Method for controlling an elevator

Abstract

A method for controlling an elevator (1) comprising: cars (9a, 9b, 9c) arranged to be movable along different vertical axes between floors (5) of a building (3); and a sensor system (12) adapted to provide sensor data (13) indicative of a current state of the elevator, the method comprising: receiving sensor data comprising a current position (15) of each car relative to a floor, a list (17) of current calls assigned to the cars and a new call (19) to be assigned to one of the cars, wherein each call indicates a desired arrival floor of the building; generating a list (33) of cars (9a, 9b) that meet the condition from the list (35) of cars using the sensor data and at least one rule that each car should meet when performing the new call; inputting the sensor data as input data (14) into an artificial neural network (37) trained to convert the input data into an output value (39) for each car, where the output value represents a probability and / or a trend that a car is assigned to a new call; determining an item (9a) as a selected car (43) from the list of cars that meet the condition using the output value; a new call is assigned to the selected car.

Description

Technical Field

[0001] The present invention relates to a method for controlling an elevator, an elevator, an elevator controller, a computer program and a computer readable medium for executing the method. Background Art

[0002] An elevator may include different cars for transporting people and / or objects between different floors of a building along different vertical axes or shafts. Given a list of calls indicating the desired arrival floors of a building, the controller of the elevator needs to be able to assign each call to one of the cars and determine at which floors the car should stop in order to execute the assigned calls in an efficient manner. This may also be called scheduling.

[0003] Especially in high-rise buildings with multiple floors and multiple elevators operating simultaneously, solving the trip scheduling problem using traditional elevator control algorithms, which are usually based on a set of known rules, can become very complex. In this case, machine learning models can be used to make better control decisions. Such models are usually trained using reinforcement learning, where the actions of a simulated elevator (the "agent") are estimated by rewards, and the algorithm controlling the agent (for example in the form of an artificial neural network) is optimized based on the rewards so that future control decisions are better than past control decisions.

[0004] US 5,767,461 A describes a method for controlling elevators with cars located in separate elevator shafts using artificial neural networks. EP 2 149 532 A1 and EP 1 733 990 A1 describe methods that use dedicated rules to select one of a plurality of elevator cars when a new call arrives. Summary of the invention

[0005] An object of the present invention is to provide an improved control method for an elevator, the elevator comprising two or more cars. Another object of the present invention is to provide a corresponding elevator, a corresponding elevator controller, a corresponding computer program and a corresponding computer readable medium.

[0006] These objects are achieved by the subject-matter of the independent claims. Advantageous embodiments are defined in the dependent claims and in the following description and the drawings.

[0007] A first aspect of the invention relates to a computer-implemented method for controlling an elevator, wherein the elevator comprises: different cars arranged to be movable along different vertical axes between different floors of a building; and a sensor system adapted to provide sensor data indicating the current state of the elevator. The method comprises: receiving sensor data, wherein the sensor data comprises the current position of each car relative to the floor, a list of current calls assigned to the cars, and a new call to be assigned to one of the cars, wherein each call indicates a desired arrival floor of the building; using the sensor data and at least one rule that each car should comply with when executing the new call, generating a list of eligible cars from the list of cars; inputting the sensor data as input data into an artificial neural network, the artificial neural network being trained to convert the input data into an output value for each car, wherein the output value indicates the probability and / or tendency of the car being assigned to the new call; using the output value to determine an item from the list of eligible cars as a selected car; and assigning the new call to the selected car.

[0008] In other words, the list of eligible cars can be used as a filter for filtering the output of the artificial neural network. This filtering can also be called masking. This significantly improves the accuracy of the method compared to elevator group control algorithms that lack this masking.

[0009] The method may be automatically performed by a processor.

[0010] The current position of the car may be the ID of a floor on or near the floor where the car is currently located.

[0011] The desired arrival floor may be a destination floor and / or a starting floor. For example, each call may be defined by a unique ID, a timestamp, a starting floor, and a destination floor.

[0012] Generating the list of qualifying cars may include, for example, estimating a future state of each car from the input data and estimating the future state and / or a transition from a current state of the car to the future state using at least one rule.

[0013] The rule(s) may be based on predefined and / or known parameters (e.g., as opposed to machine learning models that have hidden parameters).

[0014] At least two different rules can be used to generate a list of eligible cars. In this case, a car can be added to the list of eligible cars only if it meets each rule when executing a new call.

