Electromobility Implementation Challenges and Opportunities in Urban Parcel Delivery: A Case Study of a Fictive Delivery Company in Miskolc

Abstract

The growing demand for parcel delivery plays an important role in the integration of electromobility and urban logistics into urban delivery systems, especially in a mid-sized Central European city. This study investigates the challenges and opportunities of adopting electric vehicles (EVs) for last-mile delivery in the Miskolc region, Hungary. The author introduces a practical approach to describe the cost-based optimization of urban parcel delivery, formulated as an Electric Vehicle Routing Problem (EV-VRP) that builds on classical Vehicle Routing Problem (VRP) concepts. The developed model focuses on route and vehicle allocation and examines the impact of charging infrastructure and fleet composition on delivery performance, while explicitly evaluating five cost categories: vehicle (including maintenance and service), driver, infrastructure, operation center, and environmental energy. The numerical results validate the model and show that partial fleet electrification can improve cost efficiency and reduce environmental impact even in regions with limited charging capacity. The proposed approach makes it possible to analyze the operational costs of electromobility strategies on last-mile logistics under realistic routing, capacity, and energy constraints. The results confirm that the integration of electric vehicles into city logistics can contribute to more flexible, sustainable, and cost-effective delivery systems. The numerical analysis shows that under the conditions examined, the model results in approximately 20% lower total operational cost compared to the conventional vehicle fleet operating under similar conditions. The cost structure is dominated by labor and vehicle-related components, while infrastructure, operational management, and environmental–energy factors appear with lower intensity.

1. Introduction

Logistics has crucial role in the supply chain and service processes because it must handle numerous logistical tasks based on the 7 QR principle, which underlines the importance of fulfilling customer orders with the right product, in the right quality and quantity, to the right destination, at the right time, at the right costs, and to the right customer [1].

In today’s globalized market, various service industries are developing more efficient and sustainable solutions for diverse finished products distribution, which are offering alternatives to conventional urban deliveries. These activities go beyond traditional market demands and reflect a broader shift toward innovation and environmental responsibility in urban delivery systems [2].

The collaboration between electromobility and logistics offers significant advantages in finished product delivery, which can contribute to meeting sustainability goals and enhancing economic efficiency. These technologies not only solve current problems but also help to define the upcoming logistics solutions [3]. At the same time, last-mile delivery has become a more complex system, where increasing demand, route density, and time-related requirements place additional pressure on delivery operations. These characteristics are closely connected to the classical Vehicle Routing Problem (VRP), which describes how routes can be organized under distance, capacity, and service-time constraints [4]. The modeling principles of the VRP also appear in urban parcel delivery [5], since the allocation of vehicles to delivery clusters, the ordering of stops, and the available service windows follow the same logic. In the case of electric vehicles, these routing constraints must be extended with range and charging-related conditions, which form the basis of the Electric Vehicle Routing Problem (EV-VRP) [6]. The present study applies these principles and adapts them to the regional characteristics of the Miskolc delivery system.

Figure 1 shows that the urban parcel delivery process can be interpreted as a networked system where electromobility, information flow, and operational planning interact. These elements affect the practical organization and coordination of delivery tasks. Several operational aspects presented in the Figure are also closely related to routing-oriented approaches discussed in the literature. Although the Figure provides a conceptual overview, its structure is consistent with the principles commonly applied in VRP formulations, where distance, capacity, and time-related considerations play a key role in planning efficient delivery operations.

Recent research also shows that the spatial distribution and accessibility of charging infrastructure play a central role in achieving sustainable electromobility in urban transport networks [7].

In the Hungarian domestic context, the gradual transition to EV fleets is still a strategically important area, representing a significant challenge for service providers and the Hungarian economy. The availability of EVs equipped with higher-capacity batteries and increased storage space expands the practical applicability of electromobility in logistics. These developments could enable greater flexibility for larger domestic parcel delivery companies to adopt more electric delivery fleets, while also emphasizing the need to balance environmental and economic performance during the transition process [8].

2. Literature Review

In this chapter, the systematic literature review is applied to identify the research gaps, map the relevant scientific articles, and analyze their statistical and content significance. Furthermore, further research directions are presented.

The SLR follows the SLR–PDSA-based testing protocol, which is designed to improve the clarity and quality of the review, identify existing research gaps, and filter out irrelevant or inadequate studies [3,9,10].

2.1. Descriptive Analysis

There are numerous articles which address electromobility, Vehicle Routing Problems, and last-mile logistics supply systems in the context of urban delivery and service processes. To establish the methodology of the structured literature review and define conceptual framework for the connection between this research and the researcher, the most relevant results must relate to the following tasks:

• Search for scientific articles in selected literature sources (such as Scopus, ResearchGate, Google Scholar and Science Direct);
• Select the most relevant articles by carefully reading the abstract and identifying the core topic;
• Define a methodology to analyze the relevant articles;
• Describe the main scientific results and identify the scientific gaps and bottlenecks [11].

The SLR–PDSA protocol was applied following the guidelines described in the methodological literature [3,10]. The SLR–PDSA protocol supports the transparency and quality of the review process, helps identify relevant research gaps, and facilitates the exclusion of studies that do not meet the predefined methodological criteria.

In the first stage, the relevant terms were specified. It is a crucial phase of the integrated review because there are several outstanding review articles in the fields of electromobility and logistics concerning the possibility of improving urban delivery processes, and the presented methodology was applied. The first search in the Scopus database was carried out using the keywords “electromobility”, “distribution”, and “vehicle”, which resulted in 140 records. After selecting journal articles only, 69 publications remained.

Figure 2 shows these articles considering 10 subject areas. This classification shows that interest is highest in engineering, business, management, and computer sciences.

To complement this list with routing-related studies, a second search was performed using the keywords “last-mile”, “distribution”, and “VRP”. This resulted in 40 records, which were reduced to 20 journal articles after applying the same selection criteria.

Figure 3 shows these articles considering 10 subject areas. This classification shows that both technology-oriented and routing-oriented studies appear in the literature on urban delivery systems.

The two searches together provide a dataset that covers electromobility, logistics systems, and routing aspects of urban delivery. These searches were conducted in December 2025, so additional relevant articles may have been published since then.

The articles were excluded under the SLR-PDSA-based testing protocol, whose direction did not find any interest and cannot be addressed to the operation management aspects of decision-making in electromobility, Vehicle Routing Problem, logistics, and urban parcel delivery. After this reduction, I obtained 53 articles. I added 12 other highly relevant articles selected through a separate search, so the final list for classification and evaluation from the point of view of the scientific results includes 65 articles.

2.2. Content and Critical Analysis

Based on the result of the SLR, the selected papers were analyzed. This content analysis provides a comprehensive overview of the international literature, offers an opportunity to learn about the researchers’ work, and identifies related knowledge and future development potential.

Urban parcel delivery is fundamentally a routing-based process, which links these developments directly to the classical Vehicle Routing Problem (VRP). Research efforts carried out in European countries underline the potential of electromobility to redesign these delivery systems by reducing operational costs and promoting environmentally sustainable solutions. The following sub-categories have been identified in the urban delivery supply chain systems, and the analysis is based on this classification.

As a result of the literature review, the following sub-categories were identified, summarizing the areas that helped to provide a comprehensive overview of their current relevance:

• Supply Chain Impacts;
• Technology and Innovation;
• Policy and Environmental Impacts;
• Routing and Operational Methods.

2.2.1. Supply Chain Impacts

This Section examines the impacts of electromobility on production and service supply systems, emphasizing the transformation of traditional delivery systems under emerging electric mobility trends. Electromobility introduces new challenges and opportunities in urban last-mile delivery design: battery life-cycle management, the development and spatial arrangement of charging infrastructure, EV-specific route planning under range and time-window constraints, and the implementation of sustainable reverse-logistics processes.

