Abstract
In this research, the impact of applications on improving urban logistics was examined using the example of the company EX, with an emphasis on the sustainability of its business. To conduct this research, expert decision-making was used. The model used ten criteria and eight applications. To incorporate uncertainty into this research, an intuitionistic fuzzy approach was used. Based on the obtained CC values, the criteria weights were determined using the SiWeC (Simple Weight Calculation) method, while the WASPAS (Weighted Aggregated Sum Product Assessment) method ranked the applications. The results showed that “Security and data protection” and “System reliability and stability” were the most important criteria, while Application 1 achieved the best results. These results were confirmed by the consistency analysis of the WASPAS method and the sensitivity analysis, which considered 30 scenarios.
1. Introduction
The changes in business resulting from the development of digital systems have shaped digital business models. As a result, e-commerce is expanding, enabling purchases from and deliveries to customers’ homes [1], resulting in an increase in deliveries to residential addresses. Alongside this, urbanization is rising, and city populations are growing [2]. With an increasing number of people and the growing use of digital commerce, a greater focus is being placed on urban logistics.
Logistics companies aim to deliver goods to customers more quickly and make deliveries more affordable by using urban logistics. In this regard, developing logistics supply chains, especially in urban areas, is necessary. Urban logistics is studied primarily due to its high implementation costs, its negative environmental impact resulting from carbon dioxide emissions, and its contribution to traffic congestion and longer delivery times [3]. This is why emerging trends in urban logistics are focusing on greater efficiency and environmental friendliness. Achieving this is possible by applying modern technologies to deliver goods faster while reducing environmental impact [4].
It is necessary to optimize urban logistics, mainly because this type of logistics consumes the most energy in urban areas [5]. The delivery of goods in the last mile, which makes logistics inefficient and costly, generates high expenses [6]. These costs in the last mile range from 13% to 75% of total delivery costs to customers [7]. Besides increasing logistics expenses, last-mile delivery also pollutes the air, raises noise levels, and contributes to traffic congestion [8].
Urban logistics is increasingly focusing on sustainable mobility strategies [9] and the application of new solutions, as traditional logistics models did not adequately consider ecology and its environmental impact. In urban logistics, new solutions are applied to reduce CO2 emissions and avoid city congestion. In this regard, drones, electric vehicles, and bicycles are used in urban logistics [10]. Many efforts are being made to address problems in urban logistics, with greater emphasis on sustainability and environmental protection, such as reducing last-mile delivery costs by optimizing the delivery process.
Optimization in urban logistics depends on modern digital technologies, with various applications playing distinct roles [11]. The use of these applications in urban logistics improves route planning, reduces CO2 emissions, and shortens delivery times. Their significance is broad, ranging from reducing logistics costs to decreasing environmental pollution by cutting CO2 emissions from delivery vehicles to using electric delivery vehicles that do not emit CO2 [12], and from making efforts to use applications in logistics, especially for last-mile delivery.
1.1. Research Motivation
The challenge with applying applications in urban logistics, especially for last-mile delivery, is that there are many options on the market. This raises the question of which application to choose to achieve the best results in this logistics. It is important to consider that these applications must be evaluated from multiple perspectives. Therefore, it is essential to balance various criteria, including economic, environmental, technical, social, and others, when selecting applications. It is important to highlight that the subjectivity of the decision-maker (DM) during the selection process is often overlooked. DMs must make a choice and, in doing so, may doubt their decision. This doubt is always present when the human factor is involved in decision-making. This is because the DM can always ask them whether the selected application is the most suitable for a specific logistics company. As a result, decision-making uncertainty is present, linked to this doubt. Therefore, it is necessary to develop a reliable model to assist in choosing an application for urban logistics needs that considers various criteria, which motivates this research.
1.2. Research Objectives and Questions
Given the numerous problems in urban logistics, particularly in last-mile delivery, it is logical to optimize the delivery of goods to end customers. This research will focus on the company EX, which handles express mail services. This company delivers goods that customers order from online stores, and the volume of shipments it must deliver is increasing daily. This growth is attributed to the rising popularity of e-commerce among consumers, resulting in an increase in shipments that express mail companies must deliver. The company EX aims to be innovative and technology-driven, and has electric vehicles in its fleet to reduce the environmental impact of urban logistics. In this context, the company has decided to optimize goods delivery and is considering introducing an application to support this effort. To evaluate potential applications, EX will use multi-criteria decision-making methods (MCDMs). Additionally, to address uncertainty in subjective decision-making, an intuitionistic fuzzy set (IFS) will be used. From the above, the primary objective of this research is to assess digital technologies for optimizing sustainable urban logistics, particularly for last-mile delivery, by incorporating uncertainty into the decision-making process using IFS. The company’s task is to select an application that will help achieve this goal and enhance its business. To accomplish this, the following research questions are set:
- -
- How to determine the importance of criteria for evaluating applications in urban logistics?
- -
- How can uncertainty be incorporated into decision-making?
- -
- Which applications have the best indicators for the company EX?
- -
- How stable are the research results in the sensitivity analysis?
1.3. Research Contributions
This research contributes to the development of theory and practice for understanding urban logistics in the context of business digitalization. First, examining the importance of the last mile in urban logistics helps advance theory and provides guidelines for improving this critical segment. Second, the development of methodology enhances the theoretical understanding of the importance of criteria in selecting applications for urban logistics. This includes establishing the theoretical foundations for the individual importance of sub-criteria and for integrating sustainable criteria into the application selection process. In addition to analyzing individual criteria, this study also deepens the theoretical understanding of decision-making under uncertainty. Here, a model for selecting an application is developed based on research criteria, and therefore, guidelines are established for future similar studies. Third, incorporating uncertainty into decision-making is achieved through the IFS-based uncertainty analysis framework. Applying the IFS approach, uncertainty is integrated through the formation of crisp values, so using classical MCDM methods, values that are weighted with a certain level of uncertainty are used. Therefore, the contribution of this research lies not in applying the hybrid MCDM method, but in developing a decision-making model and incorporating uncertainty into the decision-making process. Fourth, the model’s practical application enables the evaluation of urban logistics solutions, with a focus on last-mile delivery. Selecting the most suitable application will help the EX company improve its operations by optimizing routes, reducing carbon dioxide (CO2) emissions, and ensuring timely delivery. Fifth, the findings of this research will provide guidelines for future studies and help shape strategies for implementing sustainable urban logistics, thereby reducing air pollution and traffic congestion in cities. Additionally, these results can support the development of smart cities by providing a foundation for selecting appropriate logistics types for specific situations.
2. Literature Review
The sustainability of urban logistics has been examined in numerous studies. Particular attention in these studies is focused on the last-mile delivery problem, which is crucial for every logistics company. To address the problems in this logistics, various applications are used. When applying these applications, the issue of selecting the most suitable application to implement sustainable urban logistics arises. Therefore, this section focuses on the last mile, the use of digital technologies, and the application of MCDM methods to select applications in urban logistics.
2.1. Sustainable Last-Mile Logistics
The importance of last-mile logistics has been addressed in numerous studies; therefore, only a select few of these studies and segments of this logistics will be covered in this section. Lauenstein and Schank [13] systematized 87 studies on green business and the sustainability of last-mile delivery. They stated that the challenges of sustainable last-mile logistics are not only for logistics companies but also require action from consumers and government bodies. Technological solutions are needed to make the last mile more sustainable and reduce environmental impacts. The authors Ramírez-Villamil et al. [14] analyzed previous research on urban logistics, reviewing 201 studies. They approached these studies through problem modeling, solution development, multimodal transport and efficiency measures, and distribution network sustainability. This provided guidelines for improving urban logistics and enhancing its sustainability. Similarly, Ferreira and Esperança [15] argue that electric vehicles should be used for last-mile delivery and that AI optimization should be implemented in urban logistics. This approach can help logistics companies reduce delivery times, improve energy efficiency, and lower CO2 emissions. The importance of electric vehicles in sustainable urban logistics is also discussed in Alsaleh’s [16] research, which shows that their use, along with autonomous vehicles, can reduce CO2 emissions, lower costs, increase efficiency, and meet consumer needs. Therefore, efforts must be made to electrify vehicles in logistics companies, especially for last-mile delivery. On the other hand, van Vliet et al. [17] focused on consumers and their preferences, applying simulation models to reduce CO2 emissions per delivered package.
