and/or specific waste streams lead to
deviations from the waste hierarchy, the waste hierarchy can be
complemented by Life Cycle Assessment (LCA), a quantitative
method to assess the environmental performance throughout the
entire life cycle of a product, in order to identify the best environ-
mental option of waste treatment. Moreover, Life Cycle Thinking, as
a conceptual approach, is considered a strategic issue to guarantee
the sustainable use of natural resources (European Commission,
2005, 2008; Manfredi and Pant, 2011).
A recent review shows LCA to be the principal method used for
evaluating solid waste management systems (SWMS) (Allesch and
Brunner, 2014) and a corresponding substantial increase in the
number of publications over the last decade (Laurent et al., 2014).
The same review highlights agreements with ISO standards and

http://dx.doi.org/10.1016/j.jclepro.2016.12.049
0959-6526/© 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Angelo, A.C.M., et al., Life Cycle Assessment and Multi-criteria Decision Analysis: Selection of a strategy for
domestic food waste management in Rio de Janeiro, Journal of Cleaner Production (2016), http://dx.doi.org/10.1016/j.jclepro.2016.12.049
2 A.C.M. Angelo et al. / Journal of Cleaner Production xxx (2016) 1e13

ILCD Handbook recommending the inclusion of all relevant impact
categories in performance of LCA (ISO, 2006; European
Commission, 2010). However, most SWMS LCAs focus on a small
number of environmental impact categories. While this might
facilitate the interpretation of the results, it also creates a risk of
excluding relevant aspects from the assessment. On the other hand,
the inclusion of a larger number of impact categories in the study
creates a risk of conflicting results, i.e. one alternative can be
preferable in relation to some aspects, while another is preferable
in relation to others. Thus, while results from a LCA can facilitate
understanding of the pros and cons of each waste management
option, determining the best alternative might be difficult.
Multi-criteria Decision Analysis (MCDA) is a powerful approach
for decision aiding, where highly diversified indicators can be
analysed in the same framework helping to organise the available
information and identify pros and cons in the decision process as
well as its aggregation (Matteson, 2014; Clímaco and Valle, 2016;
Valle and Clímaco, 2015; Recchia, 2011). Roughly speaking, MCDA
aims to help managers make better decisions. In complex problems,
MCDA interactive procedures avoid a final aggregation of the
preferences of decision agents based on a unique criterion, in some
cases proposing the combination of algorithmic protocols with the
experience and intuition of decision agents in the process of pref-
erence aggregation.
This paper aims to improve decision aiding in solid waste
management by combining LCA and MCDA techniques, using
management of food waste from households in the city of Rio de
Janeiro as a case study. MCDA seems useful for improving aggre-
gation of LCA results. However, the more commonly used aggre-
gation procedure, a simple weighted sum of the normalised values
of criteria results, is not acceptable in most situations (Rowley et al.,
2012). Therefore, other approaches should be exploited, which take
into account that waste management models should combine
MCDA and LCA harmonically, in order to maximise their strengths
and/or minimise their weaknesses (Karmperis et al., 2013). In this
paper we propose an interactive learning oriented multi-criteria
tool based on the additive model only requiring partial/imprecise
information on the weights.
This paper is organised into six sections, including this intro-
duction. Section 2 describes the SWMS in the city of Rio de Janeiro.
Section 3 presents a short overview of the use of MCDA in LCA and
SWMS. Section 4 presents the methodology, including the LCA
study of the current situation of domestic food waste management.
It also shows how the study will integrate the LCA study with the
MCDA chosen method, in our case VIP-Analysis (Variable Interde-
large quantity generators produce over 120 L or 60 kg of solid waste
each day and, for this reason, in accordance with municipal law
3723/2001 they must use a private waste management company
(SMAC, 2012). The household waste composition is presented in
Table 1. Data on food waste generation was obtained only as an
average for waste of animal and vegetable origin. However, as both
the potential biogas generation and the nutrient content differ
largely between food waste of animal origin and food waste of
vegetable origin (Riber et al., 2009), it is important, in the present
study, to establish the relation between them. Due to a lack of data
from the case study area in question, the relation was based on a
Swedish case, where detailed assessments of food waste compo-
sition were performed (Bernstad and Andersson, 2015), stating a
relation of 12% and 88% for animal and vegetable origin, respec-
tively, in domestic food waste from urban areas.
There are 160 neighbourhoods in the municipality of Rio de
Janeiro, but only 41 of them have access to formal systems for
source-segregation of dry materials, such as metals, plastics, glass,
paper and cardboard (SMAC, 2012). The source-segregation sys-
tems in these 41 neighbourhoods account for only 0.22% of the total
solid waste collected by COMLURB each year (COMLURB, 2013a).
Post-separation of recyclables is performed on a small amount of
waste collected from households in the city. The organic fraction
separated in this process is later mixed with organic waste from
commercial generators, for later treatment through composting.
Produced compost (150 tons/month), cannot be used in agricultural
activities due to the high content of inorganic matter (COMLURB,
2013b). Therefore, it is used by the municipality in landscaping
projects (SMAC, 2012).
As recommended by the PNRS waste hierarchy, alternatives to
landfills should be found for treatment of organic waste. The
Municipal Policy of Integrated Management of Solid Waste of Rio de
Janeiro (SMAC, 2012) has established guidelines and targets for
increased diversion of organic waste from landfills but, as for now,
no investigations of the environmental effects of different alter-
natives for achieving this have been made.
3. Application of MCDA in LCA and SWMS
The use of MCDA methods is no longer uncommon in
Table 1
Household waste composition in the city of Rio de Janeiro, based on data from
COMLURB, 2013a,b.

pendent Parameters - Analysis). In Section 5 LCA results are used as Waste type % Waste fractions %
inputs to the VIP-Analysis software. The aim is to select the Paper and cardboard 16.9 Other clean paper 11.7
preferred environmental option for handling the organic fractions Paper and carton containers 3.7
of household waste. Finally, conclusions are presented in Section 6. Milk cartons 1.5
Plastic 19.0 Hard plastics 4.4
Soft plastics 12.7
2. Household waste management in the city of Rio de Janeiro
Plastic bottles 1.9
Glass 3.4 Clear glass 1.6
The municipality of Rio de Janeiro has more than 2 million Brown glass 1.8
homes and 97% of its total population live in urban areas (IBGE, Food waste 52.8 Animal food waste 9.9
2010). The SWMS in the city of Rio de Janeiro is carried out by Vegetable food waste 42.9
Metal 1.6 Other metals 1.0
the Municipal Urban Waste/Cleaning Company (COMLURB), which Food cans 0.6
is responsible for collecting household solid waste, street cleaning, Inert material 1.1 Stones, concrete 0.9
transferring and final disposal of waste, and composting, among Ceramics 0.2
other services not related to regular waste management system Others 5.2 Rubber 0.3
Yard waste, flowers 1.4
(e.g. vector control). Of the 9666 tons/day of solid waste produced
Wood 0.5
in the city of Rio de Janeiro, 8511 tons/day are collected by COM- Textiles 1.9
LURB of which 44.9% is household waste which represents Shoes, leather 0.3
1,828,754 tons/year. The amount not collected by COMLURB (1155 Coconut shells 0.5
tons/day) comprises the solid waste generated by large quantity Batteries 0.3
Total 100% Total 100%
generators (not including waste from civil construction). These

