Vehicle dispatch management system

Abstract

By efficiently deploying vehicles while considering the variance of prediction error values, we can improve the vehicle utilization rate by balancing the supply and demand of vehicles within the service area. [Solution] The dispatch management system 1 comprises a vehicle V, a dispatch management device 10 for managing the placement of the vehicle V, and a data management device 2. The data management device 2 stores the demand measurement value of the vehicle V measured at each placement point a, the demand forecast value indicating the future demand for the vehicle V at each placement point a, and the forecast error value. The dispatch management device 10 includes controllers 34 and 44 that perform the following processes for each placement point a: calculate the error variance showing the variance between the demand forecast value and the forecast error value; extract placement points a whose error variance satisfies predetermined conditions as candidates for the placement of a vehicle V; and prioritize high-demand placement points among the extracted candidate placement points a where the demand for the vehicle V is determined to be relatively high based on the demand forecast value, and determine them as the placement destination for a vehicle V.

Description

Technical Field

[0001] The present disclosure relates to a vehicle dispatching management system for vehicles (service vehicles) used for user mobility services (mobility services).

Background Art

[0002] Mobility services are positioned as services that smoothly provide user mobility and luggage transportation by vehicles. In mobility services, in order to improve the utilization rate of vehicles while improving the convenience of users (passengers), it is required to predict the demand for vehicles for each location (placement location) where the vehicles are placed and create an efficient placement plan.

[0003] For example, Patent Document 1 below discloses a method for updating a predicted value based on measured values in various past periods to be predicted. Specifically, a technique for improving prediction accuracy is disclosed by calculating the occurrence probability of a prediction error value in a prediction period based on data indicating the time transition of the error (prediction error value) between the measured value and the predicted value in the past period, and correcting the predicted value with the predicted error value estimated from the calculated occurrence probability.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] The technology disclosed in Patent Document 1 above proposes correcting predicted values ​​using prediction error values. However, since this technology does not take into account the variance (dispersion) of prediction error values, even if the accuracy of the average value of the predicted values ​​improves, there is a risk that the actual values ​​(measured values) and the predicted values ​​will deviate from each other at the measurement level. For example, if the variance of the prediction error values ​​is large, the accuracy of the predicted values ​​corrected with the prediction error values ​​will decrease. Therefore, if a vehicle deployment plan is created without considering the variance of prediction error values, an oversupply or undersupply of vehicles at deployment locations can easily occur, worsening the balance of supply and demand for vehicles and potentially reducing the utilization rate of vehicles.

[0006] Therefore, this disclosure aims to provide a technology that can efficiently allocate vehicles while taking into account the variance of prediction error values, thereby improving the balance of supply and demand for vehicles within the service area of ​​mobility services and increasing the utilization rate of vehicles. [Means for solving the problem]

[0007] A dispatch management system according to one aspect of the present disclosure comprises a vehicle for providing mobility services, a dispatch management device for managing the placement of the vehicle at a designated location within a predetermined area, and a storage device, wherein the storage device stores a demand measurement value indicating the demand for the vehicle measured at each location within the area, a demand forecast value indicating the future demand for the vehicle at each location, and a forecast error value indicating the error between the demand measurement value and the demand forecast value at each location; the dispatch management device includes a controller that performs the following for each location within the area: a process for calculating the demand forecast value based on the demand measurement value and an error variance indicating the variance of the forecast error value over a predetermined past period; a process for extracting locations where the error variance satisfies predetermined conditions as candidates for the placement of one of the vehicle; and a process for prioritizing high-demand locations among the extracted candidate locations where the demand for the vehicle is determined to be relatively high based on the demand forecast value, and determining them as the placement locations for one of the vehicle. [Effects of the Invention]

[0008] According to this disclosure, it becomes possible to efficiently allocate vehicles while taking into account the variance of prediction error values, thereby improving the balance of supply and demand for vehicles within the area where mobility services are provided and increasing the utilization rate of vehicles. [Brief explanation of the drawing]

[0009] [Figure 1A] This figure illustrates an example of a dispatch management system according to the first embodiment of the present disclosure. [Figure 1B] This is a schematic diagram illustrating the locations of vehicles used for the ride-hailing service. [Figure 2] Figure 1 is a block diagram showing an example of the hardware and functional configuration of the data management device in the dispatch management system. [Figure 3A] This diagram illustrates an example of location classification in sample data. [Figure 3B] This figure illustrates an example of a demand factor in sample data. [Figure 4A] Figure 1 is a block diagram showing an example of the hardware and functional configuration of a demand forecasting device in the dispatch management system. [Figure 4B] This is a conceptual diagram illustrating an example of error time series data used in demand forecasting. [Figure 5] Figure 1 is a block diagram showing an example of the hardware and functional configuration of the dispatch management device in the dispatch management system. [Figure 6] This flowchart shows an example of a dispatch management method using the dispatch management system of the first embodiment of this disclosure. [Figure 7] This flowchart shows an example of a vehicle dispatch management method according to a first modification of the first embodiment of the present disclosure. [Figure 8] This flowchart shows an example of a vehicle dispatch management method relating to a second modification of the first embodiment of the present disclosure. [Figure 9] This flowchart shows an example of a dispatch management method using the dispatch management system of the second embodiment of this disclosure. [Figure 10] This is a conceptual diagram illustrating the grouping of vehicle placement locations.

Mode for Carrying Out the Invention

[0010] Hereinafter, embodiments of the present invention will be described with reference to the drawings. Note that each drawing is schematic and may differ from the actual one. Further, the embodiments of the present invention shown below exemplify devices and methods for embodying the technical idea of the present invention, and the technical idea of the present invention does not specify the structure, arrangement, etc. of the components as follows. The technical idea of the present invention can be variously modified within the technical scope defined by the claims described in the claims.

[0011] 1. First Embodiment (System Configuration) FIG. 1A is a diagram for explaining an example of a vehicle allocation management system according to the first embodiment (hereinafter referred to as this embodiment) of the present disclosure. FIG. 1B is a schematic diagram for explaining a service area which is an area where a vehicle allocation service is provided. The vehicle allocation management system 1 is a system that provides a vehicle allocation service for allocating a vehicle V used for the movement of a user. The vehicle allocation management system 1 includes a data management device (an example of a storage device) 2, a demand prediction device 3, a vehicle allocation planning device 4, and a vehicle V, and these are connected to each other so as to be communicable via a communication network 9. The demand prediction device 3 and the vehicle allocation planning device 4 constitute a vehicle allocation management device 10 that manages the allocation of the vehicle V. The vehicle allocation management system 1 also includes a factor data collection device 5 and a demand measurement device 6, and these are connected to the data management device 2 so as to be communicable with each other via the communication network 9.

[0012] The service area A is a regional range including one or more placement points where the vehicle V used by the user is placed. For example, the service area A is a regional range including a plurality of placement points a, and can be recognized with places (POI (Point Of Interest)) with high user usage frequency and interest such as airports, railway stations, and large-scale customer gathering facilities as the core. As will be described in detail later, the vehicle allocation management device 10 (demand prediction device 3, vehicle allocation planning device 4) manages the allocation of the vehicle V to the allocation point a within the service area A (an example of a predetermined area). For example, when there is a vehicle V in the empty state (the state where the user has already disembarked and there is no vehicle reservation) at the position P within the service area A, the vehicle allocation management device 10 allocates the vehicle V from the position P to the allocation point a within a predetermined range based on a predetermined condition. In the present disclosure, an example in which the vehicle allocation management system 1 is adapted to a mobility service for moving passengers and transporting luggage by the vehicle V will be described, but the service application is not limited to this. The vehicle V may be at least a vehicle used for providing a mobility service. Further, the allocation point may be a waiting place where the vehicle V waits before allocation to a predetermined boarding place where the user boards or after the user disembarks at a predetermined disembarking place, or may be a boarding and alighting place where the vehicle V can stop at multiple locations.

[0013] The data management device 2 is a computing device that manages data used for vehicle allocation management in the vehicle allocation management device 10. The data management device 2 functions as a storage device that comprehensively stores and holds data used for vehicle allocation management in the vehicle allocation management system 1. As will be described in detail later, in this example, the data management device 2 stores sample data used for demand prediction of the vehicle V in the service area A, vehicle data and allocation point data used for creating an allocation plan of the vehicle V to each allocation point a. The sample data is used to calculate a demand prediction value that predicts the future demand for the vehicle V in a predetermined time period at each allocation point a in the service area A. Further, the vehicle data is data related to the vehicle V used for the movement of the user, and the allocation point data is data related to each allocation point a within the service area A. The sample data includes a measured value of the past demand (demand measurement value) of the vehicle V in one service area (service area A) and one or more types of demand factor data that affect the demand. That is, in the present embodiment, the data management device 2 stores a demand measurement value indicating the demand for the vehicle V measured at each allocation point where the vehicle V is allocated within one service area.

