Leveraging network connectivity for planning and recommendation for users in the context of service applications

Abstract

A system and a computer-implemented method for generating or amending locations for service stations, the method comprising a step of acquiring, by an input interface, input information including transportation network information from a plurality of data sources. A processor generates a graph representation of a transportation network based on the acquired transportation network information and determines a set of constraints based on the acquired input information. The processor determines a scoring measure for candidate locations of service stations, and a scored list of candidate locations of service stations based on the determined scoring measure for the candidate locations. Further, the processor determines a communicability distance measure of the transportation network by performing a spectral analysis of the graph representation, determines a predetermined number of optimized locations for service stations by optimizing the determined communicability distance measure, and performs a multi-objective optimization of a multi-objective function for determining locations of the service stations based on the determined set of constraints, the scored list of candidate locations of service stations, and the predetermined number of optimized locations for service stations. The processor then generates an output signal including the determined or amended locations for service stations and outputs, via an output interface, the generated output signal.

Description

[0001] Leveraging network connectivity for planning and recommendation for users in the context of service applications

[0002] The invention is in the field of networks of service stations. In particular, the invention concerns planning of locations of service stations in networks, optimizing locations of service stations and managing operations of a network of service stations.

[0003] Providing technical infrastructure arranged over large inhabited areas needs complex planning processes and near constant adaptation to changing requirements from its users. For example, planning the infrastructure for the increasing ownership and use of electric vehicles (EV) by private citizens, private businesses and public transportation providers is a task that requires significant effort in order to cope with the equally increasing demand for charging stations. Owners and users of EVs may charge the batteries at their home using privately owned charging stations. Similarly, EVs of a business fleet or a municipal service may have a company-owned service station or municipal service station with charging stations. Many users of EVs rely on public charging stations for regular charging, or supplementary charging when travelling over long distances.

[0004] Planning locations of electric vehicle charging stations, or battery replacement stations for EVs in a network of service stations is a complex process that has to take into account many factors.

[0005] On the one hand, spatially distributing the charging stations has to ensure that the requested services are provided in a convenient way for the user of the EVs in order to be accepted. For example, the user of the EV will benefit from a short traveling distance to reach the service station, in order to minimize the energy consumption for reaching the charging station so that he can use the EV up to a low charging level of the battery of the EV without having to bother about recharging.

[0006] On the other hand, distributing additional service stations in a region in order to provide a high density of charging stations, which improves the comfort of the users of EVs, causes considerable costs for the operator of the network of charging stations (service provider). For example, an even distribution of charging stations in a region may neglect factors such as quantity of EVs in the region, traffic density of EVs in the region, availability of suitable sites for charging stations that require space and connectivity with the power grid with sufficient electric power available. Without sufficiently balancing these partially conflicting requirements, an operator of a network of charging stations may experience low acceptance by users of the EVs, low revenues from charging EVs, and hence low return on the investment into the network of charging stations. This may also result in low acceptance of EVs by private users, businesses and public services.

[0007] A service station may be a charging station or battery replacement station, or a logistic service station of a logistic service provider,

[0008] The factors influencing the planning processes for service stations may vary overtime.

[0009] Therefore, not only planning a network of service stations that includes, e.g., generating a map of locations of service stations for a region requires attention, the process of amending the map of service stations that may include adding a location of a new service station, removing a location of a service station envisaged for closure, or relocating a service station, is also of importance for the operator of the network of service stations.

[0010] Even determining the recommendation to a user for a particular service station of a network of service stations for a required service is of significance to the operator of a network of service stations. By the determined and output recommendations to users, an even use of the service stations is achievable that reduces waiting times for services for the user and may also enable to take respectively lower capacities at individual service stations into account.

[0011] There exist approaches to use automated optimization processes in order to achieve an optimized compromise between the needs of the user and the provider. For example, arranging the service stations in a region may apply methods from the fields of operations research and combinatorial optimization problems.

[0012] US 10,467,556 B2 proposes systems and methods to identify placement for an EV charging station infrastructure in the form of a deployment plan for one or more EV charging stations. An exemplary method includes gathering data for a selected specific area and forecasting a demand for EVs for the area. The example method models driving patterns in the area using available data and improving a charging infrastructure model based on the driving patterns and forecast demand for the area. The example method generates a recommendation regarding an EV charging infrastructure and a deployment strategy for the area based on the improved charging infrastructure model. The gathered data, and the forecasted demand for EVs base on geographic data, demographics, power distribution grid, car sales for geographic areas, which correspond to a selected zip code, a set of zip codes, a state, a county or city. The method uses the gathered data to predict demand as a basis for determining the spatial distribution of the charging infrastructure. The resulting charging infrastructure depends on the quality of the predicted demand, and the modeling of the estimated driving patterns of the EVs, which bases on data evaluated for zip codes. The actual road traffic network, and live traffic data of the dynamic processes is largely neglected. Thus, the object of the invention is to improve processes for determining locations of service stations in a network of service stations.

[0013] The computer-implemented methods in a first aspect, a second aspect, and a third aspect, and the system in a fourth aspect, according to the independent claims, address the object and provide advantageous solutions.

[0014] The dependent claims define further advantageous embodiments.

[0015] According to a first aspect, the computer-implemented method for generating or amending locations for service stations comprises a step of acquiring, by an input interface, input information including transportation network information from a plurality of data sources. A processor generates a graph representation of a transportation network based on the acquired transportation network information and determines a set of constraints based on the acquired input information. The processor determines a scoring measure for candidate locations for service stations, and a scored list of candidate locations of service stations based on the determined scoring measure for candidate locations. Further, the processor computes a communicability distance measure of the transportation network by performing a spectral analysis of the graph representation, determines a predetermined number of optimized locations for service stations by optimizing the determined communicability distance measure, and performs a multi-objective optimization of a multi-objective function for determining locations of the service stations based on the determined set of constraints, the scored list of candidate locations of service stations, and the predetermined number of optimized locations for service stations. The processor then generates an output signal including the determined or amended locations for service stations and outputs, via an output interface, the generated output signal.

[0016] The method according to the first aspect provides an output signal including determined locations for service stations that arranges the locations of the service stations as optimal as possible for defined targets and constraints, and simultaneously offers valuable services for the user on the one hand and on the other hand for the operator.

[0017] The method situates the service stations by integrating a robust network design with detailed information from a plurality of data sources that may, e.g. include satellite imagery, enhancing station location for users requesting services from a service station and the operator of the service stations. The approach benefits from the capability to use hybrid data. Hybrid data may include proprietary telematics and open datasets like space availability, supermarket locations, power grid details, and crime statistics. The approach uses the hybrid data to accurately model traffic flow and traffic infrastructure. Due to using spectral analysis of the graph representation, the method may identify an optimal network layout for maximum reachability regarding different objectives. The method utilizes a topology and dynamics of the transportation network for optimal placement of the service stations. The term transportation network as used in the present document does not include a supply infrastructure, e.g., service stations. Based on performing the multi-objective optimization, the spatial feasibility of the locations of the service stations can be ensured and even a social factor as well as user preferences and operator preferences may be integrated via the information acquired from the plurality of data sources, which is integrated into processing of the computer-implemented method.

