Passenger flow prediction device, passenger flow prediction system, passenger flow prediction method, and passenger flow prediction program

Abstract

To obtain a passenger flow prediction device capable of predicting a passenger flow indicating a movement situation of passengers who are customers.SOLUTION: A passenger flow prediction device 10 includes: a data collection unit 11 for collecting operation-related data regarding operation of a railroad from a plurality of railroad operators; an accumulation data storage unit 12 for accumulating the operation-related data collected by the data collection unit 11; and a prediction control unit 13 for predicting a passenger flow of railroad users using the railroad based on the operation management data collected from the plurality of railroad operators.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present disclosure relates to a passenger flow prediction device, a passenger flow prediction system, a passenger flow prediction method, and a passenger flow prediction program that predict passenger flow based on information held by railway operators.

Background Art

[0002] Railway operators are required to achieve both an improvement in the quality of transport services and efficient train operation in order to acquire customers and improve their balance of payments.

[0003] One way to improve the quality of transport services is to reduce the congestion level of trains. Increasing the number of vehicles per train formation can reduce the congestion level of trains, but there are cases where the number of vehicles per train formation cannot be increased due to restrictions on the length of the platform. In such cases, it is necessary to increase the number of train operations in order to reduce the congestion level of trains.

[0004] However, increasing the number of train operations reduces the fare revenue per train formation and increases the number of crew required for operation. Therefore, railway operators need to determine the number of train operations so that the congestion level is such that passengers, who are customers, do not become dissatisfied and the balance of payments is in the black.

[0005] The number of railway passengers is not always constant and varies depending on the season, weather, time of day, and the presence or absence of events. For this reason, railway operators are required to predict future passenger flow and then determine the number of train operations so that the congestion level is such that passengers do not become dissatisfied and the balance of payments is in the black.

[0006] Patent Document 1 discloses a demand prediction system that retains past passenger predictions of transportation means and external event information, etc., and synthesizes predicted values of annual trends, seasons, and time-of-day models to perform passenger demand prediction. By using the demand prediction system disclosed in Patent Document 1, when a railway operator expects an increase in passengers due to an event or the like, the railway operator can take measures such as increasing the number of extra trains and determine the number of train operations in accordance with the predicted future passenger flow.

Prior Art Documents

Patent Documents

[0007]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0008] In recent years, with the development of the railway network, interchangeable stations where multiple railway operators can enter have been opened, or multiple railway operators have come to mutually enter. For this reason, the operation status of a certain railway operator often affects the operations of other railway operators.

[0009] Since the demand prediction system disclosed in Patent Document 1 collects information only from one railway operator to predict the passenger flow, there is a possibility that the accuracy of predicting the passenger flow will be reduced.

[0010] The present disclosure has been made in view of the above, and an object thereof is to obtain a passenger flow prediction device that can accurately predict a passenger flow indicating the movement status of passengers who are customers.

Means for Solving the Problems

[0011] In order to solve the above-described problems and achieve the object, a passenger flow prediction device according to the present disclosure includes a data collection unit that collects operation-related data regarding railway operations from a plurality of railway operators, and a storage data storage unit that stores the operation-related data collected by the data collection unit. The passenger flow prediction device includes a prediction control unit that predicts the passenger flow of railway users who use the railway based on the operation-related data collected from the plurality of railway operators.

Effect of the Invention

[0012] According to the present disclosure, there is an effect that a passenger flow prediction device capable of accurately predicting the passenger flow indicating the movement status of passengers who are customers can be obtained.

Brief Description of the Drawings

[0013]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Modes for Carrying Out the Invention

[0014] Hereinafter, a passenger flow prediction device, a passenger flow prediction system, a passenger flow prediction method, and a passenger flow prediction program according to an embodiment will be described in detail with reference to the drawings.

[0015] Embodiment 1. FIG. 1 is a diagram showing the configuration of a passenger flow prediction system according to Embodiment 1. The passenger flow prediction system 100 according to Embodiment 1 includes a passenger flow prediction device 10 that predicts the passenger flow of railway users using railways, an operation management device 20 operated by each of a plurality of railway operators, and an external server 30.