Based on this research, it can be said that barriers to urban transport in urban logistics have the greatest impact on pollution in urban settlements. Therefore, various studies have been conducted to address problems in freight transport in urban areas. Williamsson and Moen [18] studied the role of digitalization and electrification of the delivery fleet and identified the factors that delay these innovations. They found that obstructions exist in the freight distribution sector due to the influence of various interests and interest groups. Kervall and Pålsson [19] identified 11 categories of problems affecting urban parcel delivery through a review of previous research. Additionally, they provided guidelines for making urban delivery sustainable. Meanwhile, Lee et al. [20] examined the institutional problems that impact transport within urban logistics. Their research showed that traffic congestion results from the imbalance caused by the expansion of urban centers and insufficient planning for that growth. This results in longer delivery times and increased air pollution due to traffic congestion. Route planning based on space-time relationships was also investigated by Wang et al. [21], who used existing graph networks. Deweerdt and Fabre [22] share this view, emphasizing that improved urban transport planning is necessary to reduce the impact of urban logistics on climate change. Bastida-Molina et al. [23] advocate for solving this problem by electrifying vehicles. On the other hand, Sinesio et al. [24] highlight the use of autonomous vehicles for last-mile delivery to reduce environmental impacts. On the other hand, Xing et al. [25] considered the use of bicycles for last-mile delivery. Based on this literature review, it is evident that electrifying logistics companies’ fleets is crucial for reducing environmental pollution, particularly in urban areas. Furthermore, these solutions must be applied when planning deliveries in urban regions.
2.2. Digital Technologies and Applications in Urban Logistics
Urban logistics management in developing digital solutions relies on specific applications. Teixeira et al. [26] introduced the SOLFI program, used to optimize urban logistics processes. This program reduces traditional goods flows, promotes greater use of public transportation, and supports bicycle delivery for the last mile. As a result, the number of vehicles in urban areas decreases, thereby reducing CO2 emissions. Fioravanti et al. [27] highlight congestion and air pollution in urban logistics and suggest the use of collaborative platforms. Kolasińska-Morawska et al. [28] focus on how mobile technologies can support the implementation of urban logistics, using smart city residents as a case study to foster sustainable growth. Diglio [29] addresses last-mile delivery challenges and emphasizes crowdsourcing as an alternative. They also explore three platforms across different dimensions as potential applications in urban logistics within Brazil. Zhang & Demir [30] examine parcel lockers as a digital technology solution that makes last-mile delivery more sustainable. Bouti et al. [31] provide a literature review of digital technologies in urban logistics and the application of Industry 4.0 principles to this sector, highlighting their role in supply chain management.
Further development of digital technology in urban logistics followed the introduction of machine learning and, later, artificial intelligence (AI). Bhat et al. [32] emphasize that a large amount of information must be processed in last-mile delivery, and therefore, it is necessary to use AI in urban logistics management. AI can also make predictions based on historical data, thereby reducing congestion during the delivery of goods. In line with the application of AI in urban logistics, there is a gradual transition to Industry 5.0, which should highlight social and environmental sustainability, especially in last-mile delivery [33]. Ferreira and Esperança [15] specifically emphasize the need to apply innovative technologies to reduce costs and CO2 emissions. On the other hand, with the development of AI, vehicle fleet optimization can be achieved, leading to improved shipment delivery and more efficient vehicle route planning [34]. Therefore, the emphasis is on applying digital technologies in urban logistics to reduce costs and environmental impact.
2.3. Application of MCDM Methods in Urban Logistics
To improve urban logistics, decisions must be made by evaluating different options and considering specific alternatives. Therefore, MCDM methods are applied in urban logistics decision-making. The following table displays research that used MCDM methods in urban logistics (Table 1).
Table 1.
Application of MCDM methods in urban logistics.
The application of intuitionistic fuzzy sets was previously used in urban logistics. Uyanik et al. [50], who used the DEMATEL, TOPSIS, and VIKOR (ser. Multiple Criteria Optimization Compromise Solution) methods to determine logistics centers. Gerasimenko and Kureichik [51] investigated cost reductions in port use within urban logistics. Additionally, this approach was also applied to the last-mile delivery problem. Kanık et al. [46] prioritized strategies using SWOT analysis and using the FUCOM and COPRAS methods. Ayyildiz and Erdogan [52] examined the potential of autonomous delivery robots for last-mile delivery using the PIPRECIA (Pivot Pairwise Relative Criteria Importance Assessment) method. Bilgili et al. [53] evaluated last-mile delivery strategies with the TOPSIS method. Based on these studies, it can be concluded that the IFS approach has been applied in urban logistics and last-mile delivery research; however, its use has not been as extensive as that of other methods, such as the traditional fuzzy approach. The above research indicates that the primary objective of MCDM methods is to support informed decision-making. These methods evaluate multiple options and determine how well they meet the set goals. In this way, the DM is assisted in selecting a specific alternative, and using these methods leads to improved efficiency and sustainability in business, as demonstrated by Özekenci [54]. Additionally, these methods can help improve transportation through sustainable development, as shown by Çilingir et al. [55] in their study conducted in Turkey. Furthermore, MCDM methods can be applied to various urban logistics issues, such as urban transport [56], machine learning in smart cities [57], digital technology in tourism [58], transport management [59], sustainable location selection in transport [60], and many other areas.
Based on this, the research aims to fill several gaps. First, IFS requires further development to address urban logistics challenges and to select applications that enhance last-mile delivery. Second, comprehensive criteria will be used to choose applications, providing guidelines for improving this approach in urban logistics. Third, an integrated MCDM model is developed for application selection, creating new opportunities for future research, with an emphasis on weighting through a framework of criteria and ranking applications for urban logistics.
3. Methodology and Research Methods
Given the specificity of this research and the demonstration of the selection process for an application to improve urban logistics through a practical example, the section will be divided into several segments. The company and the research problem will be described, followed by the criteria and applications. Finally, the IFS approach and the MCDM methods used in it will be explained.
3.1. Case Study
The EX company operates in Bosnia and Herzegovina (BiH) and delivers goods within this territory. The main activity of this company is express mail services. So far, more than 19,000 clients have used the company’s service, and it has a high quality of delivered packages. This company operates 17 distribution centers, 10 business centers, and a fleet of over 600 delivery vehicles, employing more than 700 staff members. The current system allows tracking shipments until delivery. The majority of packages delivered by this company are ordered through e-commerce platforms. Due to the specific nature of this type of delivery, the EX company faces several challenges, including higher delivery costs driven by rising delivery volumes. Most deliveries occur in urban areas, where vehicle traffic has increased, and some deliveries are made past the scheduled deadlines. Additionally, the company faces environmental concerns related to CO2 emissions from its vehicles. To address these issues, the EX company has already taken steps, such as replacing nearly 40% of its fleet with electric vehicles. This move helps reduce problems primarily associated with air pollution. To further reduce delivery costs and shorten delivery times, the company is considering purchasing new applications to support this process.
Based on this, the EX company aims to acquire an application that best suits the needs of urban logistics and last-mile delivery. This application should have specific features to fulfill the company’s requirements, such as optimized delivery in urban areas. It should gather data, particularly on traffic and congestion, and provide alternative routes for users. Therefore, these applications should support AI or machine learning. To make this decision, the EX company first identified which employees could participate in this decision-making process. During this selection process, an effort was made to identify employees who have the most contact with various types of applications. Especially those who have implemented existing applications for tracking shipments. Ultimately, it was decided to select 10 employees to serve as DMs in the process. Since the EX company operates across BiH, it has appointed operators who monitor and plan deliveries as DMs and use applications for tracking shipments. All of these operators have over five years of experience.
The DMs’ first task was to determine the criteria for evaluating these applications. To do this, the modified Delhi technique was used. This method was chosen because the selected employees were not in the exact location, so it was necessary to gather information from them. This technique was used to minimize bias in the application selection process. Additionally, the chosen DMs are independent of the IT supplier and are located in different locations; their anonymity was also maintained. The technique used to establish the criteria was as follows: first, the DMs proposed a set of criteria for evaluating applications. These criteria were then collected and processed to combine similar ones and eliminate duplicates. The refined criteria were then sent back to the DMs for further prioritization. This process was repeated over three iterations, with the data collected, processed, and sent back to the DMs. Ultimately, 10 criteria were selected to evaluate the applications. The reason for choosing ten criteria is that too few would limit decision-making span, while too many would weaken the importance of each criterion. Therefore, the DMs agreed to use 10 criteria (Table 2). These criteria address various aspects of the company’s business and broader social communities, including ecological, economic, social, and technological factors. To assist the DMs, questions were provided to assess how well each application satisfies these criteria (Table 2).
