Information processing device, information processing method, information processing system, and computer program

Abstract

To enable evaluation of an operation plan of a vehicle.SOLUTION: An information processing device generates first event interval delay time information indicating a delay time between a first event and a second event preceding the first event among a plurality of events in which departure or arrival at a plurality of stop positions with respect to at least one vehicle and a time of the departure or the arrival are determined, based on delay time information indicating a delay time of a third event preceding the first event, and generates delay time information indicating a delay time of the first event based on the first event interval delay time information and delay time information indicating a delay time of the second event.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an information processing device, an information processing method, an information processing system, and a computer program. [Background technology]

[0002] For railway companies and other organizations, delays in train schedules, which are planned train operations, are a serious problem that can result in reduced sales and increased costs such as fines. Therefore, it is desirable to create a timetable that is as robust as possible against delays. If it were possible to evaluate timetables using as little actual operation data as possible, such a timetable evaluation method would be highly valuable.

[0003] One method for evaluating timetables is to calculate the delay probability at each station. For example, there is a method that uses a Bayesian network technique to output the delay probability at each station. Another method is to treat arrival, departure, and passing through a station as events, and calculate the delay probability distribution of the event by inputting the delay probability distribution between the event preceding the event and the event in question (for example, the expected value of the delay time) and the required time between the events.

[0004] However, the delay probability distribution between events may change depending on the delay status of other events. To consider the change in the delay probability distribution between events using the above-mentioned method, a large amount of actual data is required, and it is difficult to evaluate the change in the delay time of the departing event due to the change in the delay probability distribution between events. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 2020-82920 [Patent Document 2] Japanese Patent Application Laid-Open No. 2015-3625 Summary of the Invention [Problem to be solved by the invention]

[0006] The embodiments of the present invention provide an information processing device, a method, an information processing system, and a computer program that enable evaluation of a vehicle operation plan. [Means for solving the problem]

[0007] The information processing device of this embodiment generates first inter-event delay time information representing the delay time between a first event and a second event preceding the first event among a plurality of events that define the departure or arrival of at least one vehicle at a plurality of stopping locations and the times of the departure or arrival, based on delay time information representing the delay time of a third event preceding the first event, and generates delay time information representing the delay time of the first event based on the first inter-event delay time information and delay time information representing the delay time of the second event. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a block diagram of a diagram evaluation device which is an information processing device according to a first embodiment. [Figure 2] FIG. 4 is a diagram showing an example of timetable information. [Figure 3] FIG. 10 is a diagram showing an example of inter-event delay time information. [Figure 4] FIG. 10 is a diagram showing an example of information on required inter-event time. [Figure 5] FIG. 10 is a diagram showing an example of acquiring the required time between events based on the slack time. [Figure 6] FIG. 1 is a diagram showing an overview of a travel time change model. [Figure 7] FIG. 10 is a diagram showing an example of a riding time proportional coefficient. [Figure 8] 10 is a flowchart illustrating an example of processing by a delay probability distribution calculation unit. [Figure 9] 10 is a flowchart illustrating an example of processing by a delay probability evaluation unit. [Figure 10] FIG. 10 is a diagram showing an example of event information. [Figure 11] A diagram showing event information in a graphical format. [Figure 12] FIG. 10 is a diagram showing a specific example of a process for generating a composite probability distribution. [Figure 13] FIG. 10 is a diagram showing an example of travel time change information. [Figure 14] FIG. 10 is a diagram showing an example of delay probability distribution information. [Figure 15] FIG. 10 is a diagram showing an example of display of evaluation information. [Figure 16] FIG. 10 is a diagram showing another example of displaying evaluation information. [Figure 17] FIG. 10 is a block diagram of a diagram evaluation device which is an information processing device according to a second embodiment. [Figure 18] FIG. 1 is a diagram showing the hardware configuration of an information processing apparatus according to each embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following embodiments, a train (car formation) including multiple cars will be described as an example, but the present invention can also be applied to a single-car train, a bus, a taxi, or other vehicle.

[0010] 1 is a block diagram of a diagram evaluation device 101 (hereinafter referred to as the device 101), which is an information processing device according to the first embodiment. The device 101 includes a diagram information input unit 110, an inter-event delay time information input unit 120, an inter-event required time input unit 130, a model information input unit 140, a delay probability distribution calculation unit (processing unit) 500, and an output unit 400.

[0011] This device 101 has multiple storage units, such as a timetable information storage unit 210, an inter-event delay time information storage unit 220, an inter-event required time storage unit 230, a model information storage unit 240, an event information storage unit 250, a delay probability distribution storage unit (delay time information storage unit) 300, and a ride time change information storage unit 310.

[0012] [Schedule information] The timetable information input unit 110 receives input operation of timetable information that defines a train operation timetable (timetable or operation plan) including multiple operations from a user who is the operator of the device 101, and acquires the timetable information.

[0013] The diagram information storage unit 210 stores the diagram information acquired by the diagram information input unit 110.

[0014] Each operation defined by the timetable information includes multiple events that define the stop locations (locations, stations) for departure, arrival, passing, etc., and the time, etc. The time of an event defined by the timetable is called the "scheduled time."

[0015] A single operation can be defined in various units, such as from departure from a depot to return to the depot, or from travelling from the starting station to the final station. Basically, one operation is assigned to one train, but there are also cases where one operation is assigned to multiple trains. In the following, we will assume that one train is assigned to one operation. A set of events belonging to the same train (i.e. operations assigned to the same train) is called a line.

[0016] Figure 2 shows an example of timetable information. Figure 2(A) shows an example of timetable information in a table format. Figure 2(B) shows the timetable information of Figure 2(A) in a graph format.

[0017] As shown in Fig. 2(A), the schedule information includes information on multiple events. The information on each event includes an event identifier (event identifier or event ID), the event time (scheduled time), the identifier of the line to which the event belongs, the location where the event will take place (occur), the type of event, etc.

[0018] In the illustrated example, the events occur only at stations, but they may also occur at other locations, such as block sections or train depots.

[0019] In the illustrated example, the event type is departure (departure) or arrival (arrival), but there may be other examples such as passing, overtaking, or waiting. A departure event will be described as a departure event, an arrival event as an arrival event, and a passing event as a passing event.

[0020] In the graph in Figure 2(B), the vertical axis represents location and the horizontal axis represents time. Events are shown as circles, and events belonging to the same route are connected by lines. The numbers inside the circles are the event identifiers (event IDs). Routes may include turnarounds or may start from intermediate stations.

[0021] Examples of event information included in the timetable information are not limited to those shown in Figures 2(A) and 2(B). For example, information on whether an event can depart early or information indicating one or more events (preceding events) preceding the event may be included. An event preceding another event means that the time (scheduled time) of the event is earlier than the time (scheduled time) of the other event. For example, a second event preceding a first event means that the time (scheduled time) of the second event is earlier than the time (scheduled time) of the first event. An early departure means that an event is executed earlier than the scheduled time. The preceding event refers to an event that is scheduled to occur earlier than the event in the same or another timeline. The range of preceding events may be defined in advance. For example, the event that occurs immediately before a certain event in the same timeline may be considered a preceding event. Alternatively, the preceding event may be an event that has a predetermined relationship with a certain event in a different timeline. For example, for an event departing at a station, the preceding events may be the most recent arrival event on the same route at the same station, or the most recent preceding departure event on another route at the same station. In this example, the preceding events for event 6 are event 5 on the same route and event 3 on a different route.

[0022] [Event delay information] The inter-event delay time information input unit 120 receives an input operation of inter-event delay time information, which is information indicating the delay time between events, from the user, and acquires the inter-event delay time information. The inter-event delay time information storage unit 220 stores the inter-event delay time information acquired by the inter-event delay time information input unit 120 .

[0023] The delay between events is the time difference between the scheduled time of the target event (target event) and the scheduled time of the event preceding the target event (preceding event). A preceding event here is an event that is scheduled earlier than the target event and that can directly affect the delay time of the target event (can become a delay propagation path). Events that can have an effect typically include events that belong to the same channel as the target event (for example, an event that takes place immediately before the target event) and events that belong to a different channel from the target event but occur at the same station (for example, an event that takes place immediately before the target event at the same station).

