DE102019109711A1 - Computer-implemented method for assigning vehicles to a vehicle fleet and vehicle assignment system - Google Patents

Abstract

The invention describes a computer-implemented method for assigning vehicles, in particular autonomous vehicles, a vehicle fleet, to inquiring customers, comprising the following steps: a) receiving a large number of inquiries from customers in a predetermined optimization period; b) determining time frames for the waiting times of each inquiring customer; c) determining an allocation with a global optimization method, whereby the waiting times of the inquiring customers and the costs incurred are minimized, as well as determining a fixed pick-up time for each inquiring customer; whereby for each inquiring customer initially the time frame for the waiting time determined in step b) and according to the global optimization process in step c) the fixed pick-up times are communicated to the requesting customers. In addition, the invention describes a computer-readable medium according to claim 15, an allocation unit according to claim 16 and a vehicle allocation system according to claim 17. The invention provides a particularly fast method for allocating vehicles from a vehicle fleet to inquiring customers, in which a reliable and time-efficient allocation of the vehicles of the vehicle fleet to be requested is reached.

Description

The invention relates to a computer-implemented method for assigning vehicles of a vehicle fleet to inquiring customers according to claim 1, a computer-readable medium according to claim 15, an allocation unit according to claim 16 and a vehicle allocation system according to claim 17.

The high number of vehicles, especially in large cities, means that traffic is becoming increasingly dense. The trend is increasingly leading to on-demand mobility (ODM) services. The advantage of ODM is that the vehicles in a car sharing vehicle fleet can be used more efficiently and unnecessary vehicle downtimes are reduced.

The challenge that arises is to find a time and cost-optimized solution for the overall system consisting of the vehicle fleet and customers. The aim is to provide the customer with precise information about the waiting time as quickly as possible and to find an optimized route for the vehicles, on the one hand to reduce the costs incurred for the vehicle fleet and on the other hand to reduce the customer's waiting time.

Due to the large solution space of the optimization problem that arises from the large number of variables, it is not computationally possible to find the global optimum of the optimization problem within a reasonable period of time. Furthermore, the dynamic nature of the problem requires a quick calculation of the optimum, since the problem changes with every customer request.

In the prior art, heuristics are used to find a solution to the optimization problem as quickly as possible. The speed is achieved at the expense of the quality. One approach that is often used for this is the Nearest Neighbor Policy (NNP). The shortest distance between all vehicles in the vehicle fleet and the customer's pick-up locations is determined.

However, only the waiting times for current customer inquiries and not the overall system are optimized, which increases the fleet costs and can increase the waiting times for subsequent inquiries.

For this reason, metaheuristics are commonly used that dynamically improve the solution in an iterative manner. However, there are also disadvantages here.

On the one hand, approaches are known from the prior art in which all customer inquiries are taken into account in the solutions that have not yet been served at the time of the calculation. The customer only finds out the waiting time shortly before the pick-up.

Furthermore, customer inquiries that have already been optimized are fixed in the state of the art, whereby these are no longer taken into account in subsequent optimizations. However, this has the disadvantage that customers only receive information about the waiting time after an optimization has been completed. On average, this leads to longer waiting times and higher customer dissatisfaction.

In general, it must be stated that there is still no satisfactory technical solution for the disadvantages described above. It is therefore the object of the invention to provide a method for assigning vehicles of a vehicle fleet to inquiring customers as quickly as possible, in which on the one hand a time-efficient allocation of the vehicles of the vehicle fleet to inquiring customers and on the other hand the best possible allocation between the inquiring customers and the vehicles of the vehicle fleet is achieved.

The object is achieved by a computer-implemented method for assigning vehicles to a vehicle fleet to be requested by customers according to claim 1, a computer-readable medium according to claim 15, an allocation unit according to claim 16 and a vehicle allocation system according to claim 17.

