A Sustainable Evaluation Processes for Investments in the Transport Sector: A Combined Multi-Criteria and Cost–Benefit Analysis for a New Highway in Italy
Abstract
:1. Introduction
- Standardize the procedures followed by decision-makers (e.g., through specific guidelines or regulations);
- Estimate the effects of a new intervention/investment through consolidated quantitative methods, also by introducing prudential hypothesis to avoid benefit overestimation;
- Enlarge consensus to reach shared choices with all the stakeholders. Indeed, shared decision-making reduces the risk of planning failure and often reduces (or does not increase) the infrastructure realization time, which have a significant impact, for example, on the investment cost.
2. Literature Review
3. A Sustainable Evaluation Methodology for Investments in the Transport Sector
- The combined use of both CBA and MCA analysis to ensure robust and rational choices for sustainable development;
- The implementation of a transportation system model (TSM) to reduce the risk of “planning fallacy”;
- Stakeholder engagement during the decision-making process to ensure shared choice.
- Transport users’ benefits, which are the impacts produced for users of the transport system, including those directly involved with the project; these impacts could be both perceived by the transport users (e.g., travel time and cost reduction, consumer surplus) and not perceived by the travelers (e.g., savings costs for car maintenance);
- Non-users’ benefits, which are those produced for those who will not use the transport system and are not involved with the project, but will still experience some positive impacts (e.g., externalities reduction, pollutant reductions, quality of life, accident reduction).
- (a)
- Evaluation criteria and/or sub-criteria definition;
- (b)
- “Evaluation matrix” estimation;
- (c)
- An estimate of benefits: of all the positive impacts produced by the infrastructure, only some benefits (those more easily observable and/or measurable) have been considered in the analyses;
- An estimate of costs: overestimation of the investment costs and underestimation of the residual value of the investment.
4. The Application Case Study: A New Highway in Italy
4.1. Identification
- Scenario 1: a “stand-alone” scenario, improving the actual road path “Romea” quality and safety;
- Scenario 2: a new toll highway scenario;
- Scenario 3: in addition to Scenario 2, there is the “Romea” conversion into a “green-way” with restrictive policies for cars and freight vehicles beyond the quality of life and tourism improvements into the area.
4.2. Traffic Demand Estimation
Vj = βtime · trj + βtime · st.dev.trj + βcost · tollj + βcost · fuel_costj
- The potential demand interested in the design scenario was about 20–25 thousand vehicles/day (cars and freight vehicles);
- The traffic demand catchable by the new highway (Scenario 2 and 3) was about 18–20 thousand vehicles/day (cars and freight vehicles), equal to about 80% of the overall potential demand.
β3 * %Var_toll (t2 − t1) + β4 * %Var_Population (t2 − t1)
4.3. Cost Estimation
4.4. Benefits Estimation in Costs–Benefits Analysis
- The benefits for the users have been estimated considering the benefits produced by the design scenario during travel time; reliability of travel time; the cost of travel and operating costs;
- The positive effects (benefits) of the intervention on economic growth have not been considered (e.g., in terms of traffic revenues, impact on tourism, increasing accessibility of existing industrial centers and other wider economic benefits) because they are not standardized by national and European guidelines;
- Underestimating the quality of life benefits of alternative design: in the analyses, the benefits in terms of the reduction in the number of accidents are estimated, but the positive effects on social and territorial cohesion are not;
- Underestimating the environmental benefits: the benefits in terms of local climate change, air pollution and noise reduction are estimated, but the positive effects of the intervention in terms of protected and restricted areas at risk (hydrogeological, landslide) are not;
- The residual value of the investment underestimation, equal to 30% of the total investment cost (prudential hypothesis);
- Underestimating the vehicle fleet renewal over the analyzed time period.
4.5. Benefits Estimation in Multi-Criteria Analysis
5. Results and Discussion
6. Conclusions
Author Contributions
Funding
Acknowledgments
Conflicts of Interest
References
- Kahneman, D.; Tversky, A. Prospect theory: An analysis of decision under risk. Econometrica 1979, 47, 363–391. [Google Scholar] [CrossRef] [Green Version]
- Guide to Cost-Benefit Analysis of Investment Projects, Economic Appraisal Tool for Cohesion Policy 2014–2020. Available online: https://ec.europa.eu/regional_policy/sources/docgener/studies/pdf/cba_guide.pdf (accessed on 12 November 2019).
- Cascetta, E.; Cartenì, A.; Pagliara, F.; Montanino, M. A new look at planning and designing transportation systems: A decision-making model based on cognitive rationality, stakeholder engagement and quantitative methods. Transp. Policy 2015, 38, 27–39. [Google Scholar] [CrossRef]
- Cartenì, A. Accessibility indicators for freight transport terminals. Arab. J. Sci. Eng. 2014, 39, 7647–7660. [Google Scholar] [CrossRef] [Green Version]
- Cartenì, A.; De Guglielmo, M.L.; Henke, I. Design of Sustainable Urban Transport Infrastructures: A Real Case Application in Italy. Int. J. Civ. Eng. Technol. 2018, 9, 2131–2147. [Google Scholar]
- Henke, I.; Cartenì, A.; Molitierno, C.; Errico, A. Decision-Making in the Transport Sector: A Sustainable Evaluation Method for Road Infrastructure. Sustainability 2020, 12, 764. [Google Scholar] [CrossRef] [Green Version]
- Roy, B.; Vincke, P. Multicriteria analysis: Survey and new directions. Eur. J. Oper. Res. 1981, 8, 207–218. [Google Scholar] [CrossRef]