[0024] The inter-event delay time information may take various forms. For example, it may be a probability distribution of the delay time between events (delay probability distribution), a statistical quantity such as the mean value or variance of the delay time between events, or a parameter that defines the delay probability distribution. The type of inter-event delay time information may differ for each event. In this embodiment, a negative binomial distribution is assumed as the delay probability distribution between events, and an expected value is given as a parameter of the negative binomial distribution. That is, the expected value of the delay time between events is acquired by the inter-event delay time information input unit 120 as the inter-event delay time information. The negative binomial distribution obtained by assigning the acquired expected value as a parameter corresponds to the delay probability distribution between events.

[0025] FIG. 3 shows an example of inter-event delay time information. In this example, an expected delay time is stored for each pair of a target event (target event) and a preceding event that precedes the target event. The expected delay time between event 4 and event 5 is 30, which is larger than the delay times between other events. Therefore, the delay time of event 5 is expected to be large.

[0026] [Time required between events] The required inter-event time input unit 130 receives an input operation from the user regarding information (required inter-event time information) regarding the required time between two events (for example, the minimum time required between two events), and acquires the required inter-event time information. The required inter-event time storage unit 230 stores the required inter-event time information acquired by the required inter-event time input unit 130.

[0027] For example, there is the time required for travel between stations (required travel time) and the time required at a station (required stop time). There is also the time required for disembarking (required disembarking time), the time required for boarding (required boarding time), and the time required for safety checks (required check time). Safety checks are tasks performed by station staff to ensure safety after a train arrives at a station and before it departs. It may also be the time required between departure and arrival events that occur consecutively at the same station and belong to different routes (hereinafter referred to as the required train interval). The value of the required train interval may differ between routes or may differ from station to station. Other examples of the time required between events that belong to different routes include the time interval between arrival events or the time interval between departure events.

[0028] 4 shows an example of the information on the required inter-event time. Information indicating the required inter-event time is stored for each pair of a target event (target event) and a preceding event that precedes the target event.

[0029] Instead of information about the required time between events, the device 101 may acquire information about the allowable slack time (margin time) between events. The slack time is the amount of time that can be delayed relative to the required time between events (for example, the maximum amount of time that can be delayed). Specifically, the slack time is the value obtained by subtracting the required time between the two events from the difference between the scheduled times of the two events. In other words, the time between the scheduled times of two events (scheduled time interval) defined in the timetable corresponds to the time obtained by adding the required time between the two events to the slack time between the two events. Therefore, when the device 101 acquires slack time information, the required time between the events can be obtained by subtracting the slack time between the events from the difference between the scheduled times of the events defined in the timetable.

[0030] 5 shows an example of acquiring the required inter-event time for each event when the device 101 acquires information on slack time. The required inter-event time is obtained by subtracting the slack time from the scheduled time interval.

[0031] Furthermore, instead of acquiring inter-event delay time information, the device 101 may acquire distribution information on travel times between events (e.g., probability distribution or expected value of travel times, etc.) and distribution information on stop times (e.g., probability distribution or expected value of stop times, etc.). Alternatively, the device 101 may acquire distribution information on disembarking times (e.g., probability distribution or expected value, etc.), distribution information on boarding times (e.g., probability distribution or expected value, etc.), and distribution information on confirmation times (e.g., probability distribution or expected value, etc.). The device 101 can acquire inter-event delay time information based on the acquired information and the required time between events described above. For example, the expected value of travel delay time can be obtained by subtracting the required time between events (required travel time) from the expected value of travel time between events.

[0032] [Event Delay Time] The event delay time is defined based on the inter-event delay time and the required inter-event time described above. The event delay time is calculated by subtracting the difference in scheduled time between the target event and the preceding event from the sum of the delay time of the preceding event, the delay time between the target event and the preceding event (inter-event delay time), and the required time between the target event and the preceding event. As described above, the value obtained by subtracting the required time between events from the difference in scheduled time between events is called the slack time. When the sum of the delay time of the preceding event and the delay time between the target event and the preceding event is greater than the slack time, the delay time value becomes positive, and a delay occurs. Examples of information representing the event delay time (event delay time information) include statistical values ​​such as the probability distribution of the event delay time (event delay probability distribution) and the expected value of the event delay time. The event delay time information (event delay time information) is calculated by the delay probability distribution calculation unit 500, which will be described later.

[0033] The model information input unit 140 receives an input operation of information including a travel time change model (travel time change model information) from a user, and acquires the travel time change model information. The model information storage unit 240 stores the travel time change model information acquired by the inter-event required time input unit 130.

[0034] The ride time change model is a model that calculates or estimates the amount of change in ride time (the time it takes for passengers to board a train from a station) between a target departure event at a station and its immediately preceding arrival event at the same station. The ride time change model corresponds to an example of a model that calculates change amount information that represents the amount of change in delay time between a target event and a preceding event. The amount of change in ride time between a target departure event and its immediately preceding arrival event at the same station corresponds to an example of the amount of change in delay time between a target departure event and its immediately preceding arrival event at the same station.

[0035] Figure 6 shows an overview of the ride time variation model. Two routes (operations) 11 and 12 are shown. Circles represent events. Consider a variation model (function) of the ride time between arrival event j and departure event k at station Q on route 12. In this variation model, arguments are distribution information of the delay time of departure event i immediately preceding station Q on a different route 11 (delay time information of departure event i, e.g., expected value or probability distribution of the delay time of departure event i) and distribution information of delay time of arrival event j (delay time information of arrival event j, e.g., expected value or probability distribution of the delay time of departure event i). Specifically, for example, the argument is the difference between the delay time information of departure event i on a different route 11 and the delay time information of arrival event j immediately preceding station Q on the same route 12 (i.e., delay time information between events i and j). If the time interval between departure event i and arrival event j is defined as the train interval, this difference corresponds to the difference from the scheduled train interval on the same line.

[0036] This difference affects the amount of change in the boarding time between arrival event j and departure event k. The amount of change in the boarding time between arrival event j and departure event k affects the change in the number of passengers boarding at station Q.

[0037] It is assumed that the change in the ride time is proportional to the change in the interval (train interval) between departure event i and arrival event j. The flow rate of passengers boarding the train [persons / second] is constant, and the proportionality coefficient (hereinafter referred to as the ride time proportionality coefficient) λ corresponds to the change in the delay time between departure event k and arrival event j. k (≧0). The inter-event delay time between departure event k and arrival event j increases by the increment (change) of the ride time, and this increment is proportional to the change in the train interval by a proportional coefficient λ k It is assumed that the delay time is proportional to the above. One of the features of this embodiment is that the increment (amount of change) in the ride time is reflected in the delay time between events to calculate the distribution of the delay time of the originating event k (delay time information of the originating event k). A specific example of the ride time change model will be described later (see, for example, equations (1) to (3) etc. described later).

[0038] Figure 7 shows an example of the riding time proportional coefficient. In this example, the target pairs of departure and arrival events are the pair of events 3 and 2 and the pair of events 6 and 5. A value of the riding time proportional coefficient (=0.5) is set for the pair of events 6 and 5. In this example, there is no event on a different route that is scheduled earlier than event 3, so no riding time proportional coefficient is defined for the pair of events 3 and 2. The value of the riding time proportional coefficient may be determined in any way, for example, by performing a simulation and adopting a highly accurate value, or by determining it empirically based on actual operation data. Specific examples of methods for determining the value of the riding time proportional coefficient will be described later.

[0039] In the example of Figure 2 described above, since there is a departure event (event 3) at the same station on a different route and an arrival event (event 5) just before it on the same route for departure event 6, the travel time proportional coefficient λ6 between events 5 and 6 is defined as 0.5. As described in the explanation of Figure 3, the delay time of event 5 is expected to be larger than the other events, so the difference between the delay time of event 3 and the delay time of event 5, i.e., the inter-event delay time between events 3 and 5, is expected to be a positive value. Therefore, the change in travel time between events 5 and 6 calculated from the travel time change model is expected to be a positive value.

[0040] The delay probability distribution calculation unit 500 calculates delay time information (for example, the delay probability distribution or expected value of an event) representing the delay time of an event based on the information stored in the storage units 210 to 240. The delay probability distribution calculation unit 500 corresponds to a processing unit that calculates delay time information representing the delay time of an event. The delay probability distribution calculation unit 500 includes an event information creation unit 510, an evaluation order determination unit 515, and a delay probability evaluation unit 520. The processing of the delay probability distribution calculation unit 500 will be described below with reference to FIGS. 8 and 9.

[0041] FIG. 8 is a flowchart of an example of the processing of the delay probability distribution calculation unit 500. The event information creating unit 510 creates event information by treating each event included in each time slot in the timetable information as a target event (first event) (Step_A). The event information storage unit 250 stores the created event information for each event.

