See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/382456918

Selection of an international distribution center location: A comparison

between stand-alone ANP and DEMATEL-ANP applications

Article in Research in Transportation Business & Management · October 2024

DOI: 10.1016/j.rtbm.2024.101135

CITATIONS READS

4 62

4 authors, including:

Victoria Muerza Emilio Larrodé

University of Zaragoza University of Zaragoza
38 PUBLICATIONS 225 CITATIONS 77 PUBLICATIONS 471 CITATIONS

SEE PROFILE SEE PROFILE

Nebojsa Bojovic
University of Belgrade
119 PUBLICATIONS 1,113 CITATIONS

SEE PROFILE

All content following this page was uploaded by Victoria Muerza on 22 July 2024.

The user has requested enhancement of the downloaded file.

 Research in Transportation Business & Management 56 (2024) 101135

Contents lists available at ScienceDirect

Research in Transportation Business & Management

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

Selection of an international distribution center location: A comparison

between stand-alone ANP and DEMATEL-ANP applications
Victoria Muerza a, *, Milos Milenkovic b, Emilio Larrodé c, Nebojsa Bojovic b
a
Zaragoza Multicriteria Decision Making Group (GDMZ), Area of Quantitative Methods for Economics and Business, Department of Applied Economics. Faculty of
Economics and Business, University of Zaragoza, 50005 Zaragoza, Spain
b
Faculty of Transport and Traffic Engineering, University of Belgrade, 11010 Belgrade, Serbia
c
Transport Engineering and Infrastructure Area, Department of Mechanical Engineering. Faculty of Engineering and Architecture, University of Zaragoza, 50018
Zaragoza, Spain

A R T I C L E I N F O A B S T R A C T

Keywords: The problem of Distribution Center (DC) strategic location is critical since it impacts the company's overall
International distribution center distribution strategy and enhances supply chain resilience. This paper compares and evaluates five locations in
Location selection Europe for the possible establishment of an international distribution center. The Analytic Network Process
Multiple criteria decision making (MCDM)
(ANP) and Decision-Making Trial and Evaluation Laboratory (DEMATEL)-Based ANP techniques were applied in
ANP
the analysis, considering 25 criteria across seven dimensions. This paper aims to provide a decision-making
DEMATEL- ANP (DANP)
framework for prioritizing distribution center locations in Europe, identifying key criteria, and ranking alter­
native locations to guide decision making processes for stakeholders and policy makers involved in DC location
selection. The results demonstrated that both approaches rank the alternatives similarly, although they assign
varying degrees of importance to decision criteria. The research was constrained by a limited number of alter­
natives and respondents, as well as imprecision in human judgments. Future research will explore additional
sustainability and social criteria, more alternative locations, and incorporate fuzziness for a more comprehensive
selection of the optimal International Distribution Center (IDC) location.

1. Introduction the optimal location of a DC. The environment, which consists of society,
culture, and infrastructure, is of crucial importance for the functionality
Nowadays the global economy is characterized by a fragile post- of the location. The competitive position of a location may be dimin­
pandemic recovery and dimmed by the ongoing war in Ukraine. Sup­ ished by cultural differences, unavailability of skilled labor, lack of space
ply chains are faced with inventory and shipping issues, while inventory and facilities, undeveloped infrastructure, as well as unfavorable tax
centralization and distribution remain a challenge (Gürbüz, Muerza, system and custom procedures. In that case, apparently short distances
Marchiori, & Zangiacomi, 2021). To mitigate these problems, and make may be less time-efficient than longer distances in an area with better
supply chains more resilient, the definition of the proper location of attributes from the point of view of the above-mentioned criteria. Other
distribution centers (DCs) is necessary, which has a fundamental impact criteria influencing DC location are political stability and territorial di­
on the distribution strategy of a company. mensions (Ayadi, Hamani, Kermad, & Benaissa, 2021; Kumar & Anba­
International Distribution Centers (IDCs) are used for global logistics nandam, 2020), interconnected business activities and network
operations. Their location is a key factor as it facilitates the optimization connections (Elevli, 2014), geographical and physical closeness
of the supply chain, reducing costs and delivery times. The optimal (Erkayman, Gundogar, Akkaya, & Ipek, 2011), regional characteristics
location of the DC can reduce transportation costs and decrease inbound and the price of connection to different means of transport (Kampf,
and outbound lead times, which will facilitate inventory control and Průša, & Savage, 2011), population and handling capacity of seaports
thus increase the level of service. Several criteria, apart from lead time (Önden, Acar, & Eldemir, 2018). Due to the heterogeneity in the ty­
and cost reduction, must be taken into consideration while determining pology of criteria affecting the decision, Multiple Criteria Decision

* Corresponding author.
E-mail addresses: [email protected] (V. Muerza), [email protected] (M. Milenkovic), [email protected] (E. Larrodé), [email protected]
(N. Bojovic).

https://doi.org/10.1016/j.rtbm.2024.101135
Received 10 August 2023; Received in revised form 21 April 2024; Accepted 14 May 2024
Available online 12 June 2024
2210-5395/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-
nc/4.0/).
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

Making (MCDM) methodologies are revealed as a suitable approach to and their performances are compared from the standpoint of IDC loca­
be applied. tion selection. Third, a detailed assessment of five regions with the
Considering that the dimensions/attributes are interdependent, the highest intensity of logistics activities in Europe is included: North-
Analytic Network Process (ANP) is selected as fa tool that addresses Rhine Westphalia region (Germany), Aragon region (Spain), South
decisions involving dependency and feedback. In the context of ANP Holland region (The Netherlands), Hauts-de-France region (France), and
applications it is assumed that the problem's network structure is known Emilia-Romagna region (Italy).
a priori, however the decision maker may encounter difficulty in readily The rest of the paper is structured as follows. Following the intro­
defining this structural framework (Milenković, Val, Lutovac, Bojović, & duction, a comprehensive literature review is presented in Section 2.
Knežević, 2021). The second issue is related to the quantification of Section 3 describes the methodology for the identification and selection
influence in ANP models. Namely, in ANP, the influence between of the most important criteria, as well as for the selection of the preferred
criteria is quantified by pairwise comparisons in which in­ IDC location in Europe, based on the application of the ANP and inte­
terdependencies are treated as reciprocal values (Golcuk & Baykasoglu, grated DEMATEL and ANP approaches. Section 4 contains an analysis of
2016), which does not correspond to the real situation. Furthermore, the results, as well as a sensitivity analysis. Section 5 presents a dis­
even though ANP allows for an assessment of influence and interde­ cussion of the results and implications for management. The last section
pendence, it may lead to a lack of understanding on the part of decision contains concluding remarks and directions for further research.
makers. Therefore, as an improvement of the ANP approach for the DC
location selection problem, DEMATEL (Decision Making trial and eval­ 2. Literature review
uation laboratory) has been integrated with ANP as one of the most
popular causal dependency models. DEMATEL allows for a better un­ The DC location problem has been studied from different perspec­
derstanding of influences by analyzing elements in cause-and-effect re­ tives and methodologies. A classification of approaches includes (Oka­
lationships (Ortíz, Felizzola, & Isaza, 2015). In the integrated tan, Peker, & Baki, 2019): (i) mathematical modeling; (ii) heuristic-
DEMATEL- ANP (DANP) approach, DEMATEL first assesses the causal meta-heuristic methods; (iii) MCDM techniques; (iv) fuzzy logic; and
relationships of the dimensions/attributes by decomposing them into (v) qualitative methods. Thus, for example, Li and Zhou (2021) pro­
groups of causes and effects. The derived causal relationships are then posed a multi-objective model for determining the location of cold-chain
used to determine the attribute weights by applying the ANP method. logistics distribution center. Zhang, Chen, She, and Li (2021) developed
DEMATEL allows the resolution of complex and interdependent prob­ a bi-level programming model for determining the optimal location of a
lems (Fontela & Gabus, 1976). Meanwhile, ANP is a generalization of cold chain distribution center in Wuhan (China) taking into account the
AHP (Saaty, 2008) and avoids the limitations related to the interde­ total social cost of the logistics system as well as the cost incurred by
pendence of nodes at different levels of the hierarchy (Saaty & Vargas, each logistics user. Dupas, Deschamps, Taniguchi, Qureshi, and Hsu
2006). (2023) modeled the problem of optimal location of urban consolidation
In this paper, ANP and DNP have been utilized to determine the most centers (UCCs) in the city of Bordeaux (France) as a multicommodity
optimal location in Europe for an IDC among five regions known for network flow problem. The objective of the resulting linear program­
their heavy logistics activity. ming model was to select UCCs based on minimal costs and CO2 emis­
DC location approaches have been mainly implemented at regional/ sions of a two-tier distribution system.
country level, e.g. Puška, Štilic, and Stevic (2023), Kumar and Anba­ Strutynska, Aftanaziv, Strogan, and Ortynska (2018) proposed an
nandam (2020), Ayadi et al. (2021); or company case study, e.g. Liang, optimization model for regional DCs. The location decision considered
Verhoeven, Brunelli, and Rezaei (2021), in different applications, e.g. as population density, goods turnover, distance to customers, infrastruc­
refugee camps (Abikova, 2020), offshore out-sourcing (Zhou & Xiao, ture, and transport in the region. Zhen, Sun, Wang, and Zhang (2019)
2019), wind farms (Gigović, Pamučar, Božanić, & Ljubojević, 2017), developed a heuristics tabu search algorithm to determine the optimal
ammunition depots (Gigović, Pamučar, Bajić, & Milićević, 2016), freight location and scale of facilities in a closed-loop supply chain network.
village locations in Turkey (Karaşan & Kahraman, 2019). Nonetheless, Zhang and Yin (2017) proposed a genetic algorithm-based method to
as far as current knowledge extends, no study has been devoted to a solve the logistics center (LC) selection problem, where the objective
comprehensive multicriteria analysis for DC location selection at a function was set up with the minimum total cost. In addition, an
global level, particularly focusing on the location of an IDC in Europe. application of qualitative methods to the location problem can be seen in
The research gap is identified based on a combination of recognizing Essaadi, Grabot, and Fénies (2016). Taouktsis and Zikopoulos (2024)
contextual challenges (complexity of supply chain dynamics, trans­ proposed a hybrid approach based on Deep Neural Network (DNN) and
portation and logistics challenges, sustainability, and environmental Farthest Insertion (FI) algorithm to obtain a DC location with minimal
concerns, among others), acknowledging the importance of specific aid distribution distance to support humanitarian logistics networks.
research questions, and undertaking comprehensive analyses to fill gaps MCDM-based research includes the application of different ap­
in existing literature related to IDC location selection. The research proaches. For instance, Kampf et al. (2011) used the Weighted Sum
questions (RQ) designed to fill gaps in existing literature are focused on Approach for the location of public LCs in the Czech Republic. Mean­
the following inquiries: while, Meidutė and Raudeliūnienė (2011) based their approach on a
three-stage process: identification, assessment, and estimation of
• RQ1. Which are the main dimensions/criteria that can be used to external and internal factors of LC location. The complex multicriteria
evaluate the location of an IDC in Europe? assessment method was used as a decisional tool. Peker, Baki, Tanyas,
• RQ2. Is there any difference in the assessment of the problem using and Ar (2016) applied the ANP and a Benefits, Opportunities, Costs &
ANP and DE-MATEL-ANP in five real locations in Europe? Risks (BOCR) analysis in a case study in Trabzon, Turkey. Pamučar,
Dorović, Božanić, and Ćirović (2012) developed a Decision-Making Trial
By addressing these RQs, the paper contributes to the previous and Evaluation Laboratory – Multi Attributive Ideal-Real Comparative
literature in three ways. First, the criteria/dimensions that should be Analysis (DEMATEL-MAIRCA) model for locating a multimodal LC by
considered by stakeholders in the IDC location selection process are the Danube River. Liang et al. (2021) applied the best-worst method
identified and, through an assessment of their influences and in­ (BWM) for the selection of an inland terminal location. Similarly, Özmen
terrelationships, their relative weights are determined. This RQ and Aydoğan (2020) used the BWM in combination with the evaluation
implicitly addresses the question of where an IDC should be located in based on distance from average solution (EDAS) method to rank
Europe. Second, two multicriteria decision making methods for the different LC locations in Kayseri, Turkey.
location of International Distribution Centers-ANP, and DANP-are used The literature focused on fuzzy applications is extensive. For