- Wang, J.J.; Jing, Y.Y.; Zhang, C.F.; Zhao, J.H. Review on multi-criteria decision analysis aid in sustainable energy decision-making. Renew. Sustain. Energy Rev. 2009, 13, 2263–2278. [Google Scholar] [CrossRef]
- Malczewski, J. GIS-based multicriteria decision analysis: A survey of the literature. Int. J. Geogr. Inf. Sci. 2006, 20, 703–726. [Google Scholar] [CrossRef]
- Cartenì, A.; De Guglielmo, M.L.; Pascale, N. Congested urban areas with high interactions between vehicular and pedestrian flows: A cost-benefit analysis for a sustainable transport policy in Naples, Italy. Open Transp. J. 2018, 12, 273–288. [Google Scholar] [CrossRef]
- Carteni’, A.; Henke, I.; Molitierno, C. A cost-benefit analysis of the metro line 1 in Naples, Italy. Wseas Trans. Bus. Econ. 2018, 15, 529–538. [Google Scholar]
- Cartenì, A.; Henke, I. The evaluation of public investments according to the cost-benefit analysis: An application to the formia-gaeta railway line [La valutazione degli investimenti in opere pubbliche attraverso l’analisi costi-benefici: Un’applicazione alla riqualificazione della linea ferroviaria formia-gaeta]. Ing. Ferrov. 2018, 74, 651–681. [Google Scholar]
- Artis, M.; Carrion, J.L.; Moreno, R.; Pons, G.; Surinach, J. Efficiency measurements in infrastructure projects: Cost-benefit analysis of the Tunnel of Cadi. Int. J. Transp. Econ. Riv. Internazionale Econ. Trasp. 2000, 1, 401–423. [Google Scholar]
- Jacyna, M. The Multiobjective Optimisation to Evaluation of the Infrastructure Adjustment to Transport Needs. In Transport Science and Technology; Emerald Group Publishing Limited: Bingley, UK, 2006; pp. 395–405. [Google Scholar]
- Marianna, J. The modeling of the external cost influence on the modal split in the transport network. In Proceedings of the 2008 19th International Conference on Systems Engineering, Las Vegas, NV, USA, 19–21 August 2008; pp. 453–457. [Google Scholar]
- Jacyna-Gołda, I.; Jolanta, Ż.; Piotr, G. Models of traffic flow distribution for various scenarios of the development of proecological transport system. Arch. Transp. 2014, 32, 17–28. [Google Scholar] [CrossRef]
- De Mare, G.; Granata, M.F.; Nesticò, A. Weak and strong compensation for the prioritization of public investments: Multidimensional analysis for pools. Sustainability 2015, 7, 16022–16038. [Google Scholar] [CrossRef] [Green Version]
- Smith, C.W.; Smithson, C.; Wilford, D.S. Strategic Risk Management. In Institutional Investor Series in Finance; Harper and Row: New York, NY, USA, 1990; pp. 45–46. [Google Scholar]
- IOP Conference Series: Earth and Environmental Science 2019. Available online: https://iopscience.iop.org/issue/1755-1315/235/1 (accessed on 24 November 2020).
- Khakzad, N.; Khan, F.; Amyotte, P. Quantitative risk analysis of offshore drilling operations: A Bayesian approach. Saf. Sci. 2013, 57, 108–117. [Google Scholar] [CrossRef]
- Wang, T.; Qu, Z.; Nichol, T.; Yang, Z.; Dimitriu, D.; Clarke, G.; Bowden, D. Impacts of climate change on rail systems: A new climate risk analysis model. In Safety and Reliability-Safe Societies in a Changing World; Taylor & Francis Group: Oxfordshire, UK, 2018. [Google Scholar]
- Wu, Y.J.; Hayat, T.; Clarens, A.; Smith, B.L. Climate change effects on transportation infrastructure: Scenario-based risk analysis using geographic information systems. Transp. Res. Rec. 2013, 2375, 71–81. [Google Scholar] [CrossRef]
- Songer, A.D.; Diekmann, J.; Pecsok, R.S. Risk analysis for revenue dependent infrastructure project. Constr. Manag. Econ. 1977, 15, 377–382. [Google Scholar] [CrossRef]
- de Almeida Guimaraes, V.; Junior, I.C.; da Silva, M.A. Evaluating the sustainability of urban passenger transportation by Monte Carlo simulation. Renew. Sustain. Energy Rev. 2018, 93, 732–752. [Google Scholar] [CrossRef]
- Salling, K.B.; Leleur, S. Modelling of transport projects uncertainties: Risk assessment and scenario analysis. Eur. J. Transp. Infrastruct. Res. 2012, 12, 21–38. [Google Scholar]
- Lemp, J.D.; Kockelman, K.M. Understanding and accommodating risk and uncertainty in toll road projects: A review of the literature. Transp. Res. Rec. 2009, 2132, 106–112. [Google Scholar] [CrossRef] [Green Version]
- Arpke, A.; Hutzler, N. Operational life-cycle assessment and life-cycle cost analysis for water use in multioccupant buildings. J. Archit. Eng. 2005, 11, 99–109. [Google Scholar] [CrossRef]
- Malnaca, K.; Gorobetz, M.; Yatskiv, I.; Korneyev, A. Decision-Making Process for Choosing Technology of Diesel Bus Conversion into Electric Bus. In Reliability and Statistics in Transportation and Communication; Springer: Berlin/Heidelberg, Germany, 2018; Volume 68, pp. 91–102. [Google Scholar]
- Itoh, R. Is transportation infrastructure cost recoverable under the risk of disasters? Transp. Res. Part A Policy Pract. 2018, 118, 457–465. [Google Scholar] [CrossRef]
- Yates, B.T. Cost-effectiveness analysis and cost-benefit analysis: An introduction. Behav. Assess. 1985, 7, 207–234. [Google Scholar]
- Beria, P. Effectiveness and monetary impact of Milan’s road charge, one year after implementation. Int. J. Sustain. Transp. 2016, 10, 657–669. [Google Scholar] [CrossRef] [Green Version]
- Geske, J.; Berghout, N.; van den Broek, M. Cost-effective balance between CO2 vessel and pipeline transport. Part I—Impact of optimally sized vessels and fleets. Int. J. Greenh. Gas Control 2015, 36, 175–188. [Google Scholar] [CrossRef]
- Geng, Y.; Ma, Z.; Xue, B.; Ren, W.; Liu, Z.; Fujita, T. Co-benefit evaluation for urban public transportation sector—A case of Shenyang, China. J. Clean. Prod. 2013, 58, 82–91. [Google Scholar] [CrossRef]
- Shafiei, E.; Leaver, J.; Davidsdottir, B. Cost-effectiveness analysis of inducing green vehicles to achieve deep reductions in greenhouse gas emissions in New Zealand. J. Clean. Prod. 2017, 150, 339–351. [Google Scholar] [CrossRef]
