**4. The methodology and the AHP model**

#### **4.1. Methodological aspects in the evaluation of transport projects**

#### *4.1.1. Generalities*

The evaluation of solutions is a key activity within feasibility studies concerning transport projects. Study guidelines and regulations have traditionally focused their attention on the so‐ called "technical and economic feasibility"; in this context, evaluation has been based on the technical, financial, and economic contents or effects of a project, and more specifically, on economic efficiency [2]. With this respect, financial methods have been adapted and used in order to judge the economic efficiency by comparing the cost of the resources employed to produce and manage the infrastructure, and the economic benefits that could be derived from it. This evaluation approach is grounded on the assumption that it is rational to choose the solution that is the best according to one criterion: the economic advantage of the stakeholders.

A standard methodology of this kind is the cost‐benefit analysis (CBA), which is focused on the welfare change of the community that bears the costs and gains the advantages of the transport infrastructure and service. CBA is widely used for infrastructure projects on account of the public expenses expected to realize and run them [3]. In the standard CBA, the capital and running financial costs and other effects (e.g., on the environment, on the community's mobility etc.) are converted into "social opportunity" costs or revenues. Different synthesis indicators, which are based on the monetization of the effects and their inter‐temporal discount (e.g., the economic net present value), can be used to evaluate the economic performance of a project solution (see [4]). The methodology can clearly support the selection of the "best" alternative in a set in that the solutions can be ranked according to their economic performance.

Even if the CBA takes into account a single criterion of evaluation, the methodology aims to assess different types of outcomes, which is always appropriate to projects in the transport sector (see [5]). Indeed, some effects (e.g., the environmental costs) can be reasonably translated into monetary terms and then aggregated into a single measure of economic efficiency. Nonetheless, the methodology becomes complicated and less transparent when it comes to assessing other effects of transport infrastructures (in particular, the "intangible" ones). In addition, some criticism about the principle of welfare theory and the monetization approach, and the need to introduce different dimensions of evaluation have encouraged the use of methods that can explicitly take into consideration several evaluation criteria [6–8].

Alternative "750 m B": The project allows the composition and circulation of trains 750 m long. In this case, an additional opening should be realized to allow the direct long train movement from and to Molo VII, while 650 m trains may use the other openings and reach Molo V and VI with simpler shunting operations.

The alternatives were analyzed with particular reference to the goals of TEN‐T policy, and it was decided to investigate the methods that could be suitable to support the evaluation process ensuring transparency and participation. The methodological aspects and the resulting choice

The evaluation of solutions is a key activity within feasibility studies concerning transport projects. Study guidelines and regulations have traditionally focused their attention on the so‐ called "technical and economic feasibility"; in this context, evaluation has been based on the technical, financial, and economic contents or effects of a project, and more specifically, on economic efficiency [2]. With this respect, financial methods have been adapted and used in order to judge the economic efficiency by comparing the cost of the resources employed to produce and manage the infrastructure, and the economic benefits that could be derived from it. This evaluation approach is grounded on the assumption that it is rational to choose the solution that is the best according to one criterion: the economic advantage of the stakeholders.

A standard methodology of this kind is the cost‐benefit analysis (CBA), which is focused on the welfare change of the community that bears the costs and gains the advantages of the transport infrastructure and service. CBA is widely used for infrastructure projects on account of the public expenses expected to realize and run them [3]. In the standard CBA, the capital and running financial costs and other effects (e.g., on the environment, on the community's mobility etc.) are converted into "social opportunity" costs or revenues. Different synthesis indicators, which are based on the monetization of the effects and their inter‐temporal discount (e.g., the economic net present value), can be used to evaluate the economic performance of a project solution (see [4]). The methodology can clearly support the selection of the "best" alternative in a set in that the solutions can be ranked according to their economic performance.

Even if the CBA takes into account a single criterion of evaluation, the methodology aims to assess different types of outcomes, which is always appropriate to projects in the transport sector (see [5]). Indeed, some effects (e.g., the environmental costs) can be reasonably translated into monetary terms and then aggregated into a single measure of economic efficiency. Nonetheless, the methodology becomes complicated and less transparent when it comes to assessing other effects of transport infrastructures (in particular, the "intangible" ones). In addition, some criticism about the principle of welfare theory and the monetization approach,

are discussed in the next section.

*4.1.1. Generalities*

**4. The methodology and the AHP model**

**4.1. Methodological aspects in the evaluation of transport projects**

120 Applications and Theory of Analytic Hierarchy Process - Decision Making for Strategic Decisions

The methods of multi‐criteria decision analysis (MCDA) can support decisions in which several aspects, points of view, or decision‐makers are relevant. Projects of transport infra‐ structures have generally more than one goal to achieve which are advocated by different interest groups or stakeholders. Indeed, they are typically the subject of "strategic" decision processes that involve several participants with different roles in and power over the process and the final decision. In these multi‐actor decision processes, the participants, explicitly or implicitly, promote their interests and often represent those of other subjects or groups (see, e.g., [9, 10]); in the following, the term "stakeholder" will be used to indicate subjects or groups that participate or are represented in the decision process.