2
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

example, Chen (2001) proposed a solution to the location problem under 3. Methodology
a fuzzy setting. Five criteria were taken into consideration: “investment
cost, expansion possibility, availability of acquirement material, human re­ The research methodology proposed in this section provides a
sources, and closeness to demand market”. Batanović, Petrović, and Pet­ structured and systematic framework for addressing the research ques­
rović (2009) developed three new algorithms built on the search among tions and providing valuable insights into the problem of determining
possible location nodes. The authors applied comparison operations on optimal locations for an IDC in Europe. More specifically, applied
discrete fuzzy sets to choose the best depot locations. Turskis and research methodology:
Zavadskas (2010) presented an Additive Ratio Assessment (ARAS) Fuzzy
Method for the location selection problem. Erkayman et al. (2011) • Provides a thorough examination of research problems through the
applied a Fuzzy Technique for Order of Preference by Similarity to Ideal identification of criteria for IDC location selection and application of
Solution (F-TOPSIS) approach in the eastern Asia Minor region of advanced MCDM approaches for location analysis.
Turkey for LC location selection. Meanwhile, Kuo (2011) developed an • Ensures that the selected criteria are based on a combination of
F-DANP method for IDC location selection, considering six major sea­ relevant literature review as well as on existing knowledge and
ports in Asia. Similarly, Karaşan and Kahraman (2019) proposed an F- practical insights from industry professionals.
DANP-TOPSIS model for freight village location selection for the city of • Involves an analysis of five of the most logistically intensive regions
Istanbul. in Europe which enhances its relevance and applicability to real-
The methodology proposed by Li, Liu, and Chen (2011) considered world scenarios.
two stages. The first was based on Axiomatic Fuzzy Sets (AFS) to eval­ • Through the use of ANP and DEMATEL-ANP, provides a systematic
uate the location of the LC. The final selection was carried out using evaluation of multiple criteria and their relationships leading to
TOPSIS. Another method was provided by Elevli (2014). The author greater robustness in decision-making outcomes.
applied a Fuzzy Preference Ranking Organization Method for Enrich­
ment Evaluation (F-PROMETHEE) to the evaluation of five possible lo­ The research methodology was divided into two main steps. The first
cations for the establishment of a LC in Samsun, Turkey. More recently, one consisted of the identification and selection of criteria for the
Ayadi et al. (2021) compared the location rankings provided by F- location of IDCs (RQ1). A number of dimensions and criteria were
PROMETHEE and F-MAIRCA after applying the fuzzy full consistency identified based on relevant literature and expert opinions from the lo­
method (F-FUCOM) to the weighting of the criteria identified for the gistics sector. The dimensions were associated with different sets of
location of a logistics platform with a sustainability perspective. The criteria that best describe the IDC location problem. The second step
sustainability perspective was also adopted by Kumar and Anbanandam involved a location analysis in five logistics-intensive regions of Europe
(2020). In contrast, the authors adopted an intuitionistic fuzzy inte­ to comply with RQ2. For this purpose, a MCDM approach based on ANP
grated Analytic Hierarchy Process (IF-AHP) and TOPSIS approach to this and DANP was developed, a performance analysis was carried out, and
end. Four Multimodal freight terminal (MFT) locations in India were findings and conclusions were derived. The following sections describe
evaluated. the procedure in detail.
Pham, Ma, and Yeo (2017) proposed an F-Delphi-TOPSIS approach
from the logisticians' viewpoint in Vietnam. Fuzzy Analytic Hierarchy
Process is combined with spatial statistics in the research provided by 3.1. Identification and selection of criteria for the location of IDCs
Önden et al. (2018). Yazdani, Chatterjee, Pamučar, and Chakraborty
(2020) applied data envelopment analysis (DEA), full consistency The construction of the hierarchical structure and its evaluation was
(FUCOM) and combined compromise solution (CoCoSo) methods under carried out in collaboration with a group of fifteen logistics experts. The
a rough setting for the LC location selection in five Spanish autonomous experts were part of a network of collaborators in different projects in
communities. Recently, Agrebi and Abed (2021) developed a fuzzy the supply chain field (e.g. Lead, Next-Net, Inspire): 1 manager of a LC
multi-attribute, multi-actor decision making approach and applied it to management entity, 2 logistics managers of a user company, 3 managers
a company interested in the selection of a new DC location. The of logistics operators, 1 full Professor of transportation, 3 researchers in
contribution of Alidrisi (2021) is based on considering effectiveness and logistics and supply chain management, 1 expert in multicriteria deci­
efficiency in the selection process. Kieu, Nguyen, Nguyen, and Ho sion making and logistics, 3 experts in logistics infrastructure, 1
(2021) focused on a model based on Spherical Fuzzy AHP and CoCoSo to administration representative.
support the distribution location selection problem of perishable agri­ The participation of the experts in the research is described in more
cultural products. Önden and Eldemir (2022) proposed an iterative so­ detail as follows: (1) A preliminary list of dimensions (D) and criteria (C)
lution approach based on fuzzy AHP and GIS's spatial and network (10 and 66, see Table 2) was proposed by the research team (the Analyst
analysis capabilities to determine the locations of logistics centers in Group – AG), as part of a project funded by the Regional Government of
metropolitan areas. Aragon. The data collection process included the evaluation of relevant
More recently, Puška et al. (2023) focused on the problem of data sources from numerous databases of different organizations (World
selecting the location of DCs in a specific district in Bosnia and Bank, Eurostat, national statistical databases, international consultancy
Herzegovina. firms, port authorities, real estate companies, transport operators, etc.)
Table 1 presents a summary of the main contributions and research as well as personal meetings with experts in the field of transport and
gaps covered by the current literature. As can be seen in the table, the logistics.
methods analyzed do not consider a comparison of ANP and DANP, and The dimensions, which are defined as follows, combined the AG's
are developed at company, city, or regional level. The applications have criteria with those of the IDC location:
been predominantly developed within the Asian context, although some
applications have also been provided in European regions. As far as • Regional system attractiveness: logistics incentives of a region as a
current knowledge extends, this is the first paper to focus on the problem result of previous investments in logistics competences that allow for
of locating distribution centers on a global level and, more specifically, a minimization of costs and a better integration in the logistics
on the location of an IDC with consideration of different European system.
countries. • Demographics: related to the characteristics of the population of a
specific region.
• Industry: linked to the peculiarities of the economic and productive
activity of companies in a region.

3
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

Table 1
Summary of the main findings in the literature related to the DC location selection problem.
Reference Problem Method Research gap covered Main contributions

Agrebi and Abed DC location Fuzzy multi-attribute and Include qualitative criteria in the The model restricts the field of study, and
(2021) multi-actor decision- decision process. focus only on the selection between the best
making approach alternativesThis row stands out from the
rest. Please unify and remove the gray
background and the separator stripe. In
addition, in the preview view, this
reference is repeated

Alidrisi (2021) DC location PROMETHEE II and Data Develop a productivity evaluation The author considers effectiveness and
envelopment analysis model efficiency in the selection process
Ayadi et al. (2021) Logistics Platform selection in F-FUCOM; F-PROMETHEE; Definition and development of fuzzy The paper studies the impact of limited and
Republic of Tunisia F-MAIRCA compensatory and partially weak sustainability on the location
compensatory composite indicators for problem. Economic and environmental
locating a logistics platform from a criteria were found to be the most
sustainable perspective. important for the location selection
Batanović et al. Generic location for LCs in a Fuzzy set theory Comparison between alternatives Three network-covering problems are
(2009) distribution company consider demand at nodes are formulated. Decision under uncertain
imprecise and uncertain. environment
Chen (2001) Urban DC location in a company Fuzzy set theory Formulation of the location problem Criteria selected by the company. Decision
that mimics, as closely as possible, the is made under fuzzy environment
reality of a broad class of facility
location tasks.
Dupas et al. (2023) Optimal location of urban Mixed-integer Generic evaluation method based on a Determination of the best configuration of
consolidation centers (UCCs) in the programming (MIP) MIP model to test scenarios the distribution network with respect to
city of Bordeaux (France) implementing UCCs compared to the certain design or use choices such as the
classical distribution scenario (i.e., capacity and the number of UCCs and the
without UCCs). delivery modes.
Elevli (2014) LC location in Turkey F-PROMETHEE The use of the chosen method (F- Results show the use of PROMETHEE is
PROMETHEE) in the selection of a LC suitable when site information is uncertain
location. and there is subjectivity
Erkayman et al. LC selection in Turkey F-TOPSIS The use of the chosen method (F- Decision is made under fuzzy environment.
(2011) TOPSIS) in the selection of a LC The model considers the vagueness of the
location. subjective decision making
Essaadi et al. Survey for hub location selection Literature review Simultaneous assessment of the Structuration of criteria by geographical
(2016) criteria location at national and subnational level and family. A specific list of criteria
level. according to problem typology is not
provided
Kampf et al. (2011) Public LCs location in the Czech Weighted Sum Approach First approach for the location of a A characterization of LC and DC in the
Republic public LC in the Czech Republic Czech Republic is provided
Karaşan and Freight village location selection for DANP-Topsis The use of the chosen method (DANP- Validation of the efficiency of the method
Kahraman the city of Istanbul, Turkey Topsis) in the selection of a freight by the Simple Additive Weighting method
(2019) village location.
Kieu et al. (2021) DC location Spherical Fuzzy AHP and Focus on perishable agricultural Incorporate uncertainty in the decision
CoCoSo methods products process.
Kumar and Multimodal freight terminal IF-AHP and IF-TOPSIS Identification of STEEP criteria for the Results revealed that technical
Anbanandam location under social, technical, location assessment. First study to sustainability is the most important criteria
(2020) economic, environmental, and propose technical and political of Multimodal freight terminal location
political (STEEP) sustainability. sustainability dimensions.
Application in India
Kuo (2011) International DC location in six F-DEMATEL-AHP/ ANP Handle the independent and The proposed method can overcome
Pacific Asia ports dependent criteria that exist problems of the characteristic of criteria
simultaneously in real-life situations. existing simultaneously as dependent and
independent
Li et al. (2011) LC location in cities AFS-TOPSIS An integration of different The model considers town planning and
methodologies to a comprehensive logistics
location decision.
Li and Zhou (2021) Location of cold chain logistics DC A non-dominated sorting A multi-objective location model of Determining the location of cold chain
genetic algorithm II (NSGA- cold chain logistics distribution center logistics DC by considering three objectives:
II) considering carbon emission minimum logistics cost, minimum carbon
emission, and maximum customer
satisfaction
Liang et al. (2021) Inland terminal location selection of BWM The inland terminal location selection Four layers are considered: transport,
the Maersk shipping line problem from the scope of a shipping logistics, infrastructure, location. An
line. efficient inland transport chain benefits all
the involved stakeholders
Meidutė and Assessment of internal and external Complex multiple criteria Assessment of internal and external Identification, assessment, and estimation
Raudeliūnienė factors of competition for LCs assessment factors of competition of logistics of external and internal factors of LC
(2011) establishment centers. location
Önden et al. Suitability degrees of 19 locations in F-AHP, spatial statistics and Approach developed in the framework Spatial suitability is analyzed. The
(2018) the framework of a LC project analysis approaches of a strategic development plan. method detect changes in the decision
Evaluation of public LC locations. environment.
Önden and Location of logistics center in Fuzzy AHP and GIS-based Holistic approach that integrates A multiple facility location decision for
Eldemir (2022) metropolitan areas spatial and network different methods for addressing logistics centers in a metropolitan area is
analysis logistics center problems studied with a comprehensive perspective
(continued on next page)

4
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

Table 1 (continued )
Reference Problem Method Research gap covered Main contributions

Agrebi and Abed DC location Fuzzy multi-attribute and Include qualitative criteria in the The model restricts the field of study, and
(2021) multi-actor decision- decision process. focus only on the selection between the best
making approach alternativesThis row stands out from the
rest. Please unify and remove the gray
background and the separator stripe. In
addition, in the preview view, this
reference is repeated

