Demand prediction support system, demand prediction support program, and recording medium on which demand prediction support program is recorded

Abstract

To provide a demand prediction support system for calculating a prediction value for correcting the demand prediction by reflecting the influence of holding an event to more accurately predict the demand of transportation means.SOLUTION: A demand prediction support system 1 specifies an event influence ratio from an attribute of an event, and specifies an event influence period from the attribute of the event and a scheduled date and time of the event. The demand prediction support system further specifies transportation means that arrives at and departs from a prescribed station in the specified event influence period, and calculates a prediction value for correcting the demand prediction of the transportation means from the specified event influence ratio. As a result, the event influence ratio as a standard for calculating the prediction value or the event influence period for reflecting the event influence ratio varies according to the attributes of the event, and the prediction value to be calculated and the transportation means reflecting the prediction value also vary accompanied by these variations. Therefore, by applying the calculated prediction value to the demand prediction, a more accurate demand prediction can be made according to the attributes of the event.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a demand forecasting support system for correcting a demand forecast for transportation means, a demand forecasting support program, and a recording medium on which the demand forecasting support program is recorded. [Background technology]

[0002] In recent years, yield management systems that automatically calculate operation (flight) plans and fares in accordance with predicted demand forecasts have been used in revenue management for transportation means such as railroads, buses, and airplanes. The yield management system automatically calculates operation (flight) plans based on predicted demand forecasts, for example, in order to supply an optimized amount of transportation means in accordance with increases and decreases in demand for the transportation means. It also automatically calculates fares based on predicted demand forecasts, for example, in order to maximize revenue per transportation means.

[0003] For example, Patent Document 1 describes a demand forecasting system that automatically extracts actual demand values ​​influenced by an event based on past actual demand values, and performs demand forecasting based on the actual demand values ​​influenced by the event. In this demand forecasting system, for example, an actual demand value whose difference from a reference value based on past actual demand values ​​exceeds a threshold value is found, and the found actual demand value is regarded as an actual demand value influenced by an event. Then, based on the actual demand value influenced by the event, predictions are made as to whether an event will occur and the effect of the event, which is the effect that the event will have on the actual demand value, and if an event occurs, demand forecasting is performed based on the event effect. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2019-168868 A Summary of the Invention [Problem to be solved by the invention]

[0005] In the demand forecasting system described in Patent Document 1, a demand actual value affected by an event is identified from a reference value based on past demand actual values, and the demand actual value affected by the event is compared with the reference value to predict whether or not an event will occur and the impact of the event on the demand forecast. Therefore, for events that occur periodically, such as an event that occurs on a specific day every year, the impact of the event can be reflected in the demand forecast. However, for events that are held irregularly or new events, it is difficult to even predict the occurrence of the event. Therefore, in this demand forecasting system, for events that are difficult to extract automatically as described above, information such as the schedule and scale of the event can be entered and registered in advance to reflect the impact of the event in the demand forecast. However, for example, if an administrator or user of the system has to enter information on an event that is held irregularly or a new event every time it is confirmed, it would require considerable effort, and it would be difficult to accurately grasp the impact of the event, making it difficult to accurately reflect the impact of the event in the demand forecast.

[0006] Therefore, the present invention aims to provide a demand forecasting support system, a demand forecasting support program, and a recording medium on which the demand forecasting support program is recorded, which calculates predicted values ​​to correct the demand forecast by reflecting the impact of an event being held, in order to more accurately forecast demand for transportation. [Means for solving the problem]

[0007] In order to solve the above problem, the demand forecast support system of the present invention is for correcting the demand forecast of a transportation means, and is characterized by comprising: an event information storage unit that stores event information, which is information about the event, including the scheduled date and time of the event; an event attribute storage unit that stores, for each attribute, event attribute information, which is information about the attributes of the event, including an event influence ratio that indicates the increase rate of demand for the transportation means due to the influence of the event and an event influence period that indicates a predetermined period for reflecting the event influence ratio; an event influence aggregation unit that identifies, based on the event information and the event attribute information, a specific influence ratio that is the event influence ratio corresponding to the attribute of the event, and a specific influence period that is the event influence period that corresponds to the attribute of the event and corresponds to the scheduled date and time; a transportation means identification unit that identifies an event target transportation means, which is the transportation means that departs and arrives at a predetermined stop during the specific influence period; and a prediction value processing unit that calculates a prediction value for correcting the demand forecast of the event target transportation means from the specific influence ratio.

[0008] In the above invention, the event information may include a type and a scale of attendance of the event, and the event influence period may be set corresponding to the type and the scale of attendance of the event.

[0009] In the above invention, the event influence ratio may be set in accordance with the type and the scale of attendance.

[0010] In the above invention, the event influence period may comprise a first influence period, which is a predetermined period before the scheduled date and time of the event, and a second influence period, which is a predetermined period after the scheduled date and time of the event, and the event target transportation means may comprise a first event target transportation means, which is the transportation means that arrives at the stop during the specific influence period corresponding to the first influence period, and a second event target transportation means, which is the transportation means that departs from the stop during the specific influence period corresponding to the second influence period.

[0011] In the above invention, the event attribute information may include an estimated event duration anticipated for the event, and the second influence period may be a predetermined period after the estimated event duration has elapsed from the scheduled event date and time.

[0012] In the above invention, the event information storage unit may store the event information of a plurality of the events, and the event influence aggregator may identify the specific influence ratio and the specific influence period for the plurality of the events, and may have a specific influence ratio summation function for, when a first overlapping period is present in the specific influence period corresponding to the first influence period of the plurality of the events, setting the specific influence ratio in the first overlapping period to a first sum which is a total value obtained by adding up the specific influence ratios corresponding to the attributes of the events having the first overlapping period in the specific influence period, and for, when a second overlapping period is present in the specific influence period corresponding to the second influence period of the plurality of the events, setting the specific influence ratio in the second overlapping period to a second sum which is a total value obtained by adding up the specific influence ratios corresponding to the attributes of the events having the second overlapping period in the specific influence period.

[0013] In the above-mentioned invention, the upper limit of each of the first total value and the second total value may be 200 percent.

[0014] In order to solve the above problem, the demand forecasting support program of the present invention calculates a predicted value for correcting the demand forecast of a transportation means, and is characterized in that it stores event information, which is information about the event including the scheduled date and time of the event, and holds event attribute information, which is information about the attributes of the event, including an event influence ratio indicating the increase ratio of demand for the transportation means due to the influence of the event and an event influence period indicating a predetermined period for reflecting the event influence ratio, for each attribute, and identifies a specific influence ratio, which is the event influence ratio corresponding to the attribute of the event, and a specific influence period, which is the event influence period corresponding to the attribute of the event and corresponds to the scheduled date and time, based on the event information and the event attribute information, and identifies an event target transportation means, which is the transportation means that departs and arrives at a predetermined stop during the specific influence period, and calculates a predicted value for correcting the demand forecast of the event target transportation means from the specific influence ratio.

[0015] The demand forecasting support program may be recorded in a computer-readable recording medium. Effect of the Invention

[0016] According to the demand forecasting support system of the present invention, an event influence ratio corresponding to the attribute of a scheduled event and an event influence period corresponding to the attribute of the event and the scheduled date and time are specified based on the event information and the event attribute information. Then, in the specified event influence period, an event target transportation means, which is a transportation means departing and arriving at a specified station, is specified, and a predicted value for correcting the demand forecast of the event target transportation means is calculated from the specified event influence ratio. In other words, in the demand forecasting support system of the present invention, an event influence ratio is specified based on the attribute of the event specified for each event, and an event influence period is specified based on the attribute of the event specified for each event and the scheduled date and time. Then, among the transportation means departing and arriving at a specified station, a transportation means departing and arriving during the specified event influence period is specified as the event target transportation means, and a predicted value for correcting the demand forecast of the event target transportation means is calculated based on the specified event influence ratio. Therefore, the event influence ratio, which is the basis for calculating the predicted value, can be varied according to the attributes of the event, and the event influence period for reflecting the event influence ratio can be varied according to the attributes and scheduled date and time of the event. In addition, in accordance with these variations, the predicted value regarding the increase in demand for transportation means due to the influence of the event and the transportation means that reflects this predicted value can also be varied. In this way, by applying the predicted value calculated by the demand forecasting support system to the demand forecast, a more accurate demand forecast tailored to the attributes of the event can be achieved.

[0017] Moreover, according to the demand forecasting support system of the present invention, the event influence period is set according to the type of event and the scale of the audience. Therefore, the event influence period, which is a predetermined period for reflecting the event influence ratio, can be varied according to the type (genre) of the event and the scale of the audience of the event. For example, if the event genre is a live event, the demand for transportation means generally increases several hours before and after the performance time of the live event. On the other hand, for example, if the event genre is a marathon, in addition to the participants and spectators of the marathon who move to the venue on the day, there are many participants and spectators who stay at accommodation facilities near the venue the day before. Therefore, the demand for transportation means may increase not only several hours before and after the marathon event period, but also the day before and the day after. This may also occur depending on the scale of the audience of the event to be held. Therefore, according to the demand forecasting support system of the present invention, the event influence period is set according to the type of event and the scale of the audience, so that it is possible to respond to the special increase in demand as described above.

