Strategy for the Siting of Electric Vehicle Charging Stations for Parcel Delivery Service Providers

Abstract

The development of urban logistics is closely related to the inhabitants of specific areas. People are changing their shopping habits, hence the growing share of online shopping. The consequence of the foregoing is increased demand for parcel delivery to customers’ addresses as well as in-store and automated pick-up points. This has increased the number of couriers that provide parcel delivery services. To achieve sustainable development, there is a real need to convert conventional combustion engine fleets using greener solutions such as electric vehicles. However, introducing such vehicles on a large scale involves constraints such as the vehicle running range on a single charge, the price of a new fleet and of its maintenance, and the location of charging stations. This article focuses on the strategies for siting of charging stations intended to be used by the providers of parcel delivery services. Different strategies as well as a method for identifying optimal goal-specific locations have been proposed. The results show the limitations and challenges facing both the operators of parcel delivery services and local authorities that should also be involved in the development of charging points. Additionally, based on the method presented in the paper, operators can choose between different strategies depending on the particular requirements and local conditions.

1. Introduction

An increasing demand for logistic services can be observed worldwide. This is related to the changes in consumer behaviour [1,2]. The shift from traditional to online shopping has increased the demand for parcel delivery directly to customers’ homes, parcel shops, and automatic parcel pick-up points [3]. The current global situation (including the impact of the COVID-19 pandemic) has further increased the consumer interest in online shopping [4]. Attention should be paid to the need to use electric vehicles for delivery in urban areas in order to achieve sustainable mobility. This will allow continuity of delivery with a relatively low environmental impact.

The intensive development of the logistics sector along with the pursuit of fulfilment of consumer demand has increased the number of trips made by carriers as well as the share of the delivery vehicle fleet in the transport system [5,6]. The majority of vehicles in use feature conventional propulsion systems, which exert a negative impact on the environment through increased emissions and noise [7,8,9,10]. These vehicles often cover the last-mile section in densely populated areas, which is precisely why it is extremely important to protect the environment there [11]. Companies using traditional technologies know their benefits: hassle-free and fast refuelling at many points. However, the use of fossil fuels in road transport accounts for around 10.2% of the global greenhouse gas emissions [12].

One can attempt to overcome these negative effects by switching to eco-friendly means of transport such as electric vehicles [13]. An additional argument for introducing electric delivery vehicles into fleets may be the process of establishing green city zones that admit only vehicles driven by eco-friendly power transmission systems [14]. In order to reach their destination points (addresses or parcel machines), businesses that provide parcel shipment services will be forced to deploy electric fleets, thereby allowing them to perform their operations without constraints [15]. A review of challenges of electric vehicle adoption in last-mile parcel delivery can be found in [16]. Infrastructure barriers and limited range are among such challenges.

There is a growing number of available electric vehicles that enable parcel delivery to customers or dedicated machines. Based on a review of the literature on this subject, the vehicles of this type have been summarised in

Table 1. Examples of electric delivery vehicles and their technical parameters [17,18,19,20,21].

Name	Manufacturer	Carrying Capacity (kg)	Max Velocity (km/h)	Travel Range (km)	Battery Charging Time (h)	Battery (kWh)
Master E-Tech	Renault	No data	100	200	6 (7.4 kW)	52
E-Ducato	Fiat	1015	100	360	1.5 to 80% (50 kW) 8 to 100% (7 kW)	47 or 79
eSprinter	Mercedes	958	80	135	0.5 to 80% (80 kW) 6 to 100% (7.4 kW)	55
eVito	Mercedes	1003	80	265	6.5 (11 kW)	60
E-deliver 3	Maxus	1095	90	210–253	8.3 (7.4 kW)	50.23
e-Crafter	Volkswagen	1709	90	115	0.5 (50 kW CCS)	35.8
e-NV200 Furgon	Nissan	703	123	200	1 (50 kW CHAdeMO)	40
Kangoo Z. E.	Renault	600	130	230	4.5 (Type 2)	33
eTGE	MAN	950	90	115	0.5 (50 kW CCS)	35.8

along with their basic parameters. The term “electric vehicle” in this article refers to battery electric vans.