- Weigel, W. The standard cost model-A critical appraisal. In Proceedings of the 25th Annual Conference of the European Association of Law and Economics, Haifa, Israel, 24–26 September 2008. [Google Scholar]
- Ponti, M. Evaluation of transport infrastructure projects beyond cost-benefit analysis. An application to Barcelona’s 4th ring road. A comment. Int. J. Transp. Econ. Riv. Internazionale Econ. Trasp. 2002, 29, 235–237. [Google Scholar]
- Dyr, T.; Misiurski, P.; Ziółkowska, K. Costs and benefits of using buses fuelled by natural gas in public transport. J. Clean. Prod. 2019, 225, 1134–1146. [Google Scholar] [CrossRef]
- Harford, J.D. Congestion, pollution, and benefit-to-cost ratios of US public transit systems. Transp. Res. Part D Transp. Environ. 2006, 11, 45–58. [Google Scholar] [CrossRef]
- Danon, G.; Tica, S.; Vasić, B. Cost benefit analysis of Belgrade tram rolling stock. J. Appl. Eng. Sci. 2006, 4, 57–68. [Google Scholar]
- Chi, S.; Bunker, J.; Teo, M. Measuring impacts and risks to the public of a privately operated toll road project by considering perspectives in cost-benefit analysis. J. Transp. Eng. Part A Syst. 2017, 143, 04017060. [Google Scholar] [CrossRef] [Green Version]
- Tsamboulas, D.A.; Yiotis, G.; Mikroudis, G. A method for multi-criteria analysis in transportation infrastructure investments. Int. J. Transp. Econ. Riv. Internazionale Econ. Trasp. 2007, 34, 113–131. [Google Scholar]
- Marletto, G. Transalpine transport policies: Towards a shared approach. Int. J. Transp. Econ. Riv. Internazionale Econ. Trasp. 2010, 37, 353–370. [Google Scholar]
- Tsamboulas, D. A tool for prioritizing multinational transport infrastructure investments. Transp. Policy 2007, 14, 11–26. [Google Scholar] [CrossRef]
- Zolfani, S.H.; Zavadskas, E.K.; Khazaelpour, P.; Cavallaro, F. The multi-aspect criterion in the PMADM outline and its possible application to sustainability assessment. Sustainability 2018, 10, 4451. [Google Scholar] [CrossRef] [Green Version]
- Hickman, R.; Saxena, S.; Banister, D.; Ashiru, O. Examining transport futures with scenario analysis and MCA. Transp. Res. Part A Policy Pract. 2012, 46, 560–575. [Google Scholar] [CrossRef]
- Nogués, S.; Gonzàlez-Gonzàlez, E. Multi-criteria impacts assessment for ranking highway projects in Northwest Spain. Transp. Res. Part A Policy Pract. 2014, 65, 80–91. [Google Scholar] [CrossRef]
- Awasthi, A.; Omrani, H.; Gerber, P. Investigating ideal-solution based multicriteria decision making techniques for sustainability evaluation of urban mobility projects. Transp. Res. Part A Policy Pract. 2018, 116, 247–259. [Google Scholar] [CrossRef]
- Decision Marking among Multiple Attribute Alternatives: A Survey and Consolidated Approach. Available online: https://www.rand.org/pubs/research_memoranda/RM4823.html (accessed on 12 November 2019).
- Roy, B. Classement et choix en prèsence de points de vue multiples. Rev. Fr. D’Informatique Rech. Oper. (Riro) 1968, 2, 57–65. [Google Scholar] [CrossRef]
- On the Internet Traffic Classification: A Multi-Criteria Decision Making Approach. Available online: http://www.sciepub.com/reference/155073 (accessed on 24 November 2020).
- Hwang, C.L.; Yoon, K. Multiple Attribute Decision Making: A State of the Art Survey. In Lecture Notes in Economics and Mathematical Systems; Springer: Berlin, Germany, 1981; Volume 186, pp. 58–191. [Google Scholar]
- Brans, J.P. La méthode PROMETHEE. In L’Ingéniérie de la Décision. Elaboration d’Instruments d’Aide à la Décision. Méthode PROMETHEE; Université LAVAL: Quebec, QC, Canada, 1982; pp. 183–213. [Google Scholar]
- Brans, J.P.; Mareschal, B.; Vincke, P. PROMETHEE: A new family of outranking methods in multicriteria analysis. In Operational Research ‘84 IFORS 84, 6–10 August 1984, Washington, DC, USA; Brans, J.P., Ed.; Elsevier Science Pub. Co.: Amsterdam, The Netherlands, 1984; pp. 477–490. [Google Scholar]
- Gomes, L.F.A.M.; Lima, M.M.P.P. TODIM: Basics and application to multicriteria ranking of projects with environmental impacts. Found. Comput. Decis. Sci. 1992, 16, 113–127. [Google Scholar]
- Saaty, T.J. Decision Making in Complex Environments, The Analytical Hierarchy Process for Decision Making with Dependence and Dependence and Feedback; RWS Publications: Pittsburgh, PA, USA, 1996. [Google Scholar]
- Saaty, T.L. Fundamentals of analytic network process. In Proceedings of the 5th International Symposium on the Analytic Hierarchy Process, Kobe, Japan, 12–14 August 1999; pp. 348–379. [Google Scholar]
- Zavadskas, E.K.; Kaklauskas, A. Determination of an efficient contractor by using the new method of multicriteria assessment. In International Symposium for “The Organisation and Management of Construction”. Shaping Theory and Practice 2: Managing the Construction Project and Managing Risk; Langford, D.A., Retik, A., Eds.; CIB W: London, UK; Weinheim, Germany; New York, NY, USA; Tokyo, Japan; Melbourne, Australia; Madras, India; E and FN SPON: London, UK, 1996; Volume 65, pp. 94–104. [Google Scholar]
- Opricovic, S. Multiple Criteria Optimization of Civil Engineering Systems; Faculty of Civil Engineering: Belgrade, Serbian, 1998. [Google Scholar]
- Zavadskas, E.K.; Turskis, Z.; Vilutiene, T. Multiple criteria analysis of foundation instalment alternatives by applying additive ratio assessment (ARAS) method. Arch. Civ. Mech. Eng. 2010, 10, 123–141. [Google Scholar] [CrossRef]
- Keršulienė, V.; Zavadskas, E.K.; Turskis, Z. Selection of rational dispute resolution method by applying new step-wise weight assessment ratio analysis (SWARA). J. Bus. Econ. Manag. 2010, 11, 243–258. [Google Scholar] [CrossRef]
- Ginevicius, R. A New Determining Method for the Criteria Weights in Multi-Criteria Evaluation. Int. J. Inf. Technol. Decis. Mak. 2011, 10, 1067–1095. [Google Scholar] [CrossRef]
- Zavadskas, E.K.; Turskis, Z.; Antucheviciene, J.; Zakarevicius, A. Optimization of weighted aggregated sum product assessment. Electron. Elektron. Ir Elektrotechnika 2012, 122, 3–6. [Google Scholar] [CrossRef]