Table 2.
Selected criteria for evaluating applications in urban logistics.
After establishing the evaluation criteria, the next step for the DMs was to identify which applications to assess. They began by listing all potential applications and then filtered those that could meet the company’s needs. The applications were evaluated based on how well they met these needs, resulting in a reduction in the number of applications. The Delphi method was then repeated using the same process used to establish criteria, resulting in the selection of eight applications. To maintain objectivity, the specific names of the applications will not be disclosed; instead, they will be labeled A through H. This anonymization helps prevent any bias, as not all applications will necessarily be the best. Given that MCDM methods rank alternatives, these applications will be ordered from most to least suitable during evaluation. The following outlines the key characteristics of these applications:
- Application A is aimed at optimizing routes for delivering goods to reduce fuel consumption and, at the same time, reduce CO2 emissions. This application uses traffic and weather data to optimize delivery routes. It is particularly suitable for medium- to large-sized logistics companies. It is an application that uses dynamic planning and real-time route optimization algorithms, which can be applied in urban logistics.
- Application B uses a crowdsourcing approach that allows data to be shared between different users. The advantage of this application is the reduction in delivery costs by sharing vehicles of several partners. This application requires greater coordination among users because data and cars are shared. It represents a collaborative and crowdsourced logistics platform that shares resources and drives cost-effectiveness in last-mile delivery operations.
- Application C is specialized for use with electric vehicles owned by logistics companies. This application enables battery monitoring and, based on the consumption of these vehicles, plans charging schedules accordingly. Since electric vehicles do not emit CO2, this application can also be used in urban areas where traffic bans apply to classic vehicles. This application uses electric vehicles for fleet management, which are only available in zero-CO2 delivery zones.
- Application D uses predictive analysis and machine learning to plan deliveries for low levels of orders. This can reduce costs and congestion in urban areas. This application is suitable for companies operating in cities with seasonal or irregular logistics demands. It belongs to the class of predictive analytics applications for forecasting logistics demand, enabling the planning of deliveries accordingly.
- Application E combines IoT sensors for tracking packages with an emphasis on transparency and secure delivery of shipments. These sensors provide data on the shipment’s location and delivery time. It can be used when high-value goods need to be delivered or when sensitive goods require special handling during delivery. It presents a real-time shipment tracking application based on IoT, an urban logistics system.
- Application F allows the use of multiple delivery modalities to deliver goods as efficiently as possible. The focus of this application is on optimizing last-mile delivery. This application is suitable for urban areas with restrictions on motor vehicle use and where deliveries need to be made. It can be used to manage multimodal last-mile delivery systems that integrate different vehicles for sustainable operations.
- Application G uses AI to analyze customer behavior and thus predict orders. Additionally, it enables dynamic route planning, allowing routes to be adjusted with new data as it becomes available. This application is suitable for large logistics companies with substantial field data. It can be categorized as an AI-powered demand forecasting platform and integrates with customer analytics in urban logistics.
- Application H uses smart city sensors and thus obtains traffic data to optimize routes. In this way, it enables the planning of sustainable delivery routes. This application offers strong compatibility with urban planning and smart cities. It falls into the category of applications for logistics coordination in smart cities, where open urban data and traffic monitoring sensor networks are used.
After identifying the criteria and alternatives using DMs, their third task was to assess the importance of the criteria and evaluate applications against them. To do this, linguistic values were used in this research, enabling decision-making with uncertain data. Uncertainty is present when human factors influence the decision process, as people often question whether ratings are realistic, too high or too low, and whether they are accurate. Therefore, uncertainty exists among DMs during decision-making. Another reason for using linguistic values in the rating function is that the criteria are qualitative, making it difficult to assign precise numerical ratings, so linguistic ratings are used instead. The same linguistic values are applied to assess the importance of criteria and evaluate applications, ranging from “absolutely bad” to “absolutely good,” with nine rating levels (Table 3).
Table 3.
Linguistic ratings for determining the importance of criteria and evaluating applications.
3.2. Application of the Intuitionistic Fuzzy Set Approach and the MCDM Method
After the DMs assess the importance of the criteria and evaluate the applications, the analytical phase proceeds to using these ratings to generate research results. Since DMs determine the ratings for criteria and applications, there is some uncertainty in those ratings. This is especially true in decision-making, when DMs lack all the necessary information for proper assessment. In such cases, they give ratings under conditions of uncertainty. For this reason, IFS was used in this research. Besides allowing for the inclusion of uncertainty in decision-making, this approach handles it more effectively than the classic fuzzy set [61]. It also enables more flexible rating, as the IFS approach provides DMs with a sense of security by incorporating uncertainty into the process [62].
To include uncertainty in decision-making, the starting point is to determine the degree of membership in the set A ( and the degree of non-membership in that set [63]. Based on the relationship between membership and non-membership in the set, the degree of uncertainty is calculated, which is defined as:
The uncertainty is defined as the difference between the value one (1) and the values of the degree of membership and non-membership. When the degree of uncertainty is also defined, then the set A can be defined in this approach as:
Atanassov [63] developed and defined the basic operations with IFS on these grounds, namely:
However, to apply the IFS approach in practice, various techniques have been developed. In this study, the procedure developed by Işık & Adalar [64] will be used. This procedure has the following steps:
Step 1. Evaluation of the importance of criteria or alternatives using linguistic values.
Step 2. In the IFS approach, aggregate decision matrices are formed by converting linguistic values into intuitionistic fuzzy numbers (IFNs).
Step 3. Determination of positive () and negative ideal solutions ().
Step 4. Determination of the distance measure.
Step 5. Calculating the closeness coefficient (CC).
Finally, once the CC coefficient values are obtained, they are treated as crisp numbers, and then classic MCDM methods are applied.
To obtain the final results, it is first necessary to determine the importance of the criteria and rank the applications. This is done using the MCDM method. In this study, the SiWeC (Simple Weight Calculation) method will be used to determine the importance of the criteria by assigning weights, and the WASPAS (Weighted Aggregated Sum Product Assessment) method will be used to rank the applications.
The SIWEC method is used to determine the weight of the criteria based on the DMs’ ratings [65]. This method is used because of its simplicity and efficiency in determining the weight of the criteria. The SIWEC method determines the weight of the criteria based on direct assessments by the DMs. The reasons for choosing this method in this study are as follows:
- Its ease of application comes from using a simple assessment of the criteria’s importance by the DMs.
- A quantitative measurement scale helps determine the importance of criteria and makes data collection and processing easier.
- It includes subjective assessments based on DMs’ ratings, where their knowledge and experience are considered.
- It is less demanding for DMs because only a rating of a single criterion is needed without comparing it to others, reducing their cognitive effort.
- Its flexibility is shown in the use of different value scales in the assessment, allowing for the use of linguistic values, which means this method can be applied in various situations.
To apply this method in practice, the following steps need to be performed:
Step 1. Normalization of CC coefficients (). The initial normalization values are the CC values obtained by converting IFN into crisp values. The reason for using this step is that, due to the uncertainty embedded in these values, the maximum is not one, but slightly less. Therefore, normalization ensures that the maximum criterion score is one.
where is the maximum value of the CC coefficient for all criteria.
Step 4. Calculation of the standard deviation (). The task of calculating the standard deviation is to evaluate the ratings of individual DMs [66]. If the ratings of individual DMs are uniform and there is little dispersion, the importance of the ratings of these DMs will be lower than those of others, where there is greater dispersion in the ratings and the ratings are uneven.
Step 5. Weighting the normalized values with the standard deviation (). This step corrects the DM’s ratings in accordance with the standard deviation values. In this way, different importance is given to the DM’s ratings.
Step 6. Calculation of the sum of the weights of the criteria (). In this step, the weighted scores for individual criteria from all DMs are summed.
