Content uploaded by Alice Denarie
Author content
All content in this area was uploaded by Alice Denarie on Dec 18, 2020
Content may be subject to copyright.
Multicriteria Approach
for a Multisource District Heating
Alice Dénarié, Marco Calderoni and Marcello Aprile
Abstract The EU project SmartReFlex—smart and flexible, 100%-renewable
district heating and cooling systems for European cities—aims to promote the
massive use of renewable sources for heating and cooling in cities through district
heating networks. Among the project activities, the analysis of real case studies
shows the potential of renewables in district heating systems. AIRU, Italian
Association of District Heating, and the Department of Energy of Politecnico di
Milano are supporting the promotion of local initiatives for renewable networks in
the Emilia Romagna region: the feasibility of a multisource DHC system in
Mirandola is assessed and presented in this paper. In Mirandola’s district heating
and cooling system, natural gas is only one among several possible energy sources:
alternative configurations integrating biomass, biogas and solar thermal have been
included in the study. The analysis deals with the extension of the network and with
the choice of the best new energy source to cover the new heat demand. The use of
MCDA has been applied in order to perform a holistic analysis of possible
energy-related choices by considering competing objectives. For instance, the use
of biomass is quite controversial: biomass is a renewable, local and a CO
2
neutral
source, able to reduce GHG emissions. However, biomass burning can have neg-
ative impacts on air quality by producing pollutants such as PM
10
, BaP, SOx and
NOx. This paper presents the multicriteria process applied to plant design, the
various alternatives and the criteria used. The result is a combination of natural gas,
biogas, solar thermal energy and biomass, which corresponds to the preference of
both the utility and municipality.
Keywords District heating Multicriteria Renewable energy
Sustainable planning
A. Dénarié(&)M. Calderoni M. Aprile
Dipartimento di Energia, Politecnico di Milano, via Lambruschini 4, Milan, Italy
e-mail: alice.denarie@polimi.it
©Springer International Publishing AG, part of Springer Nature 2018
A. Bisello et al. (eds.), Smart and Sustainable Planning for Cities and Regions,
Green Energy and Technology, https://doi.org/10.1007/978-3-319-75774-2_2
21
1 Introduction
One of the main aims of current European energy policy (Decision No 406/2009/
EC) is the minimization of fossil-fuel consumption to increase energy-supply
security and sustainability (Baños et al. 2011). If the objectives of energy policies
are clear, the path to reach them is not unique: several combinations of various
renewable energy sources can be integrated in a sustainable energy project. This is
particularly true for district heating. District heating (DH) is a system that delivers
hot water from a production site to several users distributed across the city through
a piping network. The European Commission (EC COM 2011) is mostly focusing
on buildings’heat demand savings and electrification of the heating sector. The
Heat Roadmap Europe (HRE) project (Connolly et al. 2012,2013) highlights that
the use of district heating fed by renewable-energy sources and industrial waste heat
can significantly reduce primary energy consumption and emissions’impact. This
study revealed a different scenario than the one proposed by the EU commission,
with the same energy results and at lower costs. Consistent with the HRE philos-
ophy, the SmartReFlex project promotes 100%-renewable district heating and
cooling networks by supporting local authorities. Among project activities, the
analyses of several case studies have been performed. This paper describes the
feasibility study for one of them in Mirandola, a small town of 24, 000 inhabitants
in the Po Valley of Italy. The case study deals with the extension of an existing
district heating and cooling system. It is currently based on natural gas and biogas,
and it is planned to double its capability to meet the heat demand. Consequently, it
will require the integration of new generation plants. The extension of the DH line
is foreseen by the local authority as the energy infrastructure that will bring
renewable heat to the city’s historical center. DH is an interesting way to bring
renewable energy to historical buildings, a type of energy that would otherwise
rarely be integrated because of architectural restrictions. The project’s main purpose
is to support the heat provider and the local authority in the choice among various
alternative heat sources and technologies for the new heat demand. The process is
based on a comprehensive cost-benefit analysis, which includes the environmental
impact. In particular, a combination of natural gas, biogas, solar thermal energy and
biomass has been assessed. Biomass is a renewable and local resource, and its
burning is mainly considered CO
2
neutral in energy policies (IPCC 2006), even if
various recent studies are questioning the carbon neutrality of biomass (Liu et al.