1. a) Empfangen einer Vielzahl von Anfragen von Kunden in einer vorbestimmten Optimierungsperiode;
2. b) Bestimmen von Zeitrahmen für die Wartezeiten jedes anfragenden Kunden;
3. c) Bestimmen einer optimalen Zuweisung mit einem globalen Optimierungsverfahren, wobei die Wartezeiten der anfragenden Kunden und anfallende Kosten minimiert werden, sowie Bestimmen eines festen Abholzeitpunktes für jeden anfragenden Kunden;

wobei an jeden anfragenden Kunden zunächst der in Schritt b) bestimmte Zeitrahmen für die Wartezeit und nach dem globalen Optimierungsverfahren in Schritt c) die festen Abholzeitpunkte an die anfragenden Kunden kommuniziert wird.In particular, the object is achieved by a computer-implemented method for assigning vehicles, in particular autonomous vehicles, to a vehicle fleet to inquiring customers, comprising the following steps:

1. a) receiving a plurality of requests from customers in a predetermined optimization period;
2. b) determining time frames for the waiting times of each inquiring customer;
3. c) Determining an optimal allocation with a global optimization method, the waiting times of the inquiring customers and the costs incurred being minimized, as well as determining a fixed pick-up time for each inquiring customer;

whereby the time frame determined in step b) for the waiting time and, according to the global optimization method in step c), the fixed pick-up times are communicated to the inquiring customer to each inquiring customer.

An essential core idea of the invention is based on the fact that inquiring customers are given a time frame for the waiting time. The time frame for the waiting time is used on the one hand to ensure that the inquiring customer can immediately see when a vehicle can be expected. In particular, the time frames are determined on the basis of a provisional assignment between vehicles and inquiring customers established in step b). It is therefore possible that vehicles can be dispatched directly even before the global optimization process in step c) has ended. Furthermore, the fixed pick-up times of the customers are communicated to the inquiring customers directly according to the global optimization process.

Another advantage of the present invention is that, in particular with the global optimization method, a better allocation with regard to waiting times and the costs incurred for the vehicle fleet can be found between the vehicles of the vehicle fleet and inquiring customers than with a conventional heuristic.

An autonomously driving vehicle is understood to mean a vehicle that has suitable sensors and, thanks to suitable algorithms, is able to drive with little or no human intervention.

A predetermined optimization period denotes a fixed period of time in which a large number of requests from customers can be received. For example, the optimization period can be around 2 minutes. It is also conceivable that the optimization period is between 2 and 10 minutes.

The waiting time of a requesting customer can be defined as the time that elapses between the received request from the customer and the fixed pick-up time and / or estimated pick-up times.

In particular, steps a) to c) are repeated periodically, as a result of which there is a dynamic response to new inquiries from customers.

In one embodiment, in step b) a heuristic, in particular a nearest neighbor policy method, is used to determine a vehicle of the vehicle fleet for each inquiring customer that can pick up the inquiring customer as early as possible. By using a heuristic, an initial assignment between the requesting customer and a vehicle in the vehicle fleet can be determined as quickly as possible.

Preferably, in step b) an estimated pick-up time is also determined for each inquiring customer and, in particular, a vehicle of the vehicle fleet is provisionally assigned to each inquiring customer, whereby a reliable first estimated pick-up time can be determined.

It is preferred that in step b) a predetermined time frame is placed around the estimated pick-up times of the inquiring customers. This makes it possible to set a time frame in such a way that it can actually be adhered to.

In particular, if a pick-up time estimated in step b) is before the end of the optimization period, the vehicle previously provisionally assigned to the requesting customer is permanently assigned. This speeds up the process, as inquiries from customers who can still be picked up during the optimization period are assigned directly.

In step c), a vehicle from the vehicle fleet is preferably permanently assigned to each inquiring customer. Duplicate assignments are prevented, which increases the efficiency of the assignment solution.

It is preferred that in step c) a metaheuristic, in particular a taboo search method, is used as the global optimization method, the waiting times of the inquiring customers and the costs likely to be incurred being minimized. By using a metaheuristic, a functionally better solution for the overall system can be found in a relatively short computing time. It is therefore possible to minimize the waiting times of the inquiring customer and the costs incurred for the operator of the vehicle fleet at the same time.

The inquiries from customers preferably include at least one pick-up location, and in particular one delivery location, so that global optimization of the overall system is made possible.

In a preferred embodiment, in step c) only those inquiring customers are taken into account in the global optimization method who were provisionally assigned to a vehicle in step b). If a vehicle has been permanently assigned to customers who have already made inquiries in step b), the fixed assignments are no longer taken into account again in the global optimization process. The fixed assignments therefore remain unchanged. This serves to reduce the solution space for the global optimization method, so that the method according to the invention is accelerated.

