Given the great impacts associated with the construction and maintenance of infrastructures in both the environmental, the economic and the social dimensions, a sustainable approach to their design appears essential to ease the fulfilment of the Sustainable Development Goals set by the United Nations. Multicriteria decision-making methods are usually applied to address the complex and often conflicting criteria that characterise sustainability. The present study aims to review the current state of the art regarding the application of such techniques in the sustainability assessment of infrastructures, analysing as well the sustainability impacts and criteria included in the assessments. The Analytic Hierarchy Process is the most frequently used weighting technique. Simple Additive Weighting has turned out to be the most applied decision-making method to assess the weighted criteria. Although a life cycle assessment approach is recurrently used to evaluate sustainability, standardised concepts, such as cost discounting, or presentation of the assumed functional unit or system boundaries, as required by ISO 14040, are still only marginally used. Additionally, a need for further research in the inclusion of fuzziness in the handling of linguistic variables is identified.
Sustainable development was first defined in 1987 by the Brundtland Commission as a way to meet the present needs of the society without compromising the ability of future generations to meet their own needs. Sustainable actions and decisions shall therefore be based on the simultaneous consideration of their economic, environmental, and social consequences over time. Sustainable design of products, as an application of the sustainability concept in the industry, takes particular relevance when considering the construction sector. In recent times, construction industry has become one of the main environmental stressors of our society, since it is responsible for 30% of global energy consumption, 40% of raw material extraction, and 30% of greenhouse gas emissions [
So, given the relevant implications of infrastructure design, and considering that most infrastructures are designed to serve a significant group of people over a long, intergenerational period of time, the assessment of the different dimensions of sustainability related to the infrastructure design has been in the spotlight of many researchers in recent times. Studies have been conducted on cost optimisation of infrastructure design [
To deal with the assessment of the conflicting dimensions of sustainability in a multistakeholder and long-term context like infrastructure design, the use of multicriteria decision-making (MCDM) techniques has revealed itself as the most suitable approach compared to other methods commonly used in infrastructure design, such as single- or multiobjective optimisation. MCDM techniques allow the decision maker to assess complex problems involving multiple and divergent criteria on the basis of the subjective judgements of a panel of experts or of stakeholders affected by the decision. Therefore, this paper is devoted to analysing the current trends regarding the application of MCDM techniques to the sustainability assessment of infrastructure design, paying special attention to the particular criteria considered in these assessments.
The rest of the paper is structured as follows. Section
The present study formulates two research questions, namely, how MCDM methods have been applied for the sustainability assessment of infrastructures in recent times, and what particular impact criteria and indicators have been considered in these evaluations as representative for the sustainability of an infrastructure design.
The data collection process performed in the present literature review consists of two stages, as shown in Figure
Systematic literature review.
The search is carried out through the scientific bibliographic databases SCOPUS and Web of Science. The search period is established from 1995 to 2019, since there is no evidence of relevant contributions before that date. The search algorithm used to identify the articles conforming the preliminary set consists of a combination of the terms “Multi-criteria decision making,” “MCDM,” and “Sustainability” along with other civil engineering-related terms, such as “Construction” or “Infrastructure,” by means of the Boolean operators “AND” and “OR.”
To filter the obtained results, some exclusion criteria have been followed to build the first set of papers. First, only original, peer-reviewed scientific articles and conference proceedings are included. Secondly, those manuscripts that do not clearly identify either the MCDM technique used or the sustainability criteria considered are excluded. Third, articles are required to consider at least two of the three dimensions of sustainability in the assessment through an appropriate selection of decision criteria. Finally, it should be taken into account that only articles written in English are considered in this study. This structured filtering process resulted in an initial set of 45 papers.
Once the initial set of contributions is generated, the references included in the selected manuscripts are then reviewed and analysed. The set is then expanded by applying the filtering process exposed above to the articles referenced in the papers included in the first set, which results in a final and expanded set of manuscripts. This sampling technique has been used previously in other literature review works [
Although in 2007 there was a first rebound in the number of publications regarding sustainability assessment of infrastructures, the number of contributions increases drastically in 2015 (Figure
Distribution of contributions per year (1996–2019).