[0018] For planning of locations of service stations, conventional analytical approaches often focus on abstract road connectivity and neglect essential details like parking space availability, vehicle maneuverability during charging, and social factors such as weather protection and pedestrian safety. Interests of the operator of the service stations, e.g., a proximity of e.g. charging / service stations to existing dealerships, are also vital. The method integrates these into an analytic graph assessment process, and facilitates layout planning that considers various objectives in the determined locations of service stations.

[0019] The method uses an integrated multi-objective optimization that synthesizes weighted inputs from information representation, a spatial analysis, and the computed communicability distance measure to balance aspects such as spatial feasibility, e.g. maneuverability and avoiding station deadlocks, network reachability, e.g. considering traffic flow in general to reach service stations, stakeholder interests, e.g., by bringing the user closer to dealer locations, to offer a more profitable customer service in case of malfunction of an EV or a service station, and social aspects, e.g. pedestrian safety, close proximity to common places, e.g., supermarkets, doctors, thereby supporting the decision of finding the optimal location of a service station.

[0020] In an embodiment of the computer-implemented method, the set of constraints includes constraints comprising at least one of a power grid connectivity, a parking lot availability, a safety score, a vehicle maneuverability during charging, weather protection, pedestrian safety, and a proximity to dealerships (including, e.g., repair shops).

[0021] The computer-implemented method improves planning and locating service stations in a region and the operational efficiency with respect to the constraints and the targets of the operator. As a fulfillment of user needs that the method regards when determining the locations of the service stations, even user satisfaction maybe improved during operation of the network of service stations. Furthermore, due to including constraints, e.g., considering the reachability of facilities of dealerships, costs and required time for maintenance or malfunction recovery of service stations decreases since they are in close neighborhood with the facilities of dealerships.

[0022] The computer-implemented method according to an embodiment acquires the transportation network information including open street map (OSM) data.

[0023] Hence, the basic information for generating the graph representation of the transportation network, and integrating the spatial aspects is readily available.

[0024] In an embodiment of the computer-implemented method, the scoring measure of each candidate location of the service station comprises a weighted sum of the individual constraints of the set of constraints.

[0025] The consideration of a plurality of pieces of information from a plurality of data sources and their relative weight with each other during consideration of the information is computationally efficient.

[0026] In the computer-implemented method according to an embodiment, the processor computes the communicability distance measure of the transportation network based on the adjacency matrix of the graph representation or the corresponding Taylor series.

[0027] Thus, well-established processes from spectral graph analysis are used to optimize the locations of service stations, achieving advantageous user accessibility and an efficient network topology of the locations of service stations.

[0028] The computer-implemented method according to an embodiment comprises the processor determining the predetermined number of optimized locations for service stations by selecting the predetermined number of optimized locations for service stations that maximize the communicability distance measure of the graph representation.

[0029] Thus, optimal locations for service stations are selected for further processing, thereby ensuring an efficient use of processing resources.

[0030] In an embodiment of the computer-implemented method, the multi-objective function comprises a weighted sum of a spatial feasibility measure, a network reachability measure, a stakeholder interest measure with regard to the closest service station, and a safety score.

[0031] Thus, the method is capable to identify the locations of service stations that are most advantageous with regard to an optimizing spatial feasibility, the network connectivity, alignment with stakeholder interests, and social impact due to regarding social aspects. The computer-implemented method according to an embodiment comprises the service stations including at least one of a battery exchange station for exchanging a battery or a charging station for charging batteries.

[0032] Thus, the method provides a tool for addressing the current issues in a transformation of the transportation sector to electric vehicles.

[0033] In the computer-implemented method according to an embodiment, the service stations form part of a service network that includes at least one of a mobility service network, a logistic service network, a supply network, a disposal network, or a network of base stations for autonomous devices.

[0034] The method offers an approach that solves the planning problems in different industrial networks due to its capability to integrate versatile information from a plurality of data sources with the spatial information included in the graph representation, and the efficient data processing of the integrated information using multi-objective optimization and spectral graph analysis.

[0035] The computer-implemented method for determining at least one location of a service station for removal, addition or relocating in a network of service stations or a transportation network according to the second aspect comprises a step of acquiring by an input interface, input information including transportation network information from a plurality of data sources. The method further includes acquiring, by a processor, a graph representation of the transportation network; determining, by the processor, future changes and demands to the transportation network and the network of service stations based on the acquired input information; determining, by the processor, an adjusted cumulative objective function based on the graph representation of the transportation network, the future changes and demands to the transportation network and the network of service stations; performing, by the processor, a multi-objective optimization of the adjusted cumulative objective function for determining at least one location of a service station. The processor then generates information on the determined at least one location of a service stations and outputs, via an output interface, the generated information.

[0036] Thus, the data representation of the discussed embodiment provides an advantageous solution to the problem of dynamically adding, removing, or relocating service stations in response to changing conditions to improve service quality and network efficiency in a network of service stations.

[0037] A computer-implemented method for recommending a service station in a network of service stations according to a third aspect comprises: acquiring, by an input interface, input information including transportation network information from a plurality of data sources; acquiring by an input / output interface, a request for a service by the network of service stations from a user; generating, by a processor, a graph representation of a transportation network based on the acquired transportation network information; determining, by the processor, a set of constraints based on the acquired input information; determining, by the processor, a scoring measure including a charging speed score for each candidate location for a service station; determining a scored list of candidate locations of service stations based on the determined scoring measure for each candidate location; performing, by the processor, a multi-objective optimization of a multi-objective function for determining a location of a recommended service station based on the determined set of constraints, and the scored list of candidate locations of service stations, wherein the multi-objective function includes a weighted combination of a current location of the user requesting the service, a predicted waiting time until starting to provide the service, an utilization rate, and a waiting queue length; and generating, by the processor, output information based on the determined location of the recommended service station and outputting, via the input / output interface, the generated output information to the user.

[0038] The method according to the third aspect ensures efficient management of service requests, an optimal use of service station resources, and provides maximized user satisfaction as well as to distribute equally the traffic in the network of service stations regarding the provided service.