In particular, the vehicles that were provisionally and / or permanently assigned to a requesting customer in step b) drive in the direction of the pick-up location of the assigned customer, provided that the assigned vehicle is not being used by another customer. In this way, the waiting times of inquiring customers are not unnecessarily extended.

minimiert wird, wobei die Terme T₁ und T₂ wie folgt definiert sind: $$T_1=\underset{v\in C}{\mathrm{\Sigma }}\underset{a\in V}{\mathrm{\Sigma }}\underset{b\in V}{\mathrm{\Sigma }}{c}_{a,b}{x}_{a,b}^v,$$

$$T_2=\underset{i\in P}{\mathrm{\Sigma }}{\text{d}}_i,$$

mit einem Fahrzeug v, der Menge C aller aktuellen Fahrzeugstandorte und/oder aller Abgabeorte der zuvor fest zugewiesenen Fahrzeuge, der Menge V aller Abholorte P, Abgabeorte D und C, sowie einem Abholort i, wobei eine Indikatorvariable $x_{a,b}^v$

gleich 1 ist, wenn ein Fahrzeug v von einem Ort a nach einem Ort b fährt und ansonsten gleich 0 ist, wobei eine Gewichtung c_a,b, die Kosten eines Fahrzeugs umfasst, das von Ort a und Ort b fährt, und ein Strafparameter d_i ist.It is also preferred that in step c) a target function: $$f_{Opt}={\mathrm{<figure-callout\; id="1"\; label="T"\; filenames="DE102019109711A1\_0013.png"\; state="\{\{state\}\}">T</figure-callout>}}_1+T_2$$

is minimized, where the terms T ₁ and T ₂ are defined as follows: $${\mathrm{<figure-callout\; id="1"\; label="T"\; filenames="DE102019109711A1\_0013.png"\; state="\{\{state\}\}">T</figure-callout>}}_1=\underset{v\in \mathrm{C.}}{\mathrm{\Sigma }}\underset{a\in V}{\mathrm{\Sigma }}\underset{b\in V}{\mathrm{\Sigma }}{c}_{a,b}{x}_{a,b}^v,$$

$$T_2=\underset{i\in P}{\mathrm{\Sigma }}{\text{d}}_i,$$

with a vehicle v, the amount C of all current vehicle locations and / or all delivery locations of the previously permanently assigned vehicles, the amount V of all collection locations P, delivery locations D and C, and a collection location i, with an indicator variable $x_{a,b}^v$

equals 1 if a vehicle v drives from a location a to a location b and is otherwise equal to 0, where a weighting c _{a, b} comprises the costs of a vehicle that drives from location a and location b, and a penalty parameter d _i is.

The number of all inquiries known at the time of optimization is denoted by n and the number of all vehicles by m. Each inquiring customer is assigned a pick-up location i ∈ {1, ..., n} and a drop-off location i + n. The quantities of all pick-up and drop-off locations are denoted by P and D, respectively.

For all vehicles v ∈ {2n + 1, ..., 2n + m}, the set C represents the actual locations of the vehicles (for empty trips) or the delivery location of the last assigned vehicle request at the time of optimization.

The set P ∪ D ∪ C is represented by V, which represents all nodes of an undirected graph G = (V, A). The set of all edges A contains all possible combinations of the different locations (a, b), where a ∈ C and b ∈ P, a ∈ P and b ∈ D, a ∈ D and a ∈ P can be.

$${\mathrm{\Sigma }}_{i\in P}{x}_{i,i+n}^v-{\mathrm{\Sigma }}_{i,j\in P}{x}_{i+n,j}^v=0\forall v\in C,$$

$$t_{i,d}=t_{i,p}>0\forall i\in P.$$

The solution found by minimizing f _opt must also meet the following constraints: $${\mathrm{\Sigma }}_{v\in \mathrm{C.}}{x}_{i,i+n}^v=1\forall i\in P,$$

$${\mathrm{\Sigma }}_{i\in P}{x}_{i,i+n}^v-{\mathrm{\Sigma }}_{i,j\in P}{x}_{i+n,j}^v=0\forall v\in \mathrm{C.},$$

$$t_{i,d}=t_{i,p}>0\forall i\in P.$$

The first secondary condition specifies that each request can be assigned exactly once, i.e. with a vehicle. It is also ensured that no collections of vehicles are taken into account for individual inquiries, as every pick-up of a customer must be followed by a drop-off of the vehicle. The second secondary condition stipulates that the pick-up and drop-off of each customer is carried out by the same vehicle. The third secondary condition finally stipulates that the pick-up time of each customer t _{i, p} must take place before the respective drop-off time t _{i, d} .