Step 7. Determination of the final weights of the criteria (). To make the weights’ values equal to one, the individual value of each weight is divided by the total value of all weights for every criterion.
The WASPAS method will be used to rank the alternatives. The WASPAS method was first used in practice by Zavadskas et al. [67]. The specificity of this method lies in its combination of two methods: the Weighted Sum Model (WSM) and the Weighted Product Model (WPM). The result from these methods is a compromise between them. A compromise of the results from these methods forms the final result. The reasons why this method was chosen are as follows:
- It combines two methods to determine the value of alternatives, reducing the possibility of errors and providing a more reliable ranking.
- This approach is practical for solving complex decision-making problems, where its results remain consistent even when the number of alternatives decreases.
- In addition to combining two methods, it is simple to calculate, enabling a quick assessment of alternatives.
- The method has proven to be highly reliable in practice and has been used in various decision-making problems.
The steps for implementing this method in this research are as follows:
Step 1. Formation of the initial decision matrix. This decision matrix is formed by deriving IFNs from linguistic values and then transforming them into crisp values using the CC coefficient.
Step 2. Normalization of the initial decision matrix (). Due to the specificity of the linguistic values used and the use of qualitative criteria, only one normalization is used. Unlike the normalization used in the SiWeC method, in the WASPAS method, individual data are divided by the highest value of each criterion.
Step 3. Calculating the value of the WSM method (). In this step, all weighted values for individual alternatives are summed, and the value of the WSM method is obtained.
Step 4. Calculating the value of the WPM method (). When calculating this method, the normalized values are scaled, and the product of these scaled values for individual alternatives is determined.
Step 5. Final value of the WASPAS method (). In this step, DMs should assess the importance of the results from the WSM and WPM methods in the final result of this method. This is reflected in the compromise in the results of these two methods.
Usually, the same importance is given to these two methods, so in this study, the importance of the method will be 0.5.
After the ranking order is determined, two more analyses will be conducted. The first analysis assesses the consistency of the relationship between the WSM and WPM methods. In this analysis, the importance of one method will increase, while the others will decrease. In this way, the importance of one method will be , while the importance of the other method will decrease proportionally . In this way,
After determining the consistency of the WASPAS method results, a comparative analysis will be conducted. This analysis will use the initial decision matrix for evaluating applications and the criteria weights obtained through the SiWeC method. Then, other MCDM methods will be used to rank the applications, and it will be assessed whether there are discrepancies between the results of the WASPAS method and those of the other MCDM methods.
The second analysis in this study is a sensitivity analysis. It involves decreasing the importance of one criterion while increasing the importance of the others. Each criterion will be reduced by 30%, 60%, and 90%. Since each criterion is reduced three times, the total number of scenarios will be 30. This analysis will assess the extent to which each criterion influences the final ranking of alternatives.
To accomplish all of this, the steps in the research methodology (Figure 1) are applied. These steps outline the methods and procedures used in this research.
Figure 1.
Steps of the methodology and research methods.
4. Results
In conducting this research, DMs rated the importance of the criteria and then evaluated how well individual applications met them. To obtain results from these ratings, an IFS-based approach will be used. Since calculating the rank order of alternatives requires criterion weights, these weights will first be determined using the SiWeC method. First, DMs need to assess the importance of the criteria using linguistic values (Table 4), and then follow the steps explained in the methodology of this research. When linguistic values are obtained, they must be transformed into IFN values. This process is based on the defined IFN numbers (Table 3). For example, the rating of “absolute good” (AA) is transformed into an IFN number [0.80, 0.10]. All other linguistic values are transformed in the same way.
Table 4.
Decision matrix for criteria evaluation.
Once the IFNs are formed, the uncertainty value is set. In this study, it is set to 0.10 for all answers. This is because it does not add more or less uncertainty to a specific score, as uncertainty exists throughout the decision-making process. After the transformation, the distances from the positive and negative ideal solutions are calculated (Equations (8) and (9)). In the example of the CR1 criterion and the score obtained by DM1, this is calculated as follows:
After calculating the distance from the ideal solutions, the closeness coefficient (CC) is determined (Equation (10)). The value for the same example is then calculated as follows:
The explained procedure is repeated for all criteria and DMs, and an initial crisp decision matrix is formed (Table 5), on which the criterion weight values are then calculated using the SiWeC method.
Table 5.
CC value matrix for criteria.
The first step of the SiWeC method is to perform normalization. In this step, it is necessary to identify the highest CC value, which is 0.905, and then divide all other values by this number (Equation (11)). In the same example, normalization is calculated as follows:
After normalization, the standard deviation of the DMs’ ratings is calculated. Then, the normalized values (Equation (12)) are multiplied by this value (Table 6). In the same example, it is calculated as follows:
Table 6.
Weighted decision matrix and results of the SiWeC method.
With closer examination of the weights of these criteria, it is apparent that their values fluctuate around a central point (Figure 2), and some individual deviations affect the importance of these criteria.
Figure 2.
Graphical representation of criteria weights.
Then, the aggregate value for each criterion is calculated from the weighted values (Expression 13), and the weights for each criterion are finally determined (Expression 14). Using the CR1 criterion as an example, this is calculated as follows:
Based on DM’s ratings and the steps of the SiWeC method, criterion CR 7—Security and data protection—is the most important, followed by criterion CR 3—System reliability and stability. Criterion CR 6—Flexibility and modularity—is the least important. It is worth noting that the difference between the highest and lowest criteria is a weight difference of 0.044, meaning the CR 7 criterion has a 55% greater weight than the CR 6 criterion.
After calculating the criteria weights, the application rankings are established. The first step in creating this ranking is similar to the process of calculating the criteria weights. First, the DMs evaluate the application using linguistic values (Table 7), and then the transformation to crisp values is performed by calculating the CC value.
Table 7.
Decision matrix for application evaluation.
The process for calculating CC values is identical to that for determining criterion weights, so it will not be explained here to avoid repetition. After creating the initial decision matrix (Table 8), the steps of the WASPAS method are followed. The first step in this method is normalization. Unlike the normalization used in the SiWeC method, this approach divides each value by the maximum value for a specific criterion (Equation (15)).
Table 8.
Initial matrix of CC values for application.
On the example of Application 1 and CR 1 criteria, normalization is calculated as follows:
Afterward, the WSM and WPM values are calculated (Equations (16) and (17)). In the Application 1 example, the values of these methods are determined as follows:
Ultimately, equal importance is assigned to these methods, and the final value of the WASPAS method is computed (Equation (18)). For the criterion CR 1, this is calculated as follows:
The results obtained using the WASPAS method show that Application 1 is ranked best, followed by Application 8 (Table 9), while Application 2 is ranked worst. In this case, it means the selected application is specifically designed to determine vehicle routes and uses traffic data from other databases.
Table 9.
Results of the WASPAS method.
To confirm these results, a consistency test of the WASPAS method will be performed by increasing the importance of the WSM method in determining the WASPAS value and decreasing the importance of the WPM method. In this way, 11 results from the WASPAS method will be obtained, representing this analysis (Table 10).
Table 10.
Consistency test of the results of the WASPAS method.
The results of this analysis showed that the ranking of these applications remained unchanged, indicating that the WSM and WPM methods deliver consistent results.
After the consistency test, a comparative analysis will be conducted. In this analysis, other MCDM methods are used to rank applications [68]. Four additional MCDM analyses were included. The SAW (Simple Additive Weighting) method was chosen because it essentially represents the WSM method, with the only difference being the name; their steps are identical. The ARAS (Additive Ratio Assessment) method was selected for its specific approach to calculating normalization, which uses a different normalization method, and also for its comparison with the optimal function of the alternatives. The MABAC (Multi-Attributive Border Approximation area Comparison) method also uses a different normalization and weighting of normalized data. The CORASO (COmpromise Ranking from Alternative SOlutions) method uses the same normalization as the WASPAS method, while the ranking of alternatives is based on optimal solutions. Based on these methods’ results, it was shown that the ranking order of applications remains unchanged, confirming that application 1 is the top-ranked application (Figure 3).
Figure 3.
Comparative analysis results.
