WO2023119965A1 - Supply–demand matching device and method - Google Patents

Supply–demand matching device and method Download PDF

Info

Publication number
WO2023119965A1
WO2023119965A1 PCT/JP2022/042536 JP2022042536W WO2023119965A1 WO 2023119965 A1 WO2023119965 A1 WO 2023119965A1 JP 2022042536 W JP2022042536 W JP 2022042536W WO 2023119965 A1 WO2023119965 A1 WO 2023119965A1
Authority
WO
WIPO (PCT)
Prior art keywords
passenger
demand
route
transportation
information
Prior art date
Application number
PCT/JP2022/042536
Other languages
French (fr)
Japanese (ja)
Inventor
陽平 長谷川
三揮 米原
Original Assignee
株式会社日立製作所
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 株式会社日立製作所 filed Critical 株式会社日立製作所
Publication of WO2023119965A1 publication Critical patent/WO2023119965A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G06Q50/14—Travel agencies

Definitions

  • the present invention generally relates to supply and demand matching.
  • Patent Literature 1 discloses a technique for controlling the demand and supply of resources while satisfying both resource usage conditions and provision conditions.
  • the quality of transportation services For transportation operators, it is important to achieve both the quality of transportation services and operational efficiency (for example, profitability). For example, if the means of transportation is a bus, if the number of buses is insufficient, the quality of the service will deteriorate (for example, the waiting time before boarding will increase, or the inside of the bus will be crowded), and the number of buses will be excessive. Then, the operational efficiency is lowered.
  • passenger travel demand varies. For example, some passengers just need to be seated, while others just need to reach their destination.
  • the purpose of the present invention is to realize demand-supply matching that satisfies a wide variety of travel demands of passengers and maintains the operational efficiency of transportation services.
  • the demand-supply matching device collects travel demand information, which is information representing travel demand for each passenger, provisional route information, which is information representing a provisional route (provisional route from the departure point to the destination) for each passenger, and a transportation network. It refers to traffic network information, which is information to represent.
  • a travel demand includes a point of origin, a destination, and a requested service level, which is the level requested for each of one or more items relating to the transportation of a passenger.
  • a transportation network is a graph-structured network composed of a plurality of nodes and a plurality of links. A node corresponds to a location where a passenger may pick up and/or drop off. Links correspond to transport services.
  • the transportation network information includes, for each transportation service, information regarding the capacity of the transportation service and the service level of the transportation service.
  • the demand-supply matching device receives travel demand from the target passenger terminal (passenger terminal of the target passenger) of the target passenger (one of the passengers), and includes information representing the travel demand in the travel demand information.
  • the demand-supply matching device identifies one or more problematic links, each link failing to meet at least one of operational efficiency and required service level, based on the travel demand information and the transportation network information.
  • the supply and demand matching device extracts a partial traffic network which is a graph as a part of the traffic network including the one or more problem links.
  • the demand-supply matching device Based on the temporary route information and the movement demand information, the demand-supply matching device generates clusters of passenger movement demand corresponding to the temporary route through the partial transportation network.
  • the demand-supply matching device performs demand-supply optimization to determine a solution that satisfies travel demand and maintains operational efficiency, based on each generated cluster and traffic network information.
  • the determined solution is the cluster assignment for each leg in the partial traffic network. In a partial traffic network, a leg is one or more links that differ between nodes. If the determined solution includes a section not included in the provisional route for the target passenger, the supply and demand matching device selects either the provisional route for the target passenger or a new route that includes the section not included in the provisional route. As a response to the movement demand from the target passenger terminal, information on the target passenger's confirmed route is sent to the target passenger terminal.
  • 1 shows a configuration example of an entire system according to an embodiment of the present invention
  • 1 shows an example hardware configuration of a demand-supply matching server.
  • 4 shows an example of the functional configuration of a demand-supply matching server;
  • the processing flow of the passenger application section is shown.
  • An example of a transport service request UI (User Interface) is shown.
  • 4 shows an example of an acceptance inquiry UI.
  • 4 shows an example of a ticket UI.
  • the flow of supply and demand matching batch processing is shown.
  • 4 shows the flow of processing for extracting a partial traffic network.
  • 4 shows the flow of partial traffic network generation processing.
  • An example of generation of a temporary partial traffic network is shown schematically.
  • 4 shows the flow of partial traffic network reconfiguration processing.
  • 4 shows a configuration example of an operational constraint table; 4 shows the flow of demand cluster generation processing. 4 shows a configuration example of a movement demand table; A configuration example of a demand clustering table is shown. The flow of supply and demand optimization processing is shown. 4 shows the flow of intra-cluster passenger count processing. 4 shows a configuration example of a passenger acceptance history DB; 4 shows a configuration example of an expected value management table; 1 shows an example of a partial traffic network; 4 shows a configuration example of a section management table according to the present embodiment; 4 shows a configuration example of a section management table according to a comparative example; 4 shows a configuration example of an affiliation probability management table; An example of demand-supply matching is shown schematically.
  • an "interface device” may be one or more interface devices.
  • the one or more interface devices may be at least one of the following: - An I/O interface device that is one or more I/O (Input/Output) interface devices.
  • An I/O (Input/Output) interface device is an interface device for at least one of an I/O device and a remote display computer.
  • the I/O interface device to the display computer may be a communications interface device.
  • the at least one I/O device may be any of a user interface device, eg, an input device such as a keyboard and pointing device, and an output device such as a display device.
  • - A communication interface device that is one or more communication interface devices.
  • the one or more communication interface devices may be one or more of the same type of communication interface device (e.g., one or more NICs (Network Interface Cards)) or two or more different types of communication interface devices (e.g., NIC and It may be an HBA (Host Bus Adapter).
  • NIC Network Interface Cards
  • HBA Hypervisor Adapter
  • memory refers to one or more memory devices, typically a main memory device. At least one memory device in the memory may be a volatile memory device or a non-volatile memory device.
  • a "persistent storage device” is one or more persistent storage devices.
  • a permanent storage device is typically a non-volatile storage device (for example, an auxiliary storage device), specifically, for example, a HDD (Hard Disk Drive) or SSD (Solid State Drive).
  • the “storage device” may be at least the memory of the memory and the permanent storage device.
  • a "processor” is one or more processor devices.
  • the at least one processor device is typically a microprocessor device such as a CPU (Central Processing Unit), but may be another type of processor device such as a GPU (Graphics Processing Unit).
  • At least one processor device may be single-core or multi-core.
  • At least one processor device may be a processor core.
  • At least one processor device may be a broadly defined processor device such as a hardware circuit (for example, FPGA (Field-Programmable Gate Array) or ASIC (Application Specific Integrated Circuit)) that performs part or all of processing.
  • FPGA Field-Programmable Gate Array
  • ASIC Application Specific Integrated Circuit
  • the function may be described using the expression “yyy part”, but the function may be realized by executing one or more computer programs by a processor, or may be realized by executing one or more computer programs. It may be realized by the above hardware circuits (for example, FPGA or ASIC), or may be realized by a combination thereof.
  • a function is realized by executing a program by a processor, the defined processing is performed using a storage device and/or an interface device as appropriate, so the function may be at least part of the processor. good.
  • a process described with a function as the subject may be a process performed by a processor or a device having the processor. Programs may be installed from program sources.
  • the program source may be, for example, a program distribution computer or a computer-readable recording medium (for example, a non-temporary recording medium).
  • the description of each function is an example, and multiple functions may be combined into one function, or one function may be divided into multiple functions.
  • each DB and each table
  • one DB and one table
  • may be divided into two or more DBs two or more tables.
  • all or part of two or more DBs may be one DB (one table).
  • FIG. 1 shows a configuration example of the entire system according to this embodiment.
  • a supply and demand matching server 10 communicates with traffic operation management servers 20 of each of a plurality of transportation companies and passenger terminals 30 of each of a plurality of passengers via a communication network 40 (for example, an interface).
  • Both servers 10 and 20 may be physical computer systems (one or more physical computers) or based on physical computer systems (for example, a cloud platform having multiple types of physical computing resources). It may be a logical computer system (for example, cloud computing service).
  • the traffic operation management server 20 arranges transportation services and manages means of transportation.
  • a “transport service” is a service that transports passengers in vehicles.
  • Transport means is typically a vehicle, for example, a vehicle in a regular operation type, demand traffic type, share type or DH (dynamic headway) type transportation system (e.g., railway, bus, taxi, passenger car or bicycle).
  • DH dynamic headway
  • transportation of passengers is a broad definition that includes transportation by the passengers themselves driving a vehicle (for example, a passenger car or a bicycle in a shared transportation system).
  • the passenger terminal 30 is a passenger information processing terminal (typically, a mobile information processing terminal such as a smartphone).
  • the passenger terminal 30 is used, for example, for reservations and ticketing of transportation services.