According to a relevant article by Banville et al. [11], the multi‐criteria methods are particularly appropriate for the introduction and the use of the stakeholder concept in supporting deci‐ sions, so that different viewpoints can be explicitly and clearly adopted. In particular, the concept of stakeholder can improve the initial stages of decision‐making, namely problem analysis, identification of the alternatives and of the evaluation criteria. Different stakeholders have different perceptions of the problem that is targeted by the project, its solutions and their outcomes on the mobility system, the community, and the environment; they may appreciate such aspects from different perspectives, and therefore, the levels of importance of the evaluation criteria may vary among them. Bristow and Nellthorp [3] confirmed that MCDA can be more effective than CBA to capture and assess environmental and social impacts, while other authors [7, 12] concluded that MCDA is suitable to taking into explicit account public opinion and making a transparent and participative decision.

MCDA methods give the opportunity to include in the evaluation of the project solutions several points of view and effects, some of which are not easily converted into economic values. They have been successfully applied to decision processes concerning transport systems and are promoted within guidelines on feasibility studies (see, e.g., [13]). A good number of articles have been published that present applications of MCDA to infrastructural projects (e.g., [14– 16]) and transport plans or policies (e.g., [12]). Several studies confirm the above‐mentioned advantages of MCDA with respect to economic‐based methods, but, in line with official guidelines, they usually suggest to carry out both analyses to have a complete picture of a project's effectiveness and efficiency (see, in particular, [7, 8]). As for ports, the studies pay particular attention to logistic and locational effects of these infrastructures and highlight the necessity to involve the port authority and users, governmental institutions, and community stakeholders in the decision process [17, 18].

Traditionally, the stress has been posed on the notion that any action performed on a transport system would involve public or private costs in exchange of benefits to the system's users or external to the system. In addition, in many cases, the decision concerns the choice of one item within a set of pre‐defined alternative solutions. Accordingly, the decision process and the supporting MCDA methodology have often been applied in the framework of a "rational" decision process with the following main features:

The subjects who take part in the process have specific (and possibly stable) roles;

The process can be structured in a sequence of steps;

There is a clear distinction among decision‐making, analysis and evaluation activities;

The object of the decision can be described through a set of properties or attributes that can be individually analyzed and evaluated.

Several authors have proposed guidelines for a structured application of MCDA, which should make the methodology more straightforward and, in the case of projects involving the public interest, like in the energy and transport sectors, could favor the transparency of the process (see, e.g., [15, 19]). The proposed structures or procedures can be more or less detailed but, typically, they include the following steps:

definition and structuring of the problem;

identification and development of the alternative solutions;

definition of a set of criteria for the evaluation;

identification of the stakeholders' system of preference;

evaluation of the alternatives and recommendation to the decision‐maker(s).

It is commonly accepted that the MCDA process is non‐linear and can be subject to several refinement cycles, in particular when several actors are involved. As previously remarked, the participants to decision processes concerning transport projects have usually different roles and decision power that can be broadly subsumed under the following categories:

actual decision‐makers (who can take the final or other key decisions);

stakeholders or their representatives (who can have more or less powerful influence on decision‐makers);

analysts and experts (who may be consulted to support the whole process or to provide highly technical information on specific topics).

In common practice, there is a clear distinction between the decision‐makers and the sub‐ jects who support (or "aid") the decision by means of their technical or scientific expertise. Some authors, however, have made a clearer distinction between "experts," who provide technical advice on specific topics (e.g., experts who carry out mobility analysis), and ana‐ lysts, who analyze, model and facilitate the decision process [20–22]. Furthermore, it has been remarked that the analyst can effectively support the decision‐makers if there is an interaction with them, so that the products of the analysis (specially the models) are under‐ stood and validated.

As previously noted, decisions concerning transport projects or plans are the result of a multi‐actor process in which diverse stakeholders can exercise their influence or decision power with different strength, in different stages and on different objects of the process.

For instance, a public institution can promote a project solution since the pre‐feasibility study and decide to fund it, but it cannot take the final decision about its actual implemen‐ tation. The clear determination of the actual objective of the decision, the involved stake‐ holders, and their role in the process are prerequisites to taking into account the key stakeholders' viewpoints in the application of the evaluation method, and to promoting their participation in the analysis and evaluation activities.

Some studies have indeed underscored the importance of including the stakeholders and their objectives at the beginning of the MCDA, so that it is possible to relate them to the evaluation perspectives, the criteria used in the model and their levels of importance [10, 23, 24]. Furthermore, it has been remarked that using the MCDA methods that enable to add them to the model explicitly, may improve the transparency of the whole exercise [25, 26].

In the case study discussed in this article, the analytic hierarchy process (AHP) was used because it offers some features that can facilitate the inclusion of the above‐mentioned as‐ pects in the decision aiding process. The AHP [27] has been applied since the 1980s to sup‐ port the evaluation and selection of alternative project solutions or plans in the transport field. The effective hierarchical representation of a decision problem and the ability to eval‐ uate quantitative and qualitative properties of the alternatives are just two of the most no‐ table features of the method. Its key steps can be explained by the analyst to the decision‐ maker with limited effort, establishing a solid ground for the mutual understanding and validity of the evaluation process [28].

The evaluation of the alternative infrastructural solutions of the port railway system was carried out by means of an AHP model, which offered an explicit representation of the main stakeholders, the evaluation criteria, and the alternatives. The inclusion of the stake‐ holders in the hierarchy, which has been used in some previous studies [29], allows to take into consideration their judgments in regard to the elements of the lower level (dimensions of evaluation and criteria). The experts of the project team performed the evaluation of the alternatives' performance against the criteria. This is in line with other studies, which af‐ firm that the experts' contribution to the decision process improves the technical soundness of the solutions and can take into account the viewpoints of the users of the transport sys‐ tem [30, 31]. The model and the results are discussed in the next section.