Lygizos et al. developed a reverse-logistics network for the collection and recycling of end-of-life lithium-ion batteries in Attica, Greece. The circular economic principles and carbon taxation can enhance the sustainability and economic feasibility of electromobility systems [12]. The energy flow and the battery degradation must be carefully balanced to ensure long-term system sustainability. Wohlschlager et al. examined the integration of bidirectional charging to strengthen grid stability and lower life-cycle emissions, particularly when renewable energy penetration exceeds 60% of total supply [13]. Thus, Bautista et al. expanded the sustainability performance and provided magnesium–sulfur and lithium-ion batteries life-cycle assessment. Magnesium-based systems can significantly reduce both carbon intensity and critical material dependency [14].

Zahler et al. focus on these technological and infrastructural developments, presenting methodology of assessing smart and bidirectional charging use cases across ten European countries. While vehicle-to-home applications are already achievable, grid-supporting services still face regulatory constraints [15].

Recent studies emphasize that the spatial layout of charging infrastructure significantly influences the feasibility and efficiency of electric vehicle infrastructure for urban last-mile delivery. Malíčková et al. describe the retail electromobility network possibilities in Slovakia and how it can integrate EV charging infrastructure as a tool for competitiveness and customer retention, while reflecting the growing convergence between mobility systems and commercial services [16]. Lewicki et al. also show the integration of electric vehicle power supply systems within existing urban grids in Türkiye, showing that coordinated grid regulation and smart charging strategies can reduce peak load and improve resilience. Peña et al. show that realistic energy consumption features, including elevation, stopping patterns, and regenerative braking, significantly influence feasible routing structures and the number of required tours [17]. The results support national energy transition objectives and the integration of EV systems within urban energy infrastructures [2]. Sierpiński et al. presented that the appropriate placement of charging stations can reduce non-feasible trips, minimize detours, and additional energy consumption [18], while Gülmez et al. showed that flexible time-window routing supported by multi-objective optimization can enhance green delivery performance [19,20]. Erdelic et al. highlight that the feasibility of last-mile EV delivery strongly depends on balancing partial charging opportunities with service time windows [6]. Beyazıt et al. also highlight that the real-time coordination of integrated EV and hydrogen charging infrastructure is essential for mitigating charger overstaying, improving station efficiency, and supporting the wider scalability of electromobility systems [21].

In addition to infrastructure-focused findings, a growing number of route-planning studies highlight issues specific to densely populated urban areas: energy-related route-planning decisions, limited charging options, detours due to traffic congestion, and the need for dynamic scheduling during operation. These studies underscore the need for models that consider energy, time, and spatial constraints in a real urban environment for last-mile transportation with electric vehicles.

Based on the reviewed literature, the present model incorporates supply chain impacts through explicit vehicle capacity constraints, energy feasibility conditions, and infrastructure-dependent routing decisions. Battery degradation, charging availability, and reverse-logistics considerations motivate the inclusion of energy-related constraints and infrastructure cost components in the EV-VRP formulation.

2.2.2. Technology and Innovation

This Section, technology and innovation, focuses on the interaction between electric vehicles, charging infrastructure, and power systems. Recent studies examine the role of advanced technologies in electromobility, including coordinated and AI-assisted charging, adaptive network planning, and vehicle-to-grid solutions.

Kovačević et al. examine the long-term effects of large-scale transport electrification on power distribution systems up to 2050. Their analysis examined how electromobility affects the distribution grid and highlights the need for coordinated charging and grid reinforcement in urban areas [22]. Gönül et al. propose a coordinated charging scheduling strategy for interoperable EV stations. Their results show improvements in EV infrastructure by balancing grid capacity, minimizing waiting times, and reducing operational costs through smart power allocation [23]. Saqib et al. present an adaptive planning model for charging-infrastructure expansion under uncertainty, integrating spatial and stochastic demand data to support real-time investment and accessibility decisions [24]. Połecki et al. demonstrate how artificial intelligence can support the optimal allocation and grid connection of electric bus charging stations in Poland. They present an approach to reduce grid stress and improve energy efficiency while supporting urban planning [25]. Fakhrooeian et al. determine the maximum number of battery electric vehicles that can be charged simultaneously within a low-voltage grid. The importance of electric vehicle modeling and proactive capacity management support to prevent grid bottlenecks is clear [26]. Similarly, Joglekar et al. introduce an alternative to conventional reinforcement, called Solid-State Transformers. Their study helps to reduce overloading and voltage deviation, thus supporting scalable electromobility without extensive infrastructure expansion [27]. Coban et al. explore bidirectional vehicle-to-grid interactions in the Turkish electricity system to enhance grid flexibility, energy independence, and overall profitability [28]. Chudy et al. evaluate fast DC charging for electric bus fleets in Poland. The authors identify the voltage fluctuations and harmonic distortions exceeding PN-EN 50160 standards [29] and underscore the necessity of continuous power quality monitoring [30]. Nassar et al. address the vehicle design dimension by optimizing a double wishbone suspension system for lightweight electric vehicles, achieving an 8% weight reduction while enhancing comfort and road performance [31]. Ramesh Babu et al. extend the technological perspective by modeling the thermal behavior of large battery packs in electric trucks, and their experimental validation confirms the requirements for safe and efficient battery integration [32]. At the distribution level, Konstantinidis et al. and Johansson et al. highlight how coordinated charging- and smart-grid management can mitigate transformer overloads and voltage instability, providing cost-effective alternatives to extend grid upgrades [33,34]. Braunstein et al. emphasizes that widespread use of three-phase charging and advanced power converters is essential to minimize harmonics and reactive power issues in electromobility networks [35]. Finally, Paffumi et al. analyze real-world driving and charging data from almost 28,000 vehicles to evaluate urban electromobility readiness, showing that optimized charging strategies can meet most mobility needs while keeping grid loads manageable [36].

Recent last-mile logistics articles refer to concepts such as the Physical Internet (PI), digital-twin applications, and urban-cloud platforms in the context of data sharing and operational planning. PI-related studies describe urban logistics systems characterized by higher levels of collaboration, real-time information exchange, and more flexible routing and resource allocation in dense urban environments. Li et al. introduce a Physical Internet-based urban distribution model and report lower distribution costs and higher infrastructure utilization compared to conventional city logistics systems [37]. Peng et al. analyze PI-enabled production–inventory–distribution systems and identify improved resilience and service performance under disruption conditions [38]. Luo et al. examine a metropolitan furniture delivery case and show that PI-based digitalization and optimization can perform better than traditional delivery solutions [39]. Shi et al. focus on joint distribution systems with real-time demand updates and report improved capacity utilization and lower emissions in urban operations [40]. Matusiewicz presents survey-based evidence on researchers’ and practitioners’ views, indicating that the Physical Internet is widely regarded as a promising concept for improving efficiency and sustainability in urban last-mile logistics [41].

The technological aspects discussed in the literature are reflected in the model through parameters describing vehicle energy consumption, charging time requirements, and the availability of charging infrastructure. These elements appear explicitly in the routing constraints and in the infrastructure-related cost components of the EV-VRP formulation.

2.2.3. Policy and Environmental Impacts

This Section highlights the most important electromobility policies, environmental impacts, and regulatory frameworks, which guide the sustainable transformation of delivery systems. The reviewed studies emphasize infrastructure regulation, emissions reduction, and companies’ efforts towards achieving decarbonization goals.

These elements are closely interconnected, as infrastructure regulation shapes the availability of charging facilities, which directly influences emission outcomes and the feasibility of meeting urban and national decarbonization targets.