The individual penalty parameters d _i are preferably dependent on the respective expected waiting times of the inquiring customer, as a result of which an allocation with the shortest possible waiting times is achieved.

In particular, the individual penalty parameters d _i increase more sharply as the expected waiting time of the inquiring customer increases. A greater increase in the penalty parameters with an increasing expected waiting time has the consequence that longer waiting times are weighted more heavily, whereby shorter waiting times are more preferred when minimizing the objective function.

The object of the invention is also achieved by a computer-readable medium, comprising instructions which, when executed by a processor, cause the processor to execute the method of one of the preceding claims.

Furthermore, the object of the invention is achieved by an assignment unit which is configured to receive location information from vehicles in a vehicle fleet and requests from customers, the requesting customer being assigned a vehicle from the vehicle fleet according to the method of the above type.

Furthermore, the object of the invention is achieved by a vehicle allocation system which has a vehicle fleet consisting of several vehicles, which preferably drive autonomously, and an allocation unit, with inquiring customers being allocated a vehicle from the vehicle fleet according to the method of the above type.

Further embodiments emerge from the subclaims.

In order to find solutions for the minimization of the objective function described above, a metaheuristic will be used. For example, a taboo search method can be used.

The basic idea of the taboo search procedure is not to stick to local optima (or minima) by allowing solutions that are worse in terms of the value of the objective function. At the same time, solutions that have already been found during the search are not allowed again.

The initial solution at the beginning of each optimization is generated, for example, by an NNP process. The iterative optimization then takes place.

With each iteration of the optimization process, a neighborhood is defined by applying local transformations to the current solution. This neighborhood is used to find new solutions.

For example, three different types of transformations can be applied: swaps, shifts and interchange. A swap exchanges requests within the route of a vehicle, while a shift moves the last request of the route of one vehicle to the end of the route of another vehicle. An interchange changes the solution by exchanging requests from two different vehicle routes.

All solutions of a neighborhood are ordered according to the value of the objective function f _obj . A so-called tabular list is created and checked. The tabular list shows whether a solution has already been found and therefore cannot be used as a new initial solution. A best neighbor solution is accepted, if it does not already exist, and added to the tabular list.

Accepted solutions are then compared with the best solution found, which is also the first solution in the first iteration. If the objective function value of the accepted solution is smaller than the objective function value of the best solution found, the best solution is replaced by the accepted solution.

The tab list can have a maximum number of entries. For example, the maximum number can be 10 Entries amount. When this maximum number of entries is reached, the oldest entry is replaced by the most recently added solution.

If the best neighbor solution is already on the tab list, it is not accepted and the next one on the ordered list of all neighbors is checked until i) a solution is found that is not on the tab list, ii) all neighboring solutions are checked, or iii) a maximum number of solutions already found is checked, which in this work corresponds to the maximum number of entries in the tab list.

If a solution is accepted, a new iteration begins with this solution as the initial solution. If ii) or iii) is the case, the search is canceled.

This process is carried out iteratively until it either aborts or until a maximum number of iterations is reached. For example, there can be a maximum number of iterations 100 be. This limitation of the iterations (as well as the relatively small maximum number of entries in the tab list) is chosen in order to concentrate on a fast search method, which is useful because the original solution is not generated randomly, but rather lies in a promising area of the solution space should.

• 1 ein Ablaufdiagramm von einem Ausführungsbespiel der vorliegenden Erfindung und
• 2 eine schematische Abbildung eines Ausführungsbeispiels der vorliegenden Erfindung.

In the following, the invention is explained in more detail on the basis of non-limiting exemplary embodiments with reference to the accompanying drawings. Here show:

• 1 a flow chart of an embodiment of the present invention; and
• 2 a schematic illustration of an embodiment of the present invention.

In 1 Figure 3 is a flow diagram of one embodiment of the present invention.

It will first step in 101 checked whether one or more new inquiries were received from customers. If so, the current vehicle locations and drop-off times of each assigned vehicle are checked.

In step 102 an NNP procedure is used to determine a vehicle for each new inquiring customer that could pick up the new inquiring customer at the earliest. The vehicle determined according to the NPP procedure is now provisionally assigned to the requesting customer. An associated pick-up time is estimated for each inquiring customer and a time frame with a predetermined length is placed around the estimated pick-up time. The time frame for the waiting time is communicated to the respective customer as soon as the NNP procedure has found the provisional allocation. Furthermore, the preliminary assignment is made in a list on which the queries to be taken into account in the global optimization process are collected.

