Operation plan change support device, operation plan change support method, and train operation management system

Abstract

To provide a method by which an operation project revising plan gradually executed is generated.SOLUTION: An operation project revising support device comprises a determination model generation unit and an operation project revising plan generation unit. The determination model generation unit, to which both information about an operation content of a train at or before a certain point in time and revising information on an operation project executed before the certain point in time are inputted, generates an operation project revising determination model of outputting revised information on an operation project to be executed next to the operation project already executed. The operation project revising plan generation unit accepts input of information about train operation contents at or before a point in time specified by a user and revised information on an operation project executed before the specified point in time, and repeats one time or more to generate information on a revising content of the operation project based on information outputted from the operation project revising determination model, so as to generate revised candidates that are candidates of the respective operation projects to be revised gradually at or after the specified point in time, specify a revised candidate satisfying a prescribed condition from each generated revised candidate and generate the specified revised candidate as the operation project revised plan.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an operation plan change support device, an operation plan change support method, and a train operation control system. [Background technology]

[0002] One of the tasks in train operation management is changing the operation plan, and there is an operation plan change support system that supports this.A known technology related to the operation plan change support system is to predict the operation status of each train from the current time onwards based on the running record and operation plan of each train obtained up to the current time, and if a delay is expected, to automatically create a proposed change to the operation plan and present it to the operator in charge of operation management, thereby supporting the operation plan change task.

[0003] For example, Patent Document 1 discloses an operation management system that supports vehicle operation management by operation dispatchers, and has a operation prediction unit that uses operation prediction to create a predicted train timetable up to the time of operation rescheduling, which changes the train operation plan, and a operation rescheduling support processing unit that creates a model for judging the need for operation rescheduling based on past actual timetables, which are the results of train operations in the past, and applies the model to the actual and predicted timetables, which are the results of train operations on that day, to create a result of judging the need for operation rescheduling, indicating whether or not operation rescheduling is necessary. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2020-117169 A Summary of the Invention [Problem to be solved by the invention]

[0005] However, the technology disclosed in Patent Document 1 uses a judgment model to judge whether or not operation rescheduling is necessary based on the amount of delay, and does not utilize combinations of multiple types of plan changes, the order of plan changes implemented in stages, or the presence or absence of inconsistencies that may occur along the way. Therefore, with the technology disclosed in Patent Document 1, it is difficult to create a proposed operation plan change for operation disruptions that is widely accepted by operators. In particular, in situations where operation schedules are becoming more complex, it is necessary to implement changes to operation plans in stages to prevent operation disruptions from expanding, and the current situation is that such operations still rely heavily on the knowledge and experience of operators such as dispatchers.

[0006] The present invention has been made in consideration of the current situation, and its purpose is to provide an operation plan change support device, an operation plan change support method, and a train operation management system that are capable of supporting the creation of operation plan change proposals to be implemented in stages. [Means for solving the problem]

[0007] One of the present inventions for solving the above problem includes inputting information about train operation details before a specified time point and information about changes to an operation plan that were executed before the specified time point into an operation plan change judgment model that outputs information about changes to an operation plan to be executed next to the executed operation plan, and creating information about changes to the operation plan based on the information output from the operation plan change judgment model. Multiple times By repeating this process, the operation plan change proposal is executed in stages from the point where it was first created Multiple an operation plan change proposal creation unit that creates change candidates that are candidates for each of the operation plan changes, identifies a change candidate that satisfies a predetermined condition from each of the created change candidates, and creates an operation plan change proposal from the identified change candidates; an allowable inconsistency storage unit that stores allowable inconsistency information, which is information on operation plan changes that have been executed in stages in the past, which have caused a train inconsistency due to the change but have resolved the inconsistency due to a change in a subsequent stage; and In the case of changing the plurality of operation plans The inconsistency caused by the changes indicated by the above-mentioned change candidates In the case of changing the plurality of operation plansand a change application determination unit that determines whether the inconsistency can be resolved at a later stage, and if it is determined that the inconsistency will be resolved at a later stage, determines that the created change candidate is a change candidate that satisfies the specified condition, and if it is determined that the inconsistency will not be resolved at a later stage, determines that the created change candidate is not a change candidate that satisfies the specified condition.

[0008] In addition, one aspect of the present invention for solving the above problem is to provide an information processing device that inputs information on train operation contents before a specified time point and information on changes to an operation plan executed before the specified time point into an operation plan change judgment model that outputs information on changes to an operation plan to be executed next to the executed operation plan, and creates information on changes to the operation plan based on the information output from the operation plan change judgment model. Multiple times By repeating this process, the operation plan change proposal is executed in stages from the point where it was first created Multiple an operation plan change proposal creation process that creates change candidates that are candidates for each of the operation plan changes, identifies a change candidate that satisfies a predetermined condition from each of the created change candidates, and creates an operation plan change proposal from the identified change candidates; an allowable inconsistency storage unit that stores allowable inconsistency information, which is information on operation plan changes that have been executed in stages in the past, which caused a train inconsistency due to the change but which were resolved by a change in a subsequent stage; and In the case of changing the plurality of operation plans The inconsistency caused by the changes indicated by the above-mentioned change candidates In the case of changing the plurality of operation plans and a change application determination process for determining whether the inconsistency can be resolved at a later stage, and if it is determined that the inconsistency will be resolved at a later stage, determining that the created change candidate is a change candidate that satisfies the specified condition, and if it is determined that the inconsistency will not be resolved at a later stage, determining that the created change candidate is not a change candidate that satisfies the specified condition.

[0009] In addition, one of the present inventions for solving the above problem is a method for inputting information on train operation contents before a specified time point and information on changes to an operation plan executed before the specified time point into an operation plan change judgment model that outputs information on changes to an operation plan to be executed next to the executed operation plan, and creating information on changes to the operation plan based on the information output from the operation plan change judgment model. Multiple times By repeating this process, the operation plan change proposal is executed in stages from the point where it was first created Multiple an operation plan change proposal creation unit that creates change candidates that are candidates for each of the operation plan changes, identifies a change candidate that satisfies a predetermined condition from each of the created change candidates, and creates an operation plan change proposal from the identified change candidates; an allowable inconsistency storage unit that stores allowable inconsistency information, which is information on operation plan changes that have been executed in stages in the past, which have caused a train inconsistency due to the change but have resolved the inconsistency due to a change in a subsequent stage; and In the case of changing the plurality of operation plans The inconsistency caused by the changes indicated by the above-mentioned change candidates In the case of changing the plurality of operation plans The train operation management system includes an operation plan change support device that includes an operation plan change application determination unit that determines whether the inconsistency can be resolved at a later stage, and if it is determined that the inconsistency will be resolved at a later stage, determines that the created change candidate is a change candidate that satisfies the specified condition, and if it is determined that the inconsistency will not be resolved at a later stage, determines that the created change candidate is not a change candidate that satisfies the specified condition, and a route control device that controls trains or train facilities based on the created operation plan change proposal. Effect of the Invention