Özmen and LC location in Kayseri, Turkey BMW-EDAS Application for a specific area in Methodology applied for the first time in LC
Aydoğan (2020) Turkey, proposing administrative location literature.
inference.
Pamučar, Tarle, Multimodal LC selection in eight DEMATEL -MAIRCA First approach considering DEMATEL Hybrid application considering MAIRCA, a
and Parezanovic locations by the Danube River and MAIRCA for the location selection recently developed MCDM methodology
(2018) problem.
Peker et al. (2016) LC site selection in Trabzon, Turkey ANP-BOCR First approach considering ANP/BOCR The model considers the interdependency
method for the location selection relationships among the factors
problem.
Pham et al. (2017) LC location in Vietnam F-Delphi-Topsis Benchmarking framework for the The perspective of logisticians under an
location selection problem based on environment of uncertain information is
the logisticians' perspective. adopted
Puška et al. (2023) DC selection in a specific district in Fuzzy set theory Evaluation of a potential DC location in The model combines fuzzy sets along with
Bosnia and Herzegovina BiH's Brcko District the IMF SWARA and fuzzy CRADIS
methods.
Strutynska et al. Optimization for regional DCs Theory of scheduling and The optimization model considers the The model combines geographic and social
(2018) location. Application to four cities optimization transportation networks and factors. A desirability function is proposed
in Central Ukraine availability of infrastructure in the
vicinity of the possible arrangement.
Taouktsis and Location of DC with minimal aid A hybrid approach based New hybrid approach to apply in real- An innovative approach with the
Zikopoulos distribution distance in on Deep Neural Network world humanitarian logistics network perspective of improving the effectiveness
(2024) a humanitarian logistics network (DNN) and Farthest scenarios to develop a robust and efficiency of humanitarian logistics
Insertion (FI) algorithm distribution infrastructure decision-making processes
Turskis and LC location ARAS-F First approach considering ARAS-F for The utility function value is directly
Zavadskas the location selection problem. proportional to the relative effect of
(2010) weights of the considered criteria
Yazdani et al. LC location in the Spanish R-DEA, R-FUCOM, R- First integrated evaluation approach A ranking of the most suitable communities
(2020) autonomous communities CoCoSo for logistics center selection. is provided. The model can be used for the
formulation of warehousing and
transportation strategies
Zhang and Yin LC location MATLAB genetic algorithm First approach using a genetic The algorithm solves the problem to
(2017) toolbox algorithm for the location selection minimize total transportation costs
problem.
Zhang et al. (2021) Location of a cold chain distribution Bi-level linear Determining the location of low- A scientific basis regarding decisions
center in Wuhan (China) programming and cloud carbon competitive distribution concerning the location of new cold chain
particle swarm centers by combining low-carbon and distribution centers by the government, and
optimization competitive location ideas for energy conservation improvement
and carbon emissions reduction by the
whole society.
Zhen et al. (2019) Location of (re)manufacturers and Stochastic Mixed integer First approach jointly considering the The scale of the facilities is included on the
LCs in China non-linear programming location and scale of facilities. decision model. Inspection should be
model, and an improved performed at producer locations
tabu search heuristic
algorithm

• Infrastructure: referring to the basic structures and resources avail­ dimensions and criteria on a scale ranging from 0 (very unimportant), 3
able in a region that support the functionality of firms and (important), 5 (fair), 7 (important) to 10 (very important) (Chen, Hsu, &
households. Tzeng, 2011). There were allowed intermediate values (1, 2, 4, 6, 8, 9)
• Logistics information: provides information to manage goods and representing intermediate preferences between the scales considered. As
track delivery throughout the supply chain. an additional outcome of this phase of the analysis, other criteria pro­
• Workforce: associated with the characteristics of the people available posed by the experts, which were not initially considered by the AG,
in a region to support business activity. were obtained.
• Transport costs: the monetary value associated with the different (3) The AG obtained the preferred subset of dimensions/ criteria
means of transport. using Triangular Fuzzy Numbers (TFN) from the experts' answers. A TFN
• Global shipments: related to the characteristics of the processes is a triplet (a,b,c), where a (smallest likely value) and c (largest likely
associated with the global shipment of goods. value) membership functions are 0, and b (most probable value) mem­
• Taxes: the monetary imposition that governments set on the global bership function is 1 (Pamučar et al., 2012). In the analysis, the mini­
shipping of goods. mum threshold was 7.5 (a = 6.5, b = 7 (important in the scale), c = 7.5)
• Sustainability: regulation and impact of transport activity in a (Chen et al., 2011). The preferred list of dimensions and criteria ac­
region. cording to the expert's opinion (those with a mean of 7.5 and above) is
shown in Table 3. A total of 7 dimensions and 25 criteria were selected.
(2) The set of dimensions and criteria identified were distributed in a The relationships between the 7 dimensions are depicted in Fig. 1.
questionnaire to the experts to corroborate their importance for inclu­ (4) The selected dimensions and criteria were used as the basis for
sion in the model. Each expert evaluated the importance of the developing the model. A second questionnaire was developed for the

5
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

Table 2 Table 2 (continued )

Dimensions and criteria for IDC location. Dimension Criteria Source
Dimension Criteria Source
Aydoğan, 2020; Pham et al.,
Agrebi & Abed, 2021; Ayadi 2017; Strutynska et al., 2018
Ease of access to seaport et al., 2021; Yazdani et al., Capacity of the airport Karaşan & Kahraman, 2019
2020 Quality of airport services Agrebi & Abed, 2021;
Agrebi & Abed, 2021; Ayadi Essaadi et al., 2016; Karaşan
Ease of access to airport et al., 2021; Yazdani et al., & Kahraman, 2019; Yazdani
2020; Puška et al. (2023) et al., 2020
Agrebi & Abed, 2021; Ayadi Availability of railroads Ayadi et al., 2021; Elevli,
et al., 2021; Erkayman et al., within region 2014; Essaadi et al., 2016;
2011; Kumar & Kampf et al., 2011; Kumar &
Closeness to industrial hubs Anbanandam, 2020; Özmen Anbanandam, 2020;
& Aydoğan, 2020; Pham Karaşan & Kahraman, 2019;
et al., 2017; Yazdani et al., Önden et al., 2018; Özmen
Regional logistics 2020; Kieu et al. (2021) & Aydoğan, 2020; Pham
system Agrebi & Abed, 2021; et al., 2017; Strutynska
attractiveness Erkayman et al., 2011; et al., 2018
Closeness to logistics hubs Kumar & Anbanandam, Capacity of railroads for Karaşan & Kahraman, 2019
2020; Yazdani et al., 2020; freight transportation
Kieu et al. (2021) Quality of railroad services Agrebi & Abed, 2021;
Logistics potential of the Essaadi et al., 2016; Karaşan
region (industrial demand for & Kahraman, 2019; Özmen
Authors contribution
logistics services/indirect & Aydoğan, 2020; Yazdani
businesses) et al., 2020
Karaşan & Kahraman, 2019; Level of intermodal Karaşan & Kahraman, 2019;
Intensity of transport flows in Kumar & Anbanandam, transportation Kumar & Anbanandam,
relation to different modes of 2020; Li et al., 2011; Özmen 2020; Özmen & Aydoğan,
transport (traffic congestion) & Aydoğan, 2020; Peker 2020;
et al., 2016 Cost of land acquisition Agrebi & Abed, 2021; Ayadi
Kampf et al., 2011; Önden (square meter) et al., 2021; Elevli, 2014;
Population density et al., 2018; Strutynska Kumar & Anbanandam,
et al., 2018 2020; Li et al., 2011; Özmen
Resident buying power per & Aydoğan, 2020; Peker
Kampf et al., 2011
capita et al., 2016; Pham et al.,
Demographics
Residents' average age Authors contribution 2017; Turskis & Zavadskas,
Active population Authors contribution 2010; Yazdani et al., 2020;
Kumar & Anbanandam, Kieu et al. (2021); Puška
Level of political stability in
2020; Özmen & Aydoğan, et al. (2023)
the region
2020; Cost of Rent Authors contribution
Level of industry diversity Elevli, 2014 Availability of land Essaadi et al., 2016; Özmen
Import to export ratio Authors contribution & Aydoğan, 2020;
Ratio of inbound to outbound Availability of pre-built Authors contribution
Authors contribution
freight volumes facilities
Industry
Number of carriers in the Land purchasing cost Agrebi & Abed, 2021; Ayadi
Authors contribution
region et al., 2021; Elevli, 2014;
Number of retailers in the Kumar & Anbanandam,
Authors contribution
region 2020; Li et al., 2011; Özmen
Essaadi et al., 2016; Kampf & Aydoğan, 2020; Peker
Availability of highways et al., 2011; Karaşan & et al., 2016; Pham et al.,
within the region Kahraman, 2019; Strutynska 2017; Turskis & Zavadskas,
et al., 2018 2010; Yazdani et al., 2020;
Agrebi & Abed, 2021; Ayadi Puška et al. (2023)
et al., 2021; Elevli, 2014; Ö Structure and ownership of Elevli, 2014; Özmen &
Connectivity of highways to
nden et al., 2018; Pham land Aydoğan, 2020; Peker et al.,
logistics hubs
et al., 2017; Yazdani et al., 2016
2020; Puška et al. (2023) Possibility of expansion Agrebi & Abed, 2021; Ayadi
Essaadi et al., 2016; Karaşan et al., 2021; Chen, 2001;
Quality of highways & Kahraman, 2019; Özmen Elevli, 2014; Karaşan &
& Aydoğan, 2020; Kahraman, 2019; Kumar &
Ayadi et al., 2021; Elevli, Anbanandam, 2020; Liang
2014; Essaadi et al., 2016; et al., 2021; Özmen &
Infrastructure Kampf et al., 2011; Karaşan Aydoğan, 2020; Peker et al.,
Availability of seaports
& Kahraman, 2019; Önden 2016; Pham et al., 2017;
within the region
et al., 2018; Özmen & Turskis & Zavadskas, 2010;
Aydoğan, 2020; Pham et al., Yazdani et al., 2020; Puška
2017; Strutynska et al., 2018 et al. (2023)
Capacity of the seaport Karaşan & Kahraman, 2019 Security in the area Agrebi & Abed, 2021; Ayadi
Essaadi et al., 2016; Karaşan et al., 2021; Kumar &
& Kahraman, 2019; Önden Anbanandam, 2020; Özmen
Quality of seaport services
et al., 2018; Özmen & & Aydoğan, 2020; Yazdani
Aydoğan, 2020; et al., 2020; Puška et al.
Ayadi et al., 2021; Elevli, (2023)
2014; Essaadi et al., 2016; Earthquake risk Karaşan & Kahraman, 2019
Availability of airports within
Kampf et al., 2011; Karaşan Ease of tracking shipments Authors contribution
region
& Kahraman, 2019; Önden Logistics Quality of logistics
et al., 2018; Özmen & information information systems used Özmen & Aydoğan, 2020
Workforce
(continued on next page)

6
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

Table 2 (continued ) Table 3

Dimension Criteria Source
Dimensions and criteria selected for the location of an IDC.
Dimension (D) Criteria (C)
Proximity to customers Agrebi & Abed, 2021; Chen,
2001; Liang et al., 2021; • Ease of access to seaport (C1)
Pham et al., 2017; Yazdani • Closeness to logistics hubs (C2)
et al., 2020; Puška et al. • Logistics potential of the region (industrial
(2023) Regional logistics system demand for logistics services/indirect
Level of skilled labor Essaadi et al., 2016 attractiveness (D1) businesses) (C3)
Availability of logistics • Intensity of transport flows in relation to
Authors contribution
educational programs different modes of transport (traffic congestion)
Labor regulations Authors contribution C4)
Chen, 2001; Elevli, 2014; • Level of industry diversity (C5)
Labor availability Kumar & Anbanandam, Industry (D2) • Number of carriers in the region (C6)
2020 • Number of retailers in the region (C7)
Workforce Ayadi et al., 2021; Karaşan • Availability of highways within the region (C8)
& Kahraman, 2019; Kumar • Connectivity of highways to logistics hubs (C9)
Labor costs (€/month) & Anbanandam, 2020; Ö • Quality of highways (C10)
zmen & Aydoğan, 2020; • Capacity of railroads for freight transportation
Infrastructure (D3)
Pham et al., 2017 (C11)
Level of conflicts Authors contribution • Quality of railroad services (C12)
Schedule flexibility Authors contribution • Level of intermodal transportation (C13)
Turnover rate Authors contribution • Cost of land acquisition (square meter) (C14)
Ayadi et al., 2021; Karaşan • Level of skilled labor (C15)
& Kahraman, 2019; Kumar • Availability of logistics educational programs
& Anbanandam, 2020; Liang (C16)
Medium/ long distance Workforce (D4)
et al., 2021; Özmen & • Labor regulations (C17)
Aydoğan, 2020; Peker et al., • Labor availability (C18)
2016; Pham et al., 2017 • Labor costs (€/month) (C19)
Ayadi et al., 2021; Karaşan • Last mile (C20)
Transport costs (D5)
& Kahraman, 2019; Kumar • Main haulage costs (per container) (C21)
& Anbanandam, 2020; Liang • Ease of loading/unloading (C22)
Last mile Global shipments (D6)
et al., 2021; Özmen & • Level of supply chain fluidity (C23)
Aydoğan, 2020; Peker et al., • Tax rate (C24)
Taxes (D7)
2016; Pham et al., 2017 • Possibility of VAT deferment (C25)
Main haulage costs (per
Transport costs Authors contribution
container)
Ayadi et al., 2021; Karaşan
& Kahraman, 2019; Kumar
& Anbanandam, 2020; Ö
Airplane (per kg)
zmen & Aydoğan, 2020;
Peker et al., 2016; Pham
et al., 2017
Ayadi et al., 2021; Karaşan
& Kahraman, 2019; Kumar
& Anbanandam, 2020; Ö
Railway
zmen & Aydoğan, 2020;
Peker et al., 2016; Pham
et al., 2017
Fig. 1. Relationship between the dimensions identified as defined by
Level of documentation Authors contribution
Ease of customs clearing Authors contribution the experts.
Global shipments
Ease of loading/unloading Authors contribution
Level of supply chain fluidity Authors contribution application of ANP and DANP. The dimensions and criteria were
Agrebi & Abed, 2021; Ayadi
structured into matrices. In ANP, the experts were asked to compare
et al., 2021; Karaşan &
Tax rate Kahraman, 2019; Kumar & with respect to the clustering which of the criteria in each of the rows
Anbanandam, 2020; concerning each of the criteria in each of the columns was more
Yazdani et al., 2020 important, and to assess the influence of the chosen criterion using the
Possibility of VAT deferment Authors contribution scale of Saaty (all the elements on the diagonal have the same impor­
Taxes Agrebi & Abed, 2021; Ayadi
et al., 2021; Karaşan &
tance). For example, in Table 4, the experts were asked to identify
Kahraman, 2019; Kumar & whether C5 or C6 is the most important for the industry and to weigh
Fees at ports
Anbanandam, 2020; Özmen that importance.
& Aydoğan, 2020; Yazdani
et al., 2020
Fees at customs Authors contribution Table 4
Ratio of green companies Authors contribution Example of matrices to be filled in by the experts.
Ratio of green logistics
Authors contribution Industry (D2) Level of industry Number of Number of
service providers
Level of regional diversity (C5) carriers in the retailers in the
Authors contribution region (C6) region (C7)
environmental regulations
Sustainability
Agrebi & Abed, 2021; Level of industry Filled by the Filled by the Filled by the
Karaşan & Kahraman, 2019; diversity (C5) experts experts experts
Environmental impact of
Peker et al., 2016; Pham Number of carriers
transportation activities Filled by the Filled by the Filled by the
et al., 2017; Yazdani et al., in the region
2020 experts experts experts
(C6)
Number of
Filled by the Filled by the Filled by the
retailers in the
experts experts experts
region (C7)