Following this analysis, a sensitivity analysis is conducted, in which the value of each criterion is decreased by 30%, 60%, and 90%, resulting in 30 different scenarios. The results of this test (Figure 4) showed that lowering the weight of individual criteria not only impacted the ranking of Application 2 but also altered the rankings of other applications, depending on the scenario. For instance, Application 1 ranked second when the importance of criterion CR 3 was decreased by 90%. In this case, Application 3 needs to improve this criterion (System reliability and stability) to surpass Application 1 in the ranking. This approach enables the observation of other applications and identifies the necessary improvements for them to become the top choice for company EX.
Figure 4.
Sensitivity analysis results.
To analyze the sensitivity analysis in more detail, the Spearman correlation coefficient was calculated for the rank orders obtained from this analysis (Table 11). In this analysis, the results obtained by scenarios were compared in relation to the obtained rank order. These results indicate that most of the ranks are consistent with those obtained in the original ranking. The largest deviation is in scenarios S2 and S3, where the rank order has changed for five applications. The rank order has changed the most in application A5, which moved from fifth place in the ranking to second place. Only in these scenarios is the level of statistical significance lower; however, this correlation remains statistically significant. In all other scenarios, there is a strong correlation between the rank orders (p < 0.003).
Table 11.
Results of the Spearman correlation coefficient in the sensitivity analysis.
5. Discussion
The challenges in urban logistics, particularly in last-mile delivery [69], require logistics companies to adapt and modify their business models to reduce their negative environmental impact through social responsibility. This is especially important because the increase in online sales has led to a greater number of packages needing delivery, particularly in urban areas [1]. This increases CO2 emissions, which is why new solutions to reduce air pollution are being introduced in last-mile delivery. The importance of using applications in urban logistics has been highlighted in many studies, such as Liu et al. [70], who emphasize that delivery routes should be optimized, while Abouelrous et al. [12] argue that, in addition to optimizing routes, electric vehicles should be used for last-mile delivery, a point also supported by Ferreira and Esperança [15]. This research shows that the two leading solutions for sustainable urban logistics are the use of electric vehicles and route optimization through applications. As the EX company began replacing its vehicle fleet with electric vehicles to reduce pollution, its next step was to implement applications to optimize delivery routes, which is the focus of this research.
To select the application that would produce the best results, the EX company used subjective decision-making based on expert judgment, specifically that of DMs involved in the research. The employees chosen as DMs were dispatchers responsible for handling deliveries, as they are directly involved in the delivery process. Only employees with a minimum of five years of experience in these roles were selected. These employees helped determine the criteria and applications observed in the study, using an adapted Delphi method for selection. After identifying 10 criteria and 8 applications related to last-mile delivery, the DMs evaluated them by assessing the importance of each criterion and the extent to which they met the research standards, using linguistic ratings. Since most of the criteria were qualitative, assigning precise ratings was difficult [11]. The inability to provide precise ratings introduced uncertainty into the decision-making process, as DMs lacked sufficient experience to evaluate specific applications definitively without prolonged use. Therefore, this research also includes the uncertainty related to the final decision the company will make regarding the application choice.
The EX company has not yet used delivery route optimization applications, so its decision-making process is somewhat uncertain. Addressing this issue includes integrating the uncertainty in DMs into the process, which can be achieved by applying the IFS used in this study. Gerasimenko and Kureichik [51], Ayyildiz and Erdogan [52], Bilgili et al. [53], and many others have demonstrated that IFS offers robust indicators for addressing problems in urban logistics and last-mile delivery. This is because, when defining the IFN, the degrees of membership and non-membership are specified, and a certain amount of uncertainty is calculated from them. Unlike some studies, this study used the same level of uncertainty when defining the IFN for linguistic values. This approach treats all ratings with linguistic values equally, without prioritizing any of them. Additionally, when defining ratings using linguistic values, the goal was to allow their use both for determining the importance of criteria and evaluating applications based on these criteria. This made the decision process easier for DMs, as they did not need to consider the type of criteria; instead, they could evaluate all applications uniformly and assess the importance of the criteria [71]. To obtain results using IFS, the method proposed by Işık & Adalar [64] was used, which simplifies the process by calculating deviations from the positive and negative ideal solutions to produce crisp values expressed via CC values.
Since uncertainty was incorporated into the decision-making process using IFS, the final choice was determined through the MCDM method. The SiWeC method was used to assess the importance of criteria by assigning weights. The results indicated that, according to DMs, the most important criterion relates to security and data protection. Every application must be safe for users to ensure practical usability [72], as unauthorized individuals may access key data if it is not secure [73]. Therefore, protecting the company’s business data is crucial to prevent unauthorized access and misuse. The second most important criterion was the system reliability and stability. To minimize delivery costs, the system must use multiple data sources for analysis, providing accurate and reliable information throughout the delivery process. When determining optimal last-mile delivery routes, it is necessary to use data from other sources. Additionally, the system’s stability is reflected in the application’s fast, non-blocking performance, enabling continuous use at any time.
When ranking applications using the WASPAS method, Application 1 was identified as having the best characteristics. This result was achieved by DMs comparing how the observed applications could benefit the company EX. The top-ranked application would therefore be the best fit for this company because of its functionality. According to DMs, this application would best highlight sustainability as a key aspect of modern business. These findings were also confirmed through a comparative analysis. This analysis showed a significant difference between the applications and indicated that using different MCDM methods did not lead to different rankings, even though various steps were used to obtain the results. This is likely due to its specificity, as it is designed for medium- to large-sized logistics companies and optimized for efficient delivery routes. In this way, this application can help the EX company reduce fuel consumption and CO2 emissions, especially since the company has not yet fully transitioned to electric vehicles. Additionally, the application can access data from specialized websites, such as traffic or weather information, and predict delays, enabling delivery drivers to take alternative routes. This ensures deliveries are completed as quickly as possible, improves delivery reliability, and increases customer satisfaction, which can positively impact the company’s service usage. However, this may not be the top choice for all similar companies, as each has its own goals; criteria that influence application selection, leading other companies to prioritize different factors. Sensitivity analysis revealed that this application is reliable and stable; when the importance of this criterion decreases, Application 3 performs more effectively. By using this and other applications, it is possible to digitally transform logistics, with a particular emphasis on urban logistics and last-mile delivery. In this way, express delivery companies such as EX will improve their business and implement sustainability in their operations. In addition, they will improve vehicle tracking and delivery route planning [21]. This research has demonstrated the necessity of investing in digital technology to allow for company transformation, optimize last-mile delivery routes through applications, and reduce emissions while improving delivery reliability. The most important thing is to promote sustainability in business, as it is an imperative of modern business [74,75].
5.1. Research Implications
The research was primarily aimed at improving urban logistics at the EX company, with a focus on last-mile delivery. Therefore, this research first identified the problems in this logistics area through a review of previous research. These problems are common across all logistics companies, highlighting areas that need improvement. Additionally, these issues should be studied first theoretically and then practically in logistics operations. One aspect of this challenge is addressed by applications that increasingly optimize delivery routes and select the most suitable delivery vehicles. Of course, the use of these applications is growing over time. However, choosing an application is a significant investment for each company, and selecting the wrong one can cause different issues. For instance, the application may not align with the company’s goals, and purchasing it may require significant financial resources. Therefore, the most important implication of this research is the development of a methodology for selecting urban logistics applications, especially for last-mile delivery. This research provides initial guidance for application selection, including questions to help DMs evaluate the importance of various criteria and how well applications meet them. This methodology has a significant impact on the further development of decision-making theory in logistics, as well as on practical decision-making processes. The approach used to obtain these results can also be applied by other companies involved in delivery, including those outside logistics that deliver products or provide services such as passenger transport or municipal waste collection. These companies operate in urban environments and must apply digital solutions to urban logistics. This research can serve as an example for them on how to improve their business through the development of sustainable practices and reduce the negative impact on the environment by reducing CO2 emissions. Therefore, the implications of this research are broader and relevant to all companies engaged in a specific type of transportation. Applying the IFS approach to this decision-making process also suggests that uncertainty can be incorporated into nearly all decision problems, especially those involving subjective judgments by decision-makers. In this way, this research encourages the use of this approach, representing an additional implication of this research.