[0039] The system for generating or amending locations for service stations according to the fourth aspect comprises an input interface configured to acquire input information including transportation network information from a plurality of data sources, and a processor configured to generate a graph representation of a transportation network based on the acquired transportation network information. The processor is further configured to determine a set of constraints based on the acquired input information, to determine a scoring measure for candidate locations for service stations, to determine a scored list of candidate locations of service stations based on the determined scoring measure for each candidate location, to determine a communicability distance measure of the transportation network by performing a spectral analysis of the graph representation, to determine a predetermined number of optimized locations for service stations by optimizing the determined communicability distance measure, to perform a multi-objective optimization of a multi-objective function for determining locations of the service stations based on the determined set of constraints, the scored list of candidate locations of service stations, and the predetermined number of optimized locations for service stations, and to generate an output signal including the determined or amended locations for service stations. The system further comprises an output interface configured to output the generated output signal.

[0040] The system for generating or amending a map including locations for service stations according to the fourth aspect achieves corresponding advantageous effects as discussed with reference to the computer-implemented method of the first aspect.

[0041] The following description of embodiments refers to the figures, in which

[0042] Fig. 1 displays a simplified flowchart of the computer-implemented method according to an embodiment;

[0043] Fig. 2 illustrates a graph representation of a transportation network as used in the embodiment of the computer-implemented-method;

[0044] Fig. 3 displays a simplified flowchart of the computer-implemented method according to an embodiment;

[0045] Fig. 4 displays a simplified flowchart of the computer-implemented method according to an embodiment; and

[0046] Fig. 5 provides an overview over the architecture of a system according to an embodiment.

[0047] In the figures, corresponding elements have the same reference signs. The discussion of the figures avoids discussion of same reference signs in different figures wherever considered possible without adversely affecting comprehensibility and avoiding unnecessary repetitions for sake of conciseness.

[0048] The description of an embodiment refers to a network of a plurality of service stations situated in a road transportation network. The service stations include charging stations for batteries of electric vehicles travelling in the road transportation network. Other implementations of the method and the system 10 are discussed after a presentation of the steps of the computer -implemented method.

[0049] Fig. 1 displays a simplified flowchart of the computer-implemented method according to an embodiment.

[0050] The individual method steps and specific aspects of the processes of the computer-implemented method are discussed after an overview over the processing sequence. The computer-implemented method for generating or amending locations for service stations starts with a step Si of acquiring, by an input interface 14, input information including transportation network information from a plurality of data sources 18.

[0051] The method integrates use of proprietary information (proprietary data), e.g. dealer and partner locations, GPS traffic data, and open information. Open information (open data) may include, e.g., supermarket locations, information on power grid infrastructure, parking space availability, crime statistics, weather protection infrastructure, location of street lamps. The method uses this hybrid information (hybrid data) to add information to locations of service stations and even for supporting ongoing service scheduling.

[0052] In step S2, a processor 11 generates a graph representation of a transportation network based on the acquired transportation network information.

[0053] The computer-implemented method represents the network, e.g. the transportation network by a graph representation in the format of a dynamic graph 1

[0054] Gt= (Vt, Et);

[0055] In the dynamic graph 1, the parameter Vtrepresents nodes 2, 4 at time t and the parameter Etrepresents connecting edges 3 between the nodes 2, 4 at time t.

[0056] In the example of a road transportation network, the dynamic graph 1 includes the nodes 2, 4 representing locations on a road, road intersections or crossings. The edges 3 of the dynamic graph 1 represent roads or road segments connecting the nodes 2, 4, as discussed in fig. 2.

[0057] The nodes 2, 4 include candidate locations for service stations (potential service station locations).

[0058] Each node vi,te Vtof the dynamic graph 1 is characterized by attributes:

[0059]

[0060] The attributes of each node 2, 4 include the spatial coordinates xi,t, yi,tof a location i at time t, a set of constraints Ci tsatisfied by a location i at time t, the power grid connectivity score Pi tfor location i at time t, a parking availability score ai tfor location i at time t, and a safety score γi,tfor location i at time t.

[0061] A sparse identifier tensor E =

[0062]

[0063] is used for distinguishing between different types of nodes over time:

[0064]

[0065] The different types of nodes 2, 4 may include a service station and a dealership, in order to name just two examples. In (1), ξi,1,t= 1 represents a node of type service station and ξi,2,t= 1 represents a node of type dealership. Thus, the following sets at time step t are defined:

[0066]

[0067] In expression (4), Λt⊆ Vtdenotes the set of service stations at time t. The term Δtdenotes a set of dealerships.

[0068] A distance matrix Dtis defined as:

[0069]

[0070] In expression (5), d(a, b) represents a shortest path distance between node a and node b. Alternatively, another distance measure than the shortest path distance may be used in expression (5) for the term d(a, b).

[0071] In step S3, the processor 11 determines a set of constraints based on the acquired input information;

[0072] The processor 11 then proceeds to step S4, in which the processor 11 determines a scoring measure for each candidate location for a service station.

[0073] Based on the determined scoring measure for each candidate location, the processor 11 determines a scored list of candidate locations of service stations in step S5.

[0074] The steps S3, S4, and S5 represent a process for performing a spatial analysis and infrastructure analysis.

[0075] In the spatial analysis and infrastructure analysis, the method analyzes the collected hybrid data from the plurality data sources for identifying optimal locations of service stations based on spatial constraints and infrastructural constraints.

[0076] The process starts by defining constraints in the form of a set of constraints Ci t. The constraints include, e.g., a power grid connectivity and a parking space availability. Subsequently, candidate location for service stations are scored based on a quality, an accessibility, and a strategic value, including a proximity of the candidate location to dealerships.

[0077] The process of the spatial analysis and infrastructure analysis may perform different methods in sub-processes.

[0078] A sub-process of spatial filtering eliminates candidate locations for service stations that fail to meet basic infrastructural requirements for the location of a service station. Scoring algorithms assess a suitability of each candidate location by evaluating various criteria for a location of a service station. The criteria may include an access to the power grid, the availability of a parking space, and a safety parameter (safety score). The safety score for a candidate location may be determined based on crime statistics acquired as part of the hybrid data. The crime statistics include location-related information on crimes, in particular crimes concerning property, vehicles, and traffic situations.

[0079] The set of constraints Ci tthat a location i at time t satisfies, may in particular include a constraint defining that the power grid connectivity score for the candidate location of the service exceeds a predetermined first threshold.

[0080] The set of constraints Ci tthat a location i at time t satisfies, may include a constraint defining that a parking space availability score exceeds a predetermined second threshold. The parking space availability score takes into account that the location of the service station has enough space available for vehicles to park while their batteries are being charged, which the predetermined spatial availability score threshold (second threshold) ensures.

[0081] The set of constraints Ci tthat a location i at time t satisfies, may include safety constraint defining that the candidate location of the service station is in an area for which the safety score exceeds a third predetermined threshold (safety score threshold).

[0082] The set of constraints Ci tthat a location i at time t satisfies may include a proximity constraint defining a proximity to dealerships: the proximity-to-dealership constraint defining that the candidate location of the service station is close to vehicle dealership, in particular that a proximity-to-dealership score exceeds a fourth predetermined threshold (proximity score threshold).