2017). Despite this benefit, biomass burning can have negative impacts on air
quality, in particular in the case of outdated inefficient technologies. In the regional
report about emissions, the ambivalence of biomass is evident by looking at the
various fuels impact on air quality (Fig. 1).
Primary sources of air pollution, such as PM
10
, and secondary ones, such as SO
x
and NO
x
, were found to be produced by biomass burners, mainly old firewood
boilers and open chimneys (ARPA and Regione Emilia 2010). DH systems,
compared to individual heating systems, are not necessarily less polluting a priori
(Genon et al. 2009), although the centralization and the use of one larger burner far
22 A. Dénariéet al.
from residential areas allows the use of higher stacks, higher-quality
pollution-reduction devices and higher efficiencies (Jonsson and Hillring 2006).
Moreover, DH networks allow the integration of various renewables. In this study,
the “zero emission”solar thermal technology has been chosen to reduce the bio-
mass impact on air quality in summer, since such a configuration has proved to be
an effective solution (Mathiesen et al. 2012). Biomass and solar radiation are energy
sources that imply intensive land use; despite their beneficial application in
sustainable-energy systems, their use in agricultural regions is creating some con-
cerns because of the competition over land use for food production (Bentsen and
Møller 2017). The adverse opinion by local authorities or citizens is not new
(Dénariéet al. 2017). For these reasons, the use of such land-intensive energy
sources requires an integrated approach in planning that privileges biomass residues
instead of energy crops (Calvert and Mabee 2015). A detailed calculation of the
biomass supplied to DH is crucial to assess its availability and the impact in case of
stock shortages.
2 Methodology
Dealing with renewable-energy projects at urban scale, such as DH, local author-
ities and the utilities providing heat have to answer several questions about the
choice of energy sources. The sustainability of system is now considered not only
from the economic perspective, but also from the point of view of environmental
impact, air quality and social impact (Wang et al. 2009). DH decisional process
involves multiple actors and considers a wide range of energy sources under
multiple aspects, as introduced in the previous paragraph. These are the reasons
why the Mirandola case study has been evaluated through multicriteria decision
Fig. 1 Fuels’impact in terms of CO
2
(a) and PM
10
(b) emissions in Emilia Romagna (ARPA and
Regione Emilia 2010)
Multicriteria Approach for a Multisource District Heating 23
analysis (MCDA) techniques. MCDA can break down the problem of decision
making into steps that can be evaluated with comparison criteria, meanwhile
granting mathematical robustness even to subjective evaluations (Kablan 2004). In
this work, a Multi Attribute Utility Technique (MAUT) has been applied (Neumann
and Morgenstern 1953; Savage 1954).
Application of MCDA to decisional problems follows normally the following
steps (Wang et al. 2009):
•Identification of goals !DH sustainability
•Definition of alternatives: a1;a2;...;aj;...;am!Four combination of solar
and biomass (Chap. 3)
•Identification of criteria: C1;C2;...;Ci;...;Cn!Seven criteria considered
here (Sect. 3.1)
•Scoring: calculation of performances for every alternatives according to chosen
criteria: x11;x12 ;...;xij;...;xnm !simulations’results (Chap. 3and Sect. 3.1)
•Criteria weighting w1;w2;...;wi;...;wn!according to actors’preferences
(Sect. 3.1)
•Application of multicriteria decision analysis methods !MAUT method
(Sect. 3.1 and 3.2)
•Aggregation of methods, in case of multiple Decision Makers (Sect. 3.2)
The MCDA methods are applied to the decision matrix, or performance matrix
X¼xij
resulting from the first five preliminary steps
a1ajam
C1
Ci
Cn
!
w1
wi
wn
x11 x1jx1m
xi1xij xim
xn1xnj xnm
2
43
5
ð1Þ
In the decision matrix, each element xij describes the performance of the various
alternatives j, with respect to each criterion i, which are the results of simulations
and spreadsheets. In MCDA problems, the performance matrix is filled with data
expressed in various units, and consequently they are first normalized to make them
comparable. In a MAUT application (Beinat 1997), the normalization from 0 to 1 is
done through expected utility functions. Expected utilities are mathematical func-
tions with various forms (linear, parabolic, Gaussian) that represent a preference.