#### **4.2. The implementation of the AHP method**

#### *4.2.1. The main features of the evaluation*

The subjects who take part in the process have specific (and possibly stable) roles;

There is a clear distinction among decision‐making, analysis and evaluation

The object of the decision can be described through a set of properties or attributes

Several authors have proposed guidelines for a structured application of MCDA, which should make the methodology more straightforward and, in the case of projects involving the public interest, like in the energy and transport sectors, could favor the transparency of the process (see, e.g., [15, 19]). The proposed structures or procedures can be more or less detailed but,

evaluation of the alternatives and recommendation to the decision‐maker(s).

and decision power that can be broadly subsumed under the following categories: actual decision‐makers (who can take the final or other key decisions);

It is commonly accepted that the MCDA process is non‐linear and can be subject to several refinement cycles, in particular when several actors are involved. As previously remarked, the participants to decision processes concerning transport projects have usually different roles

stakeholders or their representatives (who can have more or less powerful

analysts and experts (who may be consulted to support the whole process or to

In common practice, there is a clear distinction between the decision‐makers and the sub‐ jects who support (or "aid") the decision by means of their technical or scientific expertise. Some authors, however, have made a clearer distinction between "experts," who provide technical advice on specific topics (e.g., experts who carry out mobility analysis), and ana‐ lysts, who analyze, model and facilitate the decision process [20–22]. Furthermore, it has been remarked that the analyst can effectively support the decision‐makers if there is an interaction with them, so that the products of the analysis (specially the models) are under‐

As previously noted, decisions concerning transport projects or plans are the result of a multi‐actor process in which diverse stakeholders can exercise their influence or decision power with different strength, in different stages and on different objects of the process.

The process can be structured in a sequence of steps;

122 Applications and Theory of Analytic Hierarchy Process - Decision Making for Strategic Decisions

that can be individually analyzed and evaluated.

typically, they include the following steps:

influence on decision‐makers);

stood and validated.

definition and structuring of the problem;

definition of a set of criteria for the evaluation;

identification and development of the alternative solutions;

identification of the stakeholders' system of preference;

provide highly technical information on specific topics).

activities;

The development of the NAPA project and the key role of the evaluation have been discussed in Sections 2 and 3. The evaluation stage can be conceived as a process with specific objectives, resources, and expected results. The relevant characteristics of this process are the following:

overall goal: The project is aimed at finding a configuration of the railway infrastructure in the port that can pursue the goals of TEN‐T policy;

expected output: The result of the evaluation process is the identification of the project solution that will be submitted to the EU for funding;

object of the evaluation: The candidate solutions are given and mutually exclusive (alternatives);

key stakeholders: There is a limited set of stakeholders who can exert a significant influence on the final decision (submission of the project solution);

role of experts: The project team is entitled to support the evaluation stage and provide the technical expertise to assess the performance of the project solutions.

These aspects have significant effects on the way in which the evaluation is performed and the AHP model is built. In the AHP (see, e.g., [32]), the decision problem is structured as a functional hierarchy of connected elements (or "nodes"). The overall goal was located in the top level (focus of the decision) and is broken down into elements that are used to analyze its properties, assess their importance for the decision‐makers and gauge the level at which such properties are found in the available solutions (located at the bottom level). Each element of a level can be connected to one or several elements of the lower levels; stated differently, the elements connected to an upper node ("parent node") share the property represented by that node: it is then possible to judge the relative "priority" (or "importance") of the elements with respect to the property. Relative priorities can be obtained by direct rating or, as suggested by Saaty [33] and in all those cases in which direct assessment is not viable, by pairwise comparing the elements of a level against the parent node. The child nodes of a level that are connected to a parent node can be grouped into a "cluster"; the elements in a cluster should not have priorities that differ by orders of magnitude, and a cluster can contain all or only a number of the nodes of a certain level. Finally, the connected hierarchy enables to synthesize the priorities of the alternatives with respect to the overall goal.

#### *4.2.2. The evaluation model*

The hierarchy built in the case study consists of four levels (**Figure 3**):

first (top) level: overall goal of the decision; second level: main stakeholders; third level: criteria of evaluation; fourth (bottom) level: alternatives.

The upmost node represents the overall goal of the evaluation process, which was stated as "Select the port project alternative that can achieve best the objectives of EU's TEN‐T policy." It is worth noting that the sentence is fairly simple and contains the key aspects of the evalu‐ ation:

the evaluation is aimed at selecting a candidate in a given set of alternatives;

the evaluation solutions are evaluated on the basis of their potential to implement TEN‐T policy.

The overall goal was validated by the members of the project team on the basis that it complies with the main objective to be achieved in the NAPA project. The project team identified the stakeholders that can exert major influence or power on the implementation of the project (see Section 3):

the Regional Government of Friuli Venezia Giulia (FVG), because of its planning and control role over strategic projects in the regional territory;

the Port Authority of Trieste (APT), which plans, co‐ordinates, promotes, and monitors port operations;

the Italian Railway infrastructure manager (RFI), which has the responsibility for the planning and management of the railway infrastructure.