[0016] The passenger flow prediction device 10 includes a data collection unit 11 that collects data from the operation management device 20 and the external server 30, an accumulation data storage unit 12 that stores the data collected by the data collection unit 11 as accumulation data in a database, a prediction control unit 13 that predicts the passenger flow and creates prediction result data based on the prediction of the passenger flow, and a data transmission unit 15 that transmits the prediction result data created by the prediction control unit 13 to the operation management device 20. Generally, the passenger flow prediction device 10 is realized by a cloud server connected to the Internet, but in the case where the number of railway operators is small, an on-premises server may be used.

[0017] The data collection unit 11 collects operation-related data, which is data related to the operation of the railways of each railway operator, from the operation management devices 20 operated by each of a plurality of railway operators that are related to each other. Here, the plurality of railway operators that are related to each other are railway operators where the inflow and outflow of railway users due to transfers occur, or railway operators that conduct mutual access.

[0018] The operation-related data is schedule diagram data, boarding rate data, number of boarding and alighting passengers data, origin and destination data, crew vehicle arrangement information, accident information, and substitute transportation information.

[0019] The planned schedule data is data indicating which vehicles will operate as which types of premium trains at what times for railway operators. The ridership data is data indicating the ratio of the number of passengers actually on board to the seating capacity of the trains operated. The ridership can be calculated based on the load applied to the air springs installed on the vehicle, or based on the images from the cameras installed inside the vehicle. The passenger flow data includes the number of passengers boarding and alighting at each station of the trains operated, and the number of people entering and leaving each station. The origin-destination data is data indicating the entry station and exit station of railway users, and is also referred to as OD (Origin Destination) data. The staff vehicle allocation information is information indicating the allocation of staff and vehicles of each railway operator. The information indicating the allocation of staff includes crew allocation information indicating the allocation of crew members on board the trains, station staff allocation information indicating the allocation of station staff involved in station operations, and allocation information indicating the allocation of maintenance staff involved in track maintenance work. The accident information is information indicating where and what kind of accidents occurred on the line. The transfer transportation information is information indicating in which section of the line transfer transportation was carried out.

[0020] In addition, the data collection unit 11 collects operation impact cause data, which is data on events that cause an impact on railway operations, from the external server 30. The operation impact cause data is weather data and event information.

[0021] The weather data is data indicating actual weather changes such as weather, temperature, wind speed, etc. The event information is data indicating events related to the railway business. Events related to the railway business are, for example, events held at stations or facilities around stations. The event information is information including the type of event and the number of people gathered, such as a concert held in a venue with a seating capacity of 10,000 people. Examples of events held at stations include the opening ceremony of a new line and the retirement ceremony of old vehicles. Examples of events held at facilities around stations include sports games and concerts.

[0022] The accumulated data storage unit 12 stores the operation-related data and operation influence cause data collected by the data collection unit 11 as accumulated data.

[0023] By combining the planned train schedule data of each of a plurality of railway operators with at least one of the origin-destination data of railway users, the boarding rate data of the trains operated by each of the plurality of railway operators, and the number of boarding and alighting passengers data of each of the plurality of railway operators, it is possible to estimate from which station a railway user boarded, which train they boarded, which route they took, and at which station they alighted. That is, by combining the planned train schedule data of each of the plurality of railway operators with at least one of the origin-destination data of railway users, the boarding rate data of the trains operated by each of the plurality of railway operators, and the number of boarding and alighting passengers data of each of the plurality of railway operators, count data, residence data, and movement trajectory data can be created. Count data is data indicating the number of people passing through a certain point or the number of people within a specific range. Residence data is data indicating the number of people staying in a certain point or space for a certain period of time. Movement trajectory data is data indicating the movement trajectory of people. The four types of data, count data, residence data, origin-destination data, and movement trajectory data, are data indicating the movement status of people and are collectively referred to as people flow data. Therefore, by combining the planned train schedule data of each of the plurality of railway operators with at least one of the origin-destination data of railway users, the boarding rate data of the trains operated by each of the plurality of railway operators, and the number of boarding and alighting passengers data of each of the plurality of railway operators, it is possible to create passenger flow data, which is people flow data regarding passengers who are customers of the railway operator. Hereinafter, the number of railway users passing through a certain point within the area under the jurisdiction of the railway operator or the number of railway users within a specific range within the area under the jurisdiction of the railway operator, the number of railway users staying in a certain point or space within the area under the jurisdiction of the railway operator for a certain period of time, the entry and exit stations, and the movement trajectory of the railway users are collectively referred to as "passenger flow", and the data indicating the passenger flow is referred to as "passenger flow data".