[0083] The set of constraints Ci tmay include other constraints known from facility location problems alternatively or additionally to the discussed constraints defining the access to the power grid, the availability of a parking space, the safety of the area and the proximity to dealerships.

[0084] A scoring measure (scoring function) Score; for each candidate location of a service station vt,t is:

[0085] Score;

[0086]

[0087] In expression (6), the weights w1,t, w2, w3, and w4are weights representing the importance of each constraint of the set of constraints Ci t. The time steps t are the time steps of the optimization. The weights w1, w2, w3, w4define a set of weights with values in the range [0,1] for different constraints representing the importance of each constraint relative to the other constraints.

[0088] In expression (6), the term Pi te [0,1] is the power grid connectivity score for a candidate location i at time t. The term ai te [0,1] denotes the parking availability score for a candidate location i at time t.

[0089] The term Dv*b* is the shortest path distance from the candidate location v*i,t∈ V to the nearest dealership b* ∈ Δt. The shortest path distance is minimized in order to maximize the proximity scoring. As the distance score term Dv*,b*is unlimited, the weight factor w3cannot be compared to the other weights w1, w2, and w4. The weighting in the form of expression (6) ensures that larger distances to the next dealership are weighted lower and that for small distances a proportional weighting of the term prevails.

[0090] In expression (6), the term

[0091]

[0092] te [0,1] defines the safety score for candidate location i at time t.

[0093] Based on the determined scoring measure for each candidate location, the processor n determines the scored list of candidate locations of service stations in step S5 and then proceeds to step S6.

[0094] The spatial analysis and infrastructure analysis of steps S3, S4, and S5 analyzes the collected hybrid data for identifying optimal locations of service stations based on spatial constraints and infrastructural constraints. The result of the spatial analysis and infrastructure analysis is a filtered and scored list of candidate locations of service stations, which is optimized for spatial feasibility and strategic value. The computer-implemented method further comprises a sub-process of performing a spectral graph analysis to evaluate the reachability and connectivity of the transportation network based on the graph representation and optimizing the locations of the service for achieving a maximum user accessibility and network efficiency. Steps S6 and S7 of the flowchart depicted in fig. 1 cover the sub-process of performing a spectral graph analysis.

[0095] The sub-process of performing the spectral graph analysis uses the dynamic graph Gt= (Vt, Et) from the previous sub-process.

[0096] The spectral graph analysis performed in the sub-process includes computing the communicability distance measure (Gt) using the adjacency matrix Atof the dynamic graph graph Gt= (Vt, Et).

[0097] Alternatively and equivalently, the spectral graph analysis performed in the sub-process uses the corresponding Taylor series:

[0098]

[0099] For defining the communicability distance δpq(Gt) between nodes p and q in the dynamic graph, the elements of the matrix exponential eAtare used:

[0100]

[0101] The communicability distance δpq(Gt) represents an effective distance between nodes p and q, taking into account all possible paths along the edges of the dynamic graph connecting the nodes p and q and thereby taking into account the overall network structure of the transportation network represented by the dynamic graph.

[0102] The sub-process of performing the spectral graph analysis then identifies optimal service station locations by selecting a predetermined number k of locations of service stations that maximize the network's communicability distance measure:

[0103]

[0104] An optimization technique is used to select k candidate locations of service stations as a subset of all nodes 2, 4, maximizing the communicability distance measure (Gt). The predetermined number of candidate locations for service stations may predetermined by the operator of the network of service stations. The predetermined number of candidate locations for service stations correspond to the number of service stations that are planned for the given region of the transportation network. The predetermined number of service stations may be varied for different runs of the method.

[0105] Alternative facility placement approaches than the proposed scheme using the communicability distance measure (Gt). are also possible.

[0106] The communicability distance measure (Gt) enables to determine nodes 2, 4 of the dynamic graph to locate service stations at the corresponding locations of service stations by ensuring that the chosen subset Atof nodes 2, 4 of the dynamic graph maintains a high network connectivity, which is has the effect of optimizing reachability and efficiency in the transportation network.

[0107] The sub-process of performing the spectral graph analysis includes steps S6 and S7 of the flowchart of fig. 1.

[0108] In step S6, the processor 11 computes the communicability distance measure (Gt) of the transportation network by performing a spectral analysis of the graph representation Gt= (Vt, Et) for generating optimized locations of service stations.

[0109] In step S7, the processor 11 determines a predetermined number k of optimized locations for service stations based on the computed communicability distance measure.

[0110] After step S7, the processor 11 performs in step S8 a multi-objective optimization of a multiobjective function for determining the locations of the service stations based on the determined set of constraints, the scored list of candidate locations of service stations, and the predetermined number of optimized locations for service stations.

[0111] In particular, the computer-implemented method includes the sub-process of performing an integrated multi-objective optimization.

[0112] For performing the integrated multi-objective optimization, the computer-implemented method integrates weighted inputs from the sub-processes of data representation utilization in steps Si, S2, and S3 of the method, of spatial analysis and infrastructure analysis, and of spectral graph analysis using the communicability distance measure to balance multiple objectives with the ultimate aim to determine optimized locations of service stations in the region from the candidate locations. The integrated multi-objective optimization includes objectives, which comprise spatial feasibility, network reachability, stakeholder interests, and social aspects.

[0113] For performing the process of integrated multi-objective optimization, the problem is formulated using a multi-objective function:

[0114] max F = β1· f1+ β2· f2+ β3· f3+ β4· f4subject to Ci,t; (10)

[0115] In expression (to), the term (objective) f1represents a spatial feasibility, the term f2represents a network reachability, the term f3represents stakeholder interests with respect to the closest service station, and f4represents social aspects. The respective weights β1, β2, β3, β4in expression (10) are each in the interval [0,1] for each of the objectives in expression (10).

[0116] In expression (10), the time steps t are the time steps of the optimization.

[0117] Each individual objective function A in expression (10) is defined as follows:

[0118] f1= Σv∈Λ(Maneuverabilityi× Deadlock Avoidancei); (11)

[0119] f2= Θ(Gt); (12)

[0120]

[0121] A Xv,eAtY i,t’ (14)

[0122] In expression (11), the maneuverability score Maneuverabilityiand the deadlock avoidance score Deadlock Avoidanceiare spatial feasibility scores reflecting the ease of vehicle navigation and the mitigation of traffic congestion based on spatial availability from a visual satellite data assessment.

[0123] In expression (12), the term Θ(Gt) denotes the communicability distance measure.

[0124] In expression (13), the term Di tdenotes the shortest path distance to the nearest dealership.

[0125] In expression (14), the term

[0126]

[0127] tdenotes the safety score.