[0015] The artificial neural network can be, for example, a multi-layer perceptron (MLP), a convolutional neural network (CNN), a recurrent neural network (RNN), a long short-term memory (LSTM), a graph neural network (GNN), or a combination of at least two of these examples. In particular, the artificial neural network can be trained using a policy-based and / or value-based reinforcement learning algorithm.

[0016] The input data may be encoded in a format specifically suitable for the artificial neural network, for example in the form of a vector of sensor data. The input data may also be identical to the sensor data. In this case, the sensor system may be adapted to provide the sensor data as input data in a format specifically suitable for the artificial neural network.

[0017] The input data may be input into one, two or more artificial neural networks, such as one artificial neural network per car, to generate output value(s).

[0018] The input data may also include additional data indicating the current state of the elevators. For example, the additional data may include the current direction of movement of each car (e.g., "up", "down", "stop") and / or the current occupancy (e.g., "full", "not full"). This may improve the accuracy of the method.

[0019] For example, the input data can be an observation vector with different components, such as the floor number at or near the floor each car is currently located at (e.g., "0", "1", "2"), the current moving direction of each car (e.g., "up", "down", "stop"), the occupancy rate of each car (e.g., "full", "not full"), the current calls assigned to each car, and the new calls to be assigned to one of the cars.

[0020] A simple way to encode a car call may be to combine the following three aspects.

[0021] 1) All landing operating panel (LOP) calls assigned to the car can be provided as origin-destination tuples and aggregated into a (two-dimensional) origin-destination matrix. The matrix can then be converted to a matrix of length The first vector of is the number of floors of the building. A so-called LOP call comprises in particular information about the origin floor and information about the destination floor. Such LOP calls are often referred to as destination calls.

[0022] 2) All new LOP calls to be assigned to cars can be converted to lengths of The second vector of .

[0023] 3) All car operating panel (COP) calls from the car can be directly stored in a length of This is because there is no need to remember where the COP call came from.

[0024] Then, the first vector, the second vector, and the third vector of all cars may be linked to form an observation vector.

[0025] The output value can be a value within a discrete and / or continuous value range. For example, the output value can be a Q value and / or a bid value. The output value can also be a percentage or a Boolean value between 0 and 1.

[0026] A second aspect of the invention relates to an elevator controller comprising a processor configured to perform the above and below methods. An elevator controller is a special data processing device designed and programmed to control an elevator. An elevator controller may include an elevator group controller and a car controller. An elevator controller may include hardware and / or software modules. In addition to a processor, an elevator controller may also include a memory and one or more data communication interfaces for data communication with peripheral devices (e.g., sensors, sensor systems, or actuators of an elevator).

[0027] A third aspect of the invention relates to an elevator. The elevator comprises: at least two different cars arranged to be movable along different vertical axes between different floors of a building; a sensor system adapted to provide sensor data indicative of the current state of the elevator, wherein the sensor data comprises the current position of each car relative to the floor, a list of current calls assigned to the cars, and new calls to be assigned to one of the cars, wherein each call indicates a desired arrival floor of the building; and an elevator controller as described above and below.

[0028] For example, the car can be arranged to be able to move along different elevator shafts of a building. Specifically, each car can be arranged to be able to move in its own elevator shaft.

[0029] Other aspects of the invention relate to a computer program comprising instructions which, when executed by a processor, cause the processor to perform the method described above and below, and to a computer-readable medium in which the computer program is stored.

[0030] The computer readable medium may be a volatile or non-volatile data storage device. For example, the computer readable medium may be a hard disk, a USB (Universal Serial Bus) storage device, a RAM (Random Access Memory), a ROM (Read Only Memory), an EPROM (Erasable Programmable Read Only Memory), an EEPROM (Electrically Erasable Programmable Read Only Memory), a flash memory, or a combination of at least two of these examples. The computer readable medium may also be a data communication network, such as the Internet or a cloud, for downloading program code.

[0031] It should be noted that features of the method described above and below may also be features of the elevator controller, the computer program and the computer-readable medium (and vice versa).

[0032] Embodiments of the present invention may be considered to be based on the following ideas and findings, but not to limit the present invention.

[0033] According to one embodiment, the following steps may be performed for each car to generate a list of eligible cars: assign a new call to the car; simulate the movement of the car to perform the new call based on the car's current position and the current calls assigned to the car; analyze the movement to determine if the car meets at least one rule; if the car meets at least one rule, add the car to the list of eligible cars. Such an algorithm may be easily implemented.

