Rescheduling proposal output device

Abstract

To provide a replan output device capable of automatically making a replan that reduces the number of changes from an original plan and profit and loss caused by the changes.SOLUTION: A replan output device for outputting a replan for an operation plan in which scheduled processes are performed in order by the operation of a vehicle in charge, comprises a replan candidate determination unit that when there is a difficult process that is difficult to carry out by the vehicle in charge among the scheduled processes, determines another vehicle as a candidate for an alternative vehicle for the difficult process; and an alternative vehicle selection unit that when a plurality of alternative vehicle candidates are set, selects an alternative vehicle for a difficult process from the plurality of alternative vehicle candidates, taking into consideration the number of times when the vehicle in charge of each process is changed and the additional cost, which are stipulated in the operation plan.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a replanning proposal output device, and as a non-limiting specific example, it relates to a technology for automatically replanning an operation plan using a dedicated processor or the like when it becomes necessary to change the planned operation plan of a transportation facility. [Background technology]

[0002] In public transportation, there are often cases where it becomes impossible to operate as planned due to various unforeseen circumstances (hereinafter referred to as "accidents"), such as malfunctions of vehicles or equipment.

[0003] When an accident like the one described above occurs, it is necessary to re-plan the operation after the accident. Conventionally, the task of re-planning this operation has been largely dependent on the manual work (manual input operation) of operators, etc.

[0004] If it were possible to process the re-planning of such operations mechanically, it would be possible to perform re-planning based on objective indicators that are not dependent on individuals. For this reason, it is believed that there is a growing need for the automation of re-planning.

[0005] The present inventors have conducted extensive research and, based on the knowledge or concept that such re-planning can be solved by a computer as an "optimization problem," have come up with the following idea.

[0006] In other words, in order to quickly solve large-scale optimization problems in the real world, one possible method is to describe the problem using an "interaction model" such as the Ising model, and perform the calculations using a "computational technology dedicated to optimization problems" such as CMOS annealing. When re-planning using an interaction model, a promising method for efficiently handling complex constraints is to generate candidates for subproblems using a column generation method or the like, and then use a method for solving the set partitioning problem or set covering problem.

[0007] As a related art technique, for example, the technique described in Patent Document 1 below can be mentioned. [Prior art documents] [Patent documents]

[0008] [Patent Document 1] JP 2017-151810 A Summary of the Invention [Problem to be solved by the invention]

[0009] Patent Document 1 describes a general set partitioning problem / set covering problem by an interaction model, and describes a configuration for solving the optimization problem quickly. However, Patent Document 1 does not teach a specific policy for handling operation plan re-planning. In addition, when an operation plan is re-planned using the method described in Patent Document 1, many changes occur in the operation plan, and it is considered that the method is not suitable for actual work.

[0010] An object of the present invention is to provide a revised plan output device capable of automatically generating a revised plan that reduces the number of changes from the original plan and the profits and losses that arise due to the changes. [Means for solving the problem]

[0011] According to a preferred aspect of the present invention, there is provided a re-planning proposal output device for outputting a re-planning proposal for an operation plan in which a scheduled route is carried out in sequence by an operation of a vehicle in charge, a candidate vehicle determination unit that, when a difficult journey occurs among the scheduled journeys that is difficult for the vehicle in charge to carry out, determines another vehicle as a candidate for the difficult journey; a vehicle selection unit that selects a vehicle for the difficult route from among the vehicle candidates when a plurality of candidates for the vehicle are set, taking into consideration the number of times the vehicle for each route is changed and additional costs defined in the operation plan; Equipped with. Effect of the Invention

[0012] According to the present invention, when a difficult route occurs that is difficult for the vehicle in charge to carry out, a vehicle in charge of the difficult route can be selected from among a plurality of candidates for the vehicle in charge, taking into consideration the number of times the vehicle in charge has been changed and the additional costs. Therefore, a rescheduled plan that reduces the number of changes from the original plan and the profit and loss caused by the changes can be automatically created. [Brief description of the drawings]

[0013] [Figure 1] 1 is a block diagram showing an overview of an information processing system in which a revised plan output device of the present invention is implemented. [Diagram 2] 2 is a flowchart showing a process flow of the information processing system shown in FIG. 1. [Diagram 3] 2 is a diagram illustrating information stored in an input / output support unit in the information processing system. FIG. [Figure 4] FIG. 13 is a diagram illustrating an example of a daily operation plan (original allocation plan). [Diagram 5] FIG. 13 is a diagram for explaining an example of a constraint condition. [Figure 6] 3 is a flowchart corresponding to a subroutine of step S2 in FIG. 2. [Figure 7] FIG. 13 is a diagram showing an example of a directed graph that explains a case where an accident occurs to a vehicle 3 during the execution of an operation plan, and a route 8 becomes a difficult route to execute. [Figure 8] 11 is a flowchart illustrating an example of a process flow for generating various directed graphs. [Figure 9] 8 is a directed graph (change candidate graph) for explaining a state in which replan candidates and route candidates are generated based on the directed graph shown in FIG. 7. [Figure 10] 7 is a flowchart corresponding to a subroutine of step S22 in FIG. 6. [Figure 11] This is a directed graph to explain the differences in the number of changes, additional costs, etc. between two replanning candidates (route candidates). [Figure 12]FIG. 13 is a diagram showing an example of a plan candidate (change candidate table). [Figure 13] 11 is a directed graph for explaining an example of a processed plan candidate, plan implementation information, and change information. [Figure 14] FIG. 13 is a diagram showing an example of a portion of the plan candidate (change candidate table) in FIG. 12 that becomes a matrix (Aij) and a vector (Cj) that become coefficient values ​​of an interaction model. [Figure 15] FIG. 13 is a diagram showing an example of a portion of the change information that becomes a matrix (Dij) that becomes coefficient values ​​of an interaction model. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] The embodiments of the present invention will be described in detail with reference to the drawings. However, the present invention should not be interpreted as being limited to the description of the embodiments shown below. It will be easily understood by those skilled in the art that the specific configuration can be changed without departing from the concept or purpose of the present invention.

[0015] In the configurations of the invention described below, the same parts or parts having similar functions are denoted by the same reference numerals in different drawings, and duplicated descriptions may be omitted.

[0016] When there are multiple elements having the same or similar functions, they may be described by using the same reference numerals with different subscripts. However, when there is no need to distinguish between multiple elements, the subscripts may be omitted.

[0017] In the following description, the term "storage" refers to at least one of memory and persistent storage. The term "memory" refers to one or more memory devices, which is an example of one or more storage devices, and may typically be a primary storage device.

[0018] In the following description, a "processor" may refer to one or more processor devices. The at least one processor device may typically be a microprocessor device such as a CPU (Central Processing Unit).

[0019] Furthermore, in the following description, for the sake of convenience, it may be described that a software block mainly executes processing, but it should be noted that the actual processing is mainly executed by a processor.

[0020] <<Outline of this embodiment>> An embodiment of the present invention can be realized (implemented) within a computer or computer system as a rescheduling proposal output device that outputs a rescheduling proposal for an operation plan in which the scheduled itinerary is carried out in sequence by the operation of a vehicle in charge.

[0021] Such a replanning proposal output device includes a replanning candidate determination unit that, when a difficult process occurs within the scheduled journey that is difficult for the vehicle in charge to carry out, reorganizes the plan from the time when the difficult process occurred, and generates candidates for a plan in which the difficult process is to be carried out by another vehicle.

[0022] In addition, when multiple substitute vehicle candidates are set, the re-planning proposal output device is equipped with a substitute vehicle selection unit that selects a substitute vehicle for a difficult journey from the multiple substitute vehicle candidates, taking into consideration the number of changes in the vehicle responsible for each journey specified in the operation plan and additional costs.

[0023] The re-planning candidate determination unit and the substitute vehicle selection unit can be configured by the same processor or different processors.

[0024] Of the above, it is desirable to store the operation plan in a predetermined memory so that it can be referred to, processed, and further used to create and output a revised plan by the revised plan output device.

[0025] As a specific example (format) of each "journey" included in the operation plan, the departure time from the starting point and the arrival time to the end point are specified. In the following explanation, it is assumed that the end point of the previous journey and the starting point of the following journey are the same for each journey performed by the same vehicle, and an interval (such as a rest time) can be set between these journeys.

[0026] In this case, the re-planning candidate determination unit can determine other vehicles performing a journey that does not overlap with the time period from the departure time to the arrival time of the difficult journey to be performed as candidates for substitute vehicles for the difficult journey to be performed.

[0027] Regarding the vehicles used in the operation plan, there are no particular limitations on their types, numbers, etc., and many types and numbers of vehicles may be used in actual operations. However, the more types and numbers of vehicles there are, the more complicated the explanation becomes, so the following explanation will be given using an example of an operation plan that is as simplified as possible.