[0128] This sub-process identifies the optimized, hence the most advantageous locations of service stations based on evaluating expression (10) from the candidate locations of service stations. Thereby the computer-implemented method optimizes the locations of service stations based on their spatial feasibility, network connectivity, stakeholder interest, and social impact.

[0129] After determining the locations of the service stations (optimized locations of service stations) the processor n generates in step S9 the output signal including the determined optimized locations of service stations. The processor 11 then outputs the generated output signal via an input / output interface 15 and a human user interface 17, e.g., to a user or an operator.

[0130] Fig. 2 illustrates a graph representation of a transportation network as used in the embodiment of the computer-implemented-method. The transportation network in the example of fig. 2 is a road network in a built-up area.

[0131] Embodiments of the invention represent a transportation network by a graph representation corresponding to a dynamic graph 1

[0132] Gt= (Vt, Et, (15)

[0133] In the dynamic graph 1, the parameter Vtrepresents nodes 2, 4 at time t and the parameter Etrepresents connectivity edges 3 at time t. In the dynamic graph 1, nodes 2, 4 represent locations on a road, road intersections or crossings. The edges 3 of the dynamic graph 1 represent roads connecting the nodes 2, 4, as depicted in fig. 2. In fig. 2, points represent nodes 2, 4 and connecting lines between points represent edges 3 that connect the nodes 2, 4.

[0134] The nodes 2, 4 include candidate locations for service stations (potential service station locations).

[0135] The dynamic graph 1 may be generated from Open Street Map data (OSM data).

[0136] Each node vi,te Vtis characterized by attributes:

[0137]

[0138] The attributes of each node 2, 4 include

[0139] the spatial coordinates xi,t, yi,tof a location i at time t,

[0140] a set of constraints Ci tsatisfied by a location i at time t,

[0141] the power grid connectivity score Pi tfor location i at time t,

[0142] a parking availability score ai tfor location i at time t, and a safety score yi tfor location i at time t.

[0143] Fig. 3 displays a simplified flowchart of the computer-implemented method according to an embodiment.

[0144] In the preceding section of the discussed embodiment of the computer-implemented method, determining the locations of service stations is discussed. The data representation of the discussed embodiment provides also an advantageous solution to the problem of dynamically adding, removing, or relocating service stations in response to changing conditions to improve service quality and network efficiency.

[0145] The process of adding, removing, or relocating service stations bases on acquired real-time information and a predictive modeling to identify evolving patterns in user behavior, traffic flow, and infrastructure changes for identifying locations of service stations for adding, removing, or relocating service stations. The following section presents such approach.

[0146] The graph representation of the transportation network, in particular the dynamic graph according to expression (1), (17)

[0147] Gt= (Vt, Et, (17)

[0148] includes the nodes Vtcomprising candidate locations for service stations (potential service station locations) at time t and the edges Etconnecting nodes Vtat a time t.

[0149] The identifier tensor E =

[0150]

[0151] according to expression (2), (18) enables to distinguish between different types of nodes 2, 4 over time:

[0152] ξi,j,t{1, if node v{i,t}is of a type j at a time t; 0, otherwise

[0153] Hence, the identifier tensor E =

[0154]

[0155] ensures that dynamic changes are accurately reflected in the dynamic graph. The time steps t are the time steps of network evolvability, including an adding, removing, or relocating of service stations. The time step t has therefore a different time scale than the real-time. The time range between t and t + 1 may represent one or more months, or even one or more years depending on at least one of the application scenario, and targets and constraints of the operator of the network of service stations.

[0156] The network evolvability problem represents a dynamic optimization problem with three particular cases: A first case concerns a rearranging (relocating) of existing service stations: no new service stations are added to the network and existing service stations are relocated to new locations in order to meet changed system requirements. Rearranging existing service stations includes determining amended locations of service stations.

[0157] A second case concerns increasing the number of service stations. Additional service stations are added to the network at new locations of service stations. The new locations of service stations currently include no service stations.

[0158] A third case concerns decreasing the number of service stations: existing (operable) service stations are closed and removed from the network of service stations.

[0159] The computer-implemented method addresses the first, second and third cases and the underlying scenarios of the first, second, and third case are represented in mathematical notation as follows:

[0160] In the first case of relocating an existing service station, expression (19) applies:

[0161] max Ytotalsubject to |Λt+1| = |Λt|; (19)

[0162] In the second case, of increasing the number of the currently existing service stations, expression (20) applies:

[0163] max Ytotalsubject to |Λt+1| > |Λt|; (20)

[0164] In the third case, of decreasing the number of the currently existing service stations, expression (21) applies:

[0165] maxTtotalsubject to |At+1| < |At|; (21)

[0166] In expressions (19), (20), and (21), the term Λtrepresents the set of service stations at time t, Λt+1represents the set of service stations at the time t + 1. The term Gt+1= (Vt+1, Et+1) denotes the updated dynamic graph for the next time t + 1. The term Ytotalis the cumulative objective function over the time horizon T.

[0167] For estimating an expected evolvability of the network, a prediction function 풫(Gt, Ξt, 풟t) is introduced, which predicts future conditions based on the current dynamic graph Gt, the current identifier tensor Et, and a current set of dynamic factors

[0168]

[0169] The dynamic factors may include, e.g., a user behavior, traffic flow, and infrastructure changes. Expression (22) defines the prediction function

[0170]

[0171] The computer-implemented method then adjusts the cumulative objective function Ytotalaccording to expression (23) in order to consider the predicted evolvability over multiple time steps:

[0172]

[0173] In expression (23), the term T denotes the time horizon over which the optimization is performed.

[0174] The computer-implemented method performs the optimization process by executing a sequence of steps, which includes acquiring current information from the plurality of data sources, performing a predictive modeling of future changes and demands, and reconfiguring of the network of service stations based on the optimization process of expression (23).

[0175] Incorporating the identifier tensor E in expression (23) ensures that the type and the status of each node 2, 4 is accurately reflected in the optimization process.

[0176] The simplified flowchart of fig. 3 of the method for determining at least one location of a service station for removal, addition or relocating in a network of service stations in a transportation network starts with a step S11 of acquiring by an input interface 14, input information including transportation network information from a plurality of data sources 18.

[0177] In the subsequent step S12, the method acquires, by a processor 11, a graph representation of the transportation network. The graph representation may be stored in a memory 13 or in a server 17.

[0178] The processor 11 then determines future changes and demands to the transportation network and the network of service stations based on the acquired input information from the plurality of data sources 18.

[0179] Alternatively 0 additionally, the processor 11 determines future changes and demands to the transportation network and the network of service stations based on user input acquired via an input / output interface 15 and a human user interface 17 from an operator.