5.2. Limitations and Future Research Directions
When conducting a study, it is always possible to approach it differently, which defines the study’s own limits. First, the focus can be on selecting the criteria, then on selecting the applications, the approach, and the methods. When selecting criteria, other options can be considered, but the choice primarily depends on the study’s goals. Future studies based on this methodology should be tailored to the specific needs of companies, with criteria adapted accordingly. Of course, different criteria could have been used, which would have influenced the final decision. In this case, DMs were involved in selecting criteria and determining which ones were significant for the company. In future research, it would be necessary to include additional criteria, such as real-time data integration or application modularity. These criteria should be adapted to the needs of the companies being studied, and, together with the DMs, it should be determined which criteria are adequate for them to achieve the set development goals. Another limitation is the selection of applications observed. This limit is related to the large number of applications on the market and the impossibility of evaluating them all. In this study, applications deemed most suitable by DMs for the company were chosen. Also, due to incomplete information, there is uncertainty in the decision-making process, and a better option may not have been considered. To address this, uncertainty was incorporated into decision-making through IFS, though this approach may also be a limitation. Future research could compare different approaches to determine which produces better results. The choice of decision-making approach and MCDM method is at the researcher’s discretion. Other methods could be used to determine the criteria weights and rank alternatives. The reasons for the chosen methods are discussed within this research, and future studies could apply this methodology, especially when uncertainty is involved. An additional limitation of this research is that it is based solely on the subjective assessment of DMs. Therefore, it is necessary to add criteria that can be objectively measured in future research and incorporate this data into the decision-making process in order to obtain more objective results. In future research, it is also necessary to consider options for applications that support machine learning or AI, allowing these applications to make certain decisions independently without human influence. In this way, it would be possible to immediately influence the change in route, which would reduce fuel consumption, CO2 emissions, and make faster delivery.
6. Conclusions
This study aimed to address four key research questions. The first question focused on identifying the most important criteria for application selection. Using the SiWeC method, it was found that Security and data protection, as well as System reliability and stability, are the most important criteria. The second question involved understanding how to incorporate uncertainty into decision-making. The IFS approach demonstrated that accounting for uncertainty results in more reliable decisions than approaches that ignore it. For the third question, applications were ranked using the WASPAS method, with Application 1 ranked best overall because it met the extended sustainability criteria used in this study. To validate these findings, two additional analyses were performed. The first was a consistency analysis, which evaluated the contribution of the WSM and WPM methods within the WASPAS framework, confirming the reliability of the ranking across different scenarios. The second was a sensitivity analysis, examining how different criteria affected the rankings of the applications. Results showed that Application 1 consistently ranked first in most scenarios, with only one instance of second place, answering the fourth research question. Based on these results, Application 1 is the top choice for company EX, enabling improvements in urban logistics, particularly in last-mile delivery, and highlighting the vital role of applications in implementing sustainable urban logistics and helping companies enhance their operations. Based on all this, this research has demonstrated that companies involved in urban logistics should use applications to optimize routes, particularly in last-mile delivery. Policymakers should encourage the procurement of these applications because their implementation directly reduces CO2 emissions and delivery times. Once these applications are introduced, the task of managers will be to protect data and increase the reliability of these applications to ensure robustness. Additionally, it is necessary to test these applications to improve their operational efficiency.
Author Contributions
Conceptualization, A.P. and R.D.; methodology, A.P.; software, N.P.; validation, N.P., Đ.Ć. and N.N.B.; formal analysis, A.P.; investigation, A.Š.; resources, R.D.; data curation, N.P.; writing—original draft preparation, A.P.; writing—review and editing, A.Š.; visualization, Đ.Ć.; supervision, N.N.B.; project administration, A.Š.; funding acquisition, A.P. All authors have read and agreed to the published version of the manuscript.
Funding
This research received no external funding.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
The authors confirm that the data supporting the findings of this study are available within the article.
Conflicts of Interest
Nedeljko Prdić is employed by JKP Tržnica Novi Sad. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
References
- Krstić, M.; Tadić, S.; Čvorović, A.; Veljović, M. Optimization of Last-Mile Delivery Alternatives Using the Fuzzy FARE and ADAM Multi-Criteria Decision-Making Methods. J. Eng. Manag. Syst. Eng. 2025, 4, 98–108. [Google Scholar] [CrossRef] [Scilit]
- Hussein, A.M.; Osman, B.M. The Impact of Rapid Urbanization on Poverty Levels in the Context of Climate Change: Empirical Evidence from Somalia. Chall. Sustain. 2024, 12, 281–291. [Google Scholar] [CrossRef] [Scilit]
- Bányai, Á.; Kaczmar, I.; Bányai, T. Green Micromobility-Based Last-Mile Logistics from Small-Scale Urban Food Producers. Systems 2025, 13, 785. [Google Scholar] [CrossRef] [Scilit]
- Mohsen, B.M.; Mohsen, M. AI-Driven Energy Optimization in Urban Logistics: Implications for Smart SCM in Dubai. Sustainability 2025, 17, 8301. [Google Scholar] [CrossRef] [Scilit]
- Liu, S.; Cao, Y.; Gao, Q.; Liu, W. Spatial Accessibility in Last-Mile Logistics: A New Dimension of Urban–Rural Integration. Land 2025, 14, 1691. [Google Scholar] [CrossRef] [Scilit]
- Prdić, N.; Gajić, T.; Stepanov, S. Wholesale markets as supply channels for green markets and their impact on rural development in Serbia. Ekon. Poljopr. 2025, 72, 63–87. [Google Scholar] [CrossRef] [Scilit]
- Silva, V.; Amaral, A.; Fontes, T. Sustainable Urban Last-Mile Logistics: A Systematic Literature Review. Sustainability 2023, 15, 2285. [Google Scholar] [CrossRef] [Scilit]
- Gürel, Ö.; Serdarasan, S. Drone-Assisted Last-Mile Delivery Under Windy Conditions: Zero Pollution Solutions. Smart Cities 2024, 7, 3437–3457. [Google Scholar] [CrossRef] [Scilit]
- Simović, S.; Šaković-jovanović, J.; Ivanišević, T.; Trifunović, A. Application of the Analytic Hierarchy Process for Optimizing the Selection of Electric Vehicles in Urban Courier Services. J. Intell. Manag. Decis. 2025, 4, 256–267. [Google Scholar] [CrossRef] [Scilit]