[0028] Among the above, one specific example of an "other vehicle" that is a candidate for a substitute vehicle to take over a difficult journey is a vehicle among the multiple vehicles specified in the operation plan other than the vehicle that is having difficulty completing the journey. The number of candidate vehicles (units) determined by the replanning candidate determination unit may be, at most, the sum of all vehicles assigned to the operation plan for that day and the number of standby vehicles. If the operation plan for that day cannot be operated using only the assigned vehicles, standby vehicles will be assigned. However, if there are no standby vehicles and no vehicle can be determined (if the number of determined vehicles is zero), none of the vehicles specified in the operation plan can substitute for the difficult route that has occurred, and therefore another vehicle not specified in the operation plan must be prepared.

[0029] On the other hand, normally, the more candidate vehicles (number of units) determined by the re-planning candidate determination unit, the more processing (amount of calculations) required for the proxy vehicle selection unit to select a proxy vehicle will be, and the more complicated the explanation will be, so the following explanation will be given using examples that are as easy to understand as possible.

[0030] On the other hand, if the number of candidate vehicles determined by the re-planning candidate determination unit is small (for example, two or three), there may be cases where a substitute vehicle can be selected based on the human judgment of an operator or other person, without the substitute vehicle selection unit having to perform the selection process (calculation, etc.) of the responsible vehicle.

[0031] In consideration of the above circumstances, it is believed that the re-planning proposal output device of this embodiment is particularly useful in situations where the number of candidate vehicles determined by the re-planning candidate determination unit is large and humans cannot quickly determine which vehicle should carry out (act as a proxy) a difficult process.

[0032] In such a case, the proxy vehicle selection unit may be configured to select a vehicle that optimizes the solution of the interaction model using the number of changes and the additional cost as a proxy vehicle for the difficult journey.

[0033] Alternatively, from the viewpoint of minimizing changes to the contents of the operation plan, the substitute vehicle selection unit may be configured to select from a plurality of substitute vehicle candidates a vehicle that has been changed the least number of times for each route specified in the operation plan as a substitute vehicle for a difficult route.

[0034] On the other hand, from the viewpoint of minimizing additional costs incurred due to changes in the operation plan (e.g., various losses such as extra fuel consumption, labor costs, and time delays), the substitute vehicle selection unit should be configured to select a vehicle that will result in the least additional costs from multiple substitute vehicle candidates as a substitute vehicle for a difficult journey.

[0035] The above-mentioned method of selecting a proxy vehicle for a difficult journey by the proxy vehicle selection unit can be switched depending on the purpose, or the weights can be fine-tuned by changing the coefficients of the mathematical formula used (such as an interaction model).

[0036] Furthermore, in order to make the processes of determining rescheduling candidates and selecting substitute vehicles (and ultimately outputting rescheduling proposals), which tend to become complicated in actual operations, easier to understand visually, it is advisable to configure the system with a graph generation unit that generates a directed graph in which vehicle nodes specified in the operation plan and journey nodes assigned to each vehicle are connected by directed links in the order of execution.

[0037] In this case, the re-planning candidate determination unit is configured to connect directed links in the directed graph from the node related to the determined candidate substitute vehicle to the nodes of the journey related to the re-planning, including the difficult-to-execute journey, or to the nodes of all planned journeys (hereinafter referred to as journey nodes), thereby causing the graph generation unit to generate a graph as a candidate for the re-planning proposal.

[0038] On the other hand, the substitute vehicle selection unit is configured to cause the graph generation unit to generate a graph as a revised plan by connecting a directed link from the node related to the selected substitute vehicle to the journey node in the directed graph.

[0039] With the above-described configuration, the user (operator, etc.) can easily see vehicle breakdowns, difficult routes to complete, and even candidate substitute vehicles to take over difficult routes, the current status of selected substitute vehicles, future schedules, etc., which is expected to lead to smooth execution of replanning.

[0040] Each of the above-mentioned configurations will be described in more detail below with reference to the drawings.

[0041] <<Overall system overview>> FIG. 1 is a block diagram showing an example of the configuration of an information processing system according to this embodiment. The information processing system 10 shown in Fig. 1 can be configured by one physical computer. Alternatively, each component of the information processing system 10 may be realized by a separate physical computer. Furthermore, the information processing system 10 may be another type of system, for example, a system (for example, a cloud computing system) realized on a group of physical computing resources (for example, a cloud platform), instead of a computer system having one or more physical computers. An information processing system 10 shown in FIG. 1 corresponds to the "re-plan output device" of the present invention.

[0042] 1, the information processing system 10 includes an input / output support unit 20, a processing support unit 50, and an optimization calculation unit 80. Of these, the processing support unit 50 includes a calculation unit 40 and a storage unit 60.

[0043] Among the above, the calculation unit 40 has the functions of the above-mentioned "re-planning candidate determination unit" and "substitute vehicle selection unit", and in particular, a plan candidate creation unit 141 described later assumes these roles.

[0044] 1, input and output of information to and from the information processing system 10 is performed via an input / output support unit 20. Referring to Fig. 3, the input support unit 20 is connected to an input interface (operation input unit such as a keyboard or mouse, a communication unit such as a modem, etc.) of an external device such as a PC, receives information input by a user via the input interface, stores the received information in an input data storage unit 30 described later, and supplies the information to a processing support unit 50.

[0045] In addition, the input / output support unit 20 is connected to an output interface (a display unit such as an LCD, a printing unit such as a printer, an external memory such as an HDD, a communication unit such as a modem, etc.) of an external device such as a PC, and can receive or visually view data on the rescheduled plan automatically prepared by the information processing system 10 via such output interface.

[0046] Although not shown in FIG. 3, the input / output support unit 20 may include a hardware processor for carrying out the above-mentioned processes while communicating with an external device such as the above-mentioned PC.

[0047] Furthermore, referring to FIG. 3, the input / output support unit 20 includes an input data storage unit 30 that stores, as input data, the plan that was originally formulated (hereinafter referred to as the "original plan") and information about the accident that occurred, and an output data storage unit 90 that stores, as output data, information about a replanning proposal based on the original plan; details of this configuration and the stored information will be described later.

[0048] The optimization calculation unit 80 uses a "computational technology dedicated to optimization problems" such as CMOS annealing to process the optimization problem of replanning at high speed. Here, the optimization problem refers to a problem of finding the value of a variable that minimizes or maximizes the cost, which is the evaluation value, under required constraints. The computational technology dedicated to optimization problems is an ultra-high-speed dedicated simulator of an interaction model described in the form of the following formula 1, such as an Ising model or a Boltzmann machine.

[0049]

number

[0050] In formula 1, x and h are column vectors, and J is a symmetric matrix. In formula 1, x is a variable. The type of the variable x depends on the nature of the hardware, so it is sufficient to model it appropriately according to the nature of the hardware. Note that the subscript T is a code indicating a transposed matrix. Furthermore, in Formula 1, J and h are each a constant, and the width (magnitude) of these constants, whether J and h are integers or real numbers, and the like also depend on the properties of the hardware, but are not particularly limited thereto, and in addition, are less relevant to the features of this embodiment.

[0051] Generally, in computer-related technologies that exclusively calculate optimization problems (hereinafter, sometimes referred to as "computational technologies dedicated to optimization problems"), the optimal value of variable x is found by searching or reproducing on a computer the minimum energy state of the interaction model described in Equation 1 (the value of x that minimizes H on the left side of Equation 1). In other words, the calculation technology dedicated to optimization problems performs calculations using input values ​​such as the interaction coefficient of the interaction model (the value corresponding to J in Equation 1) and the external field coefficient (the value corresponding to h in Equation 1), and calculates and outputs the variable x that minimizes H.

[0052] Returning to FIG. 1, the optimization calculation unit 80 is an example of a parallel processing device (a device capable of parallel processing), and serves as an accelerator for solving the allocation problem by calculating an interaction model (e.g., an Ising model) as described above in Equation 1.

[0053] The optimization calculation unit 80 is typically a hardware circuit such as an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), or a Graphics Processing Unit (GPU). The optimization calculation unit 80 may be realized by a CPU (Central Processing Unit) having one or more physical or virtual cores.

[0054] Alternatively, the optimization calculation unit 80 can be, for example, a semiconductor computer (for example, a CMOS annealing machine) that simulates the operation of an Ising model using a semiconductor circuit (for example, a CMOS (Complementary Metal Oxide Semiconductor) circuit). The optimization calculation unit 80 may be realized by, for example, an algorithm (for example, momentum annealing) that runs on a GPU.

[0055] Furthermore, the optimization calculation unit 80 may use an annealing type quantum computer or a general-purpose quantum computer instead of a digital computer.