**Figure 3.** Hierarchical model built to support the evaluation.

expected output: The result of the evaluation process is the identification of the

object of the evaluation: The candidate solutions are given and mutually exclusive

key stakeholders: There is a limited set of stakeholders who can exert a significant

role of experts: The project team is entitled to support the evaluation stage and provide the technical expertise to assess the performance of the project solutions. These aspects have significant effects on the way in which the evaluation is performed and the AHP model is built. In the AHP (see, e.g., [32]), the decision problem is structured as a functional hierarchy of connected elements (or "nodes"). The overall goal was located in the top level (focus of the decision) and is broken down into elements that are used to analyze its properties, assess their importance for the decision‐makers and gauge the level at which such properties are found in the available solutions (located at the bottom level). Each element of a level can be connected to one or several elements of the lower levels; stated differently, the elements connected to an upper node ("parent node") share the property represented by that node: it is then possible to judge the relative "priority" (or "importance") of the elements with respect to the property. Relative priorities can be obtained by direct rating or, as suggested by Saaty [33] and in all those cases in which direct assessment is not viable, by pairwise comparing the elements of a level against the parent node. The child nodes of a level that are connected to a parent node can be grouped into a "cluster"; the elements in a cluster should not have priorities that differ by orders of magnitude, and a cluster can contain all or only a number of the nodes of a certain level. Finally, the connected hierarchy enables to synthesize the priorities

project solution that will be submitted to the EU for funding;

124 Applications and Theory of Analytic Hierarchy Process - Decision Making for Strategic Decisions

influence on the final decision (submission of the project solution);

(alternatives);

of the alternatives with respect to the overall goal.

first (top) level: overall goal of the decision;

second level: main stakeholders; third level: criteria of evaluation; fourth (bottom) level: alternatives.

The hierarchy built in the case study consists of four levels (**Figure 3**):

The upmost node represents the overall goal of the evaluation process, which was stated as "Select the port project alternative that can achieve best the objectives of EU's TEN‐T policy." It is worth noting that the sentence is fairly simple and contains the key aspects of the evalu‐

the evaluation is aimed at selecting a candidate in a given set of alternatives;

the evaluation solutions are evaluated on the basis of their potential to implement

*4.2.2. The evaluation model*

TEN‐T policy.

ation:

These "key stakeholders" were explicitly introduced in the hierarchy, and representatives of them were interviewed to elicit the levels of importance of the criteria used in the evaluation. The explicit introduction of the stakeholders in the model is not strictly necessary to use their judgments for weighing the criteria. Nonetheless, it makes their importance or influence over the main goal more transparent, and the input of their judgments of importance, with regard to the elements of the lower levels, easier and clearer. As for the importance of the three stakeholders with respect to the overall goal, after some discussion within the project team, it was decided to assign them equal weights and analyze the sensitivity of the overall priorities to variations of their values.

The third level of the hierarchy is made up of the criteria used for the evaluation. They are grouped into four clusters that represent the four dimensions of evaluation directly related to the EU's objectives for TEN‐T. It is worth remarking that TEN‐T policy proposes four "cate‐ gories of objectives": cohesion, efficiency, sustainability, users' benefits. Each of them is a dimension of evaluation that includes a set of "general objectives" with a common policy goal and are valid for any kind of project concerning the transport network (independently from the transport mode, location, type of action). The general objectives are then translated into "specific objectives" in light of the particular action to be carried out in a portion of the network. The categories of objectives are briefly introduced in the following points [1].

Cohesion can be carried out through the improvement of the accessibility of all EU's regions, the reduction of quality gaps that can affect the infrastructure and a stronger interconnection of long‐distance, regional, and local traffic flows.

Efficiency aims at assuring the continuity of flows and the interconnection and interoperability of transport networks and promotes intermodality and economic efficiency in the production of transport service.

Sustainability aims to provide a transport network that can reduce externalities, preserve sensitive areas, and reduce emissions.

The users' benefits should take into consideration the needs of the users of the transport system including their safety and security.

The categories were used as clusters to group the criteria that are specific to transport projects that aim to improve the Adriatic‐Baltic and Mediterranean corridors within the TEN‐T network. They were selected by the project team among the criteria proposed in the "Baltic‐ Adriatic Core Network Corridor Study" [34] taking into account those directly connected with freight multi‐modal transport. The criteria aim to evaluate the performances of a project alternative in regard to the aspects that are reported here below. It can be noted that all the criteria are aspects to be improved by means of the implementation of the project.

#### *4.2.3. Cohesion*

Interconnection of flows (IC): The development of the interconnectivity of the transport network between long‐haul sections and long‐haul/short‐haul sections.

Standard and quality of the infrastructures (SQ): The improvement of the quality and standards of the transport infrastructure with the aim of complying with the technical requirements (for the railway transport: line speed, train length, axle load).

Regional integration (IR): the development of the accessibility of European regions with particular reference to the most peripheral ones.

#### *4.2.4. Efficiency*

The third level of the hierarchy is made up of the criteria used for the evaluation. They are grouped into four clusters that represent the four dimensions of evaluation directly related to the EU's objectives for TEN‐T. It is worth remarking that TEN‐T policy proposes four "cate‐ gories of objectives": cohesion, efficiency, sustainability, users' benefits. Each of them is a dimension of evaluation that includes a set of "general objectives" with a common policy goal and are valid for any kind of project concerning the transport network (independently from the transport mode, location, type of action). The general objectives are then translated into "specific objectives" in light of the particular action to be carried out in a portion of the network.