[0180] In step S14, the processor 11 determines an adjusted cumulative objective function based on the graph representation of the transportation network, the future changes and demands to the transportation network and the network of service stations and proceeds to step S15. In step S15, the processor 11 performs a multi-objective optimization of adjusted cumulative objective function for determining at least one location of a service station. The determined at least one location of a service station represents an optimized recommendation for at least one location at which to add a service station, remove a service station, or from what location to which location to relocate a service station.

[0181] In step S16, the processor 11 then generates information on the determined at least one location of a service station and outputs, via an output interface 15, the generated information.

[0182] The data representation of the discussed embodiment provides therefore an advantageous solution to the problem of dynamically adding, removing, or relocating service stations in response to changing conditions to improve service quality and network efficiency in a network of service stations.

[0183] The discussed approach ensures a flexible and responsive network of service stations, continuously optimizing user satisfaction and operational efficiency of the network of service stations by anticipating and adapting to predicted future changes and demands.

[0184] The data representation of the discussed embodiment provides also an advantageous solution to the problem of dynamically scheduling services provided by the service stations in response to changing operating conditions thereby improving service quality and network efficiency.

[0185] A process of dynamically scheduling service stations enables to generating a recommendation for a particular service station to a user of the network of charging stations that is beneficial for the user and the operator of the network of service stations.

[0186] When the network of service stations at their respective locations in the region is in place and users use the network of service stations, a dynamic scheduling system for recommending a particular service station of the network of service stations is advantageous. On the one hand, the operator of the network of service stations improves efficiency and the operation grade when recommending a suited specific service station to the user. On the other side, the user can input their preferences and obtain a matching recommendation for a particular service station in real-time that suits the demand of the user.

[0187] The scheduling process according to an embodiment bases on an optimization that combines types of objectives in combination with predetermined constraints and the communicability distance measure Θ(Gt). The computer-implemented method discusses an approach for the dynamic scheduling process, which incorporates the communicability distance measure Θ(Gt) into a multiobjective optimization scheme with previously defined objectives.

[0188] In this context, it extends the graph representation including the locations of the service stations generated by the computer-implemented method in the dynamic scheduling process together with preferences of the operator of the network of service stations. The embodiment generates and outputs an adequate suggestion for a service station based on the user demand in a request and further taking predetermined preferences of the operator into account.

[0189] Instead of a power availability score of the computer-implemented method for generating or amending locations for service stations, the computer-implemented method for recommending a service station uses a charging speed score.

[0190] By replacing the power availability score, with the charging speed score, an impactful and novel holistic approach for an operator of the network of service stations, e.g., a mobility provider, enabling to schedule the provided services over the network of service stations is available.

[0191] The optimization method of the computer-implemented method differs from the computer-implemented method insofar as the scheduling process needs be solving real-time or at least in close to real-time. This involves implementing a real-time, computer-implemented method for the scheduling process that adapts to fluctuating demands of users for services, e.g., charging the batteries of their EVs and varying operational conditions in the transportation network.

[0192] The process of dynamic scheduling includes continuously monitoring usage of the network of service stations, acquiring the user demand, and monitoring the traffic conditions in the transportation network.

[0193] The process of dynamic scheduling allocates resources and schedules charging sessions to optimize utilization of the service stations.

[0194] The process of dynamic scheduling may be implemented in a system that includes elements configured to integrate real-time information, to perform predictive analytics, to allocate resources, and to perform queue management.

[0195] The process of dynamic scheduling includes performing an optimization process based on expression (24)

[0196]

[0197] In expression (24), b is a current location of the user requesting a service, e.g., charging the battery of their EV. The term Tirepresents the total wait time until providing the service starts, the term Ui represents an utilization rate, which describes the capability of the service station per time to fulfill the request, e.g. corresponding to the transmittable energy per time. The term represents the queue length, which corresponds to a number of users in the queue. The terms δ1, δ2, δ3, δ4are weights in a numerical range of [0,1] for each criterion.

[0198] The process of dynamic scheduling based on the proposed algorithm includes steps of acquiring information, predicting demand, optimizing resources, notifying the user. The proposed process may also include a continuous feedback loop to the user. The continuous feedback loop may enable the method continuously to provide information on determined changes in the service schedule or on changes due to delayed arrival at the service station of other users to the user. The method ensures efficient management of charging requests, optimal station resource utilization, and maximized user satisfaction as well as to distribute equally the traffic in the network regarding the service.

[0199] The human user interface 17 enabling the user to input a request for a service and to obtain a recommendation in response to the input request may be a software based interface running on hardware such as a personal mobile device of the user or integrated into a wearable device.

[0200] Alternatively or additionally, the human user interface 17 for inputting a request for a service and for obtaining a recommendation in response to the request may be a software-based interface running on hardware integrated in a vehicle controlled by the user.

[0201] Alternatively or additionally, the hardware for implementing the human user interface 17 may include another hardware or web-service that connects to the scheduling system.

[0202] Fig. 4 displays a simplified flowchart illustrating the computer-implemented method for recommending a service station in a network of service stations. The method comprises a step S21 of acquiring, by an input interface 14, input information including transportation network information from a plurality of data sources 18.

[0203] In step S22, the method acquires, by an input / output interface 17, a request for a service by the network of service stations from a user.

[0204] The method proceeds to step S23, in which a processor 11 generates a graph representation of a transportation network based on the acquired transportation network information. In step S24, the processor 11 determines a set of constraints based on the acquired input information.

[0205] The processor 11 then determines in step S25, a scoring measure including a charging speed score for each candidate location for a service station.

[0206] In step S26, the processor 11 then proceeds with determining a scored list of candidate locations of service stations based on the determined scoring measure for each candidate location.

[0207] The processor 11 then performs in step S27 a multi-objective optimization of a multiobjective function for determining a location of a recommended service station based on the determined set of constraints, and the scored list of candidate locations of service stations. The multi-objective function includes a weighted combination of a current location of the user requesting the service, a predicted waiting time until starting to provide the service, a utilization rate, and a waiting queue length.

[0208] In step S28, the processor 11 generates output information based on the determined location of the recommended service station and outputs the generated output information to the user via the input / output interface 17.

[0209] The method ensures efficient management of charging requests, an optimal use of service station resources, and provides maximized user satisfaction as well as distributes equally the traffic in the network of service stations regarding the provided service.

[0210] Fig. 5 provides an overview over the architecture of a system according to an embodiment.

[0211] The overview of fig. 5 is an overview on a high level of abstraction of an architecture of computer hardware elements suitable for running an embodiment of the computer-implemented method, and illustrates in particular interfaces to further hardware elements useful for understanding elements of the disclosure.

[0212] The system 10 of fig. 5 includes a processor 11, a data storage 13 (memory 13), an input / output interface 15, and a network interface 54, which are all connected by a data bus 52.