[0056] In the information processing system 10 of this embodiment, the processing support unit 50 includes a calculation unit 40 implemented on a processor, and a storage unit 60 for storing intermediate data (hereinafter, for ease of distinction and explanation, referred to as "the intermediate data"). The main functions of the processing support unit 50 in this embodiment are to receive information input by a user and convert it into an input (interaction model) for the optimization calculation unit 80, and to reconvert the output result of the optimization calculation unit 80 into a format that is easy for the user to view and return (output).

[0057] The calculation unit 40 of the processing support unit 50 includes a plan candidate creation unit 141, a model generation unit 142, and a compilation processing unit 143, as shown in FIG. Of these, the plan candidate creation unit 141 generates the above-mentioned intermediate data that becomes a plan candidate based on the input information received from the input / output support unit 20. Therefore, the plan candidate creation unit 141 may be called an "intermediate data generation unit."

[0058] On the other hand, the model generation unit 142 uses the intermediate data to calculate interaction coefficients and external field coefficients of the interaction model that are input to the optimization calculation unit 80, and performs processing to send the calculation results to the optimization calculation unit 80. Furthermore, the aggregation processing unit 143 uses the intermediate data and the output result of the optimization calculation unit 80 to generate output data, and performs processing to return the generated output data to the input / output support unit 20.

[0059] Here, the above-mentioned "intermediate data" (stored in the storage unit 60) can be broadly divided into the following three types of information: (60-1), (60-2), and (60-3). (60-1) Plan candidate 161 (60-2) Plan execution information 162 (60-3) Change information 163 Specific examples of the above information (60-1) to (60-3) will be described below.

[0060] <60-1: Candidate Project> In this embodiment, when re-planning a plan using an interaction model, a column generation method or the like is used to generate multiple plan candidates as a method for efficiently handling complex constraint conditions (processing in the plan candidate creation unit 141). Specifically, the plan candidates are, for example, candidates for which vehicles / equipment should be used for which flights.

[0061] The candidate plan 161 is information that lists candidates that satisfy some or all of the required constraints among the various candidates described above. The candidate plan 161 is data that is composed of binary or integers so that it can be directly used to generate an interaction model. The candidate plan 161 can also be used in an interaction model for solving a general set covering problem / set partitioning problem.

[0062] <60-2:Plan execution information> The above-mentioned candidate plan 161 includes information such as which vehicle / equipment will be used for which flight, but does not include any more detailed information (such as various information required for actual operation). This is because the candidate plan 161 is information for the purpose of calculating the coefficients of the interaction model.

[0063] In this embodiment, information necessary for actual operation is handled (separately) by the plan execution information 162, and this is one of the features of this embodiment.

[0064] Specifically, when the replanning causes a change in the flight's departure and arrival time, the plan execution information 162 includes information such as the changed time, where the flight departs and arrives from, reservation information in the case where there is a reservation, the size of the vehicle / aircraft required to operate the flight, etc. In general, the plan execution information 162 can be defined as "comprehensive information required to execute the re-drafted plan."

[0065] <60-3:Change Information> The change information 163 is information introduced to adjust the change rate caused by replanning, and is one of the features of this embodiment. In this embodiment, the change information 163 is numerical data having the same number of elements as the above-mentioned plan candidate 161. In one specific example, the change information 163 is data that quantifies whether or not each candidate constituting the plan candidate 161 has deviated from the original plan, or to what extent it has deviated.

[0066] The above-mentioned plan candidate 161, plan execution information 162, and change information 163 can each be stored in the storage unit 60 as database data (see FIG. 1).

[0067] <<Example of outputting a replan of the current day's operation plan>> The following describes how to re-plan the day's route bus operation plan with reference to the flowchart in Figure 2.

[0068] Note that the embodiment described below deals with a route bus as an example only, and the present invention is not to be interpreted as being limited to the following description. Therefore, for example, as another example, the bus vehicle described below may be replaced with an aircraft (such as a passenger plane) in aviation. Also, as another example, the bus route described below may be replaced with an airline flight in aviation.

[0069] This embodiment can be used when it becomes necessary to re-plan an operation plan in real time when one or more steps in the operation plan for one day cannot be executed by the vehicle in charge due to an error such as an accident or a vehicle breakdown (so-called an emergency occurs).Alternatively, this embodiment can be used when there is enough time and real-time performance is not required, but the operation plan has been created in advance according to certain rules, and the operation plan needs to be reorganized without violating those rules as much as possible.

[0070] In this embodiment, it is assumed that for some reason it becomes difficult to return one of the route buses to service on the same day, and the remaining vehicles (buses) take over (perform the itinerary planned in the operation plan for that day (plan candidate 161 described above).

[0071] First, an overview of the processing procedure up to output of a revised plan in the information processing system 10 will be described with reference to the flowchart shown in FIG. 2, and the details of each step (steps S1 to S6) will be supplementarily described later.

[0072] In step S1, the information processing system 10 first reads input data. In one specific example, the input data is read by reading various information related to the original plan transmitted from an external device such as a PC by the operation of the processor of the input / output support unit 20, as described above in FIG. 1 and FIG. 3. The read input data may be stored in an input data storage unit 30 such as a HDD in the input support unit 20 so as to be classified (categorised) according to the type of information, as shown in FIG. 3. In step S1 shown in FIG. 2, it is assumed that the input data includes accident information, which will be described later. In other words, there is no need to generate a new plan while no accident or difficult execution process has occurred.

[0073] In the next step S2, the processing support unit 50 (plan candidate creation unit 141 shown in FIG. 1) of the information processing system 10 uses the input data stored in the input data storage unit 30 to list the above-mentioned plan candidates, plan execution information, and change information. At this time, the plan candidate creation unit 141 outputs intermediate data so that the listed information is classified (separated) into plan candidates 161, plan execution information 162, and change information 163 according to the type of information, as shown in FIG. 1, and stored in the storage unit 60. Note that further details of the processing content in step S2 will be described later in the explanation of FIG. 6 and subsequent figures.

[0074] Hereinafter, the plan candidates, plan execution information, and change information at the stage before they become the above intermediate data (plan candidate 161, plan execution information 162, change information 163) may be collectively referred to as "intermediate information."

[0075] In the next step S3, the calculation unit 40 of the processing support unit 50 uses the intermediate information generated (listed) in the previous step to perform a model creation process, in particular, to calculate the interaction coefficients and external magnetic field coefficients of the interaction model. Such calculations are performed by the model generation unit 142 of the calculation unit 40 (see FIG. 1). Further details of steps S2 and S3 will be described later.

[0076] In the subsequent step S4, the information processing system 10 passes the interaction model generated in the previous step from the model generation unit 142 shown in FIG. 1 to the optimization calculation unit 80, and the optimization calculation unit 80 executes optimization calculations to optimize the interaction model (e.g., minimize costs).

[0077] In the next step S5, the optimization calculation unit 80 of the information processing system 10 passes (sends) the optimized (e.g., minimized) solution calculated in the previous step as an output result to the calculation processing unit 143 of the calculation unit 40. Then, the calculation processing unit 143 performs calculation processing to create output data, and passes (sends) the output data from the calculation processing unit 143 to the input / output support unit 20.

[0078] In one specific example of this tabulation process, the tabulation processing unit 143 uses the result of the optimization calculation in step S4 (e.g., the minimized solution) and the above-mentioned intermediate data (plan candidate 161, plan execution information 162, change information 163) to format, i.e., arrange the data into a format that is easy for the user to view. In addition, as shown in Fig. 3, it is preferable to provide an output data storage unit 90 in the input support unit 20, and store the output data in the output data storage unit 90 so as to classify (categorise) the data according to the type of information (in the example shown in Fig. 3, a revised plan 191, change costs 192, and change locations 193).

[0079] In the next step S6, the information processing system 10 outputs the replanned plan 191 in the output data storage unit 90 to an external device such as a PC through processing by the processor of the input / output support unit 20, and ends the processing.

[0080] Below, we will supplement some of the steps in the flowchart in Figure 2.

[0081] <1. Input / Output Data> As an example of the input data to be read in step S1, the following information (30-1), (30-2), and (30-3) may be prepared as shown in the input data storage unit 30 in FIG.

[0082] (30-1) Original operation plan 131 In this embodiment, the original operation plan 131 is the information that is the subject of replanning (more specifically, ) output by the information processing system 10, and typically refers to the information of the operation plan that was created at the beginning (original) of its formulation.

[0083] (30-2) Accident Information 132 The accident information 132 may include, for example, the time when an accident or the like that caused the rescheduling occurred, and information on vehicles that became unable to operate.

[0084] (30-3) Constraint condition 133 The constraint conditions 133 include, for example, conditions that are required when creating a normal operation plan.