Please cite this article in press as: Angelo, A.C.M., et al., Life Cycle Assessment and Multi-criteria Decision Analysis: Selection of a strategy for
domestic food waste management in Rio de Janeiro, Journal of Cleaner Production (2016), http://dx.doi.org/10.1016/j.jclepro.2016.12.049
 A.C.M. Angelo et al. / Journal of Cleaner Production xxx (2016) 1e13
environmental problems (Linkov and Moberg, 2012). Huang et al.
(2011) conducted an extensive review of the application of MCDA
methods in environmental science, showing the most important
MCDA approaches used and typifying different areas of application.
Herva and Roca (2013) also conducted a review of MCDA usage but
specific to corporate environmental evaluation. Some studies have
used MCDA methods in sustainability assessment in order to offer a
sound decision analytical framework to deal with the multi-
dimensional nature of sustainability (eg. Basson and Petrie, 2007;
Sadok et al., 2008, 2009; Castellini et al., 2012; Bachmann, 2013;
Wang et al., 2015). Generally speaking, the most commonly used
methods can be classified as those based on utility or value-
function by single synthesising criterion (eg., AHP, MAUT
methods), outranking methods (eg. ELECTRE, PROMETHEE) and
reference point approaches (eg. TOPSIS). More recently, rule-based
approaches have also been emerging in sustainable assessment
cases. For instance, for selecting the location of a municipal solid
waste plant (Abastante et al., 2012), for selection portfolios of
sustainable development projects (Zaras et al., 2012), for assessing
rural sustainable development potentialities (Boggia et al., 2014)
and for handing sustainable development of nanomaterials (Cinelli
et al., 2015). Detailing these methods is beyond the scope of this
paper (see for instance Greco et al., 2016).
The use of MCDA methods as a tool to improve decision analysis
in LCA studies has been methodologically discussed over the past
two decades, either for problem structuring in the LCA framework
or for the aggregation of LCA results (eg. Miettinen and
Hamalainen, 1997; Mazri et al., 2003). As such, MCDA methods
have been combined with LCA for many purposes. Hermann et al.
(2006) put together complementary aspects of LCA, AHP and
environmental indicators to assess the environmental performance
of a pulp industry; Rogers et al. (2008) proposed coupling LCA tools
with Stochastic Multi-Attribute Acceptability Analysis, an MCDA
technique for exploring uncertain preference information (weight
spaces); Linkov and Seager, 2011 proposed the integration of Risk
Assessment, LCA and MCDA for emergency threats; Milani et al.
(2011) used an MCDA approach to deal with decision conflicts
among environmental criteria, cost and properties for the selection
of material composites; Myllyvita et al. (2012) integrated MCDA
and LCA to assess environmental impacts of biomass production
chains; Scott et al. (2016) integrated LCA and MCDA to verify
whether carbon nanomaterials can improve photovoltaic devices.
In the area of waste management, MCDA methods are consid-
ered very effective because the explicit consideration of conflicting
criteria is crucial in this area. There are many examples of the
application of MCDA methods related to waste management that
can be identified in some reviews, namely: Morrissey and Browne,
2004; Allesch and Brunner, 2014; Soltani et al., 2015. According to
Morrissey and Browne (2004), there are three main approach cat-
egories in waste management models: those based on cost-benefit
analysis, i.e., with monetary evaluation; those based on LCA models
focusing on the comparison of waste management strategies
(recycling, incineration and disposal); and those based on the use of
MCDA for choosing between several alternatives of waste treat-
ment, finding the most preferable waste treatment strategy, or
choosing the location of treatment plants. In fact, most of the
studies applying MCDA methods in SWMS are related to solving a
location problem (Soltani et al., 2015).
The ELECTRE III method is the most commonly used method for
waste decisions (Hokkanen and Salminen, 1997; Morrissey and
Browne, 2004; Karagiannidis and Perkoulidis, 2009). However,
the Analytic Hierarchy Process (AHP) is identified as the dominant
method in the area of SWMS and is applied mainly to solve waste
treatment plant location problems (some cases combine MCDA
with a Graphical Information System - GIS). On the other hand,
Please cite this article in press as: Angelo, A.C.M., et al., Life Cycle Assessment and Multi-criteria Decision Analysis: Selection of a strategy for
domestic food waste management in Rio de Janeiro, Journal of Cleaner Production (2016), http://dx.doi.org/10.1016/j.jclepro.2016.12.049
3
ELECTRE III is the most commonly used method in treatment
strategy studies (Soltani et al., 2015).
Some studies use both LCA and MCDA for SWMS decisions. For
instance, Liu et al. (2012) integrated Risk Assessment, LCA and
MCDA to estimate significant aspects in environmental manage-
ment systems and applied this framework in a waste-recycling
factory; Karmperis et al. (2013) reviewed and discussed solid
waste management models dealing with LCA, cost-benefit analysis
and MCDA; Ulukan and Kop (2009) proposed the use of Fuzzy-
TOPSIS method to compare waste collection systems; Hung et al.
(2007) combined LCA outputs, social, economic and technological
indicators with Fuzzy-AHP as a sustainable decision making model
for SWMS; Hanandeh and El-Zein (2009, 2010) used ELECTRE-SS (a
modified stochastic ELECTRE III method) for selection of waste
management strategies; Pires and Chang, 2011 integrated two
MCDA methods (AHP and TOPSIS) for helping the decision makers
to select waste management strategies considering LCA results and
general indicators of social, economic and technical aspects.
Finally, it must be emphasised that the interactivity associated
with many MCDA methods facilitates the search for a good
compromise among a set of alternatives evaluated in several
criteria. In general, MCDA modelling begins with the identification
of the problem under discussion and the definition of the objectives
under analysis. It is then necessary to define the alternatives and
the evaluation criteria to obtain the impact matrix associated with
the problem under study. The interactive elicitation of the decision
maker preferences enables an aggregation of the results looking for
a satisfactory non-dominated solution.
4. Methodology
The major focus of this paper is to combine LCA and an inter-
active MCDA approach for decision aiding in waste management
problems. The study was performed in three steps:
1. Performing a LCA of the SWMS in Rio de Janeiro, including
modelling of the current situation and three possible future
scenarios;
2. Development of a model in order to rank the importance of the
criteria used to evaluate the outputs of a LCA study. This enables
the obtainment of a set of linear constraints on feasible values of
the weights of the criteria used to aggregate the evaluation
results.
3. Use of an interactive learning oriented multi-attribute additive
MCDA model using imprecise information in order to identify
the most preferable alternative amongst the four scenarios. In
this case study it is VIP-Analysis software.
4.1. LCA study methodology
Covering exclusively the end of life phase, the LCA study was
conducted following the international standards ISO 14040 and
14044. All scenarios were modelled with an attributional approach
based on the EASETECH software, developed by DTU, Denmark
which is a commonly used software for LCA of SWMS (Allesch and
Brunner, 2014). The software is based on material flow modelling
and enables generation of input and process specific results. Life-
cycle assessment impact categories and methods recommended by
ILCD (European Commission, 2010) were used. The following
impact categories were considered: Global Warming Potential
(GWP), Ozone Depletion Potential (ODP), Particulate matter
(PM2.5), Ionising radiation, Photochemical Oxidant Formation
(POF), Eutrophication of freshwater (EU-fresh), Marine eutrophi-
cation marine (EU-marine), Acidification, Terrestrial Eutrophication
4 A.C.M. Angelo et al. / Journal of Cleaner Production xxx (2016) 1e13
(EU-land), Ecotoxicity (Ecotox), Depletion of abiotic resources (AR),
Human Toxicity cancer effects (HT-carc), Human Toxicity non-
cancer effects (HT-noncarc). In addition to these impact cate-
gories, the recovery of the macronutrients N, P and K was included
in the assessment. As an attributional approach was used, elec-
tricity provision was modelled on the average Brazilian electricity
mix (based on Ecoinvent 3.0). Emissions of GHG (Green House Gas)
were adjusted to represent official statistics of the GHG-emissions
from the national electricity system over the period 2010e2014,
based on national statistics, resulting in 0.0533t CO2-eq/MWh
(MME/EPE, 2015).
4.1.1. LCA scenarios definition
Four scenarios were developed with special focus on the
handling of domestic food waste fractions. Although the aim of the
study is to identify the most suitable option for the organic fraction
management, sometimes it is necessary to consider the total
household solid waste stream on performing the LCA (Manfredi
and Pant, 2011). Therefore, both the scenarios use 1 ton of mixed
household solid waste. Its composition is shown in Table 1.
Scenario a1 considers the current situation in the city of Rio de
Janeiro with 0% separate collection of organics and final disposal in
the Serope dica Landfill. Scenario a2 considers segregation at source
and separate collection of 20% of the generated household food
waste. The level can be seen as conservative, based on international
experiences (Bernstad, 2012). In scenario a3, the level of segrega-
tion is increased to 50%, a level seen as realistic based on interna-
tional experiences (Bernstad, 2012). In both these scenarios, the
anaerobic digest (biosolid) is used on land as a substitute for
mineral fertilizer.
Finally, scenario a4 considers no segregation of food waste at
source. Instead, the mixed waste is separated in a Material Recovery
Facility (MRF) where separated biomass is treated through AD
(anaerobic digestion). Due to an assumed high level of contami-
nation in generated digestion, this fraction is directed to final
disposal in the Serope dica Landfill. The scenarios and their graphic
illustration are summarised in Table 2 and Fig. 1.
4.1.2. LCA - considering the 4 scenarios
We start with the data used to fix the parameters in order to
model the landfill.
Data on diesel used for collection of residual waste in the city of
Rio de Janeiro was collected from COMLURB (Angelo, 2014),
resulting in 1.3 L diesel/ton wet waste. The modelling of the landfill
was based primarily on information from the viability assessment
produced prior to construction of the Serope dica landfill (Abrelpe,
2012). Data related to energy and material used for construction
and operation of the landfill was based on Da Silva (2011). However,
for various parameters data was not available, and assumptions
were made based on the experience of the authors (Table 3).
In scenarios 2e4, it was assumed to construct an AD plant close
Table 2
Scenarios used in the LCA.
Scenario Description
Scenario a1 Current situation: 0% separate collection of organics. Landfilling
of all MSW.
Scenario a2 20% separate collection of organics for anaerobic
digestion. Landfilling of residual waste. Use of biosolids as
fertilizer on farmland.
Scenario a3 50% separate collection of organics for anaerobic digestion.
Landfilling of residual waste. Use of biosolids as fertilizer
on farmland.
Scenario a4 Anaerobic digestion of organic waste after MRF.
Landfilling of residual waste and of biosolids produced in AD.
Please cite this article in press as: Angelo, A.C.M., et al., Life Cycle Assessment and Multi-criteria Decision Analysis: Selection of a strategy for
domestic food waste management in Rio de Janeiro, Journal of Cleaner Production (2016), http://dx.doi.org/10.1016/j.jclepro.2016.12.049
to the current landfill. Due to the urban characteristics of the re-
gion, as well as an increasingly difficult fresh water situation in the
southeast region of Brazil, dry anaerobic digestion was seen as the
most interesting alternative. The amount of water added in the
digester in dry systems is between two and seven times less than in
wet systems. In addition, large parts of the water added can be
recovered in the dewatering of produced biosolids. Mechanical
pretreatment is needed prior to AD in order to remove unwanted
materials from the fraction prior to entering the digester. In this
case, a simple but robust technique, previously used for pretreat-
ment of food waste prior to dry AD was modelled (VMAB, 2012).
Parameters used in the modelling of the AD system are presented in
Table 4.
In scenarios a2 and a3, produced digestate is assumed to be of
sufficiently high quality (i.e. low concentrations of contamination
such as metals, plastics etc.) to be used as a substitution for mineral
fertilizers on farmland. Such utilisation of digestate from AD-plants,
where source-segregated household waste is treated, is common in
countries such as Sweden, where produced digestate also can be
certified to guarantee its quality (Swedish Waste Management
Organization, 2012). Digestate from the AD-plant would, in this
case study, be transported around 70 km to the Petropolis region,
an area with organic agriculture close to the city of Rio de Janeiro.
Transport vehicles (supposedly lorries, 25t loading capacity) are
assumed to return empty, but this could change with a scheme of
receiving agricultural residue, such as chicken manure produced in
the Petropolis-region, to the AD-plant.
Storage of digestate can result in vast emissions of methane
(Lantz et al., 2009). However, in the present study, storage prior to
transport for land application was assumed to take place in covered
containers with continuous recovery of methane, and thus, emis-
sions of methane during storage are not considered. Emissions due
to the use of different types of fertilizers on farmland have previ-
ously been studied in detail in specific local conditions (Bruun et al.,
2006). However, as little site-specific data was available regarding
the soil composition and other relevant parameters in the area
potentially relevant for application of digestate in the present case,
IPCC (2003) default values were used. In all cases, it was assumed
that digestate could substitute macronutrients (N, P and K) from
mineral fertilizers with a substitution factor of 100%, based on
Starberg, 2005. Nutrient content in generated digestate and pa-
rameters used in modelling of on-land application are presented in
Table 5.
4.1.3. Results of LCA of household solid waste in the city of Rio de
Janeiro
The results of the LCA assessment are presented in Figs. 2 and 3.
These results clearly show that, while one of the investigated sce-
narios is preferable in relation to some of the categories, other
scenarios are preferable in relation to others.
The highest impact on GWP is seen in scenario a4, while an
increasing amount of food waste diverted to AD results in an
increased avoidance of GWP in scenarios a2 and a3. This trend is
seen also in other categories with a strong link to energy provision,
such as ionising radiation, ozone depletion and abiotic resource
depletion. However, in relation to acidification and eutrophication,
the trend is the opposite. Macronutrients (N, P and K) are recovered
in scenarios a2 and a3 only.
4.2. VIP-analysis - the MCDA approach used in this study
4.2.1. VIP-analysis relevance
LCAs are not equipped with interpretation tools that convert
data to decision relevant information in a concise and transparent
manner. Hence, LCA results are rarely conclusive. The consensus-
 A.C.M. Angelo et al. / Journal of Cleaner Production xxx (2016) 1e13 5