[0034] According to one embodiment, the movement can be simulated according to a (known) control logic to control the car independently of any other car of the elevator. Such a single car control logic can be, for example, a selective collective logic or a down collective logic. The control logic can be based on known rules and / or known parameters. However, the control logic can also include a machine learning model with hidden (i.e. unknown) parameters, such as an artificial neural network. The same control logic can be used to simulate the movement of the car and generate control commands for actually controlling the car.

[0035] According to one embodiment, at least one rule may specify a preferred direction of movement for each car. In other words, a car can be added to the list of eligible cars only when the car moves along the preferred direction of movement when executing a new call.

[0036] According to one embodiment, at least one rule can specify the maximum number of calls that can be assigned to each car. In other words, a car can be added to the list of eligible cars only if the number of calls assigned to it when executing a new call does not exceed its maximum number of calls.

[0037] According to one embodiment, the selected car may be the eligible car corresponding to the highest output value (eg, highest Q value). Alternatively, the selected car may be the eligible car corresponding to the lowest output value (eg, lowest bid value).

[0038] According to one embodiment, an artificial neural network may be trained using training data generated by simulating an environment in which different cars of an elevator move along different vertical axes between different floors of a building in response to calls indicating desired arrival floors of the building.

[0039] The environment can be considered as a computer-computed space in which the elevator as a virtual agent takes actions. The environment can have the same and / or similar features as a real elevator and / or a real building in which the real elevator is installed.

[0040] For example, the environmental simulation may include the following aspects: elevator simulation, traffic simulation, interaction between elevator simulation and traffic simulation. The elevator simulation may be configured with at least one of the following parameters: maximum speed, acceleration, door opening time, door closing time, minimum door opening time, car capacity. The traffic simulation may be performed based on a specific traffic pattern that determines the local and / or temporal distribution of calls to be executed by the elevator. For example, as described above, each call may be characterized by a unique ID, a timestamp, a starting floor, and a destination floor. The traffic pattern may change during the simulation. The simulation of the interaction between the elevator and the call (i.e., the passenger) may be configured with parameters such as waiting time (i.e., the time the passenger waits before entering the car), travel time (i.e., the time the passenger travels before leaving the car), or destination time (i.e., the total time required to complete the call).

[0041] In general, an agent (e.g., an elevator group controller) can be configured to receive observations of the environment at each simulation step and perform actions based on the observations, such as assigning a new call to one of the cars and optionally controlling the car accordingly. In reaction, the environment can update its state and emit observations of its updated state along with a reward, which is used to train the agent to improve its actions in future simulation steps.

[0042] Rewards can be positive, negative, or zero. For example, a positive reward value might indicate a "more useful" action, while a negative reward value might indicate a "less useful" action. Rewards can be calculated periodically, or only in specific circumstances (such as when a transition from the current state to the next state of the environment is observed).

[0043] The environment state can be considered as a combination of the elevator state and the traffic state. The elevator state can be defined, for example, by at least one of the following parameters: the current floor of each car, the current moving direction of each car, the current door state of each car (open, moving, closed), the current number of passengers in each car, the next reachable floor of each car. The traffic state can be defined, for example, by at least one of the following parameters: active calls from and / or to each floor, the current number of passengers in each car, the current number of passengers waiting on each floor.

[0044] During training, actions may be chosen with some randomness, for example by using an epsilon-greedy algorithm. The degree of randomness may change, and in particular decrease, as training progresses.

[0045] According to one embodiment, the method may further comprise generating a control command to cause the selected car to execute a new call. The control command may be generated to control the movement of the selected car along one of the vertical axes, such as along a shaft of a building. The control command may be generated taking into account other calls assigned to the selected car. Examples of such control commands are "up", "down", "go to floor X" or "go to next floor". The control command may also cause the selected car to close and / or open its door(s) and / or execute other calls assigned to it.

[0046] According to one embodiment, generating a control command may include: receiving additional input data, the additional input data including the current position of the selected car and a list of all calls assigned to the selected car, including the new call; inputting the additional input data into an additional artificial neural network, the additional artificial neural network being trained to convert the additional input data into at least one additional output value, the at least one additional output value indicating a possible control command from a set of possible control commands for controlling the selected car; selecting one of the possible control commands as the control command using the at least one additional output value; applying the control command to cause the selected car to execute the new call. In other words, different artificial neural networks may be used to assign the new call to one of the cars and to control the selected car to which the new call has been assigned. This significantly improves the accuracy and / or efficiency of the method.