Struzewska et al. evaluate the influence of electromobility development on the air pollution background level in Poland using the GEM-AQ urban forecast model. Their findings indicated a decrease in specific gas concentrations with increasing EV adoption, confirming the environmental benefits of electromobility and supporting national clean-air policies [42].

Mazur et al. propose a method to identify suitable locations for electric vehicle charging stations along Poland’s TEN-T core network. Their study combines legal, financial, and technical parameters in line with EU Regulation 2019/631 and AFIR requirements. Based on this approach, they found 188 potential locations that could improve both accessibility and sustainability of zero-emission transport [43]. Zadorożny et al. examine the development of electromobility in Poland from 2019 to 2023. Their results show that despite policy support, charging infrastructure remains unevenly distributed and concentrated in large cities. The authors emphasized that expanding stations along the TEN-T network is essential for balanced national growth in sustainable mobility [44].

Kalasova et al. extend urban micromobility behavior, finding that education level significantly influences preferences toward sustainable and shared transport. Their results provide practical insights for policymakers to push traffic behavior patterns toward more active and environmentally friendly transport solutions [45].

Shaban et al. describe regional simulation models of the EV market in Greece, which showed regions with higher GDP, where population density show a higher uptake of EVs adaptation. The study provides a policy framework for designing region-specific incentives consistent with Greece’s National Energy and Climate Plan [46]. Čulík et al. apply new urban mobility regulations to taxi fleet modernization in Slovakia. They argue that digital platforms and electromobility trends have significantly changed the taxi market, creating new challenges and opportunities for sustainable urban transport [47].

Sendek-Matysiak et al. evaluate the TCO calculation to analyze the application of combustion-powered vehicles and EVs. The findings indicate that commercial electric vehicles can be economically better within infrastructure and logistics frameworks [5].

Tucki et al. forecast the EV adaptation possibility in the European Union, United States, and Japan. Their study evaluates the impact of electromobility on power systems. The findings indicate that the climate benefits of EVs are highly dependent on the decarbonization of electricity production and reinforce the need for policy alignment between transport and energy sectors [48]. Baraniak et al. examine how electric vehicle chargers affect electricity quality, with emphasis on issues related to converter systems. The authors also highlight the development potential of charging systems incorporating V2G technology [49]. Skrúcaný et al. examine the environmental impacts of electromobility by evaluating how electricity generation efficiency and energy sources influence GHG emissions in several Central European countries. Based on the EN 16258:2012 standard [50], they compare national electricity mixes and conversion efficiencies to assess the true ecological footprint of electric vehicles. Their findings show that the environmental benefits of electromobility widely diversify from country to country and do not automatically guarantee sustainable transportation [51].

Overall, the reviewed studies address related policy and environmental aspects from different angles, and together they illustrate how regulatory expectations and emission targets influence the practical conditions of electromobility.

The policy and environmental aspects discussed in the literature are reflected in the proposed model through explicit environmental and energy-related cost components, service-time constraints, and infrastructure-related parameters. These elements allow the evaluation of urban parcel delivery strategies under regulatory and sustainability considerations without extending the model beyond its cost-based EV-VRP formulation.

2.2.4. Routing and Operational Methods

A significant part of the recent electromobility and urban-logistics literature focuses on methodological and optimization-oriented approaches. These contributions include mathematical models, simulation frameworks, algorithmic development, and decision-support tools used to analyze charging demand, fleet allocation, and delivery performance. Many of these studies also relate—directly or indirectly—to the classical Vehicle Routing Problem (VRP), which provides the mathematical foundation for organizing urban delivery routes under distance, capacity, time-window and, in the case of electric vehicles, energy constraints.

Zamal et al. examine two-echelon distribution settings and show that different timing constraints narrow the range of feasible routes [52]. Zhang et al. solve scheduling problems with flexible time windows and alternative delivery locations that respond to customer needs and preferences [53]. Xu et al. analyze another two-echelon issue, where both time-window requirements and product freshness affect service quality [54]. Pingale et al. focus on the application of decentralized exchange points in collaborative last-mile delivery and report reductions in unnecessary trips [55]. Lehmann et al. consider multi-trip routing with mixed pickup and delivery tasks and show that time-window conditions strongly influence daily route scheduling [56].

Flocea et al. propose a smart charging reservation algorithm extending the OCPP standard to allow charging pre-booking. By allocating non-overlapping intervals based on historical data, the method improved charging station utilization, user satisfaction, and centralized energy distribution management [57]. This finding can be directly integrated into EV-VRP formulations, where feasible routes depend on the availability of charging windows.

The following predictive methods support reverse-logistics VRPs planning by enabling more accurate estimation of pickup volumes, fleet requirements, and temporal routing cycles in battery-collection networks. Rettenmeier et al. developed a forecasting method for end-of-life automotive traction battery packs, combining benchmark data with multiple performance metrics. The model supports better planning for recycling and reuse, promoting sustainable battery life-cycle management [58]. Lopes et al. introduced the X-Modeci method to estimate electric vehicle charging demands. The approach demonstrated a 10% improvement in station utilization in a large-scale Brazilian case study [59]. Kłos et al. propose a GIS-supported siting method for EV charging stations based on spatial accessibility and urban potential. Their results show a nearly 200% increase in coverage, validating its use for data-driven urban planning [60]. Other studies address the technological and grid-related aspects of electromobility from a different angle.

Kryzia et al. use Monte Carlo simulation to assess the feasibility of converting gasoline-powered cars into electric vehicles. The results indicate economic viability for more than 90% of spark-ignition vehicles, representing a potential demand of 1.75 TWh per year [61]. Király and Medveď simulate vehicle-to-grid (V2G) charging interactions within smart grids that include renewable generation and household loads. Their findings highlight V2G as a distributed storage option enhancing grid flexibility and voltage robustness [62]. In addition to these technology-focused results, several contributions examine planning and optimization issues relevant to charging infrastructure and network operation.

Ali et al. define different demand-based models for EV charging points at various power levels, using over two million simulated profiles. The model supports grid-capacity planning under different charging-power and spatial-demand scenarios [63]. These studies highlight different methodological approaches, but all of them address practical questions related to charging demand and infrastructure planning. Chudy et al. developed an optimization model for managing power-distribution networks with EV charging loads. They identified centralized, decentralized, and hierarchical coordination schemes for safe and efficient grid operation [64]. Lazzeroni et al. developed a simplified model to estimate EV charging demand in urban areas using floating car data. The approach provides reliable load profiles to assist planners in infrastructure expansion [65]. Gauglitz et al. proposed a raster-based geospatial model to integrate public charging infrastructure at points of interest. Their case study confirms the application of municipal-level planning [66].

Mele et al. coordinated smart charging and V2G on a non-interconnected island grid. The results showed that coordinated flexibility management can achieve 20–25% EV penetration without compromising reliability [67]. Sierpiński et al. compared three EV-charging-station siting methods using travel-planner simulations in the Upper Silesian–Dąbrowa Basin. Their findings provide decision support for selecting optimal EV charging locations that minimize range anxiety and detour time for urban drivers [18].

Figure 4 summarizes the main results of the systematic literature review and shows how the analyzed studies address supply chain impacts, technological and infrastructural developments, and policy-related aspects of electromobility. At the same time, the Figure also indicates that only a smaller part of the literature connects these findings with routing-oriented models such as VRP or EV-VRP, especially in the context of urban last-mile delivery. This highlights the motivation of the present research, which integrates EV-specific operational constraints into a cost-based parcel-delivery optimization framework.