[0010] According to the present invention, it is possible to assist in creating an operation plan change proposal that is executed in stages. Problems, configurations and effects other than those described above will become apparent from the following description of the embodiments. [Brief description of the drawings]

[0011] [Figure 1]FIG. 1 is a diagram illustrating an example of the configuration of a train traffic control system according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a diagram illustrating an example of a configuration of an operation plan change support device. [Diagram 3] FIG. 13 illustrates an example of a changed input history DB. [Figure 4] FIG. 11 is a diagram illustrating an example of a plan change rule. [Diagram 5] 1 is a flow diagram illustrating an overview of train traffic control processing performed by a train traffic control system. [Figure 6] FIG. 11 is a flow diagram illustrating an example of a teacher data creation process. [Figure 7] FIG. 11 is a diagram illustrating an example of allowable inconsistency information. [Figure 8] FIG. 11 is a flow diagram illustrating details of an operation plan change proposal creation process. [Figure 9] FIG. 13 is a diagram showing an example of a change proposal screen. [Figure 10] FIG. 11 is a flow diagram illustrating details of a change application determination process. [Figure 11] FIG. 11 is a flow diagram illustrating details of a change execution determination process. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] 1 is a diagram showing an example of the configuration of a train traffic control system 10 according to this embodiment. The train traffic control system 10 includes an operation plan management device 40 that manages information used to manage train operations, such as the operation plan (train timetable, etc.) and operation record of each train 6, a train tracking device 50 that tracks the running of each train 6 and acquires the operation record, a route control device 60 that controls the route of each train 6, and an operation plan change support device 100 that supports changes to the operation plan of each train 6.

[0013] The operation plan management device 40 manages information used to manage train operation, such as operation plans, operation results, and operation plan change history, including information related to the current day when operation management is being performed and information related to the previous day and earlier. After one day of train operation is completed, the operation plan management device 40 adds each piece of information, such as the initial operation plan for that day, the operation results for the entire day, and the plan change history for the entire day, to the past history database as past history data, and further initializes various information for operation management on the next day. Specifically, for example, the operation plan management device 40 updates the operation plan to the next day's one, empties the operation results, and empties the plan change history. When the operation plan management device 40 receives a proposed change to the operation plan from the operation plan change support device 100, it also changes the operation plan based on the received proposed change to the operation plan, and further stores the received operation plan change information and the current time as a pair as a plan change history.

[0014] The train tracking device 50 acquires information on the location of vehicles on the line from various facilities 5, and creates operation records (running history) of each train 6 based on the acquired information on the location of vehicles on the line and the operation plan. The train tracking device 50 transmits the created operation records to the operation plan change support device 100.

[0015] The route control device 60 automatically controls the running (route) of each train 6 by transmitting instruction information to the train 6 or equipment 4 such as signals and switches based on the operation plan of each train 6, etc.

[0016] When train operations are disrupted due to an event such as an accident or a natural disaster, the operation plan change support device 100 generates an operation plan change proposal for making stepwise changes to the operation plan to restore train operations in response to the disruption, based on an operation plan change determination model to be described later. A detailed configuration of the operation plan change support device 100 will be described later (FIG. 2).

[0017] The operation plan management device 40, the train tracking device 50, the route control device 60, and the operation plan change support device 100 are communicatively connected to each other via a wired or wireless communication network 3 such as a communication line (data bus), a dedicated line, a LAN (Local Area Network), a WAN (Wide Area Network), or the Internet. Next, the configuration of the operation plan change support device 100 will be described.

[0018] <Operation plan change support device> 2 is a diagram illustrating an example of a configuration of the operation plan change support device 100. The operation plan change support device 100 includes an arithmetic processing device 11 such as a central processing unit (CPU), a digital signal processor (DSP), a graphics processing unit (GPU), or a field-programmable gate array (FPGA), a memory 12 such as a read only memory (ROM), a random access memory (RAM), or a storage device 13 such as a hard disk drive (HDD) or a solid state drive (SSD), an input device 14 including a mouse, a keyboard, or the like, a display device 15 including a liquid crystal display, an organic electro-luminescence (EL) display, or the like, and a communication device 16 including a network interface card (NIC), a wireless communication module, a universal serial interface (USB) module, a serial communication module, or the like.

[0019] In addition, the operation plan change support device 100 has each functional unit (program) of a predictive schedule creation unit 101, a teacher data generation unit 102, an allowable inconsistency memory unit 1025, a judgment model generation unit 103, an operation plan change proposal creation unit 104, a change application judgment unit 105, and a plan change rule update unit 107.

[0020] The predictive timetable creation unit 101 predicts a future operation plan (predictive timetable) using the operation record of each train (if a train operation is disrupted, information on the disruption) and the current operation plan (planned timetable) as input values. The predictive timetable creation unit 101 is, for example, a simulator that performs a macro-type simulation, and calculates the arrival time or departure time of each train at each station or from each station as a decision variable. As a train operation simulator that performs a macro-type simulation, for example, one based on the idea of ​​PERT (Program Evaluation and Review Technique) is known. Note that the predictive timetable creation unit 101 also determines the arrival time or departure time of each train at each station or from each station based on the operation record for the period before the current time, and outputs the result as it is included in the simulation result.

[0021] The predictive timetable creation unit 101 outputs information about an inconsistency (route conflict) in train operation in the predictive timetable, together with the predictive timetable. An inconsistency refers to a situation in which trains cannot run due to route conflict, such as platform conflict where multiple trains enter the same platform at the same time, or a contradiction in the departure order between multiple trains.

[0022] The teacher data generating unit 102 generates information on step-by-step changes to the operation plan that have been considered (input) and executed in the past by an operator such as a dispatcher (i.e., information on the order of changes to the operation plan) as teacher data for the operation plan change judgment model. In this embodiment, the teacher data generating unit 102 generates information on a past disruption that associates information on the train operation before the disruption occurred, changes to the operation plan executed before the disruption occurred, and changes to the operation plan executed in the next stage of the executed operation plan (specifically, the content of the change, and whether or not the change was actually executed after consideration, etc.) with each other.

[0023] In this embodiment, the operation plan change judgment model is a trained model configured as a neural network consisting of multiple input / output layers, but may be other types of trained models.