[0014] The demand for vehicle V may be the number of requests for vehicle dispatch from users who wish to board near each location (number of requests), or the number of vehicles dispatched to users at each location (number of vehicles dispatched). Furthermore, demand factor data includes data relating to the characteristics of the installation locations (location factor data) and data relating to the characteristics of each period division at the installation locations (period division factor data). Here, a period division refers to a division of time and date based on predetermined seasonal conditions (season, day of the week, time of day, etc.). Sample data is transmitted from other devices to the data management device 2 via the communication network 9. For example, demand factor data from the sample data is transmitted from the factor data collection device 5, and historical demand measurements for vehicle V in service area A are transmitted from the demand measurement device 6. Furthermore, the data management device 2 searches for data and transmits the search results to other devices in response to data acquisition requests from the demand forecasting device 3.

[0015] The demand forecasting device 3 uses sample data stored in the data management device 2 to forecast the future demand for vehicles V at each location a in a service area A, which is a single service area, for a predetermined time period. Specifically, the demand forecasting device 3 calculates a forecast value, which is a predicted value of the future demand for vehicles V for the predetermined time period, based on the demand measurement values ​​in the sample data. When performing a demand forecast, the demand forecasting device 3 sends a request to the data management device 2 via the communication network 9 to acquire sample data and extract the sample data (e.g., demand measurement values) to be used for the demand forecast. The demand forecasting device 3 can also correct the calculated forecast value based on the error (forecast error value) between past forecast values ​​and actual measured values. This improves the accuracy of the demand forecast. The demand forecasting device 3 also calculates the variance (error variance) of the forecast error value at each location a. Details of the calculation, correction, and calculation of the error variance of the forecast value will be described later. Furthermore, the demand forecasting device 3 transmits the demand forecast value and error variance as demand forecasting results to the dispatch planning device 4 via the communication network 9.

[0016] The dispatch planning device 4 creates a dispatch plan for operating vehicle V from a designated pick-up location to a designated drop-off location within a service area (in this example, service area A) based on a dispatch request from a user (received dispatch request information). The dispatch planning device 4 also creates a vehicle placement plan for each placement point a within the service area (service area A) based on the demand forecasting results (demand forecast value and error variance) transmitted from the demand forecasting device 3. This allows for efficient placement of vehicle V at each placement point in the service area, taking into account the variance of the forecast error value, thereby improving the supply-demand balance of vehicle V within the service area and increasing the utilization rate. The placement plan may include vehicle data for vehicle V and placement point data indicating the placement point a. The dispatch planning device 4 instructs vehicle V to move to each placement point based on the placement plan via the communication network 9. The dispatch planning device 4 transmits dispatch request information to the demand measurement device 6. The dispatch planning device 4 is also capable of sending and receiving data with the vehicle V, and may receive updated data from the vehicle V and transmit the received updated data to the data management device 2 via the communication network 9. The updated data may include the current location and current status of the vehicle V. The updated data may also be a dispatch plan or deployment plan created by the dispatch planning device 4.

[0017] The factor data collection device 5 periodically measures the above-mentioned demand factor data (location factor data, period classification factor data) at predetermined time intervals and transmits it to the data management device 2 via the communication network 9. The demand measurement device 6 periodically measures the demand for vehicles V in a service area (service area A) at predetermined unit time intervals (e.g., 15 minutes) and transmits demand measurement data indicating the measured demand value to the data management device 2 via the communication network 9. Specifically, the demand measurement device 6 measures the number of requests per unit time at each location in a service area based on the dispatch request information transmitted from the dispatch planning device 4. As will be explained in more detail later, the dispatch request information includes at least the boarding location where the user boards the vehicle V and the date and time (reception date and time) when the dispatch planning device 4 received the dispatch request from the user. The demand measurement device 6 identifies the boarding location based on the dispatch request information and identifies the unit time in which the dispatch request was made from the reception date and time. As a result, the demand measurement device 6 measures the number of requests per unit time (number of dispatch requests) for each boarding location as a demand measurement value. In this embodiment, each deployment point is associated with a predetermined area of ​​responsibility, and the demand measurement value of the deployment point corresponding to the area of ​​responsibility is identified by summing the demand measurement values ​​of each boarding location located within that area of ​​responsibility.

[0018] (Configuration of the data management device) Next, using Figure 2, the data management device 2 in the dispatch management system 1 will be explained. The data management device 2 receives demand factor data, demand measurement data, and other-region data from the factor data collection device 5 and the demand measurement device 6, respectively, and stores them as sample data. In addition to sample data, the data management device 2 also stores vehicle data and location data. Then, in response to a data acquisition request, it transmits this data to the demand forecasting device 3 and the vehicle dispatch planning device 4. The data management device 2 consists of an input device 21, an output device 22, a communication device 23, and a controller 24. The data management device 2 may be, for example, an information processing device such as a personal computer or a server computer.

[0019] The input device 21 is a keyboard or mouse, and the output device 22 is a display or printer. The communication device 33 is equipped with a NIC (Network Interface Card) for connecting to a wireless LAN or a wired LAN. The data management device 2 sends and receives data to and from each device via the communication device 33. The controller 24 is an electronic control unit (ECU) that controls the operation of the data management device 2. The controller 24 comprises a processor 24a and a storage device 24b.

[0020] The processor 24a may be, for example, a CPU (Central Processing Unit) or an MPU (Micro-Processing Unit). The storage device 24b may include memory such as ROM (Read Only Memory) and RAM (Random Access Memory) used as main memory, as well as non-transitory tangible storage media such as memory registers and cache memory. The storage device 24b stores a database which functions as a data storage module 241. The data storage module 241 holds sample data 241a, vehicle data 241b, and location data 241c. The sample data 241a consists of demand factor data (location factor data, period segmentation factor data) transmitted from the factor data collection device 5 and demand measurement data transmitted from the demand measurement device 6. The demand measurement data includes the demand measurement value at each location in service area A and the measurement time of said demand measurement value. The measurement time only needs to include at least the start time or end time of measurement, and may include both. The demand measurement value is measured at regular intervals (e.g., 15 minutes) as described above. For this reason, the demand measurement data, including the demand measurement value, may be stored in the data storage module 241 for each unit time.

[0021] Here, we will explain the details of sample data 241a using Figures 3A and 3B. As described above, sample data 241a includes demand factor data transmitted from factor data collection device 5 and demand measurement data transmitted from demand measurement device 6. For example, sample data 241a may be data in which demand measurement data (demand measurement value, measurement time) is linked to factor classifications (location classification and period classification) for linking with demand factors. The location classification identifies each deployment location within service area A where demand measurement data was collected, and is used to associate location factor data with demand measurement data. As shown in Figure 3A, each deployment location a (for example, locations a1 to a8) is associated with a location ID (A001 to A008), and by assigning a location ID to the transmitted demand measurement data, the deployment location where the demand measurement data was collected can be identified.

[0022] Furthermore, as mentioned above, period divisions are divisions of time and date based on seasonal conditions (season, day of the week, time of day, etc.), and are used to associate period division factor data with the unit time on which demand measurement data was measured. For example, period divisions can be broadly divided into three categories: season, day of the week, and time of day. For example, among the category categories, "season" is divided into four categories: "spring, summer, autumn, winter," "day of the week" is divided into two categories: "weekdays, holidays," and "time of day" is divided into eight categories: "pre-dawn, early morning, morning, before noon, noon, afternoon, evening, night." Each time of day is, for example, divided into 3-hour intervals from 0:00 to the next 0:00 24 hours later. Period divisions may be identified by a combination of category IDs that uniquely represent each item of each category.

[0023] As shown in Figure 3B, among the demand factor data, location factor data may include, for example, population, POI clusters indicating the density of POIs in the vicinity of the location, and data indicating the convenience of public transportation (transportation convenience). These are data collected and aggregated at predetermined intervals for each location a in a service area (service area A). Note that location factor data is not limited to these, and may include various factor data indicating location characteristics that affect demand forecasting, such as the degree of gradient, bridges, wide roads, etc. Location factor data is stored in the data storage module 241, linked to location IDs. Furthermore, demand measurement data can be associated with location factor data using location IDs. This allows the demand forecasting device 3 to identify demand measurement data linked to specific location factor data when forecasting demand for vehicles V.

[0024] Furthermore, among the demand factor data, the period classification factor data may include, for example, climate data such as temperature and weather, data indicating the frequency of public transportation operations, and data indicating the degree of congestion of public transportation. However, the period classification factor data is not limited to these and may include various factor data that change depending on seasonal conditions and affect demand forecasting. The period segmentation factor data is data collected and aggregated for each period segment at each location. Therefore, in the data storage module 241, the period segmentation factor data is stored linked to the location ID and the period segmentation ID. The data storage module 241 stores sample data 241a. In other words, the sample data 241a is not overwritten, and includes past demand factor data and demand measurement data. Note that the structure of the sample data 241a shown in Figures 3A and 3B is just one example, and the structure of the sample data in this disclosure is not limited to this; any structure that allows for the correspondence between demand measurement data and demand factor data (location factor data, period segmentation factor data) is acceptable.