[0085] Next, a specific example of the “original operation plan 131” described in (1) will be described with reference to FIG. As shown in Fig. 4, a specific example of the original operation plan 131 may be a train schedule in a matrix format. In the example shown in Fig. 4, information on the itinerary (schedule) of each of vehicles 1 to 8 assigned to a specific day is recorded.

[0086] For the sake of simplicity, FIG. 4 shows a plan for eight vehicles with up to six journeys for each vehicle, but in reality, the schedule can include information on more vehicles and more journeys for each vehicle.

[0087] For ease of explanation, in the example shown in Fig. 4, vehicle 8 is not assigned a route as an accident response vehicle, but the present invention is not limited to this and any route may be assigned to the vehicle. Conversely, a configuration in which the number of accident response vehicles (vehicles to which no route is assigned) is increased may be used.

[0088] On the other hand, if the number of vehicles with no itinerary assigned is increased for accident response purposes, it becomes easier to calculate a new plan, but other problems arise, such as vehicle management and standby drivers, etc. For this reason, in this embodiment, it is assumed that the original operation plan 131 is planned to eliminate or reduce as few vehicles as possible that are not assigned to a itinerary in a day.

[0089] In addition, in the example of the original operation plan 131 shown in FIG. 4, the journey corresponding to each vehicle (vehicle 1 to vehicle 8) is represented in a table format, but this is just one example, and any configuration that can show which vehicle is responsible for which journey may be used, and as another example, it may be data for creating an operation diagram.

[0090] In other words, the numbers 1 to 24 shown in Figure 4 are simply serial numbers assigned to distinguish (identify) the routes (from the departure point to the destination) of vehicles operating (moving from the departure point to the destination) on a given day.

[0091] In addition, in the example of the original operation plan 131 shown in FIG. 4, each journey assigned to a vehicle number is executed in the order of the journey with the lowest number. For example, journey 1, journey 2, journey 3, and journey 4 assigned to vehicle 1 are executed in the order of journeys 1, 2, 3, and 4 in terms of operation time. Also, for example, journey 4 is executed earlier than journeys 5, 7, 9, 12, 18, and 21. As described above, intervals (for example, driver rest time, work time associated with driver change, etc.) can be provided between each journey (see journeys 1 to 4 in FIG. 7 as appropriate).

[0092] 4, in multiple journeys assigned to one vehicle and arranged in one row, the destination of journey n (the end of the route) and the departure point of journey n+1 (the start of the route) have the same relationship. For example, if the destination of journey 1 for vehicle 1 is point B, the departure point of journey 2 (the start of the route) is also point B.

[0093] In general, the trips assigned to vehicle numbers (one vehicle) placed in a row of the original operation plan 131 (matrix table) shown in Figure 4 represent the number of routes scheduled to be executed on a particular day and the order in which they are scheduled to be executed.

[0094] In contrast, the trips arranged in the same column in the table shown in Fig. 4 are independent of each other and have no temporal relationship with each other. For example, trip 7 of vehicle 3 may be performed before trip 5 of vehicle 2 (see Fig. 7, etc. as appropriate).

[0095] In light of the above, in this specification, a "journey" refers to something that has at least information on the operating route and operating time. Even if different terms are used depending on the industry or business, anything that has information on the operating route and operating time falls under the category of a "journey." In addition, the target period for replanning is in daily (1 day) units in this embodiment, but is not limited to this and may be, for example, in multiple day units, weekly units, monthly units, etc.

[0096] Next, the "accident information 132" as the information of (30-2) will be described in more detail. In one specific example, the accident information 132 may include information on a vehicle that has become inoperable, a route number that the vehicle has not yet completed (moved), or the time when the vehicle became inoperable. Note that the "time" information may be later than the time when the inoperable vehicle was discovered, and may be any time that can identify or determine the above-mentioned uncompleted route number, in other words, the route that needs to be replanned.

[0097] For convenience, the time input as accident information will be referred to as the "accident occurrence time" below. Also, the inoperable vehicle input as accident information will be referred to as the "accident vehicle." Also, although it will be referred to as an accident here for convenience, the cause of the need for replanning may include causes other than accidents and is not limited to accidents.

[0098] Next, the "constraint condition 133" as the information in (30-3) will be described in more detail. The constraint condition 133 is, for example, detailed information that is the breakdown of each journey required to carry out an operation. The constraint condition 133 is, for example, as shown in FIG. 5, the departure point and arrival point (hereinafter, these may be collectively referred to as "departure and arrival points") that are the breakdown (starting point and ending point) of the route of each journey (journey 1, 2, 3, ...), the departure time departing from the departure point and the arrival time stopping (arriving) at the arrival point (hereinafter, these may be collectively referred to as "departure and arrival times"), the type of vehicle required for the journey, and the like.

[0099] In the following explanation, it is assumed that the type of vehicle is the same concept as the vehicle ID that identifies each vehicle, and that different numbers are assigned to different vehicles. However, additional information may be included, such as whether the vehicle is large or small, the maximum number of passengers, the maximum loading weight, and the vehicle height. In this case, for example, the vehicle type number (1, 2) may be expressed by adding a sub-number (for example, vehicle 1-1-30, vehicle 2-2-5, etc.).

[0100] Furthermore, the constraint conditions 133 include not only detailed information about each journey, but also constraint conditions required for creating the original operation plan. For example, interval times between journeys (constraint conditions according to regulations, such as stopping for 10 minutes or more at bus stops at large railway stations) and the time of the last bus (constraint conditions such as not operating after 11 p.m. due to work regulations, etc.) are also input if there are any. The constraint conditions 133 may be stored in the input data storage unit 30 in advance in preparation for accidents.

[0101] On the other hand, as an example of the output data, the following information (90-1), (90-2), and (90-3) are stored in the output data storage unit 90 in FIG.

[0102] (90-1) Replanning Proposal 191 In this embodiment, the re-plan proposal 191 is information indicating one or more changes to the original plan. Note that specific aspects of the re-plan proposal 191 will be described later in the explanation of FIG.

[0103] (90-2) Change cost 192 In this embodiment, the change cost 192 is information indicating the cost (loss of money, time, etc.) incurred due to replanning, i.e., changing the original plan, and corresponds to the "additional cost" of the present invention. When multiple replans are generated, the change cost 192 is indicated for each replan.

[0104] (90-3) Changes 193 In this embodiment, the change portion 193 is information indicating the change content, that is, the difference (changed portion) between the original operation plan 131 and the rescheduled plan 191. In one specific example, the change portion 193 includes the difficult operation schedule that occurred and various information related thereto (time of accident, vehicle in which the accident occurred, selected substitute vehicle, etc.). In this system, it is desirable that this information is output (displayed, printed, etc.) as visually understandable information, as described later.

[0105] Specific aspects (output examples, etc.) of the output data (replanned plan 191, change cost 192, change location 193) will be described later in the explanation of FIG. 7 and subsequent figures.

[0106] <2. List of plan candidates, information for implementing the plan, and change information> The above-mentioned "intermediate information" can be generated using a method such as a column generation method. Below, an example of a procedure for generating intermediate information and a resulting data format will be described. The procedure for generating intermediate information described below is merely an example, and there are no particular limitations on the procedure and output format in which the intermediate information is generated.

[0107] An example of the procedure for generating the intermediate information will be described with reference to the flowchart of Fig. 6 and the graph of Fig. 7. The flowchart shown in Fig. 6 corresponds to the subroutine of step S2 described above with reference to Fig. 2.

[0108] After step S1 in FIG. 2 (reading of input data) has been performed, in step S20 in FIG. 6, the processor of the information processing system 10 (the plan candidate creation unit 141 in the calculation unit 40 of the processing support unit 50, step S22 is also the same) first creates data in which the connections between each vehicle and each journey are represented as a directed graph (hereinafter simply referred to as a "directed graph"), as shown in FIG. 7.

[0109] It can be seen that the directed graph shown in Fig. 7 can be created based on the above-mentioned train schedule 131 in Fig. 4, the constraint conditions in Fig. 5, and the accident information 132. In addition, in the example shown in Fig. 7, the accident information 132 indicates that vehicle 3 became an accident vehicle at 15:00, in other words, the trip 8 cannot be completed by vehicle 3.

[0110] The graph to be created does not necessarily need to be drawn in a picture or visible form as in Fig. 7, and it is sufficient to store information on each of the nodes 101 and 102 and each of the links 103 and 104 as digital data in a storage unit. Next, the processor of the information processing system 10 adds a link (a new directed link 103-Ch, described later) to the directed graph created as in Fig. 7, according to the input constraints. This process will be described later in the explanation of Figs. 8 and 9.

[0111] Furthermore, in step S22, the processor of the information processing system 10 generates the above-mentioned intermediate information (intermediate data) by performing a graph search that follows the added links, stores or updates the generated intermediate information (161 to 163) in the memory unit 60, and terminates the processing of the flowchart in Figure 6.