[0024] Here, multiple operation plan changes made to the same train at the same station (for example, changes to departure and arrival times and changes to departure and arrival platforms) may be regarded as one unit of operation plan change (operation plan changes executed simultaneously), and the same operation plan change for trains with consecutive departure orders at the same station may be regarded as one unit of operation plan change.

[0025] The allowable inconsistency memory unit 1025 stores information on changes to operation plans that were implemented in stages in the past, which caused train inconsistencies but which were resolved by changes in a subsequent stage (hereinafter referred to as allowable inconsistency information).

[0026] The judgment model generation unit 103 generates an operation plan change judgment model based on the teacher data generated by the teacher data generation unit 102. In other words, the operation plan change judgment model receives information about train operation details before a certain point in time and information about changes to the operation plan that were executed before that point in time, and outputs information about changes to the operation plan that will be executed next to the executed operation plan. Note that the information about changes to the operation plan that were executed before the above-mentioned point in time includes information about the type and number of changes to the operation plan that were executed before the above-mentioned point in time.

[0027] In this embodiment, as described above, a change to an operation plan may be a combination of multiple change elements (for example, a change to the arrival and departure time of a certain train and a change to the platform number). The operation plan change support device 100 sets a change to an operation plan consisting of multiple change elements as a single change to an operation plan (hereinafter, referred to as a change group).

[0028] The operation plan change proposal creation unit 104 creates information on changes to the operation plan at least once by sequentially inputting an operation schedule (predicted schedule) based on a new obstruction or an operation schedule (predicted schedule) calculated from the operation plan change judgment model regarding an obstruction specified by a user or a delay that naturally occurs during rush hour or the like (hereinafter referred to as a new obstruction) into the operation plan change judgment model, creates change candidates which are candidates for each of the changes to the operation plan to be executed in stages after the point in time when a new obstruction occurs, identifies change candidates that satisfy specified conditions from each of the created change candidates, and creates the identified change candidates as operation plan change proposals.

[0029] For example, the operation plan change proposal creation unit 104 calculates an evaluation value of the operation plan (predicted timetable) when the changes indicated by each change candidate are applied, judges whether the calculated evaluation value is an optimal value, and if it judges that the evaluation value is an optimal value, identifies each of the change candidates (as a change candidate that satisfies the above-mentioned predetermined conditions). Note that the predetermined conditions to be satisfied include elimination of inconsistencies, operational constraints (time required for trains to turn around at stations, conditions for the departure order of trains, etc.), and plan change rules to be described later.

[0030] When a change candidate has multiple change elements, the change application judgment unit 105 creates multiple combinations (patterns) of the change elements in the change candidate, calculates a change in evaluation value that can represent the effect of the change to the operation plan for each of the multiple combinations created, and determines a combination of change elements to be identified as a change candidate from among the multiple combinations based on the calculated change in evaluation value.

[0031] In addition, based on the allowable inconsistency information, the change application judgment unit 105 judges whether or not the inconsistency caused by the change candidate created by the operation plan change proposal creation unit 104 will be resolved at a later stage, and if it is determined that the inconsistency will be resolved at a later stage, it identifies the created change candidate (as a change candidate that satisfies the above-mentioned specified condition).

[0032] The plan change rule update unit 107 generates or updates plan change rules that are rules that are applied when changing an operation plan and are determined in advance by a manager, etc. The plan change rules are, for example, rules that forcibly eliminate a train delay or inconsistency when there is a train delay or inconsistency in the operation schedule, and are rules that can eliminate a delay or inconsistency even when it cannot be directly controlled by an operation plan change proposal created by the operation plan change judgment model.

[0033] Next, the operation plan change support device 100 stores each data of a change input history DB200, a plan change teacher data DB215, a change judgment model DB220, a plan change rule DB230, an inconsistency history DB240, an allowable inconsistency list DB250, a train schedule DB260, a running record DB270, and an operation plan change proposal DB280. Note that, although it is assumed here that these data are stored as a database, the data storage format is not particularly limited.

[0034] The change input history DB 200 stores the history of stepwise changes to the operation plan that were input in the past by an operator such as a dispatcher.

[0035] (Change input history DB) 3 is a diagram showing an example of the change input history DB 200. The change input history DB 200 has data items of a disruption 201 in which information related to a disruption is set, and a plan change input 202 in which input information related to a change in the operation plan related to the disruption 201 is set.

[0036] Of these, the disruption 201 has the following data items: date 203, which sets the date the disruption occurred; disruption occurrence time 204, which is the time the disruption occurred; disruption end time 205, which is the time the disruption ended; and disruption station 206, which is the location (station, etc.) where the disruption occurred.

[0037] The plan change input 202 has data items such as a train ID 207 in which the ID of a train (change target train) that is the subject of a change in the operation plan due to the disruption 201 is set, a station ID 208 in which the ID of a station (change target station) that is the subject of a change in the operation plan is set, a change type 209 in which the ID of a type of change in the operation plan (change type) is set, a related train 210 in which the ID of another train (related train) that is changed due to the change in the operation plan of the change target train is set, change contents 211 in which specific contents of the change in the operation plan specified by the train ID 207, the station ID 208, the change type 209, and the related train 210 are set, and execution / non-execution 212 in which information on whether or not the change in the operation contents indicated by the change contents 211 has actually been made is set. Note that, in the change type 209, information such as a change in the departure and arrival order of trains at a station and a change in the departure and arrival platform of trains is set.

[0038] Since a single disruption that occurs during train operation causes step-by-step changes to the operation plan, multiple records for a disruption on the same date may exist in the change input history DB 200. The change input history DB 200 may be set automatically, or may be input by an operator such as a dispatcher.

[0039] Next, as shown in FIG. 2, the plan change teacher data DB 215 stores the teacher data generated by the teacher data generating unit 102.

[0040] The change judgment model DB 220 stores the operation plan change judgment model generated by the judgment model generation unit 103.

[0041] The plan change rule DB 230 stores plan change rules. For example, the plan change rule DB 230 stores train operation details when a change is made to an operation plan in association with the corresponding change details.

[0042] (Plan change rules) 4 is a diagram for explaining an example of a plan change rule. When the arrival of a train at a station is delayed due to a delay in another train arriving at the station 231, the arrival order and departure order of both trains at the station are changed. The plan change rule stores information including various order information and delay information between each train before the change in the arrival order or departure order, and various orders between each train (including related trains) after the change.

[0043] Next, as shown in FIG. 2, the inconsistency history DB 240 stores information on the history of occurrence and resolution of inconsistencies in both past train operation details and predicted train operation details (predicted train schedule).