As sustainability life cycle assessments are based on the life cycle impacts derived from the different activities considered in the analysis, it is essential to define in the early stages of the decision process not only which impacts (criteria) are going to be considered in the analysis, but also how those impacts are going to be assessed. Of the analysed publications, 74.7% (62 papers) base their assessments on the impacts derived from at least two different stages of the life cycle of the infrastructure under study. To evaluate the life cycle impacts and establish coherent impact categories, an objective methodology has been standardised in the environmental field [
Out of the 83 reviewed manuscripts, only 7 do not consider economic criteria in their sustainability assessments. Among the rest, three main economic impacts have been identified, namely, the construction or implementation costs, the costs derived from maintenance and operation of the infrastructure, and the costs resulting from the end of life stage. Table
Main economic criteria and indicators.
Economic criteria | Indicator | Assessment | References |
---|---|---|---|
Installation costs | €/output unit | Quantitative | [ |
Maintenance and operation costs | €/output unit | Quantitative | [ |
Disposal costs | €/output unit | Quantitative | [ |
It shall be noted that, among the reviewed papers, only 5 explicitly present the assumed discount rates that allow to transform future costs into present currency values. In the field of building design, Mosalam et al. [
Regarding the environmental dimension of sustainability, seven main impact categories have been found to be recurrent in the reviewed studies, namely, emission of pollutants, energy consumption, resources depletion, waste generation, land use, eutrophication, and ozone layer depletion. Table
Main environmental criteria and indicators.
Environmental criteria | Indicator | Assessment | References |
---|---|---|---|
Emission of pollutants | kg CO2/output unit | Quantitative | [ |
kg SO2/output unit | Quantitative | [ | |
kg NOx/output unit | Quantitative | [ | |
€/kg pollutant removed | Quantitative | [ | |
Costs of medical care needs due to pollution (€) | Quantitative | [ | |
Oxygen, nitrogen, and phosphates emitted to water | Quantitative | [ | |
Assessment by experts through point scale | Qualitative | [ | |
|
|||
Energy consumption | MJ (MWh)/output unit | Quantitative | [ |
€/year/output unit | Quantitative | [ | |
Tonnes of oil equivalent (TOE) | Quantitative | [ | |
Assessment by experts through point scale | Qualitative | [ | |
|
|||
Raw material consumption | Consumption/output unit | Quantitative | [ |
Assessment by experts through point scale | Qualitative | [ | |
|
|||
Waste generation | kg/output unit | Quantitative | [ |
Assessment by experts through point scale | Qualitative | [ | |
|
|||
Land use | m2/output unit | [ | |
Assessment by experts through point scale | Qualitative | [ | |
Aquatic ecotoxicity, salinity, biological indices | Quantitative | [ | |
|
|||
Eutrophication | kg phosphate/output unit | Quantitative | [ |
Assessment by experts through point scale | Qualitative | [ | |
|
|||
Ozone depletion | kg CFC (Chlorofluorocarbons)/output unit | Quantitative | [ |
Assessment by experts through point scale | Qualitative | [ |
The emission of pollutants as an indicator of the environmental impact of an infrastructure is the most used criterion within the reviewed papers. It considers the emissions derived from the production of construction materials and construction works, but also from the externalities associated with the construction of infrastructure and its maintenance, such as traffic congestion [
46.3% of the articles include energy consumptions as an additional measure of the environmental impact of an infrastructure. The majority of articles consider the energy needed to produce the construction materials and to construct the particular infrastructure under assessment [
The depletion of natural resources is accepted as one of the main consequences of unsustainable construction practices. 32 studies account for the consumption of natural resources into construction materials in their sustainability assessments. Some authors take into consideration the positive environmental impact of using recycled materials [
Given that the construction industry is considered one of the greatest producers of wastes in a global scale [
Land use is an environmental concept that implies both land occupation and transformation of land. Land use derived from the construction of infrastructures results in damage to ecosystems and loss of biodiversity. From the 83 reviewed articles, 25 (30.1%) take land use into account as an indicator of the environmental damage derived from infrastructures. The effects of land use have been accounted for as local ecosystem disturbances [
Eutrophication is the consequence of the emission of particular pollutants, mainly phosphate, derived from human activities to water, promoting an uncontrolled growth of algae that shall compromise the survival of other water species. This environmental impact has been considered by nine articles (10.8% of total).