[0024] Based on the operation-related data collected from the operation management devices 20 of each of the plurality of railway operators, or based on the operation-related data collected from the operation management devices 20 of each of the plurality of railway operators and the operation impact cause data collected from the external server 30, the prediction control unit 13 predicts the passenger flow.

[0025] FIG. 2 is a diagram showing the configuration of the prediction control unit of the passenger flow prediction device according to Embodiment 1. The prediction control unit 13 includes a learning device 131, a learned model storage unit 132 that stores the learned data learned by the learning device 131, an inference device 133, and a countermeasure plan creation unit 134.

[0026] The learning device 131 learns the passenger flow using, as learning data, the planned schedule data among the accumulated data, the passenger flow data generated from at least one of the origin-destination data, the boarding rate data, and the number of boarding and alighting passengers, and at least one of the weather data, event information, accident information, crew vehicle arrangement information, and substitute transportation information. That is, when the learning device 131 acquires at least one of the weather information and the event information as learning data, it performs learning based on the operation-related data and the operation impact cause data, and when it does not acquire either the weather information or the event information as learning data, it performs learning based on the operation-related data.

[0027] In the following description, the learning process will be described by taking as an example the case where the passenger flow is learned using, as learning data, the planned schedule data, the passenger flow data generated from at least one of the origin-destination data, the boarding rate data, and the number of boarding and alighting passengers, and at least one of the weather data, event information, accident information, crew vehicle arrangement information, and substitute transportation information. However, the same learning process is performed whether all of the exemplified data is used for learning or only a part of the exemplified data is used for learning.

[0028] FIG. 3 is a diagram showing the configuration of the learning device of the passenger flow prediction device according to Embodiment 1. The learning device 131 includes a learning data acquisition unit 1311 and a model generation unit 1312.

[0029] The learning data acquisition unit 1311 acquires, as learning data, the planned train schedule data, at least one of the origin-destination data, boarding rate data, and number of boarding and alighting passengers data, and at least one of the weather data, event information, accident information, staff vehicle allocation information, and substitute transportation information from the storage data storage unit 12. That is, the learning data acquisition unit 1311 acquires the storage data stored in the storage data storage unit 12 as learning data. As described above, since passenger flow data can be created by combining the planned train schedule data and at least one of the number of boarding and alighting passengers data, boarding rate data, and origin-destination data, the learning data acquisition unit 1311 can be regarded as acquiring, as learning data, the passenger flow data and at least one of the weather data, event information, accident information, staff vehicle allocation information, and substitute transportation information.

[0030] The model generation unit 1312 learns the passenger flow based on the learning data created based on the combination of the planned train schedule data output from the learning data acquisition unit 1311, the passenger flow data generated from at least one of the origin-destination data, boarding rate data, and number of boarding and alighting passengers data, and at least one of the weather data, event information, accident information, staff vehicle allocation information, and substitute transportation information. That is, the model generation unit 1312 generates a learned model for inferring the passenger flow from the planned train schedule data of multiple railway operators, the passenger flow data generated from at least one of the origin-destination data, boarding rate data, and number of boarding and alighting passengers data, and at least one of the weather data, event information, accident information, staff vehicle allocation information, and substitute transportation information.

[0031] As the learning algorithm used by the model generation unit 1312, known algorithms such as supervised learning, unsupervised learning, and reinforcement learning can be used. As an example, the case of applying a neural network will be described.

[0032] The model generation unit 1312 learns the passenger flow, for example, by so-called supervised learning according to a neural network model. Here, supervised learning refers to a method of learning the features in the learning data by giving the learning device 131 a set of data of an input and a label which is the result, and inferring the result from the input.

[0033] A neural network is composed of an input layer consisting of a plurality of neurons, an intermediate layer consisting of a plurality of neurons, and an output layer consisting of a plurality of neurons. The intermediate layer is also referred to as a hidden layer. The intermediate layer may be one layer or two or more layers.