- Krylovas, A.; Zavadskas, E.K.; Kosareva, N.; Dadelo, S. New KEMIRA Method for Determining Criteria Priority and Weights in Solving MCDM Problem. Int. J. Inf. Technol. Decis. Mak. 2014, 13, 1119–1133. [Google Scholar] [CrossRef]
- Rezaei, J. Best-worst multi-criteria decision-making method. Omega 2015, 53, 49–57. [Google Scholar] [CrossRef]
- Keshavarz Ghorabaee, M.; Zavadskas, E.K.; Olfat, R.; Turskis, Z. Multi-Criteria Inventory Classification Using a New Method of Evaluation Based on Distance from Average Solution (EDAS). Informatica. 2015, 26, 435–451. [Google Scholar] [CrossRef]
- Keshavarz Ghorabaee, M.; Zavadskas, E.K.; Turskis, Z.; Antucheviciene, J. A new combinative distance-based assessment (CODAS) method for multi-criteria decision-making. Econ. Comput. Econ. Cybern. Stud. Res. 2016, 50, 25–44. [Google Scholar]
- Keshavarz Ghorabaee, M.; Amiri, M.; Zavadskas, E.K.; Turskis, Z.; Antucheviciene, J. Simultaneous Evaluation of Criteria and Alternatives (SECA) for Multi-Criteria Decision-Making. Informatica 2018, 29, 265–280. [Google Scholar] [CrossRef] [Green Version]
- Yazdani, M.; Zarate, P.; Zavadskas, E.K.; Turskis, Z. A Combined Compromise Solution (CoCoSo) method for multi-criteria decision-making problems. Manag. Decis. 2018, 57. [Google Scholar] [CrossRef]
- Turró, M.; Penyalver, D. Hunting white elephants on the road. A practical procedure to detect harmful projects of transport infrastructure. Res. Transp. Econ. 2019, 75, 3–20. [Google Scholar] [CrossRef]
- Annema, J.A.; Mouter, N.; Razaei, J. Cost-benefit analysis (CBA), or multi-criteria decision-making (MCDM) or both: Politicians’ perspective in transport policy appraisal. Transp. Res. Procedia 2015, 10, 788–797. [Google Scholar] [CrossRef] [Green Version]
- Cascetta, E.; Cartenì, A.; Henke, I.; Pagliara, F. Economic growth, transport accessibility and regional equity impacts of high-speed railways in Italy: Ten years ex post evaluation and future perspectives. Transp. Res. Part A Policy Pract. 2020, 139, 412–428. [Google Scholar] [CrossRef] [PubMed]
- Beria, P.; Maltese, I.; Mariotti, I. Multi-Criteria versus Cost Benefit Analysis: A comparative perspective in the assessment of sustainable mobility. Eur. Transp. Res. Rev. 2012, 4, 137–152. [Google Scholar] [CrossRef] [Green Version]
- Improving Public Investment Management for Large-Scale Government Projects: Focusing on the Feasibility Studies: Lessons and Challenges. Available online: https://orbit.dtu.dk/en/publications/improving-public-investment-management-for-large-scale-government (accessed on 24 November 2020).
- Ambrasaite, I. Comprehensive Transport Infrastructure Assessment Based on the Rail Baltica Case. Master’s Thesis, Department of Transport, Technical University of Denmark, Lyngby, Denmark, 2010. [Google Scholar]
- Barfod, M.B.; Salling, K.B.; Leleur, S. Composite decision support by combining cost-benefit and multi-criteria decision analysis. Decis. Support Syst. 2011, 51, 167–175. [Google Scholar] [CrossRef]
- Ambrasaite, I.; Barfod, M.B.; Salling, K.B. MCDA and risk analysis in transport infrastructure appraisals: The Rail Baltica case. Procedia-Soc. Behav. Sci. 2011, 20, 944–953. [Google Scholar] [CrossRef] [Green Version]
- Telicka, P. Annual Activity Report-Priority Project N° 27. 2006. Available online: http://ec.europa.eu/ten/transport/coordinators/doc/2006_07_pp27_en.pdf (accessed on 12 November 2019).
- Cartenì, A.; De Guglielmo, M.L.; Pascale, N.; Calabrese, M. An adaptive rational decision-making process for developing sustainable urban mobility plans. Int. J. Civ. Eng. Technol. 2017, 8, 1147–1156. [Google Scholar]
- Cartenì, A.; Henke, I. External costs estimation in a cost-benefit analysis: The new Formia-Gaeta tourist railway line in Italy. In Proceeding of the 17th IEEE International Conference on Environment and Electrical Engineering, Milan, Italy, 6–9 June 2017. [Google Scholar]
- Cartenì, A.; Henke, I.; Mallozzi, F.; Molitierno, C. A multi-criteria analysis as a rational evaluation process for building a new highway in Italy. In WIT Transactions on Ecology and the Environment; WIT Press: Southampton, UK, 2018; Volume 217, pp. 713–723. ISSN 1743-3541. [Google Scholar] [CrossRef] [Green Version]
- Cartenì, A.; Henke, I.; Di Bartolomeo, M.I.; Regna, M. A Cost-Benefit Analysis of a Fully-Automated Driverless Metro Line in a High-Density Metropolitan Area in Italy. In Proceedings of the 2019 IEEE International Conference on Environment and Electrical Engineering and 2019 IEEE Industrial and Commercial Power Systems Europe, EEEIC/I and CPS Europe, Genova, Italy, 11–14 June 2019. [Google Scholar]
- Cartenì, A.; Galante, G.; Henke, I. An assessment of models accuracy in predicting railways traffic flows: A before and after study in Naples. Wit Trans. Ecol. Environ. 2014, 191, 783–794. [Google Scholar] [CrossRef] [Green Version]
- de Luca, S.; Cartenì, A. A multi-scale modelling architecture for estimating of transport mode choice induced by a new railway connection: The Salerno-University of Salerno-Mercato San Severino Route [Un’architettura modellistica multi-scala per la stima delle ripartizioni modali indotte da un nuovo collegamento ferroviario: Il caso studio della tratta Salerno-Università di Salerno-Mercato San Severino]. Ing. Ferrov. 2013, 68, 447–473. [Google Scholar]
- Cartenì, A.; Russo, F. A distribution regional freight demand model. Adv. Transp. 2004, 16, 275–285. [Google Scholar]
- Decreto del Ministro delle Infrastrutture e dei Trasporti. Linee Guida per la Valutazione Degli Investimenti in Opere Pubbliche nei Settori di Competenza del Ministero delle Infrastrutture e dei Trasporti. 2017. Available online: http://www.mit.gov.it/sites/default/files/media/notizia/2017-07/Linee%20Guida%20Val%20OO%20PP_01%2006%202017.pdf (accessed on 21 October 2020).