After all new requests have been added to the requests to be optimized, step 103 checks whether the vehicles reach the pick-up locations of the provisionally assigned customers at this point in time. If so, be in step 104 These customer requests are permanently assigned to the respective vehicle by adding them to the current solution. Furthermore, these customer requests are removed from the list of requests to be taken into account by the global optimization method.

In step 105 it is checked whether the end of the optimization period has been reached. If this is not the case, step 100 proceeded to the next time step by receiving more new requests.

If the end of the optimization period has been reached, step 106 the global optimization method (the metaheuristic) optimizes those from the list of queries to be taken into account by the global optimization method according to the method described above. With the global optimization process, a solution is found that leads to a fixed allocation between the requesting customer and the vehicles.

In step 107 The exact pick-up time is now communicated to every inquiring customer. In addition, the list of queries to be taken into account by the global optimization method is emptied, so that with each subsequent optimization only new queries are taken into account.

In step 108 the current solution is replaced by the solution found by the global optimization procedure. A new optimization phase is then initiated with the current solution.

In 2 A schematic diagram of an embodiment of the present invention is depicted. The allocation unit 1 is trained to answer inquiries R1 , R2 from multiple customers K1 and K2 , during the optimization period T _{g, opt} to receive. The times of the inquiries are therefore before the optimization period T _{g, opt} . In addition, is the allocation unit 1 trained to provide location information I ₁ , I ₂ , I ₃ multiple vehicles F1 , F2 , F3 a vehicle fleet to receive.

The customer's inquiries can be sent to the assignment unit, for example, via an application and / or a program on a mobile phone and / or a user terminal. For this purpose, the assignment unit has corresponding receiving devices and / or a suitable interface in order to receive the inquiries from the customers.

The vehicles in the vehicle fleet are preferably autonomous vehicles. The current locations of the vehicles can be periodic and / or event-controlled (for example at the request of the assignment unit 1 ) to the allocation unit 1 be transmitted.

Time frames become before the end of the optimization period T _{g, opt} T1 _win , T2 _win intended for the waiting times of every inquiring customer. The time frames T1 _win , T2 _win can be determined by initially making a preliminary assignment between the vehicles using a heuristic (for example, according to the nearest neighbor policy) F1 , F2 , F3 the vehicle fleet and the inquiring customer K1 , K2 is made. An estimated pick-up time can be determined from the preliminary assignment, around which a time range with a predetermined length is placed. The timeframes so determined T1 _win , T2 _win are sent to the requesting customer before the end of the optimization period T _{g, opt} K1 , K2 Posted. In addition, the vehicles can be informed directly about the provisional assignment.

Then the global optimization process takes place (for example according to a metaheuristic such as taboo search), whereby the waiting times of the inquiring customers and the costs incurred are minimized. In the global optimization process, only those customer inquiries that were previously assigned are considered. There will also be a fixed pick-up time t _{1, p} , t _{2, p} for every inquiring customer K1 , K2 certainly. The fixed pick-up times t _{1, p} , t _{2, p} will be sent to the inquiring customer K1 , K2 communicates.

Furthermore is the assignment unit 1 trained to pick up locations p ₁ , p ₂ of the permanently assigned customer inquiries K1 , K2 to the assigned vehicle F1 , F2 , F3 to communicate. In the working example shown in 2 shown was vehicle F1 the customer K1 assigned, hence the pick-up location p ₁ on vehicle F1 transmitted. vehicle F3 became the customer K2 assigned, hence the pick-up location p ₂ on vehicle F3 transmitted.

List of reference symbols

100

Step in which a transition is made to a next iteration step;

101

Query whether new requests have been received;

102

Determine the time frame for the waiting times of the inquiring customers and communicate the waiting times to the customers;

103

Query whether the vehicles will reach the pick-up locations of the provisionally assigned customers;

104

fixed assignment;

105

Query whether the end of the optimization period has been reached;

106

global optimization process;

107

Communicating the fixed pick-up times to each requesting customer;

108

current solution;

1

Allocation unit;

F1, F2, F3

Vehicles in the vehicle fleet;

K1, K2

inquiring customers;

I ₁ , I ₂ I ₃

current locations of the vehicles in the vehicle fleet;

p ₁ , p ₂

permanently assigned pick-up location;