[0018] Furthermore, according to the demand forecasting support system of the present invention, the event influence ratio is set according to the type and scale of the event. Therefore, as described above, in addition to varying the event influence period according to the type and scale of the event, the event influence ratio can also be varied. Therefore, it is possible to calculate a forecast value that is more suitable for the attributes of the event.

[0019] According to the demand forecasting support system of the present invention, the event influence period is composed of a first influence period, which is a predetermined period before the scheduled date and time of the event, and a second influence period, which is a predetermined period after the scheduled date and time of the event, and the event target transportation means is composed of a first event target transportation means, which is a transportation means that arrives at the station during a specific influence period corresponding to the first influence period, and a second event target transportation means, which is a transportation means that departs from the station during a specific influence period corresponding to the second influence period. Therefore, the event target transportation means identified by the transportation means identification unit can be limited to transportation means that arrive at the station during a predetermined period before the scheduled date and time of the event, and transportation means that depart from the station during a predetermined period after the scheduled date and time of the event. Therefore, transportation means that depart from the station before the scheduled date and time of the event and transportation means that arrive at the station after the scheduled date and time of the event can be excluded from the event target transportation means that are targets for reflecting the event influence ratio.

[0020] Furthermore, according to the demand forecasting support system of the present invention, the event attribute information includes an expected event duration, and the second impact period is a predetermined period from the scheduled event date and time until the expected event duration has elapsed. Therefore, transportation means that depart from a station during the period from the scheduled event date and time until the expected event duration has elapsed can be excluded from targets for reflecting the event impact ratio. Therefore, transportation means that do not need to reflect the demand increase ratio due to the impact of the event can be excluded, thereby supporting more accurate demand forecasting.

[0021] Furthermore, according to the demand forecast support system of the present invention, the specific impact ratio summing function of the event impact aggregation unit sets the specific impact ratio in the first overlapping period as the first total value when the specific impact periods of the multiple events include a first overlapping period, and sets the specific impact ratio in the second overlapping period as the second total value when the specific impact periods of the multiple events include a second overlapping period. Therefore, when multiple events are held and there is a first overlapping period or a second overlapping period between the events, the specific impact ratio in the first overlapping period is set as the first total value, and the specific impact ratio in the second overlapping period is set as the second total value, and a forecast value for correcting the demand forecast can be calculated from the total value obtained by adding up the specific impact ratios corresponding to the attributes of each of the multiple events. Therefore, for example, when multiple events are held at close scheduled dates and locations, a forecast value can be calculated from the total value of the specific impact ratios corresponding to the attributes of the events, so that more accurate demand forecasting can be supported.

[0022] Furthermore, according to the demand forecasting support system of the present invention, the upper limit of each of the first total value and the second total value is 200 percent. Therefore, even when a large number of events are scheduled to be held at the same time and the specific influence ratios corresponding to the attributes of each of these events are added up to obtain the first total value or the second total value, the first total value or the second total value will not become a numerical value higher than necessary, and it is possible to prevent the forecast value calculated by the forecast value processing unit from becoming far removed from the actual increase rate of train demand due to the influence of the holding of the events.

[0023] Furthermore, according to the demand forecasting support program of the present invention and a recording medium having the demand forecasting support program recorded thereon, it is possible to achieve the same effects as those of the demand forecasting support system described above. [Brief description of the drawings]

[0024] [Figure 1]1 is a block diagram showing an outline of a demand forecasting support system according to an embodiment of the present invention, a yield management system in which the demand forecasting support system is incorporated, an event information collection system, and a reservation sales system. [Diagram 2] FIG. 2 is a diagram illustrating an example of event information in the demand forecast support system according to the embodiment of the present invention. [Diagram 3] 1 is a diagram illustrating an example of event attribute information of a demand forecast support system according to an embodiment of the present invention. [Figure 4] 1 is a diagram showing an example of event effect summary information of the demand forecast support system according to the embodiment of the present invention. FIG. [Diagram 5] 1A is a diagram showing an example of a method for calculating a first total value and a second total value in a specific influence ratio of a demand forecast support system according to an embodiment of the present invention. FIG. 1B is a diagram showing an example of a method for calculating a first total value in a specific influence ratio of a demand forecast support system according to an embodiment of the present invention when a first influence period includes a day before a scheduled date of an event. [Figure 6] FIG. 2 is a diagram showing an example of operation plan information of the demand prediction support system according to the embodiment of the present invention. [Figure 7] FIG. 2 is a diagram showing an example of target transport means information of the demand forecasting support system according to the embodiment of the present invention. [Figure 8] FIG. 2 is a diagram showing an example of forecast value change train information of the demand forecast support system according to the embodiment of the present invention. [Figure 9] FIG. 4 is a flow diagram illustrating an example of a calculation process of the demand forecast support system according to the embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0025] Next, an embodiment of a demand forecast support system according to the present invention will be described in detail with reference to the accompanying drawings.

[0026] As an embodiment of a demand forecasting support system according to the present invention, the schematic configurations of a demand forecasting support system 1 shown in FIG. 1 and a yield management system 100 incorporating the demand forecasting support system 1 will be described. In this embodiment, an example will be described in which the means of transportation is a Shinkansen train and the stops are Shinkansen stations. Note that the demand forecasting support system 1 may be one in which the means of transportation is a conventional train other than a Shinkansen and the stops are conventional train stations, or the means of transportation may be other means of transportation such as a bus where the stops are bus stops.

[0027] The demand forecast support system 1 is configured as a system incorporated as a part of a yield management system 100, for example, as shown in FIG. The yield management system 100 includes, for example, a processor such as a CPU (Central Processing Unit) that performs various processes according to a program, a ROM (Read Only Memory) that is a recording device that records electronic information such as a program for operating the CPU and can read the recorded electronic information, a RAM (Random Access Memory) that is a recording device that can read and write electronic information, and a HDD (Hard Disk Drive) that is a storage device that can read and write various other types of electronic information. If necessary, the yield management system 100 may also include input devices such as a keyboard, mouse, or touch panel, and output devices such as a liquid crystal display or printer. The CPU may be configured with other processors such as an MPU (Micro Processing Unit) or a DSP (Digital Signal Processor), and the HDD may be other storage devices such as an SSD (Solid State Drive).

[0028] The yield management system 100 includes at least a demand processing unit 101 that predicts and manages train demand, and a management system communication unit 102 that communicates with other systems independent of the yield management system 100. The demand processing unit 101 performs train demand prediction, and changes the train operation plan and fares based on the predicted demand prediction. The demand processing unit 101 performs demand prediction based on, for example, the train capacity and the predicted number of passengers, and optimizes the train operation plan based on the predicted demand prediction. Specifically, for example, an operation plan is created in which the number of train operations is adjusted so that the occupancy rate of each train is around 100 percent. In addition, the demand processing unit 101 performs demand prediction based on, for example, the number of train operations, the train capacity, and the predicted number of passengers, and changes the train fares based on the predicted demand prediction. Specifically, for example, when it is expected that the demand for a specific train is high, the sales price of the seats of the specific train is changed according to the demand. In addition, the demand processing unit 101 has a function of correcting a demand forecast value. Therefore, for example, a forecast value calculated by the demand forecast support system 1 can be further applied to the calculated demand forecast, thereby making it possible to perform a more accurate demand forecast.

[0029] The management system communication unit 102 is connected to a communication network such as the Internet, and is configured to be capable of communicating with the event information collection system 200, the advance sales system 300, and other devices, which are systems independent of the yield management system 100. The management system communication unit 102 transmits and receives various information via the communication network. Note that the communication network includes not only the Internet, but also any network, such as a network connected by a dedicated line, an intra-company LAN, an inter-company LAN, or a WAN, and the form of the communication line includes any form of communication, such as wired communication or wireless communication.

[0030] The demand forecast support system 1 is incorporated as a part of the yield management system 100, and is configured by sharing hardware such as a processor such as a CPU, a ROM, a RAM, and a HDD that are equipped in the yield management system 100. The demand forecast support system 1 includes a control unit 10 that controls various processes performed by the demand forecast support system 1, an event information storage unit 20 that stores event information, an event attribute storage unit 30 that holds event attribute information, an event influence aggregation unit 40 that identifies a specific influence ratio and a specific influence period of an event, a transport means identification unit 50 that identifies an event target transport means, a forecast value processing unit 60 that calculates a forecast value for correcting a demand forecast from the specific influence ratio, and a support system communication unit 70 that communicates with other systems and other devices.

[0031] As described above, the demand forecast support system 1 may be incorporated into a part of the yield management system 100 and configured by sharing the hardware such as a processor such as a CPU, a ROM, a RAM, and a HDD provided in the yield management system 100. Alternatively, the demand forecast support system 1 may be configured as an independent system equipped with these hardware separately, or may be configured to be realized as software in addition to the hardware configuration. When realized as software, for example, it is configured as a demand forecast support program that is stored and operated in a ROM and RAM in a computer equipped with hardware such as a processor such as a CPU, a ROM, a RAM, and a HDD. In this demand forecast support program, for example, the functions of the event information storage unit 20, the event attribute holding unit 30, the event influence aggregation unit 40, the transportation means identification unit 50, the forecast value processing unit 60, and the support system communication unit 70 are operated so that the computer performs the functions as an event information storage procedure, an event attribute holding procedure, an event influence aggregation procedure, a transportation means identification procedure, a forecast value processing procedure, and a communication procedure. Therefore, for example, the demand forecast support system 1 can be realized by recording the demand forecast support program on a recording medium such as a CD-ROM and reading the recording medium into a computer. The recording medium may be, for example, a CD-ROM, a flexible disk, an optical disk, or any other storage device such as an HDD or SSD.