[0047] For example, the additional artificial neural network may be configured to calculate an output value, such as a Q value, for each possible control command. In this case, the control command may be the possible control command corresponding to the highest output value (eg, the highest Q value).

[0048] According to one embodiment, an additional artificial neural network may be trained using training data generated by simulating an environment in which at least one car of an elevator moves between different floors of a building in response to a call indicating a desired arrival floor of the building. The environment may have the same and / or similar features as a real elevator and / or a real building in which a real elevator is installed. In particular, the environment may include only one car. The additional artificial neural network may be trained in a manner similar to that described above. However, in this case, the agent (e.g., an elevator car controller) may be configured to perform actions in the environment by applying control commands for controlling a single car of the elevator (rather than assigning new calls to different cars of the elevator).

[0049] According to one embodiment, the processor of the elevator controller can be configured to perform the method of the first aspect of the present invention according to an embodiment of generating control commands. In this case, the elevator may also include an actuator system adapted to control the car according to the control commands generated by the elevator controller. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The preferred embodiments of the present invention will be described in more detail below in conjunction with the accompanying drawings. Neither the description nor the drawings should be construed as limiting the present invention.

[0051] Figure 1 An elevator according to an embodiment of the invention is shown.

[0052] Figure 2 An elevator group controller of an elevator controller according to an embodiment of the present invention is shown.

[0053] Figure 3 A car controller of an elevator controller according to an embodiment of the present invention is shown.

[0054] The accompanying drawings are schematic diagrams only and are not drawn to scale. The same reference numerals in the drawings denote the same features or features having the same effect. DETAILED DESCRIPTION

[0055] Figure 1 An elevator 1 is shown installed in a building 3 having a plurality of floors 5 and a plurality of elevator shafts, here a first elevator shaft 7a, a second elevator shaft 7b and a third elevator shaft 7c, connecting the floors 5. The elevator 1 comprises a plurality of cars for transporting passengers 8 between the floors 5, here a first car 9a, which is arranged to be movable along the first elevator shaft 7a; a second car 9b, which is arranged to be movable along the second elevator shaft 7b; and a third car 9c, which is arranged to be movable along the third elevator shaft 7c.

[0056] The elevator 1 further comprises an actuator system 11 adapted to control the movement of each car 9a, 9b, 9c. In this example, the actuator system 11 comprises one electric drive ("M") for each car 9a, 9b, 9c.

[0057] Furthermore, the elevator 1 comprises a sensor system 12 which is adapted to provide sensor data 13 comprising: a current position 15 of each car 9a, 9b, 9c relative to the floor 5 (see Figure 2 , Figure 3 ), list 17 of current calls assigned to cars 9a, 9b, 9c (see Figure 2 ), and a new call 19 to be assigned to one of the cars 9a, 9b, 9c (see Figure 2 , Figure 3 ).

[0058] like Figure 1 As shown, an elevator controller 21 of the elevator 1 may be configured to generate control commands 22 for controlling the actuator system 11 from the sensor data 13 .

[0059] In this example, the elevator controller 21 includes: a group controller 23 (see Figure 2 ), which is configured to assign the new call 19 to one of the cars 9a, 9b, 9c; and three car controllers 25 (see Figure 3 ), wherein each car controller 25 is configured to generate a control command 22 for one of the cars 9a, 9b, 9c (ie the corresponding electric drive). The car controllers 25 can generate their control commands 22 independently of each other.

[0060] The elevator controller 21 includes a processor 27 and a memory 29. The processor 27 may be configured to perform the following method by executing a computer program stored in the memory 29.

[0061] In a first step, the elevator controller 21 , here the car estimation module 31 of the group controller 23 , receives the sensor data 13 .

[0062] In a second step, the car estimation module 31 generates a list 33 of qualifying cars from a list 35 of all existing cars 9a, 9b, 9c using the sensor data 13 and at least one rule that each car 9a, 9b, 9c should comply with when executing the new call 19. In this example, the list 33 includes cars 9a, 9b as qualifying cars.

[0063] At the same time, in a third step, the input data 14 encoding the sensor data 13 is input into a trained artificial neural network 37 (for example, by a policy-based and / or value-based reinforcement learning algorithm) to convert the input data 14 into a set of output values ​​39, here three output values ​​39, each of which represents the probability and / or trend of one of the three cars 9a, 9b, 9c.