[0213] The input / output interface 15 may in particular provide a capability to output information via visual or audible signals obtain to a human user. The input / output interface 15 provides the system 10 with a capability to communicate with a human user interface 17. The human user may include the user of a vehicle or at least one operator of the network of service stations. The system io may output via the input / output interface 15 to the human user interface 17. the output signal including information on the determined locations of service stations, e.g., including a map of determined locations of service stations in a visualization.

[0214] The input / output interface 15 also provides the system 10 with a capability to obtain information and commands from the human user via the human user interface 17.

[0215] The input / output interface 15 may provide the capability to provide information to other devices, e.g. one or more navigation systems of transport systems transporting materials and physical objects in a transportation network. The input / output interface 15 may at least in part be implemented in software modules running on the processor 11.

[0216] The input / output interface 15 represents an interface for connecting input / output devices including, but not limited to keyboards, mouse, pointing devices, displays, microphones, loudspeakers or a combination thereof. The input / output devices may include mobile personal devices, computer devices, wearable devices, or a human-machine interface 17 implemented in software running on a processing unit of a vehicle.

[0217] The processor 11 may be any type of controller or processor 11, and may even be embodied as one or more processors 11 adapted to perform the functionality discussed herein. The processor 11 may include using a single integrated circuit (IC), or may include use of a plurality of integrated circuits or other components connected, arranged or grouped together, such as controllers, microprocessors, digital signal processors (DSP), parallel processors, multiple core processors, custom ICs, application specific integrated circuits (ASIC), field programmable gate arrays (FPGAs). The processor 11 may further include adaptive computing ICs and associated memory, e.g. RAM, DRAM and ROM, and other ICs and components. Hence, the term processor 11 should be understood to equivalently mean and include a single IC, or arrangement of custom ICs, ASICs, processors, microprocessors, controllers, FPGAs, adaptive computing ICs, or some other grouping of integrated circuits which perform the functions discussed for the computer-implemented method, with associated memory, such as microprocessor memory or additional RAM, DRAM, SDRAM, SRAM, MRAM, ROM, FLASH, EPROM or E2PROM.

[0218] The processor 11 with its associated memory may be adapted or configured via programming, FPGA interconnection, or hard-wiring to perform the methodology of the computer-implemented method. For example, the method may be programmed and stored, in the processor 11 with its associated memory or the memory 13, and other equivalent components, as a set of program instructions or other code for subsequent execution when the processor 11 is operative, e.g. powered on and functioning. The processor 11 may in particular provide the hardware on which the modules and submodules of a software-implementation of the method run.

[0219] The memory 11, which may include a data repository or database, may be embodied in any number of forms, including within any computer or other machine-readable data storage medium, memory device or other storage or communication device for storage or communication of information, including, but not limited to, a memory IC, or memory portion of an integrated circuit, e.g., a resident memory within a or processor 11, whether volatile or non-volatile, whether removable or non-removable, including without limitation RAM, FLASH, DRAM, SDRAM, SRAM, MRAM, FeRAM, ROM, EPROM or E2PROM, or any other form of memory device, such as a magnetic hard drive, an optical drive, a magnetic disk or tape drive, a hard disk drive, other machine-readable storage or memory media such as a floppy disk, a CDROM, a CD-RW, digital versatile disk (DVD) or other optical memory, or any other type of memory, storage medium, or data storage apparatus or circuit, know. In addition, such computer-readable media includes any form of communication media, which embodies computer readable instructions, data structures, program modules or other data in a data signal or modulated signal, such as an electromagnetic or optical carrier wave or other transport mechanism, including any information delivery media, which may encode data or other information in a signal, wired or wirelessly, including electromagnetic, optical, acoustic, RF or infrared signals, and so on. The memory 13 may be adapted to store various look up tables, parameters, coefficients, other information and data, programs or instructions of the software of the present disclosure, and other types of tables such as database tables.

[0220] The processor 11 is programmed, using software and data structures of the disclosed computer-implemented method, for example, to perform the methodology of the present disclosure. Consequentially, the system 10 and the computer-implemented method of the present invention may be embodied as software, which provides such programming or other instructions, such as a set of instructions and / or metadata embodied within a computer readable medium, discussed above. In addition, metadata may also be utilized to define the various data structures of a look up table or a database. Such software may be in the form of source or object code, by way of example and without limitation. Source code further may be compiled into some form of instructions or object code including assembly language instructions or configuration information.

[0221] The software, metadata, or other source code of the present invention and any resulting bit file (object code, database, or look up table) maybe embodied within any tangible storage medium, such as any of the computer or other machine-readable data storage media, as computer-readable instructions, data structures, program modules or other data, such as discussed above with respect to the memory 13, e.g., a floppy disk, a CDROM, a CD-RW, a DVD, a magnetic hard drive, an optical drive, or any other type of data storage apparatus or medium, as mentioned above.

[0222] The network interface 14 provides the system 10 with the capability to link to external databases, the plurality of data sources 18 and server(s) 17 via a communication network 16. The network interface 14 in particular enables to implement the system 10 in a spatially distributed manner by performing at least some of the individual steps of the computer-implemented method at least in part remote from the system 10.

[0223] The plurality of data sources 18 provide the information acquired by the system 10.

[0224] The invention is useful in number of application scenarios that include planning of mobility-related services that require an identification of locations for a plurality of service stations in a specific region of operation.

[0225] The invention is useful in application scenarios that include managing services that require a dynamic management of service requests with regular user interactions for distributing tasks and scheduling resources according to demands for services input by users and objectives of the operator of the service-providing network.

[0226] A specific example is a battery exchange service for scooters or rickshaws using battery exchange stations (BeX stations) as well as a charging service based on charging stations for e-bikes in urban regions or during leisure time in rural regions, e.g. forests and on longdistance cycle paths.

[0227] Another example are delivery services, ground-based or air-based, in which the service station enables the user to pick up their delivered items as well as to prepare and place items for delivery by the delivery service.

[0228] The region for locating the network of service stations may include regions of different geographic scales, from a single building, a logistic areas, a storehouse, a construction site, a large building complex, a nursing home, hospital area, a company site, an urban area to a region that covers a county or an entire country.

[0229] The service stations may include e.g. service stations with battery exchange stations or service stations with charging stations for charging batteries.

[0230] The service network may include a mobility service network, logistic service network, supply network, disposal network, or a network of base stations for autonomous devices. The mobility services may include scooter-based, rickshaw-based, e-bike-based, EV-based, aerial-vehicle-based or even unmanned aerial- vehicle-based (UAV-based) services. The mobility network may include, e.g., a car sharing network, a scooter-sharing network, a bicycle-sharing network, or a car rental network.

[0231] Logistic service network may include delivery services for, e.g. objects (goods), postal mailings, food and generally a wide range of logistic services.

[0232] Supply networks, e.g. water, represent a further application area. A supply network may include a water supply network, a gas supply network, a district heat network (long-distance heat), or a district cooling network.