[0034] FIG. 4 is a diagram showing an example of a neural network of the learning device of the passenger flow prediction device in Embodiment 1. The neural network 70 shown in FIG. 4 has a three-layer structure composed of an input layer 71 consisting of neurons X1, X2, X3, an intermediate layer 72 consisting of neurons Y1, Y2, and an output layer 73 consisting of neurons Z1, Z2, Z3. When a plurality of inputs are input to the input layer 71 of the neural network 70, their values are multiplied by weights w11, w12, w13, w14, w15, w16 and input to the intermediate layer 72, and the result is further multiplied by weights w21, w22, w23, w24, w25, w26 and output from the output layer 73. This output result varies depending on the values of the weights w11, w12, w13, w14, w15, w16 and the values of the weights w21, w22, w23, w24, w25, w26.

[0035] In this embodiment, the neural network 70 learns the passenger flow by so-called supervised learning according to the learning data created based on a combination of the train schedule data acquired by the learning data acquisition unit 1311, at least one of the origin-destination data, the boarding rate data, and the number of boarding and alighting passengers data, and at least one of the weather data, event information, accident information, staff vehicle allocation information, and substitute transportation information. That is, the model generation unit 1312 learns the passenger flow by performing learning with the passenger flow data generated from the train schedule data output from the learning data acquisition unit 1311 and at least one of the origin-destination data, the boarding rate data, and the number of boarding and alighting passengers data as "labels" and the weather data, event information, accident information, staff vehicle allocation information, and substitute transportation information as "inputs".

[0036] The neural network 70 inputs at least one of the weather data, event information, accident information, staff vehicle allocation information, and substitute transportation information into the input layer 71, and adjusts the weights w11, w12, w13, w14, w15, w16 and the weights w21, w22, w23, w24, w25, w26 so that the result output from the output layer 73 approaches the passenger flow data generated from the train schedule data and at least one of the origin-destination data, the boarding rate data, and the number of boarding and alighting passengers data, thereby performing learning.

[0037] The model generation unit 1312 generates and outputs a learned model by executing the learning as described above.

[0038] The learned model storage unit 132 stores the learned model output from the model generation unit 1312.

[0039] FIG. 5 is a flowchart relating to the learning process in the learning device of the passenger flow prediction device according to Embodiment 1. The process of learning the passenger flow by the learning device 131 will be described with reference to FIG. 5.

[0040] In step S11, the learning data acquisition unit 1311 acquires, as learning data, the planned train schedule data, passenger flow data generated from at least one of the origin-destination data, boarding rate data, and number of boarding / alighting passengers data, and at least one of the weather data, event information, accident information, staff vehicle allocation information, and substitute transportation information. Although it was assumed that the planned train schedule data, passenger flow data generated from at least one of the origin-destination data, boarding rate data, and number of boarding / alighting passengers data, and at least one of the weather data, event information, accident information, staff vehicle allocation information, and substitute transportation information are acquired simultaneously, it is sufficient that the planned train schedule data, passenger flow data generated from at least one of the origin-destination data, boarding rate data, and number of boarding / alighting passengers data, and at least one of the weather data, event information, accident information, staff vehicle allocation information, and substitute transportation information can be input in association with each other. The planned train schedule data, at least one of the origin-destination data, boarding rate data, and number of boarding / alighting passengers data, and at least one of the weather data, event information, accident information, staff vehicle allocation information, and substitute transportation information may be acquired at different timings respectively.

[0041] In step S12, the model generation unit 1312 learns the passenger flow by so-called supervised learning according to the learning data created based on the combination of the planned train schedule data acquired by the learning data acquisition unit 1311, at least one of the origin-destination data, boarding rate data, and number of boarding / alighting passengers data, and at least one of the weather data, event information, accident information, staff vehicle allocation information, and substitute transportation information, and generates a learned model.

[0042] In step S13, the learned model storage unit 132 stores the learned model generated by the model generation unit 1312.

[0043] FIG. 6 is a diagram showing the configuration of the inference device of the passenger flow prediction device according to Embodiment 1. The inference device 133 includes an inference data acquisition unit 1331 and an inference unit 1332.

[0044] The inference data acquisition unit 1331 acquires, as inference data, at least one of the operation-related data newly collected by the data collection unit 11 and the operation influence cause data, namely, weather data, event information, accident information, staff vehicle arrangement information, and transfer transportation information.

[0045] The inference unit 1332 infers the passenger flow obtained by using the learned model. That is, by inputting at least one of the weather data, event information, accident information, staff vehicle arrangement information, and transfer transportation information acquired by the inference data acquisition unit 1331 into this learned model, the passenger flow inferred from at least one of the weather data, event information, accident information, staff vehicle arrangement information, and transfer transportation information can be output.