[0112] Next, specific examples of the procedures of step S20 (creating a directed graph) and step S22 (listing plan candidates, plan implementation information, and change information by route search on the directed graph), which are subroutines in the flowchart of FIG. 6, will be described.

[0113] <2-1. An example of a procedure for generating a directed graph> First, an overview of how a directed graph is generated will be described with reference to FIG.

[0114] The directed graph includes a vehicle node 101, a journey node 102, and an arrow-shaped link (directed link 103 or dotted link 104) connecting these nodes. The example shown in FIG. 7 is a graph assuming that 24 journeys, journey 1 to journey 24, are performed using eight vehicles, vehicle 1 to vehicle 8, according to the original operation plan 131 described in FIG. 4 and the constraints 133 described in FIG. 5. Vehicles do not need to be different nodes for each individual vehicle, and different vehicles can be represented by the same node if, for example, they have the same starting point or are the same type, and the routes of the graph search are considered to be the same or almost the same. This embodiment will be described assuming that a different node is created for each individual vehicle.

[0115] However, to avoid complicating the explanation, FIG. 7 omits the illustration of vehicle nodes 101 for vehicles 4 to 7 and route nodes 102 for routes 9 to 24 of the original operation plan 131 assigned to vehicles 1 to 7.

[0116] In this embodiment, the directed link 103 is used as an arrow from a vehicle node 101 to a journey node 102, or as an arrow from one journey node 102 to another journey node 102 (a later journey node 102 that does not overlap in time).

[0117] Note that the directed links 103 do not connect vehicle nodes 101 to each other, and the length of the directed links 103 has no technical meaning.

[0118] The basic meaning of the directed link 103 described above also applies to the dotted line link 104 used in the event of an accident.

[0119] The difference between the directed link 103 and the dotted link 104 is that, when an accident occurs to any vehicle specified in the original operation plan 131, the dotted link 104 is used as an arrow pointing to a journey node 102 that has not yet been completed among the vehicle node 101 and each journey node 102 assigned to the accident vehicle.

[0120] Similarly to the directed link 103, the dotted link 104 does not connect any vehicle nodes 101 to each other, and in addition, the length of the dotted link 104 has no technical meaning.

[0121] Among the above, the form of the "arrow pointing to the process node 102 that has not yet been executed" of the dotted line link 104 can be roughly classified into the following types of accident example 1 to accident example 3.

[0122] Accident example 1: In the case of an arrow pointing from the vehicle node 101 of the accident vehicle to all the journey nodes 102 that the vehicle has traveled (in the case of an accident occurring from the time of departure, a dotted link 104 is connected from the vehicle node 101 of the accident vehicle to all the journey nodes 102 up to the final journey node 102).

[0123] Accident example 2: When the vehicle becomes an accident vehicle after completing one of the journeys, a dotted line link 104 is connected between the journey nodes 102 that have been completed and can be traced back from the vehicle node to the next journey node 102 and ultimately to the final journey node 102.

[0124] Accident example 3: A dotted link 104 is connected from the completed journey node 102 to the final journey node 102 because the vehicle in question became an accident vehicle while performing the final journey or after performing the journey before the final journey.

[0125] Generally speaking, the amount of change required to the contents of the original operation plan 131 (see FIG. 4) (in other words, the amount of data of the change information 163) increases in the order of accident example 3, accident example 2, and accident example 1. In order to avoid complicating the explanation below, however, the generation of a directed graph assuming the case of accident example 3 will be described.

[0126] FIG. 8 is a flowchart showing a procedure for generating the directed graph shown in FIG. 7 according to the original operation plan 131, accident information 132, and constraint conditions 133 described above, and for generating a graph as shown in FIG. 9 (hereinafter, referred to as a "candidate change graph") based on the directed graph.

[0127] In step S201, the processor of the information processing system 10 (the plan candidate creation unit 141 of the calculation unit 40) secures a data storage area corresponding to all nodes (the vehicle nodes 101 and the journey nodes 102) defined in the original operation plan 131. This area can be secured by, for example, securing a data storage area for the plan candidate 161, the plan execution information 162, and the change information 163 in the storage unit 60.

[0128] In the next step S202, the plan candidate creation unit 141 (the above processor, the same below) stores (inputs) corresponding information for all nodes (same as above). Specifically, the plan candidate creation unit 141 stores necessary information such as the starting time, departure point, and vehicle information of the original plan read from the input data in a storage area corresponding to each vehicle node 101. In addition, the plan candidate creation unit 141 stores necessary information such as the departure and arrival points, departure and arrival times, and operation conditions of the original plan read from the input data (see step S1 in FIG. 2) in a storage area corresponding to each journey node 102, 105. Furthermore, the plan candidate creation unit 141 performs a process of connecting corresponding links (directed links 103 or dotted links 104) to all nodes (same as above) based on the accident information 132 and constraint conditions 133 read from the input data.

[0129] By performing the processes of steps S201 and S202 described above, it is possible to generate data corresponding to the directed graph described above in Fig. 7. In other words, steps S201 and S202 correspond to a subroutine of step S20 shown in Fig. 6.

[0130] Referring again to FIG. 7, it can be seen that in the directed graph generated at this point, each journey node 102 has information on departure and arrival points and departure and arrival times in accordance with the input data (constraint condition 133 described above in FIG. 5).

[0131] For example, the journey node 102 corresponding to journey 1 is a journey that departs from bus stop A (node ​​start end) at 7:10 and arrives at bus stop B (node ​​end) at 9:10, and is assigned to vehicle 1. The journey node 102 corresponding to journey 2 is a journey that vehicle 1 departs from bus stop B (node ​​start end) at 11:30 and arrives at bus stop A (node ​​end) at 13:30. The same is true for journeys 3 and 4 that vehicle 1 subsequently performs, and directed links 103 are connected so as to connect these multiple nodes 101 (vehicle 1), 102 (journey 1) to 102 (journey 4) one-to-one in chronological order.

[0132] 7 illustrates a case where vehicle 3 becomes an accident vehicle at 15:00 after the completion of journey 7, according to accident information 132 in the input data. Therefore, the journey node 102 of journey 8 that requires a change in vehicle allocation is indicated by reference symbol 102-Ch, and the journey node 102 of journey 7 and the journey node 102 (102-Ch) of journey 8 are connected by a dotted link 104.

[0133] In other words, route node 102-Ch (only route 8 in Figure 7) is a route that was originally planned to be taken by vehicle 3, but due to an accident involving vehicle 3, vehicle 3 was no longer able to take on the route, and corresponds to the "difficult route to carry out" of the present invention.

[0134] Also, the dotted link 104 can be said to be a link that would have been connected if the original plan had been followed, but is now in danger of being unable to be carried out due to the occurrence of an accident. In other words, the original plan was to connect the directed link 103 from the route 7 to the route 8, but in reality it was not connected, and instead the dotted link 104 (indicating that the route cannot be carried out by the vehicle 3) is connected. Thus, in the example shown in FIG. 7, the route node 102-Ch of the route 8 (a difficult route to carry out) that cannot be carried out by the vehicle 3 requires the connection of the directed link 103 from a node (vehicle node 101 or route node 102) assigned to any of the other vehicles (vehicles 1, 2, 4 to 8 in this example), and the other vehicle must take over the route.

[0135] Next, a description will be given of the processing from step S203 to step S212 in the flowchart of Fig. 8. This processing corresponds to processing up to generating a change candidate graph as shown in Fig. 9 based on the directed graph shown in Fig. 7. This processing can be realized by first executing only the loop passing through S205 in the branch at S204 in Fig. 8, so that the graph in Fig. 7 is generated first, and then executing the loop passing through S206. Also, if both S205 and S206 are processed simultaneously, the graph of the state in FIG. 9 can be created without creating the state in FIG.

[0136] In one specific example, the plan candidate creating unit 141 performs the following process of trying to connect the directed links 103 (steps S203 to S207) to the vehicle nodes 101 of all vehicles other than the accident vehicle (vehicles 1, 2, 4 to 8 in this example).

[0137] First, in step S203, the plan candidate creating unit 141 sets the vehicle node 101 of one vehicle (for example, vehicle 1) as an execution target (hereinafter, the one vehicle node 101 is referred to as a "node of interest").

[0138] In the next step S204, the plan candidate creating unit 141 determines whether or not the departure time of the node of interest is before the accident occurrence time (15:00 in this example). If it is determined that the departure time is before the accident occurrence time (step S204, YES), the plan candidate creating unit 141 determines that the link connection is not possible, and proceeds to step S205. On the other hand, if the result of the above determination is that the departure time is not before the accident occurrence time (step S204, NO), the plan candidate creating unit 141 determines that the link can be connected, and proceeds to step S206.