- Korzhenevych, A.; Dehnen, N.; Brocker, J.; Holtkamp, M.; Meier, H.; Gibson, G.; Varma, A.; Cox, V. Update of the Handbook on External Costs of Transport; Final Report for the European Commission, DG MOVE.; MOVE. DIW Econ, CAU.; Ricardo-AEA: London, UK, 2014. [Google Scholar]
- Cascetta, E. Transportation Systems Engineering: Theory and Methods. Applied Optimization; Springer: Boston, MA, USA, 2001. [Google Scholar]
- Cascetta, E. Transportation System Analysis: Models and Applications; Springer: New York, NY, USA, 2009. [Google Scholar]
- Cartenì, A.; Punzo, V. Travel time cost functions for urban roads: A case study in Italy. WIT Trans. Built Environ. 2007, 96, 233–243. [Google Scholar] [CrossRef] [Green Version]
- Cartenì, A. Urban sustainable mobility. Part 1: Rationality in transport planning. Transp. Probl. 2014, 9, 39–48. [Google Scholar]
- Cartenì, A.; Henke, I.; Regna, M.; Di Bartolomeo, M.I.; Di Francesco, L. A Stakeholder Engagement Process for a Rational Decision-Making Process in Transportation Planning. Int. J. Adv. Res. Eng. Technol. 2020, 11, 333–343. [Google Scholar]
- Cartenì, A.; D’Acierno, L.; Gallo, M. A Rational Decision-Making Process with Public Engagement for Designing Public Transport Services: A Real Case Application in Italy. Sustainability 2020, 12, 6303. [Google Scholar] [CrossRef]
- Bierlaire, M. BIOGEME: A free package for the estimation of discrete choice models. In Proceedings of the 3rd Swiss Transport Research Conference, Ascona, Switzerland, 19–21 March 2003. [Google Scholar]
- Bifulco, G.N.; Cartenì, A.; Papola, A. An activity-based approach for complex travel behaviour modelling. Eur. Transp. Res. Rev. 2010, 2, 209–221. [Google Scholar] [CrossRef] [Green Version]
- Henke, I. The Effect Of Railway Accessibility on the Choice of University Studies. Int. J. Transp. Dev. Integr. 2017, 1, 339–347. [Google Scholar] [CrossRef]
- Cascetta, E.; Cartenì, A.; Henke, I. Stations quality, aesthetics and attractiveness of rail transport: Empirical evidence and mathematical models [Qualità delle stazioni, estetica e attrattività del trasporto ferroviario: Evidenze empiriche e modelli matematici]. Ing. Ferrov. 2014, 69, 307–324. [Google Scholar]
- Cartenì, A.; Pariota, L.; Henke, I. Hedonic value of high-speed rail services: Quantitative analysis of the students’ domestic tourist attractiveness of the main Italian cities. Transp. Res. A Policy Pract. 2017, 100, 348–365. [Google Scholar] [CrossRef]
- Cartenì, A.; Cascetta, E.; de Luca, S. A random utility model for park & carsharing services and the pure preference for electric vehicles. Transp. Policy 2016, 48, 49–59. [Google Scholar] [CrossRef]
- Cascetta, E.; Carteni, A.; Henke, I. Acceptance and equity in advanced path-related road pricing schemes. In Proceedings of the 5th IEEE International Conference on Models and Technologies for Intelligent Transportation Systems, New York, NY, USA, 26–28 June 2017; pp. 492–496. [Google Scholar] [CrossRef]
- Cartenì, A. A new look in designing sustainable city logistics road pricing schemes. Wit Trans. Ecol. Environ. 2017, 223, 171–181. [Google Scholar] [CrossRef] [Green Version]
- Cantarella, G.E.; Cartenì, A.; de Luca, S. Stochastic equilibrium assignment with variable demand: Theoretical and implementation issues. Eur. J. Oper. Res. 2015, 241, 330–347. [Google Scholar] [CrossRef]
- Cartenì, A. Urban sustainable mobility. Part 2: Simulation models and impacts estimation. Transp. Probl. 2015, 10, 5–16. [Google Scholar] [CrossRef] [Green Version]
- Cartenì, A. Updating demand vectors using traffic counts on congested networks: A real case application. In WIT Transactions on the Built Environment; Wit Press: Southampton, UK, 2007; Volume 96, pp. 211–221. [Google Scholar]
- ISTAT. Available online: https://www.istat.it/ (accessed on 14 November 2019).
- de Luca, S. Public engagement in strategic transportation planning: An analytic hierarchy process based approach. Transp. Policy 2014, 33, 110–124. [Google Scholar] [CrossRef]
- ACI. Autoritratto. Available online: http://www.aci.it/laci/studi-e-ricerche/dati-e-statistiche/autoritratto.html (accessed on 14 November 2019).