Ozone layer is essential for life, as it hinders harmful solar ultraviolet radiation. Ozone layer depletion because of the emission of substances containing chlorine and bromine atoms has been accounted in eight studies as an additional indicator capable of measuring the environmental damage derived from infrastructures and their associated activities.
Regarding the social dimension of sustainability, the criteria assessed in the studies reviewed shall be grouped into eight main categories, namely, social wellbeing, aesthetics, job creation, development of local economies, externalities, innovation, culture, and health. Table
Main social criteria and indicators.
Social criteria | Indicator | Assessment | References |
---|---|---|---|
Social wellbeing | Increase of income of local population (€/year) | Quantitative | [ |
Assessment by experts through point scale | Qualitative | [ | |
Habitability increase (m2) | Quantitative | [ | |
Comfort (hours/year) | Quantitative | [ | |
|
|||
Aesthetics | Assessment by experts through point scale | Qualitative | [ |
|
|||
Job creation | Hours of work/output unit | Quantitative | [ |
Gross value added/hour worked | Quantitative | [ | |
Unemployment rate | Quantitative | [ | |
Employment increase (%) | Quantitative | [ | |
Assessment by experts through point scale | Qualitative | [ | |
|
|||
Development of local economies | GDP increase (€) | Quantitative | [ |
Land value degradation (€/m2) | Quantitative | [ | |
Assessment by experts through point scale | Qualitative | [ | |
|
|||
Externalities | Noise pollution (dB) | Quantitative | [ |
Traffic congestion (travel time) | Quantitative | [ | |
Vehicle operating costs (€), user delay costs (€) | Quantitative | [ | |
Assessment by experts through point scale | Qualitative | [ | |
|
|||
Innovation | Assessment by experts through point scale | Qualitative | [ |
|
|||
Culture | Assessment by experts through point scale | Qualitative | [ |
|
|||
Health and safety | Injuries/output unit | Quantitative | [ |
Fatalities/output unit | Quantitative | [ | |
Particulate matter (PM) concentration (PM2.5/PM10) | Quantitative | [ | |
Safety costs (€) | Quantitative | [ | |
Assessment by experts through point scale | Qualitative | [ |
The impact of an infrastructure on the social wellbeing is included in 34 manuscripts (41% of total) and combines aspects such as public acceptance [
Aesthetics has also been identified as a main indicator for social sustainability, which is closely related to social acceptance of the project. The aesthetic, which has been assessed in 26 articles, includes not only the aesthetical perception of the infrastructure itself, but also its integration with the urban [
Direct and indirect working opportunities derived from the construction and maintenance of an infrastructure have been considered in 16 studies (19.3% of total), which is closely related to an increase of the social welfare. Although the methodological sheets for social life cycle assessments developed by UNEP/SETAC [
16 studies take into consideration the effects of an infrastructure on the local development of a region, resulting from both the construction and maintenance activities, as well as from the serviceability provided by the infrastructure. Aspects such as the increase of the Gross Domestic Product [
Externalities derived from infrastructure construction and, mainly, from infrastructure maintenance have been considered in 33.7% of the reviewed studies. Effects such as traffic disruption [
The inclusion of innovative concepts in the infrastructure design is also accounted for as a social indicator, as it seeks to ensure the progress and technological development of the society. 9 articles have taken such aspect into account. The evaluation of this impact is based either on a binary indicator, which scores 1 if the design includes patented materials or solutions [
13.3% of the reviewed manuscripts include culture as a measure of social sustainability, paying special attention to the respect for the cultural heritage of a region [
Health and safety include both the practices of construction and industry companies to protect the lives of their workers, but also the risk of accidents for users of an infrastructure. The impact of the activities associated with the construction and maintenance of an infrastructure on the safety of the involved workers, as well as the risks to the health of the users of the infrastructures, has been considered in 42 articles (50.6% of total).