The parameters of the delivery vehicles provided above in Table 1 show how considerably they vary in terms of the operating range, which means that even fully charged vehicle batteries may not suffice to complete the delivery jobs planned for an entire workday. A courier company’s vehicle typically covers 100 to 350 km a day [22]. This all depends on the size and specifics of the area of its operations. By comparing the travel range of vehicles, one can attempt to adapt the tasks at hand to a given vehicle, but this is not always possible. A situation in which the vehicle is operated to the battery limits should be avoided. There are also external phenomena (e.g., traffic congestion and weather conditions) that need to be taken into account when performing daily tasks because they also affect the battery discharging [23,24,25]. Additionally, most vehicles have only slow charging service lines on board as a standard feature [26,27]. Currently, the average range of an electric delivery vehicle is ca. 200 km on one full charge [28]. The loading capacity of this type of vehicle is around 1000 kg [29]. The charging time from 0 to 100%, for example, is approx. 6 h for a 35 kWh battery with a 7.4 kW charger and 4 h for a 22 kW charger. The time needed for the unloading of parcels is typically 30 min; during this time, it is possible to recharge around 12.5% of the vehicle’s battery, which translates into a further 20 km [30].

Yet another matter to take into consideration is the fleet management policy of courier companies, which varies according to their needs. There are three basic delivery vehicle fleet handling strategies. One assumes that an in-house fleet of vehicles is built; in this case, once electric vehicles have been deployed, the company can equip its car park with devices to recharge the batteries to 100% overnight. Chargers distributed around the town will prove necessary when longer-distance jobs are to be performed. The company can also break a single supply chain down into smaller sections by assuming that batteries are recharged in the parking yard. Another strategy used by courier companies is hiring owners of private vehicles. Where this is the case, the employer may require that electric vehicles are to be used. This calls for a different approach to the deployment of charging stations because vehicles are not parked on the company premises at night. The third strategy involves operating a mixed fleet that is partially owned by the courier company while employing drivers with their own vehicles at the same time.

Therefore, an important part of the policy aimed at deploying electric delivery vehicles is to take care of the charging infrastructure, and this again entails different approaches [31,32,33]. Businesses may decide to use only the charging stations installed at the parking areas in the vicinity of their own distribution centres (thus limiting the extent of the vehicle operations) or to use the available open-access charging stations (which, however, involves the risk that they may be occupied as well as the distance to the parcel machines). With the above strategies in mind, it is worth investing in an in-house network of electric vehicle charging points because it allows businesses to become independent of unpredictable factors.

However, to succeed in deploying company-owned charging points across towns, one must first select adequate locations [34]. Based on the authors’ review of the methods employed when searching for optimal locations to install electric vehicle chargers, it has been established that the approaches applied to date have mainly considered the rationale behind building infrastructure intended for all inhabitants of a given area. Another approach was used in [35]. In this case, the authors conducted a strong literature review related to the parameters important to the charging station location problem. They identified 102 indicators. Finally, the number of indicators was reduced, and the main ones related in general to sustainable planning were the charging power, EV fleet distribution, existing charging infrastructure, average charging time, and population density.

The literature refers to a variety of concepts used to determine optimal sites for generally accessible electric vehicle charging stations [36]. The authors of [37] conducted a three-dimensional area analysis using a multi-criteria factor analysis and a hierarchical process analysis method. The paper elaborated upon a method used to determine the weights of specific factors; however, it disregarded the way in which ratings were assigned. A different approach to the method employed when determining the potential locations for charging stations was discussed in [38], in which the potential sites were analysed with reference to the existing infrastructure. A Voronoi diagram was used to divide the area subject to the analysis and to search for new charging station sites. In [39], the matter of accessibility of the existing charging stations was analysed against the distance in the road network. The authors of [40] proposed a two-stage method for charging station siting through an analysis first performed at a macro scale (in the case described, the territory was entirely in Hungary) and at the micro scale in the second stage (the Budapest district). The analysis they proposed was based on hexagon-shaped primary areas. A charging station siting method based on input data concerning the number of trips made and the vehicle use intensity was proposed in [41]. What was also attempted in that study was a search for locations intended for public and private charging points. The results of the relevant analyses were discussed with regard to large administrative regions. The article showed how to find a potential charging site within a large area. The method developed by the authors of this article made it possible to indicate an exact place suitable for a charging station with a tolerance of up to 10 m. A fuzzy hierarchical process analysis was employed in the study addressed in [42], the outcome of which was a comparison of the identified potential sites with the existing infrastructure of charging stations. Fifteen factors were grouped into three categories (environment, economy, and urban planning). The method presented in [43] was based on a multi-criteria analysis of input factors combined with a fuzzy hierarchical process analysis. The authors proposed 10 factors that affected the charging station siting potential of the analysed areas. Unlike in the other methods, each factor was analysed not only for whether or not it was observed in the area in question, but also in terms of the distance of the given area from the factor in a straight line, which provided grounds for scoring.