[0044] The allowable inconsistency list DB 250 stores, among the inconsistencies in the inconsistency history DB 240, the history of inconsistencies that occurred but were subsequently resolved (allowable inconsistency information).

[0045] The train schedule DB 260 stores information on past operation plans (past train schedules) and predicted train operation details (predicted train schedules).

[0046] The running record DB 270 stores the past running history (operation record) of each train.

[0047] The operation plan change proposal DB 280 stores the operation plan change proposal created by the operation plan change proposal creation unit 104.

[0048] Each function of the operation plan change support device 100 described above is realized by the calculation processing device 11 reading and executing each program stored in the memory 12 or the storage device 13. In addition, the above-mentioned program can be recorded on a recording medium and distributed, for example. Next, the process performed by the train traffic control system 10 will be described.

[0049] FIG. 5 is a flow diagram illustrating an outline of the train traffic control process performed by the train traffic control system 10. As shown in FIG.

[0050] First, the operation plan change support device 100 executes a teacher data creation process to create teacher data for an operation plan change judgment model in advance (s1). Note that the operation plan change support device 100 may execute the teacher data creation process s1 at a predetermined timing (for example, at a predetermined time such as after the end of daily operation or at a predetermined time interval), or may execute the teacher data creation process s1 at the timing when the change input history DB 200 is updated.

[0051] After that, the operation plan change support device 100 generates an operation plan change judgment model based on the teacher data created in s1. When an obstacle such as a delay or a vehicle breakdown occurs during daily train operations, the operation plan change support device 100 executes an operation plan change proposal creation process s3 to create a new operation plan (operation plan change proposal) using the generated operation plan change judgment model (s3).

[0052] The operation plan change support device 100 displays the proposed operation plan change created in s3 on the display device 15 and requests an operator or the like to input approval. When the approval input is obtained, the operation plan change support device 100 transmits the proposed operation plan change created above to the operation plan management device 40. The operation plan management device 40 changes the operation plan based on the received proposed operation plan change, and further stores the received proposed operation plan change and the current time as a pair as a plan change history. The route control device 60 controls the trains 6 and facilities 5 by transmitting instruction information to the trains 6 and facilities 5 based on the operation plan changed by the operation plan management device 40 (s5).

[0053] Thereafter, the processing from s1 or s3 onwards is repeated. The teacher data creation processing s1 and the operation plan change proposal creation processing s3 will be described in detail below.

[0054] <Teacher data creation process> FIG. 6 is a flow diagram illustrating an example of the teacher data creation process s1. The teacher data generating unit 102 extracts (s11) information (hereinafter referred to as target problem information) about one problem that occurred in the past (hereinafter referred to as target problem) and information about the corresponding change to the operation plan (hereinafter referred to as plan change information) from the change input history DB 200. Specifically, the teacher data generating unit 102 selects all records with the same problem date 203 or problem information 201 from the change input history DB 200, and acquires the contents of the selected records.

[0055] The teacher data generation unit 102 classifies the changes to the operation plans in the plan change information extracted in s11 into a plurality of groups (change groups), and orders the change groups (s13).

[0056] For example, the teacher data generation unit 102 orders each operation plan change according to the order of execution of the operation plan change (the chronological order in which the operation plan management device 40 instructed the route control device 60 to perform control). At this time, the teacher data generation unit 102 classifies each operation plan change into multiple groups for each train to be changed, station to be changed, and target type, and operation plan changes that belong to the same group are considered to be the same operation plan changes (i.e., a change group). In determining the change group, changes that do not depend on the control order related to the running of other trains can be considered to belong to the same group, but each operation plan change may be considered to be one change group. In addition, one change group is considered to include one or more operation plan changes.

[0057] Here, the teacher data generating unit 102 creates train operation details (predicted train schedule) in the event that the target trouble occurs (s15).

[0058] Specifically, the teacher data generation unit 102 creates a predicted timetable by inputting the target disruption information, operation history up to the time the target disruption occurs (running history DB270), and the train timetable (train timetable DB260), which is the operation plan after the target disruption occurs, into the predicted timetable creation unit 101.

[0059] Next, the teacher data generation unit 102 selects, from the change groups classified in s13, the change group that has not been selected so far and is the first in the order (the higher-ranking change group) (s17).

[0060] Then, the teacher data generating unit 102 creates a predicted timetable when the change group (selected group) selected in s17 is applied (s19, s21).

[0061] For example, the teacher data generation unit 102 updates the predicted schedule by inputting the current predicted schedule (the predicted schedule created in s15 or last updated in s21) and changes to the operation plan indicated by the selected group to the predicted schedule creation unit 101.

[0062] Next, based on the information of the predicted timetable before application (the predicted timetable before executing s19 and s21), the teacher data generation unit 102 calculates information regarding the operation content and operation plan of each train before the change in the operation plan, which will become the feature in the teacher data of the operation plan change judgment model (s23).

[0063] These features are features for each train at each station, and include the arrival and departure delay times at each station subject to change for each train (e.g., the train subject to change or related trains that depart before or after the train subject to change) before the change to the operation plan related to the selected group, and information on each change to the operation plan implemented before the change to the operation plan related to the selected group (e.g., information indicating the type of change to the operation plan that was implemented and the stage at which each change is implemented. Note that if the change is the first (first stage) change, this information is not set).

[0064] In addition, this feature amount includes, for example, disruption information of the target disruption corresponding to the selected group obtained from the change input history DB200, or inconsistency information before the change to the operation plan related to the selected group obtained from the inconsistency history DB240.

[0065] The feature amount includes, for example, the number of times that the change target train related to the selected group has been a change target train in the past, calculated from the change input history DB 200. The feature amount includes, for example, information on the number of times that the same type of change to the operation plan as the change to the operation plan related to the selected group has been made in the past, calculated from the change input history DB 200.

[0066] Furthermore, the teacher data generation unit 102 generates information on the changes to the operation plan that have actually been executed among the changes to the operation plan indicated by the selected group, as information that will become output labels in the teacher data of the operation plan change judgment model (s23).

[0067] For example, the teacher data generation unit 102 acquires information on changes applicable to the operation plan related to the selected group and information on whether the changes have actually been made from the change contents 211 and execution status 212 in the change input history DB 200. Note that for changes that are applicable to the operation plan but do not exist in the change input history DB 200, the output label is "no" since the changes have not actually been made.

[0068] Furthermore, the teacher data generation unit 102 compares the predicted timetables before and after the update identified or calculated in s19 and s21, and if an inconsistency has been resolved before and after the change to the operation plan for the selected group, the teacher data generation unit 102 stores the details of the resolution of the inconsistency as allowable inconsistency information in the allowable inconsistency list DB250 (s25).