7
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

For the application of DANP, the experts were asked to evaluate the structured as the coupling of two parts (Saaty, 2013). The first part in­
interdependency between criteria. For this, they gave their opinion cludes a control hierarchy or a network of criteria and sub-criteria in the
about the impact one criterion has on another using a scale from 0 to 4 system under study, where the goal can be included. The second part
(all the elements on the diagonal have no impact). For example, in identifies the network of influences between elements and groupings.
Table 4, we asked experts to identify whether C5 has an impact on C6 This stands in contrast to the Analytic Hierarchy Process (AHP) which is
and to weigh this impact. characterized by its hierarchical and linear structure wherein the goal
A total of 5 responses were obtained from the same group of experts resides at the top and the alternatives are positioned at the lower levels
in the period between June and September 2022. Responses were ob­ (Saaty & Vargas, 2006). The ANP has emerged in the literature as the
tained through a personal interview and completed surveys from 1 preferred multicriteria analysis method for modeling dependence and
university professor, 2 researchers from eminent research institutes and feedback. More than a thousand papers since 2005 in the SCOPUS
2 representatives of logistics-service providers. database have applied this methodology (Golcuk & Baykasoglu, 2016).
The network structure for the selection of the preferred IDC location Decision networks within ANP consist of clusters, their elements, and
is illustrated in Fig. 2. At the level of dimensions and criteria, the in­ the links between elements (Büyüközkan & Güleryüz, 2016). The links
terrelationships within and between dimensions/criteria (dotted lines) between elements within the same cluster indicate the inner de­
were considered. pendencies of the elements, while the links between a parent element in
one cluster and its dependent elements in another cluster represent the
3.2. ANP and DANP methods outer dependencies. Outer dependencies between two clusters in both
directions are known as feedback (Karpak & Topcu, 2010). However, the
The ANP introduced by Tomas Saaty (Saaty, 1996) presents a treatment of interdependencies in the ANP method is not objectively
framework that treats dependence within a cluster (inner dependence) addressed in relation to the considered decision-making problem. For
and between different clusters (outer dependence). The problem can be that reason, the DEMATEL method is used in combination with ANP. In

Fig. 2. Network structure of the evaluation model.

8
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

contrast to ANP, DANP modifies the pairwise comparisons and forms a interdependence between dimensions and attributes (Step 4.1–4.5):
comprehensive unweighted matrix in which the pairwise comparisons Step 4.1. Calculate the average direct relation matrix (DRM).
are not performed only within clusters, but also for the whole system Calculate the average direct relation matrix A. Experts use a scale
according to the structure of the problem (Milenković et al., 2021). from 0 to 4 (no influence (0), low influence (1), medium influence (2),
Then, the total relation matrices among the clusters are used to weigh high influence (3) and very high influence (4)), to compare dimensions
the appropriate parts of the supermatrix to get the weighted super­ and attributes pairwise in terms of influence and direction. From these
matrix. The weighted supermatrix is then raised to limiting powers to entries, a matrix A of nxn dimensions, known as the average DRM, is
obtain the final priorities (Golcuk & Baykasoglu, 2016). constructed (1). Each element aij of matrix A represents the impact of
The strategy of this paper is to apply the ANP approach for the se­ dimension/attribute i on dimension/attribute j and is calculated as the
lection of an IDC location and then combine DEMATEL with ANP to average of all experts' inputs for a given ij pair of dimensions or
improve the assessment of interdependencies at the dimension and attributes.
criteria level. The steps of both approaches are illustrated in Fig. 3. ⎡ ⎤
a11 … a1j … a1n
The following set of steps (Milenković et al., 2021) describes the ⎢ ⋮ ⋮ ⋮ ⎥
DANP approach for the selection of an IDC location: ⎢ ⎥
Step 1. Conduct a literature survey and interviews with experts
A=⎢ ⎢ ai1 aij ain ⎥
⎥ (1)
⎣ ⋮ ⋮ ⋮ ⎦
directly and indirectly involved in transportation, logistics and supply an1 anj ann
chain-related activities.
Step 2. Design the questionnaire based on the experience, the liter­ Step 4.2. Normalize the DRM. The normalized DRM (NDRM) repre­
ature survey and the expert's knowledge. sents a multiplication of the matrix A and k:
Step 3. Collect expert opinions on the dimensions and attributes and N=Axk (2)
select the preferred subset based on TFNs.
Step 4. Apply the DEMATEL methodology to determine the where

Fig. 3. Evaluation framework: A general overview.

9
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

⎡ ⎤
⎡ ⎤
⎢
⎢ 1 1 ⎥
⎥ ta1111 ⋯ ta111j ⋯ ta111m1 ⎥
k = min⎢ n ⃒ ⃒, n ⃒ ⃒⎥ (3) ⎢
⎢ ⎥
⎣ ∑ ∑ ⎦ ⎢ ⋮ ⎥
maxi ⃒aij maxj
⃒ ⃒ aij ⃒ ⎢ ⋮ ⋮ ⎥
⎢ ⎥
j=1 j=1
Ta11 =⎢ ta11i1 ⋯ ta11ij ⋯ ta11im1 ⎥ (8)
⎢ ⎥
Step 4.3. Determine the total-influence matrix. ⎢
⎢ ⋮ ⋮ ⋱ ⋮ ⎥ ⎥
⎣ ⎦
The total influence matrix T = [tij]nxn is obtained from matrix N. ta11m11 ⋯ ta11m1j ⋯ ta11m1m1
Transition theory is applied and the direct initial influence and the in­
direct influence of all components of the system are summed, therefore, Steps 5.1–5.8 (Fig. 3) contain the ANP methodology for calculating
the total influence is the sum of an infinite series (Lee, Tzeng, Yeih, the relative weights of nodes in the network (Büyüközkan & Güleryüz,
Wang, & Yang, 2013). 2016).
Step 5.1. Construction of the problem network.
T = N + N2 + N3 + ⋯ + Nm = N(I − N)− 1 , when m→∞ (4)
Step 5.2. Determination of the relative weights of the dimensions/
where I is an n x n identity matrix. attributes and construction of pairwise comparison matrix using the
Step 4.4. Obtain the Inter-Relationship Map. scale of Saaty.
The Inter-Relationship Map (IRM) represents a diagram illustrating Step 5.3. Calculate the eigenvalues and eigenvectors of the compar­
the roles that dimensions and attributes have in the evaluation of the ison matrices. In case there are N attributes or criteria (C1, …, Cj, …, Cn)
alternative location of the IDC with the horizontal axis (r + c) and the and the pairwise comparison matrix A = aij, where aij represents the
vertical axis (r-c) (Dalvi-Esfahani, Niknafs, Kuss, Nilaski, & Afrough, relative importance of criteria Ci and Cj. For all i and j it is necessary that
2019). The sum (ri + cj), known as “prominence”, represents the aij = 1 and aij = 1/aji. The row vector average method is used for the
importance that dimension/attribute i plays in the considered decision- normalization of the results. The approximate weight Wi is calculated as
making problem, while the difference (ri - cj), known as “relation”, shows follows:
the net effect that dimension/attribute i contributes to the evaluation of ⎛ ⎞
the IDC location alternative. A positive result of (ri - cj), means that the n ⎜
∑ aij ⎟
⎜∑ ⎟
factor i is a net causer, while a negative result (ri - cj) means that the ⎝n ⎠
j=1 aij
factor i is a net receiver. To obtain a suitable diagram, it is necessary to i=1

establish the threshold value of the level of influence. This threshold Wi = ∀i, j = 1, 2…, n (9)
n
reduces the complexity of the structural relation model implied by the
The largest eigenvalue λmax is obtained from the formula:
matrix T (Lee et al., 2013). Independent factors, i.e., those that have low
prominence and relationship and are relatively unconnected to the 1∑ n
(AW)i
AW = λWλmax = (10)
system can be identified as those that meet r-c < 0 and r + c < mean (Si, n i=1 Wi
You, Liu, & Zhang, 2018).
To draw the IRM it is necessary to calculate the sums of the rows r Step 5.4. Consistency test. The consistency of the pairwise compari­
and columns c of the total relation matrix T: son is evaluated using the consistency index (CI) and the consistency
[ ] ratio (CR):
∑n
r = (ri )n×1 = tij (5) λmax − n
CI = (11)
j=1 n×1 n− 1
[ ]' CI
n
( ) ( )' ∑ CR = (12)
c = cj n×1 = cj 1×n = tij (6) RI
i=1 1×n
If the CR is >0.1, the pairwise comparisons must be repeated. RI
where ri: represents the total effect, both direct and indirect, that represents the random index. Experts can work under consensus or
dimension/attribute i has on the other dimensions/attributes; cj: repre­ individually. In the case of individual judgments, it is necessary to
sents the sum of the jth column, in matrix T represents the total effects synthesize all responses (Aguarón, Escobar, Moreno-Jiménez, & Turón,
both direct and indirect, that dimension/attribute j has received from 2019).
other dimensions/attributes j. The superscript ´denotes transpose. The remaining steps represent the essence of the DANP approach.
The influence matrix T can be divided into TD based on dimensions Step 5.5. Determination of the unweighted supermatrix. The super­
and TA based on attributes. matrix in general form takes the following appearance (Supeekit, Som­
boonwivat, & Kritchanchai, 2016):

(7) (13)

Matrix T11
a represents a matrix of attributes owned by group D1 as
where Dn denotes the nth dimension (cluster), anm refers to the nth
well as the influences with respect to the attributes from dimension D1 attribute in mth dimension and Wij is the principal eigenvector of the
(8). T12
a is a matrix of attributes related to D2 and the influences with
impact of the attributes owned by the jth dimension compared to the ith
respect to the attributes from dimension D2. The rest of the elements of dimension (cluster). In case the jth dimension has no influence, then Wij
the Ta matrix can be described similarly. = [0].
To obtain the unweighted supermatrix, the total influence matrix Ta

10
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

must first be normalized. The normalized total influence matrix Tαa has where tijD represents the sum of influences of the Tija matrix. The
the following form: normalization is performed as follows.
⎡ / ⎤ ⎡ ⎤
/ 1j /
⎢ tD d1 ⋯ tD d1 ⋯ tD d1 ⎥ ⎢ tDα11
11 1n
⋯ tD1j tDα1n ⎥
α
⋯
⎢ ⎥ ⎢ ⎥
⎢ ⋮ ⋮ ⋮ ⎥ ⎢ ⋮ ⋮ ⋮ ⎥
⎢ / / / ⎥ ⎢ ⎥
TDα = ⎢ tDi1 di ⋯ tDij di ⋯ tDin di ⎥ = ⎢ tDαi1
⎢ ⎥ ⎢ ⎥
tDij tDαin ⎥
α
⋯ ⋯
(14) ⎢
⎢ ⋮
⎥ ⎢ ⎥
⎢ ⋮ ⋮ ⎥ ⎢
⎥ ⎢ ⋮ ⋮ ⋱ ⋮ ⎥ ⎥
⎣ n1 / / / ⎦ ⎣ αn1 ⎦
tD tDnj tDαnn
α
tD dn ⋯ tDnj dn ⋯ tDnn dn ⋯ ⋯