[0025] Returning to Figure 2, the vehicle data 241b is data for vehicle V used for mobility services within service area A. Vehicle data 241b includes, for example, data on vehicle ID, vehicle attributes (vehicle registration number, vehicle type, and maximum passenger capacity / maximum load capacity, manned or unmanned type, etc.), vehicle performance, other specifications, current dispatch plan, current location, and current status (service status, battery remaining capacity (driving range), and number of passengers, etc.). Vehicle data 241b may be used, for example, to create a dispatch plan for vehicle V in the dispatch planning device 4. Furthermore, the deployment location data 241c is data for each deployment location a within service area A. The deployment location data 241c includes data such as the deployment location ID, location, the number of vehicles V (deployed vehicles) deployed, and the maximum number of vehicles deployed. The deployment location data 241c may be used, for example, to create a vehicle deployment plan in the vehicle dispatch planning device 4. Vehicle data 241b and deployment location data 241c may be updated based on update data transmitted from the vehicle dispatch planning device 4.

[0026] The storage device 24b stores a program for implementing the data management module 242. The processor 24a executes this program, thereby realizing the functions of the data management device 2. The data management module 242 adds and extracts sample data 241a to the data storage module 241. Specifically, the data management module 242 adds new sample data 241a to the data storage module 241 based on the reception of demand factor data and demand measurement data, and extracts specific data from the sample data 241a based on a sample data acquisition request and transmits it to the demand forecasting device 3. Furthermore, the data management module 242 extracts specific data from the vehicle data 241b based on a vehicle data acquisition request and transmits it to the dispatch planning device 4, and extracts specific data from the location data 241c based on a location data acquisition request and transmits it to the dispatch planning device 4. Furthermore, the data management module 242 updates the vehicle data 241b and the location data 241c based on the update data received from the dispatch planning device 4.

[0027] (Configuration of a demand forecasting system) Next, the demand forecasting device 3 in the dispatch management system 1 will be explained using Figure 4. The demand forecasting device 3 acquires specific demand measurement data from the sample data 241a from the data management device 2 and calculates the demand forecast value for each vehicle V at each location in a service area (service area A) based on the demand measurement data. The calculated demand forecast value is transmitted to the dispatch planning device 4 and the data management device 2. The demand forecasting device 3 consists of an input device 31, an output device 32, a communication device 33, and a controller 34. The demand forecasting device 3 may be an information processing device such as a personal computer or a server computer. The configuration of the input device 31, output device 32, and communication device 33 is equivalent to that of the input device 21, output device 22, and communication device 23 in the data management device 2, so a detailed explanation is omitted.

[0028] The controller 34 is an electronic control unit (ECU) that controls the operation of the demand forecasting device 3. The controller 34 comprises a processor 34a and a storage device 34b. The processor 34a has the same configuration as the processor 24a in the data management device 2, so a detailed explanation is omitted. The storage device 34b has basically the same configuration as the storage device 24b in the data management device 2, but the programs it stores are different. The storage device 34b stores programs for implementing control modules such as the prediction value calculation module 341 and the error calculation module 342. The functions of the demand forecasting device 3 are realized when these programs are executed by the processor 34a.

[0029] The forecast value calculation module 341 calculates the forecast demand for vehicles V at each location a in service area A based on demand measurement data and the like stored in the data management device 2. For example, the forecast value calculation module 341 may calculate the forecast demand for the next unit time (7:15 to 7:30) at the start of a new unit time in which demand measurement values ​​are taken (for example, 7:00).

[0030] The forecast value calculation module 341 calculates a demand forecast value based on demand measurement values ​​for each location a within service area A, based on sample data 241a stored in the storage device 24b of the data management device 2, for example. Specifically, the forecast value calculation module 341 may calculate the latest demand forecast value based on past demand measurement data and demand factor data (location factor data, period division factor data) of the location a to be predicted from the sample data 241a. For example, the forecast value calculation module 341 may extract sample data 241a corresponding to the location a to be predicted and the period division to which the unit time of the prediction belongs from the data management device 2, and calculate the demand forecast value based on the extracted sample data 241a. The forecast value calculation module 341 may include a demand forecasting model that has been machine-learned according to a predetermined machine learning algorithm, for example, using the acquired sample data 241a (demand measurement data, demand factor data) as explanatory variables and the demand forecast value (e.g., the number of requests per unit time) as the dependent variable.

[0031] The forecast value calculation module 341 transmits forecast value data, including the calculated demand forecast value, to the dispatch planning device 4 via the communication device 33. The forecast value data may include the demand forecast value, the location of the vehicle to be forecasted, and the start time of the unit time of the vehicle to be forecasted. This allows the dispatch planning device 4 to create a dispatch plan for the vacant vehicle V based on the latest demand forecast value. In this embodiment, the forecast value calculation module 341 may transmit forecast value data to the dispatch planning device 4, which includes at least the forecast demand values ​​for multiple deployment locations a indicated by the deployment location group data received from the dispatch planning device 4. As will be described in detail later, the deployment location group data is data that identifies multiple deployment locations a that can be the deployment destination for a single vehicle V. The forecast value calculation module 341 may also transmit the forecast demand values ​​for all deployment locations a within the service area A to the dispatch planning device 4. Furthermore, the forecast value calculation module 341 may transmit forecast value data, including all calculated demand forecast values, to the data management device 2. The data management module 242 of the data management device 2 stores the received forecast value data in the data storage module 241 at unit time intervals. This allows the data management device 2 to maintain a correspondence between demand forecast values ​​and demand measurement data. In other words, the storage device 24b of the data management device 2 stores forecast error values, which indicate the error between the actually measured demand measurement values ​​and the demand forecast values ​​at each location a in a single district (service district A). Note that the forecast value data and forecast error values ​​may also be stored in a database separate from the data storage module 241.

[0032] As described above, in this embodiment, demand forecast values ​​are calculated based on sample data 241a. However, due to various factors, it may become difficult to forecast demand based on demand factor data and past demand measurements at the target location, and the calculated demand forecast values ​​may deviate from the actual demand measurements. In this case, the dispatch efficiency will be insufficient, and user convenience will be reduced.

[0033] Therefore, in this embodiment, the forecast value calculation module 341 may correct the latest demand forecast value calculated as described above using error time series data that shows the forecast error values ​​for a predetermined period in the past at the location of the target of forecasting. In other words, the demand forecast value included in the forecast value data may be the corrected value.

[0034] Here, an example of error time series data will be explained with reference to Figure 4B. Error time series data is data that shows the error between measured demand values ​​and forecast demand values ​​measured over a predetermined period in the past, in time series. The predetermined period in the past is, for example, the period including the previous unit time "t-1" (the immediately preceding unit time) when the current unit time is "t". If one unit time (for example, 15 minutes) is considered "1 unit", then the predetermined period in the past can be the past period of n units including the unit time "t-1", that is, a continuous period from unit time "t-1" to unit time "tn" (15 minutes × n). In other words, error time series data is data that shows the forecast error values ​​over multiple (n units) unit times including the immediately preceding unit time in time series. Furthermore, the time period to be predicted may be, for example, the current unit time "t", or for example, the next unit time "t+1" (not shown). In other words, the prediction calculation module 341 may calculate the demand forecast value for the target location a at the start of a unit time (current unit time t), or it may calculate the demand forecast value for the next unit time (next unit time t+1) before the start of a unit time. The forecast value calculation module 341 calculates the average value of the forecast error values ​​for a predetermined past period included in the error time series data, and adds this average value to the latest demand forecast value to perform the above correction. For example, if the average value to be added is a positive number, the latest demand forecast value will be higher by the average value than the calculated latest demand forecast value, and if it is a negative number, the latest demand forecast value will be lower by the average value. The forecast value calculation module 341 transmits the corrected demand forecast value to the vehicle dispatch planning device 4, as described above. This makes it possible to create an efficient vehicle allocation plan based on the corrected demand forecast value.

[0035] The forecast value calculation module 341 may also derive the mode of the forecast error values ​​for a predetermined past period included in the error time series data, and add this mode to the latest demand forecast value as the above correction. In other words, after calculating the demand forecast value for each location a, the forecast value calculation module 341 may correct the demand forecast value with a representative value (mean or mode) of the forecast error values ​​(error time series data) for a predetermined past period.