  • the demand-supply matching server 10 communicates with a plurality of traffic operation management servers 20 and a plurality of passenger terminals 30 to perform demand-supply matching, which is matching of travel demand and transportation services.
  • FIG. 2 shows a hardware configuration example of the supply and demand matching server 10.
  • the supply and demand matching server 10 has an I/O interface device 107, a communication interface device 105, storage devices (permanent storage device 104 and memory 103) for storing programs and data, and a processor 101 connected to them.
  • the I/O interface device 107 is an interface between an input device 108 such as a keyboard and a mouse, and a display device 109 such as a liquid crystal display or an organic EL display.
  • FIG. 3 shows an example of the functional configuration of the supply and demand matching server 10.
  • the supply and demand matching server 10 comprises a travel demand DB 128, a temporary route DB 129, a fixed route DB 130, a passenger acceptance history DB 131, and a transportation network DB 132. These DBs are stored, for example, in persistent storage 104 .
  • the movement demand DB 128 is a database that stores information representing the movement demand of each passenger.
  • the "movement demand” consists of elements related to the customer's movement (eg, destination, departure point, departure time, etc.) and the service level requested by the customer. That is, in the movement demand DB 128, for each passenger, the ID of the movement demand of the passenger and the information representing the movement demand are divided into the movement information of the passenger (for example, information representing the destination, departure place and departure time) and the relevant Passenger service level information (information representing the service level requested by the customer for transportation services).
  • a "service level” is a level for each of a plurality of service items relating to transportation services.
  • the ID of travel demand may be synonymous with the ID of a passenger. That is, when the same travel demand is registered by passengers A and B, the ID of the travel demand registered by passenger A and the ID of the travel demand registered by passenger B may be different.
  • the temporary route DB 129 is a database that stores information representing a temporary route from the passenger's destination to the departure point for each passenger.
  • the temporary route DB 129 may represent a temporary route for each ID of travel demand of passengers.
  • the fixed route DB 130 is a database that stores information representing the fixed route from the passenger's destination to the departure point for each passenger.
  • the fixed route DB 130 may represent a fixed route for each ID of travel demand of a passenger.
  • the passenger acceptance history DB 131 is a database that stores the history of acceptance or rejection of new route candidates.
  • the history may indicate, for example, whether or not the new route candidate is accepted for each of one or more new route candidates for each passenger travel demand ID.
  • the transportation network DB 132 is a database that stores information representing transportation networks.
  • a "traffic network” is a graph-structured network composed of a plurality of nodes and a plurality of links connecting the nodes.
  • a node is a pre-determined vehicle stopping location where passengers can board and disembark.
  • a link is a physical or logical path (eg, a railroad, road, or airway) connecting those stops.
  • a transport network consists of a plurality of sub-transport networks.
  • One sub-transportation network corresponds to one transportation service out of a plurality of transportation services (a plurality of transportation means).
  • a traffic network is one network in which a plurality of sub-transport networks are logically placed on the same layer.
  • a node corresponding to a place where any two or more modes of transportation such as railways, buses, and taxis can be used is common to the two or more modes of transportation in a plurality of sub-transportation networks. . That is, if a node is common to a plurality of transport services, only nodes are overlapped (common) between different sub-transport networks, and links are not overlapped. For example, in a transportation network, there may be multiple links between node A (station A) and node B (station B), such as a rail link and a bus link. On the other hand, in the sub-transport network, since the sub-transport network corresponds to one transportation service, there is one link connecting nodes.
  • Information representing a sub-transport network may be included in the transport network DB 132 for each transport service.
  • the transportation network DB 132 stores, for each node, information representing the details of the location corresponding to the node (for example, name, location information, whether or not return is possible for each transportation service, timetable for each transportation service, congestion status, delay status, etc.), and for each link, information representing details of the road corresponding to the link (for example, corresponding transport service (transport means), congestion status, delay status, etc.) may be included.
  • each of the two nodes sandwiching one link for one transportation service may be called an "adjacent node" for convenience. That is, if there are multiple links between adjacent nodes, there are multiple different transportation services between adjacent nodes (transport from one location corresponding to one adjacent node to another location corresponding to another adjacent node). There are multiple transportation services as a service).
  • a set of one or more continuous links from a starting point (node) to a destination (node) can be called a "route".
  • a passenger application unit 121 When the program is executed by the processor 101, a passenger application unit 121, a route search unit 122, a partial network extraction unit 123, a demand clustering unit 124, a passenger counting unit 125, a supply and demand optimization unit 126, an operation A function such as the ordering unit 127 is realized. At least one of these functions appropriately references or updates at least one of the DBs 128-132.
  • a passenger application unit 121 communicates with the passenger terminal 30 .
  • the operation arrangement unit 127 communicates with the traffic operation management server 20 . Details of these functions 121-127 will be described later.
  • FIG. 4 shows the flow of passenger application processing.
  • the passenger application unit 121 receives input of travel information and service level information from the passenger terminal 30, and registers the information in the travel demand DB 128 (S1). For example, the passenger application unit 121 displays a transportation service request UI 140 (screen for inputting a request for transportation service) illustrated in FIG. 5A on the passenger terminal 30 . Travel information (eg, origin, destination, departure time and/or arrival time) and service level information (level for each of a plurality of service items) are entered via the transportation service request UI 140 . Entering service level information may be optional.
  • the passenger application unit 121 issues a route search instruction to the route search unit 122 based on the above-described movement information (and service level information) input from the passenger terminal 30, and in response to the instruction, the route search unit 122 Information representing the searched provisional route is registered in the provisional route DB 129 (S2).
  • the search instruction includes at least a part of the input travel information (and service level information) (for example, departure place, destination, departure time and/or arrival time, and level for each of a plurality of service items). ) is associated as a search condition.
  • the route search unit 122 searches the traffic network DB 132 for a route that satisfies the search conditions associated with the search instruction, and returns information representing the searched route.
  • the passenger application unit 121 determines whether or not the deadline for route confirmation has passed (S3).
  • “Route determination” means to set one route candidate out of one or more route candidates as a definite route.
  • “One or more route candidates” is at least the provisional route selected from the provisional route registered in S2 and one or more new route candidates accepted in S5.
  • the "deadline for route determination” may be a deadline determined by a predetermined policy (for example, it may be a time when a certain period of time has passed since the provisional route was registered in S2, or it may be when a new route candidate is finally accepted. (It may be after a certain period of time has passed since, or it may be the time limit specified by the passenger).
  • the passenger application unit 121 determines whether or not a route change notification has been received (S4). If the determination result of S4 is false (S4: No), the process returns to S3.
  • the "route change notification” is a notification that a new route candidate different from the provisional route (and the notified new route candidate) has been found. This notification is issued in the demand-supply matching batch process shown in FIG. 6 (specifically, the notification issued in S67 of FIG. 15 is the route change notification).
  • the passenger application unit 121 receives from the passenger terminal 30 whether or not the new route candidate indicated in the route change notification can be accepted (S5). For example, the passenger application unit 121 displays on the passenger terminal 30 an acceptance inquiry UI 150 (a screen for inputting whether or not to accept the new route candidate) illustrated in FIG. 5B. According to the example shown in FIG. 5B, there are two new route candidates, each provided with a checkbox 52 . Passengers can select new routes to accept (check boxes). The passenger application unit 121 registers the content of the selection (information indicating whether or not each new route candidate can be accepted) in the passenger acceptance history DB 131 (S6). After that, the process returns to S3.
  • the passenger application unit 121 determines the confirmed route, deletes the information on the temporary route registered in S2 from the temporary route DB 129, and confirms the information representing the confirmed route.
  • Register in the route DB 130 (S7).
  • a "fixed route" is one route candidate out of one or a plurality of route candidates, as described above. ⁇ When information representing one or more acceptable new route candidates is registered in the passenger acceptance history DB 131, one new route candidate obtained from the one or more acceptable new route candidates is a confirmed route. becomes. This new route candidate may be the last accepted new route candidate, a randomly selected new route candidate, or a route candidate selected from passengers. - If there is no new acceptable route candidate, the tentative route becomes the final route.
  • a ticket UI 160 which is a screen displaying the issued digital ticket, is displayed on the passenger terminal 30 as illustrated in FIG. 5C.
  • the ticket UI 160 shows a fixed route "A ⁇ C: railroad, C ⁇ E: demand bus" (route from point A to point C by rail, and from point C to point E by demand bus). Information is displayed.
  • the ticket UI 160 also displays a digital ticket object 163 of the fixed route. When object 163 is specified, details of the contents of the ticket are displayed.
  • Fig. 6 shows the flow of supply and demand matching batch processing. This process is performed periodically.
  • a partial traffic network extraction process is performed (S11). Next, demand cluster generation processing is performed (S12). Finally, supply and demand optimization processing is performed (S13). The processing of these S11 to S13 will be described in detail below.