(19)

where:
where Tαa11 represents a normalized sum of the influences of factors a11,
n
…, a1m1 related to the attributes belonging to dimension D1 and calcu­ ∑
di = tdij , for i = 1, 2, …, n (20)
lated as follows: j=1
⎡ ⎤
/ 11 / 11 / 11 The weighted supermatrix WW is based on the integration of the
t 11
⎢ a11 1 d ⋯ t 11
d ⋯ t 11
d 1 ⎥
⎢ a1j 1 a1m1 ⎥ unweighted matrix W into the normalized matrix of attribute influences
⎢ ⋮ ⋮ ⋮ ⎥
⎢
⎢ 11 / 11 / / ⎥ TDα according to (17).
11 ⎥
Ta = ⎢ tai1 di
α11
⋯ ta11ij d11 i ⋯ t 11
aim1 d i ⎥ ⎡ ⎤
⎢ ⎥
⎢ ⋮ ⋮ ⋱ ⋮ ⎥
⎢ tD11 × W11 tDα21 × W12 ⋯ tDαn1 × W1n ⎥
⎢ ⎥ α
⎣ 11 / 11 / / ⋯
11 ⎦
tam11 di ⋯ ta11m1j d11 i ⋯ t 11
am1m1 d i
⎢ α
⎢ t 12 ×W 21
tD × W
α22 22
⋮ ⋮
⎥
⎥
⎢ D ⎥
⎢ in ⎥
Ww = ⎢ tDij × Wij ⋯ tD × W ⎥ (21)
⎡ ⎤ α αnj
⋮ ⋯
⎢ ⎥
⎢ taα11
11
⋯ taα1j11 ⋯ taα1m
11
1 ⎥
⎢
⎢ ⋮ ⋮ ⋱ ⋮ ⎥
⎥
⎢ ⎥ ⎣ α1n ⎦
⎢
⎢ ⋮ ⋮ ⋮ ⎥ ⎥ tD × Wn1 tDα2n × Wn2 ⋯ ⋯ tDαnn × Wnn
⎢ tai1 taij ⋯ taαim111 ⎥ (15)
α11 α11
=⎢ ⋯ ⎥
⎢ ⎥
⎢
⎢ ⋮ ⋮ ⋱ ⋮ ⎥ ⎥ Step 5.7. Computation of the limit supermatrix. The limit supermatrix
⎣
taαm11
11
⋯ taαm1j
11
⋯ taαm1m1
11 ⎦ is calculated by raising the weighted supermatrix WW to a sufficient
power k, until the supermatrix has converged to a long-term stable
supermatrix. The vectors of the limit supermatrix represent the relative
where d11
i represents the sum of the influences of factors a11, …, a1m1 weights of each attribute with regards to the evaluation of the IDC
associated with the first dimension (D1): location.
m1
∑
d11 ta11ij , for i = 1, 2, …, m1 (16) lim Wwk (22)
i = k→∞
j=1
Step 5.8. Determination of the final weights of the alternatives. The
The elements ta1111 represent the values of the attribute impacts a11, …, final weights of the alternatives are found in the corresponding columns
a1m1 regarding the attributes pertaining to the dimension D1, while the of the limit supermatrix. The alternative with the highest overall priority
elements taα11
11
are their normalized values. value is selected.
The unweighted matrix W (13) is composed of the normalized values
of the attribute influences Tcαnn which are calculated following the pro­ 4. Case study and results
cedure explained above. The component matrices within the super­
matrix W represent the values of the attribute influences among the This section details the application of the proposed approaches for
different dimensions. For example, matrix W11 represents the values of locating an International Distribution Center (IDC) in Europe. The pro­
the attribute impacts from dimension D1 in relation to the attributes cess of selecting alternatives and determining the criteria values is
from dimension D1 according to (17). described in Subsections 4.1 and 4.2, respectively. Subsections 4.3 and
⎡ ⎤c11 ⋯ c1j ⋯ c1m1 4.4 focus on the application of ANP and DANP for international DC
⎢ taα11
11
⋯ taαj111 ⋯ taαm111 1 ⎥ location selection. Subsection 4.4 provides a summary and comparison
c11 ⎢ ⎥
⋮ ⎢
⎢ ⋮ ⋮ ⋮ ⎥ of the results obtained from ANP and DANP methodologies. The final
⎥
W11 = c1i ⎢
⎢ taα1i11 ⋯ taαji11 ⋯ α11 ⎥
tam1 i ⎥ (17) subsection conducts a sensitivity analysis to assess the robustness of the
⋮ ⎢
⎢
⋮ ⋮ ⋱ ⋮ ⎥
⎥ selected approach.
⎢ ⎥
c1m1 ⎣ ⎦
taα1m1
11
⋯ taαjm1
11
⋯ taαm1m1
11

4.1. Selection of alternatives for international DC location in Europe

Step 5.6. Calculation of the weighted supermatrix. The weighted
supermatrix WW is obtained by the same procedure, employing the Potential alternatives for international DC location are chosen using
normalized total influence matrix TD and the unweighted supermatrix. the Logistics Performance Index (LPI), an interactive benchmarking tool
⎡ ⎤ developed by the World Bank. Updated regularly, the LPI helps countries
⎢ tD11 ⋯ tD1j ⋯ tD1n ⎥ assess their trade logistics performance, identifying challenges and op­
⎢ ⎥
⎢ ⋮
⎢ ⋮ ⋮ ⎥ ⎥ portunities. According to the 2018 global ranking (Arvis et al., 2018),
⎢
TD = ⎢ tDi1 tDij
⎥
tDin ⎥ (18) several countries rank highly on the LPI scale, including Germany
⎢
⎢ ⋮
⎥ (3.83), the Netherlands (3.74), France (3.84), Spain (4.02), and Italy
⎢ ⋮ ⋮ ⎥ ⎥
⎣ n1 ⎦ (4.20) on a scale of 1 to 5. Logistics-intensive regions within these
tD ⋯ tDnj ⋯ tDnn
countries are evaluated as potential IDC locations.
More precisely, the selected alternatives (A) are (Fig. 4):

11
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

• A1. Germany: North Rhine-Westphalia (NRW) represents one of the • C6. Number of carriers in the region (in this research figures are
most important logistics hubs in Europe. NRW plays a key role in the provided at national level).
hinterland traffic of ZARA's seaports (Ports of Zeebrugge, Amster­ • C7. Number of retailers in the region (data provided does not show
dam, Rotterdam, and Antwerp) not only as a transit region, but also the number of shops, but rather the number of companies).
as a region of origin and destination. • C8. Availability of highways within the region measured as number
• A2. The Netherlands: the South Holland region, with the Rotterdam of connections if highways are available.
logistics cluster, represents one of the central hubs for interconti­ • C9. Connectivity of highways to logistics hubs – qualitative indicator
nental maritime import and export flows to and from the European based on the number of highways within the region.
Union. • C10. Quality of highways, measured by the Road Quality Indicator,
• A3. France: the Hauts-de-France region (formerly Northern France), provided by the World Economic Forum.
and its capital Lille, are pivotal in the third-richest consumer area of • C11. The capacity of railroads for freight transportation, measured as
the world. the railway density (km of railway lines per 1000 km2), by NUTS 2
• A4. Spain: the region of Aragon is characterized by its strategic regions.
location at the convergence of the main transportation routes: • C12. Quality of railroad services (1(low) - 7(high)) based on the WEF
Madrid-Barcelona, and the Atlantic-Mediterranean corridor, and its (World Economic Forum) Executive Opinion Survey which includes
connection to the South of France. the opinions of over 14,000 business leaders in 144 countries.
• A5. Italy: the Emilia-Romagna region represents a focal platform and • C13. Level of intermodal transportation measured in the availability
gateway for freight traffic. It is located in one of the most productive of different modes of transportation such as Very high (rail, airport,
areas of Europe. inland shipping, seaport, road and pipeline modes are available) or
High (rail, seaport, airport, road, pipelines modes are available).
The data collection process for each of the criteria considered in the • C14. Cost of land acquisition measured in € per square meter.
model (Table 3) included the evaluation of relevant data sources from • C15. Level of skilled labor (low (1) – high (5)) measured by the
numerous databases of various organizations (Eurostat, World Bank, “Logistics Competence & Quality” of the World Bank on NUTS-
national statistical databases, real estate companies, port authorities, 0 level.
transport operators, international consultancy companies, etc.), as well • C16. Availability of logistics educational programs measured in
as personal meetings with experts in the field of transport and logistics. quality of educational programs based on ranking in Eduniversal
Best Logistics Masters Ranking.
• C17. Labor regulations, measured by the Labor freedom index
4.2. Determining the values of criteria for evaluation of international DC
(0− 100).
location
• C18. Labor availability, measured as the regional unemployment
rate (%).
Values of selected criteria for each alternative are given in Table 5.
• C19. Labor costs measured in € per month.
Valuation of criteria is made as follows:
• C20. Last mile, measured as the cost in € per 100 km.
• C21. Main haulage costs (per container), measured as the freight
• C1. Ease of access to seaport measured as the volume of containers
rates in euros (table shows average costs for export/ import to and
(million TEUs), data provided at national level.
from Shanghai).
• C2. Closeness to logistics hubs measured as the distance to logistics
• C22. Ease of loading/unloading, measured as the efficiency at the
hub and number of logistics hubs in the area.
customs: speed, simplicity, predictability of formalities.
• C3. Logistics potential (industrial demand for logistics services/in­
• C23. Level of supply chain fluidity, measured as the timeliness of
direct businesses) measured as the amount of freight transported by
shipments in reaching destination within the scheduled or expected
road (million ton-kilometers).
delivery time.
• C4. Intensity of transport flows in relation to different modes of
• C24. Tax rate (Corporate (%).
transport measured in % of traffic congestion).
• C25. Possibility of VAT deferment.
• C5. Level of industry diversity, measured by the number of
manufacturing companies.
4.3. ANP for international DC location selection

Results of the execution of Steps 5.1 through 5.8 (Section 3.2) are
given. Table A1 shows the unweighted supermatrix, Table A2 the
weighted supermatrix and Table A3 the limit supermatrix containing
dimensions and criteria (see Tables in Appendix A). Expert Choice™
software was used to obtain the eigenvectors and consistency tests, and
Excel for the calculation of the unweighted, weighted, and limit
matrices. It can be seen that the most important criteria for evaluating
the location of an IDC are C23 (9%), C7 (7.34%), C6 (7.13%), and C5
(6.13%).
The eigenvectors of the five alternatives for the 25 criteria were
obtained from Table 5. Results are provided in Table 6. The final weights
of the alternatives can be extracted from the vector of priorities, which
was aggregated from Table A3, and the vector of priorities provided in
Table 6, as seen in (23).
The results obtained from the model (refer to Eq. 23) indicate the
ranking of the preferred alternatives as follows: A1 (North Rhine-
Westphalia) > A4 (Aragon) > A3 (Hauts-de-France) > A5 (Emilia-
Romagna) > A2 (South Holland) (also depicted in Fig. 5). This ranking
highlights North Rhine-Westphalia as the most favorable location for the
Fig. 4. Locations under study. International Distribution Center (IDC), followed by Aragon, Hauts-de-

12
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

Table 5
Information on the criteria for each of the alternatives considered (source: adapted from Zaragoza Logistics Center, ZLC (2018).
D C A1 (North Rhine-Westphalia) A2 (South Holland) A3 (Hauts-de-France) A4 (Aragon) A5 (Emilia-Romagna)

C1 14.7 13.88 5.9 17 10.86

3 logistics hubs in a 495 km 5 logistics hubs in a radius of 5 logistics hubs in a radius of 3 logistics hubs in a radius of 3 logistics hubs in a radius of
C2
D1 radius 134 km 300 km 150 km 200 km
C3 311,869 44,418 181,400 249,555 137,986
C4 19 19 20 13 18
C5 148,551 1089 133,778 292,302 284,136
D2 C6 34,537 12,000 119,000 58,026 90,000
C7 597 441 392 451 285
C8 5 connections 2 connections 4 connections 6 connections 2 connections
C9 Very high Medium High Very High Medium
C10 5.30 6.40 5.40 5.70 4.40
D3 C11 ≥ 120 98 63 38 63
C12 4.37 4.21 4.00 3.84 3.85
C13 Very high Very high High High High
C14 100 202.5 51.56 92.5 100.5
C15 4.31 4.09 3.84 3.8 3.66
Yes (Ranking 52nd Business Yes (Ranking 21st Business Yes (Ranking 5th Business Yes (ranking 1st Business Yes (Ranking 3th Business
C16
School) School) School) School) School)
D4
C17 53 60 45 58 51
C18 6.3 6.3 8.8 12.4 5.3
C19 4174 4002 4652 3029 3121
C20 120 120 140 213 215
D5
C21 694 / 881 469 / 781 737 / 774 345 / 765 525 / 1069
C22 4.09 3.92 3.59 3.62 3.47
D6
C23 4.39 4.25 4.15 4.06 4.13
C24 30 25 26.50 25 24
D7
C25 Yes Yes Yes Yes Yes