Fig. 1. Graphic illustration of the scenarios used in the LCA.

Table 3
Parameters used in modelling of landfill.

Parameter Value Unit Reference

Years of utility 18 Years Abrelpe, 2012

Leachate collection (year 1e30) 90 % Assumed
Leachate collection (year 31e100) 50 % Assumed
Methane oxidation in top cover 10 % Assumed
Density 1 t/m3 Assumed
Height 50 m Calculated, based on Abrelpe, 2012
Diesel use (operation) 0.003 kg/t waste Da Silva, 2011
Diesel use (construction) 3.4E-06 kg/t waste Da Silva, 2011
Concrete (construction) 0.00011 kg/t waste Da Silva, 2011
PEAD (construction) 0.003 kg/t waste Da Silva, 2011
Gas collection (year 1e50) 50 % Abrelpe, 2012
Gas collection (year 51e100) 0 % Abrelpe, 2012
Utilisation of collected LFG e Flare 40 % Abrelpe, 2012
Utilisation of collected LFG e Electricity generation 27 % Abrelpe, 2012
Utilisation of collected LFG e Upgraded and injected to gas grid 33 % Abrelpe, 2012
Use of electricity in upgrading of LGF 0.341 kWh/Nm3 LFG SGC, 2013
Methane losses during upgrading 1 % of CH4 SGC, 2013
Methane content in upgraded LGF 97 % SGC, 2013
Use of electricity in leachate treatment 0.443 kWh/t waste Yoshida et al., 2014
Carbon storage in landfill 0 % of C-bio Assumed

Table 4
Parameters used in the modelling of anaerobic digestion (ww ¼ wet weight).

Process Value Unit Reference

Pretreatment 12.6 kWh/t ww VMAB, 2012

Post-separation of organics 32 kWh/t ww Karagiannidis, 2012
Eletricity use in dry AD 14 kWh/t ww DWMT, 2014
Methane emissions in dry AD 0.5% % of produced CH4 DWMT, 2014
DS in biosolids (prior to dewatering) 35 % DWMT, 2014
Electricity use in dewatering 35 kWh/t ww r, 2012
Balme
DS in biosolids (post dewatering) 60 % r, 2012
Balme
Transport of biosolids 0.007 L diesel/km, t Angelo, 2014
Electric energy recovery 36 % Genrup and Jonshagen, 2011

based ISO methods for conducting Life Cycle Impact Assessment are weighted sums of normalised outputs obtained by direct elicitation
inadequate for identifying salient trade-offs (Prado-Lopez et al., procedures. Nevertheless, compensatory methods (where any
2014; Prado-Lope z, 2015). Most LCA studies are based on weak performance in a given criterion can always be compensated

Please cite this article in press as: Angelo, A.C.M., et al., Life Cycle Assessment and Multi-criteria Decision Analysis: Selection of a strategy for
domestic food waste management in Rio de Janeiro, Journal of Cleaner Production (2016), http://dx.doi.org/10.1016/j.jclepro.2016.12.049
6 A.C.M. Angelo et al. / Journal of Cleaner Production xxx (2016) 1e13

Table 5
Nutrient content in generated digestate and parameters used in modelling of on-land application.

Parameter Value Unit Reference

Digestate and mineral fertilizers on farmland

N2OeN emissions (direct) 1.25 % of N-tot IPCC, 2003
NH3eN emissions 15 % of N-tot OSB/WSTB, 2000
NO3 eN emissions 20 % of N-tot OSB/WSTB, 2000
Incoming N contenta 4.85 kg N-tot/ton ww Based on modelling
Incoming P contenta 0.65 kg P-tot/ton ww Based on modelling
Incoming K contenta 1.48 kg K-tot/ton ww Based on modelling
Application of digestate 20 MJ/t digestate Borgshed et al., 2003
Application of mineral fertilizers 0.36 MJ/kg N-tot Finnveden et al., 2000
a
In food waste only, as kg per ton mixed incoming waste.