[0042] FIG. 10 shows an example of event information. The event information includes an event identifier, information on whether the event can be started early, and information on the preceding event. FIG. 11 shows the event information of FIG. 10 in a graph format. In the graph of FIG. 11, an event is connected to its preceding event by a solid line or an arc with an arrow. A preceding event in the same line is connected by a solid line, and a preceding event in another line is connected by an arc with an arrow.

[0043] The early departure permission information indicates whether early departure is permitted. Early departure means that the delay time of the departing event is negative, i.e., departing earlier than the scheduled time of the departing event. In the example of FIG. 10 , early departure is prohibited for all departing events 1, 3, 4, and 6. Since arrival events 2 and 5 are not departing events, early departure is permitted. While arrival events are permitted to arrive earlier than the scheduled time, whether early arrival is permitted for each arrival event may also be defined for each event. Whether early passing is permitted for each passing event may also be defined for each event. The early departure permission information may be included in the timetable information or may be stored in advance in the event information storage unit 250 as data in a separate file. Alternatively, the user may input the event information at the start of this process. Alternatively, all departing events may be uniformly and pre-defined as not allowing early departure. In this embodiment, it is assumed that the timetable information includes early departure permission information.

[0044] Information on preceding events is stored only when there is a preceding event of at least one of the following types 1 to 3 for the target event. However, the type of preceding event is not limited to these types and can be determined arbitrarily depending on the purpose of evaluating the diagram information. Type 1: The latter event in a pair of consecutive events (connected by a solid line in the graph in Figure 2) that belong to the same thread. In other words, the event that occurs immediately before the target event, connected by a solid line. Type 2: When the target event is an arrival event, the immediately preceding departure event at the same station (the event is called a departure-arrival event). The departure-arrival event belongs to a different line from the target event. Type 3: When the target event is a departure event, there is an departure event (called a departure-departure event) that takes place at the same station as the departure event. The departure-departure event belongs to a different line from the target event.

[0045] As an example other than types 1 to 3, if there are multiple platforms at which a train can arrive at a station and the target event is an arrival event, the target event may be the departure-arrival event on the same platform as the arrival event. Also, if the target event is a departure event, the target event may be the departure-departure event on the same platform as the departure event. Furthermore, when there are two traveling directions (for example, an up direction and a down direction), a departure-arrival event or a departure-departure event may be set as the preceding event for the same direction and the opposite direction, respectively. If there is a return service, the arrival and departure events before and after the return service can be considered in the same way as arrival and departure events in the same direction. Furthermore, when the target event is a passing event, the preceding event may be the immediately preceding departure event (called a departure-passing event) at the same station (or the same platform if there are multiple platforms), as in the case of an arrival event. The immediately preceding departure event may be in the same direction as the passing event or in the opposite direction.

[0046] The information on the preceding event includes the expected inter-event delay time between the target event and the preceding event, and the margin time between the target event and the preceding event. Furthermore, if the target event is an outgoing event and there are an arrival event at the same station on the same route and an outgoing-out event (the above-mentioned Type 3 event) as preceding events, the information on the preceding event includes a riding time proportional coefficient.

[0047] The evaluation order determination unit 515 creates an event list in which the identifiers (event IDs) of events are arranged in the evaluation order (Step_B). The event IDs are topologically sorted, and the order after the sorting is set as the evaluation order. In this example, as an example of topological sorting, the event IDs are sorted in descending order of the scheduled time of the event, and the order after the sorting is set as the evaluation order of the events. By performing subsequent processing in the evaluation order, it is guaranteed that when processing the event information of each event, the processing of the event information of all preceding events has been completed.

[0048] The delay probability evaluation unit 520 receives the first element (event N) of the event list (Step_C), and calculates delay time information (in this example, delay probability distribution) representing the delay time of the event based on the event information of the event (Step_D).

[0049] The delay probability evaluation unit 520 includes a convolution calculation unit 521 , a margin shift unit 522 , a composite probability calculation unit 523 , a round-up processing unit 524 , and a travel time change calculation unit 525 .

[0050] FIG. 9 is a flowchart illustrating an example of the process of the delay probability evaluation unit 520. First, in steps D0 to D2, the delay probability distribution between all preceding events {S} for event N is calculated. Below, we will mainly explain an example where event N is an outgoing event and event S is an incoming event that is the immediately preceding event in the same thread.

[0051] The travel time change calculation unit 525 determines whether event N is an originating event. If event N is not an originating event, the travel time change calculation unit 525 does not perform any processing. If event N is an originating event, the travel time change calculation unit 525 determines whether an arrival event exists immediately before at the same station and the aforementioned origin-origin event exists. If either one does not exist, no processing is performed. If both events exist, the travel time change calculation unit 525 calculates the amount of change ΔSN Calculate the change amount Δ SN As an example of this, the expected value of the change ΔE SN Calculate the expected value ΔE SN An example of a travel time change model for calculating is shown in equation (1). ΔE SN =λ N (E[X S ]―E[X M ]) (1)

[0052] In equation (1), the arrival event immediately before the target event (event N) on the same route is S, and the departure event immediately before the target event (event N) on the same route is M. Xi represents the random variable for the delay time of event i. E[·] represents the expected value according to the probability distribution that the random variable “·” follows. Therefore, E[X S ] is the expected time delay of the event M, E[X M ] is the expected time delay of the arrival event S.

[0053] The model of equation (1) is E[X S ] and E[X M ] and the difference (E[X S ]―E[X M ]) as an argument, and the proportional coefficient is λ N This difference is the expected value of the difference (change) in delay time between arrival event S and departure event M. This difference can be either a positive or negative value. The model in equation (1) is the expected value ΔE SN is the difference (E[X S ]―E[X M ])

[0054] Instead of equation (1), we use the random variable (X S , X M ) as an argument, the expected value of equation (2) may be calculated. λ N (X S -X M ) (2) That is, based on the following formula (3), the expected value of the change in the riding time ΔESN、 may be calculated. ΔE SN =E[λ N (Xs—X M )] (3)

[0055] In this example, the travel time change calculation unit 525 calculates the expected value of the delay probability distribution of each of the preceding events S and M, which are scheduled earlier than the event N, as E[X S ] and E[X M ] and the expected value of the change in travel time ΔE SN The preceding event S corresponds to, for example, a second event whose scheduled time is earlier than the event N, and the preceding event M corresponds to, for example, a third event whose scheduled time is earlier than the event N.

[0056] The expected value ΔE of the change in the travel time calculated by the travel time change calculation unit 525 SN are stored in the travel time change information storage unit 310 as travel time change information in association with the identifiers of the event N and the preceding event S (see FIG. 13, which will be described later).

[0057] The convolution calculation unit 521 calculates a probability distribution (f Y′SN ) is generated (Step_D1). That is, the delay probability distribution of event N when passing through event S is generated as a probability distribution (f Y′SN ) "SN" may be a subscript of "Y'".

[0058] This probability distribution f Y′SN is the probability distribution of the delay time of the event S, XS and the inter-event delay probability distribution (f DSN "SN" can be a subscript of "D'". Y' SN is a random variable that represents the delay time of event N. DSNis the distribution of delay times between events given by the event delay time information, and the change in ride time Δ SN This is a distribution that represents the delay time between events S and N, generated based on (the amount of change in the delay time between events). f DSN is the change in travel time Δ SN corresponds to the first inter-event delay time information representing the delay time between events S and N, which is reflected in the delay time between events S and N. The distribution representing the delay time between events given by the inter-event delay time information is SN corresponds to second inter-event delay time information that represents the delay time between events S and N, which is not reflected in the delay time between events S and N.

[0059] In this embodiment, the inter-event delay probability distribution is a negative binomial distribution, and the expected value of the inter-event delay time is given as a parameter that defines the distribution. In this case, the amount of change in the ride time is taken into consideration in this example. Therefore, the expected value of the inter-event delay time between event N (in this example, the departure event) and the immediately preceding arrival event (event S) on the same route is calculated by multiplying the expected value of the inter-event delay time given from the inter-event delay time information by the expected value of the change in the ride time, ΔE SN The inter-event delay time information (inter-event delay time given by inter-event delay time information) is changed by the amount of change in the ride time, and the changed inter-event delay time information is given as a parameter of the negative binomial distribution. As a result, the amount of change in the ride time Δ SN The inter-event delay probability distribution is obtained, which is reflected in the delay time between events S and N.