- Srdelić, L. Transition Risks of Climate Change: An Analysis of Greenhouse Gas Emissions in Croatia and the Euro Area. Oeconomica Jadertina 2024, 14, 74–90. [Google Scholar] [CrossRef] [Scilit]
- Nozari, H.; Szmelter-Jarosz, A.; Weiland, D. A Fuzzy Multi-Objective Sustainable and Agile Supply Chain Model Based on Digital Twin and Internet of Things with Adaptive Learning Under Environmental Uncertainty. Appl. Sci. 2025, 15, 10399. [Google Scholar] [CrossRef] [Scilit]
- Abouelrous, A.; Bliek, L.; Zhang, Y. Digital Twin Applications in Urban Logistics: An Overview. Urban Plan. Transp. Res. 2023, 11, 2216768. [Google Scholar] [CrossRef] [Scilit]
- Lauenstein, S.; Schank, C. Design of a Sustainable Last Mile in Urban Logistics—A Systematic Literature Review. Sustainability 2022, 14, 5501. [Google Scholar] [CrossRef] [Scilit]
- Ramírez-Villamil, A.; Montoya-Torres, J.R.; Jaegler, A. Trends and Advances in Urban Logistics Research: A Systematic Literature Review. J. Adv. Transp. 2025, 2025, 8859606. [Google Scholar]
- Ferreira, J.C.; Esperança, M. Enhancing Sustainable Last-Mile Delivery: The Impact of Electric Vehicles and AI Optimization on Urban Logistics. World Electr. Veh. J. 2025, 16, 242. [Google Scholar] [CrossRef] [Scilit]
- Alsaleh, A. Electric and Autonomous Vehicles in Italian Urban Logistics: Sustainable Solutions for Last-Mile Delivery. World Electr. Veh. J. 2025, 16, 338. [Google Scholar] [CrossRef] [Scilit]
- van Vliet, J.E.; de Bok, M.A.; Atasoy, B.; de Almeida Correia, G.H. The evolution of consumer preferences in last-mile delivery methods and the impact on urban logistics—A simulation study in the Rotterdam-The Hague region. Res. Transp. Econ. 2025, 113, 101624. [Google Scholar] [CrossRef] [Scilit]
- Williamsson, J.; Moen, O. Barriers to business model innovation in the Swedish urban freight transport sector. Res. Transp. Bus. Manag. 2022, 45, 100799. [Google Scholar] [CrossRef] [Scilit]
- Kervall, M.; Pålsson, H. Barriers to change in urban freight systems: A systematic literature review. Eur. Transp. Res. Rev. 2022, 14, 29. [Google Scholar] [CrossRef] [Scilit]
- Lee, J.; Arts, J.; Vanclay, F. Investigating institutional barriers and opportunities to an integrated approach for transport and spatial development: Mega urban transport development in a rapidly developing city, Seoul. J. Urban Aff. 2022, 46, 40–62. [Google Scholar] [CrossRef] [Scilit]
- Wang, T.; Chen, J.; Lu, J.; Liu, K.; Zhu, A.; Snoussi, H.; Zhang, B. Synchronous spatiotemporal graph transformer: A new framework for traffic data prediction. IEEE Trans. Neural Netw. Learn. Syst. 2023, 34, 10589–10599. [Google Scholar] [CrossRef] [Scilit]
- Deweerdt, T.; Fabre, A. The role of land use planning in urban transport to mitigate climate change: A literature review. Adv. Environ. Eng. Res. 2022, 3, 033. [Google Scholar] [CrossRef] [Scilit]
- Bastida-Molina, P.; Ribó-Pérez, D.; Gómez-Navarro, T.; Hurtado-Pérez, E. What is the problem? The obstacles to the electrification of urban mobility in Mediterranean cities. Case study of Valencia, Spain. Renew. Sustain. Energy Rev. 2022, 166, 112649. [Google Scholar] [CrossRef] [Scilit]
- Sinesio, E.P.; Fontana, M.E.; de Guimarães, J.C.F.; Marques, P.C. Autonomous Last-Mile Logistics in Emerging Markets: A Study on Consumer Acceptance. Logistics 2025, 9, 106. [Google Scholar] [CrossRef] [Scilit]
- Xing, X.; Wan, L.; Luo, F. Demand prediction for shared bicycles around metro stations incorporating STAGCN. PLoS ONE 2025, 20, e0328452. [Google Scholar] [CrossRef] [Scilit]
- Teixeira, L.; Ramos, A.L.; Costa, C.; Pedrosa, D.; Faria, C.; Pimentel, C. SOLFI: An Integrated Platform for Sustainable Urban Last-Mile Logistics’ Operations—Study, Design and Development. Sustainability 2023, 15, 2613. [Google Scholar] [CrossRef] [Scilit]
- Fioravanti, R.; Lima, O.F., Jr.; Giset, M.M.; Pinto, J.A. ULaaS: Urban Logistics as a Service—A conceptual platform for digital transformation of logistics services in urban areas. Transp. Res. Procedia 2025, 82, 1369–1383. [Google Scholar] [CrossRef] [Scilit]
- Kolasińska-Morawska, K.; Brzozowska, M.; Morawski, P. Mobile digital technologies as an integrator of the logistics of transport services in a Smart City. Eng. Manag. Prod. Serv. 2024, 16, 96–115. [Google Scholar] [CrossRef] [Scilit]
- Diglio, A.; Piccolo, C.; Pipicelli, E. The effective design of self-collection networks in last-mile delivery: Evidence from the Italian case. Int. J. Syst. Sci. Oper. Logist. 2025, 12, 2513647. [Google Scholar] [CrossRef] [Scilit]
- Zhang, Q.; Demir, E. Parcel locker solutions for last mile delivery: A systematic literature review and future research directions. Front. Future Transp. 2025, 6, 1654621. [Google Scholar] [CrossRef] [Scilit]
- Bouti, N.; Boukallal, I.; El Khoukhi, F. Sustainable Urban Logistics and Supply Chain 4.0: A Literature Overview and Future Directions. Lect. Notes Netw. Syst. 2024, 1104, 335–344. [Google Scholar] [CrossRef] [Scilit]
- Bhat, V.; Rincon-Guio, C.; Rabelo, L.; Elkamel, M.; Laynes, V.; Gutierrez-Franco, E. Last-Mile Delivery in High-Density Emerging Economy Cities Using Crowd-Generated Data and Artificial Intelligence. Commun. Comput. Inf. Sci. 2025, 2557, 55–70. [Google Scholar] [CrossRef] [Scilit]
- Rubino, G.; Gattuso, D.; Gronalt, M. Modeling the Interactions Between Smart Urban Logistics and Urban Access Management: A System Dynamics Perspective. Appl. Sci. 2025, 15, 7882. [Google Scholar] [CrossRef] [Scilit]
- Ieva, S.; Bilenchi, I.; Gramegna, F.; Pinto, A.; Scioscia, F.; Ruta, M.; Loseto, G. Enhancing Last-Mile Logistics: AI-Driven Fleet Optimization, Mixed Reality, and Large Language Model Assistants for Warehouse Operations. Sensors 2025, 25, 2696. [Google Scholar] [CrossRef] [Scilit]
- Nila, B.; Pamucar, D.; Roy, J. Designing and analyzing drone-based city logistics solutions for Kochi using a Fuzzy Multi-Criteria decision making framework. Environ. Dev. Sustain. 2024, in press. [Google Scholar] [CrossRef] [Scilit]
- Wang, N.; Xu, Y.; Puška, A.; Stević, Ž.; Alrasheedi, A.F. Multi-Criteria Selection of Electric Delivery Vehicles Using Fuzzy–Rough Methods. Sustainability 2023, 15, 15541. [Google Scholar] [CrossRef] [Scilit]
- Moslem, S.; Pilla, F. A hybrid decision making support method for parcel lockers location selection. Res. Transp. Econ. 2023, 100, 101320. [Google Scholar] [CrossRef] [Scilit]
- Bonilla, M.A.M.; Da Silva, B.S.; Schmitt, M.B.; Bouzon, M. Sustainable practices in last-mile logistics of small Brazilian e-retailers: An analysis using an MCDM approach. Acad. Rev. Latinoam. Adm. 2024, 37, 383–408. [Google Scholar] [CrossRef] [Scilit]
- Veljović, M.; Tadić, S.; Krstić, M. Last Word in Last-Mile Logistics: A Novel Hybrid Multi-Criteria Decision-Making Model for Ranking Industry 4.0 Technologies. Mathematics 2024, 12, 2010. [Google Scholar] [CrossRef] [Scilit]
- Pourmohammadreza, N.; Akbari Jokar, M.R. Efficient last-mile logistics with service options: A multi-criteria decision-making and optimization methodology. Int. J. Ind. Eng. Comput. 2024, 15, 367–386. [Google Scholar] [CrossRef] [Scilit]
- Tadić, S.; Krstić, M.; Radovanović, L. Assessing Strategies to Overcome Barriers for Drone Usage in Last-Mile Logistics: A Novel Hybrid Fuzzy MCDM Model. Mathematics 2024, 12, 367. [Google Scholar] [CrossRef] [Scilit]
- Silva, V.; Amaral, A.; Fontes, T. Towards sustainable last-mile logistics: A decision-making model for complex urban contexts. Sustain. Cities Soc. 2023, 96, 104665. [Google Scholar] [CrossRef] [Scilit]