[0036] The error calculation module 342 calculates the error variance, which shows the variance of the predicted error values ​​over a predetermined past period. The error variance is calculated using the predicted error values ​​included in the error time series data. The error calculation module 342 may obtain the average value of the predicted error values ​​included in the error time series data calculated by the predicted value calculation module 341, and calculate the error variance based on this average value. For example, the error calculation module 342 may calculate the error variance using the following equation (1). Note that in equation (1), the error variance is "S 2 ", and the prediction error values ​​from unit time "t-1" to unit time "tn" are each "x 1、 x2···x n Let "n" be the number of data points for the prediction error values, and "x" be the average value of the prediction error values.

number

[0037] The error calculation module 342 may also calculate the error variance from the distribution of multiple predicted error values ​​(predicted error values ​​included in the error time series data). The error calculation module 342 may output the calculated error variance to the prediction value calculation module 341. The prediction value calculation module 341 may transmit the prediction value data for each of the above-mentioned placement points a and the error variance for each placement point a to the vehicle dispatch planning device 4 as demand forecast results. The error calculation module 342 may calculate the error variance for placement point a indicated by the placement point group data, or it may calculate the error variance for all placement points a in service area A.

[0038] (Configuration of the vehicle dispatch planning system) Next, the dispatch planning device 4 in the dispatch management system 1 will be described using Figures 5A and 5B. The dispatch planning device 4 obtains forecast data including the forecasted demand value and the error variance showing the variance of the forecast error value as the demand forecast result for each location a from the demand forecasting device 3. Based on the demand forecast result and the error variance, it creates a dispatch plan that includes the location a to which the vehicle V, which is currently empty, will be placed, and then places the vehicle V to the specific location a.

[0039] Furthermore, the dispatch planning device 4 creates a dispatch plan for vehicle V based on dispatch requests from users who wish to ride in vehicle V, and dispatches vehicle V from each deployment point a to the user's desired boarding location. The dispatch planning device 4 may also transmit dispatch request information, which associates the dispatch request with the user's boarding location (boarding point), to the demand measurement device 6. This allows the demand measurement device 6 to aggregate the number of requests per unit time (number of dispatch requests) for each deployment point a that includes the boarding location in its service area as a demand measurement value. However, it is not limited to this, and the demand measurement device 6 may also extract dispatch request information and dispatch plan information stored in the storage device 44b of the dispatch planning device 4 (described later), and measure the number of requests per unit time for each deployment point a based on these.

[0040] The dispatch planning device 4 consists of an input device 41, an output device 42, a communication device 43, and a controller 44. The dispatch planning device 4 may be an information processing device such as a personal computer or a server computer. The configuration of the input device 41, output device 42, and communication device 43 is equivalent to that of the input device 21, output device 22, and communication device 23 in the data management device 2, so a detailed explanation is omitted.

[0041] The controller 44 is an electronic control unit (ECU) that controls the operation of the dispatch planning device 4. The controller 44 comprises a processor 44a and a storage device 44b. The processor 44a has the same configuration as the processor 24a in the data management device 2, so a detailed explanation is omitted. The storage device 44b has basically the same configuration as the storage device 24b in the data management device 2, but the programs it stores are different. The storage device 44b stores programs for implementing control modules such as the vehicle management module 441, the destination determination module 442, and the plan creation module 443. The functions of the vehicle dispatch planning device 4 are realized when the processor 44a executes these programs.

[0042] The vehicle management module 441 manages the status of vehicle V. For example, when the vehicle management module 441 receives identification information (vehicle ID), location information indicating the current location, and data indicating the service status from vehicle V, it may transmit this data as update data to the data management device 2 via the communication device 43. As a result, the vehicle data 241b and the location data 241c are updated in the data management device. The data indicating the service status indicates whether or not vehicle V is in service (a user is riding in it). Furthermore, if a vehicle V is in an empty state, the vehicle management module 441 determines that it is necessary to determine a waiting location (deployment location a) for the vehicle V to wait, and outputs information identifying the vehicle V (e.g., "vehicle ID") to the deployment location determination module 442, requesting that the deployment location be determined. For example, the vehicle management module 441 may determine whether each vehicle V is empty or not based on data indicating the service status of each vehicle V received from each vehicle V, a dispatch plan, etc. If there are multiple empty vehicle Vs, the vehicle management module 441 may sequentially output the vehicle IDs of the multiple vehicle Vs to the deployment location determination module 442, requesting that the deployment location be determined. In addition, the vehicle management module 441 may store the location information received from the vehicle V in a predetermined storage area of ​​the storage device 44b during the status management of the vehicle V, and output this location information to the deployment location determination module 442 when requesting that the deployment location be determined.

[0043] The placement determination module 442 acquires vehicle data for a vehicle V from the data management device 2 based on the vehicle information (e.g., vehicle ID) output from the vehicle management module 441. Based on the acquired vehicle data, the placement determination module 442 acquires the demand forecast results (demand forecast value and error variance) for each placement location a from the demand forecasting device 3 and determines the placement location a to which the vehicle V will be placed. The following describes in detail the processing performed by the placement determination module 442.

[0044] <Identification of deployment locations> The destination determination module 442 obtains the vehicle data of a vehicle V from the vehicle data 241b stored in the data management device 2 based on the vehicle ID of a vehicle V output from the vehicle management module 441. The destination determination module 442 may send the vehicle ID of a vehicle V to the data management device 2 as a request to obtain vehicle data, and obtain the vehicle data of a vehicle V (at least the current location information of a vehicle V) from the data management device 2. The placement location determination module 442 identifies placement locations a within a predetermined range from the current position of a vehicle V as a group of placement locations, and obtains the demand forecast value and error variance for each placement location a included in the group of placement locations from the demand forecasting device 3. The group of placement locations represents multiple placement locations a that are candidates for the placement of a vehicle V. In other words, the placement location determination module 442 may identify placement locations a within a predetermined range from the position of a vehicle V (for example, position P) as a group of placement locations, and extract candidate placement locations from the placement locations a included in the group of placement locations. For example, the placement determination module 442 designates an area within a predetermined range from the current position of a vehicle V within service area A as the placement area, and extracts data for placement points a located within the placement area from the data management device 2. The placement determination module 442 may send location information indicating the placement area to the data management device 2 as a request to acquire placement point data, and acquire placement point data (data of a group of placement points) related to placement points a included in the placement area from the data management device 2. The above placement area may be an area within a predetermined distance from the current position of a vehicle V, or an area that can be reached within a predetermined time from the current position of a vehicle V. In other words, the above placement area may be an area that can be reached within a predetermined range with a travel cost (distance or time).

[0045] The placement location determination module 442 may send data of the group of placement locations extracted from the data management device 2 to the demand forecasting device 3 as a forecast result request, and obtain from the demand forecasting device 3 forecast value data (demand forecast values) and demand forecast results including error variance for multiple placement locations a included in the group of placement locations.

[0046] <Extraction of potential placement locations> The placement determination module 442 extracts placement locations a that satisfy predetermined conditions for the error variance in the acquired demand forecast results as candidate placement locations for a single vehicle V. This makes it possible to determine the placement location of vehicle V while taking into account the variance (scattering) of the forecast error values. More specifically, the placement location determination module 442 may extract placement locations a whose error variance is below a predetermined threshold as candidate placement locations for a single vehicle V. In other words, the placement location for a single vehicle V may be determined from among the placement locations a included in the group of placement locations, where the error variance is below a predetermined threshold. This makes it possible to determine the placement location for a vehicle V from among placement locations a with small variance (variability) of the predicted error value. Therefore, it is possible to place the vehicle V with greater certainty while considering the variance of the predicted error value.

[0047] <Decision on placement> The deployment location determination module 442 prioritizes high-demand deployment locations, which are determined to have relatively high demand based on demand forecasts, among the deployment locations a extracted as candidate locations, and determines them as the deployment locations for one vehicle V. As a result, the vehicle dispatch management system 1 can improve the balance of supply and demand for vehicles within the service area and increase the vehicle utilization rate. More specifically, the deployment location determination module 442 may subtract the number of vehicles currently deployed at each candidate deployment location a from the demand forecast value to calculate the number of vehicles V currently lacking at each deployment location a, and identify high-demand deployment locations based on this number of vehicles lacking. The deployment location determination module 442 only needs to obtain the number of vehicles currently deployed at each candidate deployment location a from the data management device 2.

[0048] The magnitude of the shortage value indicates the high demand for vehicles V at each deployment location a. In other words, the deployment location determination module 442 may identify deployment locations a with a relatively large shortage value among the deployment locations a extracted as candidate locations as high-demand deployment locations. This allows the dispatch management system 1 to further improve the utilization rate of vehicles V within the area where mobility services are provided. Furthermore, high-demand placement locations may be the locations with the largest shortage of vehicles among the candidate placement locations a. This allows the dispatch management system 1 to more reliably improve the utilization rate of vehicles V within the service area of ​​the mobility service.