[0032] Furthermore, the demand forecasting support system 1 can be realized by downloading the demand forecasting support program to a computer via a communication network such as the Internet, other than by reading the demand forecasting support program from a recording medium. Furthermore, the demand forecasting support system 1 may be configured to be realized as SaaS (Software as a Service) by an ASP (Application Service Provider) without downloading the demand forecasting support program to a computer. In this case, the demand forecasting support system 1 is provided as a cloud service configured to be accessible to the above-mentioned hardware configuration or software via a communication network.

[0033] The control unit 10 is realized by a processor such as a CPU equipped in the yield management system 100, and controls the event information storage unit 20, the event attribute retention unit 30, the event impact aggregation unit 40, the transportation means identification unit 50, the prediction value processing unit 60, and the support system communication unit 70, for example, using an OS (Operating System) program or an application program, part of which is stored in RAM, for execution by the CPU in accordance with a program stored in ROM.

[0034] The event information storage unit 20 stores event information, which is information related to events. Specifically, in an event information table 21 shown in FIG. 2, information related to events scheduled to be held in the future is stored as event information. The event information stored in the event information table 21 by the event information storage unit is acquired by information collected by the event information collection system 200. The event information may be collected by the event information collection system 200, or may be registered by an administrator or user of the demand forecast support system 1. In addition, the demand forecast support system 1 may be provided with an event information collection unit (not shown) that collects event information on its own, and the event information may be collected by this event information collection unit. In the event information table 21, event information relating to one specific event is stored in one record. Therefore, event information relating to multiple events is aggregated in multiple records in the event information table 21. Each record constituting the event information table 21 has, for example, an event date column 22, a station code column 23, an event name column 24, a genre name column 25, a customer scale column 26, an event time column 27, and a venue name column 28.

[0035] As an explanation of the records of the event information table 21, for example, taking the record 21A showing the event information of a specific event "Live A" as an example, the date of event 22 stores "2022-07-20" indicating July 20, 2022, which is the scheduled date of "Live A", and the station code column 23 stores "0000001", which is a station code indicating Shin-Osaka Station, among station codes that are identifiers preset for each stop station, assuming that the stop station corresponding to "Nagai Stadium", the scheduled venue of "Live A", is Shin-Osaka Station. The event name column 24 stores "Live A", which is the name of the event. The genre name column 25 stores "Live event" indicating the type (genre) of "Live A", and the audience size column 26 stores "Medium", which indicates the audience size of "Live A", among the indicators "Large", "Medium", or "Small" indicating the audience size of the event. The event time column 27 stores "17:00", which indicates the scheduled time of "Live A", and the venue name column 28 stores "Nagai Stadium", which indicates the name of the venue where "Live A" is scheduled to be held. The scheduled date and time of "Live A" is determined to be 17:00 on July 20, 2022, based on the scheduled date ("2022-07-20") stored in the event date column 22 and the scheduled time ("17:00") stored in the event time column 27.

[0036] Also, for example, taking the record 21B indicating the event information of another specific event "Baseball B" as an example, the date of event 22 stores "2022-07-20" indicating July 20, 2022, which is the scheduled date of "Baseball B", the station code column 23 stores "0000001" indicating Shin-Osaka Station, since the stop corresponding to "Kyocera Dome", the venue where "Baseball B" is scheduled to be held, is Shin-Osaka Station, and the event name column 24 stores "Baseball B", which is the name of the event. In addition, the genre name column 25 stores "Baseball", which indicates the type (genre) of "Baseball B", and the customer scale column 26 stores "Large", which is an index indicating the customer scale of "Baseball B". The time of event column 27 stores "15:00", which indicates the scheduled time of "Baseball B", and the venue name column 28 stores "Kyocera Dome", which indicates the name of the venue where "Baseball B" is scheduled to be held. The scheduled date and time of "Baseball B" is 15:00 on July 20, 2022, based on the scheduled date ("2022-07-20") stored in the date column 22 and the scheduled time ("15:00") stored in the time column 27.

[0037] In addition, the types (genres) of events stored in the genre name column 25 of the event information table 21 include, in addition to the above-mentioned "live event" and "baseball", for example, "idol" indicating a concert by an idol, "enka" indicating a live event by an enka singer, "pop music" indicating a live event of pop music by an artist, "classical" indicating a classical concert, "Western music" indicating a live event by a foreign artist, and "marathon", "soccer", "tennis", "rugby", "volleyball", "basketball", "badminton", "bicycle", "table tennis", "swimming", or "figure skating" indicating various sports competitions. In addition, types of events other than live events and sports competitions, such as "musical", "exhibition", "academic conference", or "cultural festival", are also included. Therefore, the types of events stored in the genre name column 25 include all types of events that are expected to increase the demand for trains due to the influence of the event. Similarly, the event information stored by the event information storage unit 20, in other words, the event information aggregated in the event information table 21, is also aggregated with respect to all events.

[0038] The event attribute storage unit 30 stores event attribute information, which is information related to the attributes of an event. Specifically, in the event attribute table 31 shown in FIG. 3, for example, an event influence ratio indicating the increase ratio of demand for transportation means due to the influence of the event and an event influence period indicating a predetermined period for reflecting the event influence ratio are stored as event attribute information. In the event attribute table 31, event attribute information set in accordance with the type (genre) of the event and the scale of the customer gathering is stored in one record for each attribute, and a plurality of pieces of event attribute information are aggregated. In addition, each record constituting the event attribute table 31 includes, for example, a genre name column 32, a scale of the customer gathering column 33, an event influence ratio column 34, an expected event duration column 35, and an event influence period column 36. In addition, the event influence period column 36 includes a first influence period column 36A and a second influence period column 36B, and the first influence period column 36A includes a previous night column 36C. The event attribute information may be preset by an administrator or user of the demand forecast support system 1 or the yield management system 100, or may be set each time the demand forecast support system 1 is used. In addition, the administrator or user of the demand forecast support system 1 or the yield management system 100 can also change the event attribute information at will.

[0039] As an example of the records in the event attribute table 31, for example, record 31A showing the attributes of a specific event is stored in the genre name column 32, which stores "live event" indicating the type (genre) of the event, and in the scale of attendance column 33, "large" which is an index selected from the indices "large", "medium", and "small" showing the scale of attendance of the event is stored. Therefore, it can be seen that record 31A is event attribute information related to the attributes of an event whose type (genre) of the event is "live event" and whose scale of attendance of the event is "large" among the attributes of a plurality of events. The type (genre) of the event stored in the genre name column 32 is the same as that stored in the genre name column 25 of the event information table 21, and the index showing the scale of attendance of the event stored in the scale of attendance column 33 is the same as that stored in the scale of attendance column 26 of the event information table 21. Therefore, the event information table 21 and the event attribute table 31 can specify records corresponding to each other by matching the type (genre) of the event and the scale of attendance of the event as a set.

[0040] The event impact ratio column 34 stores an event impact ratio indicating the increase ratio of train demand due to the impact of the event. The event impact ratio is set according to the type (genre) and scale of the event. In the record 31A, "100" is stored as the event impact ratio when the type (genre) of the event is "live event" and the scale of the event is "large". "100" in the event impact ratio column 34 of this record 31A indicates that the demand for trains is expected to increase by an additional 100 percent from the normal demand forecast. The expected event duration column 35 stores the expected event duration (hours) estimated from the type (genre) and scale of the event. In the record 31A, "3" is stored, indicating that the expected event duration (hours) estimated for the event with the type (genre) of "live event" and the scale of the event is "large" is 3 hours.

[0041] The event influence period column 36 stores an event influence period indicating a predetermined period for reflecting the event influence ratio. This event influence period is also set corresponding to the type (genre) and scale of the event. The event influence period column 36 has a first influence period column 36A and a second influence period column 36B. The first influence period column 36A stores a first influence period, which is a predetermined period before the scheduled date and time of the event, among the event influence periods. The second influence period column 36B stores a second influence period, which is a predetermined period after the scheduled date and time of the event and is a predetermined period after the expected event period has elapsed from the scheduled date and time of the event. In the record 31A, as the event influence period when the type (genre) of the event is "live event" and the scale of the event is "large", the first influence period column 36A stores "5" indicating that the first influence period (time) is 5 hours, and the second influence period column 36B stores "2" indicating that the second influence period (time) is 2 hours.

[0042] From the above, record 31A shows that if the type (genre) of the event is "live event" and the scale of the event's attendance is "large," then there is a first impact period of 5 hours before the scheduled date and time of the live event, then the live event is held for 3 hours afterwards, and then there is a second impact period of 2 hours after the end of the live event. In this case, the first and second impact periods are the predetermined periods for reflecting the event impact rate, and the demand for trains is expected to increase by an additional 100 percent from the normal demand forecast.