[0139] In step S205, the plan candidate creation unit 141 maintains the state in which the directed link 103 is connected in the direction of the first journey handled (for example, if the node of interest is vehicle 1, then journey node 102 of journey 1) as per the original plan, and proceeds to step S207.

[0140] On the other hand, in step S206, the plan candidate creation unit 141 sets up a directed link 103-Ch (see Figure 9 as appropriate) to connect from the vehicle node 101 of the target node to the route node 102-Ch (of route 8 whose execution is in doubt) that is the destination of the dotted line link 104 described above, and proceeds to step S207.

[0141] In step S207, the plan candidate creating unit 141 determines whether or not an attempt has been made to connect the directed link 103-Ch to the journey node 102-Ch for the vehicle nodes 101 of all vehicles other than the accident vehicle (vehicles 1, 2, 4 to 8 in this example). If the result of such determination is that connection attempts have not yet been made to the vehicle nodes 101 of all vehicles other than the accident vehicle (step S207, NO), the plan candidate creation unit 141 returns to step S203 and repeats the processing of steps S204 to S207 described above, focusing on the unattempted vehicle nodes.

[0142] On the other hand, if the plan candidate creating unit 141 determines that connection attempts have been made to the vehicle nodes 101 of all vehicles other than the accident vehicle (step S207, YES), the program candidate creating unit 141 proceeds to step S208.

[0143] In this example, by performing the above processing, a directed link 103-Ch is connected from the vehicle node 101 of the vehicle 8 waiting in case of an emergency to the route node 102-Ch of the route 8, as shown in the lower part of Figure 9.

[0144] However, since vehicle 8 is at point B when the accident of vehicle 3 occurs (see FIG. 9), it is necessary for vehicle 8 to move to point C once in order to complete journey 8.

[0145] In step S208, the plan candidate creating unit 141 performs the process of trying to connect the following directed links 103 (steps S208 to S212) to all the route nodes (route nodes 102 other than the route node 102-Ch indicated by the dotted line link 104, and so on below).

[0146] In step S208, the plan candidate creating unit 141 sets a course node 102 of one course (for example, course 1) as an execution target (similar to the above, the one course node 102 is referred to as a "node of interest").

[0147] In the next step S209, the plan candidate creating unit 141 determines whether or not the departure time of the node of interest is before the accident occurrence time (15:00 in this example). If it is determined that the departure time is before the accident occurrence time (step S209, YES), the plan candidate creating unit 141 determines that the link connection is not possible, and proceeds to step S210. On the other hand, if the result of the above determination is that the departure time is not before the accident occurrence time (step S209, NO), the plan candidate creating unit 141 determines that the link can be connected, and proceeds to step S211.

[0148] In step S210, the plan candidate creating unit 141 maintains a state in which the directed link 103 is connected to the journey as originally planned (for example, if the node of interest is journey 1, then journey node 102 of journey 2), and proceeds to step S212.

[0149] Meanwhile, in step S211, the plan candidate creation unit 141 sets a directed link 103-Ch (see FIG. 9 as appropriate) to be connected from the plan node 102 of the target node to the plan node 102-Ch (of the plan 8 whose execution is in doubt) which is the connection destination of the above-mentioned dotted line link 104, and proceeds to step S212. Here, the "allowable delay time" in FIG. 8 will be explained. For example, in the example of the original plan shown in FIG. 7 and FIG. 9, plan 6 has a departure time of 14:30 and an arrival time of 15:30, so if a link is connected from plan 6 to plan 8 (departure time 16:20, arrival time 17:20) and this is used as a replanned plan, plan 8 can be executed without changing the time. However, if the arrival time of trip 6 is scheduled to be later than the departure time of trip 8, for example departure time 15:30, arrival time 16:30, when rescheduling from trip 6 to trip 8, the departure time of trip 8 must be set to 16:30 or later, so if the interval time is set to 5 minutes, the departure time of trip 8 will be 16:35, which is a rescheduled trip that is 15 minutes later than the original plan. The extent to which this delay is tolerated is called the allowable delay time here. The allowable delay time differs depending on the operator, so it should be set as necessary.

[0150] In step S212, the plan candidate creating unit 141 judges whether or not an attempt has been made to connect the directed links 103-Ch to all the journey nodes 102. If the result of such determination is that connection attempts have not yet been made for all of the journey nodes 102 (step S212, NO), the plan candidate creation unit 141 returns to step S208 and repeats the processing of steps S209 to S212 described above, focusing on the unattempted journey nodes 102.

[0151] On the other hand, if the plan candidate creating unit 141 determines that the connection has been attempted for all the journey nodes 102 (step S212, YES), it ends the routine of FIG.

[0152] In this example, by performing the processes in steps S208 to S212, a directed link 103-Ch is connected from the route node 102 of route 6 to the route node 102-Ch of route 8, as shown on the right side of FIG.

[0153] By performing this series of processes, the information processing system 10 generates a graph as shown in FIG. 9, that is, a change candidate graph in which directed links 103-Ch (hereinafter also referred to as "change candidate links") serving as change candidates are connected.

[0154] <2-2. An example of a procedure for enumerating intermediate data using graph search> Next, an example of a procedure for generating change information 163 from the intermediate data described above after generating a change candidate graph as described above in Fig. 9 will be described. The change information 163 can be generated in the process of performing a graph search according to a certain rule along the links on the change candidate graph shown in Fig. 9.

[0155] An example of this graph search procedure is shown in the flowchart of Fig. 10. Before describing the details of the flowchart of Fig. 10, an overview of generating intermediate data from a graph search will first be described with reference to Figs.

[0156] Here, Fig. 11 shows a change candidate graph similar to the graph shown in Fig. 9, but the directed link 103 and the change candidate link 103-Ch related to the process are highlighted with thick lines. Furthermore, in Fig. 11, an additional journey (journey node indicated as "Add 1" in the figure) for moving vehicle 8 from point B to point C is inserted on directed link 103-Ch connected from vehicle node 101 of vehicle 8 to journey node 102-Ch of journey 8.

[0157] Meanwhile, Fig. 12 is a table showing an example in which the original plan and candidate plans for change are expressed as binary values ​​of 0 / 1. For reference and ease of understanding, the third column from the left in Fig. 12 (i.e., the columns for plan candidates j=1, 2, 3, especially the column for j=3) shows the binary values ​​when each process (process 1 to process 8 in this example) is performed according to the original plan.

[0158] The rows constituting node number i in the table shown in FIG. 12 (hereinafter sometimes referred to as the "plan candidate table") correspond to each node (vehicle node 101 and journey node 102) of the change candidate graph shown in FIG. 11, and the columns constituting plan candidate j in the table correspond to routes (routes that have already been traveled or will be traveled) that start from vehicle node 101 as the starting point of the graph and are connected according to the rule described above (the end point and the next starting point must be the same).

[0159] The values ​​(binary numbers) stored in each column of the table in Figure 12 are determined and stored according to the rule that if the route shown in each column (plan candidate j) passes through the node shown in each row, the value is 1, and if it does not pass through the node, the value is 0.

[0160] However, for reference, step 8 (i=16) in the column for plan candidate j=3 is shown in parentheses and underlined with a 1 because it cannot actually be executed. The sequence of plan candidates j can also be generated for each route (node ​​number i).

[0161] 12 and 11, the plan contents included in the route following the change candidate link 103-Ch for vehicle 2 on the change candidate graph shown in FIG. 11 can be expressed as the string shown for plan candidate j=4 in FIG. 12. On the other hand, the plan contents included in the route that follows the change candidate link 103-Ch for the vehicle 8 on the change candidate graph shown in FIG. 11 can be expressed as the string shown in plan candidate j=5 in FIG.

[0162] Generally speaking, the values ​​of 0 / 1 (binary numbers) in each column in the table in FIG. 12 are elements constituting the plan candidate 161, which is one of the intermediate data described above.

[0163] For example, as shown by the thick line in the middle of the graph in Fig. 11, there is a route that starts at vehicle node 101 of vehicle 2 and follows a route (thick line) in the order of journey 5, step 6, and step 8. Such a route becomes the value in column j=4 of the modification candidate table in Fig. 12. The values ​​(binary values) in each field of this j=4 column correspond to values ​​indicating the first modification candidate.

[0164] Specifically, since the route starts from vehicle 2 and passes through routes 5, 6, and 8, in the j=4 column (first change candidate), the rows for vehicle 2 and routes 5, 6, and 8 are 1, and the rest are 0. In addition, the route in the j=4 column (first change candidate) corresponds to the candidate operation plan for vehicle 2 after replanning.

[0165] On the other hand, as shown by the bold line at the bottom of the graph in Fig. 11, there is a route that starts at vehicle node 101 of vehicle 8 and follows a route (bold line) in the order of additional step 1 and step 8. This route becomes the value in the column j=5 in the modification candidate table in Fig. 12. The values ​​(binary values) in each field of this j=5 column correspond to values ​​indicating the second modification candidate.