Once the indicators are selected that properly characterise the problem and condition the decision, the following step in a multicriteria decision-making problem consists in transforming them into quantitative values. While the numerical assessment of quantitative variables is straightforward, handling with qualitative criteria requires a certain preprocessing so as to transform such values into numerical ones. When dealing with qualitative criteria, such as aesthetics or comfort, many studies require the experts to evaluate such variables by assigning them scores on different scales ranging from 0 to 1, or from 0 to 10 in the most of the reviewed cases [
In other cases, experts are required to evaluate qualitative criteria by choosing one of the different answer options provided by the decision maker in a closed form, which are then directly related to specific numerical values. This approach is often preferred when dealing with complex problems, where experts find it easier to reflect their judgements in linguistic terms rather than in the form of precise numbers. For example, Gumus et al. [
Kripka et al. [
When dealing with indicators that are measured in different units, and prior to proceed to their aggregation into a final score, indicator values shall be normalised into dimensionless, comparable values. The most basic normalisation technique used is the so-called linear normalisation and consists in dividing the indicator value
Such approach is followed by the vast majority of studies reviewed [
Such approach is followed by [
Other studies, such as [
Particular value functions have been also used to normalise the indicator values into dimensionless values. So, studies based on the Simple Additive Weighting technique called MIVES [
The aggregation technique PROMETHEE also bases the normalisation step on the construction of preference functions. Vincke and Brans [
Finally, it shall be highlighted that the normalisation of indicator values only makes sense when the involved indicators are measured in different units. Thus, those studies based on the qualitative criteria assessment of experts, who set scores for each criterion based on their expertise [
Weighting the criteria is an essential step in a decision-making process, as it will condition the results of an assessment. Figure
Weighting techniques applied within the reviewed papers.
The direct allocation of weights has been identified as the second most used method (16 papers, 19.3% of the publications). By using this technique, the evaluator directly sets the score that represents the importance of each criterion on the decision-making problem. Shanon entropy methods are used to provide weightings less based on the subjectivity inherent in the previously mentioned techniques by measuring the uncertainty associated with the provided judgements [
At last, three other techniques have been marginally used, namely, the Best-Worst method (BWM) [
Once the relevance of each criterion is established, the next step in a decision-making process is to assess the obtained results. Figure
MCDM assessment techniques applied within the reviewed papers.
Thus, to overcome such limitations when handling with more complex criteria, other MCDM methods are used. Among them, the most applied one is TOPSIS (Technique for Order Preference by Similarity to Ideal Solution), used by 15.7% (13 studies) of the reviewed papers. TOPSIS allows to rank different alternatives in a multicriteria context, considering the fact that the most preferred solution should have the shortest geometric distance to the positive ideal solution and the longest distance to the less preferred solution [
PROMETHEE has been applied in 8.4% of the papers reviewed. This outranking method has suffered different modifications over the course of time, so as to overcome some of its initial limitations. PROMETHEE III, for example, does not even require the variables to be normalised and is applicable when information is missing [
7.2% of the studies reviewed use ELECTRE in their assessments. ELECTRE is another outranking technique based on concordance analysis. This noncompensatory method is particularly useful when ordinal scales are used to measure criteria [
COPRAS has also turned out to be one of the most used techniques in sustainability assessment of infrastructures. As ELECTRE method, COPRAS has been applied in 7.2% of the papers reviewed. COPRAS is recognised to be simple to calculate and, in contrast with SAW, adequate when dealing with both maximising and minimising criteria values [
It shall be noted that the proportions found in the present review regarding the use of MCDM techniques for the sustainability assessment of infrastructures have also been found in other fields of application. As an example, Kaya et al. [
Besides SAW, TOPSIS is revealed as the most used method to assess MCDM problems in different fields, such as supplier selection [
An important step in MCDM problems is to perform sensitivity analyses on those aspects that might alter significantly the conclusions of the assessment, so as to ensure the consistency of the final decision. From the total of the reviewed studies, only 18 (21.7%) include a sensitivity analysis in their assessments.