[0060] The margin shift unit 522 calculates the delay probability distribution (probability distribution f Y′SN ) is shifted (Step_D2). Y′SN The slack time m SN By shifting the time by f, the delay probability distribution f of event N when passing through event S, which takes into account the slack time, is YSN"SN" can be a subscript of "Y", where Y SN :=Y′ SN -m SN That is, Y SN is Y′ SN From m SN is defined as the sum of the

[0061] Next, in steps D3 and D4, the delay probability distribution of event N is calculated. More details are as follows.

[0062] The composite probability calculation unit 523 calculates the number of events preceding the event N (f YSN If the number of events S for which the delay probability is calculated is 1, the delay probability distribution f YSN Let W be the delay probability distribution of the event N. The delay probability distribution may be expressed as W(k) using a random variable k of delay time (same below).

[0063] The number of events preceding event N (f YSN If the number of events S for which σ is calculated is 0, the composite probability calculation unit 523 sets a predetermined initial delay distribution as the delay probability distribution W of the event N. For example, since the first selected event does not have a preceding event, the initial delay distribution becomes the delay probability distribution for the first selected event.

[0064] When there are multiple preceding events for event N (f YSN (When the number of events S for which the delay probability distribution f is calculated is plural), the composite probability calculation unit 523 calculates the delay probability distribution f for each of all preceding events S of the event N. YSN This gives the delay probability distribution W for event N. Specifically, the following process is performed:

[0065] The composite probability calculation unit 523 calculates a probability distribution (composite probability distribution) in which the maximum delay time when the event N passes through each of all preceding events S is used as a random variable, and sets the calculated probability distribution as the delay probability distribution W of the event N (Step_D3).

[0066] For example, if there are S1 to Sh (h is an integer equal to or greater than 2) as preceding events S, generate combinations of values ​​of k (random variable) between these preceding events S1 to Sh. If the number of possible values ​​of k is g, h × g combinations are obtained. For each combination, f YSN Calculate the product of (k) and select the maximum value of k. YSN (k) is the f for random variable k YSN This is the value of k. For each combination, a pair of the k value and the product is obtained. The obtained pairs are classified by the k value to obtain multiple groups. The sum of the products is calculated for each group, and the sum is used as the delay probability (composite probability) corresponding to that group. This gives the delay probability (composite probability) for each k value, and the set of delay probabilities (composite probabilities) for each k value is obtained as the composite probability distribution (delay probability distribution W for event N).

[0067] Figure 12 shows a specific example of the process of generating a composite probability distribution in Step_D3. For simplicity, let k be a discrete value in 1-minute increments within the range of -1 to 2. Let the number h of preceding events S be 2. These events will be called preceding event 1 and preceding event 2, respectively.

[0068] The upper part of Figure 12(A) shows the probability distribution f YSN (k) and the probability distribution f when passing through preceding event 2 YSN (k) are shown respectively.

[0069] FIG. 12(B) shows a table with the k value of preceding event 1 as the horizontal item and the k value of preceding event 2 as the vertical item. The upper row of each cell in the table stores the maximum value of the k value of preceding event 1 and the k value of preceding event 2 (or any one of the values ​​if they are the same). The lower row of each cell stores the f for the k value of preceding event 1. YSN (k) and f for the corresponding value of k of preceding event 2 YSN (k) is stored. This corresponds to the occurrence probability when it is assumed that preceding event 1 and preceding event 2 are independent.

[0070] For example, in the top right cell of the table, the value of k corresponding to preceding event 1 is 2, and the value of k corresponding to preceding event 2 is -1. Therefore, the larger of 2 and -1, 2, is stored. Also, the delay probability f when the value of k for preceding event 1 is 2, is YSN (2) is 10%, and the delay probability f when the value of k of the preceding event 2 is -1 YSN (-1) is 50%. Therefore, the product of these is 10% x 50% = 5%. Therefore, 5% is stored in this cell. Similarly, the maximum value of k and the product are stored in the other cells.

[0071] The cells in the table in Figure 12(B) are classified into groups with the same value of k (maximum value), and the products contained in the cells for each group are summed up. This gives the delay probability for each value of k.

[0072] Figure 12(C) shows a table in which the sum is calculated for each group. For example, an example of calculating the sum for a group when k is 2 is shown below. First, all cells containing 2 are identified from the table in Figure 12(B). In other words, the group of cells when k is 2 is identified. The eight cells in the rightmost column and bottom row of the table are identified. Calculating the sum of the products in these cells (values ​​in the bottom row) gives 5 + 4 + 1 + 0 + 0 + 0 + 0 = 10%. Similarly, when k is -1, the sum is calculated to be 20%, when k is 0, 43%, and when k is 1, 27%. In this way, the sum (delay probability) is obtained for each value of k. This set corresponds to the composite probability distribution.

[0073] The combination of the preceding event S and the event N that has the longest delay time may be called the delay propagation path. In the following explanation, the delay probability distribution of event N is the delay probability distribution W calculated in Step_D3, and the delay time of event N is the delay time (e.g., expected value) based on the delay probability distribution W calculated in Step_D3.

[0074] The rounding-up processing unit 524 determines whether the event N cannot be started early, and if it cannot be started early, rounds up the probability that the delay time of the event N will be negative in the delay probability distribution W of the event N (Step_D4). The probability that the delay time of the event will be less than 0 is added to the probability that the delay time will be 0, and instead the probability that it will be less than 0 is set to 0.

[0075] The calculated delay probability distribution W of the event N (including both the case where rounding up is performed and the case where rounding up is not performed) is stored in the delay probability distribution storage unit 300 as delay probability distribution information.

[0076] All events are selected in order as event N, and the above steps (D0 to D4) are repeatedly executed for event N.

[0077] 13 shows an example of the travel time change information stored in the travel time change information storage unit 310 as a result of processing by the delay probability evaluation unit 520. When the event N is the event 6 and the preceding arrival event is the arrival event 5, the expected value of the amount of change in the travel time ΔE SN (Note that the departure-departure event for event N is event 3.) When event N is event 3, the preceding arrival event is event 2, but since there is no departure-departure event for event 3, the expected value of the change in the ride time for event 3 is ΔE SN However, the expected value of the change in the travel time ΔE SN It is also possible to not require the calculation of the expected value ΔE SN may also be calculated.

[0078] FIG. 14 shows an example of delay probability distribution information stored in the delay probability distribution storage unit 300 as a result of processing by the delay probability evaluation unit 520. The horizontal rows represent delay times, and the vertical columns represent event (event N) identifiers. For example, for event 6, the probabilities of delay times of 0, 1, 2, 3, 4, and 5 are 0, 0.001, 0.001, 0.002, 0.002, and 0.003, respectively. Since early departure is prohibited for event 6, the delay time value cannot be negative (the same applies to events 1, 3, and 4). On the other hand, events 2 and 5 are arrival events, and early arrival is permitted, so the delay time value can be negative. In the example of FIG. 14, for event 2, the probability of a delay time of -1 is 0.296.

[0079] The processing of the delay probability evaluation unit 520 described above may involve multiple trials of simulation, such as the Monte Carlo method (for example, by selecting delay times using random numbers for the inter-event delay probability distribution and the event delay probability distribution and performing processing). In this case, the processing from Step C onwards is repeated multiple times, and the delay probability distribution of event N is obtained from the histogram. The delay probability distribution storage unit 300 and the travel time change information storage unit 310 may store observed values ​​of the delay time and travel time change amount for each trial. Furthermore, the delay probability distribution storage unit 300 may store information on the propagation path for each trial. According to this method, the processing of each unit of the delay probability evaluation unit 520 is simplified, allowing each trial to be executed quickly.

[0080] The output unit 400 reads information from at least one of the delay probability distribution storage unit 300 and the travel time change information storage unit 310, generates evaluation information for the bus schedule, and displays it on the screen. The evaluation information may be generated using not only at least one of the delay probability distribution information and the travel time change information, but also information obtained by the processing of the delay probability distribution calculation unit 500, information acquired by the input units 110 to 140, etc.

[0081] Figure 15 shows an example of displaying evaluation information. Graphs G1 and G2 showing the delay probability distribution of event 6 are displayed. Graph G1, shown by a solid line, is the delay probability distribution of event 6 when the travel time proportionality coefficient λ6 between events 5 and 6 is set to 0.5. Graph G2, shown by a dashed line, is the delay probability distribution of event 6 when the travel time proportionality coefficient λ6 is set to 0, i.e., the delay probability distribution of event 6 when the amount of change in travel time is not taken into account (the amount of change in travel time is set to zero).