Table 6
Eigenvectors of the location alternatives according to the criteria.
D C A1 A2 A3 A4 A5

Ease of access to seaport (C1) 0.2358 0.2226 0.0946 0.2727 0.1742

Closeness to logistics hubs (C2) 0.1212 0.3030 0.1818 0.2121 0.1818
Regional logistics system attractiveness (D1)
Logistics potential of region (C3) 0.3371 0.0480 0.1961 0.2697 0.1491
Intensity of transport flows in relation to different modes (C4) 0.2135 0.2135 0.2247 0.1461 0.2022
Level of industry diversity (C5) 0.1728 0.0013 0.1556 0.3399 0.3304
Industry (D2) Number of carriers in the region (C6) 0.1101 0.0383 0.3795 0.1851 0.2870
Number of retailers in the region (C7) 0.2756 0.2036 0.1810 0.2082 0.1316
Availability of highways within the region (C8) 0.2632 0.1053 0.2105 0.3158 0.1053
Connectivity of highways to logistics hubs (C9) 0.2500 0.1500 0.2000 0.2500 0.1500
Quality of highways (C10) 0.1949 0.2353 0.1985 0.2096 0.1618
Infrastructure (D3) Capacity of railroads for freight transportation (C11) 0.2941 0.2353 0.1765 0.1176 0.1765
Quality of railroad services (C12) 0.2156 0.2077 0.1973 0.1894 0.1899
Level of intermodal transportation (C13) 0.2273 0.2273 0.1818 0.1818 0.1818
Cost of land acquisition (square meter) (C14) 0.2045 0.1571 0.2244 0.2095 0.2045
Level of skilled labor (C15) 0.2188 0.2076 0.1949 0.1929 0.1858
Availability of logistics educational programs (C16) 0.1739 0.1739 0.2174 0.2174 0.2174
Workforce (D4) Labor regulations (C17) 0.1985 0.2247 0.1685 0.2172 0.1910
Labor availability (C18) 0.1426 0.1426 0.2384 0.3337 0.1426
Labor costs (€/month) (C19) 0.2199 0.2109 0.2451 0.1596 0.1645
Last mile (C20) 0.2632 0.2632 0.1053 0.2105 0.1579
Transport costs (D5)
Main haulage costs (per container) (C21) 0.2248 0.1717 0.2163 0.1589 0.2281
Ease of loading/unloading (C22) 0.2188 0.2097 0.1921 0.1937 0.1857
Global shipments (D6)
Level of supply chain fluidity (C23) 0.2092 0.2026 0.1978 0.1935 0.1969
Tax rate (C24) 0.2299 0.1916 0.2031 0.1916 0.1839
Taxes (D7)
Possibility of VAT deferment (C25) 0.2000 0.2000 0.2000 0.2000 0.2000

France, Emilia-Romagna, and South Holland. priorities (Table B3 of Appendix B). The total influence matrix and the
network relationship map are shown in Fig. 6 and Table 7. It can be seen
that all perspectives of the IDC location selection problem are interde­
4.4. DANP for international DC location selection
pendent. The computations were carried out in Excel.
The causal diagram depicting the total relationship is illustrated in
According to the evaluation framework (Fig. 3), the combined DANP
Fig. 7, derived from the values presented in Table 8, which summarize
approach includes the modification of pairwise comparisons and the
the influences given and received across various criteria.
formation of the unweighted supermatrix (Table B1 of Appendix B)
based on the direct influence matrix where pairwise comparisons are
carried out for the entire problem rather than just at the cluster level.
Dimensions/criteria with positive ri-ci values are known as net
The weighted supermatrix (Table B2 of Appendix B) is obtained by
causers (Chen et al., 2011). Positive values of ri-ci imply that a specific
integrating the unweighted supermatrix and the total influence matrix.
dimension/criterion has a significant impact on other dimensions/
Raising the weighted supermatrix to the limiting power leads to the final

13
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

Fig. 5. Total ranking of alternatives using the ANP approach.

logistics system (D1), transport costs (D5), industry (D2) or global

shipments (D6), or indirectly, as in the case of the workforce (D4) or
infrastructure (D3). Developed infrastructure (D3) positively influences
the attractiveness of the regional logistics system (D1), transport costs
(D5), global shipments (D6) and workforce (D4). Apart from D7 and D3,
all other dimensions represent net receivers, which means that they are
under the strong influence of other dimensions. On the other hand, the
attractiveness of the regional logistics system (D1) has the highest ri-ci
value since it has the strongest influence (11.275 in rD1+cD1). This im­
plies that D1 is the most influencing aspect of the DC location selection
problem, whereas D7 has the weakest relationship with other di­
mensions (8.194 in total sum rD7+cD7).
Fig. 6. Relationships between dimensions.
Similarly, at the level of the regional logistics system attractiveness
(D1), criterion C4 related to the intensity of transport flows is the first in
criteria. Negative ri-ci value characterizes those dimensions/criteria that terms of the index of the strength of influence given and received. The
are highly dependent on other dimensions/criteria. Dimensions/criteria highest ri-ci value is given to criteria C3, which represents the logistics
with negative ri-ci value are known as net receivers. The ri-ci value re­ potential of the region. The strongest influence given and received in D2
flects the degree of relationship between dimensions/criteria. A higher (Industry) is given to criterion C7 (Number of retailers in the region),
ri-ci implies a stronger relationship between a pair of dimensions/ whereas C5 (industry diversity) is the main net causal criterion. Crite­

(23)

criteria. The highest ri-ci value is found for dimension D7 (Taxes). This rion C13 (Level of intermodal transportation) has the highest impact on
dimension has a strong unidirectional impact on all other dimensions. the other criteria in dimension D3 (Infrastructure) compared to criterion
Tax policy contributes significantly to all other aspects of DC location C14 which has the lowest (the least sum of ri-ci). Criterion C12 (Quality
selection, directly, as in the case of the attractiveness of the regional of railroad services) has the greatest direct impact on the other criteria

14
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

Table 7
Total influence matrix: seven dimensions.
D1 D2 D3 D4 D5 D6 D7 Row sum(ri) Column sum (ci) ri + ci ri-ci

D1 0.765 0.859 0.773 0.736 0.819 0.817 0.543 5.313 5.962 11.275 − 0.649
D2 0.957 0.763 0.811 0.796 0.857 0.837 0.591 5.612 5.631 11.243 − 0.018
D3 0.971 0.901 0.667 0.755 0.887 0.850 0.573 5.604 4.906 10.510 0.699
D4 0.737 0.719 0.572 0.489 0.662 0.630 0.434 4.243 4.744 8.987 − 0.501
D5 0.859 0.813 0.709 0.673 0.648 0.763 0.511 4.975 5.393 10.368 − 0.418
D6 0.886 0.825 0.736 0.680 0.801 0.649 0.539 5.117 5.241 10.358 − 0.124
D7 0.786 0.752 0.638 0.616 0.718 0.694 0.399 4.603 3.591 8.194 1.012

(rC12-cC12 = 0.682). Within dimension D4 (Workforce), criterion C19 The relative importance of dimensions and criteria obtained by
(Labor costs) represents the criterion with the highest impact on the DANP is presented in Table 9. The most important dimension for IDC
other criteria. The main net causer is criterion C17 (Labor regulations). location selection is D1 (Regional logistics system attractiveness), while
Dimension D5 (Transport costs) is characterized by a strong impact of the least important is dimension D7 (Taxes). Regarding individual
criterion C21 (Main haulage costs per container). Likewise, in the case of criteria, the five most important are criteria C20 (Last mile), C23 (Level
D6 (Global shipments), criterion C22 (Easy of loading and unloading) of supply chain fluidity), C22 (Ease of loading/unloading), C6 (Number
has a significant impact on the level of supply chain fluidity (C23). In the of carriers in the region) and C21 (Main haulage costs (per container)).
case of D7 (Taxes), criterion C24 (Tax rate) represents the most influ­ On the other hand, criterion C8 (Availability of highways within the
ential factor. region), C16 (Availability of logistics educational programs), C17 (Labor

Fig. 7. Causal diagram of total relationships.

15
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

Table 8 Table 9
Sum of given and received influences on criteria. Relative ranking of dimensions and criteria obtained by limn→∞ (W α )n .
Dimension (i)/Criteria (i) Row Column ri + ci ri-ci Dimension (D) Relative Criteria (C) Relative
sum (ri) sum (ci) weight of weight of
dimension criteria
Regional logistics system
5.313 5.962 11.275 ¡0.649
attractiveness (D1) Ease of access to seaport
0.035
Ease of access to seaport (C1) 2.605 2.013 4.618 0.593 (C1)
Closeness to logistics hubs (C2) 2.395 2.939 5.334 − 0.544 Closeness to logistics hubs
0.052
Logistics potential of region (C3) 2.694 2.092 4.786 0.601 (C2)
Intensity of transport flows in Regional logistics Logistics potential of region
relation to different modes of 2.429 2.964 5.392 − 0.535 system (industrial demand for
0.180 0.038
transport (C4) attractiveness logistics services/indirect
Industry (D2) 5.612 5.631 11.243 ¡0.018 (D1) businesses) (C3)
Level of industry diversity (C5) 2.474 2.164 4.638 0.311 Intensity of transport flows
Number of carriers in the region in relation to different
2.336 2.655 4.991 − 0.319 0.055
(C6) modes of transport (traffic
Number of retailers in the region congestion) C4)
2.482 2.617 5.100 − 0.135
(C7) Level of industry diversity
0.046
Infrastructure (D3) 5.604 4.906 10.510 0.699 (C5)
Availability of highways within Number of carriers in the
2.163 1.621 3.784 0.543 Industry (D2) 0.175 0.068
the region (C8) region (C6)
Connectivity of highways to Number of retailers in the
2.496 1.861 4.357 0.635 0.061
logistics hubs (C9) region (C7)
Quality of highways (C10) 2.094 1.681 3.775 0.413 Availability of highways
0.017
Capacity of railroads for freight within the region (C8)
2.069 1.450 3.519 0.618
transportation (C11) Connectivity of highways to
0.019
Quality of railroad services (C12) 2.104 1.423 3.527 0.682 logistics hubs (C9)
Level of intermodal transportation Quality of highways (C10) 0.018
2.434 2.173 4.608 0.261
(C13) Capacity of railroads for
0.018
Cost of land acquisition (square Infrastructure (D3) 0.120 freight transportation (C11)
1.593 1.495 3.088 0.098
meter) (C14) Quality of railroad services
0.014
Workforce (D4) 4.243 4.744 8.987 ¡0.501 (C12)
Level of skilled labor (C15) 1.300 1.564 2.864 − 0.264 Level of intermodal
0.024
Availability of logistics transportation (C13)
0.769 1.293 2.061 − 0.524
educational programs (C16) Cost of land acquisition
0.014
Labor regulations (C17) 1.473 1.158 2.631 0.314 (square meter) (C14)
Labor availability (C18) 1.428 1.783 3.210 − 0.355 Level of skilled labor (C15) 0.020
Labor costs (€/month) (C19) 1.704 2.009 3.714 − 0.305 Availability of logistics
0.017
Transport costs (D5) 4.975 5.393 10.368 ¡0.418 educational programs (C16)
Last mile (C20) 1.080 2.642 3.723 − 1.562 Workforce (D4) 0.100 Labor regulations (C17) 0.015
Main haulage costs (per container) Labor availability (C18) 0.023
1.851 1.895 3.746 − 0.044
(C21) Labor costs (€/month)
0.026
Global shipments (D6) 5.117 5.241 10.358 ¡0.124 (C19)
Ease of loading/unloading (C22) 0.681 1.895 2.810 − 1.449 Last mile (C20) 0.095
Transport costs
Level of supply chain fluidity 0.161 Main haulage costs (per
2.254 2.640 4.894 − 0.386 (D5) 0.065
(C23) container) (C21)
Taxes (D7) 4.603 3.591 8.194 1.012 Ease of loading/unloading
0.073
Tax rate (C24) 2.638 1.480 4.118 1.158 Global shipments (C22)
0.167
Possibility of VAT deferment (D6) Level of supply chain
1.452 1.256 2.709 0.196 0.094
(C25) fluidity (C23)
Tax rate (C24) 0.051
Taxes (D7) 0.096 Possibility of VAT
0.044
deferment (C25)
regulations), C11 (Capacity of railroads for freight transportation), C14
(Cost of land acquisition (square meter), and C12 (Quality of railroad
services) represent the least important criteria.
(Number of carriers in the region), C5 (Level of industry diversity), and
Now, considering the value of alternatives against each criterion, the
C1 (Ease of access to seaport) (0.0900 > 0.0734 > 0.0713 > 0.0613 >
ranking of alternatives can be calculated. Fig. 8 represents the relative
0.0553) for the ANP approach. (ii) C20 (Last mile), C23 (Level of supply
ranking of the alternatives obtained using the DANP approach.
chain fluidity), C22 (Ease of loading/unloading), C6 (Number of carriers
in the region) and C21 (Main haulage costs (per container)) (0.095 >
0.094 > 0.073 > 0.068 > 0.065) were reported as the most important
4.5. Comparison of results obtained by ANP and DANP
criteria for the DANP approach. Only C6 and C23 are among the top five
in order of importance in both approaches.
Fig. 9 presents the combined results of both approaches. Interest­
The analysis also showed a different ranking of importance of the
ingly, both approaches yielded the same order of preference for the al­
dimensions considered (Fig. 11): (i) D1 (0.1566) > D3 (0.1525) > D6
ternatives, albeit with different weighting. The ranking remains
(0.1458) > D5 (0.1442) > D2 (0.1386) > D7 (0.1327) > D4 (0.1296) for
consistent: North-Rhine Westphalia (A1) > Aragon (A4) > Hauts-de-
the ANP approach; and (ii) D1 (0.1789) > D2 (0.1748) > D6 (0.1668) >
France (A3) > Emilia-Romagna (A5) > South Holland (A2). The quan­
D5 (0.1609) > D3 (0.1196) > D4 (0.1001) > D7 (0.0958) for the DANP
tification of these preferences is as follows: 0.2176 (ANP) > 0.2151
approach. In both approaches, D1 (regional logistics system attractive­
(DANP) for A1, 0.2160 (ANP) > 0.2095 (DANP) for A4, 0.1992 (ANP) >
ness), D6 (global shipments) and D5 (transport costs) have the same
0.1984 (DANP) for A3, 0.1931 (ANP) > 0.1933 (DANP) for A5, and
order of importance (1st, 3rd, 4th positions) in the ranking.
0.1741 (ANP) > 0.1837 (DANP) for A2.
The same ranking of alternatives was obtained; however, different
criteria were identified as the most important (Fig. 10): (i) C23 (Level of
supply chain fluidity), C7 (Number of retailers in the region), C6