[0069] For example, the teacher data generation unit 102 acquires all inconsistency information from information output when the predicted timetable is created before s19 and s21 are executed. The teacher data generation unit 102 also acquires all inconsistency information from information output when the predicted timetable is created after s19 and s21 are executed. The teacher data generation unit 102 then compares the acquired inconsistency information and stores the inconsistency information resolved by executing s19 and s21 in the allowable inconsistency list DB250.

[0070] (Allowed inconsistencies) 7 is a diagram for explaining an example of allowable inconsistency information. First, a conflict between the arrival and departure platforms of two trains occurs as an inconsistency 401 (a contradiction in the order of arrival and departure events, such as the platform usage order in which train A arrives after train B departs, the departure order in which train B departs after train A departs, and train A departs after arriving at station P). When a change in the operation plan causes the arrival and departure platforms of these two trains to differ 402, the inconsistency 401 is resolved. The allowable inconsistency information stores such inconsistencies 401 and the trains in the state 403 in which the inconsistency has been resolved, a set of arrival and departure events at stations, and their order relationships.

[0071] Next, as shown in FIG. 6, if there is a change group in the next order (stage) of the currently selected change group, the teacher data generation unit 102 repeats the processes from s17 onwards to select that change group (s27).

[0072] The teacher data generating unit 102 repeats the above process for all errors stored in the changed input history DB 200. This ends the teacher data creation process s1.

[0073] Thereafter, the judgment model generation unit 103 updates the operation plan change judgment model by inputting each piece of teacher data generated in the teacher data creation process s1 into the operation plan change judgment model and performing machine learning.

[0074] <Processing for creating operation plan changes> 8 is a flow diagram illustrating details of the operation plan change proposal creation process s3. The operation plan change proposal creation process s3 is started automatically or in response to a user instruction, for example, when a new disruption occurs in train operation.

[0075] The operation plan change proposal creation unit 104 sets information about a newly occurring problem (hereinafter, referred to as a new problem) (s31). For example, the operation plan change proposal creation unit 104 may accept an input of information about a new problem from a user, or may automatically generate information about a new problem based on information acquired from a predetermined monitoring system or a train tracking device 50, etc.

[0076] The operation plan change proposal creation unit 104 creates predicted information (predicted timetable) of train operation contents after the occurrence of the new obstruction (s33). Specifically, the operation plan change proposal creation unit 104 creates the predicted timetable by inputting the information on the new obstruction set in s31, the operation record (running record DB 270) up until the occurrence of the new obstruction, and the current operation schedule to the predicted timetable creation unit 101.

[0077] The operation plan change proposal creation unit 104 calculates the feature amount (similar to the feature amount calculated in s23) of the operation plan change judgment model from the predicted timetable generated in s33 (or the predicted timetable calculated in s41 described later) (s35). Specifically, the operation plan change proposal creation unit 104 extracts the feature amount of each train at each station based on the data extracted from the predicted timetable.

[0078] The operation plan change proposal creation unit 104 inputs the feature values ​​calculated in s35 into the operation plan change judgment model, thereby acquiring information regarding changes to the operation plan output from the operation plan change judgment model (i.e., change candidates that are judged as “necessary” by the operation plan change judgment model out of the applicable operation plan change candidates) (s37).

[0079] The change application determination unit 105 executes a change application determination process s39 for determining an operation plan that satisfies a predetermined condition and is actually changed from among the changes (change candidates) of the operation plan output in s37. The change application determination process s39 will be described in detail later.

[0080] The operation plan change proposal creation unit 104 creates a next-stage predicted timetable based on the changes to the operation plan (change group) determined in the change application determination process s39 (s41). For example, the operation plan change proposal creation unit 104 creates a predicted timetable by inputting information on the operation plan in which the changes to the operation plan determined in the change application determination process s39 have been applied to the current predicted timetable, and operation results, to the predicted timetable creation unit 101.

[0081] The operation plan change proposal creation unit 104 calculates an evaluation value of the predicted timetable created in s41 (s43). For example, the operation plan change proposal creation unit 104 calculates a total delay time by summing up the arrival and departure delay times of all trains at all stations in the predicted timetable.

[0082] The operation plan change proposal creation unit 104 may calculate other types of evaluation values, such as the number of canceled trains of all trains, the transportation volume of all trains, the impact on operations, the recovery time from the new disruption to returning to the original schedule, or a combination of these. The change in the evaluation value before and after the change in the predicted schedule represents the effect of the change in the operation plan.

[0083] The operation plan change proposal creation unit 104 determines whether the current evaluation value indicates the optimal value (whether it has converged to the optimal value) based on the evaluation value calculated in s43 and the evaluation values ​​calculated in s43 before this time, and identifies (confirms) the changes to each operation plan (change candidates) decided so far in the change application determination process s39 as changes to the operation plan that constitute the operation plan change proposal (s45).

[0084] For example, if the difference between the evaluation value calculated in s45 last time and the evaluation value calculated in s45 this time is equal to or smaller than a predetermined threshold value, or if a maximum value is found from the evaluation values ​​calculated in s45 a predetermined number of times or more, the operation plan change proposal creation unit 104 determines the changes (change candidates) to the operation plans determined in the change application determination process s39 so far as the changes to the operation plans that constitute the operation plan change proposal.

[0085] If the change to the operation plan determined in s39 so far is identified as a change to the operation plan that constitutes the operation plan change proposal (s45: Yes), the operation plan change proposal creation unit 104 executes the processing of s47, and if the change to the operation plan determined in s39 so far is not identified as a change to the operation plan that constitutes the operation plan change proposal (s45: No), the operation plan change proposal creation unit 104 creates a new change candidate for the operation plan based on the timetable prediction result created in s41, that is, repeatedly executes the processing from s35 onwards.

[0086] In s47, the operation plan change proposal creation unit 104 checks whether or not train delays or inconsistencies remain in the predicted timetable as a result of the changes to each operation plan determined in the processes up to this point, and if delays or inconsistencies remain, it forcibly changes the current operation plan to eliminate the delays or inconsistencies.

[0087] For example, the operation plan change proposal creation unit 104 refers to the plan change rule DB 230 to search for changes to the operation plan that correspond to the operation content extracted from the predicted schedule, and applies the searched changes to the operation plan to the predicted schedule.

[0088] Furthermore, the operation plan change proposal creating unit 104 accepts an input of further changes to the operation contents from the user (s49). Note that the operation plan change proposal creating unit 104 may store the input changes in the plan change rule DB 230 as new plan change rules.