[0166] Specifically, since the route starts from vehicle 8 and passes through additional journey 1 and process 8, in the column j=5 (second change candidate), the rows for vehicle 8, additional journey 1, and process 8 are 1, and the rest are 0. Also, the route in the column j=5 (second change candidate) corresponds to the candidate operation plan for vehicle 8 after replanning.

[0167] Then, the change candidate table as shown in FIG. 12 can be generated in parallel by the processor (plan candidate generating unit 141) of the information processing system 10 when the change candidate graph in FIG. 11 is generated.

[0168] Furthermore, the processor (plan candidate creation unit 141) of the information processing system 10 calculates the cost for each of the above-mentioned paths (node ​​number i) and plan candidates (j), and calculates the additional cost (C j ) and store the calculated values ​​in a table.

[0169] In one specific example, the plan candidate creating unit 141 calculates an evaluation value according to the loss incurred by the user, such as the amount of loss when a delay occurs due to a change in the plan (for example, 0 yen per minute of delay time), as the additional cost, and stores the calculated value in the table. In addition, the plan candidate creating unit 141 may calculate an evaluation value according to extra fuel costs, labor costs, etc. incurred due to the change in the route, as the additional cost, and store the calculated value in the table.

[0170] For example, the routes in the columns j=1 and j=2 are as per the original plan, so no additional costs are incurred. On the other hand, the route in the column j=3 will incur repair costs if vehicle 3 breaks down, but since it is considered that no additional costs associated with the change in plan have yet been incurred, a value of 0 or null is stored.

[0171] In contrast, the route in the j=4 column (first change candidate) is compared to the route in the j=2 column, in that the journey 8 that was originally to be performed by the vehicle 3 is performed by the vehicle 2, and therefore the cost of the fuel consumption of the vehicle 2 increases. However, since the journey 8 can be performed without any delay, it is considered that there is no loss suffered by the user. For this reason, the additional cost (C j ) and 0.1 is calculated and stored in the table.

[0172] On the other hand, the route in the j=5 column (second change candidate) requires vehicle 8 to actually carry out two journeys (additional journey 1 and journey 8) in order to substitute for journey 8, which means that the fuel consumption cost and personnel cost will be higher than when journey 8 is substituted by vehicle 2. However, even in this case, journey 8 can be carried out without any delay, so it is considered that there will be no loss suffered by the user. For this reason, the additional cost (C j ) and 0.5 is calculated and stored in the table.

[0173] Thus, the processor (plan candidate creation unit 141) of the information processing system 10 calculates the additional cost (C j ) and the actual number of steps to perform step 8 (corresponding to the "number of changes" of the present invention), it is determined which of the change candidates should be adopted as the change proposal.

[0174] In this example, the additional cost (C j ) and the actual number of steps to perform step 8 are both smaller for the path in the column j=4 (first change candidate). Therefore, the processor (plan candidate creation unit 141) of the information processing system 10 adopts the path in the column j=4 (first change candidate) as the change plan. The digraph after adoption is shown in FIG. 13.

[0175] In other words, this process selects vehicle 2, which corresponds to the first change candidate, from multiple substitute vehicle candidates (in this example, vehicles 2 and 8), as a substitute vehicle for route 8 (difficult route to complete) (function of the substitute vehicle selection unit).

[0176] At this time, the processor (plan candidate creating unit 141) of the information processing system 10 can also generate the plan implementation information 162 and the change information 163 described in Fig. 1. As described above, in Fig. 9 and Fig. 11, etc., pictures (graphs) are used for visualization purposes, but in the practice of the present invention, the search for links does not necessarily need to be performed by creating a graph.

[0177] The change information 163 is data for distinguishing between itineraries that are changed from the original plan and itineraries that are not changed among the plan candidates 161. As an example, it can be data that stores the values ​​of the table shown in FIG. 15. Note that FIG. 13 includes change information when a substitute vehicle is finally selected in the accident occurrence case described in FIG. 7. Here, the plan candidates j=1, j=2, and j=3 shown in FIG. 12 are the original plan, the plan candidates j=4 and j=5 are changed plan candidates, and the plan candidate j=4 is a plan proposal in which a substitute vehicle is selected. Note that, for simplicity, no changes that would cause delays have occurred in the example shown here, but as described above, it is also possible to create a plan in which the departure time and arrival time are changed within the range of the delay allowable time.

[0178] An example of step S22 outlined in Fig. 6, that is, the procedure for searching the directed graph and listing the plan candidates, plan implementation information, and change information, is shown in the flowchart of Fig. 10. Unless otherwise specified, proceed to the procedure with the following number.

[0179] In step S221, a processor (for example, the plan candidate creating unit 141, the same below) of the information processing system 10 tries the following steps S222 to S228 for all vehicle nodes 101 (in this example, the vehicle nodes 101 of vehicles 1 to 8).

[0180] In step S222, the processor of the information processing system 10 performs a depth-first search of all routes, starting from the focused vehicle node 101. For example, based on the change table shown in Fig. 13, in the case of vehicle 1, the "depth 4" is determined from step 1, journey 2, journey 3, and journey 4. Similarly, in the case of vehicle 2, the "depth 3" is determined, vehicle 3 is determined to be "depth 1" (due to an accident), and vehicle 8 is determined to be "depth 0" (because it has been decided not to operate).

[0181] In step 223, the processor of the information processing system 10 reserves a necessary memory area for the above intermediate data for one candidate for the operation plan of the vehicle of interest (for example, one column of the table in FIG. 10 or the table in FIG. 11).

[0182] In step S224, the processor of the information processing system 10 records information on the passed nodes (already completed journeys). For example, if a constraint is imposed on the departure time of the next journey, such as an interval being provided because the end of the journey is a major station, the departure and arrival times are adjusted and stored as the replanned times in the plan execution information 162. By performing such processing, it becomes possible to smoothly respond to the occurrence of an accident vehicle (difficult journey).

[0183] In step S225, the processor of the information processing system 10 determines whether the rule defined by the constraint condition has been violated, and if it is determined that the rule has been violated (step S225, YES), the processor ends the search for the route of the vehicle of interest. Here, cases in which it is determined that the "rule has been violated" include cases in which the upper limit number of trips is exceeded, and cases in which the arrival time after replanning exceeds the final time.

[0184] In step S226, the processor of the information processing system 10 calculates and stores the cost of the route of interest according to a calculation method defined by the user.

[0185] In step S227, the processor of the information processing system 10 determines whether the search attempt for all routes has been completed, and if it is determined that the search attempt has been completed (step S22, YES), the process proceeds to step S228. On the other hand, if the processor of the information processing system 10 determines that the search attempt for all routes has not yet been completed (step S227, NO), the process returns to step S223 to select an untried route, and the process from step S223 described above is repeated.

[0186] In step S228, the processor of the information processing system 10 determines whether the depth-first search trial for all vehicle nodes has been completed, and if it is determined that the depth-first search trial for all vehicle nodes has been completed (step S228, YES), the process proceeds to step S229. On the other hand, if the processor of the information processing system 10 determines that the depth-first search trial for all vehicle nodes has not yet been completed (step S228, NO), the process returns to step S221 to focus on untried vehicle nodes.

[0187] In step S229, the processor of the information processing system 10 calculates and stores the change information 163 in accordance with a predetermined rule, and ends the series of processes.

[0188] <3. Model Creation> The above-mentioned example has been explained assuming a relatively simple case for ease of understanding, and in actual operation, it may be necessary to modify the original operation plan in a more complicated form. For this reason, the information processing system 10 of this embodiment is provided with an optimization calculation unit 80, so that the above-mentioned processes related to the candidate determination and selection of substitute vehicles can be performed using a method of calculating an optimal solution of the interaction model.

[0189] Here, the interaction model solved by the optimization calculation unit 80 is created using the plan candidates 161 and change information 163 described above. An example of an interaction model that can be considered based on a set partitioning problem (a mathematical model of the equality constraints of Formula 2) will be described below. The model shown here is merely one form of an embodiment, and is not limited to the form of a set partitioning problem. Depending on the method of creating the plan candidates, it can be considered based on a set covering problem (the inequality constraints of Formula 2), or it can be established with a completely different formulation. Also, depending on the problem, the direction of the inequality in Formula 2 (greater than or less than) can change.

[0190] In a general set partitioning problem (where the constraints in Equation 2 are equality and b=1), the binary variable x determines which combination of columns j in the candidate plan 161, such as the data in the table in FIG. 10, should be selected. jSpecifically, when selecting a candidate plan for column j, x j = 1, x if not selected j = 0. x j The actual replanning plan is the one that extracts only the routes for which A = 1. ij In the example of FIG. 14, this is a matrix of binary values ​​of 0 / 1 surrounded by a thick frame 401.