[0049] The deployment location determination module 442 outputs information indicating the determined deployment location (deployment location a) (for example, location information of the deployment location) to the plan creation module 443. This allows a vehicle V to be deployed to high-demand locations first. More specifically, within the group of deployment locations, a vehicle V can be deployed to location a, which has relatively high demand (a large shortage of vehicles), first.

[0050] The planning module 443 creates a deployment plan for the deployment location (deployment point a) determined by the deployment location determination module 442. The deployment plan includes a "vehicle ID" that identifies at least one vehicle V, the current status of the vehicle V, and location information of the deployment location a. The planning module 443 transmits the contents of the created deployment plan as update data to the data management device 2, and also transmits an instruction to move to the deployment location (movement instruction information) to the vehicle V. The movement instruction information includes at least location information of the destination deployment location a. This allows the vehicle V to be deployed to the deployment location a. In addition, the vehicle data 241b related to the vehicle V and the deployment location data 241c related to the deployment location a (for example, the number of vehicles to be deployed at the deployment location a) are updated in the data management device 2.

[0051] Vehicle V is equipped with a communication device (not shown) and an output device (not shown) capable of outputting information from the dispatch planning device 4, and is configured to receive and output instructions from the dispatch planning device 4. For example, vehicle V displays movement instruction information on its output device (e.g., display device). Upon receiving movement instruction information, vehicle V uses its navigation function to search for a route to the destination location and guides the driver. Alternatively, in the case of a fully autonomous vehicle, vehicle V, upon receiving movement instruction information, moves to the destination location autonomously. In this way, the dispatch planning device 4 places vehicles V at each location within a single area (service area A) based on the demand forecast values ​​calculated by the demand forecasting device 3 as described above. This makes it possible to place vehicles V at each location in a way that increases the utilization rate of vehicles V while improving user convenience.

[0052] Furthermore, the planning module 443 creates a vehicle dispatch plan based on dispatch requests from the user's terminal device (not shown) received via the communication device 43, and dispatches the vehicle V from each dispatch point a to the boarding location desired by the user. The dispatch management system 1 provides, for example, an on-demand mobility service (dispatch service) using multiple vehicles V targeting a specific service area (e.g., service area A). For this reason, the dispatch request includes information on the boarding location where the user wishes to board the vehicle V (desired boarding location), the disembarking location where the user wishes to disembark from the vehicle V (desired disembarking location), and the number of people using the vehicle V, but does not include information on the date and time the user wishes to board the vehicle V (desired boarding date and time). The dispatch plan may include, for example, the location of the vehicle V to be dispatched to the user, the travel schedule, and the travel route. The travel schedule may include the estimated arrival time at the pick-up location where the user boards and the estimated arrival time at the drop-off location where the user alights. The travel route shows the route taken from the pick-up location to the drop-off location.

[0053] For example, when the planning module 443 creates a dispatch plan, it sends the plan to a designated vehicle V located at the nearest dispatch point a to the pick-up location, and moves vehicle V to the pick-up location by the desired pick-up date and time. As a result, vehicle V is dispatched to the user. For example, when the planning module 443 creates a dispatch plan and dispatches a vehicle V to a user, it associates the boarding location in the dispatch plan with the date and time the dispatch request was received and transmits this as dispatch request information to the demand measurement device 6. This allows the demand measurement device 6 to measure the demand measurement value (number of requests) per unit time at each deployment location that includes the boarding location in its service area. The dispatch request information may also include the content of the dispatch request. Furthermore, when the planning module 443 creates a dispatch plan, it transmits the contents of the dispatch plan as update data to the data management device 2. This updates the vehicle data 241b related to a single vehicle V and the deployment location data 241c related to the deployment location a where the vehicle is deployed (for example, the number of vehicles deployed at deployment location a where the vehicle V dispatched to the user was waiting).

[0054] (Operation of the dispatch management system) Figure 6 is a flowchart showing an example of a dispatch management method in the dispatch management system 1 according to the first embodiment of this disclosure. This method is realized by each device constituting the dispatch management system 1 executing a dispatch management program under the control of a processor, and thereby cooperating with hardware resources. In this example, the processing performed by each controller (controllers 34, 44) of the dispatch management device 10 (demand forecasting device 3 and dispatch planning device 4) in the dispatch management system 1 will be described.

[0055] As shown in Figure 6, the demand forecasting device 3 (forecast value calculation module 341) calculates the forecast demand for each location a in the service area A based on sample data 241a extracted from the data management device 2 (S601), and corrects the calculated forecast demand using representative values ​​of the forecast error (error time series data) for a predetermined past period at each location a (S602).

[0056] Next, the demand forecasting device 3 (error calculation module 342) obtains the average value of the predicted error values ​​included in the error time series data and calculates the error variance for each placement point a (S603). The dispatch planning device 4 (vehicle management module 441) determines whether or not there are any empty vehicles V (S604). For example, if the vehicle management module 441 determines that a vehicle V is empty based on the service status of the vehicle V in the vehicle data 241b, the dispatch plan, etc. (Yes in S604), it outputs the identification information (vehicle ID) of the vehicle V to the destination determination module 442 and requests the determination of the destination location a for the vehicle V. On the other hand, if the dispatch planning device 4 (vehicle management module 441) determines that there are no empty vehicles V (No. in S604), it waits for an empty vehicle V to become available.

[0057] The dispatch planning device 4 (determination module 442) acquires vehicle data 241b of a vehicle V from the data management device 2 based on the fact that a vehicle V has become vacant and a determination of its destination has been requested (S605), and identifies a destination point a within a predetermined range (determination area A1) from the location information (current position) of the vehicle V in the acquired vehicle data (S606). Specifically, the determination module 442 extracts data of destination points a included in the destination area A1 identified based on the current position of the vehicle V from the destination point data 241c stored in the data management device 2, and identifies them as a group of destination points. The dispatch planning device 4 (location determination module 442) transmits data of the group of locations as a forecast result request to the demand forecasting device 3, and the demand forecasting device 3 obtains the forecasted demand value and error variance for each location a in the group of locations as the demand forecast result (S607).

[0058] Next, the vehicle dispatch planning device 4 (determination module 442) extracts from the identified group of dispatch locations a that have an error variance below a predetermined threshold as a candidate dispatch location for one vehicle V (S608), and obtains the number of vehicles V dispatched at the extracted dispatch location a (candidate dispatch location) from the data management device 2 (S609). Furthermore, the vehicle dispatch planning device 4 (determination module 442) identifies the location a where the shortage of vehicles is greatest (S610), and determines the identified location a as the destination for one vehicle V (S611). For example, the placement determination module 442 calculates the number of vehicles needed by subtracting the number of vehicles needed from the demand forecast value for each of the extracted placement locations a (candidate placement locations), identifies the placement location a with the largest number of vehicles needed (high-demand placement location), and determines that this is the placement location for one vehicle V. As a result, the vacant vehicle V is placed at the placement location a with the highest demand among the placement locations a with the smallest error variance.

[0059] (First variation) In the first embodiment described above, locations where the error variance is below a predetermined threshold are extracted as candidate locations for a single vehicle V, but the disclosure is not limited thereto. For example, the placement determination module 442 may calculate the probability (demand occurrence probability) that demand of a predetermined value or higher occurs at each placement location a based on the demand forecast value, forecast error, and error variance, and extract placement locations where the demand occurrence probability is greater than or equal to a predetermined value as candidate placement locations for a single vehicle V. In this modified example, the placement determination module 442 may obtain, as a demand forecast result, the forecast error (error time series data) for each placement location a in the group of placement locations, in addition to the demand forecast value and error variance from the forecast value calculation module 341. Furthermore, in this modified example, the error variance calculated by the error calculation module 342 may be the standard deviation of the forecast error value in the error time series data.

[0060] More specifically, demand exceeding a predetermined value means that the vehicle V demand (dispatch requests) for each deployment location a included in the deployment location group during a unit of time (t) for which demand forecasting is being conducted is equal to or greater than the current number of vehicles deployed at each deployment location a plus 1 (number of vehicles deployed + 1). The placement determination module 442 may apply the demand forecast results (demand forecast value, forecast error, and error variance) to a cumulative distribution function to calculate the probability (exclusion probability) that the demand for vehicle V in a unit time (t) is less than or equal to a predetermined value, and then calculate the demand occurrence probability Dp based on the calculated exclusion probability. The demand occurrence probability Dp is calculated by the following equation (2), assuming that the distribution of the forecast error values ​​for each placement location a included in the placement location group is a normal distribution. In equation (2) below, the predetermined value (number of units to be installed + 1) is denoted by "y", the demand forecast value by "z", the mean value of the forecast error by "μ", the standard deviation of the forecast error by "σ", and the cumulative distribution function by "Φ".