  • a set of one or more continuous links can be called a "section".
  • a “segment” can be part of a route.
  • the “one or more continuous links” that make up the section there is only one link between the same adjacent nodes, and if the section consists of two or more links, the link and the link are connected through the node It is connected.
  • a "section” may typically be one link or two consecutive links (two links connected to the same node).
  • one or more consecutive sections in a partial traffic network can be called a "path".
  • a “path” can be part of a route.
  • FIG. 7 shows the flow of the partial traffic network extraction process (S11 in FIG. 6).
  • the partial network extraction unit 123 calculates the number of passengers satisfying the required service level for each link of the transport service (sub-transport network) (S21). Specifically, for example, the following is performed. - The partial network extraction unit 123 refers to the movement demand DB 128 and the fixed route 130, and identifies a movement demand for which the route is not fixed. - The partial network extraction unit 123 identifies possible routes (provisional routes and new routes (new route candidates)) from the provisional route DB 129 and the passenger acceptance history DB 131 for each travel demand whose route is not fixed.
  • the partial network extraction unit 123 extracts, for each possible route that has been specified, from the travel demand corresponding to the route and the traffic network DB 132 (for example, the configuration of the traffic network and information for each link), for each link , the number of passengers for which the required service level is met is calculated for each link of the transport service (sub-transport network).
  • the partial network extraction unit 123 calculates the operation efficiency of the transportation service for each link of the transportation service (S22). For example, operational efficiency may be the ratio of expected revenue to cost. Specifically, for example, the partial network extraction unit 123 identifies the number of passengers and the number of vehicle operations in a time period for each link of the transportation service based on the transportation network DB 132 and the fixed route 130, Operational efficiency may be calculated based on the number of vehicles in operation. Note that the number of passengers in a time slot for each transport service link may be determined by predicting the number of passengers in a time slot based on past operation management information available from each traffic operation management server.
  • the partial network extraction unit 123 extracts a link with an absolute or relatively small number of passengers (the number of passengers satisfying the service level) calculated in S21, and/or an absolute Or list relatively low links (S23).
  • the link with the small number of passengers calculated in S21 is the link with the number of passengers below the threshold ⁇ passenger .
  • a link with a low operational efficiency calculated in S22 is a link whose operational efficiency is below the threshold ⁇ transportation .
  • ⁇ transportation may be thresholded to the number of transportation services for which the expected revenue is less than the cost.
  • the partial network extraction unit 123 performs partial traffic network generation processing based on the link information listed in S23 (S24).
  • each link listed in S23 may be collectively referred to as "problem link”.
  • a "problem link” is a link with a low number of passengers meeting its service level or a link with low operational efficiency.
  • FIG. 8 shows the flow of the partial traffic network generation process (S24 in FIG. 7).
  • the partial network extraction unit 123 constructs a temporary partial traffic network (S31).
  • a temporary partial traffic network is a network in which links in the vicinity of each listed problem link are connected.
  • a "neighboring link” may be a node connected to either of the nodes on either end of the problem link.
  • a link near the problem link may be another problem link or a non-problem link.
  • S31 for example, as shown in the upper part of FIG. 9, a partial traffic network as shown in the lower part of FIG. 9 is constructed by connecting neighboring links to the problem link represented by the thick line.
  • Nodes A and E are end points of the partial traffic network illustrated in the lower part of FIG. Dashed arrows denote passengers entering or exiting a partial transportation network from outside the partial transportation network.
  • the partial network extraction unit 123 expands the temporary partial traffic network so that the nodes at the endpoints of the temporary partial traffic network satisfy the operational constraints (S32).
  • “End point node satisfies operational constraints” means that the terminal node satisfies the operational constraints represented by the operational constraint table 170 illustrated in FIG. 11 for all transport services related to the node.
  • the operational constraint table 170 is included in the traffic network DB 132, for example.
  • the operation constraint table 170 indicates whether or not a return can be made at a node for each pair of a node and transportation service. According to the example shown in FIG. 11, it is taken into consideration that railroads and fixed-route buses have limited places where they can turn back, and there are restrictions on the setting of sections for temporary services. In addition, transportation services such as demand transportation and taxis are managed so that all locations can be turned back.
  • the number or ratio of transport services that cannot be returned at the endpoint node is less than or equal to ⁇
  • ⁇ may be an operational constraint. That is, “ ⁇ ” may be the number of links (the number of transport services) connected to the endpoint node, or the ratio of the number of transport services that cannot be returned to the number of links connected to the endpoint node.
  • the value of " ⁇ " is normally 0. That is, in this embodiment, in principle, for each transportation service, the end point node of the partial transportation network is a node that can be turned back.
  • the partial network extraction unit 123 combines these partial traffic networks (S33). . As a result, these partial traffic networks are made into one partial traffic network. If there are no such multiple partial traffic networks, S33 is skipped.
  • the partial network extraction unit 123 performs partial traffic network reconfiguration processing for each partial traffic network (S34). As a result, if there is a reconfigured partial traffic network (S35: Yes), the process returns to S33.
  • FIG. 10 shows the flow of the partial traffic network reconfiguration process (S34 in FIG. 8).
  • one partial traffic network is taken as an example (“target network” in the explanation of FIG. 9).
  • the partial network extraction unit 123 calculates the number of alternative paths for the route that passes through the most problematic links in the target network (S41).
  • the number of alternative paths is "3", that is, the alternative paths are the pair of links T4 and T7, A pair of links T4 and T8 and a link T6.
  • the partial network extraction unit 123 determines whether or not the number of alternative paths calculated in S41 is equal to or less than ⁇ max (S42).
  • ⁇ max is the upper bound on the number of alternative paths in the partial traffic network.
  • the partial network extraction unit 123 determines whether or not the number of alternative paths calculated in S41 is equal to or greater than ⁇ min (S43). ⁇ min is the lower bound on the number of alternative paths in the partial traffic network. Setting this value prevents the failure to set alternate paths for problem optimization. If the determination result of S43 is true (S43: Yes), the process ends. If the determination result of S43 is false (S43: No), the partial network extraction unit 123 expands the target network from the nodes at both ends of the target network to the nearest loopable node (S44). For the expanded network, it is expected that the number of alternative paths will be greater than or equal to ⁇ min .
  • the partial network extraction unit 123 determines whether or not there is a node that satisfies the operational constraints in the target network (S45). If the determination result of S45 is false (S45: No), the process ends. It should be noted that the presence or absence of "nodes satisfying operational constraints" is identified from the operational constraint table 170 shown in FIG.
  • the partial network extraction unit 123 determines whether or not the number of alternative paths is ⁇ min or more even if the target network is divided at the nodes that satisfy the operational constraints ( S46). If the determination result of S46 is false (S46: No), the process ends. If the determination result of S46 is true (S46: Yes), the partial network extraction unit 123 divides the target network along the nodes that satisfy the operational constraints (S47).
  • the above is the description of the partial traffic network extraction process (S11 in FIG. 6).
  • Information representing the partial traffic network extracted in this process is stored in the traffic network DB 132 .
  • the partial traffic network extracted here is a network constructed based on the problem links listed in S23 of FIG. 7, as described above.
  • information on each partial traffic network is stored in the traffic network DB 132 .
  • a partial traffic network is a network as a range that can be affected by one or more problem links, and a graph as a part of the traffic network. Without such a partial traffic network, many routes would have to be searched each time any link in the traffic network becomes a problem link. As a result, the number of routes becomes enormous, and there is a risk that a solution cannot be obtained in subsequent demand/supply optimization (or it will take a very long time to obtain a solution). By extracting the partial traffic network as the part related to the problem link, the optimization problem can be made smaller and the solution can be obtained (easily solved). In other words, in this embodiment, areas other than partial traffic networks are excluded from the optimization problem, thus avoiding processing areas without problem links, thus saving computational time and resources. be able to.
  • FIG. 12 shows the flow of the demand cluster generation process (S12 in FIG. 6).
  • the demand clustering unit 124 refers to the movement demand DB 128, the transportation network DB 132, and the temporary route DB 129, and identifies the movement demand (passengers) corresponding to the temporary route passing through the partial transportation network (S51). For one partial traffic network, there can be one or more temporary routes as temporary routes passing through the partial traffic network.
  • the demand clustering unit 124 identifies the time for each node included in the tentative route in the partial transportation network (S52).
  • the traffic network DB 132 includes time management information, which is information representing the arrival time and departure time at each node, the average speed of vehicles, etc., for each transportation service. Based on this time management information, time specification in S52 may be performed.
  • the demand clustering unit 124 identifies the service level for the movement demand (passengers) identified in S51 from the movement demand DB 128 (S53).
  • the demand clustering unit 124 clusters the travel demand identified in S51 from the service level identified in S53 and the time for each node identified in S52 (S54).