[0233] The disposal network may include a sewage disposal network (waste water), or a material disposal network.

[0234] Each supply network needs adequate maintenance. In future, mobile robots may support or even autonomously provide the maintenance within the supply network.

[0235] The network of base stations for autonomous devices may serve autonomous devices that perform maintenance of infrastructure systems, or operate in a logistics facility.

[0236] The operator of the supply network needs to arrange service stations within the supply network at optimal locations such that the autonomous devices can maintain themselves, e.g. charge their batteries, or pick up and release tools required for performing their allotted maintenance task.

[0237] A recommendation system considering specific characteristics of such network, e.g. flow velocities at specific routes in the network for specific times can be similarly used as traffic flow analysis for thoroughfares.

[0238] All steps which are performed by the various entities described in the present disclosure as well as the functionalities described to be performed by the various entities are intended to mean that the respective entity is adapted to or configured to perform the respective steps and functionalities.

[0239] In the claims as well as in the description the word “comprising” does not exclude the presence of other elements or steps.

[0240] The indefinite article “a” or “an” does not exclude a plurality. A single element or other unit may fulfill the functions of several entities or items recited in the claims. The mere fact that different dependent claims recite certain measures and features of the control circuit does not exclude that a combination of these measures and features cannot combined in an advantageous implementation.

[0241] The features described in the discussion of specific embodiments and depicted in the figures may be combined with each other for the invention defined in the attached claims.

Claims

1. Claims:

1. Computer-implemented method for generating or amending locations for service stations, the method comprising:3.acquiring (Si), by an input interface (14), input information including transportation network information from a plurality of data sources (18);4.generating (S2), by a processor (11), a graph representation of a transportation network based on the acquired transportation network information;5.determining (S3), by the processor (11), a set of constraints based on the acquired input information;6.determining (S4), by the processor (11), a scoring measure for candidate locations for a service station;7.determining a scored list of the candidate locations of service stations based on the determined scoring measure for the candidate locations (S5);8.determining (S6), by the processor (11), a communicability distance measure of the transportation network by performing a spectral analysis of the graph representation;9.determining (S7), by the processor (11), a predetermined number of optimized locations for service stations by optimizing the determined communicability distance measure;10.performing (S8), by the processor (11), a multi-objective optimization of a multiobjective function for determining locations of the service stations based on the determined set of constraints, the scored list of candidate locations of service stations, and the predetermined number of optimized locations for service stations;11.generating (S9), by the processor (11), an output signal including the determined locations for service stations and outputting, via an output interface (15), the generated output signal.

2. Computer-implemented method according to claim 1, wherein13.the set of constraints includes constraints comprise at least one of a power grid connectivity, a parking lot availability, a safety score, a vehicle maneuverability during charging, weather protection, pedestrian safety, and a proximity to dealerships.

3. Computer-implemented method according to claim 1 or 2, whereinthe transportation network information includes open street map data.

4. Computer-implemented method according to any one of the preceding claims, wherein16.the scoring measure of each candidate location of the service station comprises a weighted sum of the individual constraints.

5. Computer-implemented method according to any one of the preceding claims, wherein18.the processor (11) computes the communicability distance measure of the transportation network based on the adjacency matrix of the graph representation or the corresponding Taylor series.

6. Computer-implemented method according to any one of the preceding claims, wherein20.the processor (11) determines the predetermined number of optimized locations for service stations by selecting the predetermined number of optimized locations for service stations that maximize the communicability distance measure of the graph representation.

7. Computer-implemented method according to any one of the preceding claims, wherein22.the multi-objective function comprises a weighted sum of a spatial feasibility measure, a network reachability measure, a stakeholder interest measure with regard to the closest service station, and a safety score.

8. Computer-implemented method according to any one of the preceding claims, wherein24.the service stations comprise at least one of a battery exchange station for exchanging a battery or a charging station for charging batteries.

9. Computer-implemented method according to any one of the preceding claims, wherein26.the service stations form part of a service network that includes at least one of a mobility service network, a logistic service network, a supply network, a disposal network, or a network of base stations for autonomous devices.

10. Computer-implemented method for determining at least one location of a service station for removal, addition or relocating in a network of service stations in a transportation network, the method comprising:28.acquiring (S11), by an input interface (14), input information including transportation network information from a plurality of data sources (18); acquiring (S12), by a processor (11), a graph representation of the transportation network;29.determining (S13), by the processor (11), future changes and demands to the transportation network and the network of service stations based on the acquired input information;30.determining (S14), by the processor (11), an adjusted cumulative objective function based on the graph representation of the transportation network, the future changes and demands to the transportation network and the network of service stations;31.performing (S15), by the processor (11), a multi-objective optimization of adjusted cumulative objective function for determining at least one location of a service station.32.generating (S16), by the processor (11), information on the determined at least one location of a service stations and outputting, via an output interface (15), the generated information.

11. Computer-implemented method for recommending a service station in a network of service stations, the method comprising:34.acquiring (S21), by an input interface (14), input information including transportation network information from a plurality of data sources (18);35.acquiring (S22), by an input / output interface (17), a request for a service by the network of service stations from a user;36.generating (S23), by a processor (11), a graph representation of a transportation network based on the acquired transportation network information;37.determining (S24), by the processor (11), a set of constraints based on the acquired input information;38.determining (S25), by the processor (11), a scoring measure including a charging speed score for each candidate location for a service station;39.determining a scored list of candidate locations of service stations based on the determined scoring measure for each candidate location (S26); performing (S27), by the processor (11), a multi-objective optimization of a multiobjective function for determining a location of a recommended service station based on the determined set of constraints, and the scored list of candidate locations of service stations,40.wherein the multi-objective function includes a weighted combination of a current location of the user requesting the service, a predicted waiting time until starting to provide the service, an utilization rate, and a waiting queue length;41.generating (S28), by the processor (11), output information based on the determined location of the recommended service station and outputting, via the input / output interface (17), the generated output information to the user.

12. System for generating or amending locations for service stations, the system comprising:43.an input interface (14) configured to acquire input information including transportation network information from a plurality of data sources (18);44.a processor (11) configured to generate a graph representation of a transportation network based on the acquired transportation network information,45.to determine a set of constraints based on the acquired input information,46.to determine a scoring measure for candidate locations for a service station,47.to determine a scored list of candidate locations of service stations based on the determined scoring measure for candidate locations,48.to determine a communicability distance measure of the transportation network by performing a spectral analysis of the graph representation,49.to determine a predetermined number of optimized locations for service stations by optimizing the determined communicability distance measure,50.to perform a multi-objective optimization of a multi-objective function for determining locations of the service stations based on the determined set of constraints, the scored list of candidate locations of service stations, and the predetermined number of optimized locations for service stations, and51.to generate an output signal including the determined locations for service stations; and52.an output interface (15) configured to output the output signal.

Images