[0046] In addition, in this embodiment, although it has been described that the passenger flow is output using the learned model learned by the model generation unit 1312 of the passenger flow prediction device 10, a learned model may be acquired from outside the passenger flow prediction device 10, and the passenger flow may be output based on this learned model.

[0047] FIG. 7 is a flowchart showing the flow of the inference process in the inference device of the passenger flow prediction device according to Embodiment 1. Using FIG. 7, the process for obtaining the passenger flow using the inference device 133 will be described. The process shown in FIG. 7 is performed in response to a request for prediction data from the operation management device 20 owned by the railway operator.

[0048] In step S21, the inference data acquisition unit 1331 acquires, as inference data, at least one of the weather data, event information, accident information, staff vehicle arrangement information, and transfer transportation information.

[0049] In step S22, the inference unit 1332 inputs at least one of the weather data, event information, accident information, staff vehicle arrangement information, and transfer transportation information into the learned model stored in the learned model storage unit 132, and infers the passenger flow.

[0050] In step S23, the inference unit 1332 outputs passenger flow data indicating the passenger flow obtained by the learned model to the countermeasure plan creation unit 134.

[0051] Based on the passenger flow data output from the inference device 133, the countermeasure plan creation unit 134 creates a countermeasure plan for the predicted passenger flow. The countermeasure plan created by the countermeasure plan creation unit 134 is, for example, a railway operation plan, creating a plan for the operation schedule based on the predicted passenger flow or creating a plan for operation arrangement based on the predicted passenger flow. At this time, the operation schedule refers to the operation plan of train crew, station staff, maintenance staff, or vehicles from the day after tomorrow. Also, operation arrangement refers to the operation plan of train crew, station staff, maintenance staff, or vehicles on the current day. The railway operation plan or passenger flow data created by the countermeasure plan creation unit 134 is output to the data transmission unit 15 as prediction result data.

[0052] The data transmission unit 15 transmits the prediction result data to the operation management device 20, which is the requester of the prediction data.

[0053] Regardless of the request from the operation management device 20, the passenger flow prediction device 10 may periodically create a countermeasure plan and transmit it to the operation management device 20 together with the passenger flow data.

[0054] By receiving the prediction result data, the railway operator can operate the trains based on the railway operation plan created by the countermeasure plan creation unit 134, operate the trains so that the congestion level is such that passengers do not get dissatisfied and the balance is in the black, or optimize the arrangement of staff and vehicles.

[0055] For example, when predicting the passenger flow based on the operation-related data collected from a single railway operator when a large-scale event is held, at least some of the line sections compete with other railway operators. These other operators may increase the number of train operations, resulting in the actual number of passengers on one's own trains being less than expected. In such cases, if the number of train operations is increased based on the predicted passenger flow, the number of train operations will become excessive. In contrast, the passenger flow prediction system 100 according to Embodiment 1 predicts the passenger flow based on the operation-related data collected from multiple railway operators, so it can predict the passenger flow with high accuracy. Furthermore, since the passenger flow prediction system 100 according to Embodiment 1 transmits a countermeasure plan based on the highly accurate passenger flow predicted based on the operation-related data collected from multiple railway operators to the operation management device 20, the railway operator can operate the trains with a number of operations that has a small excess or deficiency compared to the actual number of passengers.

[0056] In the above example, the case where the railway operation plan based on the predicted passenger flow is transmitted from the data transmission unit 15 to the operation management device 20 as prediction data has been described. However, the passenger flow data indicating the predicted passenger flow may be transmitted from the data transmission unit 15 to the operation management device 20 as prediction data. In this case, the passenger flow prediction device 10 can have a configuration in which the countermeasure plan creation unit 134 is omitted.

[0057] In this embodiment, the case where supervised learning is applied to the learning algorithm used by the model generation unit 1312 has been described, but it is not limited to this. Regarding the learning algorithm, in addition to supervised learning, semi-supervised learning or the like can also be applied.

[0058] It is also possible to add or remove the railway operator from which the learning device 131 collects learning data during the process. Furthermore, the learning device 131 that has learned the passenger flow for a certain railway operator may be applied to another passenger flow prediction device 10, and the passenger flow for the other passenger flow prediction device 10 may be relearned and updated.