[0089] The operation plan change proposal creation unit 104 integrates the changes to the train operation plan determined in the previous processes and generates an operation plan change proposal. The operation plan change proposal creation unit 104 creates change groups for the generated operation plan change proposal in the same manner as in the process of s13 in the teacher data creation process, and displays each change group on a predetermined screen (change proposal screen) (s51). The user modifies the operation plan change proposal as necessary while viewing the change proposal screen.

[0090] After that, the teacher data generation unit 102 acquires the predicted schedule (operation plan) corrected as described above, and updates the teacher data based on the acquired predicted schedule. This ends the operation plan change proposal creation process s3.

[0091] <Change proposal screen> 9 is a diagram showing an example of the change proposal screen 300. The change proposal screen 300 has an operation plan change display field 301, a change execution designation field 302, and a re-proposal field 303.

[0092] In the operation plan change display field 301, a list of operation plan changes (change groups) at each stage constituting the operation plan change proposal generated by the operation plan change proposal creation process s3 is displayed in the order of the change groups (i.e., in the order of train or facility control). Specifically, in the operation plan change display field 301, the contents 311 of the change of each operation plan in the change group 310, the trains 312 to be changed in each operation plan change, the number of inconsistencies 313 existing in the predicted timetable after each operation plan change is applied, and the increase or decrease 314 of the delay time after each operation plan change is applied are displayed. Here, the number of inconsistencies and the delay time are displayed, but the values ​​of other evaluation indexes and change values ​​required when the user decides whether to apply each operation plan change may be displayed. In addition, the operation plan change display field 301 includes an adoption / non-adoption selection field 315 selected by the user when adopting each change group (or each change in the change group) to confirm the operation plan change.

[0093] The change execution specification field 302 is specified by the user when applying (transmitting to the operation plan management device 40) a change group (or a change within a change group) for which the adoption / non-adoption selection field 314 is selected in the operation plan change display field 301.

[0094] When discarding a change group (or a change within a change group) displayed in the operation plan change display field 301, the user specifies the change group to be discarded by deselecting the adoption / non-adoption selection field 315. When the user deselects the adoption / non-adoption selection field 315, the selection in the adoption / non-adoption selection field 315 of change groups (or changes within a change group) that are subsequent (lower) in order than the deselected change group is also automatically deselected.

[0095] The re-proposal field 303 is specified by the user when, after the user has confirmed in the change execution specification field 302 the changes selected in the adoption / rejection selection field 315 from among the change groups (or changes within the change group) displayed in the operation plan change display field 301, the user wishes to execute the processing of the operation plan change proposal creation unit 104 again without making the changes not selected in the adoption / rejection selection field 315.

[0096] <Change application decision process> Next, FIG. 10 is a flow diagram illustrating the details of the change application determination process s39.

[0097] First, the change application decision unit 105 extracts change candidates that were determined in s37 to be necessary and that satisfy the constraint conditions.

[0098] Then, the change application judgment unit 105 selects from the extracted change candidates the change candidate that is the earliest among the change candidates that have not been selected so far (the change candidate related to the station where the train needs to be controlled the earliest) (s71).

[0099] When there are multiple change candidates for the same train at the same station selected in s71 (for example, a platform change and an order change), the change application determination unit 105 selects one of the possible combinations of the change candidates (s73) (when there is only one change candidate, that one is selected). The change application determination unit 105 executes the following processes from s75 to s79 for the selected combination.

[0100] That is, the change application determination unit 105 calculates an evaluation value of the predicted timetable before applying the combination selected in s73 (hereinafter referred to as the selected combination candidate) (hereinafter referred to as the pre-application timetable) and an evaluation value of the predicted timetable after applying the selected combination candidate (hereinafter referred to as the post-application timetable) (s75). In addition, the change application determination unit 105 calculates inconsistency information (the number of inconsistencies) in the pre-application timetable and inconsistency information (the number of inconsistencies) in the post-application timetable.

[0101] Next, the change application determination unit 105 executes a change execution determination process s77 for determining whether or not to determine the selected combination candidate as a change candidate based on the increase or decrease in the evaluation value due to the change in the operation plan related to the selected combination candidate calculated in s75 and the increase or decrease in the number of inconsistencies related to the selected combination candidate. The change execution determination process s77 will be described in detail later.

[0102] The change application determination unit 105 executes the above-mentioned processes of s73 to s77 for all combinations of change candidates related to the running stations of the train selected in s71 (s79).

[0103] Then, the change application decision unit 105 identifies the combination of change candidates with the highest evaluation value from among the combinations of change candidates determined in s77 (s81).

[0104] The change application determination unit 105 selects a change candidate related to the running station of the next (next highest) train, and repeats the process of s71 to s81 (s83). The change application determination unit 105 executes the process of s71 to s83 up to the change candidate related to the running station of the train with the lowest rank among the change candidates extracted at the beginning of the change application determination process s39 (i.e., for all change candidates from the highest to the lowest), and then ends the change application determination process s39.

[0105] <Change execution decision process> 11 is a flow diagram illustrating the details of the change execution determination process s77. The change application determination unit 105 determines the effect of the selected combination candidate by determining the number of inconsistencies in the predicted schedule and the increase or decrease in the evaluation value, and determines whether or not to determine the operation plan related to the selected combination candidate as a change candidate.

[0106] First, the change application judgment unit 105 judges whether or not either of the following conditions is met: the number of inconsistencies for all trains in the predicted timetable is reduced by the change in the operation plan related to the selected combination candidate (the number of inconsistencies in the timetable after application is reduced compared to the number of inconsistencies in the timetable before application), or the number of inconsistencies for all trains in the predicted timetable is not increased or decreased by the change in the operation plan related to the selected combination candidate, but the evaluation value of the predicted timetable is improved (s91).

[0107] If any of the above conditions is met (s91: Yes), the change application decision unit 105 executes the process of s93, and if none of the above conditions is met (s91: No), the change application decision unit 105 executes the process of s95.

[0108] In s93, the change application determination unit 105 determines the selected combination candidate as a change candidate, and the change execution determination process s77 ends.

[0109] In s95, the change application judgment unit 105 judges whether the condition that the number of inconsistencies for all trains in the predicted timetable has increased due to a change in the operation plan related to the application of the selected combination candidate and the evaluation value of the predicted timetable has improved is met.

[0110] If the above condition is satisfied (s95: Yes), the change application decision unit 105 executes the process of s99, and if the above condition is not satisfied (s95: No), the change application decision unit 105 executes the process of s97.

[0111] In s97, the change application determination unit 105 determines not to confirm the selected combination candidate as a change candidate (to exclude it from the change candidates), and the change execution determination process s77 ends.