[0043] The event influence period column 36 also includes an overnight stay column 36C. The overnight stay column 36C stores the number of days of the overnight stay if an overnight stay is expected to start from the day before the scheduled date of the event. The number of days stored in this overnight stay column 36C is a predetermined period before the scheduled date and time of the event, and is therefore included in the first influence period of the event influence period. Specifically, for example, if an overnight stay is expected to start from the day before the scheduled date of the event, "1" is stored, indicating that the number of days of the overnight stay is one day, and the first influence period includes the period from midnight to midnight on the day before the scheduled date of the event. In record 31A, if the type (genre) of the event is "live event" and the scale of the event to be attended is "large," the field is left blank since an overnight stay is not anticipated.

[0044] As an example of the records in the event attribute table 31, record 31D showing the attributes of a specific event is stored in the genre name column 32, and in the scale of customer attendance column 33, "large" is stored. Therefore, it can be seen that record 31D is event attribute information related to the attributes of an event whose type (genre) is "marathon" and whose scale of customer attendance is "large" among the attributes of a plurality of events. In addition, in the event impact ratio column 34, "200" is stored as the event impact ratio when the type (genre) of the event is "marathon" and the scale of customer attendance of the event is "large". "200" in the event impact ratio column 34 of this record 31D indicates that the demand for trains is expected to increase by an additional 200 percent from the normal demand forecast. In the expected duration column 35, "10" is stored, indicating that the expected duration (time) of the event expected from the type (genre) and scale of customer attendance of the event is 10 hours. In addition, when the type (genre) of the event is "marathon" and the scale of the event's attendance is "large", the first influence period column 36A stores "2" indicating that the first influence period (time) is two hours, and the second influence period column 36B stores "2" indicating that the second influence period (time) is two hours. Furthermore, in the overnight stay column 36C, when the type (genre) of the event is "marathon" and the scale of the event's attendance is "large", it is assumed that an overnight stay will start the day before the scheduled event, and "1" is stored indicating that the number of days for the overnight stay is one day.

[0045] From the above, record 31D shows that if the type (genre) of the event is "marathon" and the scale of the event's attendance is "large," then there is a first influence period of two hours before the scheduled date and time of the marathon, which is then held for 10 hours, and a second influence period of two hours after the end of the marathon. In addition, since the number of days of overnight stay is one day, the first influence period further includes the period from 0:00 to 24:00 on the day before the scheduled date of the event. In this case, these first and second influence periods are the predetermined periods for reflecting the event influence rate, and the demand for trains is expected to increase by an additional 200 percent from the normal demand forecast.

[0046] The event influence aggregator 40 identifies a specific influence ratio and a specific influence period based on the event information stored in the event information storage unit 20 and the event attribute information held in the event attribute holding unit 30. The specific influence ratio is an event influence ratio corresponding to the attribute of a specific event, and the specific influence period is an event influence period corresponding to the attribute of a specific event and the scheduled date and time of the event. In addition, the event influence aggregator 40 creates an event influence aggregation table 41, which is a temporary table that reflects the event attribute information in the event information, from the event information and the event attribute information (FIG. 4). For example, taking "Live A" in record 21A of event information table 21 as an example, since the type (genre) of "Live A" is "live event" and the scale of attendance for "Live A" is "medium", the event influence aggregator 40 identifies a record in event attribute table 31 in which the event type (genre) is "live event" and the scale of attendance is "medium". Then, record 31B in event attribute table 31 corresponds. As a result, since the event influence ratio of record 31B is "70" stored in event influence ratio column 34, it can be identified that the specific influence ratio corresponding to the attribute of "Live A" is 70 percent.

[0047] In addition, for the specific influence period corresponding to the attribute and scheduled date and time of "Live A", it is specified that the first influence period is 3 hours from "3" in the first influence period column 36A of record 31B, and the second influence period is 1 hour from "1" in the second influence period column 36B. Furthermore, it is specified that the scheduled time of "Live A" is 17:00 from the start time column 27 of record 21A of the event information table 21, and that the expected event period (hours) expected for "Live A" is 2 hours from "2" in the expected event period column 35 of record 31B of the event attribute table 31. Furthermore, it is specified that the specific influence period corresponding to the first influence period among the specific influence periods of "Live A" is the three hours before 17:00, which is the scheduled time of "Live A", that is, from 14:00 to 17:00 (FIG. 5). Furthermore, from the above-mentioned second impact period, scheduled event time, and expected event duration (hours), it can be determined that the specific impact period of "Live A" that corresponds to the second impact period is the two hours after 7:00 p.m., which is after the expected event duration (hours) has elapsed from the scheduled event time of "Live A" (after two hours of the expected event duration (hours) have elapsed from 5:00 p.m.), i.e., from 7:00 p.m. to 9:00 p.m. In this manner, a specific impact ratio, which is an event impact ratio corresponding to the attributes of "Live A", and a specific impact period, which is an event impact period corresponding to the attributes and scheduled date and time of "Live A", are identified.

[0048] Also, taking "Baseball B" of record 21B as an example, since its type (genre) is "baseball" and its customer scale is "large", record 31E corresponds to it in the event attribute table 31. Then, for "Baseball B", if the specific impact rate and specific impact period are determined in the same manner as for "Live A" described above, it can be determined that the specific impact rate is 30 percent, the specific impact period corresponding to the first impact period is the two hours before 15:00, which is the scheduled time for "Baseball B", i.e., from 13:00 to 15:00, and further, the specific impact period corresponding to the second impact period is the two hours after the expected duration (hours) of three hours has elapsed from 15:00, which is the scheduled time for "Baseball B", i.e., from 18:00 to 20:00. In this manner, a specific influence ratio, which is an event influence ratio corresponding to the attribute of "Baseball B", and a specific influence period, which is an event influence period corresponding to the attribute of "Baseball B" and the scheduled date and time of the event, are identified. For "Live A" and "Baseball B," the overnight stay column 36C is blank and therefore not specified, but for example, in record 31D, as described above, if the type (genre) of the event is "marathon" and the scale of the event's attendance is "large," an overnight stay is expected, and "1" indicating the number of days for the overnight stay column 36C is stored. In this case, in addition to the specific influenced period corresponding to the first influenced period described above, the period from 0:00 to 24:00 on the day before the event is scheduled to be held is further specified as the specific influenced period corresponding to the first influenced period.

[0049] The event influence aggregator 40 also has a specific influence ratio summing function. For example, when there is a first overlapping period in a specific influence period corresponding to the first influence period of a plurality of events, this specific influence ratio summing function sets the specific influence ratio in the first overlapping period to a first sum which is a total value obtained by adding up the specific influence ratios corresponding to the attributes of each event having the first overlapping period in the specific influence period. Specifically, for example, the above-mentioned "Live A" and "Baseball B" are scheduled to be held on July 20, 2022, and the respective specific influence periods (first influence period and second influence period) are shown in FIG. 5(A). In this case, "Live A" and "Baseball B" overlap for one hour from 14:00 to 15:00 in the specific influence periods corresponding to each other's first influence periods. This one hour is the first overlap period 47, which is an overlapping period in the specific influence periods corresponding to the first influence periods of multiple events. Then, when there is a first overlap period 47, the specific influence ratio summing function of the event influence aggregation unit 40 sets the specific influence ratio in the first overlap period 47 to 100%, which is the sum of the specific influence ratio of 70% corresponding to the attribute of "Live A" having the first overlap period 47 in the specific influence period and the specific influence ratio of 30% corresponding to the attribute of "Baseball B" having the first overlap period 47 in the specific influence period. Therefore, the first total value in this case is 100 percent, which is the specific impact percentage in the first overlap period 47.

[0050] In addition, the specific impact ratio summation function of the event impact aggregation unit 40, for example, when there is a second overlapping period in a specific impact period corresponding to the second impact periods of multiple events, sets the specific impact ratio in the second overlapping period to a second sum, which is the total value obtained by adding up the specific impact ratios corresponding to the attributes of each event that has a second overlapping period in the specific impact period. For example, in the above-mentioned example of "Live A" and "Baseball B", one hour from 19:00 to 20:00 overlaps in the specific influence periods corresponding to the second influence periods of each event. This one hour is the second overlap period 48, which is a period of overlap in the specific influence periods corresponding to the second influence periods of multiple events. If there is a second overlap period 48, the specific influence ratio summing function of the event influence aggregator 40 sets the specific influence ratio in the second overlap period 48 to 100%, which is the sum of the specific influence ratio of 70% corresponding to the attribute of "Live A" having the second overlap period 48 in the specific influence period and the specific influence ratio of 30% corresponding to the attribute of "Baseball B" having the second overlap period 48 in the specific influence period. Therefore, the second sum value in this case is 100%, which is the specific influence ratio in the second overlap period 48. The upper limits of the first total value and the second total value can be arbitrarily determined. For example, since a certain time interval is required between trains that are operated, the upper limits of the first total value and the second total value are preferably 200 percent. In addition, when many events are scheduled to be held at the same time, if the total value obtained by adding up the specific influence ratios corresponding to the attributes of each of these many events is set as the first total value or the second total value, it is preferable that the upper limits of the first total value and the second total value are 200 percent so that the first total value or the second total value does not become an unnecessarily high value. This 200 percent indicates that the demand for trains is expected to increase by an additional 200 percent from the normal demand forecast, in other words, that there is three times the demand of the normal demand forecast.