  • the demand clustering unit 124 calculates a representative value for each cluster of travel demand (S55).
  • a cluster is a set of identical or similar travel demands.
  • each passenger has a travel demand, so a cluster can also be said to be a set of passengers having the same or similar travel demands. Since clusters are generated only for passengers whose routes (typically tentative routes) pass through the extracted partial traffic network, the optimization problem can be reduced.
  • the origin and destination may not necessarily be used.
  • at least one of the origin and destination may be used for clustering of travel demands that have an origin or destination (or both) within a partial transportation network.
  • the origin and/or destination outside may not always be used for clustering, or the origin and destination outside may not be used for clustering.
  • One cluster is, for example, the origin and/or destination within the partial transportation network, the distance of the departure and/or destination outside the partial transportation network, and the required service level (e.g., degree of congestion, delay It is a group of passengers with the same or similar movement demand factors such as degree of movement, movement preference such as seating certainty.
  • the demand cluster generation processing shown in FIG. 12 will be described in more detail below.
  • FIG. 13 shows a configuration example of the movement demand table 180.
  • the travel demand table 180 is referred to in the travel demand clustering in S54.
  • a movement demand table 180 is stored in the movement demand DB 128 .
  • the movement demand table 180 has a record for each movement demand. Each record has information such as ID 181 , origin 182 , departure time 183 , destination 184 , arrival time 185 and requested service level 186 .
  • the requested service level 186 is information representing the level of each of a plurality of service items with respect to the requested service level, such as allowable congestion 187, allowable delay 188, and guaranteed seating 189.
  • FIG. Take one movement demand as an example (“target movement demand” in the description of FIG. 13).
  • the ID 181 represents the ID of the target travel demand.
  • the departure point 182 represents the departure point (for example, the ID of the node) in the target travel demand.
  • the departure time 183 represents the departure time (time at which the passenger departs from the departure point) in the target travel demand.
  • the destination 184 represents the destination (eg ID of the node) in the target travel demand.
  • the arrival time 185 represents the arrival time (time at which the passenger arrives at the destination) in the target travel demand.
  • Service items include, for example, allowable congestion, allowable delays, and seat guarantees.
  • the permissible congestion 187 represents a permissible level of congestion on a railroad or the like for passengers (for example, "150" when a congestion of up to 150% is permissible).
  • the permissible delay 188 represents a level (eg, "10" when a delay of up to 10 minutes is permissible) that a passenger can tolerate as a delay time of transportation services such as railroads.
  • the seating guarantee 189 is a level requesting whether or not a passenger can accept no seat for transportation services such as railways (for example, if a passenger can accept no seating for a transportation service, "None"). represents
  • the demand clustering unit 124 clusters a set of travel demands in which the service level identified in S53 and the time of each node identified in S52 are similar. Specifically, for example, the demand clustering unit 124 calculates the feature amount of each of the plurality of types of information 182 to 189 for each travel demand, and clusters a set of travel demands having similar feature amounts of various types of information. .
  • a demand clustering table representing each cluster is constructed.
  • FIG. 14 shows a configuration example of the demand clustering table 190.
  • the "statistical value" is, for example, the average value, but may be the maximum value, minimum value, median value, or the like.
  • the demand clustering table 190 has a record for each cluster.
  • Each record has information such as a cluster number 191, a cluster affiliation number 192, a departure point 193, a departure time 194, a destination 195, an arrival time 196, and a requested service level 197.
  • Required service level 197 includes acceptable congestion 198, acceptable delay 199, seating guarantee 200, and the like. Take one cluster as an example (“target cluster” in the description of FIG. 14).
  • the cluster number 191 represents the identification number of the target cluster.
  • the cluster belonging number 192 represents the number of travel demands belonging to the target cluster.
  • the departure point 193 represents the departure point (statistical value of the departure point 182 in FIG. 13, for example) in the travel demand belonging to the target cluster.
  • Departure time 194 represents the departure time (for example, the statistical value of the departure point in FIG. 13) in travel demand belonging to the target cluster.
  • the destination 195 represents the destination (statistical value of the destination 184 in FIG. 13, for example) in the travel demand belonging to the target cluster.
  • the arrival time 196 represents the arrival time (statistical value of the arrival time 185 in FIG. 13, for example) of the travel demand belonging to the target cluster.
  • Allowable congestion 198 of requested service level 197 represents the allowable congestion (for example, the statistics of allowable congestion 187 in FIG. 13) for travel demands belonging to the target cluster.
  • Acceptable delay 199 of requested service level 197 represents the acceptable delay (eg, statistics of acceptable delay 188 in FIG. 13) for travel demands belonging to the target cluster.
  • Seating guarantee 200 of requested service level 197 represents the seating guarantee (eg, statistics of seating guarantee 189 in FIG. 13) for travel demand belonging to the target cluster.
  • the demand clustering unit 124 calculates a representative value for each cluster based on multiple types of information 192 to 200 corresponding to the cluster.
  • the representative value calculated for each cluster is used as a feature of the movement demand to which the transport service is allocated. That is, in the subsequent supply and demand optimization, the mobile demands are considered as clusters, and thus the representative values are used to solve the optimization problem. For example, based on the representative value of the cluster, it is possible to determine that a bus or railroad is suitable as a transport service for the cluster.
  • FIG. 15 shows the flow of the supply and demand optimization process (S13 in FIG. 6).
  • the supply and demand optimization unit 126 refers to the traffic network DB 132 and the demand clustering table 190, searches for a section that satisfies the movement demand (information 193 to 200) belonging to the cluster for each cluster, and The candidate combinations of means are listed (S61). In S61, the demand/supply optimization unit 126 also searches for sections including temporary traffic that has not yet been arranged as candidates. Information on temporary traffic (for example, information representing types of transportation services as temporary traffic and nodes to be routed through) is stored in, for example, the traffic network DB 132 .
  • the supply and demand optimization unit 126 causes the passenger count unit 125 to perform intra-cluster passenger count processing (S62).
  • the passenger count unit 125 calculates an expected value, which is the ratio of passengers passing through the section in the cluster, for each set of cluster and section, assuming that different passengers are accepted for each section.
  • the supply and demand optimization unit 126 solves a multi-objective optimization problem that satisfies the movement demand (required service level) and operational efficiency while recombining the sections to determine which section each cluster uses, and the calculated one One of the solution candidates (solutions) is selected (S63).
  • One solution candidate is, for example, as shown in FIG.
  • the demand/supply optimization unit 126 determines whether or not additional transportation service arrangements are necessary for the selected candidate solution based on, for example, the transportation network DB 132 (S64). For example, if the section route of the solution candidate includes a temporary flight, it is determined that additional arrangements are necessary (that is, the determination result of S64 is true). In other words, in addition to regular transport services, there may be non-regular transport services in candidate section routes for each cluster. is deemed necessary.
  • the demand/supply optimization unit 126 selects a transportation service that requires additional
  • the operation arrangement unit 127 is caused to present the information to the traffic operation management server 20 (S65).
  • the operation arrangement unit 127 requests additional arrangements from the traffic operation management servers 20 of all transportation companies. Specifically, for example, if a section route of a certain cluster included in the selected solution candidate includes a special bus or a special train (S64: Yes), in S65, the special bus or special train is Additional arrangements will be requested from the transportation company in charge.
  • the demand/supply optimization unit 126 determines whether or not all transportation operators presented in S64 have accepted the presentation (S66). If the determination result of S66 is false (S66: No), the next solution candidate is selected (S67), and the process returns to S65.
  • the supply and demand optimization unit 126 regards the selected solution candidate as the determined solution, and sends a route change notification to the passenger application unit 121 for this solution (S68 ).
  • a route change notification representing a new route, which is a route including sections according to the determined solution, is output for the passenger corresponding to the tentative route. Outputting the route change notification may include adding to the passenger acceptance history DB 131 a record containing information about the new route represented by the route change notification.
  • the demand/supply optimization unit 126 registers the temporary transportation as arranged in the transportation network DB 132 (S69).
  • the supply and demand optimization unit 126 regards the selected solution candidate as the determined solution, and sends a route change notification to the passenger application unit 121 for this solution (S70 ).
  • FIG. 16 shows the flow of the intra-cluster passenger count process (S62).
  • the passenger counting unit 125 refers to the travel demand DB 128 and the demand clustering table 190, and lists clusters near the cluster to which the travel demand belongs for each travel demand (passenger) (S71).
  • a "cluster near the cluster to which the travel demand belongs” is, for example, a cluster having a representative value within a certain range from the representative value of the cluster to which the travel demand belongs.
  • the passenger count unit 125 calculates a feature quantity that is input to the passenger acceptance model for each movement demand (passenger) (S72).
  • the "feature amount” here is a feature amount specified based on the travel demand DB 128 and the demand clustering table 190, and is a set of passengers, clusters, and sections (sections found in S61).