[0059] In addition, as the learning algorithm used in the model generation unit 1312, deep learning that learns the extraction of the feature quantities themselves, such as passenger flow data, weather data, event information, accident information, staff vehicle arrangement information, and transfer transportation information, can also be used. Machine learning may also be executed according to other known methods, such as genetic programming, functional logic programming, support vector machines, and the like.

[0060] Note that the learning device 131 and the inference device 133 are used to learn the passenger flow of each of a plurality of railway operators. For example, they may be separate devices from the passenger flow prediction device 10 connected to the passenger flow prediction device 10 via a network. Further, the learning device 131 and the inference device 133 may exist on a cloud server different from the passenger flow prediction device 10.

[0061] The hardware configuration of the prediction control unit 13 of the passenger flow prediction device 10 according to Embodiment 1 will be described.

[0062] FIG. 8 is a diagram showing the hardware configuration of the prediction control unit of the passenger flow prediction device according to Embodiment 1. The prediction control unit 13 is realized by a computer system including a processor 91 that executes various processes, a memory 92 that is a main memory, and a storage device 93 that stores information.

[0063] The processor 91 may be an arithmetic means such as an arithmetic unit, a microprocessor, a microcomputer, a CPU (Central Processing Unit), or a DSP (Digital Signal Processor). Also, as the memory 92, a non-volatile or volatile semiconductor memory such as a RAM (Random Access Memory), a ROM (Read Only Memory), a flash memory, an EPROM (Erasable Programmable Read Only Memory), or an EEPROM (registered trademark) (Electrically Erasable Programmable Read Only Memory) can be used. The storage device 93 stores a program for executing a process of learning the passenger flow, a process of inferring the passenger flow, and a process of creating a countermeasure plan.

[0064] In the above computer system, the processor 91 reads out a program corresponding to the processing of each component stored in the storage device 93 into the memory 92 and executes it, thereby realizing the function of the prediction control unit 13. Also, the memory 92 is also used as a temporary memory in each process executed by the processor 91. The program executed by the processor 91 may be provided in a state stored in a storage medium or may be provided via a network.

[0065] The configurations shown in the above embodiments are examples of the content, and it is possible to combine them with other known technologies, and it is also possible to omit or change a part of the configuration without departing from the gist.

[0066] Hereinafter, various aspects of the present disclosure will be summarized and described as appendices.

[0067] (Appendix 1) A data collection unit that collects operation-related data regarding railway operations from a plurality of railway operators, An accumulated data storage unit that accumulates the operation-related data collected by the data collection unit, A prediction control unit that predicts the passenger flow of railway users who use the railway based on the operation-related data collected from the plurality of railway operators. A passenger flow prediction device, characterized by comprising the above. (Appendix 2) The operation-related data is at least one of planned schedule data, origin-destination data, boarding rate data, and number of boarding and alighting passengers data, and at least one of accident information, staff vehicle allocation information, and substitute transportation information. The passenger flow prediction device according to Appendix 1, characterized in that. (Appendix 3) The prediction control unit Using a learned model that performs learning based on the operation-related data, predicts the passenger flow. The passenger flow prediction device according to Appendix 1 or Appendix 2, characterized in that. (Appendix 4) The data collection unit further collects operation influence cause data that causes an impact on the operation of the railway. The accumulated data storage unit further accumulates the operation influence cause data collected by the data collection unit. The prediction control unit predicts the passenger flow of railway users who use the railway based on the operation-related data and the operation influence cause data collected from the plurality of railway operators. The passenger flow prediction device according to Appendix 1, characterized in that. (Appendix 5) The operation-related data is at least one of planned schedule data, origin-destination data, boarding rate data, and number of boarding and alighting passengers data. The operation influence cause data is at least one of weather data and event information. The passenger flow prediction device according to Appendix 4, characterized in that. (Appendix 6) The prediction control unit Using a learned model that performs learning based on the operation-related data and the operation influence cause data, predicts the passenger flow. The passenger flow prediction device according to Appendix 4 or Appendix 5, characterized in that. (Appendix 7) The passenger flow prediction device according to any one of Supplementary Notes 1 to 6, wherein the data collection unit collects the operation-related data from the plurality of railway operators whose at least a part of the line sections conflict with each other. (Supplementary Note 8) The passenger flow prediction device according to any one of Supplementary Notes 1 to 7, further comprising a data transmission unit that transmits prediction result data indicating the passenger flow predicted by the prediction control unit. (Supplementary Note 9) The passenger flow prediction device according to Supplementary Note 8, wherein the prediction result data includes passenger flow data indicating the passenger flow predicted by the prediction control unit. (Supplementary Note 10) The passenger flow prediction device according to Supplementary Note 9, wherein the prediction result data includes a countermeasure plan based on the passenger flow data indicating the passenger flow predicted by the prediction control unit.