- Cascetta, E.; Cartenì, A. A quality-based approach to public transportation planning: Theory and a case study. Int. J. Sustain. Transp. 2014, 8, 84–106. [Google Scholar] [CrossRef] [Green Version]
- Cascetta, E.; Cartenì, A.; Carbone, A. The quality in public transportation. The campania regional metro system. Ing. Ferrov. 2013, 68, 241–261. [Google Scholar]
- Cascetta, E.; Cartenì, A.; Montanino, M. A behavioral model of accessibility based on the number of available opportunities. J. Transp. Geogr. 2016, 51, 45–58. [Google Scholar] [CrossRef]
- Reid, N. The Delphi technique: Its contribution to the evaluation of professional practice. Prof. Competence Qual. Assur. Caring Prof. 1988, 230, 262. [Google Scholar]
- Akins, R.B.; Tolson, H.; Cole, B.R. Stability of response characteristics of a Delphi panel: Application of bootstrap data expansion. BMC Med. Res. Methodol. 2005, 5, 37. [Google Scholar] [CrossRef] [Green Version]
- De Brucker, K.; de Winne, N.; Peeters, C.; Verbeke, A.; Winkelmans, W. The economic evaluation of public investments in transport infrastructure: The use of multicriteria analysis. Int. J. Transp. Econ. Riv. Internazionale Econ. Trasp. 1995, 22, 261–281. [Google Scholar]
- Figueira, J.; Greco, S.; Ehrgott, M. Multiple criteria decision analysis. In International Series in Operations Research and Management Science, 78; Springer: New York, NY, USA, 2005. [Google Scholar]
- Le Pira, M.; Inturri, G.; Ignaccolo, M.; Pluchino, A. Modelling consensus building in Delphi practices for participated transport planning. Transp. Res. Procedia 2017, 25, 3725–3735. [Google Scholar] [CrossRef]
- Cartenì, A.; Henke, I.; Molitierno, C.; Di Francesco, L. Strong Sustainability in Public Transport Policies: An e-Mobility Bus Fleet Application in Sorrento Peninsula (Italy). Sustainability 2020, 12, 7033. [Google Scholar] [CrossRef]
- Cartenì, A. A cost-benefit analysis based on the carbon footprint derived from plug-in hybrid electric buses for urban public transport services. Wseas Trans. Environ. Dev. 2018, 14, 125–135. [Google Scholar]
Methods | Reference |
---|---|
SAW (Simple Additive Weighting) | MacCrimon [48] |
ELECTRE (Elimination and Choice Expressing Reality) | Roy [49] |
AHP (Analytic Hierarchy Process) | Saaty [50] |
TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) | Hwang & Yoon [51] |
PROMETHEE (Preference Ranking Organization Method for Enrichment of Evaluations) | Brans [52]; Brans et al. [53] |
TODIM (Interactive and Multi-Criteria Decision Making) | Gomes & Lima [54] |
ANP (Analytic Network Process) | Saaty [55,56] |
COPRAS (Complex Proportional Assessment) | Zavadskas & Kaklauskas [57] |
VIKOR (Multi-criteria Optimization and Compromise Solution) | Opricovic [58] |
ARAS (Additive Ratio Assessment) | Zavadskas et al. [59] |
SWARA (Step-Wise Weight Assessment Ratio Analysis) | Kersuliene et al. [60] |
FARE (Factor Relationship) | Ginevicius [61] |
WASPAS (Weighted Aggregated Sum Product Assessment) | Zavadskas et al. [62] |
KEMIRA (Kemeny Median Indicator Ranks Accordance) | Krylovas et al. [63] |
BWM (Best-Worst Method) | Rezaei [64] |
EDAS (Evaluation Based on Distance from Average Solution) | Keshavarz Ghorabaee et al. [65] |
CODAS (A New, Combinative Distance-Based Assessment) | Keshavarz Ghorabaee et al. [66] |
SECA (Simultaneous Evaluation of Criteria and Alternatives) | Keshavarz Ghorabaee et al. [67] |
CoCoSo (Combined Compromise Solution) | Yazdani et al. [68] |
Scenario 1 “Stand-Alone” | Scenario 2 New Toll Highway | Scenario 3 | ||
---|---|---|---|---|
Green-Way | New Toll Highway | |||
Population directly involved (Mlns inhabitants) | 1.5 | 1.0 | 2.5 | |
Road length (km) | 127 | 108 | 235 | |
Average speed (km/h) | 60 | 110 | 50 | 110 |
Average travel time saved with respect to the current scenario | up to 21% | up to 65% | up to 5% | up to 65% |
Average fuel cost (€/liter) | 1.3 | 1.3 | 1.3 | |
Highway toll (cent.€/km) | 0 | 0.73 | 0 | 0.73 |
Vehicle Types | Value of Time (VOT) [€/hour] |
---|---|
(βtime/βcost)Passenger cars | 12 |
(βtime/βcost)Freight vehicles | 30 |
Attributes | Passenger Cars | Freight Vehicles | ||||
---|---|---|---|---|---|---|
βtime | Robust Std err | Robust t-test | 0.048 | −2.070 | 0.030 | −2.011 |
βcost | Robust Std err | Robust t-test | 0.204 | −7.800 | 0.050 | −2.121 |
Parameters | Passenger Cars | Freight Vehicles | ||||
Adjusted rho-square | 0.659 | 0.581 | ||||
Final log-likelihood | −34.142 | −31.103 | ||||
Null log-likelihood | −106.52 | −106.52 |
Trend | Prudential Hypotheses | Optimistic Hypotheses |
---|---|---|
Resident population (from 2016 to 2056) | Linear according to ISTAT population trends (from 2000 to 2016) [104] | 15% higher than the historical ISTAT 2000–2016 years trend |
Gross Domestic Product (GDP) | Linear regression starting the data observed in the period 2000–2019. Data observed in the International Monetary Fund provisions in the World Economic Outlook Database from January | Equal to the average GDP of the European countries EU4 until 2019 and then constant (annual growth equal to that of 2019) for the analysis time period; according to a GDP growth of Italy which can make it closer to the main European countries |