Cohesion can be carried out through the improvement of the accessibility of all EU's regions, the reduction of quality gaps that can affect the infrastructure and a stronger interconnection of long‐distance, regional, and local traffic flows.

Efficiency aims at assuring the continuity of flows and the interconnection and interoperability of transport networks and promotes intermodality and economic

Sustainability aims to provide a transport network that can reduce externalities,

The users' benefits should take into consideration the needs of the users of the

The categories were used as clusters to group the criteria that are specific to transport projects that aim to improve the Adriatic‐Baltic and Mediterranean corridors within the TEN‐T network. They were selected by the project team among the criteria proposed in the "Baltic‐ Adriatic Core Network Corridor Study" [34] taking into account those directly connected with freight multi‐modal transport. The criteria aim to evaluate the performances of a project alternative in regard to the aspects that are reported here below. It can be noted that all the

Interconnection of flows (IC): The development of the interconnectivity of the transport network between long‐haul sections and long‐haul/short‐haul sections.

Standard and quality of the infrastructures (SQ): The improvement of the quality and standards of the transport infrastructure with the aim of complying with the technical requirements (for the railway transport: line speed, train length, axle

Regional integration (IR): the development of the accessibility of European regions

criteria are aspects to be improved by means of the implementation of the project.

The categories of objectives are briefly introduced in the following points [1].

126 Applications and Theory of Analytic Hierarchy Process - Decision Making for Strategic Decisions

efficiency in the production of transport service.

preserve sensitive areas, and reduce emissions.

*4.2.3. Cohesion*

load).

transport system including their safety and security.

with particular reference to the most peripheral ones.

Removal of rail and road bottlenecks (CB): The reduction or removal of existing transport bottlenecks in the network or, conversely, the creation of a new bottle‐ neck.

Intermodality (IM): The optimal integration and interconnection of different transport modes, with particular reference to the connections to ports, airports, and rail‐road terminals.

Economic efficiency (EE): The promotion of economically efficient and high‐ quality transport and the efficient use of existing infrastructures.

#### *4.2.5. Sustainability*

Reduction of emissions (RE): The reduction of emissions and noise from the transport activity.

Reduction of externalities (RX): The reduction of external costs of transport and protection for environmentally sensitive areas.

#### *4.2.6. Users' benefits*

Congestion (CG): The reduction of congestion in the transport network.

Safety and security (SS): The enhancement of safe, secure and high‐quality transport standards, for both passenger and freight.

The criteria were then presented to the representatives of the stakeholders within the project team who considered them valid in the specific context. The hierarchical structure enables to relate the set of criteria to the solutions and the key stakeholders. It can be noted (**Figure 3**) that the key stakeholders are connected to the clusters and the criteria. This feature of the model can be practically implemented by means of the supporting software tool that was chosen: SuperDecisions [35]. In this way, it is straightforward to input in the model the stakeholders' pairwise comparisons of importance:

between the clusters;

between the criteria with respect to their cluster.

The relative weight of each criterion is then the synthesis of the judgments of all three stakeholders on both clusters and criteria.

The fourth level includes the four alternative solutions that are described in Section 3, namely:

"Do minimum";

"650 m": project which allows trains 650 m long;

"750 m A": project which allows trains 750 m long, mode A;

"750 m B": project which allows trains 750 m long, mode B.

The alternatives have been analyzed by the project team's experts and they can be characterized and judged with respect to all the criteria. However, considering that the association of the criteria with specific measures of performance is not viable at this stage of the project, it was decided to elicit the alternatives' priorities by pairwise comparing them against the criteria.

The execution of the model produces the overall priority of the alternatives: it is then possible to rank them or identify the most effective. Nonetheless, it can be noted that the criteria of the model do not include a measure that accounts for the investment cost. In fact, as already remarked, the criteria represent the positive outcomes that are expected from the implemen‐ tation of the solutions. As suggested by Saaty (cf. e.g., [32]), the cost of each alternative was introduced after the evaluation of the overall performance. This approach allowed the stakeholders to focus their attention only on the effects of the alternatives when they judged the weights of the criteria. Eventually, the benefit‐cost ratio was used to obtain a measure of resource spending efficiency: the alternatives could then be ranked, from best to worst, in decreasing order of the ratio.

#### *4.2.7. The evaluation activity*

The evaluation activity was based on two methodologies:

face‐to‐face interview with the representatives of the key stakeholders to assess the weights of the clusters and criteria;

group discussion within the experts of the project team to evaluate the perform‐ ances of the alternatives against the criteria.

The first consisted of the following stages:

preparation of the interview;

first‐pass interview;

consistency check and second‐pass interview (if needed);

input data to the model.

Two members of the project team, one of whom assumed the role of facilitator, conducted the interviews. As a first step, the main contents and objectives of NAPA project were recalled and the goals of the evaluation process were remarked. Secondly, the procedure of multi‐criteria evaluation was presented, and the categories of objectives and the criteria were introduced and illustrated. Before the assessment exercise, its main steps were introduced with the help of the propositions reported below:

*"In order to obtain the overall evaluation of each alternative, it is necessary to establish the level of importance ("weight") of each criterion in regard to the main goal: the higher the weight of a criterion, the higher will be that criterion's contribution to the overall performance of an alternative."*

*"The levels of importance are derived from pairwise comparison of the criteria. For each pair of criteria (C1 and C2), you will be asked two questions:*

*Considering criteria C1 and C2, which is more important with respect to the main goal (or are they equally important)?*

*How much more important is that criterion?"*

The second judgment was based on the verbal scale proposed by Saaty that was briefly introduced to the respondents. Each stakeholder's evaluations were recorded in two different tables (see **Tables 1**–**6**): one for the pairwise comparisons of clusters and one for pairwise comparisons of criteria within a cluster. The results were then input to the software SuperDe‐ cisions for the consistency check.