[0051] Further, for example, taking the example of "Live C" in record 21C and "Marathon D" in record 21D in the event information table 21, "Live C" in record 21C is scheduled to be held on July 22, 2022, and "Marathon D" in record 21D is scheduled to be held on July 23, 2022. Since the event type (genre) of "Marathon D" is "Marathon" and the scale of the event to be attracted is "Large", the event attribute information corresponds to record 31D in the event attribute table 31. Here, "1" is stored in the overnight stay column 36C of record 31D, indicating the number of days of the overnight stay, assuming that an overnight stay is expected. Therefore, in addition to the first influence period of the scheduled date of "Marathon D" (July 23, 2022), the specific influence period corresponding to the first influence period of "Marathon D" is specified as a specific influence period corresponding to the first influence period, the first influence period from 0:00 to 24:00 on July 22, 2022, the day before the scheduled date of "Marathon D" (July 23, 2022). On the other hand, the specific influence period corresponding to the first influence period of "Live C" is specified as 15:00 to 20:00 on July 22, 2022 from the event information and event attribute information. Then, as shown in FIG. 5(B), "Live C" and "Marathon D" overlap for five hours from 15:00 to 20:00 in the specific influence periods corresponding to each other's first influence periods, and these five hours become the first overlap period 49. In this case, the specific impact ratio summing function of the event impact aggregator 40 determines 300 percent, which is the sum of 100 percent, which is the specific impact ratio corresponding to the attribute of "Live C" having the first overlapping period 49 in the specific impact period, and 200 percent, which is the specific impact ratio corresponding to the attribute of "Marathon D" having the first overlapping period 49 in the specific impact period. Here, since 300 percent exceeds the upper limit of 200 percent of the first total value, the specific impact ratio in the first overlapping period 49 becomes the upper limit of 200 percent. Therefore, the first total value in this case is 200 percent, which is the specific impact ratio in the first overlapping period 49. In this manner, the event influence aggregator 40 identifies the specific influence ratio and the specific influence period.

[0052] Further, the event influence aggregation unit 40 creates an event influence aggregation table 41, which is a temporary table in which the event attribute information is reflected in the event information, from the event information and the event attribute information (FIG. 4). At this time, in the event influence aggregation table 41, the "scheduled date of event", "station code", "specific influence ratio", and "specific influence period" are aggregated as event influence aggregation information. Therefore, each record constituting the event influence aggregation table 41 has, for example, an event date column 42, a station code column 43, a specific influence ratio column 44, a specific influence period column 45, and a departure / arrival flag column 46. Furthermore, the specific influence period column 45 has a time From column 45A indicating the start time when the specific influence period is divided at a predetermined time, and a time To column 45B indicating the end time when the specific influence period is divided at a predetermined time. Note that the predetermined time may be a time divided according to the time when the specific influence ratio changes as shown in FIG. 4, or may be divided in one-hour units, and can be determined arbitrarily.

[0053] As an explanation of the records of the event influence aggregation table 41, for example, taking record 41A as an example, the event date column 42 stores "2022-07-20" indicating July 20, 2022, and the station code column 43 stores "0000001" which is the station code indicating Shin-Osaka Station, which is a stop station. In addition, the specific influence ratio column 44 stores "30" indicating 30%, assuming that the specific influence ratio corresponding to the attribute of "Baseball B" is 30%. Furthermore, the time From column 45A stores "2022-07-20 13:00:00" indicating 13:00:00 on July 20, 2022, assuming that the specific influence period corresponding to a part of the first influence period of the specific influence period of "Baseball B" starts from 13:00 on July 20, 2022. In addition, the time To column 45B stores "2022-07-20 13:59:59", indicating 13:59:59 on July 20, 2022, since the specific impact period of "Baseball B" that corresponds to part of the first impact period ends at 13:59:59 on July 20, 2022.

[0054] For example, taking record 41B as an example, the date column 42 stores "2022-07-20" indicating July 20, 2022, and the station code column 43 stores "0000001" which is a station code indicating Shin-Osaka Station, which is a stop station. In addition, the specific impact ratio column 44 stores "100" indicating 100%, assuming that the first total value, which is the total value of the specific impact ratio corresponding to the attribute of "Live A" and the specific impact ratio corresponding to the attribute of "Baseball B", is 100%. Furthermore, the time From column 45A stores "2022-07-20 14:00:00" indicating 14:00:00 on July 20, 2022, assuming that the specific impact period corresponding to the first overlap period 47 of the specific impact period of "Live A" and the specific impact period of "Baseball B" starts from 14:00 on July 20, 2022. In addition, the time To column 45B stores "2022-07-20 14:59:59", indicating 14:59:59 on July 20, 2022, since the specific impact period of "Live A" and the specific impact period of "Baseball B" that corresponds to the first overlap period 47 ends at 14:59:59 on July 20, 2022.

[0055] For example, taking record 41C as an example, the event date column 42 stores "2022-07-20" indicating July 20, 2022, and the station code column 43 stores "0000001" which is the station code indicating Shin-Osaka Station, which is a stop station. In addition, the specific impact ratio column 44 stores "70" indicating 70%, assuming that the specific impact ratio corresponding to the attribute of "Live A" is 70%, and further, the time From column 45A stores "2022-07-20 15:00:00" indicating 15:00:00 on July 20, 2022, assuming that the specific impact period corresponding to a part of the first impact period of the specific impact period of "Live A" starts from 15:00 on July 20, 2022. In addition, the time To column 45B stores "2022-07-20 16:59:59", indicating 16:59:59 on July 20, 2022, since the specific impact period of "Live A" that corresponds to part of the first impact period ends at 16:59:59 on July 20, 2022.

[0056] In addition, in the event influence aggregation table 41, in addition to the records corresponding to the first influence period such as the above-mentioned records 41A, 41B, and 43C, records corresponding to the second influence period such as records 41D and 41E are also aggregated in the same manner. Therefore, the event influence aggregation unit 40 has an identifier assignment function that assigns an identifier to each record in the event influence aggregation table 41 to identify whether the record corresponds to the first influence period or the second influence period. The identifier assignment function of the event influence aggregation unit 40 assigns a "departure / arrival flag" as an identifier. The departure / arrival flag includes an "arrival" flag that is assigned to a record corresponding to the first influence period and an "arrival" flag that is assigned to a record corresponding to the second influence period, and these are stored in the departure / arrival flag column 46 provided in the event influence aggregation table 41. Specifically, for example, an "arrival" flag is stored in the departure / arrival flag column 46 of record 41A corresponding to the first impact period, and a "departure" flag is stored in the departure / arrival flag column 46 of record 41D corresponding to the second impact period.

[0057] Records 41F and 41H are records corresponding to a part of the first affected period of "Marathon D" (the first affected period on the day before the scheduled event), record 41G is a record corresponding to the first overlapping period 49 between "Live C" and "Marathon D", and record 41J is a record corresponding to the first affected period of "Marathon D" (the first affected period on the scheduled event date). Therefore, in these records 41F, 41G, 41H, and 41J, an "arrival" flag is stored in the departure / arrival flag column 46. On the other hand, record 41I is a record corresponding to the second affected period of "Live C," and record 41K is a record corresponding to the second affected period of "Marathon D." Therefore, in these records 41I and 41K, the "depart" flag is stored in the departure / arrival flag column 46. In this manner, the event effect aggregation table 41 is created.

[0058] The transportation means identification unit 50 identifies trains that meet a predetermined condition from the event impact aggregation information, which is information aggregated in the event impact aggregation table 41, and the operation plan information, which is information on the train operation plan. The predetermined condition is, for example, "a train that departs from and arrives at a predetermined stop station during a specific impact period," and more specifically, "a train that arrives at a predetermined stop station during a specific impact period corresponding to a first impact period" and "a train that departs from a predetermined stop station during a specific impact period corresponding to a second impact period." The predetermined stop station is a stop station identified by a station code stored in the station code column 43 of the event impact aggregation table 41 in the event impact aggregation information, and is a stop station corresponding to the planned location of the event, more specifically, a stop station closest to the planned location of the event. In addition, trains that depart and arrive at specified stations during a specific impact period are event-subject transportation means, and among these event-subject transportation means, trains that arrive at specified stations during a specific impact period corresponding to a first impact period are first event-subject transportation means, and trains that depart from specified stations during a specific impact period corresponding to a second impact period are second event-subject transportation means. In this embodiment, since the transportation means is a train (Shinkansen) as an example, the event target transportation means will be described below as the "event target train", the first event target transportation means will be described as the "first event target train", and the second event target transportation means will be described as the "second event target train".

[0059] The operation plan information used by the transport means specification unit 50 is, for example, information collected in an operation plan table 51 shown in Fig. 6. In the operation plan table 51, for example, the "operation date" of the train, the "service number" of the train number, the "station code" indicating the station where the train stops, the "arrival time" of the train at the station, and the "departure time" of the train at the station are collected as operation plan information. Therefore, each record constituting the operation plan table 51 has, for example, an operation date column 52, a service number column 53, a station code column 54, an arrival time column 55, and a departure time column 56. The station code stored in the station code column 54 is the same as the station code in the station code column 23 in the event information table 21 or the station code in the station code column 43 in the event influence aggregation table 41 . The operation plan information used by the transport means identification unit 50 may be stored in the yield management system 100, or may be stored by the transport means identification unit 50 having an information storage function. Also, for example, the transport means identification unit 50 may have an operation plan information creation function, and the operation plan information may be created by this operation plan information creation function. In this case, for example, the transport means identification unit 50 extracts or aggregates necessary information from information on train operation such as train information and stop station information held by the railway company, and creates an operation plan table 51 as operation plan information.