This article focuses on the strategies aimed at finding locations for charging stations to be used by the providers of parcel delivery services. Different strategies are proposed, and a method for identifying optimal locations depending on specific purposes is suggested. The analyses provided in the paper were performed using an open-source tool for geographic information management—QGIS. The results showed the limitations and challenges facing both the operators of parcel delivery systems and local authorities that should also be involved in the development of charging points. Additionally, based on the method presented in the article, operators can choose between different strategies depending on specific requirements and local conditions.

2. Method

With reference to a literature search, the method proposed by the authors assumed two main strategies for introducing a fleet of electric vehicles in courier companies: using a public network of charging stations and developing an in-house charging infrastructure and relying solely on these stations. Despite the existence of many different methods of locating charging stations, the specificity of the subject addressed to electric vans made it necessary to use a different approach. The methodology proposed in this paper contains a specific approach that allowed us to take into account not only the location of charging stations related to delivery parcels and the budget of the delivery company, but also the location process (including uniform distribution of the sited chargers). This means that the method will support the delivery company even when the driver is forced to charge the vehicle on the route because uniform distribution makes the distance between chargers shorter.

This method may prove useful to persons in charge of implementing green solutions at delivery companies. Depending on the specific needs, the method’s parameters can be modified and adjusted as demanded. A decision maker can additionally use this method to plan the growth of the in-house charging station network and the deployment of electric vehicles in stages. The method proposed requires the use of open-access data and the open-source QGIS software to perform calculations based on geographic information and data.

Figure 1 is a diagram representing the first strategy, in which public charging stations are assumed to be used. This strategy comprises two steps: the first entails testing the feasibility of using public stations, while the second makes it possible to determine the potential sites for in-house charging stations if the first step fails to meet the relevant expectations.

The former of the strategies in question requires the decision maker to conduct identification analyses (I) at the beginning of the first step in order to:

-

Locate the public charging stations (PCS);

-

Locate all parcel service points operated by a given company (both machines and special pick-up and drop-off points) (PSP).

Two main criteria must be defined at the start by the delivery company. The first is the selected number of parcel service point and charging station (NP) pairs. The second is the minimisation of the number of private charging stations implemented, which means that at first the public charging stations will be searched to find the pairs and then the planning of the company’s own charging stations will be realised.

The components of the model can be defined as follows:

$$M= <PCS,PSP,PL,PrCS>$$

(1)

where:

M—model of the system;

PCS—set of public charging stations;

PSP—set of parcel service points;

PL—set of parking lots;

PrCS—set of private charging stations (built by delivery company).

Each of the elements can be described by geographical coordinates:

$$\begin{gathered} PC{S}_i=\left(X_i,Y_i\right), PS{P}_j=\left(X_j,Y_j\right), P{L}_k=\left(X_k,Y_k\right), {PrCS}^{\prime }\subset PL i= \\ \left\{1,2,\dots ,a\right\}, j=\left\{1,2,\dots ,b\right\}, k=\left\{1,2,\dots ,c\right\}, a,b,c ϵ N \end{gathered}$$

(2)

where:

X—latitude;

Y—longitude.

In the case of PL, the coordinates refer to the centre of gravity of the parking space. PrCS (if needed) will be located in parking spaces (methodology assumption), meaning that set of PrCS will be a subset of PL.