The assessment of the alternatives' priorities against the criteria was the product of group discussions that involved the technical experts of the project team. Their contribution to the decision process assured a good level of technical soundness to the evaluation and the inclusion of the viewpoints of the port users and the promoters of the solutions.

### **5. Results and discussion**

The alternatives have been analyzed by the project team's experts and they can be characterized and judged with respect to all the criteria. However, considering that the association of the criteria with specific measures of performance is not viable at this stage of the project, it was decided to elicit the alternatives' priorities by pairwise comparing them against the criteria. The execution of the model produces the overall priority of the alternatives: it is then possible to rank them or identify the most effective. Nonetheless, it can be noted that the criteria of the model do not include a measure that accounts for the investment cost. In fact, as already remarked, the criteria represent the positive outcomes that are expected from the implemen‐ tation of the solutions. As suggested by Saaty (cf. e.g., [32]), the cost of each alternative was introduced after the evaluation of the overall performance. This approach allowed the stakeholders to focus their attention only on the effects of the alternatives when they judged the weights of the criteria. Eventually, the benefit‐cost ratio was used to obtain a measure of resource spending efficiency: the alternatives could then be ranked, from best to worst, in

128 Applications and Theory of Analytic Hierarchy Process - Decision Making for Strategic Decisions

face‐to‐face interview with the representatives of the key stakeholders to assess

group discussion within the experts of the project team to evaluate the perform‐

Two members of the project team, one of whom assumed the role of facilitator, conducted the interviews. As a first step, the main contents and objectives of NAPA project were recalled and the goals of the evaluation process were remarked. Secondly, the procedure of multi‐criteria evaluation was presented, and the categories of objectives and the criteria were introduced and illustrated. Before the assessment exercise, its main steps were introduced with the help

*"In order to obtain the overall evaluation of each alternative, it is necessary to establish the level of importance ("weight") of each criterion in regard to the main goal: the higher the weight of a criterion, the higher will be that criterion's contribution to the overall*

*"The levels of importance are derived from pairwise comparison of the criteria. For each*

decreasing order of the ratio.

*4.2.7. The evaluation activity*

The evaluation activity was based on two methodologies:

ances of the alternatives against the criteria.

consistency check and second‐pass interview (if needed);

*pair of criteria (C1 and C2), you will be asked two questions:*

the weights of the clusters and criteria;

The first consisted of the following stages: preparation of the interview;

first‐pass interview;

input data to the model.

of the propositions reported below:

*performance of an alternative."*

The results of the evaluation activity are reported in **Tables 1**–**16** starting with the data collected from the face‐to‐face interviews (**Tables 1**–**6**). Each of the three interviews with the key stakeholders' representatives produced two tables: the first shows the results of pairwise comparisons (in terms of importance) between clusters (or "categories of objectives"); and the second shows those between criteria of a cluster. The notation "A 5> B" means that the element A is five times more important ("moderately more important") than B, while the notation "A <5 B" means that the element B is five times more important than A, and "A = B" means that A and B are equally important. The tables show the measure of inconsistency produces by the software for the pairwise comparisons between three or more elements; Saaty remarks that inconsistency can be acceptable if it is not >10%.


**Table 1.** Pairwise comparisons of clusters by FVG.

Only for stakeholder FVG, a second pass interview was necessary, due to the high inconsis‐ tency obtained after the first pass. Arguably, this situation can be attributed to the more limited knowledge of the NAPA project and its relation with TEN‐T policy that the FVG's represen‐ tative had with respect to the other two interviewees. Indeed, while APT and RFI had been involved in the NAPA project at an early stage, FVG was consulted only after the complete definition of the four alternatives. A second possible reason can be the fact that the consultation with FVG was the first of the three, and therefore, the two interviewers could improve their command over the interview flow in the second and third sessions.

The results of the pairwise comparisons of the alternatives against the criteria are reported in **Tables 7**–**16**. A total of 60 comparisons were made by the experts of the project team, which were based on the alternatives' features and expected performances in regard to the criteria. The grounds of the experts' judgments are briefly reported in the following paragraphs.


**Table 2.** Pairwise comparisons between the criteria of a cluster (FVG).


**Table 3.** Pairwise comparisons of clusters by APT.

Interconnection (IC—**Table 7**). The candidate solutions do not have any significant effects in terms of the improvement of the interconnection between long‐haul and short‐haul sections of the Adriatic‐Baltic and Mediterranean corridors.

Standard and quality of the infrastructures (SQ—**Table 8**). Both solutions "750 m A" and "750 m B" comply with the European standard, because they allow the 750 m‐long train operations without disruptions, while "650 m" and "Do minimum" alternatives do not. In fact, "650 m" could deal with long trains but with an important increase in track occupancy and conflicts in the port.