[0061]

number

[0062] Figure 7 is a flowchart showing the first variation of the dispatch management method according to the first embodiment. The processes from steps S701 to S706 are equivalent to the processes from steps S601 to S606 shown in Figure 6, so their explanation is omitted. The dispatch planning device 4 (location determination module 442) transmits the data of the identified group of locations to the demand forecasting device 3 as a forecast result request, and the demand forecasting device 3 obtains the forecasted demand value, forecast error value (error time series data), and error variance (standard deviation) for each location a in the group of locations as the demand forecast result (S707).

[0063] Next, the vehicle dispatch planning device 4 (location determination module 442) obtains the number of vehicles V to be deployed at each deployment location a included in the deployment location group from the data management device 2 (S708), and calculates the probability (demand occurrence probability) that the demand for vehicles V at each deployment location a will be greater than or equal to a predetermined value (demand occurrence probability) (S709). For example, the deployment location determination module 442 applies the acquired demand forecast result to the above equation (2) to calculate the demand occurrence probability Dp at each deployment location a in the deployment location group. The dispatch planning device 4 (destination location determination module 442) extracts a location a from the group of locations where the probability of demand occurrence Dp is equal to or greater than a specified value (for example, 95%) as a candidate location (S710). Then, similar to step S610, it identifies the location a with the greatest shortage of vehicles among the candidate location a locations (S711), and similar to step S611, it determines the identified location a as the destination for one vehicle V (empty) (S712).

[0064] Thus, in this modified example, the controller 44 of the dispatch planning device 4 calculates the probability of demand occurring at each location a based on the demand forecast results (demand forecast value, forecast error, and error variance) (S709 above), and may extract location a where the probability of demand occurring is 95% or higher as a candidate location for a vehicle V (S710 above). By using the probability of demand occurring, the dispatch management system 1 can analyze the supply and demand balance at each location a with high accuracy while considering the variance of the forecast error value, efficiently allocate vehicles V to each location a, and reliably improve the operating rate of vehicles V.

[0065] (Second variation) For example, the placement determination module 442 may calculate an assumed value for the demand for vehicle V under specific conditions at each placement location a based on the demand forecast value, forecast error, and error variance, and extract placement locations a with a relatively large assumed value as a candidate placement location for one vehicle V. In this modified example, the placement determination module 442 may obtain the demand forecast value, error variance (standard deviation), and forecast error value (error time series data) for each placement location a in the group of placement locations from the forecast value calculation module 341 as the demand forecast result, similar to the first modified example. In this modified example, the demand forecast value is not corrected with a representative value of the forecast error (steps S602 and S702 are not performed), which is different from the first embodiment and the first modified example.

[0066] More specifically, the placement determination module 442 may calculate an assumed value by correcting the demand forecast value for each placement location a with a downward deviation of the forecast error value over a predetermined past period. In other words, the assumed value is a value that assumes a downward deviation in the demand for vehicles V at each placement location a. The placement determination module 442 calculates the assumed value Sv from the demand forecast value, the mean value of the forecast error value (error time series data), and the error variance (standard deviation) of each placement location a. The assumed value Sv is calculated by the following equation (3), assuming that the distribution of the forecast error values ​​for each placement location a included in the group of placement locations follows a normal distribution. In equation (3) below, the demand forecast value is denoted by "z", the mean value of the forecast error is denoted by "μ", and the standard deviation of the forecast error is denoted by "σ". In this example, the lower bound of the forecast error is defined as the threshold of the bottom 5% of forecast error values ​​over a predetermined past period (the maximum value among the bottom 5%).

[0067]

number

[0068] Figure 8 is a flowchart showing the first variation of the dispatch management method according to the first embodiment. The processes from steps S801 to S805 are equivalent to the processes from steps S601 and S603 to S606 shown in Figure 6, so their explanation is omitted. The dispatch planning device 4 (location determination module 442) transmits the data of the identified group of locations as a forecast result request to the demand forecasting device 3, and the demand forecasting device 3 obtains the forecasted demand value, forecast error value (error time series data), and error variance (standard deviation) for each location a in the group of locations as the demand forecast result (S806).

[0069] Next, the dispatch planning device 4 (location determination module 442) calculates an assumed value Sv by correcting the demand forecast value of each location a included in the group of location locations with the lower limit of the forecast error value (the lower 5% threshold) as shown in equation (3) above (S807), and extracts location a with a relatively large assumed value Sv as a candidate location for one vehicle V (S808). Next, the vehicle dispatch planning device 4 (determination module 442) obtains the number of vehicles V to be deployed at the extracted deployment location a (candidate deployment location) from the data management device 2 (S809), identifies the deployment location a with the greatest shortage of deployed vehicles (S810), and determines the identified deployment location a as the deployment location for one vehicle V (S811). For example, the deployment location determination module 442 calculates the number of vehicles to be deployed by subtracting the calculated estimated value Sv from the number of vehicles to be deployed at each of the candidate deployment locations a, identifies the deployment location a with the largest number of vehicles to be deployed, and determines that the identified deployment location a is the deployment location for one vehicle V (empty).

[0070] Thus, in this modified example, the controller 44 (location determination module 442) of the dispatch planning device 4 calculates an assumed value Sv based on the demand forecast results (demand forecast value, forecast error, and error variance), which assumes the demand for vehicles under specific conditions at each location a (S807), and may extract location a with a relatively large assumed value Sv as a candidate for a single vehicle V (S808). By using the assumed value Sv based on the lower end of the forecast error value, the dispatch management system 1 can more reliably suppress the discrepancy between the actual demand value (measured demand value) and the demand forecast value. Therefore, by efficiently allocating vehicles while considering the variance of the forecast error value, the supply and demand balance of vehicles V within the service area A can be improved, and the utilization rate of vehicles V can be reliably increased. In this modified example, the calculation of the assumed value Sv may be performed by the demand forecasting device 3 (forecast value calculation module 341). In this case, it is sufficient that the assumed value Sv is included as a demand forecast value in the demand forecast result output by the forecast value calculation module 341 to the dispatch planning device 4.

[0071] 2. Second Embodiment In the first embodiment described above, the vehicle dispatch planning device 4 extracts location points a whose error variance satisfies a predetermined condition (for example, the error variance is less than or equal to a predetermined value) as candidates for the placement of one vehicle V, and from the extracted candidates (location points a), it determines the location point a with a relatively large number of vehicles in short supply (high-demand location point) as the placement of one vehicle V, but the disclosure is not limited thereto. For example, when a vehicle becomes vacant, there may be no location a within service area A where the error variance satisfies the predetermined conditions, or there may be no shortage of vehicles V at location a selected as a candidate location (for example, more vehicles V than the demand forecast value are already deployed). In such cases, the vehicle dispatch planning device 4 (location determination module 442) may group each location a in the group of location locations based on error variance and classify them into multiple location groups, and then determine the group among these multiple location groups in which the demand for vehicle V is relatively high as the location for one vehicle V. This allows for more flexible determination within the group of location locations while taking into account the variance of the prediction error value. Grouping can be performed using known methods such as cluster analysis (e.g., hierarchical clustering). The placement determination module 442 may include a grouping model that performs clustering processing according to a predetermined machine learning algorithm.

[0072] <Identification of reference points> The placement determination module 442 identifies a placement point a, which will be the center (cluster nucleus) of a placement group (cluster), as a reference point, and generates a placement group for each of these reference points. The characteristics of the placement group are determined by the characteristics of the reference point (error variance, demand forecast value). The placement determination module 442 may identify multiple placement locations a with relatively small error variances as reference points for multiple placement groups. By using placement locations a with relatively small error variances as reference points, it is possible to generate placement groups with suppressed overall error variances. Preferably, the placement determination module 442 identifies multiple placement locations a with relatively small error variances and relatively large demand forecast values ​​as reference points. This makes it possible to generate placement groups centered around placement locations a where the overall error variance is suppressed and the demand for vehicles V is predicted to be high.

[0073] The placement location determination module 442 may divide the placement area A1 (see Figure 10 described later) within a single service area (for example, service area A) into subdivided areas of a predetermined size (default area) or less, and identify one or more placement points a as reference points for each subdivided area. This makes it possible to generate well-balanced placement groups within the placement area A1.

[0074] The placement location determination module 442 may extract multiple placement locations a with relatively small error variances (primary extraction), and then, from the multiple placement locations a extracted in the primary extraction, extract multiple placement locations a in descending order from the location with the largest demand forecast value (secondary extraction) and identify them as reference locations. In the secondary extraction, one or more placement locations a may be identified as reference locations for each divided area. Multiple reference locations identified within a placement area (placement area A1) are called a "reference location group".