In summary, previous studies have examined numerous aspects of electric vehicle systems, battery technology, and charging infrastructure. Although a wide range of analytical tools have been applied, such as simulation, predictive modelling, GIS-based approaches, and mathematical optimization, only a limited subset of the literature explicitly incorporates routing constraints or provides modeling foundations suitable for EV-based last-mile delivery.

This gap further justifies the development of the routing-oriented, cost-based EV-VRP framework introduced in the present research.

2.3. Research Gap Identification

Due to the large amount of research in the field of electromobility and logistics, the most important and relevant literature must be summarized to establish a foundation for understanding the operation of sustainable urban delivery systems before elaborating on the model and optimization solution framework.

Table 1. The main topics of electromobility and urban logistics-related research fields.

Main Research Field	Key Focus Areas	Representative Studies
System and Supply Chain Impacts	Urban delivery systems; Grid interaction (V2G); Battery life-cycle and reverse logistics; Feasibility of EV-based operations	Lygizos et al. [12]; Peña et al. [17]; Király et al. [62]; Ali et al. [63]; Chudy et al. [64]
Technology and Smart Infrastructure	Smart charging; AI-based grid support; Infrastructure siting; Digitalization; Physical Internet (PI) principles; Infrastructure innovation	Gönül et al. [23]; Saqib et al. [24]; Połecki et al. [25]; Li et al. [37]; Peng et al. [38]; Shi et al. [40] Flocea et al. [57]; Mele et al. [67]
Environmental and Regulatory Aspects	Sustainability; Policy frameworks; Energy transition effects; Regulatory constraints	Struzewska et al. [42]; Mazur et al. [43]; Zadorożny et al. [44]; Baraniak et al. [49]; Skrúcaný et al. [51]
Analytical Methods and Decision Support	Simulation; GIS and modeling; Optimization; Decision support	Sierpiński et al. [18]; Rettenmeier et al. [58]; Lopes et al. [59]; Kryzia et al. [61]; Kłos et al. [60]; Lazzeroni et al. [65]
Parcel-Delivery and Vehicle Routing Modeling Foundations	Vehicle capacity; Service-time and time-window constraints; EV energy feasibility; VRP settings	Erdelic et al. [6]; Peña et al. [17]; Gülmez et al. [19]; Beyazıt et al. [21]; Zamal et al. [52]; Zhang et al. [53]; Lehmann et al. [56]

represents the main topics of the above-mentioned electromobility and urban logistics-related research fields. The identified research directions and methodological approaches provide the conceptual basis and practical justification for the model developed in this study.

However, despite the breadth of the reviewed literature, several methodological limitations remain evident—particularly in how electromobility-related findings are connected to routing-oriented decision problems. Most existing studies address technological, infrastructural or environmental aspects in isolation, while only a smaller group incorporates route design, capacity constraints, time windows or EV-specific energy limitations into a coherent VRP or EV-VRP modeling logic.

Besides the general electromobility literature, several parcel-delivery and VRP-based studies support the modeling structure used in this paper. These works commonly apply route length, vehicle capacity, time-window constraints, and, in the case of EVs, energy-related feasibility conditions. However, only a limited subset of these studies integrates these routing constraints with detailed cost structures and region-specific EV infrastructure characteristics, which are essential for realistic last-mile planning in medium-sized cities. These elements also appear in the present cost-based model, which therefore follows established modeling principles used in last-mile delivery optimization while adapting them to the regional characteristics of the case study. At the same time, the literature offers limited evidence on cost-based EV-VRP models explicitly developed for regional parcel-delivery conditions, especially in medium-sized Central European cities. This gap provides the methodological motivation for the modeling framework developed in this study.

The main goal of this paper is to develop a cost-based operational model that integrates electromobility and logistics for finished product delivery in the Miskolc area. The model focuses on optimizing route design, vehicle allocation, and charging infrastructure under regional economic and infrastructural constraints.

The main contributions of this manuscript are the following. First, it presents a cost-based modeling framework for urban parcel delivery that embeds EV-specific routing constraints into VRP-oriented operational planning. Second, it introduces a set of decision variables and constraints that reflect regional depot structures, delivery clusters, and EV-specific limitations. Third, a case study from the Miskolc region demonstrates how the model can be applied to evaluate operational costs and support practical delivery-planning decisions.

This paper is organized as follows: Section 1 describes the research background and goals, highlighting the main challenges of integrating electromobility into urban logistics systems. Section 2 presents a literature review, which summarizes the related results in electromobility, green logistics, and urban parcel delivery. Section 3 describes the model framework of optimization of cost-based urban delivery problems. Section 4 provides numerical results to demonstrate the applicability of the proposed model in Miskolc areas. Section 5 discusses the conclusions and further research directions.

3. Materials and Methods

This Section introduces a mathematical model of traditional parcel-handling processes in urban logistics. Just-in-time and just-in-sequence supplies support internal and external logistics deliveries, especially those parcel providers who are directly involved in urban delivery processes. The proposed formulation is a mathematical cost-optimization model designed to minimize the total operational cost of the urban parcel-delivery system.

The present model extends the previously published electromobility-based parcel-delivery cost framework [68] by incorporating regional delivery clusters, EV-specific operational constraints, and a more detailed representation of infrastructure-related cost components, thereby integrating urban and regional delivery processes into a single structure. This operational model is formulated as a Mixed-Integer Linear Programming (MILP) model, where binary variables represent routing and charging decisions and continuous variables describe parcel flows, service times, and energy levels.

Supply chain participants require up-to-date information to be able to fulfill delivery requirements and related logistical tasks. It requires developing a cost-based model where all participants can communicate with each other and serve each other’s needs—based on the sequenced customer orders and different service levels [3]. The goal is to find the optimal parcel delivery operations of the Miskolc Depot, which is responsible for processing customer requests and delivery resource allocations to each route.

As Figure 5 represents, the logistics system structure elements must be defined, which basically determine the supply chain’s delivery material flow. The general protocol of logistics processes can be structured into the following key stages: sender, pickup address, handling tour, sorting depot, operation center processing, delivery depot, delivery tour, delivery address, and recipient. It is important to note that the pickup and delivery address may be different from the sender or recipient because they are not necessarily identical and may involve dedicated third- or fourth-party logistics services providers or parcel points.

The system structure represents the finished products distribution material in the urban regional delivery supply chain, such as in Miskolc, Hungary or Northern Hungary region. This network structure forms the mathematical basis for the cost-minimization model presented in the subsequent equations. Within the frame of this article, the local delivery problems are known.

The following vector describes the groups of components in the relation matrix, hence, the number of supply chain decisions elements:

$${\mathrm{V}}_{\mathsf{\psi }}=\left[{\mathrm{S}}_{\mathrm{i}},{PA}_{\mathrm{j}},{HT}_{\mathrm{k}},{SD}_{\mathrm{m}},{OC}_{\mathrm{p}},{DD}_{\mathrm{q}},{DT}_{\mathrm{r}},{DA}_{\mathrm{s}},R_{\mathrm{t}}\right]$$

(1)

Each element of the ${\mathrm{V}}_{\mathsf{\psi }}$ corresponds to one functional logistics component. These parameters are shown in

Table 2. Notation and decision variables of the MILP-based urban parcel delivery model.