[0112] In s99, the change application judgment unit 105 judges whether or not a change to the operation plan for the selected combination candidate has increased the number of inconsistencies in the operation of trains related to the selected combination candidate, i.e., whether or not an inconsistency that will hinder the operation of trains related to the selected combination candidate has occurred.

[0113] If the number of inconsistencies in train operations related to the selected combination candidates has increased (s99: Yes), the change application judgment unit 105 executes processing of s97, and if the number of inconsistencies in train operations related to the selected combination candidates has not increased (s99: No), the change application judgment unit 105 executes processing of s101.

[0114] In s101, the change application determination unit 105 determines whether an inconsistency that has increased due to a change in the operation plan related to the selected combination candidate is a resolvable inconsistency. For example, the change application determination unit 105 determines whether the inconsistency has been resolved in the past by searching for allowable inconsistency information in the inconsistency allowable list DB 250. In addition, for example, the change application determination unit 105 determines whether the inconsistency is an inconsistency that has been repeatedly resolved and generated (a predetermined number of times or more) during the repeated processing from s35 to s45 in the processing flow (operation plan change proposal creation processing s3) of the operation plan change proposal creation unit 104.

[0115] If the inconsistency increased due to the change to the operation plan related to the selected combination candidate is a resolvable inconsistency (s101: Yes), the change application judgment unit 105 executes the processing of s93 and terminates the change execution judgment process s77, and if the inconsistency increased due to the change to the operation plan related to the selected combination candidate is an irresolvable inconsistency (s101: No), the change application judgment unit 105 executes the processing of s97 and terminates the change execution judgment process s77.

[0116] As described above, the operation plan change support device 100 of this embodiment generates an operation plan change judgment model that uses information about train operation details before a certain point in time and information about changes to the operation plan executed before that as feature quantities, and uses information about changes to the operation plan to be executed next as output values. When an interruption occurs, a new interruption such as an interruption specified by the user or information about delays that naturally occur during rush hour or the like is input to the operation plan change judgment model, and the information output from this is used to sequentially input further information into the operation plan change judgment model, thereby creating change candidates that are candidates for each of the changes to the operation plan to be executed in stages after a new interruption occurs, and from the created change candidates, a change candidate that satisfies specified conditions is created as an operation plan change proposal.

[0117] In this way, the operation plan change support device 100 of the present embodiment identifies operation plan candidates using an operation plan change judgment model that outputs information on the necessity of stepwise changes to the operation plan to be made in response to an occurrence of a disruption, and from among the candidates, a change that satisfies a predetermined condition is set as an operation plan change proposal. In this way, by modeling the know-how of stepwise changes to the operation plan made by a dispatcher or the like in the past in response to an occurrence of a disruption, it is possible to appropriately change the operation plan and recover from the disruption.

[0118] As described above, the operation plan change support device 100 of this embodiment can support the creation of an operation plan change proposal to be executed in stages.

[0119] In addition, the operation plan change judgment model can output appropriate operation plan changes to be executed in stages with high accuracy by using the type and number of operation plan changes that were executed prior to a certain point in time as input values.

[0120] Furthermore, with regard to the input to the operation plan change judgment model, the operation plan change support device 100 sequentially inputs new operation details (operation details of each train in the predicted timetable) created based on information on changes to the operation plan output from the operation plan change judgment model to the operation plan change judgment model. In this way, by sequentially calculating the train operation details at each stage from the operation plan change judgment model and the predicted timetable, it is possible to appropriately predict changes to the train operation plan that should be executed.

[0121] In addition, the operation plan change support device 100 of this embodiment calculates an evaluation value of the operation plan when the changes indicated by each change candidate are applied, and if it determines that the calculated evaluation value is an optimal value (converges to the optimal value), it identifies each change candidate as a change candidate for the operation plan change proposal.

[0122] This makes it possible to more reliably discover gradual changes to the operation plan that can resolve problems, etc.

[0123] In addition, the operation plan change support device 100 of the present embodiment sets, as operation plan changes to be executed simultaneously, changes that do not depend on the control sequence related to the running of trains, such as multiple operation plan changes to be made at the same station for the same train, or the same operation plan change for trains that depart consecutively from the same station. In this way, by grouping the operation plan changes for each train, it is possible to efficiently find gradual operation plan changes that can more effectively resolve problems, etc.

[0124] In addition, the operation plan change support device 100 of the present embodiment creates multiple combinations of change elements in change candidates, and calculates the change-based fluctuation of the evaluation value for each of the multiple combinations created, thereby determining a combination of change elements to be identified as a change candidate. In this way, even if the change candidate is made up of multiple elements and there are various change patterns for the operation plan, it is possible to identify the most effective combination of change elements.

[0125] Furthermore, when the operation plan change support device 100 of the present embodiment determines that an inconsistency caused by a change indicated by a change candidate will be resolved at a later stage, it specifies the change candidate as a change candidate for an operation plan change proposal, and when it determines that the inconsistency will not be resolved, it excludes the change candidate from the change candidates. In this way, for example, if an inconsistency occurs due to a change candidate at a certain stage but will be resolved later, the inconsistency is tolerated and the change candidate is adopted, thereby increasing the range of selection of change candidates that can be used as an operation plan change proposal, and making it easier to identify an appropriate operation plan change proposal.

[0126] In addition, the operation plan change support device 100 of the present embodiment displays each change in the operation plan change proposal on the screen in the order in which the changes are to be executed step by step. This allows the user to check the changes to the operation plan at each step to be executed.

[0127] In addition, the train traffic control system 10 of the present embodiment includes a route control device 60 that controls trains or train facilities based on the created operation plan change proposal. As a result, even if an interruption occurs and the operation plan needs to be changed in stages, appropriate train traffic control can be performed and operation can be restored.

[0128] The present invention is not limited to the above-described embodiment, and can be implemented using any components without departing from the scope of the present invention. The above-described embodiments and modifications are merely examples, and the present invention is not limited to these contents as long as the features of the invention are not impaired. In addition, although various embodiments and modifications have been described above, the present invention is not limited to these contents. Other aspects conceivable within the scope of the technical idea of ​​the present invention are also included in the scope of the present invention.

[0129] For example, some of the functions provided in each device in the embodiments may be provided in another device, and functions provided in another device may be provided in the same device.

[0130] Furthermore, in the present embodiment, the constraint conditions and the elimination of inconsistencies have been described as the predetermined conditions that the change candidates must satisfy, but other types of constraint conditions may be provided.