A spatial analysis was performed to check the distance between the charging stations and the parcel service points. Given the current conditions, couriers tend to arrive as close as possible to such a point to avoid the extra time needed to handle parcels. A distance between charging points and parcel machines or pick-up and drop-off points of 50 m was assumed for a courier to be able to move parcels onto a special transport trolley as well as to place parcels inside the machine and to remove those already dispatched from the locker while the electric vehicle is charging. If the decision maker should modify their approach, the 50 metre value can be changed to any chosen one (while the strategy remains functional).

Firstly, the distance for each pair of $PC{S}_i$ and $PS{P}_j$ is calculated by using a formula addressed to the spherical distance of points on the Earth:

$$\underset{PC{S}_i}{\wedge }\underset{PS{P}_j}{\wedge }arc\cos \left(sin\left(Y_i\right)·sin\left(Y_j\right)+cos\left(Y_i\right)·cos\left(Y_j\right)·cos\left(X_j-X_i\right)\right)·6371$$

(3)

In this way, a distance matrix is built that—after checking the 50 m condition—identifies a subset of the system consisting of specific PSP and PCS:

$$PC{S}^{\prime }\subset PCS, PS{P}^{\prime }\subset PSP$$

(4)

where D_{Pu_ij} means distance between the selected pairs of PSP and PCS, and D₀ means a 50 m distance. In this case:

$$\begin{array}{l}\underset{\left(PS{P}_j, PC{S}_i\right)}{\wedge }\left(D_{Pu\_ij}\le D_0\to \left(PS{P}_j, PC{S}_i\right)\in S_{Pu}\right)\wedge (D_{Pu\_ij}>D_0\to \left(PS{P}_j, PC{S}_i\right)\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\notin S_{Pu})\end{array}$$

(5)

where:

$$S_{Pu}=\left\{\left(PS{P}_j, PC{S}_i\right)| PS{P}_j\in PSP,PC{S}_i\in PCS \right\}$$

(6)

where S_Pu—set of pairs of PSP and PCS that meet the condition of 50 m of distance.

The successive values (from the first step) found to meet a given distance condition are marked as points that enable vehicle charging and parcel handling at the same time. At this stage, the decision maker must have additional information about the size of the electric vehicle fleet planned to be deployed to estimate whether the number of designated points meets the relevant quantitative and spatial criteria. The quantitative criterion is associated with the capacity to recharge the vehicle fleet while other tasks are being performed. The spatial criterion pertains to how evenly the points that meet the distance condition are distributed, thereby enabling a larger area to be covered while delivering services. If either of the quantitative or spatial conditions are not met, the second step is initiated to ensure that electric vehicle drivers are capable of performing the tasks they have been assigned. This can be described as follows—check how many elements contain the S_Pu set and if it is enough according to the input assumption. Let n(S_Pu) be the cardinal number of set and:

$$\left(n\left(S_{Pu}\right)\ge N_P\to S_{Pu}=S_F\right)\wedge \left(n\left(S_{Pu}\right)<N_P\to S_{Pu}\ne S_F\right)$$

(7)

where:

S_F—set of final pairs of PSP and PCS;

N_P—selected number of pair of parcel service points and charging stations (first criteria of the method defined by delivery company).

If the number of S_Pu (pairs of PSP and PCS that meet the condition of 50 m of distance) is smaller than N_P, this means that next step of the method is needed—the selected number of private charging stations has to be implemented.

The second step involves establishing sites suitable for an in-house charging system to make up for the shortage in the generally available charging stations. An additional advantage of in-house charging stations is the possibility of rendering them available to employees after working hours when the operating model envisaged by the company involves employing drivers with their own vehicles at their disposal. Step two makes use of the information obtained in the first step and enables scouting for optimal charging station locations in areas devoid of such infrastructure. Additional information needed for this strategy to be implemented concerns publicly available parking spaces (PL), which can be obtained, for instance, from open sources such as OpenStreetMap. Next, a spatial analysis should be conducted to check the distance between the relevant parcel service points and free parking spaces. Based on the database of parking spaces, this stage is related to calculation of the spherical distance of points on the Earth for each pair of $PS{P}_j$ and $P{L}_k$:

$$\underset{PS{P}_j}{\wedge }\underset{P{L}_k}{\wedge }arc\cos \left(sin\left(Y_j\right)·sin\left(Y_k\right)+cos\left(Y_j\right)·cos\left(Y_k\right)·cos\left(X_k-X_j\right)\right)·6371$$