They are used in decision problems under uncertainty and in risk analysis, and they
can quantify the decision maker’s subjectivity. A utility function uixij
enables
affirming that the performance result xij1of a certain alternative j1is better than
another alternative j2’s result xij2with respect to the same criteria iand its objective
(maximization or minimization):
uixij1
uixij2
ð2Þ
24 A. Dénariéet al.
Linear utility functions are used in this paper: they consist in giving utility values
to all performances by simple interpolation between extreme values 0 and 1; if a
criteria objective is its minimization, the score xij;min with the minimum value has
uixij;min
¼1 and uixij;max
¼0. The intermediate values are linearly interpolated.
See the application in the case study in Table 2.
The last step of the process is the weight attribution to each criterion. Criteria
weights reflect the relative importance of criteria given by the decision maker. In
this paper, a rank-order weight method has been applied, called SMARTER method
(Edwards and Barron 1994). The SMARTER method uses the #centroid method to
obtain criteria weights from decision makers, in this case, the heat provider and
local authority. It consists of giving weights to each criterion through an algorithm
that follows decision makers’rankings based on its preference. Being that i¼1is
the first preferred criteria and i¼nthe last, the less important one, the criteria
weights are calculated as:
wi¼1
nX
n
k¼i
1
kð3Þ
with Pn
i¼1wi¼1. After having scored and normalized alternatives, utility values
uixij
substitute nonhomogeneous scores xij in the decision matrix, which is now
filled with comparable values between 0 and 1. After having defined weights for
each criteria, the MCDA can be applied to combine results. The last step here is the
weighted sum of the utility values of all the alternatives for the various criteria.
sj¼X
n
i¼1
wiuij ð4Þ
At the end of this process, every alternative has a final weighted utility score sj
which enables alternatives’ranking, the final decision being the solution with the
highest value.
3 The Case Study
The existing district heating network covers an energy demand of 4010 MWh/year
of heat and 360 MWh/year of cold (data refers to 2014). The system recovers heat
from a natural-gas cogeneration heat and power plant (56% heat needs) and from a
third-party biogas internal-combustion engine (32% of heating needs), whereas a
natural gas back-up boiler (12% of heating needs) covers the peaks. The district
cooling system delivers chilled water produced by an absorption chiller (Fig. 2).
The local authority is willing to recover local residual biomass, currently treated
as waste, to be exploited in the new plant through a biomass boiler with high
efficiency. Solar-heat integration is of particular interest to reduce air-pollutant
Multicriteria Approach for a Multisource District Heating 25
emissions associated with biomass burning, especially in times when heat demand
is low. The feasibility assessment has been done considering energy performances,
economic aspects and environmental impacts. Such performances have been
quantified through simulations of future possible alternatives with energy PRO, a
simulation software for technical and economic analyses (EMD 2014). A model of
the current configuration of district heating with cogeneration systems, including
Italian electricity market interaction, has been built and validated through moni-
toring data (Dénariéet al. 2016). The validated model has been used to simulate
four configurations of various types of renewable-energy integrations for the
extension of the system (Fig. 3).
Four alternative plant configurations have been analyzed to exploit biomass and
solar radiation, maximizing the various aspects. Alternative 1 considers the simple
addition of a biomass boiler with a storage tank to guarantee good boiler man-
agement. This solution is the one originally foreseen by the heat provider. Being
aware of the air-quality impact of biomass and having as a main goal the reduction
of fossil-fuel consumption, the add-on of a solar thermal collector field has been
proposed in order to reduce the operational hours of the biomass boiler. Therefore,
two alternative solutions, n. 2 and n. 3 mixing biomass and solar thermal energy
have been added, with solar fields of, respectively, surfaces of 1600 m
2
and
Fig. 2 Existing district
heating and cooling system
Case. Biomass
boiler
Solar
thermal Storage
1 1 MW - 200 m3
2 1 MW 1600 m2160 m3
3 1 MW 2500 m21000 m3
4 - 1600 m2160 m3
Fig. 3 Alternative solutions for the system’s extension
26 A. Dénariéet al.
2500 m
2
. The first option is sized considering the space available, while the second
option maximizes economic incentives (D.M. 2016). Lastly, alternative n. 4 fore-
sees only a 1600 m
2
solar thermal plant integrated into the existing district heating
system, which is mainly based on natural gas, this being the solution with lowest
PM
10
impact. The four alternatives have been simulated in order to calculate the
energy, environmental and economic results. Details of the simulation’s input data
and hypothesis can be found in Dénariéet al. (2016).