The majority of them (13 out of 18 manuscripts) focus their attention on the results sensitivity against the chosen criteria weights. This evidences that the weighting is considered as a great source of uncertainty in MCDM problems, usually derived from the subjectivity inherent to weighting based on experts’ judgements [
The usual way to proceed is to make one of the involved decision criteria predominant with respect to the rest and compare the results with the ones obtained after the conventional weighting [
Ignatius et al. [
MCDM problems have a highly subjective component, since they are generally based on the cognitive capacity of the decision makers, who are usually required to provide the relevance of each criterion and even to assign performance values to the selected criteria indicators, as derived from the results shown in the present literature review. However, during the application of the described steps inherent in a decision-making process, it is common practice to handle with so-called crisp or bivalued data. This is proved by 72.3% of the analysed manuscripts (60 papers), as shown in Figure
Handling of linguistic variables within the reviewed papers.
The fuzzy theory was further developed by Atanassov [
This study presents a systematic literature review on the sustainability assessment of infrastructure projects and designs developed by means of MCDM techniques. Given the complex characterisation of sustainability, MCDM is revealed as a useful tool to integrate decision criteria related to the different dimensions of sustainability, namely, economy, environment, and society. MCDM has gained relevance to evaluate sustainability mainly since 2015, when the Sustainable Development Goals were set by the United Nations. In particular, MCDM has been found to be mainly applied for the assessment of buildings (38.6%), bridges (15.7%), energy infrastructure (14.5%), hydraulic infrastructure (13.3%), and transport infrastructure (7.2%). In view of the results, more efforts should be put in the sustainability analysis of infrastructures where long-lasting, intergenerational service lives are required, such as bridges or dams. In those cases, where the required service life frequently exceeds 100 years, and where the magnitude of the impacts is not negligible given the dimensions of the infrastructures and their material and maintenance demands, evaluating the sustainability throughout their life cycle acquires an essential relevance.
AHP is revealed as the most used weighting technique to identify the relevance of the decision criteria, being applied in 65.1% of the analysed studies. Regarding the assessment technique used to evaluate the final sustainability scores of the design alternatives under consideration, SAW has resulted to be by far the preferred option, used by 51.8% of the authors. This technique, despite its undoubted advantages, such as its ease of use, is limited by the fact that it can only deal with positive defined, maximising criteria. Given the complex relations between sustainability criteria, and their often conflicting nature, other techniques have been used by the scientific community, being TOPSIS the most applied (15.7% of the contributions).
Regarding the mathematical handling of the linguistic variables involved in MCDM process, where the main variables to derive the criteria weights are usually the judgements and opinion of experts, it has been found that the vast majority of manuscripts assume a crisp approach. It is first since 2007 when authors have started to implement the fuzziness of human judgements into the decision-making process. Although fuzzy sets theory, and to some extent even intuitionistic fuzzy sets theory, have been applied in the sustainability assessment of infrastructures, the recently developed and more generalised neutrosophic sets have not been used to date for such purpose.
Regarding the criteria considered in the assessments, it shall be said that 74.7% base their definition on the framework of the life cycle of the infrastructure, which is in good accordance with the temporal dimension of sustainability. However, although recognised standards exist that provide guidelines for coherent and robust life cycle analyses, it has been found that only 4.8% of the publications base their studies on such standards, properly defining basic concepts such as the functional unit or the system boundaries assumed in the evaluation. It shall also be noted that none of the studies base the definition of the social criteria and indicators on the “Guidelines for social life cycle assessment of products,” which provides, at present, the most recognised methodology to perform social life cycle assessments. With regard to life cycle costing, only 6.1% of the authors take into consideration the discounting of costs related to time.
In view of the obtained review results, further research is required to integrate the existing life cycle impact assessment methodologies (economic, environmental, and social) into the multicriteria sustainability assessment of infrastructures, so as to provide robust and integral assessment tools based on a universal, systematic, and transparent methodology. In addition, further efforts should be made to consider the fuzziness of experts’ judgements in future assessment models.
The authors declare that there are no conflicts of interest regarding the publication of this paper.
The authors acknowledge the financial support of the Spanish Ministry of Economy and Competitiveness, along with FEDER funding (Project no. BIA2017-85098-R).