[0082] It is expected that the delay of Event 6 will increase as the ride time between Event 5 and Event 6 increases due to a large delay in Event 5. From the solid line graph G1 in Figure 15, by taking into account changes in ride time, the mode of the delay probability distribution (delay time value of approximately 30) becomes larger than the mode of the dashed line graph G2 (delay time value of approximately 15), and the tail of graph G1 also becomes wider than that of dashed line graph G2.

[0083] Furthermore, the solid vertical line L1 in Figure 15 represents the expected value of the delay time of event 6 when λ6 = 0.5. The dashed vertical line L2 represents the expected value of the delay time of event 6 when λ6 = 0. The length of the horizontal line E1 extending in the negative direction from the solid vertical line L1 represents the expected value of the change in travel time. The end point of the solid horizontal line E1 coincides with the expected value (dashed vertical line L2) when the change in travel time is not taken into account. Therefore, the change in the expected value of the delay time of event 6 when the change in travel time is taken into account coincides with the expected value of the change in travel time.

[0084] FIG. 16 shows another example of displaying evaluation information. The user sets the value of the parameter (λ6 in this embodiment) of the travel time change model by moving the slider R, which is an input interface, left and right on the screen. In the example shown, the user can set the value of λ6 within the range of 0 to 1. A delay probability distribution for event 6 corresponding to the parameter value set by the user is generated by the delay probability distribution calculation unit 500 (processing unit). The output unit 400 displays a graph of the generated delay probability distribution and a vertical line representing the expected value of the delay time of event 6 on the screen. This allows the user to easily try and error various scenarios for the value of λ.

[0085] Variations of the evaluation information displayed on the output unit 400 are shown in (1) to (5) below. (1) The probability that each event will be delayed by more than or less than a threshold time may be displayed. Statistical values ​​such as the expected value, variance, and quantile of the delay time of each event may also be displayed.

[0086] (2) Statistical values ​​such as the expected value of the delay time for a specific section (between the start and end points of a series of consecutive events) may be displayed.

[0087] (3) Statistical values ​​such as the covariance and correlation coefficient of the delay time between an event preceding a target event and the target event may be displayed. Also, a delay propagation path estimated based on the statistical values ​​may be displayed.

[0088] (4) When information about an event (transfer event) that transfers to a target event is given, statistical values ​​such as the probability that a transfer from the transfer event to the target event will not be possible may be displayed. For example, the target event may be a departure event at a certain station, the transfer event may be an arrival event at the same station that arrives before the departure event, and a passenger may transfer from the transfer event train to the departure event train. In this case, if a delay in the transfer event that exceeds the allowance time for the time specified as the time required for the passenger to transfer occurs, the transfer will not be possible. Therefore, statistical values ​​such as the probability that a delay exceeding the allowance time will occur are displayed as the probability that the transfer will not be possible. Information about the transfer event may be acquired using the configuration shown in the second embodiment described below, or may be stored in timetable information in advance.

[0089] (5) Statistical values ​​such as the ratio or covariance between the expected value of delay time and the expected value of boarding time may be displayed. Based on the statistical values, the results of estimating the degree of influence of changes in boarding time on delay time may be displayed.

[0090] [Method for determining parameters of the travel time change model] Two methods for determining the parameter (ride time proportionality coefficient) λi of the ride time change model will be described below as Method 1 and Method 2. As described above, the ride time proportionality coefficient λi is the proportionality coefficient when the amount of change in ride time is assumed to be proportional to the interval (train interval) between the arrival event immediately preceding departure event at the same station (same line) as departure event i and the departure event on another line immediately preceding departure event at the same station. However, the parameter of the ride time change model is not limited to the ride time proportionality coefficient. For example, as shown in the second embodiment described later, a coefficient of a constant term (correction term) may be present.

[0091] (Method 1) λi is calculated by linear regression of the change in train intervals in actual operation data for event i. For example, λi can be calculated by regressing the change in delay time of event i based on actual operation data with the change in train intervals (the time interval between the time of the arrival event immediately before event i on the same route and the time of the departure event immediately before event i on a different route, as described above).

[0092] (Method 2) Assume that the change in the number of passengers boarding at event i is proportional to the change in the train interval. From this assumption, the change in the number of passengers boarding (number of boarding passengers) per unit time will be constant. Also, assume that the passenger flow rate (number of people boarding a train per unit time) is roughly constant. Therefore, λi is calculated by transforming the equation into the following equation (3).

[0093] λi = (change in number of boarding passengers per unit time) × (boarding time per unit passenger) = [(Number of passengers boarding during planned operation) / (Train interval during planned operation)] x [(Time required for boarding during planned operation) / (Number of passengers boarding during planned operation)] = (Train travel time during planned operation) / (Train interval during planned operation) (3)

[0094] Planned operation refers to when the train is running on schedule. The train intervals during planned operation are determined by the timetable and are obtained from the timetable information. Regarding the required boarding time during planned operation, if the required boarding time during planned operation has been actually measured, the actual data may be used. Alternatively, it may be assumed that the required boarding time during planned operation is approximate to the product of the stop time when a delay occurs and the ratio of the boarding time to the stop time, and the stop time when a delay occurs may be identified from actual operation data, and the ratio of the boarding time may be identified from actual data.

[0095] [Variations of the riding time change model] The ride time change model is a function having an argument that depends on the delay time information of a preceding event that is scheduled earlier than the target departure event. In the first embodiment, the preceding events that are the target of the ride time change model are a departure event on a different route immediately before the same station and an arrival event immediately before the same route, but are not limited to these events. For example, only one of these events may be the preceding event. For example, when delay time information representing the delay time between a second event that precedes the target event (first event) and the first event is generated based on delay time information representing the delay time of a third event that precedes the first event, the third event may be the same as or different from the second event.

[0096] Furthermore, an event different from the two preceding events described above may be the preceding event. For example, the first event occurring on the same line, or the consecutive events occurring two (or more) before that at the same station may be the preceding event. The number of preceding events is not limited to one or two, but may be three or more. Another example of a preceding event is an event occurring immediately before the same platform when there are multiple platforms at a station. Yet another example of a preceding event is an event occurring immediately before each of the same and opposite directions when there are two directions of travel.

[0097] Furthermore, the travel time change model does not have to be in a form in which the difference between the delay time information (e.g., expected value) of two events is proportional to the proportionality coefficient, as shown in equations (1) to (3). For example, the travel time change model may be a nonlinear function instead of a linear function. Furthermore, the travel time change model may be a regression model of another type, such as a neural network.

[0098] Examples of delay time information that can be used as an argument include delay time, expected value of delay time, probability distribution of the delay time, and a function indicating the probability distribution.Another example of an argument is the amount of change in train intervals (the time interval between a departure event on another line immediately preceding the same station and a arrival event on the same line immediately preceding the same station).

[0099] [Step_D0 processing variations] Variations of the processing of Step_D0 performed by the travel time change calculation unit 525 will be described. In the first embodiment described above, the amount of change in the travel time is expressed by a function of the expected value of the difference in delay time (train interval) between preceding events. As another method, the amount of change in the travel time may be calculated by a function of the observed value in one trial in a simulation that performs multiple trials, such as the Monte Carlo method, as described above. In this case, for example, the random variable X S ,X M is generated by random numbers etc. (i.e., random variable X S ,X M ), random variable X S ,X M By substituting the value of into equation (2), the change in the travel time is calculated.

[0100] In the first embodiment, the expected value of the change in the travel time is calculated as the change amount information representing the change in the travel time, but the probability distribution of the change in the travel time may be calculated. In this case, the change in the travel time is a function that depends on the delay time information (delay time) of one or more preceding events that are scheduled earlier than the target event, so the probability distribution of the change in the travel time is a convolution of the delay probability distribution of the one or more preceding events. In this embodiment, the random variable X S ,X M Each probability distribution is convolved according to equation (2). Specifically, λ is added to the random variables of the probability distribution obtained as a result of the convolution. N By multiplying by , we can obtain the probability distribution of the change in travel time.