Figures 4and 5show the four alternatives’energy production, environmental
impact and cash flow resulting from investment costs, operational costs, revenues
and incentives. Alternative n. 1 with the simple integration of a biomass boiler, has
the best financial results with lowest PBT (payback time) and highest NPV (net
present value), but it has the highest production of PM
10
. Solution n. 3 has the
highest renewable-energy production with the lowest environmental impact, both in
terms of greenhouse gas and particulate emissions. Nevertheless, its initial costs are
the highest, even if the NPV is not the lowest because of the practically nonexistent
costs of solar thermal operation. Solution n. 2 is quite similar to n. 3, performing
less well from an environmental point of view but with better economic results.
Solution 4 has the largest use of natural gas, which causes the highest production of
CO
2
, but with the lowest impact on air quality (PM
10
). It’s interesting to note that
the highest natural-gas consumption, despite the low investment costs, causes the
lowest NPV at the end of the system’s lifetime, clearly highlighting the high
operating costs associated with cogeneration. In conclusion, none of the solution
emerges as the best, and there are pros and cons for the three analyzed dimensions,
so a decision cannot be made without using a comprehensive holistic method.
3.1 MCDA Applied to the Case Study
The decision problem of selecting the new renewable-energy source to integrate in
the existing system can be better addressed by the MCDA method. In order to
compare the four alternative combinations of renewable energy and existing sources
Fig. 4 Energy (a) and economic (b) performances of the four alternative solutions
Multicriteria Approach for a Multisource District Heating 27
according to decision makers’preferences, different criteria could be considered.
Since the sustainability of the project is the main goal, usually four main criteria are
considered: technical, economic, environmental and social. Here, a total of seven
sub-criteria, have been considered to better address the main goal and considering
the project framework:
•Technical: Renewable Energy Ratio (RER)
•Economic: Investment costs, Payback Time (PBT), Net Present Value (NPV)
•Environmental: CO
2
,PM
10
emissions
•Social: Land use
Renewable Energy Ratio: RER is defined as the ratio of renewable-energy
production to total energy load, which gives an idea of the fossil-fuel savings in
favor of renewable energy.
Investment costs include all costs of technologies, installation and engineering of
the district heating extension project. This criterion is crucial since these costs are
very high compared to the individual system’s costs.
Pay Back Time is the period of time needed for the initial investment to be
repaid. For district heating projects, PBTs are quite long and this parameter is
usually crucial for private companies.
Net Present Value: is defined as the total cash-flow value at the end of system’s
lifetime, a commonly used indicator for long-term projects. It includes all costs and
measures financial feasibility.
CO
2
emissions quantify the system’s greenhouse-gas impact (Biomass is con-
sidered carbon neutral).
PM
10
quantifies the impact on air quality, a crucial parameter for projects
including biomass (see Chap. 1).
Land use is considered here as a social criterion because of aesthetic issues and
controversial opinions with regard to competition between agricultural and ener-
getic use of land (Chap. 1). Because of this, it can be an indicator of the social
acceptability of the project.
The performance matrix of the four alternatives is presented in Table 1.
Fig. 5 Environmental impact
of the four alternative
solutions
28 A. Dénariéet al.
Details of calculation are in Dénariéet al. (2016). The need to deal with the
problem of finding the best alternative with MCDA techniques is clearly demon-
strated by the fact that the best alternative is not always the same, when based on
different criteria. By using linear utility function, as explained in Chap. 2, for each
criterion every alternative’s score xij in Table 1is substituted by its normalized
utility value uixij
in Table 2. The performance matrix became the decision matrix:
The weights have been obtained through the SMARTER method based on the
preferences of the heat provider running the plant (survey result) and the local
authority (estimated) respectively:
•Criteria importance for heat provider: 1. Economy > 2. Energy > 3.