  • the "passenger acceptance model” is a model that inputs feature values for each type 1 pair and outputs a belonging probability (see FIG. 21) for each type 1 pair.
  • the model is typically a machine learning model (eg neural network).
  • the passenger count unit 125 inputs the feature quantity for each type 1 group calculated in S72 into the passenger acceptance model, thereby calculating the belonging probability (see FIG. 21) for each type 1 group.
  • an expected value (probability) (see FIG. 18A) is calculated for each type 2 pair, which is a pair of a cluster and an interval, based on the membership probability for each type 1 pair (S73).
  • the passenger count unit 125 identifies the section with the highest expected value (probability) for each cluster, and calculates the number of passengers (predicted ) is calculated (S74).
  • FIG. 17 shows a configuration example of the passenger acceptance history DB 131.
  • the passenger acceptance history DB 131 has a record for each pair of passenger and new route. Each record has information such as record number 131a, distance 131b to cluster representative value, number of transfers 131c, distance to outside of partial traffic network 131d, congestion rate 131e, guaranteed seating 131f, and acceptability 131g. . Take one passenger and one new route new route as an example (“target passenger” and “target new route” in the description of FIG. 17).
  • the record number 131a represents the identification number of the record of the target passenger and target new route.
  • the distance 131b to the cluster representative value represents the distance between the characteristic value calculated based on the record of the target passenger and the representative value of the cluster to which the target passenger belongs.
  • the number of transfers 131c represents the number of transfers on the target new route.
  • the distance 131d to the outside of the partial transportation network represents the distance between the departure point or destination of the target passenger and the partial transportation network.
  • the congestion rate 131e represents the congestion rate on the target new route.
  • the seating guarantee 131f represents the seating guarantee on the target new route.
  • the acceptability 131g indicates whether or not the target new route has been accepted by the target passenger ("1" means that the target new route is acceptable).
  • the passenger counting unit 125 may perform learning such as building or updating the passenger acceptance model described above based on the data stored in the passenger acceptance history DB 131.
  • each record of the passenger acceptance history DB 131 includes information representing passengers, clusters, and sections, and the passenger acceptance model is characterized by each set of passenger, cluster, section, required service level, and acceptability. Quantities may be calculated, their feature quantities may be used as inputs, and belonging probabilities for each set of the first type, which is a set of passengers, clusters, and sections, may be output.
  • the passenger counting unit 125 can obtain the belonging probability for each type 1 pair by inputting the feature amount for each type 1 pair to the trained passenger acceptance model.
  • FIG. 18A shows a configuration example of the expected value management table 210.
  • FIG. 18B shows an example of a partial traffic network.
  • the expected value management table 210 has a record for each type 2 set (a set of a cluster and an interval). Each record has information such as a cluster number 211 , an interval 212 and an expected value (probability) 213 .
  • the cluster number 211 represents the identification number of the cluster.
  • a section 212 represents a sequence of one or more links forming the section.
  • the expected value (probability) 213 represents the ratio of passengers passing through the section for the cluster.
  • an expected value (probability) 213 is adopted.
  • the interval management table 220 illustrated in FIG. 19 can be constructed.
  • the section management table 220 has a record for each section, and each record has information such as section 221 , allocation cluster 222 , cost 223 , fare 224 , passenger capacity 225 , and number of passengers (estimate) 226 .
  • Information such as the number of passengers (measured) 227, revenue 228, and congestion rate 229 is information prepared for explaining an example of the expected effects of this embodiment, and does not need to be included in the section management table 220. None.
  • a section 221 represents a row of links forming the section.
  • Assigned cluster 222 represents the identification number of the cluster assigned to the interval.
  • Cost 223 represents the cost of traveling along the section.
  • the fare 224 represents the fare for the section.
  • the capacity 225 represents the capacity of the transportation service corresponding to the section.
  • the number of passengers (forecast) 226 represents the predicted number of passengers passing through the section.
  • the number of passengers (actual) 227 represents the expected number of passengers passing through the leg.
  • Revenue 228 represents the expected revenue for the leg and is typically the number of passengers (actual) 227 multiplied by the fare 224 minus the cost 223 .
  • the congestion rate 229 represents the expected congestion rate in the section and is typically a value obtained by dividing the number of passengers (measured) 227 by the capacity 225 .
  • Information such as the cost 223 , the fare 224 , and the capacity 225 may be included in the transportation network DB 132 for each section instead of the expected value management table 210 .
  • the value as the passenger capacity 225 may be a value as an actual passenger capacity, or may be a value after the actual passenger capacity is adjusted according to the number of passengers registered in the movement demand DB 128 .
  • the belonging probability 254 is calculated for each type 1 pair (a pair of passenger, cluster, and section). In other words, it is calculated what probability the passenger passes through which section when the passenger belongs to any cluster, assuming that the passenger can belong to any cluster.
  • An expected value (probability) 213 is adopted for each type 2 pair, and a cluster corresponding to the expected value (probability) 213 is assigned to each section.
  • the number of passengers (226) is calculated from the expected value (probability) 213 and the number of passengers belonging to the cluster assigned to the section.
  • the profit 228, which is one element of operational efficiency becomes a negative value (value meaning unprofitable), or the congestion rate 229, which is one element of the service level, exceeds "100%".
  • the relationship between clusters and passengers is fixed.
  • the revenue which is one element of operational efficiency
  • the congestion rate which is one element of service level
  • a method of calculating the expected value (probability) 213 in FIG. 18A is, for example, as follows.
  • a affiliation probability management table 250 illustrated in FIG. 21 is constructed.
  • the belonging probability management table 250 has a record for each type 1 pair (a pair of a passenger, a cluster, and a section), and each record contains a passenger number 251, a cluster number 252, a section 253, and a belonging probability 254. consists of
  • the passenger number 251 represents the identification number of the passenger.
  • Cluster number 252 represents the identification number of the cluster.
  • a section 253 represents a row of links forming the section.
  • the belonging probability 254 represents the probability of accepting (passing) the segment when the passenger belongs to the cluster. This probability is a value obtained based on a passenger acceptance model (a model constructed by learning the passenger acceptance history DB 131).
  • the expected value management table 210 illustrated in FIG. 18A is constructed.
  • the expected value (probability) 213 for each type 2 pair is the sum of the belonging probabilities 254 of all passengers corresponding to the type 2 pair.
  • the expected value (probability) 213 of the interval T1 of the cluster C1 is 0.81 + 0.75 + 0.93 + . . . according to the example shown in FIG. (Probability) 213 is "60".
  • Fig. 22 schematically shows an example of supply and demand matching.
  • I have a graph in a two-dimensional Cartesian coordinate system.
  • the vertical axis (example of the first axis) is the service level evaluation function w 0 f (R, S), and the horizontal axis (example of the second axis orthogonal to the first axis) is the operational efficiency evaluation function.
  • R is the section of each cluster (the section that can include a combination of transportation means to be used).
  • S is the requested service level of each cluster.
  • R * 1 is obtained as the optimum solution as a result of the search. If supply and demand matching is not established with the solution R * 1 , the second optimal solution R * 2 is selected as the next candidate. At this time, if there are multiple Pareto-optimal solution candidates, the service level may be prioritized.
  • a specific example of the processing performed by the supply and demand optimization unit 126 is as follows.
  • Equation (1) The service level and travel demand of each cluster c are represented by Equation (1).
  • Equation 2 the demand S of all clusters can be expressed as Equation 2.
  • rc be a list of section candidates of cluster c to be subjected to section search.
  • M c intervals are obtained for each cluster.
  • Each r c,j contains an interval for each cluster.
  • r c can be expressed as in Equation 3.
  • the supply and demand optimization unit 126 generates a path R selected by each cluster as a combination of sections to be taken by each cluster.
  • R can be expressed as in Equation 4.
  • the supply and demand optimization unit 126 solves the minimization problem according to Equation 5 by changing the section of each cluster in the path R.
  • f i is a cost function that increases if the service level is not met. For example, in addition to setting the number of sections with a congestion rate exceeding the service level standard based on the result of a simulation on the flow of people, the operator's operating cost is also set. Note that the minimization problem composed of w i f i is linear. Depending on the situation, w i fi may be added or deleted, and each w i fi may be improved (simulation accuracy improvement, weight update, etc.). In addition, about the top three candidates for R * may be presented. At this time, it is possible to extract traffic (temporary flights and demand traffic) requiring new arrangements for R * , and notify the target operators of the extracted traffic of the arrangements. Also, if the approval of all the target business operators cannot be obtained, it is possible to move to the processing of the next candidate for R * .
  • the operation arrangement unit 127 transmits and receives information to and from each traffic operation management server 20, and executes processing for arranging operation of each means of transportation used for transportation services provided by each transportation operator.