Parameter	Description	Index Range
Model elements
${\mathrm{S}}_{\mathrm{i}}$	$\mathrm{Sender}—\mathrm{decision} \mathrm{or} \mathrm{load} \mathrm{variable} \mathrm{for} \mathrm{shipment} \mathrm{origin} \mathrm{i}$	$i=1,\dots , n_{S_i}$
${PA}_{\mathrm{j}}$	$\mathrm{Pickup} \mathrm{Address}—\mathrm{parcel} \mathrm{collection} \mathrm{or} \mathrm{activation} \mathrm{variable} j$	$j=1,\dots , n_{PA}$
${HT}_{\mathrm{k}}$	$\mathrm{Pickup} \mathrm{Handling} \mathrm{Tour}—\mathrm{route} \mathrm{utilization} \mathrm{variable} k$	$k=1,\dots , n_{HT}$
${SD}_{\mathrm{m}}$	$\mathrm{Sorting} \mathrm{Depot}—\mathrm{sorting} \mathrm{or} \mathrm{transfer} \mathrm{variable} m$	$m=1,\dots , n_{SD}$
${OC}_{\mathrm{p}}$	$\mathrm{Operation} \mathrm{Center}—\mathrm{scheduling} \mathrm{and} \mathrm{coordination} \mathrm{variable} p$	$p=1,\dots , n_{OC}$
${DD}_{\mathrm{q}}$	$\mathrm{Delivery} \mathrm{Depot}—\mathrm{local} \mathrm{cluster} \mathrm{assignment} \mathrm{variable} q$	$q=1,\dots , n_{DD}$
${DT}_{\mathrm{r}}$	$\mathrm{Delivery} \mathrm{Tour}—\mathrm{vehicle} \mathrm{route} \mathrm{activation} \mathrm{variable} r$	$r=1,\dots , n_{DT}$
${DA}_{\mathrm{s}}$	$\mathrm{Delivery} \mathrm{Address}—\mathrm{delivery} \mathrm{completion} \mathrm{variable} s$	$s=1,\dots , n_{DA}$
$R_{\mathrm{t}}$	$\mathrm{Recipient}—\mathrm{demand} \mathrm{fulfillment} \mathrm{variable} t$	$t=1,\dots , n_R$
$v$	Vehicle Category	$v=1, \dots , n_v$
$\tau$	Delivery order within route	$\tau =1, \dots , n_{\tau }$
$S$	$\mathrm{Set} \mathrm{of} \mathrm{delivery} \mathrm{addresses}—D{A}_s$ must be assigned to delivery tours	-
$R$	$\mathrm{Set} \mathrm{of} \mathrm{delivery} \mathrm{tours}—D{T}_{\mathrm{r}}$ considered in the model	-
${load}_s$	$\mathrm{Total} \mathrm{parcel} \mathrm{load} \mathrm{assigned} \mathrm{to} \mathrm{delivery} \mathrm{tour} \mathrm{r}$ $\mathrm{for} \mathrm{delivery} \mathrm{address} s$ used in the capacity constraint in Equation (11)	$r,s$
$Q_{veh}(r)$	$\mathrm{Maximum} \mathrm{carrying} \mathrm{capacity} \mathrm{of} \mathrm{the} \mathrm{electric} \mathrm{vehicle} \mathrm{used} \mathrm{on} \mathrm{delivery} \mathrm{tour} r$; the load on each tour cannot exceed this limit (Equation (11))	$r$
$L_{ij}$	$\mathrm{Distance} \mathrm{between} \mathrm{node} i \mathrm{and} \mathrm{node} j$	$i,j$
$e_r$	$\mathrm{Energy} \mathrm{consumption} \mathrm{rate} \mathrm{of} \mathrm{the} \mathrm{electric} \mathrm{vehicle} \mathrm{on} \mathrm{delivery} \mathrm{tour} r$	$r$
$E_r^{max}$	Maximum battery capacity available for tour r	$r$
${\eta }_{rs}$	$\mathrm{Amount} \mathrm{of} \mathrm{energy} \mathrm{recharged} \mathrm{at} \mathrm{station} \mathrm{s}$ during tour r	$r,s$
$d_{F_{isv\tau }}$	$\mathrm{Forecasted} (\mathrm{requested}) \mathrm{parcel} \mathrm{quantity} \mathrm{from} \mathrm{origin} \mathrm{i}$ $\mathrm{to} \mathrm{delivery} \mathrm{address} s, \mathrm{served} \mathrm{by} \mathrm{vehicle} \mathrm{category} \mathrm{v}$ in sequence τ	$i,s,v,\tau$
$d_{B_{jtv\tau }}$	$\mathrm{Delivered} \mathrm{parcel} \mathrm{quantity} \mathrm{from} \mathrm{node} \mathrm{j}$$\mathrm{to} \mathrm{recipient} \mathrm{t}$$, \mathrm{served} \mathrm{by} \mathrm{vehicle} \mathrm{category} v$ in sequence τ	$j,t,v,\tau$
$t_{F_{isv\tau }}$	$\mathrm{Requested} \mathrm{service} \mathrm{time} \mathrm{at} \mathrm{delivery} \mathrm{address} s$	$i,s,v,\tau$
$t_{B_{jtv\tau }}$	$\mathrm{Actual} \mathrm{delivery} \mathrm{completion} \mathrm{time} \mathrm{at} \mathrm{recipient} t$	$j,t,v,\tau$
$a_{j\tau }$	$\mathrm{Start} \mathrm{of} \mathrm{allowable} \mathrm{service} \mathrm{time} \mathrm{window} \mathrm{for} \mathrm{delivery} \mathrm{point} j$	j,τ
$b_{j\tau }$	$\mathrm{End} \mathrm{of} \mathrm{allowable} \mathrm{service} \mathrm{time} \mathrm{window} \mathrm{for} \mathrm{delivery} \mathrm{point} j$	j,τ
Decision variables
$y_{ij}$	$\mathrm{Binary} \mathrm{decision} \mathrm{variable}—\mathrm{indicating} \mathrm{whether} \mathrm{flow} \mathrm{from} \mathrm{node} i$ $\mathrm{to} \mathrm{node} j$ is activated	$i, j$
$q_{ij}$	$\mathrm{Continuous} \mathrm{decision} \mathrm{variable}—\mathrm{representing} \mathrm{the} \mathrm{quantity} \mathrm{transported} \mathrm{from} \mathrm{node} i$$\mathrm{to} \mathrm{node} j$	$i, j$
$Z_{rs}$	$\mathrm{Binary} \mathrm{decision} \mathrm{variable}—1 \mathrm{if} \mathrm{tour} r$$\mathrm{performs} \mathrm{intermediate} \mathrm{charging} \mathrm{at} \mathrm{station} s$	$r,s$

.

To follow standard modeling conventions, the assignment and flow components are represented by two separate decision variables:

$${\mathrm{y}}_{ij}\in \left\{\mathrm{0,1}\right\}$$

(2)

This is a binary variable indicating whether the connection or assignment between nodes is $i$ and $j$ active.

$${\mathrm{q}}_{ij} \ge 0$$

(3)

This is a continuous variable representing the quantity, load, or flow transported between nodes $i$ and $j$.

The ${\mathrm{y}}_{ij}$ and ${\mathrm{q}}_{ij}$ variables represent the assignment and routing relationships between network nodes in Northern Hungary.

In the extended model, the indices $i, j, s$ and $t$ may represent different hierarchical nodes of the delivery network.

The origin $i$ and destination $j$ can refer to different depot levels, while the delivery address $s$ and recipient $t$ can vary if multiple end users are aggregated under a single delivery zone.

Additionally, $\mathrm{v}$ denotes vehicle categories, and $\tau$ represents the delivery sequence order in each route.

The objective of the optimization model is to minimize the total operational cost of the electromobility-based parcel delivery system. The total cost function has five main components:

$${\mathrm{C}}_T={\mathrm{C}}_{veh}+{\mathrm{C}}_{drv}+{\mathrm{C}}_{inf}+{\mathrm{C}}_{op}+{\mathrm{C}}_{nat}\to min$$

(4)

The ${\mathrm{C}}_T$ variable represents the operational costs of the delivery systems. These costs determine the overall urban parcel delivery processes and include the most related cost components, as summarized in

Table 3. Electromobility-based urban parcel delivery model operational costs.