Fig. 2. Results of the LCA, as contribution to non-toxicological and non-resource impact categories. In all cases, impacts are related to treatment of 1 ton of mixed solid waste from
the city of Rio de Janeiro as displayed in Table 1. Units were adjusted in order to present results in the same diagram, details can be found in Table A1 - Appendix A.

by good performances in other criteria) have shown to be very
disputable, especially in environmental problems.
In order to overcome these shortcomings and to improve the
understanding of LCA studies by decision or policy makers, MCDA
methods have been applied to aggregate LCA results, and thus to
improve the decision analysis. Some applications of MCDA and LCA
integration have been detailed before in Section 3. Yatsalo et al.
(2007) and Linkov and Moberg (2012) argue that it may be diffi-
cult to choose the most adequate MCDA method, and so they have
carried out tests with different methods. Yatsalo et al., for example,
applied three different MCDA methods (a MAUT approach, an
outranking approach and the AHP method), coupled with LCA, in
two case studies involving contaminated sediment management,
and even with unique theoretical background and calculation al-
gorithms of each MCDA method, similar ranking orders for alter-
natives were pointed out. In these circumstances, they argue that
the choice of the MCDA approach may be not very relevant in many
cases. On the other hand, Prado-Lope z points out two key factors
when selecting a MCDA method to aggregate LCA results. Firstly,
the method should avoid full compensation. Secondly, the assess-
ment should be context dependent.
Until recently, LCA practitioners could basically choose between
three major MCDA approaches which promise to overcome prac-
tical hindrances such as the high uncertainty associated with
weights or scaling constants, the incommensurability of indicators
units, and the subjectivity behind the elicitation of multi-criteria
analysis parameters.
(a) Multi-Attribute Utility Theory (MAUT) allows a more scien-
tific weights elicitation, but decision makers remain depen-
dent on accurate information about them and so the true
causes of the problem are not addressed (see, for instance:
Benoit and Rousseaux, 2003; Huppes et al., 2012; Reap et al.,
2008). MAUT's ease of use may be very handy for sustain-
ability decision makers, but they seldom hold the accurate
values for the scale constants or for the weights MAUT
demands.
(b) Considering the limitations of the additive model, some au-
thors proposed avoiding total compensation approaches by
using outranking based models such as the ELECTRE and
PROMETHEE families (Benoit and Rousseaux, 2003; Munda,
2008). Nevertheless, these approaches require fixing a
certain number of parameters that reflect the decision maker
preferences, and changes in these parameters can drastically
influence the results. That is why efforts have been made on
methods that emphasise the construction of models with

partial information (Dias and Clímaco, 2000; Jacquet-Lagreze

Please cite this article in press as: Angelo, A.C.M., et al., Life Cycle Assessment and Multi-criteria Decision Analysis: Selection of a strategy for
domestic food waste management in Rio de Janeiro, Journal of Cleaner Production (2016), http://dx.doi.org/10.1016/j.jclepro.2016.12.049
 A.C.M. Angelo et al. / Journal of Cleaner Production xxx (2016) 1e13 7

Fig. 3. Results of the LCA, as contribution to toxicological and resource impact categories. In all cases, impacts are related to treatment of 1 ton of mixed solid waste from the city of
Rio de Janeiro as displayed in Table 1. Units were adjusted in order to present results in the same diagram, details can be found in Table A1 - Appendix A.

and Siskos, 2001), helping the decision maker in establishing
criteria weights or even parameters of analysis.
(c) The uncertainty associated with the quantification of the
weights/scaling constants may be dealt with by using an
alternative approach that consists of the combination of
Stochastic Multiattribute Analysis (SMAA) method with LCA
(Rogers et al., 2008; Prado-Lope z, 2015). Decision makers/
stakeholders use to feel more comfortable with deterministic
models, but it is plausible (Domingues et al., 2015) that they
would prefer to express uncertainty by linear constraints
regarding the weights rather than fixing point values, and
not depending on any assumption on stochastic distribution
of the weights nor on simulation techniques to estimate the
corresponding results distribution.
The interactive choice support system VIP-Analysis can meet
such requirements. Although based on the additive model, it offers
a good compromise between the need for an easily understandable
technique (so allowing genuine participation from stakeholders)
and the mitigation of the drawbacks associated with full compen-
sation. It is well known that the most difficult step in the decision
aiding process is setting values of the weights/scaling constants and
the uncertainty/imprecision in weight has to be tackled. Therefore,
instead of looking for a sharp value for the weights, i.e., leading to a
unique score for each scenario/alternative, VIP-Analysis simply
requires the introduction of context dependent, well-accepted
linear constraints on the weights. Furthermore, this tool seems
very adequate to support a group of stakeholders meeting face-to-
face around a computer (Phillips and Phillips, 1993), which is also
very adequate in the LCA context. However, a facilitator is needed as
proficient use of VIP-Analysis requires the sharing of knowledge
about the tool potentialities and limitations with stakeholders and
decision makers.
developed by Dias and Clímaco (2000). Several recent studies have
adopted it (see, for instance, Ventura et al., 2010, 2014; Yue et al.,
2009). The principal characteristic of this tool is that no precise
values for the scaling constants/weights are required. Instead, it can
accept imprecise information (i.e. ordinal ranking of the weights,
intervals and/or linear constraints) on these values, usually iden-
tified in indirect ways, for example by comparing swings, by
ordering scaling constants, etc. The major objectives are the iden-
tification of robust conclusions (holding for every feasible combi-
nation of the scaling constants), and secondly, identifying what the
variability of the results is due to the imprecision in the parameter
values, trying to identify the most satisfactory alternative rather
than a preference ranking of all alternatives. Of course, the
robustness analysis will depend on the degree of restriction on the
weights variability introduced by linear constraints, and these
depend on the context and on the decision makers’ points of view.
VIP-Analysis considers an additive value function under
imprecise information as described below:
X
n
Vðai ; kÞ ¼ kj vj ðai Þ: (1)
j¼1
where ai and vj represent the ith alternative and the jth normalised
global performance measure, kj is the scaling constant of vj. and k
represents the vector of scaling constants, k¼(k1, k2, …, kn).
Considering T as the set of acceptable values of the vector k of
scaling constants, the regret associated with alternative aj, when
compared with ai, here denoted as regij, is the maximal difference:
  
regij ¼ max Vðai ; kÞ  V aj ; k : (2)
k2T

4.2.2. The method description If regij is negative or null then Vðaj ; kÞ  Vðai ; kÞck2T and at
VIP-Analysis is an interactive software package dedicated to least one inequality is strict, aj dominates ai, (or ai is dominated by
choice problems regarding the evaluation of a discrete set of al- aj). A relaxation to this dominance relation, by a tolerance param-
ternatives according to a multi-attribute additive value function, eter ε means that:

Please cite this article in press as: Angelo, A.C.M., et al., Life Cycle Assessment and Multi-criteria Decision Analysis: Selection of a strategy for
domestic food waste management in Rio de Janeiro, Journal of Cleaner Production (2016), http://dx.doi.org/10.1016/j.jclepro.2016.12.049
8 A.C.M. Angelo et al. / Journal of Cleaner Production xxx (2016) 1e13
 
V aj ; k  Vðai ; kÞ  εcK2T:
In this case, aj quasi dominates ai with tolerance ε.
Furthermore, another stricter concept is the concept of absolute
dominance: aj dominates ai absolutely when
 
V aj ; k  Vðai ; k0 Þck; k0 2T
where at least one inequality is strict.
For every alternative ai the maximum regret associated with it,
when it is compared with all other alternatives which may have a
higher additive value for given T, is:
n o  
  
regmax ðai Þ ¼ max regji ¼ max max V aj ; k  Vðai ; kÞ :
jsi k2T jsi
In a first phase, some tools enable filtering of the alternatives.
Firstly, VIP can use maximum regret as a filter. For instance, if the
maximum regret of an alternative is negative or null, then ai is
optimal. Secondly, VIP module calculates the range of values of each
alternative:

minfVðai ; kÞg; maxfVðai ; kÞg :
k2T k2T
This enables, for example, the elimination of those with a
minimum value below a certain threshold, fixed by the decision
makers, while the knowledge of the maximum regret concerning
each alternative enables the elimination of those with a max regret
beyond a threshold fixed by the decision makers and the domi-
nated solutions. The alternatives passing the filtering phase are
analysed using the matrix of regrets, designated as pairwise
confrontation table, together with the relaxation of the concepts of
optimality and dominance, exploring the concepts of quasi-
optimality and the concept of quasi-dominance, enabling the de-
cision makers to identify robust conclusions in order to help in the
search for the best alternative. All the details about VIP-Analysis
methodology and the corresponding package can be read in Dias
and Clímaco (2000). They can also be found in the following link:
http://hdl.handle.net/10316/31247, where the reader can also find a
complete example describing the use of VIP-Analysis package.
4.3. Ordering scaling constants/weights
As mentioned in the last section, VIP-Analysis does not require
precise values for scaling constants/weights. It just uses imprecise
information about them, requiring just the introduction of linear
constraints on the weights, and of course, the decision makers only
introduce constraints on scaling constants/weights they feel
confident to do. In this case study, the decision makers have
decided only to consider linear constraints on the weights associ-
ated with the ordinal ranking of the 16 criteria, i.e., impact cate-
gories within the context of waste management in the city of Rio de
Janeiro.
The procedure used in this paper to order the scaling constants/
weights is inspired in the weighing procedure used in the LCA
method EDIP (Stranddorf et al., 2005). In fact, it is a simplification of
it, taking into account the availability of the data for Rio de Janeiro
in many areas is clearly insufficient. Arguably, ordering the criteria
is much less exigent than identifying the weights precisely and so
the results obtained using the procedure here outlined are surely
acceptable. Furthermore, as we will see in Section 5, the determi-
nation of any ordinal ranking of the criteria involve subjectivities,
other ordinal rankings plausible for experts/decision makers were
Please cite this article in press as: Angelo, A.C.M., et al., Life Cycle Assessment and Multi-criteria Decision Analysis: Selection of a strategy for
domestic food waste management in Rio de Janeiro, Journal of Cleaner Production (2016), http://dx.doi.org/10.1016/j.jclepro.2016.12.049
Table 6
(3) Impact Categories ranking.
Ordering Impact categories
1 Global warming
2 Ozone depletion potential
3 Human toxicological (non-carcinogenic)
Photochemical ozone formation
(4) N-recovery
P-recovery
K-recovery
4 Ecotoxicity
Abiotic resources
5 Human toxicological (carcinogenic)
Eutrophication (freshwater)
Eutrophication (marine)
6 Acidification
7 Particulate matter
Ionising radiation
(5) Eutrophication (land)
tested.
In summary, the following procedure was applied to create an
ordering of the scaling constants/weights:
1. Assessment of the current status in relation to each impact
(6)
category through official documents on local (city of Rio de
Janeiro), region (Rio de Janeiro state) or national level. Each
impact category/criterion is classified as “low”, “medium” or
“high” in relation to the current status. A value is attributed to
each level being: 1 ¼ low, 2 ¼ medium and 3 ¼ high.
2. Identification of official targets, strategies, policies, international
declarations and laws on the national or regional/local level
regarding the improvement of the current status. Each impact
category is classified as “low”, “medium” or “high” in relation to
international, national and regional/local efforts to improve the
impacts. A value is attributed to each level being: 1 ¼ low,
2 ¼ medium and 3 ¼ high.
3. A sum of the above obtained scores enables an ordinal ranking
of the impact categories/criteria to be obtained with reasonable
confidence, and so, an ordering of the corresponding weights/
scaling constants.
Both data used in the creation of the ordinal ranking and results
from the proposed method are presented in Table A2, in Appendix
A. Table 6 shows the final ranking of the criteria.
It is important to note that, in this case, the subjectivity is
introduced by the decision makers when they use the classification
of “low, medium or high” on the current status of the impacts and
the international, national and local/regional improvement efforts
regarding each environmental impact category/criterion.
Finally, it must be remarked that in the literature, there are
several procedures to define weights that can be done by stake-
holder consultation or by eliciting decision makers’ preferences.
Arguably, public participation in the weighting step has an
important role in MCDA. However, it may make the weighting
process more difficult due to the time required and transparency
involved, since it requires the definition of which actors will be
consulted and if they have the same importance in the decision
making.
5. VIP-analysis case study results
The evaluated alternatives are the four scenarios developed in
the LCA study (a1 to a4). The criteria are the sixteen selected impact
categories considered in the LCA study. In order to enable the use of
an additive aggregation, outcome values of the LCA criteria/impact
 A.C.M. Angelo et al. / Journal of Cleaner Production xxx (2016) 1e13 9

Table 7 Note that Zik denotes the value of alternative I for criterion k, and
Criteria normalisation. then, after normalisation, all criteria are transformed to maximise
Equation Range extreme Criteria objective or minimise evaluation. Three out of sixteen are maximising
Zik  Zmin
k
Maximum ¼ 1 Maximise
criteria, namely the 14, 15 and 16. The others are minimising criteria
ðZik ÞN ¼ k
Zmax  Zmin
k
Minimum ¼ 0 and, in both cases, the ideal value corresponds to number 1
ðZik ÞN ¼
k
Zmax  Zik Maximum ¼ 0 Minimise (Table A3, in Appendix A).
k
Zmax  Zmin
k
Minimum ¼ 1 We begin using the order obtained in Section 4.3. The VIP-
Analysis case study results have shown that three scenarios are
absolutely dominated (Table 8). The same conclusion can be ob-
Table 8 tained from Fig. 4, where we can see that the worst case for sce-
Range of results.
nario a3 is better than the best values for the other scenarios; this
Alternative Min value Max value Max regret Dominated? configures absolute dominance, a sort of relation that is stronger
Scenario a1 0.213 0.414 0.744 YES (Abs) than the dominance introduced in Section 4.2. This means that
Scenario a2 0.566 0.635 0.373 YES (Abs) scenarios a1, a2 and a4 can indubitably be discarded. The same
Scenario a3 0.656 1 0.09 result can also be obtained from the maximum regret of scenario
Scenario a4 0 0.251 1 YES (Abs) a3, which is negative (Fig. 5) and therefore implies that a3 is
optimal. In these circumstances, it can be concluded that scenario
a3 should be chosen even without analysing the pairwise
categories were normalised using a differences ratio normalisation. confrontation table in detail (Table 9). If the decision makers have
An example of this type of normalisation is that used in the reasons to believe that some qualitative issues not included in the
calculation of the Better Life Index (OECD, 2011). It is carried out model should also be considered, scenario a2 should be analysed in
according to the type of criterion (Z), i.e. whether it is a minimising more detail because the results indicate that it is preferable to a1
or maximising criteria (Table 7). and a4 (see Figs. 4 and 5).

Fig. 4. Range of results.

Fig. 5. The Maximum Regret of alternatives.

Please cite this article in press as: Angelo, A.C.M., et al., Life Cycle Assessment and Multi-criteria Decision Analysis: Selection of a strategy for
domestic food waste management in Rio de Janeiro, Journal of Cleaner Production (2016), http://dx.doi.org/10.1016/j.jclepro.2016.12.049
10 A.C.M. Angelo et al. / Journal of Cleaner Production xxx (2016) 1e13

Table 9
Pairwise confrontation table.
Scenario 1 Scenario 2 Scenario 3 Scenario 4
Scenario a1 - 0.152 0.242 0.367
Scenario a2 0.388 - 0.09 0.632
Scenario a3 0.744 0.373 - 1
Scenario a4 0.025 0.315 0.404 - ation in the dominance relation, by a tolerance ε, trying to identify
Max Regret 0.744 0.373 0.09 1
Results can be very sensitive to the ordering of scaling constants.
Any ordinal ranking of the criteria is subjective. Under these cir-
cumstances, six other plausible criteria rankings were tested ac-
cording to the priorities of the decision makers. In five of them the
principal conclusion of the previous case was confirmed as they
point out a3 as the optimal alternative. Nevertheless, when the
ordinal ranking is based on EPA Science Advisory Board (Lippiatt,
2007), the VIP-Analysis tools lead to contrasting results deserving
further analysis as there is a close dispute between a2 and a3 (the
table concerning the ordinal rankings tested is in Appendix B). This
can be seen by analysing the max regrets and the ranges of the
results corresponding to this ordinal ranking (Figs. 6 and 7).
Alternatives a1 and a4 are the two worst in both cases. This
seems to justify their elimination before further inspection,
whenever it may add the subsequent decision process phases. This
is not the case, however, as we are just dealing with 4 alternatives.
A preliminary reduction of alternatives/scenarios is not a key issue
in this case. In the second phase, the decision maker is confronted
with the corresponding pairwise confrontation table (in Fig. 8).
Again, a visual inspection is enough to conclude that dominated
alternatives do not exist in this case. Should they exist, they would
correspond to negative cells marked in red, namely to cells where
the row alternative is dominated by the column alternative. From
the pairwise confrontation table, it is possible to draw some robust
conclusions, such as “scenario a2 never loses to a3 by a difference
greater than 0.373” and “scenario a3 never loses to a2 by a differ-
ence greater than 0.173”. Furthermore, from the last row of the
pairwise confrontation table, it is possible to confirm that scenario
a3 is the one with min max regret (0.182). Of course, the other max
regrets corresponding to the other scenarios are also available in
the same row. Fig. 9 below summarises the numeric values of the
results:
Utilising an interactive tool (an elevator) provided by the VIP-
Analysis software, it is possible to test the introduction of a relax-
quasi-dominated alternatives: a2 quasi-dominates a1 for an ε
relaxation greater than 0.009 and a4 also quasi-dominates a1 for an
ε relaxation greater than 0.042.
Figs. 10 and 11 show considered tolerances of 0.01 and 0.05,
respectively. A cell is red when the row alternative is quasi-
dominated by the column alternative.
In summary, from the range of the results, the max regrets and
the identification of quasi-dominated alternatives for very small
tolerances, it is possible to conclude that scenarios a1 and a4 can be
discarded. From the pairwise confrontation table and from the
summary of the results, it is possible to conclude that a2 and a3 are
neither dominated nor quasi-dominated for very small tolerances
with a2 being better than a3 considering the max min (Fig. 7) and
the opposite when considering the min max regret (Fig. 6). It is still
possible to get some robust conclusions from Fig. 11 regarding the
pairwise comparison of a2 and a3. The alternative a3 is preferable if
the decision makers privilege the min max regret, i.e. a measure of
equity, but a2 is preferable if they favour the best worst case and
the variability of the results (the minimum and the maximum value
for each alternative/scenario).
It must be remarked that the robustness analysis, in the way it is
used in VIP-Analysis, can substitute sensitivity procedures.
Whereas in sensitivity analysis small variations of the parameters
around a previously calculated proposed solution are tested, in
robustness analysis a set of acceptable weights, defined by linear
constraints that the decision makers feel enough confidence to
introduce, is considered. The idea is to identify the robust conclu-
sions, i.e. those compatible with the incomplete information on the
weights provided by the decision makers. Note that sensitivity
analysis is often conduced varying one parameter at a time,

Fig. 6. VIP-Analysis screen showing the maximum regret of alternatives.