[0060] As an explanation of the records of the operation plan table 51, for example, in the record 51A, the operation date column 52 stores "2022-07-20" indicating that the train operation date is July 20, 2022, and the service number column 53 stores "1" indicating the train number of the train. In addition, the station code column 54 stores the station code "0000001" indicating Shin-Osaka Station, which is a stop station of the train. Furthermore, the arrival time column 55 stores "2022-07-20 13:50:00" indicating that the arrival time of the train at Shin-Osaka Station is 13:50:00 on July 20, 2022, and the departure time column 56 stores "2022-07-20 14:05:00" indicating that the departure time of the train from Shin-Osaka Station is 14:05:00 on July 20, 2022, is 14:05:00.

[0061] In addition, as an explanation of the records of the operation plan table 51, for example, in the record 51B, the operation date column 52 stores "2022-07-20" indicating that the operation date of the train is July 20, 2022, and the train number column 53 stores "1" indicating the train number of the train. In addition, the station code column 54 stores the station code "0000002" indicating Kanazawa Station, which is a stop station of the train. Furthermore, the arrival time column 55 stores "2022-07-20 14:25:00" indicating 14:25:00 on July 20, 2022, assuming that the arrival time of the train at Kanazawa Station is 14:25:00 on July 20, 2022. In addition, "(null)" indicating a blank is stored in the departure time column 56. In this case, "(null)" indicates that the train is not available from Kanazawa Station, for example, because the train does not depart from Kanazawa Station with passengers on board.

[0062] Next, a method for identifying an event target train by the transportation means identification unit 50 will be described. The transport means identification unit 50 identifies records corresponding to the event target trains from the operation plan table 51, and creates an event target transport means table 57, which is a temporary table that aggregates target transport means information, which is information related to the event target trains (Figure 7). The transportation means identification unit 50 first compares the scheduled event date in the event date column 42 in the event influence aggregation table 41 with the operation date in the operation date column 52 in the operation plan table 51, and identifies trains whose dates match from among the records in the operation plan table 51. Specifically, for example, when the scheduled event date in the event date column 42 in the event influence aggregation table 41 is "2022-07-20" indicating July 20, 2022, the operation date column 52 is identified from among the records in the operation plan table 51. The operation plan table 51 in FIG. 7 shows a state in which the operation date column 52 is identified as "2022-07-20".

[0063] Next, from among the records identified as trains with matching dates, the station code in the station code column 43 in the event impact aggregation table 41 is compared with the station code in the station code column 54 in the operation plan table 51 to identify trains that will stop at specified stations. Specifically, in the case of the event impact aggregation table 41, for example, the station code column 43 is "0000001" in all records, so records in the operation plan table 51 whose station code column 54 is "0000001" are identified. In the operation plan table 51 shown in Fig. 7, for example, records 51A, 51C, 51D, and 51F correspond to this.

[0064] Next, from among the records identified as trains stopping at a predetermined stop station, trains whose arrival time and departure time fall within the specific impact period are identified based on the specific impact period indicated by the times stored in the time From column 45A and the time To column 45B of the specific impact period column 45 of the event impact aggregation table 41, the departure and arrival flags stored in the departure and arrival flag column 46 of the event impact aggregation table 41, and the arrival and departure times stored in the arrival time column 55 and the departure time column 56 of the operation plan table 51. More specifically, trains whose arrival time falls within the specific impact period for which an "arrival" flag is assigned in the departure and arrival flag column 46 of the event impact aggregation table 41, and trains whose departure time falls within the specific impact period for which a "departure" flag is assigned in the departure and arrival flag column 46 of the event impact aggregation table 41 are identified. The trains identified at this time are the event target trains. Among these event target trains, trains whose arrival time falls within the specific impact period to which an "arrival" flag has been assigned are trains that arrive at a specified stop station during the specific impact period corresponding to the first impact period, and are first event target trains. Also, among the event target trains, trains whose departure time falls within the specific impact period to which a "departure" flag has been assigned are trains that depart from a specified stop station during the specific impact period corresponding to the second impact period, and are second event target trains.

[0065] Specifically, for example, in the record 41A of the event influence aggregation table 41, since the time From column 45A is "2022-07-20 13:00:00" and the time To column 45B is "2022-07-20 13:59:59" in the record 41A, the specific influence period is from 13:00:00 to 13:59:59 on July 20, 2022. Also, in the record 41A, "arrival" is stored in the departure / arrival flag column 46. Therefore, from among the records in the operation plan table 51, a train whose arrival time is included in the specific influence period between 13:00:00 and 13:59:59 on July 20, 2022 is identified. In the operation plan table 51 shown in FIG. 6, for example, the train in record 51A corresponds. Therefore, the train in record 51A is the first event target train.

[0066] Also, for example, taking record 41B of the event influence aggregation table 41 as an example, in record 41B, since the time From column 45A is "2022-07-20 14:00:00" and the time To column 45B is "2022-07-20 14:59:59", the specific influence period is from 14:00:00 to 14:59:59 on July 20, 2022. Also, in record 41B, "arrival" is stored in the departure / arrival flag column 46. Therefore, from among the records in the operation plan table 51, a train whose arrival time is included in the specific influence period, that is, from 14:00:00 to 14:59:59 on July 20, 2022, is identified. In the operation plan table 51 shown in FIG. 6, for example, the train in record 51C corresponds. Therefore, the train in record 51C is also a first event target train. In this manner, the first event target train is identified, and similarly, the second event target train is identified, thereby making it possible to identify the event target trains.

[0067] Then, the transport means specification unit 50 extracts records corresponding to the specified first event target train and second event target train from the operation plan table 51, and creates an event target transport means table 57 by aggregating necessary information (FIG. 7). In the event target transport means table 57, for example, the "operation date" of the train, the "service number" of the train number, the "station code" indicating the train's stop station, the "arrival time" of the train at the stop station, and the "departure time" of the train at the stop station are aggregated as target transport means information related to the first event target train and the second event target train from the operation plan information in the operation plan table 51. Therefore, each record constituting the event target transport means table 57 has an operation date column 52, a service number column 53, a station code column 54, an arrival time column 55, and a departure time column 56, similar to the operation plan table 51. Furthermore, when creating the event target transport means table 57, the transport means specification unit 50 refers to the specific impact ratio from the event impact aggregation information of the event impact aggregation table 41, and aggregates the specific impact ratio corresponding to each record of the event target transport means table 57. For this reason, the event target transport means table 57 is provided with a specific impact ratio column 58, and the specific impact ratio corresponding to each record is stored in this specific impact ratio column 58.

[0068] As an explanation of the records in the event target transportation means table 57, for example, record 57A will be taken as an example. Record 57A corresponds to record 41A in the event influence aggregation table 41 and record 51A in the operation plan table 51. More specifically, the train shown in record 57A arrives at Shin-Osaka Station, which is the nearest Shinkansen stop to the scheduled venue of "Baseball B" (Kyocera Dome), during a specific influence period (from 13:00:00 to 13:59:59 on July 20, 2022) corresponding to a part of the first influence period of "Baseball B." Therefore, in record 57A, "2022-07-20" is stored in the operation date column 52, indicating that the train's operation date is July 20, 2022, "1" is stored in the train number column 53, the station code column 54 stores the station code "0000001" indicating Shin-Osaka Station, which is the stop station, and "2022-07-20 13:50:00" is stored in the arrival time column 55, indicating the arrival time at Shin-Osaka Station, which is July 20, 2022, at 13:50:00. On the other hand, "(null)" is stored in the departure time column 56. This indicates that although the departure time of the train from Shin-Osaka Station is 14:05:00 (FIG. 6), this departure time is not included in the specific impact period to which the "departure" flag is attached. Therefore, in record 57A, the departure time of the train is not extracted as unnecessary information. Also, in record 57A, the arrival time is stored only in the arrival time column 55 out of the arrival time column 55 or the departure time column 56. Therefore, it can be seen that the train in record 57A is the first event target train arriving at Shin-Osaka Station. Also, in record 57A, since this record corresponds to record 41A in the event effect aggregation table 41, "30" is stored in the specific effect proportion column 58, indicating that the specific effect proportion is 30 percent.