R1, R2

Customer inquiries;

T1 _win , T2 _win

Time frame for the waiting time of the inquiring customer; and

t _{1, p} , t _{2, p}

fixed pick-up time for inquiring customers;

Claims (17)

Computer-implemented method for assigning vehicles, in particular autonomous vehicles, a vehicle fleet, to inquiring customers, comprising the following steps: a) receiving a plurality of requests from customers in a predetermined optimization period; b) determining time frames for the waiting times of each inquiring customer; c) Determining an assignment with a global optimization method, the waiting times of the inquiring customers and the costs incurred being minimized, as well as determining a (fixed) pick-up time for each inquiring customer; whereby the time frame determined in step b) for the waiting time and, according to the global optimization method in step c), the fixed pick-up times are communicated to the inquiring customer to each inquiring customer. Procedure according to Claim 1 , steps a) to c) being repeated periodically. Procedure according to Claim 1 or 2 , wherein in step b) a vehicle of the vehicle fleet is determined for each inquiring customer with a heuristic, in particular with a nearest neighbor policy method, which vehicle can pick up the inquiring customer as early as possible. Method according to one of the preceding claims, wherein in step b) an estimated pick-up time is also determined for each inquiring customer and in particular a vehicle of the vehicle fleet is provisionally assigned to each inquiring customer. Method according to one of the preceding claims, wherein in step b) a predetermined time frame is placed around the estimated pick-up times of the inquiring customers. Method according to one of the preceding claims, wherein, if a pick-up time estimated in step b) is before the end of the optimization period, the vehicle previously provisionally allocated to the requesting customer is permanently assigned. Method according to one of the preceding claims, wherein in step c) a vehicle of the vehicle fleet is permanently assigned to each inquiring customer. Method according to one of the preceding claims, wherein in step c) a metaheuristic, in particular a taboo search method, is used as the global optimization method, the waiting times of the inquiring customers and the likely costs being minimized. Method according to one of the preceding claims, wherein the inquiries from the customers include at least one collection location, and in particular one delivery location. Method according to one of the preceding claims, in particular according to Claims 4 to 6th , whereby in step c) only those inquiring customers are considered in the global optimization process who were provisionally assigned to a vehicle in step b). Method according to one of the preceding claims, wherein the vehicles that were temporarily and / or permanently assigned to a requesting customer in step b) drive in the direction of the pick-up location of the assigned customer, provided the assigned vehicle is not being used by another customer. Method according to one of the preceding claims, wherein in step c) a target function is achieved by the global optimization method: $$f_{Opt}=T_1+T_2$$

is minimized, where the terms T ₁ and T ₂ are defined as follows: $$T_1=\underset{v\in \mathrm{C.}}{\mathrm{\Sigma }}\underset{a\in V}{\mathrm{\Sigma }}\underset{b\in V}{\mathrm{\Sigma }}{c}_{a,b}{x}_{a,b}^v$$

$$T_2=\underset{i\in P}{\mathrm{\Sigma }}{\text{d}}_i,$$

with a vehicle v, the amount C of all current vehicle locations and / or all delivery locations of the previously permanently assigned vehicles, the amount V of all collection locations P, delivery locations D and C, and a collection location i, with an indicator variable $x_{a,b}^v$

equals 1 if a vehicle v drives from a location a to a location b and is otherwise equal to 0, where a weighting c _{a, b} comprises the costs of a vehicle that drives from location a and location b, and a penalty parameter d _i is. Method according to one of the preceding claims, in particular according to Claim 11 , whereby the individual penalty parameters d _{i are} dependent on the respective expected waiting times of the inquiring customer. Method according to one of the preceding claims, in particular according to Claim 12 or 13 , whereby the individual penalty parameters d _i increase more strongly with increasing expected waiting time of the inquiring customer. A computer readable medium comprising instructions which, when executed by at least one processor, cause the processor to carry out the method of any preceding claim. Assignment unit which is configured to receive location information from vehicles of a vehicle fleet and inquiries from customers, wherein the requesting customer according to the method according to one of Claims 1 to 14th a vehicle from the vehicle fleet is assigned. Vehicle allocation system which has a vehicle fleet consisting of several vehicles, which preferably drive autonomously, and an allocation unit, with inquiring customers according to the method according to one of the Claims 1 to 14th a vehicle from the vehicle fleet is assigned.

Images