[0064] In the fourth step, the list 33 of eligible cars and the output value 39 are input to the car selection module 41, which determines the selected car 43 from the eligible cars based on the output value 39 assigned to each item in the list 33. In this example, car 9a is the item with the highest output value 39 and is therefore determined to be the selected car 43.

[0065] In a fifth step, the new call 19 is assigned to the selected car 43 in the call assignment module 45 .

[0066] In a sixth step, the car controller 25 of the selected car 43 additionally generates a control command 22 which causes the actuator system 11 , ie the electric drive of the selectable car 43 , to move the selected car 43 so that it executes the new call 19 .

[0067] Generally speaking, if there is The last layer of the neural network 37 can generate An output value 39 , such as a Q value, may then be determined by the car selection module 41 as the selected car 43 with the highest Q value.

[0068] As described above, the selected car 43 may follow a single car logic, such as a selection collective logic or a down collective logic, to satisfy the call assigned to it. The single car logic may be based on a set of known rules and / or implemented as an artificial neural network (see below).

[0069] For example, by ”) executes the following algorithm to generate a list 33 of eligible cars.

[0070] i) Assign new call 19 to .

[0071] ii) Follow the single car logic simulation movement.

[0072] iii) Check whether one or more rules are violated at point (ii).

[0073] iv) If there is no violation of the rules, Added to the list 33 of eligible cars.

[0074] The list 33 can then be used to mask the output of the neural network 37, i.e., to filter the output values ​​39 (" ”):

[0075]

[0076] If the list 33 is empty, the new call 19 may be delayed for some time.

[0077] For example, the following two rules can be used to evaluate simulated moves.

[0078] 1. Wrong direction: If the car moves in the opposite direction to the preferred direction of movement to serve a new call, the new call cannot be assigned to the car.

[0079] 2. Full Car: If the car exceeds the maximum number of calls that can be assigned to it when executing a new call, no new call can be assigned to the car. In other words, if the car is full, new passengers will not be able to enter the car at the starting floor specified by the new call.

[0080] Optionally, the car controller 25 may be configured to estimate the control command 22 for the selected car 43 using an additional artificial neural network 47 that may be trained in a similar manner to the neural network 37 (i.e., using a reinforcement learning algorithm) to convert additional input data 49 including the current position 15 of the selected car 43 and a list 51 of all calls currently assigned to the selected car 43 (including the new call 19) into a set of additional output values ​​53, wherein each additional output value 53 indicates a control command from a set of possible control commands for controlling the selected car 43. The additional input data 49 may have been generated using the sensor data 13 and / or the input data 14.

[0081] The additional output values ​​53 may then be input into a command selection module 55 , which selects the possible control command corresponding to the highest additional output value 53 as the control command 22 , and applies the control command 22 to the actuator system 11 .

[0082] As mentioned above, the additional neural network 47 may also be used to simulate the movement of one of the cars 9a, 9b, 9c, for example to evaluate the control commands for one of the cars 9a, 9b, 9c to determine whether this car meets the conditions.

[0083] The two neural networks 37, 47 may have been trained using training data generated by simulating the behavior of an elevator 1 in a virtual building responding to a virtual call.

[0084] The above modules may be software and / or hardware modules.

[0085] Finally, it is noted that the terms "comprises", "comprising", "having" or "with" do not exclude other elements or steps and the indefinite article "a" or "an" does not exclude a plurality. It is also noted that features or steps described with reference to one of the above embodiments may also be used in combination with features or steps described with reference to any other embodiment described above. Reference signs in the claims shall not be construed as limiting the scope of the claims.

Claims

1. A computer-implemented method for controlling an elevator (1), wherein the elevator (1) comprises: Different cars (9a, 9b, 9c) arranged to be movable along different vertical axes between different floors (5) of a building (3); and a sensor system (12) adapted to provide sensor data (13) indicative of a current state of the elevator (1), wherein the method comprises: receiving sensor data (13), wherein the sensor data (13) includes a current position (15) of each car (9a, 9b, 9c) relative to a floor (5), a list (17) of current calls assigned to the cars (9a, 9b, 9c), and new calls (19) to be assigned to one of the cars (9a, 9b, 9c), wherein each call indicates a desired arrival floor of the building (3); Generating a list (33) of eligible cars (9a, 9b) from the list (35) of cars (9a, 9b, 9c) using the sensor data (13) and at least one rule that each car (9a, 9b, 9c) should comply with when executing the new call (19); Inputting the sensor data (13) as input data (14) into an artificial neural network (37), the artificial neural network being trained to convert the input data (14) into an output value (39) for each car (9a, 9b, 9c), wherein the output value (39) indicates the probability and / or tendency of the car (9a, 9b, 9c) being assigned to the new call (19); Using the output value (39) to determine an item (9a) from the list (33) of eligible cars (9a, 9b) as a selected car (43); The new call (19) is assigned to the selected car (43).