[0101] Specifically, for example, if the random variables are X and Y and their respective probability distributions are pX(x) and pY(y), convolution is performed by calculating the probability distribution pZ(z) of the function Z = f(X, Y) as shown in the following equation (4). pZ(z)=∫Δ(z―f(x,y))pX(x)pY(y)dxdy (4)

[0102] [Variations of Step_D1 processing] Variations of the processing of Step_D1 performed in the convolution calculation unit 521 will be described. In the first embodiment, a negative binomial distribution is assumed, and the delay probability distribution of the target event (event N) is a distribution whose expected value is shifted by the amount of change in travel time from the expected value of the distribution determined from inter-event delay time information (hereinafter referred to as the original distribution). The original distribution corresponds to the delay probability distribution calculated without considering the amount of change in travel time. As an alternative method, instead of shifting the expected value, a distribution obtained by shifting the original distribution itself by the amount of change in travel time may be used as the delay probability distribution of the target event. The shifting process may be the same as that of the margin shift unit 522.

[0103] Furthermore, instead of a distribution obtained by shifting the original distribution, a probability distribution in which the sum of the random variable of the original distribution and the random variable of the amount of change in travel time is used as a new random variable may be used as the delay probability distribution of the target event. In this case, similar to the processing in the convolution calculation unit 521, the delay probability distribution of the target event can be obtained by convolving the original distribution with the probability distribution of the amount of change in travel time.

[0104] As described above, the first embodiment can achieve the following effects. The number of people boarding a train at a station between its arrival and departure varies depending on the delay time of the preceding train's event and the delay time of the train's arrival at the station. In this embodiment, this change affects the inter-event delay between the event at which the train departs and the event at which the train arrives, thereby affecting the delay time of the event at which the train departs. In this embodiment, this change in the number of people boarding a train is reflected in the inter-event delay, thereby enabling the delay time of the departing event to be properly evaluated. This allows the train schedule to be properly reflected in the evaluation. That is, the number of people boarding a train between its arrival and departure from a station varies depending on the delay time of the preceding train's departure from the station and the delay time of the train's arrival at the station. For example, if the departure of the preceding train is delayed and the train arrives on schedule, the inter-event delay will be smaller than expected, and the number of people boarding the train is expected to decrease. Furthermore, if the departure of a preceding train is delayed and the arrival of the train is significantly delayed, the delay between events will be larger than expected, and the number of people boarding the train is likely to increase. These changes in the number of people boarding the train will affect the delay between the event from which the train departs and the event to which the train arrives, and will affect the delay of the event from which the train departs. In this embodiment, this change in the number of people boarding the train is reflected in the delay between events, allowing the delay of the departing event to be properly evaluated.

[0105] In this embodiment, the change in delay time from arrival to departure at a station is affected by a change in the number of passengers. However, the change in delay time is not limited to a change in the number of passengers. For example, the change in delay time may be affected by a change in the time it takes for station staff to check the train status, or by a change in the number of passengers getting off a train that has arrived at the station. In either case, it is possible to create a probability distribution of the delay time of a departure event that reflects the change in the time it takes for station staff to check the train status, by treating the change in the number of passengers in the same way as a change in the number of passengers.

[0106] The user may modify the timetable information based on the delay time of the event evaluated by this embodiment. The timetable information is modified, for example, by inputting using an input interface. When the timetable information is modified, train operation may be controlled by a train control device based on the modified timetable information. An information processing system may be configured that includes this information processing device and a train (vehicle). Alternatively, an information processing system may be configured that includes this information processing device, a train (vehicle), and a train control device.

[0107] 17 is a block diagram of a diagram evaluation device 101A (hereinafter referred to as the present device 101A) which is an information processing device according to the second embodiment. A configuration for performing correction processing of a travel time change model is added. Specifically, the present device 101A includes, in addition to the device 101 of FIG. 1, an information input unit 150, an information storage unit 260, a model correction unit 530, and a correction model information storage unit 270.

[0108] The information input unit 150 acquires information for correcting the travel time change model. The information input unit 150 may acquire information through a user's input operation, or may acquire information from a sensor or a computer such as a server. An example of a sensor is a camera installed on the platform or inside the train. The computer or the information input unit 150 may acquire information such as the number of people on the platform, the number of people getting off the train, and the number of people getting on the train as performance data from an image taken by the camera.

[0109] The information storage unit 260 stores the information acquired by the information input unit 150 .

[0110] The model correction unit 530 corrects the riding time variation model based on the information stored in the information storage unit 260. For example, the model correction unit 530 performs processing such as correcting the parameter values ​​of the riding time variation model, adding a correction term to the riding time variation model, or changing the function form of the riding time variation model itself. This correction has the effect of increasing the accuracy of the riding time variation model.

[0111] The correction model information storage unit 270 stores information on the corrected riding time change model.

[0112] The information stored in the correction model information storage unit 270 is read by the model information input unit 140 and stored in the model information storage unit 240. The information stored in the model information storage unit 240 is read by the delay probability distribution calculation unit 500, and the corrected ride time change model indicated in the information is used in the processing of the delay probability distribution calculation unit 500. In the processing of Step_D0, the delay probability distribution calculation unit 500 calculates the amount of change in ride time (the amount of change in inter-event delay time) using the corrected ride time change model. That is, the delay probability distribution calculation unit 500 calculates the amount of change in ride time (the amount of change in inter-event delay time) based on the information acquired by the information input unit 150.

[0113] Below, variations of information acquired by the information input unit 150 are described along with examples of correction processing performed by the model correction unit 530. Two or more correction processing based on the information of each variation described below may be combined. Furthermore, the information acquired and the correction processing may differ for each event.

[0114] [Event attribute information] The information input unit 150 acquires attribute information of an event (event attribute information). The event attribute information includes, for example, attributes of the event, such as the station, line, direction of travel, type of service (such as local or express), and the time of the event. Part of the event attribute information may also be included in the timetable information. For example, the time of the event is also included in the timetable information.

[0115] The model correction unit 530 corrects the travel time variation model based on this information. For example, if the type of operation of the target event is an express train, it may be assumed that the increase in the number of passengers per unit time due to delay is greater than that of a local train, and the proportionality coefficient may be multiplied by a weight greater than 1. Alternatively, other corrections may be made, such as adding a correction term that is a predetermined positive value to the travel time variation model. In the processing of Step_D0, the delay probability distribution calculation unit 500 calculates the amount of change in travel time (the amount of change in inter-event delay time) using the corrected travel time variation model. That is, the delay probability distribution calculation unit 500 calculates the amount of change in travel time (the amount of change in inter-event delay time) based on the event attribute information acquired by the information input unit 150.

[0116] [Transfer Information] The information input unit 150 acquires the transfer information. The transfer information includes, for example, information about the transfer event (the arrival event from which the target event is transferred) and the train route including the transfer event (the train to which the train route is assigned is called a “transfer train”).

[0117] Specific examples include the degree of congestion at the platform of a transfer event, the number of passengers disembarking (the number of passengers disembarking from the train at the transfer event), the ratio of the number of passengers disembarking to the number of passengers on the train, the time required to transfer to the target event (transfer time required), etc. The transfer information may also include information on events other than the event (the target event) that is the target of transfer from the transfer event and on routes including the event.

[0118] The model correction unit 530 corrects the ride time variation model based on the transfer information. For example, assuming that the higher the platform congestion level, the longer the ride time, the proportional coefficient may be multiplied by a weight corresponding to the congestion level. Alternatively, other corrections may be performed, such as adding a correction term corresponding to the congestion level to the ride time variation model. The congestion level may be calculated, for example, by dividing the number of people present on the platform, including the number of passengers disembarking, by the platform's capacity, or by other methods. The platform for a transfer event may be the same as or different from the platform for the target event.

[0119] Furthermore, the model correction unit 530 may determine whether a delay in the connecting event causes the time interval between the time of the target event and the time of the connecting event (delay time) to be equal to or less than the required connecting time (whether the connecting event is impossible). If the connecting event is impossible, the model correction unit 530 corrects the ride time change model so that the change in the ride time of the target event becomes smaller due to a decrease in connecting passengers. The proportional coefficient may be multiplied by a weight smaller than 1. Alternatively, other corrections may be performed, such as adding a correction term that is a predetermined negative value to the ride time change model. Furthermore, the model correction unit 530 corrects the ride time change model for the departing event so that the change in the ride time of the departing event on a different route immediately following the target event on the same track (same route) becomes larger due to an increase in connecting passengers. The proportional coefficient may be multiplied by a weight larger than 1. Alternatively, other corrections may be performed, such as adding a correction term that is a predetermined positive value to the ride time change model.