Fuel pump price | Linear regression starting the data observed in the period 2000–2019. | Linear but 50% lower than that observed |
Motorway tolls | As planned from transport operator in the business plan (annual increase of 1.50%) | Reduced by 70% to the planned (annual increase of 1.05%) |
Variables (βi) | Passenger Cars | Freight Vehicles | ||||
---|---|---|---|---|---|---|
Traffic Demand Elasticity (βi) | Standard Error | t-Test | Traffic demand Elasticity (βi) | Standard Error | t-Test | |
GDP | 1.458 | 0.447 | 3.260 | 1.383 | 0.334 | 7.124 |
Fuel | −0.417 | 0.159 | −2.631 | −0.319 | 0.119 | −2.692 |
Toll | −0.266 | 0.114 | −1.423 | −0.180 | 0.086 | −1.420 |
Population | 1.745 | 0.977 | 1.738 | 1.681 | 0.780 | 2.251 |
Aggregate Statistics | Passenger Cars | Freight Vehicles |
---|---|---|
Adjusted rho-square | 0.594 | 0.709 |
Standard error | 0.032 | 0.024 |
Root Mean Square Error (RMSE) | 0.028 | 0.021 |
Mean Squared Error (MSE) | 0.001 | 0.000 |
t-test | 11.810 | 14.423 |
C1 Investment Costs | C2 Operative and Maintenance Costs | C3 Residual Value of the Investment | Total Costs | |
---|---|---|---|---|
2018 prices and with fiscal corrections (M€) | −2923.5 | −679.7 | 336.4 | −3603.2 |
Constant prices and with fiscal corrections (M€) | −3788.4 | −1525.8 | 1136.5 | −4177.7 |
Air Pollution Marginal Costs [€ct per Vehicle*km, 2016 Prices] | 2016 | 2026 | 2036 | 2046 | |
---|---|---|---|---|---|
Urban area | Car | 1.16 | 0.69 | 0.57 | 0.56 |
Freight vehicles | 5.25 | 2.83 | 2.25 | 1.95 | |
Bus | 19.58 | 13.08 | 7.26 | 4.24 | |
Sub-Urban Area | Car | 0.62 | 0.34 | 0.26 | 0.25 |
Freight vehicles | 2.82 | 1.43 | 1.08 | 0.88 | |
Bus | 13.06 | 8.98 | 5.19 | 2.76 | |
Rural Area | Car | 0.39 | 0.21 | 0.16 | 0.15 |
Freight vehicles | 1.82 | 0.82 | 0.57 | 0.43 | |
Bus | 9.00 | 6.00 | 3.28 | 1.61 | |
Motorways | Car | 0.41 | 0.21 | 0.16 | 0.15 |
Freight vehicles | 1.66 | 0.71 | 0.47 | 0.33 | |
Bus | 7.45 | 4.82 | 2.49 | 1.18 |
The External Costs Saved | Design Scenario with Optimistic Hypothesis 2018 Prices (M€) | % | Design Scenario with Prudential Hypothesis 2018 Prices (M€) | % |
---|---|---|---|---|
Climate Change | 63.6 | 6% | 38.4 | 5% |
Air Pollution | 23.3 | 2% | 14.0 | 2% |
Noise | 2.4 | 0% | 1.4 | 0% |
Congestion | 72.6 | 7% | 43.0 | 5% |
Accident | 872.0 | 81% | 693.8 | 85% |
Up and downstream | 44.1 | 4% | 26.2 | 3% |
Total | 1077.9 | 100% | 816.8 | 100% |
Objectives | Criteria | Scenario 1- Stand Alone | Scenario 2- New Highway | Scenario 3- New Highway + Green-Way |
---|---|---|---|---|
ROAD EFFICIENCY | Investment costs (millions of €) | −33 | −2824 | −2924 |
Management and maintenance costs (millions of €) | N.A. | −680 | −680 | |
SERVICE PROVIDER BENEFITS | Revenues (millions of total vehicles) | N.A. | 243 | 285 |
Business diversification (high = 1, medium = 0.66, low = 0.33) | N.A. | 0.66 | 1 | |
USERS BENEFITS | Cars generalized transport costs (millions of hours saved) | 8.5 | 250 | 275 |
Freight vehicles generalized transport costs (millions of hours saved) | 10.9 | 79 | 112 | |
ECONOMIC GROWTH | Tourism promotion: accessibility to inland historical centres (number) | N.A. | 5 | 5 |
Tourism promotion: a new green-way road (high, medium, low) | 0.33 | 0.66 | 1 | |
Freight centres and firms accessibility (number) | N.A. | 8 | 8 | |
QUALITY OF LIFE | Accidents risk (millions of €) | 81 | 1366 | 1608 |
Social cohesion (high, medium, low) | 0.33 | 0.66 | 1 | |
LOCAL ENVIRONMENTAL BENEFITS | Pollution (millions of €) | 1 | 67 | 72 |
Protected environmental areas (number) | N.A. | 2284 | 2867 | |
Hydrological risk areas (number) | 1218 | 1529 | ||
Landslide risk areas (number) | 1066 | 1338 | ||
GLOBAL ENVIRONMENTAL BENEFITS | Global warming (tons of C02*1000) | 148 | 3016 | 3459 |
Disability-Adjusted Life Years and usage of primary resources (millions of €) | 61 | 1297 | 1480 |
Intensity of Importance | Value of Importance |
---|---|
Same weight | 1 |
Weakly greater weight compared to another one | 3 |
Greater weight compared to another one | 5 |
Demonstrated relevance compared to another one | 7 |
Absolute dominance compared to another one | 9 |
Intermediate values between two judgements | 2, 4, 6, 8 |
Objectives | Criteria | Criteria Weight for Stakeholder Group 1 (Public) | Criteria Weight for Stakeholder Group 2 (Private) |
---|---|---|---|
ROAD EFFICIENCY | Investment costs | 1.2 | 15.7 |
Management and maintenance costs | 1.8 | 7.9 | |
SERVICE PROVIDER BENEFITS | Revenues from traffic | 1.3 | 11.6 |
Business diversification | 2.8 | 2.6 | |
USERS BENEFITS | Cars generalized transport costs | 8.8 | 5.5 |
Freight vehicles generalized transport costs | 12.3 | 5.5 | |
ECONOMIC GROWTH | Tourism promotion: accessibility to inland historical centres | 5.5 | 3.3 |
Tourism promotion: a new green-way road | 4.4 | 5.1 | |