16
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

Fig. 8. Total ranking of alternatives by DANP approach.

Fig. 9. Total ranking of alternatives by ANP and DANP approach.

Fig. 10. Total ranking of the most important criteria by ANP and DANP approach.

17
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

4.6. Sensitivity analysis Table 10

Results of the sensitivity analysis.
The stability and validity of the research findings are ensured by a Scenario Applied Alternatives
sensitivity analysis. The sensitivity analysis is performed on five sce­ approach
A1 A2 A3 A4 A5
narios simulating a random sample of n = 1000 cases. The first three
scenarios include varying the criteria weights by +/− 10, 20 and 50%, Scenario ANP 0.2176 0.1741 0.1992 0.2160 0.1931
1 DANP 0.2151 0.1837 0.1983 0.2095 0.1933
respectively. The fourth and fifth scenarios include increasing and Scenario ANP 0.2177 0.1740 0.1992 0.2160 0.1931
decreasing the weights of the five most important criteria (Fig. 10) by 2 DANP 0.2151 0.1836 0.1984 0.2095 0.1934
50%, respectively. The outcomes derived from both ANP and DNP Scenario ANP 0.2177 0.1744 0.1990 0.2158 0.1931
methods demonstrate consistency in all scenarios. The ranking of the 3 DANP 0.2152 0.1837 0.1984 0.2094 0.1932
Scenario ANP 0.2134 0.1648 0.2014 0.2205 0.1999
alternatives remains the same in all cases. The results of the sensitivity
4 DANP 0.2565 0.2203 0.2396 0.2471 0.2340
analysis for all scenarios can be seen in Table 10. Scenario ANP 0.2164 0.1713 0.1997 0.2174 0.1952
5 DANP 0.1737 0.1471 0.1572 0.1718 0.1526
5. Discussion

This paper proposes an application of ANP and DANP to select the haulage costs (per container), ease of loading/unloading, level of supply
preferred location of an IDC in five regions with the highest intensity of chain fluidity, and the possibility of VAT deferment.
logistics activities in Europe: North-Rhine Westphalia region (Ger­ The results showed different priorities of the criteria with different
many), South Holland region (The Netherlands), Hauts-de-France region orders of importance in both approaches. Only the level of supply chain
(France), Aragon region (Spain), and Emilia-Romagna region (Italy). fluidity (C23), and the number of carriers in the region (C6) were among
The initial model was composed of 10 dimensions and 66 criteria. the five most important. However, in terms of dimensions, both ap­
The analysis provided a final set of 25 criteria considered as the most proaches considered regional logistics system attractiveness (D1), global
important for selecting the location of a new IDC grouped into 7 di­ shipments (D6) and transport costs (D5) to be the first, third and fourth
mensions (regional logistics system attractiveness, industry, infrastruc­ most important, respectively. Transport costs were also considered as a
ture, workforce, transport costs, global shipment, and taxes). It is key dimension in Zhang and Yin (2017).
pertinent to acknowledge that while sustainability has garnered Both approaches ranked the region of North-Rhine Westphalia (A1)
increasing attention as a focal point in certain methodologies (e.g., as the most preferable alternative for an IDC location followed closely by
Ayadi et al., 2021; Kumar & Anbanandam, 2020), particularly in the the Aragon region (A4), while the South Holland region (A2) was the
assessment of the impact of transport activities (Agrebi & Abed, 2021; least preferable location for an IDC.
Pham et al., 2017), it has been deliberately excluded from the initial North Rhine-Westphalia is dominant from the aspect of the level of
array of dimensions by experts engaged in this study, as per the meth­ supply chain fluidity (C23), and the number of retailers in the region
odology discussed in Section 3.2.1, within the context of IDC location. (C7) and is also very competitive concerning last mile costs (C20).
Of the 25 criteria used, 9 of them were proposed by the experts who Compared to this alternative, the Aragon region ranked preferred
participated in the evaluation in the framework of an IDC location, and regarding some criteria: C1 (ease of access to seaport), C2 (closeness to
the rest were identified from relevant literature (Table 2). The new logistics hubs), C5 (level of industry diversity), C6 (number of carriers in
criteria identified are as follows: logistics potential of the region, num­ the region), C8 (availability of highways within the region), C10 (quality
ber of carriers in the region, number of retailers in the region, avail­ of highways), C14 (cost of land acquisition), C16 (availability of logistics
ability of logistics educational programs, labor regulations, main educational programs), C17 (labor regulations) and C18 (labor avail­
ability), but it must improve in terms of C3 (logistics potential of the

Fig. 11. Total ranking of dimensions by ANP and DANP approach.

18
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

region), C4 (intensity of transport flows), C7 (number of retailers in the Action could also be taken on D4 (Workforce) by supporting educational
region), C11 (capacity of railroads for freight transportation), C12 programs on logistics (C16) in North Rhine-Westphalia and the South
(quality of railroad services), C13 (level of intermodal transportation), Holland region and improving labor regulations (C17) in the Hauts-de-
C15 (level of skilled labor), C19 (labor costs), C20 (last mile costs), C21 France region. Also, in D7 (Taxation), the Emilia-Romagna region can
(main haulage costs), C22 (ease of loading/ unloading), C23 (level of act on improving the tax rate for companies.
supply chain fluidity), and C24 (tax rate). Both alternatives are equally Finally, the study offers reliable decision-making support tools for
preferred in terms of C9 (connectivity of highways to logistics hubs) and IDC location selection. The DANP method, in particular, offers a higher
C25 (possibility of VAT deferment). level of comprehensiveness in the decision-making process compared to
South Holland region, despite being the least preferred location op­ ANP, enhancing its utility for managers seeking a robust approach in
tion for an IDC in our analysis, ranked first in relation to some criteria: complex decision scenarios.
C2, C10, and C17, and obtained the same importance as A1 in C4. Hauts-
de-France region (A3) obtained the highest score regarding the other 7. Conclusion and further research
alternatives in C4, C6, C14, and C19. Emilia-Romagna (A5) is dominant
in C21 regarding the other alternatives. One of the key factors in global logistics operations is the location of
From a theoretical viewpoint, this paper contributes to the research the IDC because of the impact on costs and delivery times. This paper
on the DC location by identifying nine new criteria already discussed in contributes to literature by providing a decision-making framework for
this section. In addition, the comparison of ANP and DANP has not been prioritizing the location of distribution centers in Europe, to guide the
performed before in this context. The decision-making framework pro­ decision-making process of stakeholders and policy makers involved in
posed in this paper and the results of the analysis are of great importance the selection of DC locations. Two different approaches have been used:
to decision-makers. By considering potential location sites for their IDC, ANP and DANP. Managers should note that the results of this analysis
decision-makers can define which locations are most preferable and may change depending on the situation related to the 25 criteria
which can be less important. This ranking facilitates decision-makers in considered in the evaluation model. The research carried out provides
judiciously allocating resources and financial investments effectively. some practical implications that should facilitate the process of the IDC
Furthermore, the approach provides a general decision-making frame­ location problem solving for decision makers: (i) the identification of the
work that incorporates a comprehensive set of criteria which can be used main dimensions and criteria to be used in the evaluation of an IDC
by managers to plan, analyze, and calculate the relative importance of location in Europe; (ii) the development of two multicriteria ap­
alternative locations for their IDC positioning. proaches, based on ANP and DANP to evaluate the problem; and (iii) the
The results obtained allow the regions analyzed to gain knowledge of application of both approaches in five locations in Europe.
their performance in terms of locating an IDC; the methodology illus­ This research has some limitations: the ranking of the alternatives is
trates how one region can be more competitive than another based on very close, which can be related to the number of alternative locations.
different criteria. The process extracts knowledge that is fundamental In addition, there is vagueness and imprecision in human judgments. In
for the improvement of the regions in the seven dimensions analyzed. future work, additional locations will be explored, and fuzziness and
In terms of robustness, managers can rely on both ANP and DANP fuzzy randomness will be applied in the selection of the best IDC loca­
methods, however, in the case of DANP, managers can rely on a higher tion. Additional social criteria may be considered, such as the quality of
level of comprehensiveness of the decision-making process compared to life of the territory. The final model was evaluated by 5 of the 15
the ANP technique. participating experts, which implies that there might have been small
variations in the criteria weights. Nevertheless, the sensitivity analysis
6. Practical implications carried out ensured the stability and validity of the results.
The final model does not consider sustainability as one of the criteria,
The practical importance of this study is multifaceted. The proposed despite this being a global logistics trend. This may be due to the fact
approach offers stakeholders a decision-making framework for plan­ that the experts participating in the study considered minimizing costs
ning, analyzing, and prioritizing alternative IDC locations. as the main objective. Sustainability may include additional investments
The study identifies a set of 25 most important criteria, grouped into that may increase costs in the short term. Decision-makers may have
seven dimensions which are essential for the selection of an optimal perceived the sustainability dimension as contradictory to efficiency or
location for an International Distribution Center (IDC) in Europe. A profitability.
comprehensive analysis including five logistically most intensive re­
gions provides valuable insights for stakeholders involved in IDC loca­ Funding
tion decision making.
Ranking of criteria and alternative locations help decision-makers to This work was supported by the Regional Government of Aragon
efficiently allocate resources and financial investments based on the under Grant number S35_23R; and the Serbian Ministry of Education
relative importance of alternative locations. and Science under Grant number I36022.
From the aspect of policy making, the approach provides guidelines
for implementing targeted policies and measures to improve specific CRediT authorship contribution statement
criteria in different regions. Recommendations may span from
enhancing infrastructure and workforce to taxation policies based on Victoria Muerza: Conceptualization, Data curation, Formal anal­
identified criteria. In this way, policy makers may understand and define ysis, Investigation, Methodology, Supervision, Writing – original draft,
a set of measures to improve the performance and competitiveness of Writing – review & editing. Milos Milenkovic: Conceptualization, Data
specific regions for attracting IDCs. For example, within dimension D1 curation, Formal analysis, Investigation, Methodology, Supervision,
(Attractiveness of the regional logistics system), the Public Adminis­ Writing – original draft, Writing – review & editing. Emilio Larrodé:
tration should focus on improving the ease of access to the seaport (C1) Supervision, Validation, Writing – review & editing. Nebojsa Bojovic:
in the Hauts-de-France region. In dimension D3 (Infrastructure), actions Supervision, Validation, Writing – review & editing.
should focus on improving the availability of highways (C8) and their
connectivity to logistics hubs (C9) in the South Holland and Emilia-
Romagna regions, investing in highways to improve their quality Declaration of competing interest
(C10) in the Emilia-Romagna region, and improving the capacity of
railroads for freight transport (C11) in the Hauts-de-France region. None.