[0075] <Identifying potential members> The placement determination module 442 identifies multiple reference points (reference point groups) and, for each reference point, identifies multiple placement points a other than the reference point that are candidates to be members of the placement group. The candidate placement points a only need to be within a predetermined range from each reference point (within a predetermined distance or time range), and may be placement points a within the divided region to which each reference point belongs. Placement points a from other divided regions may also be included as candidate members.

[0076] The placement determination module 442 identifies candidate members for each reference point and generates a placement group with each central (core) reference point and the candidate member. Since each placement point a that is a candidate member is not yet confirmed as a member of the placement group, the placement group including the candidate member is sometimes referred to as the "initial group".

[0077] <Adjusting potential members> The placement determination module 442 adjusts the member candidates in the initial group so that the error variance is below a predetermined threshold. The adjustment of member candidates is performed for each initial group. Based on the error variance of the member candidates in the initial group (its own group), the placement determination module 442 may decide whether or not to add each member candidate in the initial group to the placement group. The combination of member candidates after adjustment is sometimes referred to as an "adjusted group".

[0078] The placement determination module 442 may decide to include a candidate member (placement location a) as a member of the placement group if the addition of that candidate member reduces the error variance of the adjustment group. This allows placement location a, which can reduce the error variance of the adjustment group, to be classified into the placement group. For example, the placement determination module 442 compares the average value A of the error variances of placement points a (reference point and all member candidates) included in the initial group (itself group) with the average value B of the error variances of the adjustment group consisting of the remaining member candidates after excluding one member candidate (placement point a) from the initial group. If the average value A is smaller than the average value B (average value A < average value B), it may decide to add (classify) that one member candidate (placement point a) to the placement group. This makes it possible to generate placement groups (clusters) that add as many neighboring placement points a as possible, within a range where the average value of the error variance for each reference point is below a predetermined threshold.

[0079] <Decision on placement> The deployment location determination module 442 generates deployment location groups (confirmed groups) and classifies each deployment location a within the deployment location group into multiple deployment location groups based on error variance. It then determines the level of demand for vehicles V among the multiple deployment location groups based on the demand forecast value and may prioritize deployment location a in the high-demand group where the demand is relatively high, and determine it as the deployment location for one vehicle V. As a result, even if there are no deployment locations a in the deployment location group that satisfy the predetermined error variance, or if there is no shortage of vehicles V at the deployment location a extracted as a candidate deployment location, the vehicle dispatch management system 1 can efficiently deploy vehicles by considering the variance of the forecast error value, thereby improving the supply and demand balance of vehicles within the area where the mobility service is provided and improving the vehicle utilization rate.

[0080] Specifically, the placement determination module 442 may calculate a total forecast value by summing the demand forecast values ​​for each placement point a (reference point and member) within each placement group (confirmed group), and calculate a total number of units by summing the number of units to be placed at each placement point a within each placement group (confirmed group).

[0081] The deployment determination module 442 may subtract the total number of units from the predicted total to calculate a group demand value indicating the demand for vehicles V in each deployment group (confirmed group). The deployment determination module 442 can then determine the level of demand for vehicles V among multiple deployment groups by comparing these group demand values. In other words, among multiple deployment groups (confirmed groups), the group with the relatively higher group demand value may be determined as the high-demand group. For example, the confirmed group with the highest group demand value may be determined as the high-demand group. This allows the dispatch management system 1 to more reliably identify dispatch groups with high demand for vehicles V as high-demand groups.

[0082] The deployment location determination module 442 determines the deployment location a for one vehicle V from among the deployment locations a included in the high-demand group. The destination determination module 442 may designate a reference point of a destination group that is a high-demand group as the destination for one vehicle V. In other words, the dispatch planning device 4 may place an empty vehicle V at a reference point (for example, reference point a1) in a high-demand group. By placing an empty vehicle V at a reference point, which is a destination point a with a relatively large demand forecast value within each divided area, the utilization rate can be improved. Furthermore, since each destination point a, which is a member of each destination group (confirmed group), is within a predetermined range from the reference point, even if the demand for vehicle V (dispatch request) increases at a destination point a other than the reference point, the vehicle V can be quickly moved to the destination point a where the demand has increased. In addition, by prioritizing the placement of vehicle V at the reference point, it becomes possible to smoothly distribute vehicle V based on the demand at each destination point a within the same group, thereby efficiently improving the utilization rate. Furthermore, if location a, where the demand forecast value is greater than that of the reference point, is included in the high-demand group, location a may be designated as the location.

[0083] As another example, the deployment location determination module 442 may calculate the location demand, which indicates the demand for vehicle V at each deployment location a within the high-demand group, and prioritize deployment locations a with relatively high location demand within the high-demand group to determine the deployment location for a single vehicle V. This allows for the deployment of an empty vehicle V to a deployment location a that has relatively high demand for vehicle V, even among deployment locations a belonging to a deployment group with high demand for vehicle V, where the current error variance is low and there is little risk of discrepancy between the demand forecast value and the actual value (measured demand). In other words, vehicle V can be deployed to the deployment location a that has the highest potential to improve the utilization rate at this time.

[0084] Figure 9 is a flowchart showing an example of a vehicle dispatch management method according to the second embodiment. Figure 10 is a schematic diagram illustrating the grouping of dispatch points a included in the dispatch area A1 within service area A. The grouping of dispatch points a will be explained below with reference to Figures 9 and 10. Note that the processes from steps S901 to S908 are equivalent to the processes from steps S601 to S608 shown in Figure 6, so the explanation will be omitted.

[0085] The vehicle dispatch planning device 4 (determination module 442) determines whether or not one or more location locations a with an error variance of less than or equal to a predetermined value have been extracted from the group of location locations as candidates for the location of a single vehicle V (S909). If the vehicle dispatch planning device 4 (location determination module 442) determines that the group of locations includes location a with an error variance of less than or equal to a predetermined value and that one or more candidate locations have been extracted (Yes in S909), it obtains the number of vehicles V to be deployed at the extracted location a (candidate location) from the data management device 2, similar to step S609 (S910), and determines whether or not there is a shortage of vehicles at location a, which has become a candidate location (S911). If the dispatch planning device 4 (determination module 442) determines that there is a shortage of one or more vehicles (= forecast demand - number of vehicles to be allocated) at a candidate allocation location a (Yes in S911), it identifies the allocation location a with the largest shortage of vehicles (number of vehicles to be allocated) in the same way as in step S610 (S912), and determines the identified allocation location a as the allocation location for one vehicle V (empty) in the same way as in step S611 (S913).

[0086] On the other hand, if the dispatch planning device 4 (determination module 442) determines that one or more dispatch locations a have not been extracted as candidate dispatch locations (No. in S909) or if it determines that there is no shortage of dispatch locations among the candidate dispatch locations (No. in S911), it groups the dispatch locations a included in the dispatch location group and classifies them into multiple dispatch location groups (S914).

[0087] As shown in Figure 10, for example, the placement determination module 442 divides the placement area A1, which is set within a predetermined range (e.g., within a predetermined distance) from the position P of an empty vehicle V, into divided areas A11, A12, A13, and A14 with an area less than or equal to a predetermined area, and identifies one or more placement points a as reference points for each divided area. In this example, one reference point (reference points a1, a2, a3, a4) is identified in each of the divided areas A11 to A14. Once the placement determination module 442 identifies member candidates for each reference point (reference points a1, a2, a3, a4), it generates an initial group with each central (core) reference point and the aforementioned member candidates (placement points a other than the reference points), and adjusts the member candidates within the initial group to determine the members of the placement group. As a result, placement points a within the placement area A1 (placement points a within the group of placement points) are classified into multiple placement groups with reference points a1, a2, a3, a4 as the core, and placement groups are generated.

[0088] When the dispatch planning device 4 (determination module 442) generates a group of destinations, it identifies a high-demand group to which a vehicle V will be allocated (S915), and determines that one allocation point a (e.g., a reference point) within the identified high-demand group is the destination for the vehicle V (empty) (S916).

[0089] Thus, in this embodiment, if no placement location a satisfying a predetermined error variance is extracted as a candidate placement location for a vehicle V, or if there are no placement locations among the extracted candidate placement locations a that are short of one or more vehicles, the controller 44 (placement location determination module 442) of the vehicle dispatch planning device 4 may identify placement locations a within a predetermined range from the location P of a vehicle V as a group of placement locations, group each placement location a in the group of placement locations based on the error variance to classify them into multiple placement location groups, and determine one placement location a that is included in the high-demand group among the multiple placement location groups where the demand for vehicle V is relatively high as the placement location for a vehicle. By grouping placement locations a based on error variance and selecting placement locations a within the high-demand group that include placement locations a with high demand for vehicle V, it is possible to place vehicle V in areas (divided regions) with potentially high demand for vehicle V while considering the variance of prediction error values. Therefore, it is possible to improve the balance of supply and demand for vehicles within the service area and increase the vehicle utilization rate.