Please cite this article in press as: Angelo, A.C.M., et al., Life Cycle Assessment and Multi-criteria Decision Analysis: Selection of a strategy for
domestic food waste management in Rio de Janeiro, Journal of Cleaner Production (2016), http://dx.doi.org/10.1016/j.jclepro.2016.12.049
 A.C.M. Angelo et al. / Journal of Cleaner Production xxx (2016) 1e13 11

Fig. 7. VIP-Analysis screen showing the range of the results.

Fig. 8. VIP-Analysis screen showing the pairwise confrontation table.

Fig. 10. VIP-Analysis screen showing the pairwise confrontation table with 0.01 of
tolerance.

Fig. 9. VIP-Analysis screen showing the summary of results.

ignoring the interaction effects that could appear when more than
one parameter is changed at the same time.

Fig. 11. VIP-Analysis screen showing the pairwise confrontation table with 0.05 of
6. Conclusions tolerance.

This study assesses the potentialities of combining an MCDA

categories, some analysts avoid the optional steps determined by
approach with LCA towards decision support. In many cases results
the ISO Standards (2006) of weighting and aggregating the results.
from LCA of solid waste management are difficult to interpret
Without an aggregation of LCA outputs, however, the interpretation
because there is no completely satisfying alternative, since while
phase of LCA becomes much more difficult for decision making. The
one alternative might be preferable in relation to some aspects and
use of an appropriate MCDA approach not only provides clarity to
impact categories, other alternatives might be preferable in relation
the interpretation of LCA results, but also helps elucidating the
to others. Besides this, as LCA studies work with many impact