Environment > 4. Land use
•Criteria importance for local authority: 1. Energy > 2. Environment > 3. Land
use > 4. Economy
Table 1 Performance matrix of alternatives
Criteria Subcriteria Goal Unit Alternatives
a
1
a
2
a
3
a
4
Technical RER Max % 56 67 74 38
Economic Investment costs Min M€0.64 1.26 1.77 0.81
NPV Max M€2.17 1.99 1.80 1.40
PBT Min Year 6 7 9 7
Environmental CO
2
Min t 1551 1173 916 1877
PM
10
Min kg 256 231 226 58
Social Land use Min m
2
500,000 475,158 488,609 4000
Table 2 Decision matrix of alternatives
x
local
x
heat prov
a
1
a
2
a
3
a
4
RER 0.52 0.27 0.5 0.8 1 –
Investment costs 0.02 0.17 1 0.5 –0.8
NPV 0.02 0.17 1 0.7 –0.7
PBT 0.02 0.17 1 0.9 0.7 –
CO
2
0.10 0.07 0.3 0.7 1 –
PM
10
0.10 0.07 –0.1 0.1 1
Land use 0.21 0.06 0 0.05 0.02 1
Multicriteria Approach for a Multisource District Heating 29
As expected, the most important criterion for the private company providing heat
is the project’s economics, while the local authority’s priority is to increase the
renewable-energy ratio in the town. According to formula (3). the heat provider’s
weights for the four main criteria are: 0.52 for economy, 0.27 for energy aspects,
0.14 for environmental aspects and 0.06 for social aspect. Subcriteria weights have
been obtained simply by dividing the relative main criteria weights by the number
of sub-criteria. The same process has been applied for local authority ranking.
3.2 Results
By applying the averaged sum of Eq. (4) for the two decision makers, final scores
of the four alternatives are calculated, as shown in Fig. 5. The outcomes of the
MCDA show the mathematical results of the stakeholder preferences.
According to the heat provider’s criteria preferences, economic performance is
the most important parameter, and the best solution is the one with the lower costs
and shorter PBT, n. 1. The second preferred criterion is the renewable-energy ratio.
Alternative n. 3 is actually the best solution in terms of the renewable-energy
fraction only, so the second in ranking is alternative n. 3, which has good perfor-
mance according to every criterion. The same happens for the ranking according to
the local authority: the best solution is n. 3, the one with the highest RER.
According to the second criterion, environmental impact, the best solution would be
n. 4, although it has very poor performances in other criteria. The second is again
alternative n. 2. By averaging the rankings of the two decision makers (giving the
same weight to the local authority’s and the heat provider’s rankings), a unique
expected utility value can be obtained. Figure 5shows the final ranking obtained by
the application of MCDA for the two stakeholders involved, the heat provider and
the local authority (Figs. 6and 7).
Fig. 6 Expected utility of the four alternative solutions according to heat provider (a) and local
authority (b)
30 A. Dénariéet al.
The best alternative is n2, with a difference in expected utility with the second
alternative n. 3 of only 0.029. A sensitivity analysis on the weighting methods
could state in what case these two alternatives could be equal or when a reverse
ranking could happen. The last alternative, the one which uses no biomass and has a
higher use of fossil fuels, is quite distant from the 3rd position, but this result
strongly depends on the price of biomass: the economic figures have been calcu-
lated here considering a price given by the utility, considering residuals biomass
recovery and a short supply chain. Because a systematic, public analysis of the
potential of residual-biomass recovery at the local level is not available, conse-
quently, in the past, some district heating projects based on biomass have faced
problems of stock shortage. They are forced to buy biomass from the market, with a
heavy impact on the economics of the project and on the environmental impact due
to transport issues. A second round of simulations has been done for this case study
showing that, considering the biomass market price instead of residuals price, the
economic performances of alternatives 1, 2 and 3 worsen, so that alternative n.
1 becomes comparable to n. 4. It is important to note that the weighting method has
a significant impact: the used SMARTER method has the advantage of being easier,
requiring just the ranking of criteria importance from the decision makers.
However, it creates a very marked distribution of weights with substantial differ-
ences between the most and the least important one. Other methods, such as
Analytic Hierarchy Process (Saaty 2008), give weights through pairwise compar-
ison of criteria, obtaining more homogenous weights.
4 Conclusions
In this paper, a multicriteria process has been applied to assess four alternative
configurations of renewable-energy integration in an existing district heating and
cooling system to cover the new heat demand. A combination of natural gas,
biogas, solar thermal energy and biomass has been defined which corresponds to
Fig. 7 Aggregated final
expected utility of the four
alternative solutions
Multicriteria Approach for a Multisource District Heating 31
the preferences of the two decision makers involved: the heat provider running the
systems and the local authority. The alternatives have been ranked following seven
criteria: renewable-energy ratio, investment costs, net present value, payback time,
CO
2
emissions, PM
10
emissions and land use. Criteria weights have been assigned
according to the SMARTER method, and the ranking has been done using linear
utility functions. The application of MCDA techniques is an effective way to deal
with environmental sustainability of DH projects, by assessing through a holistic
approach a broad set of alternative solutions, with the various stakeholders involved
and their conflicting objectives. The application of MCDA techniques can give
mathematically rigorous shape to subjective evaluations, though they are strongly
dependent on weights and utility functions, which are very sensitive to subjective
parameters.