(8)

The result gives the opportunity to obtain a ranking of distances and identify the subset of PL that will contain only parking lots that meet the specific condition. For deliveries, it is important to locate new charging stations as close as possible to parcel service points (the charging process will be possible during pack reloading).

In order to meet the pre-set condition, a distance of less than 10 m was assumed to eliminate the additional walking distances to be covered by couriers. Let D_{Pr_ij} be the distance between the selected pairs of PSP and PL and D₁₀ be the 10 m distance. In this case:

$$\begin{array}{l}\underset{\left(PS{P}_j, P{L}_k\right)}{\wedge }\left(D_{\mathrm{Pr}\_ij}\le D_{10}\to \left(PS{P}_j, P{L}_k\right)\in S_{Pr}\right)\wedge (D_{\mathrm{Pr}\_ij}>D_{10}\to \left(PS{P}_j, P{L}_k\right)\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\notin S_{Pr})\end{array}$$

(9)

where:

$$S_{Pr}=\left\{\left(PS{P}_j, P{L}_k\right)| PS{P}_j\in PSP,P{L}_k\in PL \right\}$$

(10)

where S_Pr—set of pairs of PSP and PL that meet the condition of 10 m of distance (D₁₀).

The next step is to check how many elements contain the S_Pr set and whether that number is sufficient (together with the number of S_Pu) according to the input assumption. Let n(S_Pr) be the cardinal number of the set and:

$$\left(n\left(S_{Pu}\right)+n\left(S_{Pr}\right)\ge N_P\to S_{Pu}\cup S_{Pr}=S_F\right)\wedge \left(n\left(S_{Pr}\right)+n\left(S_{Pr}\right)<N_P\to S_{Pu}\cup S_{Pr}\ne S_F\right)$$

(11)

If the condition has not been met, the above value can be freely changed while keeping in mind that it may cause the employees’ working conditions to deteriorate. This step can be performed several times with the distance criteria increased each time (e.g., 10 m, 20 m, etc.). Usually, because of the high number of parking spaces in urban areas, one or two steps will probably result in a higher number of potential private charging station locations than the first assumption (N_P).

Another step involves a spatial analysis of the points found to comply with the distance condition. Considering the points thus established and the points identified in the first step of the strategy, one should make such a selection in order to make the area to be served as large as possible. To realise this step, the distance to the nearest neighbour is checked, and the ones with the shortest distance between each other are removed (one of the pairs from set $S_{Pr}$—the one that had the longest distance to the parking lot). This reduction process is repeated until the desired number of pairs of parcel service points and charging stations is reached.

Figure 2 illustrates the second strategy, whereby a courier company relies solely on its own charging stations.

The second strategy is fundamentally similar to the second step of the first strategy, while it differs from the latter in its complete independence from publicly available charging stations.

3. Case Study

Using the method proposed for the above two strategies, a case study was performed in the city of Katowice, located in the central part of Silesian Voivodeship in southern Poland. The case study considered machines operated by a delivery company where parcels can be dropped off and picked up. The delivery company in question was a nationwide operator. The territory subject to the analysis was narrowed down to that of a single city with an area of 164.6 km² and a population of about 290,000 [44]. Figure 3 shows the case study area against the entire country and the whole of the voivodeship (province), and it highlights the neighbouring towns that altogether form an urban area known as Metropolis GZM (the Metropolis of the Upper Silesia and the Dąbrowa Basin), which is inhabited by ca. 2 million people.

An additional advantage of the method proposed is that the number of charging stations can be adjusted to match the growing fleet of electric vehicles. For purposes of the case study, it was additionally assumed that only 20% of the machines where parcels could be dropped off and picked up would be covered by the charging stations’ operating range. This assumption can be modified as needed. In the case analysed, such a level of implementation was assumed on account of the initial phase of electromobility development. Additionally, the stations should be evenly distributed over the area so that the supply chain tasks can be performed in the best possible way. This aspect is extremely important because a fleet in motion handles not only parcel machines, but also reaches specific addresses of businesses and households.