19
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

Data availability Kieu, P. T., Nguyen, V. T., Nguyen, V. T., & Ho, T. P. (2021). A spherical fuzzy analytic
hierarchy process (SF-AHP) and combined compromise solution (CoCoSo) algorithm
in distribution center location selection: A case study in agricultural supply chain.
Not applicable. Axioms, 10, 53.
Kumar, A., & Anbanandam, R. (2020). Location selection of multimodal freight terminal
Acknowledgments under STEEP sustainability. Research in Transportation Business and Management, 33,
Article 100434.
Kuo, M. S. (2011). Optimal location selection for an international distribution center by
We would like to acknowledge the experts that were part in this using a new hybrid method. Expert Systems with Applications, 38, 7208–7221.
research for providing the data used in this study. Lee, H.-S., Tzeng, G.-H., Yeih, W., Wang, Y.-J., & Yang, S.-C. (2013). Revised DEMATEL:
Resolving the infeasibility of DEMATEL. Applied Mathematical Modelling, 37(10− 11),
6746–6757.
Appendix A. Supplementary data Li, X., & Zhou, K. (2021). Multi-objective cold chain logistic distribution center location
based on carbon emission. Environmental Science and Pollution Research, 28,
32396–32404.
Supplementary data to this article can be found online at https://doi. Li, Y., Liu, X., & Chen, Y. (2011). Selection of logistics center location using axiomatic
org/10.1016/j.rtbm.2024.101135. fuzzy set and TOPSIS methodology in logistics management. Expert Systems with
Applications, 38, 7901–7908.
Liang, F., Verhoeven, K., Brunelli, M., & Rezaei, J. (2021). Inland terminal location
References selection using the multistakeholder best-worst method. International Journal of
Logistics Research and Applications, 1–23.
Abikova, J. (2020). Application of fuzzy DEMATEL–ANP methods for siting refugee Meidutė, I., & Raudeliūnienė, J. (2011). Evaluation of logistics centres establishment:
camps. Journal of Humanitarian Logistics and Supply Chain Management, 10, 347–369. External and internal factors. Bus. Theory Into Practice, 12, 175–182.
Agrebi, M., & Abed, M. (2021). Decision-making from multiple uncertain experts: Case of Milenković, M., Val, S., Lutovac, D., Bojović, N., & Knežević, N. (2021). Evaluation of the
distribution center location selection. Soft Computing, 25, 4525–4544. innovative value proposition for the rail freight transport: An integrated
Aguarón, J., Escobar, M. T., Moreno-Jiménez, J. M., & Turón, A. (2019). AHP-group DEMATEL–ANP approach. Transport, 1–27.
decision making based on consistency. Mathematics, 7, 242. Okatan, B. S., Peker, I., & Baki, B. (2019). An integrated Dematel-ANP-Vikor approach
Alidrisi, H. (2021). DEA-based PROMETHEE II distribution-center productivity model: for food distribution center site selection: A case study of Georgia. JMML, 6, 10–20.
Evaluation and location strategies formulation. Applied Sciences, 11, 9567. Önden, İ., Acar, A. Z., & Eldemir, F. (2018). Evaluation of the logistics center locations
Arvis, J.-F., Ojala, L., Wiederer, C., Shepherd, B., Raj, A., Dairabayeva, K., & Kiiski, T. using a multi-criteria spatial approach. Transport, 33, 322–334.
(2018). Connecting to compete 2018. Trade logistics in the global economy. The Önden, İ., & Eldemir, F. (2022). A multi-criteria spatial approach for determination of the
logistics performance index and its indicators. The International Bank for logistics center locations in metropolitan areas. Research in Transportation Business
Reconstruction and Development/ the World Bank, 2018. Available at: https:// and Management, 44, Article 100734.
openknowledge.worldbank.org/bitstream/handle/10986/29971/LPI2018.pdf. Ortíz, M. A., Felizzola, H. A., & Isaza, S. N. (2015). A contrast between DEMATEL-ANP
Ayadi, H., Hamani, N., Kermad, L., & Benaissa, M. (2021). Novel fuzzy composite and ANP methods for six sigma project selection: A case study in healthcare industry.
indicators for locating a logistics platform under sustainability perspectives. BMC Medical Informatics and Decision Making, 15(3). https://doi.org/10.1186/1472-
Sustainability, 13, 1–37. 6947-15-S3-S3
Batanović, V., Petrović, D., & Petrović, R. (2009). Fuzzy logic based algorithms for Özmen, M., & Aydoğan, E. K. (2020). Robust multi-criteria decision making methodology
maximum covering location problems. Information Sciences, 179, 120–129. for real life logistics center location problem. Artificial Intelligence Review, 53,
Büyüközkan, G., & Güleryüz, S. (2016). An integrated DEMATEL-ANP approach for 725–751.
renewable energy resources selection in Turkey. International Journal of Production Pamučar, D., Dorović, B., Božanić, D., & Ćirović, G. (2012). Modification of the dynamic
Economics, 182, 435–448. scale of marks in analytic hierarchy process (AHP) and analytic network approach
Chen, C. T. (2001). A fuzzy approach to select the location of the distribution center. (ANP) through application of fuzzy approach. Scientific Research and Essays, 7,
Fuzzy Sets and Systems, 118, 65–73. 24–37.
Chen, F.-H., Hsu, T.-S., & Tzeng, G.-H. (2011). A balanced scorecard approach to Pamučar, D. S., Tarle, S. P., & Parezanovic, T. (2018). New hybrid multi-criteria decision-
establish a performance evaluation and relationship model for hot spring hotels making DEMATEL-MAIRCA model: Sustainable selection of a location for the
based on a hybrid MCDM model combining DEMATEL and ANP. International Journal development of multimodal logistics Centre, Ekon. Istraz, 31, 1641–1665.
of Hospitality Management, 30, 908–932. Peker, I., Baki, B., Tanyas, M., & Ar, I. (2016). Logistics center site selection by ANP/
Dalvi-Esfahani, M., Niknafs, A., Kuss, D. J., Nilaski, M., & Afrough, S. (2019). Social BOCR analysis: A case study of Turkey. Journal of Intelligent Fuzzy Systems, 30,
media addiction: Applying the DEMATEL approach. Telematics and Informatics, 43, 2383–2396.
1–14. Pham, T. Y., Ma, H. M., & Yeo, G. T. (2017). Application of fuzzy Delphi TOPSIS to locate
Dupas, R., Deschamps, J.-C., Taniguchi, E., Qureshi, A. G., & Hsu, T. (2023). Optimizing logistics centers in Vietnam: The Logisticians’ perspective. The Asian Journal of
the location selection of urban consolidation centers with sustainability Shipping and Logistics, 33, 211–219.
considerations in the city of Bordeaux. Research in Transportation Business and Puška, A., Štilic, A., & Stevic, Ž. (2023). A comprehensive decision framework for
Management, 47, 1–17. selecting distribution center locations: A hybrid improved fuzzy SWARA and fuzzy
Elevli, B. (2014). Logistics freight center locations decision by using fuzzy-PROMETHEE. CRADIS approach. Computation, 11, 73.
Transport, 29, 412–418. Saaty, T. L. (1996). Decision making for leaders. Pittsburgh, USA: RWS Publications.
Erkayman, B., Gundogar, E., Akkaya, G., & Ipek, M. (2011). A fuzzy Topsis approach for Saaty, T. L. (2008). Decision making with the analytic hierarchy process. International
logistics center location selection. Journal of Business Studies, 7, 49–55. Journal of Services Sciences, 1, 83–98.
Essaadi, I., Grabot, B., & Fénies, P. (2016). Location of logistics hubs at national and Saaty, T. L. (2013). Analytic network process. In S. I. Gass, & M. C. Fu (Eds.), Encyclopedia
subnational level with consideration of the structure of the location choice. IFAC- of operations research and management science. Boston, MA: Springer. https://doi.org/
PapersOnLine, 49, 155–160. 10.1007/978-1-4419-1153-7_32.
Fontela, E., & Gabus, A. (1976). The DEMATEL observer. Battelle Geneva Research Center: Saaty, T. L., & Vargas, L. G. (2006). Decision making with the analytic hierarchy process:
Geneva, Switzerland. Economic, political, social and technological applications with benefits, opportunities, costs
Gigović, L., Pamučar, D., Bajić, Z., & Milićević, M. (2016). The combination of expert and risks. New York, USA: Springer.
judgment and GIS-MAIRCA analysis for the selection of sites for ammunition depots. Si, S.-L., You, X.-Y., Liu, H.-C., & Zhang, P. (2018). DEMATEL technique: A systematic
Sustainability, 8, 372. review of the state-of-the-art literature on methodologies and applications.
Gigović, L., Pamučar, D., Božanić, D., & Ljubojević, S. (2017). Application of the GIS- Mathematical Problems in Engineering, 3696457.
DANP-MABAC multi-criteria model for selecting the location of wind farms: A case Strutynska, L. R., Aftanaziv, I. S., Strogan, O. I., & Ortynska, N. V. (2018). Determining
study of Vojvodina, Serbia. Renewable Energy, 103, 501–521. the sites of optimal location of regional logistics centers. Nauk, 6, 148–155.
Golcuk, L., & Baykasoglu, A. (2016). An analysis of DEMATEL approaches for criteria Supeekit, T., Somboonwivat, T., & Kritchanchai, D. (2016). DEMATEL-modified ANP to
interaction handling within ANP. Expert Systems with Applications, 46, 346–366. evaluate internal hospital supply chain performance. Computers and Industrial
Gürbüz, M. C., Muerza, V., Marchiori, I., & Zangiacomi, A. (2021). Unveiling the Engineering, 102, 318–330.
challenges of future supply chains: An explorative analysis. In R. Fornasiero, Taouktsis, X., & Zikopoulos, C. (2024). A decision-making tool for the determination of
S. Sardesai, A. C. Barros, & A. Matopoulos (Eds.) (1st ed.,Lecture Notes in Management the distribution center location in a humanitarian logistics network. Expert Systems
and Industrial EngineeringNext Generation Supply Chains. A Roadmap for Research and with Applications, 238, 1–17.
Innovation (pp. 103–126). Cham, Switzerland: Springer. Turskis, Z., & Zavadskas, E. K. (2010). A new fuzzy additive ratio assessment method
Kampf, R., Průša, P., & Savage, C. (2011). Systematic location of the public logistic (ARAS-F)v. Transport, 25, 423–432.
centres in Czech Republic. Transport, 26, 425–432. Yazdani, M., Chatterjee, P., Pamučar, D., & Chakraborty, S. (2020). Development of an
Karaşan, A., & Kahraman, C. (2019). A novel intuitionistic fuzzy DEMATEL – ANP – integrated decision making model for location selection of logistics centers in the
TOPSIS integrated methodology for freight village location selection. Journal of Spanish autonomous communities. Expert Systems with Applications, 148, Article
Intelligent Fuzzy Systems, 36, 1335–1352. 113208.
Karpak, B., & Topcu, I. (2010). Small medium manufacturing enterprises in Turkey: An Zaragoza Logistics Center, ZLC. (2018). Estudio de indicadores de la región de Aragón como
analytic network process framework for prioritizing factors affecting success. hub logístico, y benchmark frente a hubs europeos. Aragon exterior (AREX) private
International Journal of Production Economics, 125, 60–70. document.

20
V. Muerza et al. Research in Transportation Business & Management 56 (2024) 101135

Zhang, S., Chen, N., She, N., & Li, K. (2021). Location optimization of a competitive Zhen, L., Sun, Q., Wang, K., & Zhang, X. (2019). Facility location and scale optimisation
distribution center for urban cold chain logistics in terms of low-carbon emissions. in closed-loop supply chain. International Journal of Production Research, 57,
Computers and Industrial Engineering, 154, 1–13. 7567–7585.
Zhang, X., & Yin, Y. (2017). Research on the application of genetic algorithm in logistics Zhou, L., & Xiao, F. (2019). DCM: D number extended cognitive map. Application on
location’. In ICCSS 2017–2017 International Conference on Information, Cybernetics, location selection in SCM. International Journal of Computers Communications, 14,
and Computational Social Systems (pp. 435–438). 753–771.

21

View publication stats