[0191]

number

[0192] In the above formula 2, C j is the cost, and is the value of the part surrounded by a thick frame 402 in Fig. 14. In addition, in Formula 2, M is the number of rows, and in the example of Fig. 14, M = 11. In Formula 2, N is the number of columns (i.e., the number of plan candidates), and in the example of Fig. 14, M = 8.

[0193] The interaction model input to the optimization calculation device 80 can be simply written as the following Equation 3 for the set partitioning problem.

[0194]

number

[0195] However, it is believed that the above description of Equation 3 still does not fully reflect the characteristics of re-planning.

[0196] Therefore, in this embodiment, an interaction model that utilizes the change information 163 is used. By utilizing such a model, the processor of the information processing system 10 (the proxy vehicle selection unit of the present invention) can select, as a proxy vehicle, a vehicle that optimizes the solution of the interaction model using the number of changes and the additional cost. As an example of such an interaction model (calculation formula), the interaction model shown in the following formula 4 can be considered.

[0197]

number

[0198] In formula 4, A ij Similarly to the above-mentioned formula 2, x in formula 4 is a matrix of binary values ​​of 0 / 1 surrounded by a thick frame 401 in FIG. j Similarly to Equation 2, x is a variable that takes two values, 0 / 1, and when selecting the column of the table in Figure 12, that is, when selecting the plan candidate j, j = 1, or 0 if no selection is made.

[0199] Also, C j is the same as Equation 2, and is the cost when an arbitrary plan candidate j is selected. Therefore, the first term of Equation 4 is x j = 1, the cost C j Since only the This is the sum of the costs of the above. It is the same as the objective function shown in Equation 2.

[0200] λ in Equation 4 i , γ i , η are arbitrary parameters. By appropriately adjusting these parameters, the calculations in the optimization calculation unit 80 can reach an optimal solution as quickly as possible. These parameters can be adjusted manually, but may also be automatically adjusted in the model calculation unit 142 using an existing method (e.g., a machine learning method or a black-box optimization method). In addition, when the optimization calculation unit can handle continuous variables, γ i may be a variable.

[0201] Also, γ i = 1, this corresponds to the usual set partitioning problem, Equation 3.

[0202] In other words, if all vehicles and trips (vehicles 1 to 8 and trips 1 to 8 in the above example) are included in one of the selected plan candidates, and there is no overlap of vehicles or trips between the different selected plan candidates, then Σ j A ij x j =1.

[0203] For example, in Figure 14, for j=1, j=6, and j=8, xj=1, and for other j, x j If =0, then Σ j A ij x j =1. In this case, the value in the parentheses of the second term in Formula 4 is 0.

[0204] However, in all other cases, the second term will have a finite value, resulting in a larger value of H than when it is zero.

[0205] In the optimization calculation unit 80, the value of H is minimized by x j In order to calculate the optimal solution, the second term is a finite value x j However, the value of γ i = 1. This is because γ i ∈(0.5,1], then Σ j A ij x j = 1 can be the minimum. Also, depending on the problem, if there are trips or vehicles that can be included or excluded in the replanning, γ i ∈(0,1].

[0206] The third term in Equation 4 is a characteristic term in this embodiment, and the value in the parentheses increases as the deviation from the original plan increases. j This means that the combination of

[0207] The third item D ijThe value of the change information 163 is used for this. Specifically, it corresponds to the part framed in bold in FIG. 15. In FIG. 15, when the allocation is according to the original plan, 0 is used, so j=1, 2, and 3 are all 0. In addition, the square that substitutes the process 8, which is a difficult process to execute, is set to -1 to reduce the overall cost, and when a process that was not originally assigned to the vehicle is assigned, such as addition 1, it is set to 1 to increase the overall cost. In this way, since Aij and Dij are sparse matrices, in problems where the number of vehicles or the number of processes is several tens to several thousands in an actual business, it is possible to apply speed-up that takes advantage of the characteristics of sparse matrices as appropriate. In addition, K in Equation 4 is the number of vehicles, and the reason for specifying the range of sums for i of the three items is to take the sum only for the rows related to the processes.

[0208] In this way, D ij By setting constraints on, when looking at the plan as a whole, if many plan candidates are selected that assign routes that were not in the original plan to vehicles (for example, additional route 1 in the above example), the overall cost (the value of H) will increase, so rescheduling proposals that involve many changes to the plan are less likely to be selected.

[0209] The form of the formula 4 and the range of the sum may be changed depending on the problem so that the calculation in the optimization calculation unit 80 converges quickly.

[0210] As described above in detail, according to the information processing system 10 equipped with the re-planning proposal output device of the present invention, when a difficult-to-perform process occurs that is difficult for the vehicle in charge to perform, a substitute vehicle for the difficult-to-perform process can be selected from among multiple substitute vehicle candidates, taking into consideration the number of times the vehicle in charge is changed and the additional cost. Therefore, a re-planning proposal that reduces the number of changes from the original plan and the profit and loss caused by the changes can be automatically created. In addition, according to the information processing system 10, when it becomes necessary to re-plan the operation plan in real time due to an error such as an accident, it can be used to re-plan the operation plan for the day. [Explanation of symbols]

[0211] 10 Information processing system (rescheduling output device) 20 Input / Output Support Department 30 Input data storage unit 40 Calculation section 50 Processing Support Department 60 Storage section 80 Optimization Calculation Section 101 Vehicle Node 102 Path Node 102-Ch Change Node (Difficult Execution Process) 103 Directed Links 103-Ch Change Link 104 Dotted Link 131 Original operation plan 132 Accident Information 133 Constraints 90 Output data storage unit 50 Processing Support Department 40 Calculation section 141 Planning Candidate Creation Department (Vehicle Candidate Determination Department, Vehicle Selection Department) 142 Model Generation Unit 143 Counting Processing Unit 60 Storage section 161 Plan Candidates 162 Planning Execution Information 163 Changes 192 Change costs (additional costs)

Claims

1. A rescheduled plan output device that outputs a rescheduled plan for an operation plan in which a scheduled route is carried out in sequence by an operation of a vehicle in charge, A replanning candidate determination unit that, when a difficult-to-perform process occurs among the scheduled processes and the vehicle in charge has difficulty in performing the difficult-to-perform process, determines another vehicle as a candidate for a substitute vehicle for the difficult-to-perform process; a substitute vehicle selection unit that selects a substitute vehicle for the difficult-to-perform journey from the substitute vehicle candidates when a plurality of substitute vehicle candidates are set, taking into consideration the number of changes in the vehicle in charge of each journey defined in the operation plan and additional costs; Equipped with Each of the journeys has a departure time from a starting point and an arrival time at a terminal point specified, The re-planning candidate determination unit determines, as a candidate for the substitute vehicle, the other vehicle that performs a journey that does not overlap with a time period from a departure time to an arrival time of the difficult journey to be performed; The substitute vehicle selection unit is a re-planning proposal output device that uses an interaction model having coefficients representing the number of changes and the additional cost and having as a variable whether or not to select the substitute vehicle candidate, and selects the substitute vehicle candidate that is selected based on the solution that minimizes the value of the interaction model as the substitute vehicle.

2. 2. The re-planning proposal output device according to claim 1, The substitute vehicle selection unit selects, from among a plurality of substitute vehicle candidates, a vehicle that has a small number of changes in the vehicle in charge of each of the journeys defined in the operation plan as the substitute vehicle. Rescheduling proposal output device.

3. 2. The re-planning proposal output device according to claim 1, The substitute vehicle selection unit selects, as the substitute vehicle, a vehicle that has a smaller additional cost from among a plurality of substitute vehicle candidates. Rescheduling proposal output device.

4. 2. The re-planning proposal output device according to claim 1, a graph generation unit that generates a directed graph in which nodes of vehicles defined in the operation plan and nodes of the journeys assigned to each of the vehicles are connected by directed links in an execution order; the re-planning candidate determination unit causes the graph generation unit to generate a graph as a candidate for the re-planning plan by connecting the directed link from a node related to the determined candidate for the substitute vehicle to a node of the difficult process to be executed in the directed graph; Rescheduling proposal output device.

5. 5. The re-planning proposal output device according to claim 4, the substitute vehicle selection unit causes the graph generation unit to generate a graph as the re-planned plan by connecting the directed link from the node related to the selected substitute vehicle to the node of the difficult process in the directed graph; Rescheduling proposal output device.

6. 2. The re-planning proposal output device according to claim 1, An input data storage unit that stores, as input data, information on the operation plan, a time when the difficult process occurred, and accident information including a vehicle in charge of the difficult process, Rescheduling proposal output device.

Images