  • an example where demand for commuting to school and demand for events overlap can be considered.
  • a cluster of students who cannot be late a cluster of participants (students) who want to go to the event early, and a cluster of participants (students) who want to sit down until the event venue are generated.
  • demand buses for schools arrange for extended school bus services to event venues for schools, and arrange taxis to event venues.
  • nodes such as taxis and some demand buses are not fixed, existing nodes such as nearby bus stops belonging to the route can be used, and virtual stops can be set as nodes at fixed distance intervals.
  • a bicycle parking lot can be set as a node.
  • the passenger application unit 121 receives travel demand from the target passenger terminal 30 (passenger terminal of the target passenger) of the target passenger (one of the passengers), and stores information representing the travel demand in the travel demand DB 128 (an example of travel demand information). include.
  • the partial network extraction unit 122 identifies one or more problem links that are links that do not meet at least one of operational efficiency and required service level, based on the movement demand DB 128 and the transportation network DB 132 (an example of transportation network information). and extract a partial traffic network containing the one or more problematic links.
  • the demand clustering unit 124 clusters passenger movement demand corresponding to the temporary route through the partial transportation network. to generate The demand/supply optimization unit 126 performs demand/supply optimization to determine a solution that satisfies travel demand and maintains operational efficiency, based on the generated clusters and the transportation network DB 132 .
  • the determined solution is the cluster assignment for each leg in the partial traffic network. If the determined solution includes a section not included in the provisional route for the target passenger, the passenger application unit 121 selects either the provisional route for the target passenger or a new route that includes the section not included in the provisional route. is the definite route, and information on the definite route of the target passenger is transmitted to the target passenger terminal 30 as a response to the movement demand from the target passenger terminal 30 .
  • the extracted partial transportation network is a network in which the number or ratio of transportation services that allow return nodes at the end points is less than or equal to a predetermined value. Therefore, even if there is a change in the number of operations per unit time in the section (for example, the section containing the problematic link) in the partial traffic network, it is possible for the vehicle to turn back at the end point node where it is possible to turn back. As a result, the impact of the problem link outside the partial traffic network can be eliminated or reduced. This is one of the technical effects inherent in the allocation of transport services (means of transport).
  • the partial network extraction unit 122 connects the link adjacent to the problem link to the problem link within a range where the number or ratio of transport services that can return the end node node is a predetermined value or less. , may generate one or more partial traffic networks.
  • the partial network extraction unit 122 When the partial network extraction unit 122 generates two or more partial traffic networks that at least partially overlap each other, the two or more partial traffic networks may be regarded as one partial traffic network. As a result, it is possible to avoid the possibility that the solutions of the optimization problems of two or more partial traffic networks do not match even though they overlap each other at least partially.
  • the extracted partial traffic network may be a network in which the number of alternative paths for the relevant temporary route, which is the temporary route that passes through the problem links the most, is equal to or greater than the lower limit.
  • the alternative path may be one or more links that bypass the problem link and are not the problem link for each problem link in the relevant tentative route.
  • the partial network extraction unit 122 determines the number or ratio of transportation services that can return the node at the end point.
  • the partial traffic network may be expanded to the extent that it is equal to or less than the value. This prevents the inability to set alternative paths for problem optimization.
  • the extracted partial traffic network may be a network in which the number of alternative paths of the relevant provisional route, which is the provisional route passing through the problem links the most, is equal to or less than the upper limit value.
  • the partial network extraction unit 122 determines whether the following conditions are satisfied, and If the result is true, the generated partial traffic network may be divided along the corresponding intermediate node. This limits the size of the partial traffic network and thus prevents combinatorial explosion in supply and demand optimization.
  • the partial transportation network has an intermediate node, which is a node other than the end point node and the number or ratio of transport services that can be returned is less than or equal to a predetermined value.
  • each of the plurality of partial traffic networks satisfies that the number of alternative paths for the corresponding temporary route is equal to or greater than the lower limit value.
  • the passenger counting unit 125 calculates the probability that a passenger will pass through a section for each type 1 group (a group of a passenger, a cluster, and a section in the partial transportation network) when the passenger belongs to the cluster. Membership probabilities may be calculated. For each type 2 set, which is a set of a cluster and a section in the extracted partial transportation network, based on the belonging probability of all passengers obtained for the type 2 set, the section in the cluster An expected value, which is the percentage of passengers passing through The demand optimization unit 126 may determine the solution based on the expected value for each type 2 set.
  • the probability of belonging to each type 1 group is calculated, and the probability of belonging to each type 1 group is calculated. Based on this, the expected value for each type 2 pair is calculated. This is expected to improve the probability that the solution of the optimization problem is appropriate.
  • the passenger application unit 121 notifies the target passenger terminal 30 of the new route (route including the section that is assigned to the cluster to which the target passenger belongs and is not included in the tentative route of the target passenger and that follows the above-determined solution).
  • Information about the route may be included in the passenger acceptance history DB 131 (an example of passenger acceptance history information).
  • the passenger application unit 121 may receive a response as to whether or not the new route can be accepted from the target passenger terminal 30, and include the response in the passenger acceptance history DB 131 for the notified new route.
  • the calculation of the belonging probability for each type 1 group may be performed using a machine learning model learned based on the passenger acceptance history DB 131 .
  • the passenger acceptance history DB 131 includes information about new routes and information indicating whether or not the new routes have been accepted.
  • the passenger acceptance history DB 131 contains information indicating that the route has been rejected. . In other words, whether or not to accept the new route is left to the passenger's discretion, and the result of the passenger's discretion is reflected in the passenger acceptance history DB 131 . Based on the machine learning model learned based on such passenger acceptance history DB 131, the belonging probability for each type 1 group is calculated. This is expected to improve the probability that the solution of the optimization problem is appropriate.
  • the demand optimization unit 126 calculates one or more solutions based on the number of passengers predicted for each type 2 group and the traffic network information, and selects one unselected one of the one or more solutions One solution may be selected, and based on transportation network information, it may be determined whether the selected solution is one that requires additional arrangements for transportation services. If the result of this determination is false, the demand optimization unit 126 may set the selected solution as the determined solution. On the other hand, if the result of this determination is true, the operation arrangement unit 127 requests additional arrangements to one or more traffic operation management servers corresponding to each transportation service for which additional arrangements are required. When receiving approval responses from all the traffic management servers to which the request has been made, the demand optimization unit 126 may set the selected solution as the determined solution. This makes it possible to issue an appropriate notification of the new route (appropriate route change notification).

Landscapes

  • Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Marketing (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

In this invention, a travel demand includes an origin, a destination, and a requested service level, that is, the level requested for each of one or more items related to passenger transportation. A transportation network is a graph comprising multiple nodes and multiple links. One or more links corresponding to one or more transportation services exist between the same nodes. The device identifies a problematic link, that is, a link where at least one of the operational efficiency or the requested service level is not met, extracts a partial transportation network including the problematic link, generates clusters of passenger travel demand corresponding to virtual routes going through the partial transportation network, and determines, on the basis of each generated cluster, a solution (cluster assignment in each section of the partial transportation network) to meet the travel demand and maintain operational efficiency.

Description

需給マッチング装置及び方法Supply and demand matching device and method
　本発明は、概して、需給マッチングに関する。 The present invention generally relates to supply and demand matching.
　交通手段（例えば、バス、電車及びタクシー）を用いた輸送サービスの供給と、交通手段を利用する旅客の需要との適切なマッチングが行われることが望ましい。特許文献１には、リソースの利用条件と提供条件との双方を満たしつつ、リソースの需要と供給を制御するための技術が開示されている。 　It is desirable to appropriately match the supply of transportation services using means of transportation (for example, buses, trains, and taxis) with the demand of passengers who use the means of transportation. Patent Literature 1 discloses a technique for controlling the demand and supply of resources while satisfying both resource usage conditions and provision conditions.
ＷＯ２０１９／０９３２５５WO2019/093255
　交通事業者にとって、輸送サービスの品質と運用効率（例えば採算性）を両立することは重要である。例えば、交通手段がバスの場合、バスの本数が不十分であると、サービスの品質が低下し（例えば、乗車までの待ち時間が長くなり、或いは、車内が混雑し）、バスの本数が過剰であると、運用効率が低下する。 For transportation operators, it is important to achieve both the quality of transportation services and operational efficiency (for example, profitability). For example, if the means of transportation is a bus, if the number of buses is insufficient, the quality of the service will deteriorate (for example, the waiting time before boarding will increase, or the inside of the bus will be crowded), and the number of buses will be excessive. Then, the operational efficiency is lowered.
　一方、旅客の移動需要は様々である。例えば、着席できさえすれば良いという旅客もいれば、目的地に着きさえすれば良いという旅客もいる。 On the other hand, passenger travel demand varies. For example, some passengers just need to be seated, while others just need to reach their destination.