Please cite this article in press as: Angelo, A.C.M., et al., Life Cycle Assessment and Multi-criteria Decision Analysis: Selection of a strategy for
domestic food waste management in Rio de Janeiro, Journal of Cleaner Production (2016), http://dx.doi.org/10.1016/j.jclepro.2016.12.049
12 A.C.M. Angelo et al. / Journal of Cleaner Production xxx (2016) 1e13
acceptability of alternatives.
Certainly, every MCDA procedure requires the definition of
some parameters that reflect the decision makers' preferences
which brings some subjectivity to the decision process, and the
main challenge is how to minimise this subjectivity. The main hy-
pothesis of this study was that using MCDA approaches that miti-
gate the drawbacks of weighting procedures increases the potential
for using LCA as a decision support tool in the sense of helping in
analysing results to choose a satisfactory alternative. Looking for a
straightforward MCDA method that accepts imprecise information,
this study has chosen VIP-Analysis, an interactive choice support
system that doesn't demand the values of the criteria weights, but
just their order.
Although the described waste management decision required a
complex comparison of scenarios and criteria, using VIP-Analysis to
aggregate the outputs of the LCA not only allowed the selection of
the preferable environmental option to treat the waste, but also the
assessment of different and conflicting criteria such as social and
economic aspects. Outcomes of the LCA showed that while the
ranking of investigated scenarios were the same in relation to GWP,
ozone depletion, human toxicity and eco-toxicity, the opposite
trend was obtained in relation to acidification and eutrophication.
By simply prioritising the criteria importance, it was possible to
identify an alternative preferable in relation to the assessed impact
categories. VIP-Analysis results have shown scenario a3 as the
optimal alternative for treating domestic food waste in the city of
Rio de Janeiro.
As the results can be very sensitive to the ordering of scale
constants, other criteria importance rankings were tested (see
Appendix B). Most of them confirmed the previous results in which
scenario a3 is optimal and scenarios a1 and a4 are bad, showing
scenario a4 as the worst alternative. It was expected that scenario
a3 would appear as the best alternative because it considers more
quantity of organic waste treated by AD, i.e., more quantity of
organic waste diverted from landfill. In addition, this alternative
delivers a potential positive environmental impact due to the
nutrient recovery. On the contrary, scenario a4 appeared as the
worst option for the case study, having the highest impact on
several categories including GWP and ODP. Although this alterna-
tive also uses AD for organic waste treatment, it considers no
segregation of food waste at source as mixed waste is separated in
an MRF, resulting in a contamination of digestate. Thus, in this
scenario, digestate must be directed to final disposal in landfill
instead of being used as fertilizer.
Only one of the criteria rankings tested, namely the one based
on EPA Science Advisory Board (Lippiatt, 2007), gave different
conclusions. Scenarios a1 and a4 are still worse than the other ones,
but it is not possible to make a definitive choice between scenarios
a2 and a3. This can be explained by the fact that these two scenarios
only differ by the percentage of organic waste treated by AD. Thus,
the conclusion is not in contradiction with previous results. As with
many other MCDA approaches, VIP-Analysis does not always lead to
a final result, but to a drastic reduction of the acceptable
alternatives.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.jclepro.2016.12.049.
References
Abastante, F., Bottero, M., Greco, S., Lami, I.M., 2012. A dominance-based rough set
approach model for selecting the location for a municipal solid waste plant.
GEAM. Geoing. Ambient. E Mineraria 137, 43e54.
Please cite this article in press as: Angelo, A.C.M., et al., Life Cycle Assessment and Multi-criteria Decision Analysis: Selection of a strategy for
domestic food waste management in Rio de Janeiro, Journal of Cleaner Production (2016), http://dx.doi.org/10.1016/j.jclepro.2016.12.049
Abrelpe, 2012. Atlas of GHG Emission and Energy Potential by Waste Destination in
Brazil, Case Study 2: CTR Santa Rosa - Rio de Janeiro state. Brazilian Association
~o Paulo, SP, Brazil. http://
of Public Cleaning and Special Waste Companies. Sa
www.abrelpe.org.br/atlas/index.cfm. accessed 07 05 15.
Allesch, A., Brunner, P.H., 2014. Assessment methods for solid waste management: a
literature review. Waste Manag. Res. 32, 461e473.
Angelo, A.C.M., 2014. Contribuiço ~es para o Inventario do Ciclo de Vida dos resíduos
orga ^nicos provenientes da coleta domiciliar na cidade do Rio de Janeiro. Master
thesis. Alberto Luiz Coimbra Institute Graduate School and Research in Engi-
neering, Federal University of Rio de Janeiro, Brazil.
Bachmann, T.M., 2013. Towards life cycle sustainability assessment: drawing on the
NEEDS project ’ s total cost and multi-criteria decision analysis ranking
methods. Int. J. Life Cycle Assess. 18, 1698e1709.
Balme r, P., 2012. Strategier Fo €r Slamanva €ndning. VA-strategi AB. Presentation, VA-
m€ assan. September 2012.
Basson, L., Petrie, J., 2007. A critical systems approach to decision support for pro-
cess engineering. Comput. Chem. Eng. 31, 876e888.
Benoit, V., Rousseaux, P., 2003. Aid for aggregating the impacts in life cycle
assessment. Int. J. Life Cycle Assess. 8, 74e82.
Bernstad, S.S.A., 2012. Household Food Waste Management e Evaluations of Cur-
rent Status and Potential Improvements Using Life-cycle Assessment Method-
ology. PhD-thesis. LTH, Lund University, Sweden.
Bernstad, S.S.A., Andersson, T., 2015. Food waste minimization from a life-cycle
perspective. J. Environ. Manag. 147, 219e226.
Boggia, A., Rocchi, L., Paolotti, L., Musotti, F., Greco, S., 2014. Assessing rural sus-
tainable development potentialities using a dominance-based rough set
approach. J. Environ. Manag. 144, 160e167.
Borgshed, J., Leander, J., Ro € nnquist, E.M., Steinwall, P., 2003. Systems Analysis of
Household Waste Management in the Kalmar Region. Carl Bro Energikonsult
AB.
Brasil, 2010. Law 12,305 of 2 August 2010-Institutes the National Policy on Solid
Waste, Alters Law No. 9,605 of 12 February 1998 and Makes Other Provisions.
Ministry of the Environment of Brazil, Brasília, DF, Brazil. http://www.mma.gov.
br. accessed 12 01 15.
Bruun, S., Lund Hansen, T., Christensen, T.H., Magid, J., Jensen, L.S., 2006. Application
of processed organic municipal solid waste on agricultural land e a scenario
analysis. Environ. Model. Assess. 11 (3), 251e265.
Castellini, C., et al., 2012. A multicriteria approach for measuring the sustainability
of different poultry production systems. J. Clean. Prod. 37, 192e201.
Cinelli, M., Coles, S.R., Nadagouda, M.N., Błaszczyn  ski, J., Słowin ski, R., Varma, R.S.,
Kirwan, K., 2015. A green chemistry-based classification model for the synthesis
of silver nanoparticles. Green Chem. 17, 2825e2839.
Clímaco, J., Valle, R., 2016. MCDA and LCSA - a note on the aggregation of prefer-
ences. In: Kunifuji, S. (Ed.), Knowledge, Information and Creativity Support
Systems pp 105e116, Advances in Intelligent Systems and Computing 416.
Springer, Switzerland, 2016.
COMLURB, 2013a. Gravimetric and Microbiological Characterization of Household
Solid Waste in Rio de Janeiro City. Municipal Urban Waste/Cleaning Company,
Municipal Secretariat of Environment, Rio de Janeiro, RJ, Brazil.
COMLURB, 2013b. Assessment of Caju Organic Compost Quality. Municipal Urban
Waste/Cleaning Company, Municipal Secretariat of Environment, Rio de Janeiro,
RJ, Brazil.
Da Silva, G.F., 2011. Aterro Sanit ~o Jo~
ario Sa ao: Estudo Dos Indicadores Ambientais
em Emergia. Master Thesis. Programa de Po s-Graduaça ~o (Mestrado e Doutor-
ado) em Engenharia de Produça ~o. The University of S~ ao Paulo, Brazil.
Dias, L., Clímaco, J., 2000. Additive aggregation with interdependent parameters:
the VIP-analysis software. J. Oper. Res. Soc. 51, 1070e1082.
Domingues, A.R., Marques, P., Garcia, R., Freire, F., Dias, L.C., 2015. Applying multi-
criteria decision analysis to the life-cycle assessment of vehicles. J. Clean.
Prod. 107, 749e759.
DWMT, 2014. Database of Waste Management Technologies AD 2: Dry AD Followed
by Covered Windrows Composting. http://www.epem.gr/waste-c-control/
database/html/AD-01.htm.
European Commission, 2005. European Waste Framework Directive. http://eur-lex.
europa.eu/legal-content/EN/TXT/?uri¼CELEX:32008L0098. accessed 08 06 15.
European Commission, 2008. Thematic Strategy on the Sustainable Use of Natural
Resources. http://eur-lex.europa.eu/legal-content/EN/TXT/?uri¼CELEX:
52005DC0670. accessed 08 06 15.
European Commission, 2010. International Reference Life Cycle Data System (ILCD)
Handbook e General Guide for Life Cycle Assessment - Detailed Guidance, first
ed. Publications Office of the European Union, Luxembourg, LU. March 2010.
EUR 24708 EN.
Finnveden, G., Johansson, J., Lind, P., Moberg, Å., 2000. Life Cycle Assessments of
Energy from Solid Waste. FMS Report 137.
Genrup, M., Jonshagen, K., 2011. Integration of Bio-fired Gas Turbines in Combined
Heat and Power Generation. Va €rmeforsk Report 1164. Stockholm, Sweden.
Greco, S., Ehrgott, M., Figueira, J. (Eds.), 2016. Multiple Criteria Decision Analysis:
State of Art Surveys, International Series in Operations Research & Management
Science. Springer, New York, NY.
Hanandeh, A. El, El-zein, A., 2009. Strategies for the municipal waste management
system to take advantage of carbon trading under competing policies : the role
of energy from waste in Sydney. Waste Manag. 29 (7), 2188e2194.
Hanandeh, A.E., El-Zein, A., 2010. The development and application of multi-criteria
decision-making tool with consideration of uncertainty: the selection of a
management strategy for the bio-degradable fraction in the municipal solid
 A.C.M. Angelo et al. / Journal of Cleaner Production xxx (2016) 1e13
waste. Bioresour. Technol. 101, 555e561.
Hermann, B.G., Kroeze, C., Jawjit, W., 2006. Assessing environmental performance
by combining life cycle assessment, multi-criteria analysis and environmental
performance indicators. J. Clean. Prod. 15, 1787e1796.
Herva, M., Roca, E., 2013. Review of combined approaches and multi-criteria anal-
ysis for corporate environmental evaluation. J. Clean. Prod. 39, 355e371.
Hokkanen, J., Salminen, P., 1997. Choosing a solid waste management system using
multicriteria decision analysis. Eur. J. Oper. Res. 98, 19e36.
Huang, I., Keisler, J., Linkov, I., 2011. Multi-criteria decision analysis in environ-
mental sciences: ten years of applications and trends. Sci. Total Environ. 409,
3578e3594.
Hung, M., Ma, H., Yang, W., 2007. A novel sustainable decision making model for