[0090] (Effects of the embodiment) (1) The dispatch management system 1 according to this embodiment comprises a vehicle V for providing mobility services, a dispatch management device 10 for managing the placement of the vehicle V to placement points a within a predetermined area, and a data management device 2, wherein the data management device 2 stores demand measurement values ​​indicating the demand for the vehicle V measured at each placement point a within the service area A, demand forecast values ​​indicating the future demand for the vehicle at each placement point a, and forecast error values ​​indicating the error between the demand measurement value and the demand forecast value at each placement point a. The dispatch management device 10 includes controllers 34 and 44 that perform the following processes for each dispatch point a within the service area A: calculating a demand forecast value based on demand measurement values ​​and an error variance showing the variance of the forecast error value over a predetermined past period; extracting dispatch points a whose error variance satisfies predetermined conditions as candidates for the dispatch destination of a vehicle V; and prioritizing high-demand dispatch points among the extracted candidate dispatch points a where the demand for vehicle V is determined to be relatively high based on the demand forecast value, and deciding on them as the dispatch destination for a vehicle V. This configuration allows for the determination of vehicle V's placement location by considering the variance (scatter) of the predicted error value, thereby selecting locations where the error variance satisfies predetermined conditions as potential placement locations. Furthermore, by prioritizing high-demand placement locations among these candidates, it is possible to improve the balance of vehicle supply and demand within the service area and increase vehicle utilization.

[0091] (2) The data management device 2 stores the number of vehicles currently deployed at each deployment location a, The controller 44 may calculate the number of units needed at each location a extracted as a candidate by subtracting the number of units needed at each location a from the demand forecast value, and identify high-demand locations based on the number of units needed. This configuration allows for the identification of high-demand locations based on the current shortage of vehicles V, further improving the balance between vehicle supply and demand and increasing the utilization rate of vehicles V within the service area. (3) The controller 44 may identify a group of placement locations a within a predetermined range from the position P of a vehicle V, and extract candidate placement locations a from the placement locations a included in the group of placement locations. With this configuration, placement locations a that are far from the current location of a vehicle can be excluded from the list of possible placement locations, and a placement location a can be quickly determined from among the placement locations a that are within a range that can be reached quickly. (4) After calculating the demand forecast value for each location a, the controller 34 may correct the demand forecast value with a representative value of the forecast error value over a predetermined past period. This configuration allows for the creation of an efficient vehicle allocation plan based on corrected demand forecasts, thereby improving dispatch efficiency and enhancing user convenience. (5) The controller 44 may extract placement locations a whose error variance is less than or equal to a predetermined threshold as candidate placement locations for a single vehicle V. This configuration allows for the selection of vehicle V's placement location from among placement locations a with low variance (dispersion) of prediction error values. Therefore, it is possible to more reliably consider the variance of prediction error values ​​when efficiently positioning vehicle V, thereby improving vehicle utilization.

[0092] (6) The controller 44 may calculate the probability (demand generation probability) that a demand for a vehicle V of a predetermined value or more will occur at each placement location a based on the demand forecast value, forecast error, and error variance, and may extract placement locations where this probability is greater than or equal to a predetermined value as candidate placement locations for a single vehicle V. This configuration allows for highly accurate analysis of the supply-demand balance at each deployment location a, taking into account the variance of prediction error values, and enables efficient deployment of vehicles V to each deployment location a, thereby reliably improving the utilization rate of vehicles V. (7) The controller 44 may calculate an assumed value for the demand for vehicle V under specific conditions at each placement location a based on the demand forecast value, the forecast error, and the error variance, and may extract a placement location a with a relatively large assumed value as a candidate for the placement of one vehicle V. This configuration allows for more reliable suppression of the discrepancy between actual demand (measured demand) and forecast demand in the dispatch management system 1. Therefore, by efficiently allocating vehicles while considering the variance of forecast error values, the supply-demand balance of vehicles V within service area A can be improved, and the utilization rate of vehicles V can be reliably increased. (8) The controller 44 may calculate an assumed value by correcting the demand forecast value for each location a with the lower limit of the forecast error value over a predetermined past period. This configuration makes it possible to more reliably suppress the discrepancy between actual demand (measured demand) and forecasted demand.

[0093] (9) If no location is extracted as a candidate location for a vehicle V that satisfies the predetermined conditions for error variance, or if there is no location among the extracted candidate locations a that is short of one or more vehicles, the controller 44 may identify locations a within a predetermined range from the location P of a vehicle V as a group of locations, group each location a in the group of locations based on error variance and classify them into multiple destination groups, and decide that one location a included in the group of destination groups where the demand for vehicle V is relatively high will be the destination for a vehicle V. This configuration allows for the placement of vehicles V within areas (divided regions) where there is potentially high demand for vehicles V, while taking into account the variance of prediction error values. Therefore, it is possible to improve the balance of supply and demand for vehicles within the service area and increase vehicle utilization rates. [Explanation of Symbols]

[0094] 1. Dispatch Management System 2. Data Management Device 21 Input device 22 Output device 23 Communication equipment 24 controllers 24a processor 24b Storage device 241 Data Storage Module 241a Sample data 241b Vehicle Data 241c Placement location data 242 Data Management Modules 3. Demand forecasting device 31 Input device 32 Output device 33 Communication equipment 34 controllers 34a processor 34b Storage device 341 Prediction Calculation Module 342 Error Calculation Module 4. Vehicle dispatch planning device 41 Input device 42 Output device 43 Communication equipment 44 controllers 44a processor 44b Storage device 441 Vehicle Management Module 442 Deployment location determination module 443 Planning Module 5-Factor Data Acquisition Device 6. Demand measurement device 9. Communication Network 10. Dispatch management device a Placement point A1 placement area A11, A12, A13, A14 divided area V Vehicle

Claims

1. A dispatch management system comprising: vehicles used to provide mobility services; a dispatch management device for managing the placement of said vehicles at designated locations within a specified area; and a storage device, The aforementioned storage device is The system stores demand measurement values ​​indicating the demand for the vehicles measured at each location within the district, demand forecast values ​​indicating the future demand for the vehicles at each location, and forecast error values ​​indicating the error between the demand measurement value and the demand forecast value at each location. The aforementioned dispatch management device is For each location within the district, a process is performed to calculate the demand forecast value based on the demand measurement value and the error variance showing the variance of the forecast error value over a predetermined past period. A process to extract a location where the error variance satisfies a predetermined condition as a candidate location for one of the vehicles, A vehicle dispatch management system comprising a controller that performs the process of determining the destination of one vehicle by prioritizing high-demand locations among the candidate locations extracted, where the demand for the vehicle is determined to be relatively high based on the demand forecast value.

2. The aforementioned storage device stores the number of vehicles currently deployed at each deployment location. The aforementioned controller, The vehicle dispatch management system according to claim 1, wherein the number of vehicles at each of the candidate placement locations is subtracted from the demand forecast value to calculate the number of vehicles needed at each placement location, and the high-demand placement locations are identified based on the number of vehicles needed.

3. The aforementioned controller, A group of placement locations is identified from the position of one of the aforementioned vehicles, and candidate placement locations are extracted from the placement locations included in the group of placement locations. The dispatch management system according to claim 1.

4. The controller calculates the demand forecast value for each location and then corrects the demand forecast value using a representative value of the forecast error value over a predetermined past period. The dispatch management system according to claim 1.

5. The aforementioned controller, A vehicle dispatch management system according to any one of claims 1 to 4, wherein the error variance is less than or equal to a predetermined threshold, and a location is selected as a candidate location for a vehicle.

6. The aforementioned controller, A vehicle dispatch management system according to any one of claims 1 to 4, wherein, based on the demand forecast value, the forecast error value, and the error variance, the probability that a demand for the vehicle will occur at each dispatch location that is equal to or greater than a predetermined value is calculated, and dispatch locations for which this probability is equal to or greater than a predetermined value are extracted as candidates.

7. The aforementioned controller, Based on the aforementioned demand forecast value, forecast error value, and error variance, an estimated value is calculated that represents the expected demand for the vehicle under specific conditions at each deployment location, and deployment locations with relatively large estimated values ​​are selected as candidates. A vehicle dispatch management system according to any one of claims 1 to 3.

8. The aforementioned controller, The assumed value is calculated by correcting the demand forecast value for each location with the lower bound of the forecast error value over a predetermined past period. The dispatch management system according to claim 7.

9. The aforementioned controller, If no location is selected as a candidate where the error variance satisfies the predetermined conditions, or if there is no location among the selected candidate locations where the number of units to be short is one or more, A group of placement locations is identified from the location of one of the aforementioned vehicles, each placement location in the group is grouped based on the error variance and classified into multiple placement groups, and one placement location that is included in the group with relatively high demand for the vehicle is determined to be the placement location for the aforementioned vehicle. The dispatch management system according to claim 2.

Images