[0120] In the processing of Step_D0, the delay probability distribution calculation unit 500 calculates the amount of change in the ride time (the amount of change in the delay time between events) using the corrected ride time change model. That is, the delay probability distribution calculation unit 500 calculates the amount of change in the ride time (the amount of change in the delay time between events) based on the connection information acquired by the information input unit 150.

[0121] [Weather Information] The information input unit 150 acquires weather information. The weather information includes, for example, forecast values ​​such as the temperature at the time the target event occurs and the presence and amount of rainfall. If rain is forecast at the time the target event occurs, the model correction unit 530 corrects the ride time variation model so that the change in ride time decreases depending on the amount of rainfall, assuming that the number of boarding passengers will decrease. For example, a correction term is added to the ride time variation model so that the change in ride time decreases depending on the amount of rainfall. Alternatively, the proportional coefficient may be multiplied by a weight according to the amount of rainfall. In the processing of Step_D0, the delay probability distribution calculation unit 500 calculates the change in ride time (the change in inter-event delay time) using the corrected ride time variation model. That is, the delay probability distribution calculation unit 500 calculates the change in ride time (the change in inter-event delay time) based on the weather information acquired by the information input unit 150.

[0122] [Equipment information] The information input unit 150 acquires facility information. The facility information is, for example, facility information of a location related to the stop position (platform number) of the target event. As a specific example, the facility information includes the number of stairs and / or elevators that exist on the route from the platform to the ticket gate or the transfer destination line (transfer destination platform). In addition to the number, the location of at least stairs and / or elevators may also be included.

[0123] Assume that passengers concentrate at the car door closest to the stairs on the train of the target event, and that the boarding time at that door dominates the change in the boarding time for the target event. The model correction unit 530 predicts the ratio of the number of passengers at the door with the largest number of boarding passengers compared to all other doors (the concentration of boarding passengers). The prediction may be performed based on past performance data, on image data captured inside the train, or directly from the number of passengers boarding and alighting at each door in the passenger information described below. The model correction unit 530 corrects the boarding time change model based on the prediction result. The proportional coefficient may be multiplied by a weight corresponding to the concentration. Alternatively, other corrections may be performed, such as adding a correction term corresponding to the concentration to the boarding time change model. In Step_D0, the delay probability distribution calculation unit 500 calculates the change in boarding time (the change in inter-event delay time) using the corrected boarding time change model. That is, the delay probability distribution calculation unit 500 calculates the change in boarding time (the change in inter-event delay time) based on the facility information acquired by the information input unit 150.

[0124] [Information about large-scale events] The information input unit 150 acquires information about large-scale events. The information about large-scale events includes, for example, information about whether or not a large-scale event is being held near the station where the target event is held, the date and time, the number of participants in the event, and the like.

[0125] The model correction unit 530 determines whether the target event falls within a certain time period relative to at least one of the start and end times of a large-scale event near the station (whether the target event is close to at least one of the start and end times of the large-scale event). If the target event falls within the certain time period (is close), a correction term that increases the amount of change in the travel time for the target event according to the number of event participants is added to the travel time variation model. Alternatively, the proportionality coefficient may be multiplied by a weight according to the number of participants. In the processing of Step_D0, the delay probability distribution calculation unit 500 calculates the amount of change in travel time (the amount of change in inter-event delay time) using the corrected travel time variation model. That is, the delay probability distribution calculation unit 500 calculates the amount of change in travel time (the amount of change in inter-event delay time) based on the information about the large-scale event acquired by the information input unit 150.

[0126] [Vehicle attribute information] The information input unit 150 acquires vehicle attribute information. The vehicle attribute information includes, for example, the number of cars in the train that will be operating in the target event, and the type of each car, such as a weakly air-conditioned car.

[0127] When target events are operated by trains with different numbers of cars on the same track (same route), it is assumed that the ride time for trains with fewer cars is longer than that for trains with more cars. The model correction unit 530 corrects the ride time variation model so that the change in ride time for the target event is changed depending on the number of cars in the train of the target event. For example, a correction term depending on the number of cars in the train is added to the ride time variation model, or a proportionality coefficient is multiplied by a weight depending on the number of cars in the train. In the processing of Step_D0, the delay probability distribution calculation unit 500 calculates the change in ride time (change in inter-event delay time) using the corrected ride time variation model. That is, the delay probability distribution calculation unit 500 calculates the change in ride time (change in inter-event delay time) based on the vehicle attribute information acquired by the information input unit 150.

[0128] [Passenger Information] The information input unit 150 acquires passenger information. The passenger information includes, for example, the number and ratio of boarding passengers, and the number and ratio of disembarking passengers. The passenger information may also include the number and ratio of passengers for each vehicle or each vehicle door.

[0129] For departure events where the ratio of disembarking passengers to the number of passengers on the train is high at the immediately preceding arrival event, the number of passengers on board is small, so it is assumed that the impact of the stop time (the time interval between the departure event and the immediately preceding arrival event) on the boarding time is small. The model correction unit 530 corrects the boarding time variation model so that the change in the boarding time for the departure event (target event) decreases according to the ratio of the number of disembarking passengers at the immediately preceding arrival event. For example, a correction term according to this ratio is added to the boarding time variation model, or a proportionality coefficient is multiplied by a weight according to this ratio. In the processing of Step_D0, the delay probability distribution calculation unit 500 calculates the change in boarding time (the change in inter-event delay time) using the corrected boarding time variation model. That is, the delay probability distribution calculation unit 500 calculates the change in boarding time (the change in inter-event delay time) based on the passenger information acquired by the information input unit 150.

[0130] [Other passenger information] The information input unit 150 acquires information about other passengers. Other passengers are passengers other than those boarding and disembarking passengers for the target event. For example, there are passengers who neither board nor disembark on the train, or passengers at the platform who do not board the train. The other passenger information includes, for example, the number of other passengers on the train, or their proportion (degree of congestion). These values ​​may be included for each train car or car door. The other passenger information may include the number of other passengers on the platform, or their proportion (degree of congestion).

[0131] It is assumed that the higher the level of congestion inside the train, the longer the ride time (the time required to get inside the train). The model correction unit 530 corrects the ride time variation model so that the amount of change in the ride time for the target event increases according to the level of congestion. For example, a correction term according to the level of congestion is added to the ride time variation model, or a proportionality coefficient is multiplied by a weight according to the level of congestion. In the processing of Step_D0, the delay probability distribution calculation unit 500 calculates the amount of change in ride time (the amount of change in delay time between events) using the corrected ride time variation model. That is, the delay probability distribution calculation unit 500 calculates the amount of change in ride time (the amount of change in delay time between events) based on the information of other passengers acquired by the information input unit 150.

[0132] [Camera footage information] The information input unit 150 acquires camera image information. The camera image information includes, for example, images taken by cameras installed at ticket gates, platforms, or inside trains at stations, and information extracted from the images.

[0133] The model correction unit 530 estimates the number of passengers boarding from the camera image and corrects the ride time variation model according to the estimated value. For example, a correction term according to the number of passengers boarding is added to the ride time variation model, or a proportionality coefficient is multiplied by a weight according to the number of passengers boarding. In the processing of Step_D0, the delay probability distribution calculation unit 500 calculates the amount of change in ride time (the amount of change in delay time between events) using the corrected ride time variation model. That is, the delay probability distribution calculation unit 500 calculates the amount of change in ride time (the amount of change in delay time between events) based on the camera image information acquired by the information input unit 150.

[0134] (Hardware configuration) 18 shows the hardware configuration of an information processing device according to each embodiment. The information processing device is configured by a computer device 600. The computer device 600 includes a CPU 601, an input interface 602, a display device 603, a communication device 604, a main memory device 605, and an external memory device 606, which are interconnected by a bus 607.

[0135] The CPU (Central Processing Unit) 601 executes an information processing program, which is a computer program, on the main memory device 605. The information processing program is a program that realizes each of the above-mentioned functional components of the information processing device. The information processing program may be realized not as a single program, but as a combination of multiple programs and scripts. Each functional component is realized by the CPU 601 executing the information processing program.

[0136] The input interface 602 is a circuit for inputting operation signals from input devices such as a keyboard, a mouse, a touch panel, etc. to the information processing apparatus. The input interface 602 corresponds to the input unit of the information processing apparatus according to each embodiment.

[0137] The display device 603 displays data output from the information processing device. The display device 603 is, for example, but not limited to, an LCD (liquid crystal display), an organic electroluminescence display, a CRT (cathode ray tube), or a PDP (plasma display). Data output from the computer device 600 can be displayed on the display device 603. The display device 603 corresponds to the output unit of the information processing device according to each embodiment.