Freight centres and firms accessibility | 4.3 | 8.8 | |
LIFE QUALITY | Accidents risk | 11.7 | 8.6 |
Social cohesion | 4 | 2.3 | |
LOCAL ENVIRONMENTAL BENEFITS | Pollution decrease | 13.0 | 4.0 |
Protected environmental areas | 7.2 | 1.8 | |
Hydrological risk areas | 5.6 | 5.6 | |
Landslide risk areas | 5.6 | 5.6 | |
GLOBAL ENVIRONMENTAL BENEFITS | Global warming | 4.5 | 1.0 |
Disability-Adjusted Life Years and usage of primary resources | 6.1 | 5.0 | |
TOTAL | 100 | 100 |
Objectives | Criteria | Criteria “Weight” | Objectives “Weight” |
---|---|---|---|
ROAD EFFICIENCY | Investment costs | 8.5 | 13.3 |
Management and maintenance costs | 4.8 | ||
SERVICE PROVIDER BENEFITS | Revenues from traffic | 6.5 | 9.2 |
Business diversification | 2.7 | ||
USERS BENEFITS | Cars generalized transport costs | 7.2 | 16.1 |
Freight vehicles generalized transport costs | 8.9 | ||
ECONOMIC GROWTH | Tourism promotion: accessibility to inland historical centres | 4.4 | 15.7 |
Tourism promotion: a new green-way road | 4.8 | ||
Freight centres and firms accessibility | 6.5 | ||
LIFE QUALITY | Accidents risk | 10.2 | 13.3 |
Social cohesion | 3.2 | ||
LOCAL ENVIRONMENTAL BENEFITS | Pollution decrease | 8.5 | 24.2 |
Protected environmental areas | 4.5 | ||
Hydrological risk areas | 5.6 | ||
Landslide risk areas | 5.6 | ||
GLOBAL ENVIRONMENTAL BENEFITS | Global warming | 2.8 | 8.3 |
Disability-Adjusted Life Years and usage of primary resources | 5.5 | ||
TOTALE | 100 | 100 |
Scenario 1 (Stand Alone) | Scenario 2 (New Highway) | Scenario 3 (New Highway + Green-Way) | |
---|---|---|---|
Benefit/Cost ratio | 1.0 | 1.8 | 1.9 |
% variation with respect to the best scenario | 90.0% | 5.6% | 0% |
Optimistic Hypothesis | Prudential Hypothesis | |
---|---|---|
rate of return r | 3.0% | 3.0% |
NPV | 5113.9 M€ | 1075.3 M€ |
B/C | 2.4 | 1.3 |
IRR | 8.9% | 5.0% |
PBP | 21 years | 13 years |
ALTERNATIVES | ||||
---|---|---|---|---|
Objectives | Criteria | Scenario 1 Stand Alone | Scenario 2 New Highway | Scenario 3 New Highway + Green-Way |
ROAD EFFICIENCY | Investment costs | 0.085 | 0.003 | 0.000 |
Management and maintenance costs | 0.048 | 0.000 | 0.000 | |
SERVICE PROVIDER BENEFITS | Revenues from traffic | 0.000 | 0.055 | 0.065 |
Business diversification | 0.000 | 0.013 | 0.027 | |
USERS BENEFITS | Cars generalized transport costs | 0.000 | 0.065 | 0.072 |
Freight vehicles generalized transport costs | 0.000 | 0.060 | 0.089 | |
ECONOMIC GROWTH | Tourism promotion: accessibility to inland historical centres | 0.000 | 0.044 | 0.044 |
Tourism promotion: a new green-way road | 0.000 | 0.024 | 0.048 | |
Freight centres and firms accessibility | 0.000 | 0.065 | 0.065 | |
QUALITY OF LIFE | Accidents risk | 0.000 | 0.086 | 0.102 |
Social cohesion | 0.000 | 0.016 | 0.032 | |
LOCAL ENVIRONMENTAL BENEFITS | Pollution | 0.000 | 0.080 | 0.085 |
Protected environmental areas | 0.000 | 0.036 | 0.045 | |
Hydrological risk areas | 0.000 | 0.045 | 0.056 | |
Landslide risk areas | 0.000 | 0.045 | 0.056 | |
GLOBAL ENVIRONMENTAL BENEFITS | Global warming | 0.000 | 0.024 | 0.028 |
Disability-Adjusted Life Years and usage of primary resources | 0.000 | 0.024 | 0.028 | |
TOTAL (SUM OF THE IMPACTS) | 0.133 | 0.683 | 0.840 | |
% VARIATION WITH RESPECT TO THE BEST SCENARIO | 532% | 23% | 0% |
CBA vs. MCA | ||
---|---|---|
CBA result: % variation of the B/C ratio estimated for the first-best alternative (Scenario 3) and the second-best ones (Scenario 2) | 6% | |
MCA result: % variation of the weighted overall impacts produced by the first-best alternative (Scenario 3) and the second-best ones (Scenario 2) | 23% | |
Benefits percentage incidence | CBA | MCA |
User’s Benefits | 85.6% | 20.7% |
Non-Users Benefits | 14.4% | 79.3% |
ECONOMIC Growth (Tourism, Firms) | - | 20.20% |
| 11.90% | 17.20% |
| 1.90% | 31.20% |
| 0.50% | 10.70% |
TOTAL | 100.0% | 100.0% |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Henke, I.; Cartenì, A.; Di Francesco, L. A Sustainable Evaluation Processes for Investments in the Transport Sector: A Combined Multi-Criteria and Cost–Benefit Analysis for a New Highway in Italy. Sustainability 2020, 12, 9854. https://doi.org/10.3390/su12239854
Henke I, Cartenì A, Di Francesco L. A Sustainable Evaluation Processes for Investments in the Transport Sector: A Combined Multi-Criteria and Cost–Benefit Analysis for a New Highway in Italy. Sustainability. 2020; 12(23):9854. https://doi.org/10.3390/su12239854
Chicago/Turabian StyleHenke, Ilaria, Armando Cartenì, and Luigi Di Francesco. 2020. "A Sustainable Evaluation Processes for Investments in the Transport Sector: A Combined Multi-Criteria and Cost–Benefit Analysis for a New Highway in Italy" Sustainability 12, no. 23: 9854. https://doi.org/10.3390/su12239854
APA StyleHenke, I., Cartenì, A., & Di Francesco, L. (2020). A Sustainable Evaluation Processes for Investments in the Transport Sector: A Combined Multi-Criteria and Cost–Benefit Analysis for a New Highway in Italy. Sustainability, 12(23), 9854. https://doi.org/10.3390/su12239854