[0069] In addition, as an explanation of the records of the event target transportation means table 57, for example, record 57B will be taken as an example. Record 57B corresponds to record 41B of the event influence aggregation table 41 and record 51C of the operation plan table 51. More specifically, the train shown in record 57B arrives at Shin-Osaka Station, which is the nearest Shinkansen stop to each of the scheduled venues of "Live A" (Nagai Stadium) and "Baseball B" (Kyocera Dome), during the specific influence period (from 14:00:00 to 14:59:59 on July 20, 2022) corresponding to the first overlap period 47 of "Live A" and "Baseball B". Therefore, in record 57B, "2022-07-20" is stored in the operation date column 52, indicating that the train's operation date is July 20, 2022, "2" is stored in the train number column 53, the station code column 54 stores the station code "0000001" indicating Shin-Osaka Station, which is the stop station, and "2022-07-20 14:10:00" is stored in the arrival time column 55, indicating the arrival time at Shin-Osaka Station, which is 14:10:00 on July 20, 2022. On the other hand, "(null)" is stored in the departure time column 56. This indicates that although the departure time of the train from Shin-Osaka Station is 14:20:00 (FIG. 6), this departure time is not included in the specific affected period to which the "departure" flag is attached. Therefore, in record 57B, the departure time of the train is not extracted as unnecessary information. Also, in record 57B, the arrival time is stored only in the arrival time column 55 out of the arrival time column 55 and departure time column 56, so it can be seen that the train in record 57B is the first event target train arriving at Shin-Osaka Station. Also, since record 57B corresponds to record 41B in event effect aggregation table 41, the specific effect proportion column 58 stores "100" indicating that the specific effect proportion is 100 percent. In this manner, the event target transport means table 57 shown in FIG. 7 is created.

[0070] The forecast value processing unit 60 calculates a forecast value for correcting the demand forecast of the event target train from the specific influence ratio. Specifically, the forecast value processing unit 60 extracts necessary information from the event target transportation means table 57 created by the transportation means specification unit 50, calculates a forecast value for each event target train based on the specific influence ratio, and aggregates the data in a forecast value table 61 shown in FIG. 8. In the forecast value table 61, for example, as forecast value change train information, information such as a serial number for each record (train) stored in the forecast value change train ID column 62, a "train number" of the train number stored in the train number column 64, a "station code" indicating the train's stop station stored in the station code column 65, and a "forecast ratio" indicating a forecast value for correcting the demand forecast stored in the forecast ratio column 66 is aggregated. Therefore, each record constituting the forecast value table 61 has, for example, a forecast value change train ID column 62, a forecast value change ID column 63, a train number column 64, a station code column 65, and a forecast ratio column 66. Furthermore, the station code column 65 has a departure station code column 65A and an arrival station code column 65B, and if the train departs from a specific stop station, the "station code" is stored in the departure station code column 65A, and if the train arrives at a specific stop station, the "station code" is stored in the arrival station code column 65B. Therefore, if a "station code" is stored in the departure station code column 65A, it indicates that the specific stop station is a departure station, and if a "station code" is stored in the arrival station code column 65B, it indicates that the specific stop station is an arrival station.

[0071] As an explanation of the records in the forecast value table 61, for example, taking record 61A as an example, the forecast value changed train ID column 62 stores "1" which indicates the serial number of record 61A among the serial numbers in the forecast value table 61, the departure station code column 65A stores "null" which indicates a blank, and the arrival station code column 65B stores "0000001" which is the station code indicating Shin-Osaka Station, which is a stop station. Therefore, it can be seen that this train is a train arriving at Shin-Osaka Station. In addition, in the forecast ratio column 66, "130" is stored as a forecast value for correcting the demand forecast, indicating that the "forecast ratio" is 130 percent. This "forecast ratio" is calculated by the forecast value processing unit 60, which, for example, refers to the specific impact ratio of each record in the event target transport means table 57, and calculates the "forecast ratio" by adding 100 percent to this specific impact ratio. Specifically, for example, since the record 61A corresponds to the record 57A in the event target transport means table 57, the specific impact ratio of the record 57A is referred to, and "100" is added to "30" stored in the specific impact ratio column 58 of the record 57A to calculate "130". In this way, the forecast value processing unit 60 calculates the "forecast ratio" in the forecast ratio column 66.

[0072] As an example of the records in the forecast value table 61, record 61B is taken as an example. In the forecast value changed train ID column 62, "2" is stored, which indicates the serial number of record 61B in the forecast value table 61. In the departure station code column 65A, "null" is stored, which indicates a blank. In the arrival station code column 65B, "0000001", which is the station code indicating Shin-Osaka Station, which is a stop station, is stored. Therefore, it can be seen that this train is also a train arriving at Shin-Osaka Station. Furthermore, the predicted ratio column 66 stores "200" indicating that the "predicted ratio" is 200 percent. This "200" is calculated by adding "100" to the "100" stored in the specific impact ratio column 58 of record 57B, by referencing the specific impact ratio of record 57B, since record 61B corresponds to record 57B in the event target transport means table 57. The "prediction ratio" calculated as described above is set to 100 percent of the normal demand forecast predicted by the yield management system 100. Therefore, by multiplying the demand forecast predicted by the yield management system 100 by this "prediction ratio," it is possible to correct the demand forecast to reflect the increase in train demand due to the holding of the event.

[0073] In addition, the forecast value change ID column 63 stores, for example, a unique identification number for each management company that manages the trains shown in each record in the forecast value table 61. In each of the records 61A, 61B, and 61C in the forecast value table 61, "1" is stored in the forecast value change ID column 63, and in the record 61D, "2" is stored in the forecast value change ID column 63. This indicates that the trains shown in the records 61A, 61B, and 61C are all managed by the same management company, and that the train shown in the record 61D is managed by a different management company. Therefore, for example, when it is necessary to provide forecast values ​​for correcting demand forecasts for each management company, forecast value change train information can be aggregated based on the identification number in the forecast value change ID column 63 and provided for each management company. In addition, the forecast value processing unit 60 transmits the forecast value change train information collected in the forecast value table 61 to the yield management system 100.

[0074] The support system communication unit 70 is configured to be capable of communicating with the event information collection system 200 and the advance sales system 300, which are systems independent of the yield management system 100, and transmits and receives various information via a communication network. In this embodiment, since the demand forecast support system 1 is configured as a part of the yield management system 100, the support system communication unit 70 is configured in common with the management system communication unit 102 of the yield management system 100. Note that, when the demand forecast support system 1 is configured as a system independent of the yield management system 100, the support system communication unit 70 itself is configured to be capable of communicating with the yield management system 100, the event information collection system 200, and the advance sales system 300, and transmits and receives various information via a communication network.

[0075] <Configuration of event information collection system 200> The event information collection system 200 is configured with a processor such as a CPU, a storage device such as a ROM, a RAM, and a HDD, and includes at least an information collection unit 201 that collects event information, an information processing unit 202 that outputs the collected event information as a file, and a collection system communication unit 203 that communicates with other systems independent of the event information collection system 200. The information collection unit 201 performs web scraping to collect information on event schedules from web pages published on the Internet, and stores the collected information as event information. In addition, the information processing unit 202 converts the information collected by the information collection unit 201 into a file in a format that can be imported by the demand forecast support system 1, such as a CSV (Comma Separated Values) format, and outputs the converted information. The collection system communication unit 203 transmits the file related to the event information output by the information processing unit 202 to the demand forecast support system 1 (yield management system 100) via a communication network.

[0076] <Configuration of reservation sales system 300> The reservation sales system 300 is configured with a processor such as a CPU, and a storage device such as a ROM, a RAM, and a HDD, and is equipped with at least a reservation sales management unit 301 that manages information related to the reservation sales of train seats, a product information generation unit 302 that generates product information for train seats based on the demand forecast predicted by the yield management system 100, and a reservation sales system communication unit 303 that communicates with other systems independent of the reservation sales system 300 and communication terminals such as mobile terminals of purchasers who wish to reserve and purchase train seats. The reservation sales management unit 301 reserves and sells train seats in response to reservation purchase requests from purchasers, and manages information related to reservation sales. The product information generation unit 302 receives the demand forecast predicted by the yield management system 100, and generates product information for selling train seats based on the received demand forecast, such as product information such as train operation schedules, seat types, and seat sales prices. The reservation sales system communication unit 303 receives demand forecasts predicted by the yield management system 100, and transmits and receives information necessary for reserving and purchasing train seats between the communication terminals of potential buyers, for example, via a communication network.

[0077] Next, a calculation process in which the demand forecast support system 1 calculates a forecast value and outputs the forecast value as a file will be described with reference to the flow diagram shown in FIG.

[0078] First, the event information collected by the event information collection system 200 is received via the support system communication unit 70 of the demand forecast support system 1 and is input into the event information storage unit 20 (S400). Then, the event information is stored by the event information storage unit 20 (S401). Next, the event influence aggregator 40 identifies a specific influence ratio and a specific influence period based on the event information stored by the event information storage unit 20 and the event attribute information held by the event attribute holder 30 (S402). Next, the event influence aggregator 40 reflects the event attribute information in the event information, and creates an event influence aggregation table 41, which is a collection of event influence aggregation information including the specific influence ratio and the specific influence period (S403).

[0079] Next, the transportation means identification unit 50 identifies an event target train (event target transportation means) that is a train that departs and arrives at a predetermined stop station during a specific impact period from the event impact summary information aggregated in the event impact summary table 41 and the train operation plan information (S404). At this time, the transportation means identification unit 50 identifies a first event target train (first event target transportation means) that is a train that arrives at a predetermined stop station during a specific impact period corresponding to a first impact period of the specific impact period, and a second event target train (second event target transportation means) that is a train that departs from a predetermined stop station during a specific impact period corresponding to a second impact period of the specific impact period. Next, the transportation means identification unit 50 creates an event target transportation means table 57 that is a collection of target transportation means information that is information about the identified first event target train and second event target train (S405).