Parameter	Description
${\mathrm{C}}_{veh}$	Specific cost of vehicle operation (distance, maintenance, etc.)
${\mathrm{C}}_{drv}$	Specific cost of driver hours and labor
${\mathrm{C}}_{inf}$	Infrastructure use and charging costs
${\mathrm{C}}_{op}$	Operation center planning, IT, and handling costs
${\mathrm{C}}_{nat}$	Environmental and energy cost of EV operations

. These parameters are essential for analyzing and comparing different delivery strategies.

Each cost component is defined as follows. Vehicle costs include both fixed and variable expenditures related to the type of vehicle and its utilization across tours, which are shown in

Table 4. Electromobility-based urban parcel delivery model—vehicles costs.

Parameter	Description
$C_{veh}^{type}$	Fixed cost of vehicle type (EV, van, cargo bike, etc.)
$C_{main}$	Specific cost of maintenance and service
$C_{dist}$	Specific cost of tour distance

:

$${\mathrm{C}}_{veh}=\sum _r(C_{veh}^{type}({DT}_{\mathrm{r}})+\sum _r{C}_{main}({DT}_{\mathrm{r}})+\sum _r{C}_{dist}({DT}_{\mathrm{r}}))$$

(5)

Driver costs represent the direct and indirect labor costs of operating each delivery solution, collected in

Table 5. Electromobility-based urban parcel delivery model—driver costs.

Parameter	Description
$C_{hr}$	Fixed driver salary costs/hourly
$C_{ins}$	Specific insurance and employment-related costs
$C_{misc}$	Additional driver-related costs (parking, tolls, etc.)

. These costs include both fixed and variable labor components associated with parcel handling and delivery operations.

$${\mathrm{C}}_{drv}=\sum _r(C_{hr}({DT}_{\mathrm{r}})+\sum _r{C}_{ins}({DT}_{\mathrm{r}})+\sum _r{C}_{misc}({DT}_{\mathrm{r}}))$$

(6)

Infrastructure costs (see

Table 6. Electromobility-based urban parcel delivery model—infrastructure costs.

Parameter	Description
$C_{ins}$	Installation or rental costs of a delivery depot or charger
$C_{oper}$	Operational and maintenance costs per facility
$C_{wait}$	Cost of vehicle waiting time for loading or charging
$C_{energy}$	Electricity or fuel cost of refueling/charging

) cover all operating costs which are related to physical depots, charging stations, and facility use:

$${\mathrm{C}}_{\mathrm{i}\mathrm{n}\mathrm{f}}=\sum _{\mathrm{q}}({\mathrm{C}}_{\mathrm{i}\mathrm{n}\mathrm{s}}({DD}_{\mathrm{q}})+\sum _{\mathrm{q}}({\mathrm{C}}_{\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}}({DD}_{\mathrm{q}})+\sum _{\mathrm{q}}({\mathrm{C}}_{\mathrm{w}\mathrm{a}\mathrm{i}\mathrm{t}}({DD}_{\mathrm{q}})+\sum _{\mathrm{q}}({\mathrm{C}}_{\mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y}}({DD}_{\mathrm{q}}))$$

(7)

These include rental or depreciation costs of depots and warehouses, maintenance and utility costs of operational facilities, and the installation and energy costs of charging infrastructure for electric vehicles.

Operation center costs include all management, digital, and coordination functions necessary for the logistics system to operate efficiently. These are not physical infrastructure costs, but organizational and IT-related expenditures. The collected costs are shown in

Table 7. Electromobility-based urban parcel delivery model—operation center costs.

Parameter	Description
$C_{hand}$	Parcel handling and sorting costs
$C_{smart}$	Smart and digital technology costs
$C_{cross}$	Transport coordination costs
$C_{plan}$	Delivery tour planning costs
$C_{cust}$	Customer management and service systems costs
$C_{clus}$	Specific cost of cluster management and inverse transportation

.

$${\mathrm{C}}_{\mathrm{o}\mathrm{p}}=\sum _{\mathrm{p}}({\mathrm{C}}_{\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{d}}({OC}_{\mathrm{p}})+\sum _{\mathrm{p}}({\mathrm{C}}_{\mathrm{s}\mathrm{m}\mathrm{a}\mathrm{r}\mathrm{t}}({OC}_{\mathrm{p}})+\sum _{\mathrm{p}}({\mathrm{C}}_{\mathrm{c}\mathrm{r}\mathrm{o}\mathrm{s}\mathrm{s}}({OC}_{\mathrm{p}})+\sum _{\mathrm{p}}({\mathrm{C}}_{\mathrm{p}\mathrm{l}\mathrm{a}\mathrm{n}}({OC}_{\mathrm{p}})+\sum _{\mathrm{p}}({\mathrm{C}}_{\mathrm{c}\mathrm{u}\mathrm{s}\mathrm{t}}({OC}_{\mathrm{p}})+\sum _{\mathrm{p}}({\mathrm{C}}_{\mathrm{c}\mathrm{l}\mathrm{u}\mathrm{s}}({OC}_{\mathrm{p}}))$$

(8)

This term represents the environmental and energy externalities associated with the electromobility system (see

Table 8. Electromobility-based urban parcel delivery model—natural and environmental costs.

Parameter	Description
$C_{em}$	Specific cost of emission rates/route
$C_{net}$	Network cost of electricity/delivery provider
$C_{uns}$	Specific cost of unscheduled natural resources/delivery provider

):

$${\mathrm{C}}_{nat}=\sum _r\sum _s(C_{em}({DT}_{\mathrm{r}}{,DA}_{\mathrm{s}})+\sum _p{C}_{net}({OC}_{\mathrm{p}})+\sum _q{C}_{uns}({DD}_{\mathrm{q}}))$$

(9)

Mathematical constraints ensure the feasibility and operational consistency of the model under electromobility conditions. The following three constraints are defined to represent the basic operational and physical restrictions of the model:

$$\forall \mathrm{s}\in \mathrm{S}:\sum _{r\in \mathbb{R}}{DT}_{\mathrm{r}}, {DA}_{\mathrm{s}} =1$$

(10)

Each delivery address $\mathrm{D}{\mathrm{A}}_{\mathrm{s}}$ must be assigned to exactly one delivery tour $\mathrm{D}{\mathrm{T}}_{\mathrm{r}}$. This constraint ensures the fulfillment of the customer request; no delivery address is omitted or served multiple times. A value of one means the address $\mathrm{s}$ is served by route $\mathrm{r}$, otherwise it is zero.

The total load ${load}_{s }$ assigned to each delivery tour $r$ must not exceed the maximum carrying capacity $Q_{veh}(r)$ of the electric vehicle used on that route. This means that every vehicle can transport only as many parcels as its technical limit allows.

$$\forall r\in \mathbb{R}:\sum _{s\in \mathrm{S}}{load}_s{(DT}_{\mathrm{r}},{ DA}_{\mathrm{s}}) \le Q_{veh}(r)$$

(11)

This constraint ensures that each electric vehicle operates within its battery energy or driving range limit. It guarantees route feasibility under electromobility conditions. If intermediate charging is allowed, the constraint can be extended as follows:

$$\forall r\in \mathbb{R}:\sum _{ij \in A}{e}_r{L}_{ij}{y}_{ijr}- \sum _{s\in S}{\eta }_{rs}{Z}_{rs}\le E_r^{max}$$

(12)

The total energy consumption $E_r$ of each delivery tour $r$ must not exceed the maximum available battery capacity $E_r^{max}$ of the electric vehicle operating on that route. Here, $L_{ij}$ the route distance and $y_{ijr}$ the binary routing variables equal 1 if vehicle $r$ travels directly from node $i$ to node $j$; otherwise, they equal 0.