　本発明は、旅客の多種多様な移動需要を満たすことと輸送サービスの運用効率を維持することとを両立した需給マッチングを実現することを目的とする。 The purpose of the present invention is to realize demand-supply matching that satisfies a wide variety of travel demands of passengers and maintains the operational efficiency of transportation services.
　需給マッチング装置が構築される。需給マッチング装置は、旅客毎の移動需要を表す情報である移動需要情報と、旅客毎の仮経路（出発地から目的地までの仮の経路）を表す情報である仮経路情報と、交通ネットワークを表す情報である交通ネットワーク情報とを参照する。移動需要は、出発地と、目的地と、旅客の輸送に関する一つ又は複数の項目の各々について要求するレベルである要求サービスレベルとを含む。交通ネットワークは、複数のノードと複数のリンクとで構成されたグラフ構造のネットワークである。ノードは、旅客が乗車及び降車のうちの少なくとも一つを行い得る場所に対応する。リンクは、輸送サービスに対応する。同一のノード間に、一つ以上の輸送サービスに対応した一つ以上のリンクが存在する。交通ネットワーク情報は、輸送サービス毎に、輸送サービスの定員と輸送サービスのサービスレベルに関する情報とを含む。需給マッチング装置は、対象旅客（いずれかの旅客）の対象旅客端末（対象旅客の旅客端末）から移動需要を受け付け、当該移動需要を表す情報を移動需要情報に含める。需給マッチング装置は、移動需要情報及び交通ネットワーク情報を基に、それぞれが運用効率と要求サービスレベルの少なくとも一つが満たされないリンクである一つ以上の問題リンクを特定する。需給マッチング装置は、当該一つ以上の問題リンクを含み交通ネットワークの一部分としてのグラフである部分交通ネットワークを抽出する。需給マッチング装置は、仮経路情報及び移動需要情報を基に、部分交通ネットワークを通る仮経路に対応した旅客の移動需要のクラスタを生成する。需給マッチング装置は、生成された各クラスタと交通ネットワーク情報とを基に、移動需要を満たし運用効率を維持する解を決定する需給最適化を行う。決定された解は、部分交通ネットワークにおける区間毎のクラスタ割当てである。部分交通ネットワークにおいて、区間は、ノード間が異なる一つ以上のリンクである。決定された解が、対象旅客について仮経路に含まれない区間を含む場合、需給マッチング装置は、対象旅客の仮経路と、当該仮経路に含まれない区間を含んだ新経路とのいずれかを確定経路とし、対象旅客端末からの移動需要に対する応答として、対象旅客の確定経路に関する情報を対象旅客端末に送信する。 A supply and demand matching device will be built. The demand-supply matching device collects travel demand information, which is information representing travel demand for each passenger, provisional route information, which is information representing a provisional route (provisional route from the departure point to the destination) for each passenger, and a transportation network. It refers to traffic network information, which is information to represent. A travel demand includes a point of origin, a destination, and a requested service level, which is the level requested for each of one or more items relating to the transportation of a passenger. A transportation network is a graph-structured network composed of a plurality of nodes and a plurality of links. A node corresponds to a location where a passenger may pick up and/or drop off. Links correspond to transport services. One or more links corresponding to one or more transport services exist between the same nodes. The transportation network information includes, for each transportation service, information regarding the capacity of the transportation service and the service level of the transportation service. The demand-supply matching device receives travel demand from the target passenger terminal (passenger terminal of the target passenger) of the target passenger (one of the passengers), and includes information representing the travel demand in the travel demand information. The demand-supply matching device identifies one or more problematic links, each link failing to meet at least one of operational efficiency and required service level, based on the travel demand information and the transportation network information. The supply and demand matching device extracts a partial traffic network which is a graph as a part of the traffic network including the one or more problem links. Based on the temporary route information and the movement demand information, the demand-supply matching device generates clusters of passenger movement demand corresponding to the temporary route through the partial transportation network. The demand-supply matching device performs demand-supply optimization to determine a solution that satisfies travel demand and maintains operational efficiency, based on each generated cluster and traffic network information. The determined solution is the cluster assignment for each leg in the partial traffic network. In a partial traffic network, a leg is one or more links that differ between nodes. If the determined solution includes a section not included in the provisional route for the target passenger, the supply and demand matching device selects either the provisional route for the target passenger or a new route that includes the section not included in the provisional route. As a response to the movement demand from the target passenger terminal, information on the target passenger's confirmed route is sent to the target passenger terminal.
　本発明によれば、旅客の多種多様な移動需要を満たすことと輸送サービスの運用効率を維持することとを両立した需給マッチングを実現することができる。 According to the present invention, it is possible to realize demand-supply matching that satisfies a wide variety of travel demands of passengers and maintains the operational efficiency of transportation services.
本発明の一実施形態に係るシステム全体の構成例を示す。1 shows a configuration example of an entire system according to an embodiment of the present invention; 需給マッチングサーバのハードウェア構成例を示す。1 shows an example hardware configuration of a demand-supply matching server. 需給マッチングサーバの機能構成例を示す。4 shows an example of the functional configuration of a demand-supply matching server; 旅客アプリケーション部の処理の流れを示す。The processing flow of the passenger application section is shown. 輸送サービス要求ＵＩ（User Interface）の例を示す。An example of a transport service request UI (User Interface) is shown. 受入れ問合せＵＩの例を示す。4 shows an example of an acceptance inquiry UI. チケットＵＩの例を示す。4 shows an example of a ticket UI. 需給マッチングバッチ処理の流れを示す。The flow of supply and demand matching batch processing is shown. 部分交通ネットワークの抽出処理の流れを示す。4 shows the flow of processing for extracting a partial traffic network. 部分交通ネットワーク生成処理の流れを示す。4 shows the flow of partial traffic network generation processing. 仮の部分交通ネットワークの生成の一例を模式的に示す。An example of generation of a temporary partial traffic network is shown schematically. 部分交通ネットワーク再構成処理の流れを示す。4 shows the flow of partial traffic network reconfiguration processing. 運用制約テーブルの構成例を示す。4 shows a configuration example of an operational constraint table; 需要クラスタ生成処理の流れを示す。4 shows the flow of demand cluster generation processing. 移動需要テーブルの構成例を示す。4 shows a configuration example of a movement demand table; 需要クラスタリングテーブルの構成例を示す。A configuration example of a demand clustering table is shown. 需給最適化処理の流れを示す。The flow of supply and demand optimization processing is shown. クラスタ内旅客カウント処理の流れを示す。4 shows the flow of intra-cluster passenger count processing. 旅客受け入れ履歴ＤＢの構成例を示す。4 shows a configuration example of a passenger acceptance history DB; 期待値管理テーブルの構成例を示す。4 shows a configuration example of an expected value management table; 部分交通ネットワークの例を示す。1 shows an example of a partial traffic network; 本実施形態に係る区間管理テーブルの構成例を示す。4 shows a configuration example of a section management table according to the present embodiment; 一比較例に係る区間管理テーブルの構成例を示す。4 shows a configuration example of a section management table according to a comparative example; 所属確率管理テーブルの構成例を示す。4 shows a configuration example of an affiliation probability management table; 需給マッチングの一例を模式的に示す。An example of demand-supply matching is shown schematically.
　以下の説明では、「インターフェース装置」は、一つ以上のインターフェースデバイスでよい。当該一つ以上のインターフェースデバイスは、下記のうちの少なくとも一つでよい。
・一つ以上のＩ／Ｏ（Input/Output）インターフェースデバイスであるＩ／Ｏインターフェース装置。Ｉ／Ｏ（Input/Output）インターフェースデバイスは、Ｉ／Ｏデバイスと遠隔の表示用計算機とのうちの少なくとも一つに対するインターフェースデバイスである。表示用計算機に対するＩ／Ｏインターフェースデバイスは、通信インターフェースデバイスでよい。少なくとも一つのＩ／Ｏデバイスは、ユーザインターフェースデバイス、例えば、キーボード及びポインティングデバイスのような入力デバイスと、表示デバイスのような出力デバイスとのうちのいずれでもよい。
・一つ以上の通信インターフェースデバイスである通信インターフェース装置。一つ以上の通信インターフェースデバイスは、一つ以上の同種の通信インターフェースデバイス（例えば一つ以上のＮＩＣ（Network Interface Card））であってもよいし二つ以上の異種の通信インターフェースデバイス（例えばＮＩＣとＨＢＡ（Host Bus Adapter））であってもよい。 In the following description, an "interface device" may be one or more interface devices. The one or more interface devices may be at least one of the following:
- An I/O interface device that is one or more I/O (Input/Output) interface devices. An I/O (Input/Output) interface device is an interface device for at least one of an I/O device and a remote display computer. The I/O interface device to the display computer may be a communications interface device. The at least one I/O device may be any of a user interface device, eg, an input device such as a keyboard and pointing device, and an output device such as a display device.