[0080] Next, the forecast value processing unit 60 calculates a forecast value for correcting the demand forecast of the event target train (event target transportation means) from the specific impact ratio (S406). At this time, the forecast value processing unit 60 calculates a forecast value based on the specific impact ratio of the event target transportation means table 57 created by the transportation means identification unit 50. Next, the forecast value processing unit 60 extracts necessary information from the event target transportation means table 57 and creates a forecast value table 61 which is a collection of forecast value change train information including a "forecast ratio" for each event target train calculated based on the specific impact ratio (S407). The "forecast ratio" included in this forecast value table 61 is a forecast value for correcting the demand forecast predicted by the yield management system 100. Then, the forecast value processing unit 60 converts the forecast value change train information aggregated in the created forecast value table 61 into a file in a format that can be imported by the yield management system 100, and outputs it to the yield management system 100 (S408). In this manner, the demand forecast support system 1 calculates and outputs a forecast value for correcting the demand forecast predicted by the yield management system 100. The yield management system 100 reflects the forecast value output from the demand forecast support system 1 in the demand forecast to calculate a more accurate demand forecast value, and modifies the train operation plan and fares based on the new demand forecast value reflecting the forecast value. The yield management system 100 then transmits the train operation plan and fares based on the new demand forecast to the reservation and sales system 300.

[0081] Although the embodiment of the present invention has been described in detail above, the present invention is not limited to the above-mentioned embodiment, and various design modifications can be made to the present invention without departing from the scope of the claims.

[0082] For example, in the embodiment of the present invention, an example has been shown in which the event impact period and event impact ratio are set in accordance with the type (genre) of the event and the scale of the audience, but the event impact period and event impact ratio do not have to be set in accordance with the type (genre) and the scale of the audience, and may be set in accordance with other information, for example, the venue where the event is scheduled to be held.

[0083] Also, for example, in the embodiment of the present invention, an example in which the event influence period is composed of the first influence period and the second influence period has been shown, but the event influence period does not have to be divided into the first influence period and the second influence period. Similarly, in the embodiment of the present invention, an example in which the event target train (event target transportation means) is composed of the first event target train (first event target transportation means) and the second event target train (second event target transportation means) has been shown, but the event target train does not have to be divided into the first event target train and the second event target train, and in that case, the transportation means identification unit 50 identifies all trains that depart and arrive at a specified stop station during the specific influence period, in other words, all trains whose arrival time or departure time at a specified stop station falls within the specific influence period, as event target trains.

[0084] Furthermore, for example, in the embodiment of the present invention, when an overnight stay from the day before the scheduled event date is expected, an example was shown in which the first influence period includes the period from 0:00 to 24:00 on the date before the scheduled event date, in accordance with the number of days of the overnight stay before the event, but when an overnight stay after the scheduled event is expected, the second influence period may be configured to include the period from 0:00 to 24:00 on the date after the scheduled event date, in accordance with the number of days of the overnight stay after the event. In this case, for example, a "after stay" field is provided in the event influence period field 36, and the number of days of the overnight stay after the event is stored in this "after stay" field. Furthermore, when an overnight stay before the event or an overnight stay after the event is expected, the time included in the first influence period or the second influence period is not limited to 0:00 to 24:00, and any time can be set. In addition, in an embodiment of the present invention, an example was shown in which the second impact period is a specified period after the expected event period has elapsed from the scheduled event date and time, but the second impact period does not have to be after the expected event period has elapsed as long as it is after the scheduled event date and time.

[0085] Furthermore, for example, in the embodiment of the present invention, a demand forecasting support system 1 that calculates predicted values ​​corresponding to multiple events scheduled to be held at the same time has been described as an example, but the demand forecasting support system 1 may also be configured so that, for one specific event, the event information storage unit 20 stores event information, the event influence aggregation unit 40 identifies a specific influence proportion and a specific influence period, and the demand forecasting support system 1 calculates a predicted value for the event-target train based on the identified specific influence proportion and specific influence period. [Explanation of symbols]

[0086] 1. Demand forecast support system 10 Control section 20 Event information storage unit 21 Event Information Table 21A~21E Records (Event Information Table) 22 Event Date Column 23 Station code column 24 Event name column 25 Genre name column 26 Customer size column 27 Event Time Column 28 Venue Name 30 Event attribute storage unit 31 Event Attribute Table Records 31A to 31E (Event Attribute Table) 32 Genre name column 33 Customer size column 34 Event Impact Percentage Column 35 Expected event period 36 Event impact period column 36A First period of impact 36B Second impact period column 36C Overnight stay column 40 Event Impact Aggregation Department 41 Event Impact Aggregate Table Records 41A to 41K (Event Impact Aggregation Table) 42 Event Date Column 43 Station code column 44 Specific Impact Percentage Column 45 Specific impact period column 45A Time From column 45B Time To column 46 Arrival and Departure Flag Column 47 First overlapping period 48 Second overlapping period 49 First overlapping period 50 Transportation Means Specification Department 51 Operation plan table 51A~51F Records (Operation Plan Table) 52 Operation date column Flight No. 53 column 54 Station code column 55 Arrival time column 56 Departure time column 57 Event Target Transport Means Table Records 57A to 57D (Event target transport means table) 58 Specific Impact Percentage Column 60 Prediction value processing section 61 Prediction Table Records 61A to 61D (Prediction table) 62 Prediction value change train ID column 63 Prediction value change ID field Flight No. 64 column 65 Station code column 65A Departure station code field 65B Arrival station code column 66 Prediction ratio column 70 Support System Communication Department 100 Yield Management System 101 Demand processing section 102 Management System Communication Department 200 Event Information Collection System 201 Information Gathering Department 202 Information Processing Section 203 Collection System Communication Department 300 Reservation Sales System 301 Reservation Sales Management Department 302 Product information generation department 303 Reservation Sales System Communication Department

Claims

1. A demand forecast support system for correcting a demand forecast for a transportation means, comprising: an event information storage unit that stores event information that is information about the event including a scheduled date and time of the event; an event attribute storage unit that stores, for each attribute, event attribute information that is information on attributes of the event, including an event influence ratio that indicates an increase ratio of demand for the transportation means due to the influence of the event and an event influence period that indicates a predetermined period for reflecting the event influence ratio; an event influence aggregator that identifies a specific influence ratio, which is the event influence ratio corresponding to the attribute of the event, and a specific influence period, which is the event influence period corresponding to the attribute of the event and the scheduled event date and time, based on the event information and the event attribute information; A transportation means identification unit that identifies an event target transportation means that departs from and arrives at a predetermined station during the specific impact period; a prediction value processing unit that calculates a prediction value for correcting the demand prediction for the event target transportation means from the specific influence ratio.

2. The event information includes a type and a scale of attendance of the event, The event influence period is set according to the type and the scale of the event.

2. The demand forecast support system according to claim 1 .

3. The event influence ratio is set corresponding to the type and the scale of the customer gathering, 3. The demand forecast support system according to claim 2.

4. the event influence period is comprised of a first influence period which is a predetermined period before the scheduled event date and time, and a second influence period which is a predetermined period after the scheduled event date and time, The event-target transportation means includes a first event-target transportation means that is the transportation means that arrives at the stop during the specific impact period corresponding to the first impact period, and a second event-target transportation means that is the transportation means that departs from the stop during the specific impact period corresponding to the second impact period.

3. The demand forecast support system according to claim 1 or 2.

5. The event attribute information includes an expected holding period of the event, The second influence period is a predetermined period after the expected event period has elapsed from the scheduled event date and time.

5. The demand forecast support system according to claim 4.

6. the event information storage unit stores the event information of a plurality of the events; the event influence aggregator has a specific influence proportion summing function for identifying the specific influence proportion and the specific influence period for the plurality of events, and when there is a first overlapping period in the specific influence period corresponding to the first influence period of the plurality of events that is an overlapping period, setting the specific influence proportion in the first overlapping period to a first sum which is a sum of the specific influence proportions corresponding to the attributes of the events having the first overlapping period in the specific influence period, and when there is a second overlapping period in the specific influence period corresponding to the second influence period of the plurality of events that is an overlapping period, setting the specific influence proportion in the second overlapping period to a second sum which is a sum of the specific influence proportions corresponding to the attributes of the events having the second overlapping period in the specific influence period.

6. The demand forecast support system according to claim 5.

7. The upper limit of each of the first total value and the second total value is 200%.

7. The demand forecast support system according to claim 6.

8. A demand forecasting support program for correcting a demand forecast for a transportation means, comprising: storing event information that is information about the event including a scheduled date and time of the event; holding, for each attribute, event attribute information that is information on an attribute of the event, the event influence ratio indicating an increase ratio of the demand for the transportation means due to the influence of the event and an event influence period indicating a predetermined period for reflecting the event influence ratio; Based on the event information and the event attribute information, a specific influence ratio is identified as the event influence ratio corresponding to the attribute of the event, and a specific influence period is identified as the event influence period corresponding to the attribute of the event and to the scheduled event date and time, and Identifying an event target transportation means which is the transportation means departing and arriving at a predetermined stop during the specific impact period; Calculating a predicted value for correcting the demand forecast of the event target transportation means from the specific impact ratio; A demand forecasting support program characterized by:

9. A system in which the demand forecast support program according to claim 8 is recorded. A computer-readable recording medium comprising:

Images