When intermediate charging is possible, the model also considers the energy replenished at charging stations, represented by the binary decision variable $Z_{rs}$ (which equals 1 if vehicle $r$ charges at station $s$ and the corresponding recharged energy amount ${\eta }_{rs}$).

The demand constraint ensures that the quantity of parcels delivered fully satisfies all customer requests between active origin and destination nodes.

$$\forall \mathrm{i},\mathrm{j},\mathrm{s},\mathrm{t},v,\tau :\left|d_{F_{isv\tau }}-d_{B_{jtv\tau }}\right|$$

(13)

$d_{F_{isv\tau }}$ denotes the forecasted (requested) parcel quantity from origin node $\mathrm{i}$ to delivery address $s$, and $d_{B_{jtv\tau }}$ the delivered parcel quantity from node $\mathrm{j}$ to recipient $\mathrm{t}$ both transported by vehicle of category $v$ in delivery sequence $\tau$.

This constraint expresses the difference between requested and delivered quantities within the delivery network, which is minimized through the operational optimization process.

$$\forall \mathrm{i},\mathrm{j},\mathrm{s},\mathrm{t},v,\tau :\left|t_{F_{isv\tau }}-t_{B_{jtv\tau }}\right|$$

(14)

The $t_{F_{isv\tau }}$ variable refers to the requested service time at delivery address $\mathrm{s}$, while $t_{B_{jtv\tau }}$ denotes the actual delivery completion time at recipient $t$.

$$\forall \mathrm{j}\tau :a_{j\tau } \le t_{B_{j\tau }}\le b_{j\tau }$$

(15)

Each delivery point $\mathrm{j}$ operates within a predefined service window $\left[a_{j\tau } b_{j\tau }\right]$.

The model ensures that the actual delivery time $t_{B_{j\tau }}$ remains within this allowed time interval. For the Northern Hungary case study, urban delivery points are assigned 2 h service windows (e.g., 09:00–11:00 or 13:00–15:00), while regional and rural delivery zones use longer intervals of 3–4 h within the overall daily operational horizon (08:00–20:00).

The result of the above ideas is that a cost function for evaluating the urban delivery strategy can be defined as a cost function that seeks to minimize the overall cost of the electromobility-based parcel delivery participants:

$$\begin{array}{c}{\mathrm{C}}_{\mathrm{T}}={\displaystyle \sum _{\mathrm{r}}}({\mathrm{C}}_{\mathrm{v}\mathrm{e}\mathrm{h}}^{\mathrm{t}\mathrm{y}\mathrm{p}\mathrm{e}}({DT}_{\mathrm{r}})+{\displaystyle \sum _{\mathrm{r}}}{\mathrm{C}}_{\mathrm{m}\mathrm{a}\mathrm{i}\mathrm{n}}({DT}_{\mathrm{r}})+{\displaystyle \sum _{\mathrm{r}}}{\mathrm{C}}_{\mathrm{d}\mathrm{i}\mathrm{s}\mathrm{t}}({DT}_{\mathrm{r}}))+{\displaystyle \sum _{\mathrm{r}}}({\mathrm{C}}_{\mathrm{h}\mathrm{r}}({DT}_{\mathrm{r}})+{\displaystyle \sum _{\mathrm{r}}}{\mathrm{C}}_{\mathrm{i}\mathrm{n}\mathrm{s}}({DT}_{\mathrm{r}})+{\displaystyle \sum _{\mathrm{r}}}{\mathrm{C}}_{\mathrm{m}\mathrm{i}\mathrm{s}\mathrm{c}}({DT}_{\mathrm{r}}))\\ +{\displaystyle \sum _{\mathrm{q}}}({\mathrm{C}}_{\mathrm{i}\mathrm{n}\mathrm{s}}({DD}_{\mathrm{q}})+{\displaystyle \sum _{\mathrm{q}}}({\mathrm{C}}_{\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}}({DD}_{\mathrm{q}})+{\displaystyle \sum _{\mathrm{q}}}({\mathrm{C}}_{\mathrm{w}\mathrm{a}\mathrm{i}\mathrm{t}}({DD}_{\mathrm{q}})+{\displaystyle \sum _{\mathrm{q}}}({\mathrm{C}}_{\mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y}}({DD}_{\mathrm{q}}))\\ +{\displaystyle \sum _{\mathrm{p}}}({\mathrm{C}}_{\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{d}}({OC}_{\mathrm{p}})+{\displaystyle \sum _{\mathrm{p}}}({\mathrm{C}}_{\mathrm{s}\mathrm{m}\mathrm{a}\mathrm{r}\mathrm{t}}({OC}_{\mathrm{p}})+{\displaystyle \sum _{\mathrm{p}}}({\mathrm{C}}_{\mathrm{c}\mathrm{r}\mathrm{o}\mathrm{s}\mathrm{s}}({OC}_{\mathrm{p}})+{\displaystyle \sum _{\mathrm{p}}}({\mathrm{C}}_{\mathrm{p}\mathrm{l}\mathrm{a}\mathrm{n}}({OC}_{\mathrm{p}})\\ +{\displaystyle \sum _{\mathrm{p}}}({\mathrm{C}}_{\mathrm{c}\mathrm{u}\mathrm{s}\mathrm{t}}({OC}_{\mathrm{p}})+{\displaystyle \sum _{\mathrm{p}}}({\mathrm{C}}_{\mathrm{c}\mathrm{l}\mathrm{u}\mathrm{s}}({OC}_{\mathrm{p}}))+{\displaystyle \sum _{\mathrm{r},\mathrm{s}}}({\mathrm{C}}_{\mathrm{e}\mathrm{m}}({DT}_{\mathrm{r}}{DA}_{\mathrm{s}})+{\displaystyle \sum _{\mathrm{p}}}{\mathrm{C}}_{\mathrm{n}\mathrm{e}\mathrm{t}}({OC}_{\mathrm{p}})\\ +{\displaystyle \sum _{\mathrm{q}}}{\mathrm{C}}_{\mathrm{u}\mathrm{n}\mathrm{s}}({DD}_{\mathrm{q}}))\to \mathrm{m}\mathrm{i}\mathrm{n}\end{array}$$

(16)

The model integrates objective function components designed to improve the efficiency of supply chain operations. The electromobility-based urban parcel delivery system operates under practical limitations that ensure route feasibility, vehicle utilization efficiency, and sustainable energy usage.

In the numerical analysis, the performance of the mathematical model for electric vehicle routing was evaluated in the Microsoft Excel environment using custom VBA routines. Furthermore, the results were validated by analyzing the model outputs and identifying the optimal routing and charging strategy derived from the proposed formulation.

4. Results

This case study focuses on the evaluation of the presented operational cost-based logistics model. The analysis highlights the cost composition such as vehicle usage, driver labor, infrastructure, operational management, and environmental indicators.

In this case, several model variants can be defined, so it is necessary to narrow down and several examples for testing. The analysis applies simulated company data to represent the operational environment. The simulated input data was generated based on company-level operational information, using parameter ranges that reflect typical values observed in practice. The minimum and maximum intervals applied in

Table 9. Example of the generated input values.