**Table 4.** Pairwise comparisons between the criteria of a cluster (APT).

knowledge of the NAPA project and its relation with TEN‐T policy that the FVG's represen‐ tative had with respect to the other two interviewees. Indeed, while APT and RFI had been involved in the NAPA project at an early stage, FVG was consulted only after the complete definition of the four alternatives. A second possible reason can be the fact that the consultation with FVG was the first of the three, and therefore, the two interviewers could improve their

The results of the pairwise comparisons of the alternatives against the criteria are reported in **Tables 7**–**16**. A total of 60 comparisons were made by the experts of the project team, which were based on the alternatives' features and expected performances in regard to the criteria. The grounds of the experts' judgments are briefly reported in the following paragraphs.

Interconnection (IC—**Table 7**). The candidate solutions do not have any significant effects in terms of the improvement of the interconnection between long‐haul and short‐haul sections

Standard and quality of the infrastructures (SQ—**Table 8**). Both solutions "750 m A" and "750 m B" comply with the European standard, because they allow the 750 m‐long train operations without disruptions, while "650 m" and "Do minimum" alternatives do not. In fact, "650 m" could deal with long trains but with an important increase in track occupancy and

command over the interview flow in the second and third sessions.

130 Applications and Theory of Analytic Hierarchy Process - Decision Making for Strategic Decisions

**Cohesion Efficiency Sustainability**

**Inconsistency (cohesion)** 0.06237 **Inconsistency (efficiency)** 0.06239

**Table 2.** Pairwise comparisons between the criteria of a cluster (FVG).

**Cluster i Cluster j Cohesion** 3> **Efficiency Cohesion** 5> **Sustainability Cohesion** 7> **Users' benefits Efficiency** 3> **Sustainability Efficiency** 5> **Users' benefits Sustainability** 3> **Users' benefits**

**Inconsistency** 0.04381

of the Adriatic‐Baltic and Mediterranean corridors.

**Table 3.** Pairwise comparisons of clusters by APT.

conflicts in the port.

**IC** 5> **SQ CB** <3 **IM RE** = **RX IC** <3 **IR CB** <7 **EE Users' benefits SQ** <7 **IR IM** <5 **EE CG** 3> **SS**


**Table 5.** Pairwise comparisons of clusters by RFI.


**Table 6.** Pairwise comparisons between the criteria of a cluster (RFI).

Regional integration (IR—**Table 9**). "Do minimum" keeps the present configuration of the port railway infrastructure essentially unchanged, and some local studies have observed that it presents some weak points that affect the accessibility to other European regions (e.g., Central and Northern Europe). The other three solutions certainly improve the situation and "750 m B" should have a slightly better performance because it allows a direct connection (without decomposition) between the Port and the main network for long trains.

Removal of rail and road bottlenecks (CB—**Table 10**). The present infrastructure presents several bottlenecks (e.g., the limitation in the number of trains that can be dispatched to the network) that can be reduced by the new projects. However, the alternative "750 m B" is more effective than "650 m" and "750 m A" with this respect, because it may allow a higher number of long trains thus reducing the number of trains given the demand flows.


**Table 7.** Comparisons between the alternatives (IC).


**Table 8.** Comparisons between the alternatives (SQ).


**Table 9.** Comparisons between the alternatives (IR).


**Table 10.** Comparisons between the alternatives (CB).

Removal of rail and road bottlenecks (CB—**Table 10**). The present infrastructure presents several bottlenecks (e.g., the limitation in the number of trains that can be dispatched to the network) that can be reduced by the new projects. However, the alternative "750 m B" is more effective than "650 m" and "750 m A" with this respect, because it may allow a higher number

of long trains thus reducing the number of trains given the demand flows.

132 Applications and Theory of Analytic Hierarchy Process - Decision Making for Strategic Decisions

650 m = 750 m A 650 m = 750 m B 650 m = Do minimum 750 m A = 750 m B 750 m A = Do minimum 750 m B = Do minimum

Inconsistency 0.00

650 m <5 750 m A 650 m <5 750 m B 650 m 3> Do minimum 750 m A = 750 m B 750 m A 7> Do minimum 750 m B 7> Do minimum Inconsistency 0.02752

**Table 7.** Comparisons between the alternatives (IC).

**Table 8.** Comparisons between the alternatives (SQ).

**Table 9.** Comparisons between the alternatives (IR).

650 m = 750 m A 650 m <3 750 m B 650 m 5> Do minimum 750 m A <3 750 m B 750 m A 5> Do minimum 750 m B 7> Do minimum Inconsistency 0.02752

**Regional integration**

**Standard and quality of the infrastructures**

**Interconnection**


**Table 11.** Comparisons between the alternatives (IM).


**Table 12.** Comparisons between the alternatives (EE).

Intermodality (IM—**Table 11**). The infrastructure layout introduced by the alternative "650 m" may pose several problems to land‐sea intermodal transport because the new terminal is quite far from the sea; therefore, the internal road traffic would increase causing potential congestion within the port road network. In this regard, the layout of "Do minimum" is better than that of "650 m"; nonetheless, "750 m A" and "750 m B" increase the capacity of the railway infrastructure and improve the current layout. "750 m B" is better than "750 m A" as it gives the opportunity to manage without disruptions trains 650 m‐long too.

Economic efficiency (EE—**Table 12**). The alternatives "750 m A" and "750 m B" can make the port more attractive to freight forwarders that need to send their goods to Central and Northern Europe, because of the train capacity increase, their compatibility with European standards and the reduction of unproductive times (due to operational disturbances, for instance). "750 m B" is more flexible than "750 m A" as it allows to operate 650 m‐long trains too without disruptions.