- Ninović, M.; Dobrodolac, M.; Bošković, S.; Dupljanin, Đ.; Lazarević, D.; Dumnić, S. An Extended FullEX method: An application to the selection of online orders distribution modes based on the shared economy. J. Theor. Appl. Electron. Commer. Res. 2025, 20, 207. [Google Scholar] [CrossRef] [Scilit]
- Veličković, M.; Stojanović, Đ.; Ilin, V.; Mirčetić, D. A Combined Multi-Criteria Decision-Making and Social Cost–Benefit Analysis Approach for Evaluating Sustainable City Logistics Initiatives. Sustainability 2025, 17, 884. [Google Scholar] [CrossRef] [Scilit]
- Nong, N.-M.T.; Ha, D.-S.; Tran, D.-T. An Application of Triple Bottom Line Model and MCDM Method to Develop Sustainable Last-mile Delivery in B2C E-commerce. J. Distrib. Sci. 2025, 23, 69–79. [Google Scholar] [CrossRef]
- Kanık, Z.B.; Erişkan, S.; Soysal, M.; Ömürgönülşen, M. A hybrid approach based on qualitative and quantitative techniques for analyzing last-mile parcel delivery. Opsearch 2025, in press. [Google Scholar] [CrossRef] [Scilit]
- Belil, S.; Jghamou, A.; Kouki, S. Optimizing crowdshipping effectiveness in B2C deliveries: A multi-criteria approach for crowd selection. Supply Chain Forum Int. J. 2025, 26, 379–394. [Google Scholar] [CrossRef] [Scilit]
- Kiba-Janiak, M.; Cheba, K.; Mucowska, M.; de Oliveira, L.K.; Piecyk, M.; Evangelista, P.; Prockl, G.; Rześny-Cieplińska, J. How to design a sustainable last-mile delivery and returns business model from E-Customers’ expectations perspective? Res. Transp. Bus. Manag. 2024, 56, 101194. [Google Scholar] [CrossRef] [Scilit]
- Önden, İ.; Deveci, M.; Kara, K.; Yalçın, G.C.; Önden, A.; Eker, M.; Hasseb, M. Supplier selection of companies providing micro mobility service under type-2 neutrosophic number based decision making model. Expert Syst. Appl. 2024, 245, 123033. [Google Scholar] [CrossRef] [Scilit]
- Uyanik, C.; Tuzkaya, G.; Kalender, Z.T.; Oguztimur, S. An integrated dematel–if-topsis methodology for logistics Centers’ location selection problem: An application for Istanbul Metropolitan area. Transport 2020, 35, 548–556. [Google Scholar] [CrossRef] [Scilit]
- Gerasimenko, E.; Kureichik, V. The Minimum Cost Flow Finding under Fuzzy Intuitionistic Conditions. J. Phys. Conf. Ser. 2020, 1703, 012040. [Google Scholar] [CrossRef] [Scilit]
- Ayyildiz, E.; Erdogan, M. Addressing the challenges of using autonomous robots for last-mile delivery. Comput. Ind. Eng. 2024, 190, 110096. [Google Scholar] [CrossRef] [Scilit]
- Bilgili, F.; Yılmaz, Ş.F.; Zaralı, F.; Demirel, N. Evaluation of integrated sustainable last-mile delivery methods in an intuitionistic fuzzy environment. Environ. Dev. Sustain. 2024, in press. [Google Scholar] [CrossRef] [Scilit]
- Özekenci, E.K. A Multi-Criteria Framework for Economic Decision Support in Urban Sustainability: Comparative Insights from European Cities. Int. J. Econ. Sci. 2025, 14, 162–181. [Google Scholar] [CrossRef] [Scilit]
- Çilingir, D.; Öztürk, H.H.; Kaya, K. Invisible Walls to Green Mobility: Barriers to Sustainable Transportation in Izmir. Spectr. Decis. Mak. Appl. 2025, 3, 151–163. [Google Scholar] [CrossRef] [Scilit]
- Anjum, R.; Mirza, M.U.; Kausar, N.; Ali, R. Decision-making Framework for Urban Transportation Using Linear Diophantine Fuzzy Z-numbers with Dombi Aggregation, TOPSIS and VIKOR Methods. Spectr. Oper. Res. 2025, 4, 1–34. [Google Scholar] [CrossRef] [Scilit]
- Lipianina-Honcharenko, K.; Komar, M.; Melnyk, N.; Komarnytsky, R. Sustainable Information System for enhancing virtual company resilience through machine learning in smart city Socio-economic scenarios. Economics 2024, 12, 69–96. [Google Scholar] [CrossRef] [Scilit]
- Wang, J.; Zhao, H. Smart Tourism Cities and Income Equity: How Digital Economy Reshapes Urban-Rural Disparities. Int. J. Econ. Sci. 2025, 14, 311–332. [Google Scholar] [CrossRef] [Scilit]
- Ali, S.; Jamil, M.K. Exchange Property in Double Edge Resolving Partition Sets and Its Use in City Development. Spectr. Decis. Mak. Appl. 2024, 1, 84–98. [Google Scholar] [CrossRef] [Scilit]
- Basuri, T.; Gazi, K.H.; Bhaduri, P.; Das, S.G.; Mondal, S.P. Decision-analytics-based Sustainable Location Problem-Neutrosophic CRITIC-COPRAS Assessment Model. Manag. Sci. Adv. 2025, 2, 19–58. [Google Scholar] [CrossRef] [Scilit]
- Sarfraz, M.; Božanić, D. Optimization of Market Risk via Maclaurin Symmetric Mean Aggregation Operators: An Application of Interval-Valued Intuitionistic Fuzzy Sets in Multi-Attribute Group Decision-Making. J. Eng. Manag. Syst. Eng. 2024, 3, 100–115. [Google Scholar] [CrossRef] [Scilit]
- Iqbal, Q.; Kalsoom, S. Advanced Logarithmic Aggregation Operators for Enhanced Decision-Making in Uncertain Environments. Int. J. Knowl. Innov. Stud. 2024, 2, 57–69. [Google Scholar] [CrossRef] [Scilit]
- Atanassov, K.T. Intuitionistic fuzzy sets. Fuzzy Sets Syst. 1986, 20, 87–96. [Google Scholar] [CrossRef] [Scilit]
- Işık, Ö.; Adalar, İ. A multi-criteria sustainability performance assessment based on the extended CRADIS method under intuitionistic fuzzy environment: A case study of Turkish non-life insurers. Neural Comput. Appl. 2025, 37, 3317–3342. [Google Scholar] [CrossRef] [Scilit]
- Akgül, Y. A Novel MCDM Framework for Evaluating Corporate Financial Performance: Evidence from the Turkish Insurance Sector. J. Oper. Strateg. Anal. 2025, 3, 65–79. [Google Scholar] [CrossRef] [Scilit]
- Şimşek, E.; Eti, S.; Yüksel, S.; Dinçer, H. Evaluation of Purchasing Process in Solar Energy Investment Projects via SIWEC Methodology. Spectr. Oper. Res. 2025, 3, 81–86. [Google Scholar] [CrossRef] [Scilit]
- Zavadskas, E.K.; Turskis, Z.; Antucheviciene, J.; Zakarevicius, A. Optimization of Weighted Aggregated Sum Product Assessment. Elektron. Ir Elektrotechnika 2012, 122, 3–6. [Google Scholar] [CrossRef] [Scilit]
- Tešić, D.; Khalilzadeh, M. Development of the rough Defining Interrelationships Between Ranked criteria II method and its application in the MCDM model. J. Decis. Anal. Intell. Comput. 2024, 4, 153–164. [Google Scholar] [CrossRef] [Scilit]
- Amirova, A.; Iskendirova, S.; Daueshova, A.; Zhanseitov, A.; Abylasimov, M. The impact of digital HRM implementation on public sector performance: A quantitative analysis of service delivery efficiency. Economics 2025, 13, 303–318. [Google Scholar] [CrossRef] [Scilit]
- Liu, Y.; Wu, F.; Liu, Z.; Wang, K.; Wang, F.; Qu, X. Can language models be used for real-world urban-delivery route optimization? Innovation 2023, 4, 100520. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Biswas, A.; Gazi, K.H.; Bhaduri, P.; Mondal, S.P. Neutrosophic fuzzy decision-making framework for site selection. J. Decis. Anal. Intell. Comput. 2024, 4, 187–215. [Google Scholar] [CrossRef] [Scilit]
- Torre, D.; Alferez, M.; Soltana, G.; Sabetzadeh, M.; Briand, L. Modeling data protection and privacy: Application and experience with GDPR. Softw. Syst. Model. 2021, 20, 2071–2087. [Google Scholar] [CrossRef] [Scilit]
- Prdić, N. Mobile devices in function of the sales at the marketplace. Ekonomist 2022, 1, 7–15. [Google Scholar] [CrossRef] [Scilit]
- Elhassan, T. Green technology innovation, green financing, and economic growth in G7 countries: Implications for environmental sustainability. Economics 2025, 13, 69–91. [Google Scholar] [CrossRef] [Scilit]
- Charchafa, I.; Kimouche, B. Understanding environmental transparency: A study of Saudi Arabian companies. Oeconomica Jadertina 2024, 14, 22–40. [Google Scholar] [CrossRef] [Scilit]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).