Depot	Demands	Vehicles	Delivery Orders	${\mathit{d}}_{{\mathit{F}}_{\mathit{i}\mathit{s}\mathit{v}\mathit{\tau }}}$ [pcs]	${\mathit{d}}_{{\mathit{B}}_{\mathit{j}\mathit{t}\mathit{v}\mathit{\tau }}}$ [pcs]	${\mathit{t}}_{{\mathit{F}}_{\mathit{i}\mathit{s}\mathit{v}\mathit{\tau }}}$ [pcs]	${\mathit{t}}_{{\mathit{B}}_{\mathit{j}\mathit{t}\mathit{v}\mathit{\tau }}}$ [h]	${\mathit{t}}_{{\mathit{B}}_{\mathit{j}\mathit{\tau }}}$ [h]
1	1	1	1	1	1	0.38	0.13	0.08
.	.	.	.	.	.	.	.	.
2	6	5	10	27	23	0.21	0.50	0.17
.	.	.	.	.	.	.	.	.
3	15	5	15	23	20	0.04	0.25	0.04

correspond to depot capacities, delivery zone structures, service time windows, and EV performance limits that are characteristic of the actual last-mile operations analyzed in the study. Demand volumes, delivery periods, and cost components were produced within these empirically grounded ranges to ensure that the resulting dataset remains consistent with real operational conditions while allowing systematic evaluation of the mathematical model.

Table 9 shows that these cost-based urban parcel delivery input values are generated within predefined minimum and maximum ranges to capture the variability of market demand and specified indicators. There are three different Central Depots to fulfill the regional or cluster-based parcel demands within optimal routes in Northern Hungary.

To support the interpretation of the results, the Table also summarizes the parcel volumes, available vehicles, and time-related parameters that form the basis of the case study evaluation.

The next step is to summarize values according to the SUM $t$ delivery orders within each route (see

Table 10. Example of the generated delivery orders values.

SUM t	${\mathit{d}}_{{\mathit{F}}_{\mathit{i}\mathit{s}\mathit{v}\mathit{\tau }}}$ [pcs]	${\mathit{d}}_{{\mathit{B}}_{\mathit{j}\mathit{t}\mathit{v}\mathit{\tau }}}$ [pcs]	${\mathit{t}}_{{\mathit{F}}_{\mathit{i}\mathit{s}\mathit{v}\mathit{\tau }}}$ [pcs]	${\mathit{t}}_{{\mathit{B}}_{\mathit{j}\mathit{t}\mathit{v}\mathit{\tau }}}$ [h]	${\mathit{t}}_{{\mathit{B}}_{\mathit{j}\mathit{\tau }}}$ [h]
1 1 1 1	186	177	2.13	2.00	1.21
.	.	.	.	.	.
1 3 3 3	239	235	2.71	2.46	1.13
.	.	.	.	.	.
1 5 3 5	283	255	2.79	2.50	0.88

). A delivery route in this case represents the ordered sequence of delivery addresses assigned to the same vehicle within one cluster. The SUM $t$ values therefore show the total number of stops, distance, and time required for completing each delivery route.

The urban system element can also be determined for the value of the operational cost. The applied model satisfies the predefined operational constraints to ensure efficient vehicle utilization and balanced service performance across all delivery clusters.

Table 11. Details of delivery orders costs.

SUM t	${\mathbf{C}}_{\mathit{v}\mathit{e}\mathit{h}}$ [EUR]	${\mathbf{C}}_{\mathit{d}\mathit{r}\mathit{v}}$ [EUR]	${\mathbf{C}}_{\mathit{i}\mathit{n}\mathit{f}}$ [EUR]	${\mathbf{C}}_{\mathit{o}\mathit{p}}$ [EUR]	${\mathbf{C}}_{\mathit{n}\mathit{a}\mathit{t}}$ [EUR]	${\mathbf{C}}_{\mathit{t}\mathit{o}\mathit{t}}$ [EUR]
1 1 1 1	0.43	0.96	0.11	0.22	0.01	1.73
.	.	.	.	.	.
1 3 3 3	0.47	0.94	0.14	0.28	0.02	1.85
.	.	.	.	.	.
1 5 3 5	0.38	0.91	0.15	0.24	0.01	1.70

demonstrates that the related delivery combinations are evaluated by assigning operational costs to each sender–recipient relation.

Figure 6 illustrates how the total operational costs are broken into specific cost indicators and their percentage distribution across each order. According to these calculations, driver-related manpower (DRV) costs represent the largest share at 53%, confirming that last-mile delivery is particularly work-intensive. Vehicle operation (VEH) costs are the second biggest result at 25%, highlighting vehicle route distance and efficiency. Operational management (OP) costs account for 14%, while infrastructure and charging (INF) contribute 7% to the total operational cost. The environmental and energy component (NAT) remains marginal at 1%, as it reflects regulated externalities rather than direct operational expenditures.

Based on the sequenced route calculations, the total operational cost of the urban EV-based delivery system is 400.97 EUR, corresponding to a cost reduction of approximately 20% compared to traditional vehicle fleets operating under similar conditions. A detailed breakdown of the cost components is provided in Appendix A (Figure A1).

5. Discussion and Conclusions

The results of this study demonstrate that traditional vehicles remain a feasible alternative in cases where EV infrastructure is underdeveloped or initial investment costs are prohibitively high. These findings confirm the higher cost efficiency of EVs in urban delivery operations. The optimization of these operations addresses economic and environmental objectives aiming to mitigate emissions and seeking to utilize government funding.

The proposed model has practical value for Miskolc areas, where urban delivery systems operate under various infrastructure constraints and diverse customer demands. Integrating spatial and cost-based planning principles enhances efficiency and sustainability in urban logistics systems. The cost results outline how the different components appear in practice. Here, the driver-related costs form the largest share, efficiency, mainly depends on how the delivery routes are organized. With more balanced route sequencing, the grouping of adjacent delivery zones and the planned timing of charging can reduce unnecessary distance and waiting. In the case of Miskolc, these aspects appear through the city’s clustered delivery zones and the differing service densities, which the model uses to evaluate route alternatives and charging needs under local conditions. These aspects show that the model can support practical planning by helping identify where operational adjustments have the greatest impact. These results provide a useful foundation for adapting electromobility strategies to regional cities with limited EV infrastructure. There are several advantages and limitations linked to this study.

The following points highlight the main advantages of the research.

• Adaptable framework for cities with different levels of urban mobility and electric vehicle infrastructure development;
• Modeling regional urban delivery processes;
• Cost-based modeling and optimization which considers diverse demands, infrastructure, route planning, and resource allocation;
• Committed to EU’s sustainability and emission reduction.
• The main constraints and limitations of the study are outlined below. These factors highlight the methodological limitations of the model and indicate where further refinement or validation under real-world conditions may be needed.
• Simplified energy consumption methods to reflect average electric vehicle usage;
• Differences in charging infrastructure for different types of chargers and power capacities;
• Fluctuations in station availability due to temporary overloads or breakdowns;
• Extended traffic conditions influenced by congestion and road variability;
• Real-time network variability such as congestion, dynamic station availability affecting operational stability;
• Optimization is based on predefined patterns limiting real-time adaptation.

Although the case study is based on data from the Miskolc region, the model is not limited to this specific location. The clustered delivery structure, vehicle-related constraints, and cost components are comparable in other medium-sized cities. The approach can be applied in other urban or regional environments by tailoring the local depot layout, demand areas, and the availability of charging facilities.

The diversity of logistical processes highlights the necessity of new methods and solutions development within parcel delivery supply chains in urban mobility. During these developments, companies must strive to achieve optimal operational performance, which has become one of the most important and critical tasks in modern logistics, especially in the field of parcel delivery processes and related services.

However, further research is needed to address the design aspects of more complex urban delivery systems in electromobility and logistics by utilizing different algorithms and heuristics to solve NP-hard transportation problems.