3.1. Method’s Input Data

Data concerning public electric vehicle charging stations were downloaded from the OpenStreetMap portal using the amenity key of charging_station. The data were then manually verified. There are currently only 73 electric vehicle charging stations in the city of Katowice.

Next, locations of the machines used to drop off and pick up parcels were identified. There are 209 such points within the territory of Katowice. Since the envisaged fraction of the parcel machines covered by the charging stations was 20%, it was further assumed that 42 charging stations should be deployed (which pertained to both the first and the second strategy).

Data on the public parking spaces were downloaded from the OpenStreetMap portal using the amenity key of parking_space, and then the data were manually verified. There are currently 3976 polygons describing parking spaces in the city of Katowice.

3.2. First Strategy for Establishing Potential Electric Vehicle Charging Sites for Courier Companies Using the Existing Publicly Available Charging Infrastructure

For purposes of the case study pertaining to the first of the strategies proposed, the input data previously acquired were used. A spatial analysis of access within a walking distance of 50 m was performed for each point where a parcel machine was situated.

Thus established, the access polygons were linked with the locations of the public charging stations. It was found that six charging stations met the condition of being within a distance smaller than 50 m from a parcel pick-up and drop-off point location. Figure 4 shows the locations in question.

In order to better illustrate the example of the siting of the charging stations that met the condition of a walking distance of less than 50 m from the location of a machine where a parcel could be dropped off or picked up, a magnified view of one of the six conforming stations is provided in Figure 5.

Due to the insufficient number of charging stations; i.e., 6 out of the 42 needed to be deployed, step two was performed under the first strategy to establish sites for the remaining 36 charging stations. In order to analyse the accessibility of the parcel machines within a walking distance of 10 m, the available parking areas in the vicinity were examined. Parking spaces that met the foregoing condition were found for 118 machine locations. Of the 118 locations, 3 were eliminated because those for which the condition related to the distance from a public charging station had previously been met, thus conforming with both conditions. Figure 6 shows the 115 locations of the parcel pick-up and drop-off machines that met the condition involving the set distance from a parking space and the 6 public stations that met the condition concerning the distance to a parcel machine.

Of the 115 locations, 36 sites were selected for charging stations with their spatial distribution in mind. The six public stations were used as the basis for the distribution analysis. For all the potential points, a distance matrix was developed for the three nearest neighbouring sites. The potential locations linked in terms of distance with the six stations were removed in the first place (18 points). For the remaining elements, the average distance between the points was calculated as 571 m in a straight line. Twenty-one points with values above that threshold were left (due to the largest distance between them). For the remaining elements, a comparative analysis of the distance between points was performed (points for which the distance was the smallest were removed) to ensure an optimal distribution of the charging stations across the city. Figure 7 shows the 6 public stations that met the distance conditions and the 36 prospective sites where new in-house charging stations could be installed by the courier company.

The average value of the distance between the vehicle charging stations was 903 m, while the minimum value of the distance between the two closest stations was 470 m. The first quartile of distance was 701 m, while the third distance quartile was 1093 m.

3.3. Second Strategy for Establishing Potential Electric Vehicle Charging Sites for Courier Companies by Disregarding the Existing Public Charging Infrastructure

The second strategy disregarded the available public charging stations. In order to analyse the accessibility of the parcel machines within a walking distance of 10 m, the available parking areas in the vicinity were examined. Parking spaces meeting the foregoing condition were found for 118 machine locations. Of the 118 locations, 42 locations were chosen as those considered to be optimally distributed within the case study area while assuming their even distribution in space. For the selected points, a distance matrix was established for the three nearest neighbouring locations to eliminate clusters of the points meeting the condition. The average distance between the points was calculated as 561 m in a straight line. Twenty-four points with values above this level were left. For the remaining points, the distance to the nearest neighbouring locations was checked, and the closest point was eliminated. Following the first iteration of calculations, 61 points remained. For the remaining points, the procedure that involved calculating the distance to the three nearest neighbouring locations was repeated. The average distance came to 863 m. Points with values above this threshold (20 in total) were left. For the remaining points, the distance to the nearest neighbouring locations was checked, and the closest point was eliminated. Following the first iteration of calculations, 42 points remained. Figure 8 shows the selected locations.