Acknowledgements The work presented in this paper is a result of the work undertaken in
SmartReFlex project (IEE/13/434/SI2.674873) that was co-financed by the Intelligent Energy
Europe (IEE) EU programme.
References
ARPA, & Regione Emilia, R. (2010). Aggiornamento Inventario Regionale delle Emissioni in
Atmosfera Anno 2010 (Vol. 2010).
Baños, R., Manzano-Agugliaro, F., Montoya, F. G., Gil, C., Alcayde, A., & Gómez, J. (2011).
Optimization methods applied to renewable and sustainable energy: A review. Renewable and
Sustainable Energy Reviews, 15(4), 1753–1766.
Bentsen, N. S., & Møller, I. M. (2017). Solar energy conserved in biomass: Sustainable bioenergy
use and reduction of land use change. Renewable and Sustainable Energy Reviews, 71, 954–
958.
Calvert, K., & Mabee, W. (2015). More solar farms or more bioenergy crops? Mapping and
assessing potential land-use conflicts among renewable energy technologies in eastern Ontario,
Canada. Applied Geography, 56, 209–221.
Connolly, D., Mathiesen, B. V., Østergaard, P. A., Möller, B., Nielsen, S., Lund, H., et al. (2012).
Heat roadmap Europe 2050—First pre-study for EU 27 for Euro heat & power. Brussels:
Aalborg University.
Connolly, D., Mathiesen, B. V., Østergaard, P. A., Möller, B., Nielsen, S., Lund, H., et al. (2013).
Heat roadmap Europe 2050—Second pre-study for EU27 for Euro heat & Power. Brussels:
Aalborg University.
Dénarié, A., Calderoni, M., & Muscherà, M. (2017). Technical, financial and Urban potentials for
solar district heating in Italy. In A. Bisello, D. Vettorato, R. Stephens, & P. Elisei (Eds.), Smart
and sustainable planning for cities and regions: Results of SSPCR 2015 (pp. 15–32). Cham:
Springer.
Dénarié, A., Calderoni, M., Nourrice, K. S., & Acerbis, S. (2016). Solar and biomass for Italian
communities: Analysis of two case studies. In International Solar District Heating Conference,
Billund.
Edwards, W., & Barron, F. H. (1994). SMARTS and SMARTER: Improved simple methods for
multiattribute utility measurement. Organizational Behavior and Human Decision Processes,
60, 306–325.
32 A. Dénariéet al.
Genon, G., Torchio, M. F., Poggio, A., & Poggio, M. (2009). Energy and environmental
assessment of small district heating systems: Global and local effects in two case-studies.
Energy Conversion and Management, 50(3), 522–529.
Jonsson, A., & Hillring, B. (2006). Planning for increased bioenergy use—Evaluating the impact
on local air quality. Biomass and Bioenergy, 30(6), 543–554.
Kablan, M. M. (2004). Decision support for energy conservation promotion: An analytic hierarchy
process approach. Energy Policy, 32(10), 1151–1158.
Liu, W., Zhang, Z., Xie, X., Yu, Z., von Gadow, K., Xu, J., et al. (2017). Analysis of the global
warming potential of biogenic CO
2
emission in life cycle assessments. Scientific Reports, 7
(January), 484–489.
Mathiesen, B. V., Lund, H., & Connolly, D. (2012). Limiting biomass consumption for heating in
100% renewable energy systems. Energy, 48(1), 160–168.
Neumann, J. V., & Morgenstern, O. (1953). Theory of games and economic behavior. Princeton,
NJ: Princeton University Press.
Saaty, T. L. (2008). Decision making with the analytic hierarchy process. International Journal of
Services Sciences, 1(1), 83.
Savage, L. J. (1954). The foundations of statistics. New York: Wiley.
Wang, J. J., Jing, Y. Y., Zhang, C. F., & Zhao, J. H. (2009). Review on multi-criteria decision
analysis aid in sustainable energy decision-making. Renewable and Sustainable Energy
Reviews, 13(9), 2263–2278.
Multicriteria Approach for a Multisource District Heating 33