[0131] In addition, in the present embodiment, the operation plan change judgment model is configured as one trained model, but may be configured as multiple trained models. For example, an operation plan change judgment model may be provided for each change type of the operation plan, or the operation plan change judgment model may be different depending on the stage at which the change is made in a step-by-step operation plan change. [Explanation of symbols]

[0132] 10 Train operation management system, 100 Operation plan change support device, 103 Judgment model generation unit, 104 Operation plan change proposal creation unit

Claims

1. A processor and a memory, an operation plan change proposal creation unit that inputs information about train operation details before a specified time point and information about changes to the operation plan executed before the specified time point into an operation plan change judgment model that outputs information about changes to the operation plan to be executed next to the executed operation plan, and creates information about changes to the operation plan based on the information output from the operation plan change judgment model, by repeating this process multiple times, creates change candidates that are candidates for each of a plurality of operation plan changes to be executed in stages after the start of creation of the operation plan change proposal, identifies change candidates that satisfy predetermined conditions from each of the created change candidates, and creates an operation plan change proposal from the identified change candidates; an allowable inconsistency storage unit that stores allowable inconsistency information, which is information on changes to operation plans that have been executed in stages in the past, and which have caused train inconsistencies due to the changes but have been resolved by changes in a subsequent stage; a change application determination unit that determines whether or not an inconsistency caused by a change indicated by the created change candidate in a change to the plurality of operation plans can be resolved at a later stage in the change to the plurality of operation plans based on the allowable inconsistency information, and determines that the created change candidate is a change candidate that satisfies the predetermined condition when it is determined that the inconsistency will be resolved at a later stage, and determines that the created change candidate is not a change candidate that satisfies the predetermined condition when it is determined that the inconsistency will not be resolved at a later stage; An operation plan change support device comprising:

2. a judgment model generation unit that receives information about train operation details before a certain time point and information about changes to the operation plan executed before the time point and generates an operation plan change judgment model that outputs information about changes to the operation plan to be executed next to the executed operation plan, The operation plan change support device according to claim 1 .

3. the operation plan change proposal creation unit calculates an evaluation value of the operation plan when the changes indicated by each of the created change candidates are applied, determines whether the calculated evaluation value is an optimal value, and, when it is determined that the evaluation value is an optimal value, identifies each of the change candidates as a change candidate that satisfies the predetermined condition. The operation plan change support device according to claim 1 .

4. The operation plan change support device according to claim 2 , wherein the judgment model generation unit sets a plurality of changes to the operation plan that do not depend on a control order related to train running as changes to the operation plan to be executed simultaneously.

5. 5. The operation plan change support device according to claim 4, wherein the judgment model generation unit sets multiple operation plan changes to be made for the same train at the same station, or the same operation plan change for trains with consecutive departure orders at the same station, as operation plan changes to be executed simultaneously.

6. The change candidate is a combination of a plurality of change elements that constitute a change to the operation plan, The change application determination unit further includes a change application determination unit that creates a plurality of combinations of change elements in the change candidate, calculates a change in an evaluation value representing an effect of the change for each of the created combinations, and determines a combination of change elements to be identified as the change candidate from among the plurality of combinations based on the calculated change in evaluation value. The operation plan change support device according to claim 3.

7. the judgment model generation unit receives input of information on train operation details before a certain point in time, at least one of information on the type or number of changes to the operation plan executed before the point in time, and information on train inconsistencies occurring at the point in time, and generates an operation plan change judgment model that outputs information on changes to the operation plan to be executed next to the executed operation plan. The operation plan change support device according to claim 2.

8. The operation plan change proposal creation unit, Inputting information on train operation contents before a time point designated by a user and information on changes to the operation plan executed before the designated time point into the operation plan change judgment model; creating the change candidate by further inputting, into the operation plan change judgment model, information on new operation content created based on the information on the change to the operation plan output from the operation plan change judgment model, and the information on the change to the operation plan output from the operation plan change judgment model; The operation plan change support device according to claim 1 .

9. The operation plan change assistance device according to claim 1 , further comprising a display unit that displays each of the changes in the operation plan change proposal on a screen in an order in which the changes are to be executed step by step.

10. An information processing device, an operation plan change proposal creation process in which information about train operation contents before a specified time point and information about changes to the operation plan executed before the specified time point are input into an operation plan change judgment model that outputs information about changes to the operation plan to be executed next to the executed operation plan, and information about changes to the operation plan is created based on the information output from the operation plan change judgment model, by repeating this process multiple times, to create change candidates that are candidates for each of a plurality of operation plan changes that will be executed in stages after the creation of the operation plan change proposal is started, a change candidate that satisfies predetermined conditions is identified from each of the created change candidates, and an operation plan change proposal is created from the identified change candidates; an allowable inconsistency storage unit that stores allowable inconsistency information, which is information on changes to operation plans that have been executed in stages in the past, and which have caused train inconsistencies due to the changes but have been resolved by changes in a subsequent stage; a change application determination process for determining whether or not an inconsistency caused by a change indicated by the created change candidate in the change to the plurality of operation plans can be resolved at a later stage in the change to the plurality of operation plans based on the allowable inconsistency information, determining that the created change candidate is a change candidate that satisfies the specified condition when it is determined that the inconsistency will be resolved at a later stage, and determining that the created change candidate is not a change candidate that satisfies the specified condition when it is determined that the inconsistency will not be resolved at a later stage; A method for supporting changes to operation plans.

11. A processor and a memory, an operation plan change proposal creation unit that inputs information about train operation details before a specified time point and information about changes to the operation plan executed before the specified time point into an operation plan change judgment model that outputs information about changes to the operation plan to be executed next to the executed operation plan, and creates information about changes to the operation plan based on the information output from the operation plan change judgment model, by repeating this process multiple times, creates change candidates that are candidates for each of a plurality of operation plan changes to be executed in stages after the start of creation of the operation plan change proposal, identifies change candidates that satisfy predetermined conditions from each of the created change candidates, and creates an operation plan change proposal from the identified change candidates; an allowable inconsistency storage unit that stores allowable inconsistency information, which is information on changes to operation plans that have been executed in stages in the past, and which have caused train inconsistencies due to the changes but have been resolved by changes in a subsequent stage; an operation plan change support device including: a change application determination unit that determines whether or not an inconsistency caused by a change indicated by the created change candidate in a change to the plurality of operation plans can be resolved at a later stage in the change to the plurality of operation plans based on the allowable inconsistency information, and determines that the created change candidate is a change candidate that satisfies the predetermined condition when it is determined that the inconsistency will be resolved at a later stage, and determines that the created change candidate is not a change candidate that satisfies the predetermined condition when it is determined that the inconsistency will not be resolved at a later stage; A train operation management system having a route control device that controls trains or train facilities based on the created operation plan change proposal.

Images