The average value of the distance between the vehicle charging stations is 1025 m, while the minimum value of the distance between the two closest stations is 743 m. The first distance quartile is 867 m, while the third distance quartile is 1085 m.

4. Discussion

The methods proposed in this paper are characterised by a high flexibility. It should be noted that the foregoing pertains to both the minimum distance conditions (50 m from the existing urban charging stations and 10 m for the company’s newly designed in-house charging stations were envisaged) and the degree of expansion. Towns are at different stages of electromobility implementation. In the above case study, only 20% of the parcel pick-up and drop-off machines were covered by the operating range of the available charging stations. As per the principle of small steps, a one-time installation of a very large number of charging stations is not considered possible, while their deployment should proceed in stages as the relevant needs continue to grow. In this case, a larger number of charging stations will encourage investment in the electric vehicle fleet, and fleet expansion will in turn increase the need for charging stations.

Given the lack of integrity in terms of the spatial distribution of parcel machines and charging stations, an unquestionable downside of the first strategy is the need for the delivery vehicle to be left at a greater distance. This forces the courier to use an additional trolley to deliver parcels to and to collect them from the machine. On the other hand, this strategy reduces the costs that the courier company would otherwise have to incur to install in-house charging stations.

A compromise solution could entail more extensive cooperation between municipal authorities and parcel shipping companies as well as planning of the charging station siting that would take the needs of delivery vehicles into account. An increasing number of municipalities are implementing strategies aimed at promoting green solutions that—although investment-intensive—can translate into an improved quality of life for residents. A specific incentive that can be offered may involve special parking spaces located in the most densely populated places that are dedicated to green delivery vehicles, which could be adapted to the specific needs of electric vehicles (by installing charging stations). Furthermore, one could also assume a compromise pertaining to the availability of such spaces; for instance, during the evening and night hours, thereby enabling local inhabitants to charge their private cars.

This article did not consider any legal aspects of the acquisition of the land currently occupied by the available public parking spaces [45]. As mentioned previously, it would undoubtedly be better if municipal authorities were open and willing to cooperate in this regard because not only would it reduce the costs to be incurred by the courier company, but it would definitely also translate into an improved quality of life for local communities and better management of urban space. Similar arrangements are often made by municipal authorities and car-sharing or bike-sharing companies; this is precisely why the authors assumed that such cooperation is also possible in the case of parcel shipment service providers.

It should be noted that the method discussed in this paper did not take transfer routing into account. The solution proposed by the authors was based on an assumption that the charging stations available in the urban space were evenly distributed. It is planned that future research should take into account both routing operations and the siting of charging stations depending on how frequently orders concerning individual parcel pick-up and drop-off machines are placed. At the same time, it should be noted that uniform distribution of the locations in question yielded more resistance to any changes in terms of the parcel machine selection behaviour and subsequent deployments of charging stations based on the method proposed and ensuring larger coverage of parcel machines (gradually up to 40%, 50%, 70%…) will make the aspect analysed in this paper increasingly independent from the said factors.

5. Conclusions

The method proposed by the authors can prove particularly important in the urban areas where traffic has been confined to that of pedestrians, cyclists, collective transport, and low- or zero-emission vehicles. This is precisely where parcel deliveries could be handled by means of electric cars. In the future, one should expect an increase in the number of zones dedicated solely to environmentally friendly modes of transport.

Since charging stations are only used for a limited period of time, future efforts to be undertaken should definitely take into account the possibilities of collaboration with courier companies and the sharing of charging stations. By that means, a shared platform for the management of charging stations could be established. This makes all the more sense because in many cases the destination points are common to a number of parties or are situated in close proximity to one another. Other possibilities for future research would be using an econometric statistical-based approach to make predictions for charging sites.