2. The method according to claim 1, The following steps are performed for each car (9a, 9b, 9c) to generate a list (33) of cars (9a, 9b) that meet the conditions: assigning the new call (19) to the car (9a, 9b, 9c); Simulating the movement of the car (9a, 9b, 9c) for executing the new call (19) based on the current position of the car (9a, 9b, 9c) and the current call assigned to the car (9a, 9b, 9c); analyzing the movement to determine whether the car (9a, 9b, 9c) complies with the at least one rule; If the car (9a, 9b, 9c) meets the at least one rule, the car (9a, 9b, 9c) is added to a list (33) of eligible cars (9a, 9b).

3. The method according to claim 2, wherein said movement is simulated according to a control logic for controlling said car (9a, 9b, 9c) independently of any other car (9a, 9b, 9c).

4. A method as claimed in any one of the preceding claims, The at least one rule specifies a preferred direction of movement for each car (9a, 9b, 9c).

5. A method as claimed in any one of the preceding claims, The at least one rule specifies a maximum number of calls that can be assigned to each car (9a, 9b, 9c).

6. A method as claimed in any one of the preceding claims, wherein the selected car (43) is the eligible car (9a, 9b) corresponding to the highest output value (39); or The selected car (43) is the eligible car (9a, 9b) corresponding to the lowest output value (39).

7. A method as claimed in any one of the preceding claims, The artificial neural network (37) has been trained using training data generated by simulating an environment in which different cars (9a, 9b, 9c) of an elevator (1) move along different vertical axes between different floors (5) of the building (3) in response to calls indicating desired arrival floors of the building (3).

8. The method according to any one of the preceding claims, further comprising: A control command (22) is generated to cause the selected car (43) to execute the new call (19).

9. The method according to claim 8, The generating control command (22) includes: receiving additional input data (49), the additional input data comprising a current position (15) of the selected car (43) and a list (51) of all calls assigned to the selected car (43), including the new call (19); inputting the additional input data (49) into an additional artificial neural network (47) that is trained to convert the additional input data (49) into at least one additional output value (53) that indicates one of a set of possible control commands for controlling the selected car (43); selecting one of the possible control commands as a control command (22) using the at least one additional output value (53); Applying the control command (22) causes the selected car (43) to execute the new call (19).

10. The method according to claim 9, The additional artificial neural network (47) has been trained using training data generated by simulating an environment in which at least one car (9a, 9b, 9c) of an elevator (1) moves between different floors (5) of the building (3) in response to a call indicating a desired arrival floor of the building (3).

11. An elevator controller (21) comprising a processor (27) configured to execute the method of any one of claims 1 to 6 or 8 to 9.

12. An elevator (1), comprising: Different cars (9a, 9b, 9c) arranged to be movable along different vertical axes between different floors (5) of a building (3); a sensor system (12) adapted to provide sensor data (13) indicative of a current state of the elevator (1), wherein the sensor data (13) comprises a current position (15) of each car (9a, 9b, 9c) relative to a floor (5), a list (17) of current calls assigned to the cars (9a, 9b, 9c), and new calls (19) to be assigned to one of the cars (9a, 9b, 9c), wherein each call indicates a desired arrival floor of the building (3); An elevator controller (21) as claimed in claim 11.

13. The elevator (1) according to claim 12, wherein the processor (27) of the elevator controller (21) is configured to perform the method as claimed in any one of claims 8 to 9; The elevator (1) further comprises an actuator system (11), which is suitable for controlling the car (9a, 9b, 9c) according to a control command (22) generated by the elevator controller (21).

14. A computer program comprising instructions which, when executed by a processor (27), cause the processor (27) to perform the method of any one of claims 1 to 6 or 8 to 9.

15. A computer readable medium comprising instructions which, when executed by a processor (27), cause the processor (27) to perform the method of any one of claims 1 to 6 or 8 to 9.

Images