Reduction of emissions (RE—**Table 13**) and reduction of externalities (RX—**Table 14**). The performances of each alternative with respect to the two criteria are similar, even if the first one is mainly related to the infrastructure's operation and the second one to its layout and land use. The "Do minimum' alternative is the worst against both criteria, and the "750 m B" is the best because it can assure a larger modal shift from road to railway than "750 m A" and "650 m." These last two are similar with respect to RE, while the layout of "750 m A" is slightly more beneficial than that of "650 m" as far as external costs of transport are con‐ cerned.


**Table 13.** Comparisons between the alternatives (RE).


**Table 14.** Comparisons between the alternatives (RX).

Congestion (CG—**Table 15**). The alternative "750 m B" can produce the most valuable results in terms of reduction of congestion in the port and in the surrounding area as it can significantly increase modal shift from road to railway transport. A similar effect is expected from "750 m A," though to a less extent. "650 m" introduces some additional movements of the trains within the terminal that imply just a slight improvement of train circulation with respect to "Do minimum," which is the worst.

Safety and security (SS—**Table 16**). The modal shift that is expected from alternatives "750 m A" and '750 m B' will provide benefits in terms of safety and security of freight and, because of the reduction in the number of trucks on the network, passengers. "650 m" has a slightly lower performance than the aforementioned two, but it is surely better than the current situation.


**Table 15.** Comparisons between the alternatives (CG).

infrastructure and improve the current layout. "750 m B" is better than "750 m A" as it gives

Economic efficiency (EE—**Table 12**). The alternatives "750 m A" and "750 m B" can make the port more attractive to freight forwarders that need to send their goods to Central and Northern Europe, because of the train capacity increase, their compatibility with European standards and the reduction of unproductive times (due to operational disturbances, for instance). "750 m B" is more flexible than "750 m A" as it allows to operate 650 m‐long trains too without

Reduction of emissions (RE—**Table 13**) and reduction of externalities (RX—**Table 14**). The performances of each alternative with respect to the two criteria are similar, even if the first one is mainly related to the infrastructure's operation and the second one to its layout and land use. The "Do minimum' alternative is the worst against both criteria, and the "750 m B" is the best because it can assure a larger modal shift from road to railway than "750 m A" and "650 m." These last two are similar with respect to RE, while the layout of "750 m A" is slightly more beneficial than that of "650 m" as far as external costs of transport are con‐

the opportunity to manage without disruptions trains 650 m‐long too.

134 Applications and Theory of Analytic Hierarchy Process - Decision Making for Strategic Decisions

650 m = 750 m A 650 m <3 750 m B 650 m 5> Do minimum 750 m A <3 750 m B 750 m A 5> Do minimum 750 m B 7> Do minimum Inconsistency 0.02752

650 m <3 750 m A 650 m <5 750 m B 650 m 3> Do minimum 750 m A <3 750 m B 750 m A 5> Do minimum 750 m B 7> Do minimum Inconsistency 0.04381

**Table 13.** Comparisons between the alternatives (RE).

**Table 14.** Comparisons between the alternatives (RX).

disruptions.

cerned.

**Reduction of emissions**

**Reduction of externalities**


**Table 16.** Comparisons between the alternatives (SS).

**Table 17** (second column) displays the overall priorities of the alternatives that were obtained from the execution of the model. The alternative "Do minimum" is the worst of the four: in fact, congestion and bottlenecks are just the most evident problems that the current infra‐ structure causes and that triggered the project. In addition, the present configuration of the railway infrastructure of the port would not allow its inclusion in the TEN‐T system. This same consequence would follow if the alternative "650 m" were implemented, as it does not comply with the European TEN‐T's standards. Further, the layout and track circuits of this solution would increase the congestion in the port terminal. Conversely, the two alternatives "750 m" assure seamless integration of the port in the Baltic‐Adriatic and Mediterranean corridors, the removal of the current bottlenecks and higher economic efficiency. Still, the solution "750 m B" is more flexible than "750 m A" because it enables to operate trains 650 m‐ and 750 m‐long without increasing the occupancy of tracks or delaying train dispatching.

In order to evaluate the sensitivity of the alternatives' priorities to changes in the stakeholders' weights with respect to the overall goal, some different combinations of weights were tested. The interval of variation was set between 0.2 and 0.6, as these were reputed the minimum and maximum relative levels of importance of the key stakeholders that could be acceptable in the context of the decision. Even if there were some changes in the values of priorities, the overall ranking proved to be stable.

As a final step, the estimated capital costs of the alternatives were taken into consideration (**Table 17**, third column). The benefit‐cost ratio, namely the ratio between an alternative's overall priority and the normalized capital cost, was used as an indicator of efficiency. **Ta‐ ble 17** (fifth column) shows the results: the alternative "750 m B" keeps the first position, while "Do minimum" gains two positions. In fact, this second alternative has a cost that is eight times lower than the capital cost of "750 m B," but B's expense is repaid by its high level of overall benefits. The value of the ratio is low for "650 m," and this seems to suggest that it should not be taken into consideration in the future stages of the project.

At the conclusion of the evaluation, the results were illustrated and discussed with the decision‐makers, who approved the proposal of further investigating and refining the technical contents of solution "750 m B," in order to seize opportunities for funding it.


**Table 17.** Overall results.