Explanation of Signs

[0068] 10 Passenger flow prediction device, 11 Data collection unit, 12 Accumulated data storage unit, 13 Prediction control unit, 15 Data transmission unit, 20 Operation management device, 30 External server, 70 Neural network, 71 Input layer, 72 Intermediate layer, 73 Output layer, 91 Processor, 92 Memory, 93 Storage device, 100 Passenger flow prediction system, 131 Learning device, 132 Trained model storage unit, 133 Inference device, 134 Countermeasure plan creation unit, 1311 Learning data acquisition unit, 1312 Model generation unit, 1331 Inference data acquisition unit, 1332 Inference unit.

Claims

1. A data collection unit that collects operation-related data regarding railway operations from a plurality of railway operators, An accumulated data storage unit that stores the operation-related data collected by the data collection unit, A prediction control unit that predicts the passenger flow of railway users who use the railway based on the operation-related data collected from the plurality of railway operators. A passenger flow prediction device, characterized in that it comprises:

2. The operation-related data is at least one of schedule diagram data, origin-destination data, boarding rate data, and number of boarding / alighting passengers data, and at least one of accident information, crew vehicle allocation information, and substitute transport information. The passenger flow prediction device according to claim 1, characterized in that it is:

3. The prediction control unit: The passenger flow prediction device according to claim 1, characterized in that it predicts the passenger flow using a learned model that performs learning based on the operation-related data.

4. The data collection unit further collects operation influence cause data that is a cause affecting the operation of the railway, The accumulated data storage unit further stores the operation influence cause data collected by the data collection unit, The prediction control unit predicts the passenger flow of railway users who use the railway based on the operation-related data and the operation influence cause data collected from the plurality of railway operators. The passenger flow prediction device according to claim 1, characterized in that it is:

5. The operation-related data is at least one of schedule diagram data, origin-destination data, boarding rate data, and number of boarding / alighting passengers data, The operation influence cause data is at least one of weather data and event information. The passenger flow prediction device according to claim 4, characterized in that it is:

6. The prediction control unit: The passenger flow prediction device according to claim 4, characterized in that it predicts the passenger flow using a learned model that performs learning based on the operation-related data and the operation influence cause data.

7. The data collection unit collects the operation-related data from the plurality of railway operators where at least some line sections compete. The passenger flow prediction device according to claim 1, characterized in that it is:

8. The passenger flow prediction device according to any one of claims 1 to 7, characterized in that it comprises a data transmission unit that transmits prediction result data indicating the passenger flow predicted by the prediction control unit.

9. The passenger flow prediction device according to claim 8, wherein the prediction result data includes passenger flow data indicating the passenger flow predicted by the prediction control unit.

10. The passenger flow prediction device according to claim 9, wherein the prediction result data includes a countermeasure plan based on the passenger flow data indicating the passenger flow predicted by the prediction control unit.

11. A passenger flow prediction system having an operation management device owned by each of a plurality of railway operators and a passenger flow prediction device that predicts the passenger flow of railway users who use the railway, wherein the passenger flow prediction device includes: a data collection unit that collects operation-related data regarding the operation of the railway from the plurality of railway operators; an accumulated data storage unit that stores the operation-related data collected by the data collection unit; and a prediction control unit that predicts the passenger flow of railway users who use the railway based on the operation-related data collected from the plurality of railway operators.

12. A step of collecting operation-related data regarding the operation of the railway from a plurality of railway operators; A step of storing the collected operation-related data; A passenger flow prediction method comprising a step of predicting the passenger flow of railway users who use the railway based on the operation-related data collected from the plurality of railway operators.

13. A step of collecting operation-related data regarding the operation of the railway from a plurality of railway operators; A step of storing the collected operation-related data; A step of predicting the passenger flow of railway users who use the railway based on the operation-related data collected from the plurality of railway operators; A passenger flow prediction program characterized by causing a computer to execute the steps.

Images