[0138] The communication device 604 is a circuit for the information processing device to communicate with an external device wirelessly or via a cable. Data can be input from the external device via the communication device 604. The data input from the external device can be stored in the main memory device 605 or the external memory device 606. The communication device 604 corresponds to the input unit of the information processing device according to each embodiment.

[0139] The main memory device 605 stores an information processing program, data required for executing the information processing program, data generated by executing the information processing program, etc. The information processing program is deployed and executed on the main memory device 605. The main memory device 605 is, for example, a RAM, a DRAM, or an SRAM, but is not limited to these. Each storage unit or database of the information processing device according to each embodiment may be constructed on the main memory device 605.

[0140] The external storage device 606 stores information processing programs, data required for executing the information processing programs, data generated by executing the information processing programs, etc. These information processing programs and data are read into the main storage device 605 when the information processing programs are executed. The external storage device 606 is, for example, a hard disk, an optical disk, a flash memory, or a magnetic tape, but is not limited to these. Each storage unit or database of the information processing device may be constructed on the external storage device 606.

[0141] The information processing program may be pre-installed in the computer device 600, or may be stored in a storage medium such as a CD-ROM. The information processing program may also be uploaded onto the Internet.

[0142] Furthermore, the information processing device may be configured as a single computer device 600, or may be configured as a system made up of multiple computer devices 600 connected to each other.

[0143] The present invention is not limited to the above-described embodiments, and the components can be modified and embodied in practice without departing from the spirit of the invention. Furthermore, various inventions can be created by appropriately combining multiple components disclosed in the above-described embodiments. For example, configurations in which some components are omitted from all the components shown in each embodiment may also be considered. Furthermore, components described in different embodiments may be appropriately combined. [Explanation of symbols]

[0144] 101, 101A Diameter evaluation device (information processing device) 110 Timetable information input section 120 Inter-event delay time information input section 130 Required time between events input section 140 Model information input section 150 Information input section 210 Diagram information storage unit 220 Inter-event delay time information storage unit 230 Event-to-event required time memory section 240 Model information storage unit 250 Event information storage unit 260 Information storage section 270 Correction model information storage unit 300 Delay probability distribution memory unit 310 Ride time change information storage unit 400 Output Section 500 Delay probability distribution calculation unit (processing unit) 510 Event Information Creation Department 515 Evaluation Order Determination Unit 520 Delay Probability Evaluation Unit 521 Calculation Department 522 Margin Shift Section 523 Composite Probability Calculation Unit 524 Processing Unit 525 Travel time change calculation unit 530 Model Correction Unit 600 Computer equipment 601 CPU 602 Input Interface 603 Display device 604 Communication equipment 605 Main storage 606 External storage device 607 Bus

Claims

1. generating first inter-event delay time information representing a delay time between a first event and a second event preceding the first event among a plurality of events defining departure or arrival times at a plurality of stopping positions for at least one vehicle and the departure or arrival times at the plurality of stopping positions, based on delay time information representing a delay time of a third event preceding the first event; generating delay time information representing a delay time of the first event based on the first inter-event delay time information and delay time information representing a delay time of the second event; Processing section Equipped with the processing unit calculates change amount information representing a change amount of the delay time between the first event and the second event based on the delay time information of at least one of the third event and the second event, changes an expectation value of the delay time between the first event and the second event according to the change amount information, and generates the first inter-event delay time information by providing the changed expectation value as an input parameter of a negative binomial distribution. Information processing device.

2. the first event is an event including the departure of a first vehicle from a first stopping location; The change amount information indicates a change amount of the number of people getting on the first vehicle at the first stop position. The information processing device according to claim 1 .

3. The second event includes an event that precedes the first event and includes the arrival of the first vehicle at the first stop location. The information processing device according to claim 2 .

4. The third event precedes the second event and includes an event in which a second vehicle departs from the first stop location.

4. The information processing device according to claim 2 or 3.

5. The processing unit calculates the change amount information according to a difference between an expected value of the delay time of the second event and an expected value of the delay time of the third event. The information processing device according to any one of claims 1 to 4.

6. The processing unit calculates the change amount information by multiplying the difference by a proportionality coefficient corresponding to a change in the delay time between the first event and the second event. The information processing device according to claim 5 .

7. The processing unit generates the delay time information of the first event by convolving the first inter-event delay time information and the delay time information of the second event. The information processing device according to any one of claims 1 to 6.

8. The processing unit calculates the change amount information based on attribute information of the first event. The information processing device according to any one of claims 1 to 7.

9. the first event is an event including the departure of a first vehicle from a first stopping location; The processing unit calculates the change amount information based on information about a transfer event that is an event of arrival of a second vehicle from which a transfer to the first vehicle occurs in the first event. The information processing device according to any one of claims 1 to 8.

10. The processing unit calculates the change amount information based on weather information corresponding to the time of the first event. The information processing device according to any one of claims 1 to 9.

11. The processing unit calculates the change amount information based on facility information of a location related to a stop position of the first event. The information processing device according to any one of claims 1 to 10.

12. The processing unit calculates the change amount information based on the status of events being held near the stopping position of the first event at the time of the first event. The information processing device according to any one of claims 1 to 11.

13. The processing unit calculates the change amount information based on attribute information of a vehicle operating in the first event. The information processing device according to any one of claims 1 to 12.

14. The processing unit calculates the change amount information based on information about passengers getting on or off a vehicle operating in the first event. The information processing device according to any one of claims 1 to 13.

15. The processing unit calculates the change amount information based on information about passengers who are in a vehicle operating the first event or near a stopping position of the first event and who do not get on or off the vehicle. The information processing device according to any one of claims 1 to 14.

16. The processing unit calculates the change amount information based on an image captured by a camera installed at least one of a platform including a stop position of the first event, a ticket gate of a station including the stop position, and inside a train operating for the first event. The information processing device according to any one of claims 1 to 15.

17. an output unit that displays an input interface on a screen that allows a user to input a value of the proportionality coefficient, and displays on the screen the delay time information of the first event that is generated by the processing unit based on the value of the proportionality coefficient that is input via the input interface; The information processing device according to claim 6 , comprising:

18. The vehicle is a train including a formation of multiple vehicles. The information processing device according to any one of claims 1 to 17.

19. at least one vehicle; generating first inter-event delay time information representing a delay time between a first event and a second event preceding the first event among a plurality of events defining departure or arrival times at a plurality of stopping positions for the at least one vehicle and the departure or arrival times at the plurality of stopping positions, based on delay time information representing a delay time of a third event preceding the first event; generating delay time information representing a delay time of the first event based on the first inter-event delay time information and delay time information representing a delay time of the second event; a processing unit; Equipped with the processing unit calculates change amount information representing a change amount of the delay time between the first event and the second event based on the delay time information of at least one of the third event and the second event, changes an expectation value of the delay time between the first event and the second event according to the change amount information, and generates the first inter-event delay time information by providing the changed expectation value as an input parameter of a negative binomial distribution. Information processing system.

20. generating first inter-event delay time information representing a delay time between a first event and a second event preceding the first event among a plurality of events defining departure or arrival times at a plurality of stopping positions for at least one vehicle and the departure or arrival times at the plurality of stopping positions, based on delay time information representing a delay time of a third event preceding the first event; generating delay time information representing a delay time of the first event based on the first inter-event delay time information and delay time information representing a delay time of the second event; calculating change amount information representing a change amount of the delay time between the first event and the second event based on the delay time information of at least one of the third event and the second event; changing an expected value of the delay time between the first event and the second event according to the change amount information, and providing the changed expected value as an input parameter of a negative binomial distribution, thereby generating the first inter-event delay time information. A computer-implemented information processing method.

21. generating first inter-event delay time information representing a delay time between a first event and a second event preceding the first event among a plurality of events defining departure or arrival times at a plurality of stopping positions for at least one vehicle and the departure or arrival times at the plurality of stopping positions, based on delay time information representing a delay time of a third event preceding the first event; generating delay time information representing a delay time of the first event based on the first inter-event delay time information and delay time information representing a delay time of the second event; calculating, based on the delay time information of at least one of the third event and the second event, change amount information representing a change amount of the delay time between the first event and the second event; on the computer, the step of generating the first inter-event delay time information includes changing an expectation value of the delay time between the first event and the second event in accordance with the change amount information, and providing the changed expectation value as an input parameter of a negative binomial distribution, thereby generating the first inter-event delay time information. Computer program.

Images