municipal solid waste management. Waste Manag. 27, 209e219.
Huppes, G., van Oers, L., Pretato, U., Pennington, D.W., 2012. Weighting environ-
mental effects: Analytic survey with operational evaluation methods and a
meta-method. Int. J. Life Cycle Assess. 17, 876e891.
IBGE, 2010. Population Census. Brazilian Institute of Geography and Statistic. Rio de
Janeiro, RJ, Brazil. http://www.ibge.gov.br. accessed 12 12 2014.
IPCC, 2006, 2003 (Chapter 11), Agriculture, Forestry and Other Land Use. Borgshed,
€ nnquist, E.-M. and Steinwall, P.. N2O Emissions from Managed
J., Leander J., Ro
Soil as and CO2 Emissions from Lime and Urea Application. Guidelines for
National Greenhouse Gas Inventories, vol. 4. Systems analysis of household
waste management in the Kalmar region (in Swedish). Carl Bro Energikonsult
AB.
IPEA, 2014. PNRS Observatory: Solid Waste Final Destination. Institute for Applied
Economic Research, Brasília, DF, Brazil. http://observatoriopnrs.org/destinacao-
final/. accessed 07 04 15.
ISO, 2006. ISO 14044 International standard. In: Environmental Management e Life
Cycle Assessment e Requirements and Guidelines. International Organization
for Standardization, Geneva, CH.
Jacquet-Lagre ze, E., Siskos, Y., 2001. Preference disaggregation : 20 years of MCDA
experience. Eur. J. Oper. Res. 130, 233e245.
Karagiannidis, A., 2012. Waste to Energy. Opportunities and Challenges for Devel-
oping and Transit Economies. Springer Editor, London.
Karagiannidis, A., Perkoulidis, G., 2009. A multi-criteria ranking of different tech-
nologies for the anaerobic digestion for energy recovery of the organic fraction
of municipal solid wastes. Bioresour. Technol. 100, 2355e2360.
Karmperis, A.C., Aravossis, K., Tatsiopoulos, I.P., Sotirchos, A., 2013. Decision support
models for solid waste management: review and game-theoretic approaches.
Waste Manag. 33, 1290e1301.
Lantz, M., Ekman, A., Bo €rjesson, P., 2009. Systems Optimized Production of Vehicle
GaseAn environmental and energy Assessment of the So €deråsen Biogas Pro-
duction Plant (In Swedish). Report 69. Environmental and Energy Systems
Studies. Lund University.
Laurent, A., Bakas, I., Clavreul, J., Bernstad, A., Niero, M., Gentil, E., Hauschild, M.Z.,
Christensen, T.H., 2014. Review of LCA studies of solid waste management
systems e Part I: lessons learned and perspectives. Waste Manag. 34, 573e588.
Linkov, I., Moberg, E., 2012. Multi-criteria Decision Analysis: Environmental Ap-
plications and Case Studies. CRC Press.
Linkov, I., Seager, T.P., 2011. Coupling multi-criteria decision analysis, life-cycle
assessment, and risk assessment for emerging threats. Environ. Sci. Technol.
45, 5068e5074.
Lippiatt, B.C., 2007. BEES 4.0: Building for Environmental and Economic Sustain-
ability Technical Manual and User Guide (NISTIR 7423). National Institute of
Standards and Technology, Gaithersburg, MD. http://fire.nist.gov/bfrlpubs/
build07/PDF/b07018.pdf.
Liu, K.F.-R., Ko, C.-Y., Fan, C., Chen, C.-W., 2012. Combining risk assessment, life cycle
assessment, and multi-criteria decision analysis to estimate environmental
aspects in environmental management system. Int. J. Life Cycle Assess. 17,
845e862.
Manfredi, S., Pant, R., 2011. Supporting environmentally Sound Decisions for Bio-
waste Management a Practical Guide to Life Cycle Thinking (LCT) and Life Cy-
cle Assessment (LCA). European Commission, Joint Research Centre, Institute
for Environment and Sustainability, Luxembourg. Publications Office.
Matteson, S., 2014. Methods for multi-criteria sustainability and reliability assess-
ments of power systems. Energy 71, 130e136.
Mazri, C., Ventura, A., Jullien, A., Bouyssou, D., 2003. Life cycle analysis and decision
aiding: an example for roads evaluation. In: Proceedings of the 4th Interna-
tional Conference on Decision Making in Urban and Civil Engineering, pp. 1e11.
Miettinen, P., Hamalainen, R.P., 1997. How to benefit from decision analysis in
environmental life cycle assessment (LCA). Eur. J. Oper. Res. 102, 279e294.
Milani, A.S., Eskicioglu, Robles, K., Bujun, K., Hosseini-Nasab, H., 2011. Multiple
criteria decision making with life cycle assessment for material selection of
composites. eXPRESS Polym. Lett. 5 (12), 1062e1074.
MME/EPE, 2015. 2015 Statistical Yearbook of Electricity 2014 Baseline Year. http://
www.epe.gov.br.
Morrissey, A.J., Browne, J., 2004. Waste management models and their application
to sustainable waste management. Waste Manag. 24, 297e308.
Munda, G., 2008. Social Multi-criteria Evaluation for Sustainable Economy. ISBN
978-3-540-73702-5. Springer-Verlag Berlin Heideberg.
Myllyviita, T., Holma, A., Antikainen, R., L€ ahtinen, K., Leskinen, P., 2012. Assessing
environmental impacts of biomass production chains e application of life cycle
assessment (LCA) and multi-criteria decision analysis (MCDA). J. Clean. Prod.
29e30, 238e245. July 2012.
OECD - Organisation for Economic Co-operation and Development, 2011. Better Life
Please cite this article in press as: Angelo, A.C.M., et al., Life Cycle Assessment and Multi-criteria Decision Analysis: Selection of a strategy for
domestic food waste management in Rio de Janeiro, Journal of Cleaner Production (2016), http://dx.doi.org/10.1016/j.jclepro.2016.12.049
13
Index. http://www.oecdbetterlifeindex.org.
OSB/WSTB, 2000. Ocean Studies Board and Water Science and Technology Board.
Phillips, L., Phillips, M.C., 1993. Facilitated work groups: theory and practice. J. Oper.
Res. Soc. 44, 533e549.
Pires, A., Chang, N., 2011. Resources, Conservation and Recycling an AHP-based
fuzzy interval TOPSIS assessment for sustainable expansion of the solid waste
management system in Setúbal Peninsula, Portugal. Resour. Conserv. Recycl. 56,
7e21.
Prado-Lope z, V., 2015. Stochastic Multi Attribute Analysis for Comparative Life Cycle
Assessment. PhD thesis. Arizona State University, Arizona, USA.
Prado-Lopez, V., Seager, T.P., Chester, M., Laurin, L., Bernardo, M., Tylock, S., 2014.
Stochastic multi-attribute analysis (SMAA) as an interpretation method for
comparative life cycle assessment (LCA). Int. J. Life Cycle Assess. 19 (2),
405e416.
Reap, J., Roman, F., Duncan, S., Bras, B., 2008. A survey of unresolved problems in life
cycle assessment: Part 2: impact assessment and interpretation. Int. J. Life Cycle
Assess. 13, 374e388.
Recchia, L. (Ed.), 2011. Multicriteria Analysis and LCA Techniques: with Applications
to Agro-engineering problems, Green Energy and Technology. Springer-Verlag,
London; New York.
Riber, C., Pedersen, C., Christiansen, T.H., 2009. Chemical composition of material
fractions in Danish household waste. Waste Manag. 29 (4), 1251e1257.
Rogers, K., Seager, T.P., Linkov, I., 2008. Multicriteria Decision Analysis and Life Cycle
Assessment: Applications under High Uncertainty. Springer, Amsterdam, The
Netherlands.
Rowley, H.V., Peters, G.M., Lundie, S., Moore, S.J., 2012. Aggregating sustainability
indicators: beyond the weighed sum. J. Environ. Manag. 111, 24e33.
Sadok, W., Angevin, F., Bergez, J-e ., Bockstaller, C., Colomb, B., Guichard, L., Reau, R.,
Dore , T., 2008. Ex ante Assessment of the Sustainability of Alternative Cropping
Systems: implications for Using Multi-criteria Decision-Aid Methods e a Re-
view. Agron. Sustain. Dev. 28 (2008), pp.163e174.
Sadok, W., Bergez, J., Bockstaller, C., Colomb, B., Guichard, L., Reau, R., Mess, A.,
Dore , T., 2009. MASC, a qualitative multi-attribute decision model for ex ante
assessment of the sustainability of cropping systems. Agron. Sustain. Dev. 29,
447e461.
Scott, R.P., Cullen, A.C., Fox-lent, C., Linkov, I., 2016. Devices? Evaluation of Perfor-
mance and Impacts Using Integrated Life-cycle Assessment and Decision
Analysis. Risk Analysis (published on-line).
SGC, 2013. Biogas Upgrading e Review of Commercial Technologies. Svenskt Gas-
tekniskt Centrum. Report 2013, p. 270.
SMAC, 2012. Municipal Plan of Solid Waste Management of Rio de Janeiro City.
Municipal Secretariat of Environment. Rio de Janeiro, RJ, Brazil. http://www.rio.
rj.gov.br/web/smac/residuos-solidos. accessed 05 10 14.
Soltani, Atousa, Hewage, Kasun, Reza, Bahareh, Sadiq, Rehan, 2015. Multiple
stakeholders in multi-criteria decision-making in the context of municipal solid
waste management: a review. Waste Manag. 35, 318e328.
Starberg, K., 2005. Evaluation of Large Scale Systems for Compost and Digestion of
Source Separated Biowaste (In Swedish). Report RVF Utveckling 2005:06.
Malmo €, Sweden.
Stranddorf, H.K., Hoffmann, L., Schmidt, A., 2005. Impact Categories, Normalisation
and Weighting in LCA. Updated on selected EDIP97-data. Danish EPA Report 78.
Swedish Waste Management Organization, 2012. Certifierad Återvinning Av Bio-
go€dsel Och Kompost. SPCR120 och SPCR152. ISSN 1103e4092, Malmo € , Sweden.
Ulukan, H.Z., Kop, Y., 2009. Multi-criteria decision making (MCDM) of solid waste
collection methods using life cycle assessment (LCA) outputs. In: Computers &
Industrial Engineering. CIE, International Conference, pp. 584e589.
Valle, R., Clímaco, J.N., 2015. A New Tool to Facilitate the Quantitative Assessment of
Green Activities - a Trial Application for Rio de Janeiro. Technol. Forecast. Soc.
Change 28, 336e344.
Ventura, A., Dias, L., Clímaco, J., 2010. Using action research on the process of de-
cision support with VIP-analysis software. In: Respício, A., Adam, F., Philips-
Wren, G., Teixeira, C., Telhada, J. (Eds.), Bridging the Socio-technical Gap in
Decision Support Systems e Challenges for the Next Decade. IOS Press, 2010.
Ventura, A., Dias, L., Clímaco, J., 2014. On facilitating group decision making pro-
cesses with VIP-Analysis. In: Zarate , P., Kersten, G., Hern andez, J. (Eds.), Joint
INFORMS Group Decision and Negotiation and EWG-dss International Confer-
ence, GDN 2014, Toulouse, 10-13 June 2014. In Proceedings, pp. 246e253. LNBIP
180.
VMAB, 2012. Based on the use of Uni-Cut Technology in VMAB Dry Anaerobic
Digestion Plant in Karlshamn, Sweden. http://vmab.se/.
Wang, X., Chan, H.K., Li, D., 2015. A case study of an integrated fuzzy methodology
for green product development. Eur. J. Oper. Res. 241, 212e223.
Yatsalo, B., Kiker, G., Kim, J., Bridges, T., Seager, T., Gardner, K., Satterstrom, K.,
Linkov, I., 2007. Application of multi-criteria decision analysis tools for man-
agement of contaminated sediments. Integr. Environ. Assess. Manag. 3,
223e233.
Yoshida, H., Monster, J., Scheutz, C., 2014. Plant-integrated measurement of
greenhouse gas emissions from a municipal wastewater treatment plant. Water
Res. 61, 108e118.
Yue, D., Tian, E., Zhang, Y., 2009. A piecewise analysis method to stability analysis of
linear continuous/discrete systems with time-varying delay. Int. J. Robust
Nonlinear Control 1493e1518.
Zaras, K., Marin, J.-C., Boudreau-Trude, B., 2012. Dominance-based rough set
approach in selection of portfolio of sustainable development projects. Am. J.
Oper. Res. 2, 502e508.