- A communication interface device that is one or more communication interface devices. The one or more communication interface devices may be one or more of the same type of communication interface device (e.g., one or more NICs (Network Interface Cards)) or two or more different types of communication interface devices (e.g., NIC and It may be an HBA (Host Bus Adapter).
　また、以下の説明では、「メモリ」は、一つ以上のメモリデバイスであり、典型的には主記憶デバイスでよい。メモリにおける少なくとも一つのメモリデバイスは、揮発性メモリデバイスであってもよいし不揮発性メモリデバイスであってもよい。 Also, in the following description, "memory" refers to one or more memory devices, typically a main memory device. At least one memory device in the memory may be a volatile memory device or a non-volatile memory device.
　また、以下の説明では、「永続記憶装置」は、一つ以上の永続記憶デバイスである。永続記憶デバイスは、典型的には、不揮発性の記憶デバイス（例えば補助記憶デバイス）であり、具体的には、例えば、ＨＤＤ（Hard Disk Drive）又はＳＳＤ（Solid State Drive）である。 Also, in the following description, a "persistent storage device" is one or more persistent storage devices. A permanent storage device is typically a non-volatile storage device (for example, an auxiliary storage device), specifically, for example, a HDD (Hard Disk Drive) or SSD (Solid State Drive).
　また、以下の説明では、「記憶装置」は、メモリと永続記憶装置の少なくともメモリでよい。 Also, in the following description, the "storage device" may be at least the memory of the memory and the permanent storage device.
　また、以下の説明では、「プロセッサ」は、一つ以上のプロセッサデバイスである。少なくとも一つのプロセッサデバイスは、典型的には、ＣＰＵ（Central Processing Unit）のようなマイクロプロセッサデバイスであるが、ＧＰＵ（Graphics Processing Unit）のような他種のプロセッサデバイスでもよい。少なくとも一つのプロセッサデバイスは、シングルコアでもよいしマルチコアでもよい。少なくとも一つのプロセッサデバイスは、プロセッサコアでもよい。少なくとも一つのプロセッサデバイスは、処理の一部又は全部を行うハードウェア回路（例えばＦＰＧＡ（Field-Programmable Gate Array）又はＡＳＩＣ（Application Specific Integrated Circuit））といった広義のプロセッサデバイスでもよい。 Also, in the following description, a "processor" is one or more processor devices. The at least one processor device is typically a microprocessor device such as a CPU (Central Processing Unit), but may be another type of processor device such as a GPU (Graphics Processing Unit). At least one processor device may be single-core or multi-core. At least one processor device may be a processor core. At least one processor device may be a broadly defined processor device such as a hardware circuit (for example, FPGA (Field-Programmable Gate Array) or ASIC (Application Specific Integrated Circuit)) that performs part or all of processing.
　また、以下の説明では、「ｙｙｙ部」の表現にて機能を説明することがあるが、機能は、一つ以上のコンピュータプログラムがプロセッサによって実行されることで実現されてもよいし、一つ以上のハードウェア回路（例えばＦＰＧＡ又はＡＳＩＣ）によって実現されてもよいし、それらの組合せによって実現されてもよい。プログラムがプロセッサによって実行されることで機能が実現される場合、定められた処理が、適宜に記憶装置及び／又はインターフェース装置等を用いながら行われるため、機能はプロセッサの少なくとも一部とされてもよい。機能を主語として説明された処理は、プロセッサあるいはそのプロセッサを有する装置が行う処理としてもよい。プログラムは、プログラムソースからインストールされてもよい。プログラムソースは、例えば、プログラム配布計算機又は計算機が読み取り可能な記録媒体（例えば非一時的な記録媒体）であってもよい。各機能の説明は一例であり、複数の機能が一つの機能にまとめられたり、一つの機能が複数の機能に分割されたりしてもよい。 In addition, in the following description, the function may be described using the expression “yyy part”, but the function may be realized by executing one or more computer programs by a processor, or may be realized by executing one or more computer programs. It may be realized by the above hardware circuits (for example, FPGA or ASIC), or may be realized by a combination thereof. When a function is realized by executing a program by a processor, the defined processing is performed using a storage device and/or an interface device as appropriate, so the function may be at least part of the processor. good. A process described with a function as the subject may be a process performed by a processor or a device having the processor. Programs may be installed from program sources. The program source may be, for example, a program distribution computer or a computer-readable recording medium (for example, a non-temporary recording medium). The description of each function is an example, and multiple functions may be combined into one function, or one function may be divided into multiple functions.
　また、以下の説明では、「ｘｘｘＤＢ」や「ｘｘｘテーブル」といった表現にて、入力に対して出力が得られる情報を説明することがあるが、当該情報は、どのような構造のデータでもよいし（例えば、構造化データでもよいし非構造化データでもよいし）、入力に対する出力を発生するニューラルネットワーク、遺伝的アルゴリズムやランダムフォレストに代表されるような学習モデルでもよい。従って、「ｘｘｘＤＢ」や「ｘｘｘテーブル」を「ｘｘｘ情報」と言うことができる。また、以下の説明において、各ＤＢ（及び各テーブル）の構成は一例であり、一つのＤＢ（及び一つテーブル）は、二つ以上のＤＢ（二つ以上のテーブル）に分割されてもよいし、二つ以上のＤＢ（二つ以上のテーブル）の全部又は一部が一つのＤＢ（一つのテーブル）であってもよい。 Also, in the following description, expressions such as "xxxDB" and "xxx table" may be used to describe information that can be obtained as an output for an input, but the information may be data of any structure. (For example, structured data or unstructured data may be used.) Learning models such as neural networks, genetic algorithms, and random forests that generate outputs in response to inputs may also be used. Therefore, "xxx DB" and "xxx table" can be called "xxx information". Also, in the following description, the configuration of each DB (and each table) is an example, and one DB (and one table) may be divided into two or more DBs (two or more tables). However, all or part of two or more DBs (two or more tables) may be one DB (one table).
　以下、図面を用いて本発明の一実施形態を説明する。 An embodiment of the present invention will be described below with reference to the drawings.
　図１は、本実施形態に係るシステム全体の構成例を示す。 FIG. 1 shows a configuration example of the entire system according to this embodiment.
　一般に、地域には、複数種類の交通手段を用いた複数種類の輸送サービスがあり、それら複数種類の輸送サービスは、異なる交通事業者によって提供される。 In general, there are multiple types of transportation services using multiple types of transportation in an area, and these multiple types of transportation services are provided by different transportation operators.
　本実施形態では、需給マッチングサーバ１０が設けられる。需給マッチングサーバ１０は、複数の交通事業者の各々の交通運行管理サーバ２０と、複数の旅客の各々の旅客端末３０と、通信ネットワーク４０（例えばインターフェース）を介して通信する。サーバ１０及び２０のいずれも、物理的な計算機システム（一つ以上の物理的な計算機）でもよいし、物理的な計算機システム（例えば、複数種類の物理的な計算リソースを有するクラウド基盤）に基づく論理的な計算機システム（例えば、クラウドコンピューティングサービス）でもよい。 In this embodiment, a supply and demand matching server 10 is provided. The demand-supply matching server 10 communicates with traffic operation management servers 20 of each of a plurality of transportation companies and passenger terminals 30 of each of a plurality of passengers via a communication network 40 (for example, an interface). Both servers 10 and 20 may be physical computer systems (one or more physical computers) or based on physical computer systems (for example, a cloud platform having multiple types of physical computing resources). It may be a logical computer system (for example, cloud computing service).
　交通運行管理サーバ２０は、輸送サービスの手配や交通手段の管理を行う。「輸送サービス」は、車両に乗った旅客を輸送するサービスである。「交通手段」は、典型的には車両であり、例えば、定期運行型、デマンド交通型、シェア型又はＤＨ（ダイナミックヘッドウェイ）型の交通システムにおける車両（例えば、鉄道、バス、タクシー、乗用車又は自転車）である。本明細書において、旅客の「輸送」は、旅客自身が車両（例えば、シェア型の交通システムにおける乗用車又は自転車）を運転して移動することも含む広義の輸送である。 The traffic operation management server 20 arranges transportation services and manages means of transportation. A “transport service” is a service that transports passengers in vehicles. "Transportation means" is typically a vehicle, for example, a vehicle in a regular operation type, demand traffic type, share type or DH (dynamic headway) type transportation system (e.g., railway, bus, taxi, passenger car or bicycle). In the present specification, "transportation" of passengers is a broad definition that includes transportation by the passengers themselves driving a vehicle (for example, a passenger car or a bicycle in a shared transportation system).
　旅客端末３０は、旅客の情報処理端末（典型的には、スマートフォンのようなモバイル型の情報処理端末）である。旅客端末３０は、例えば、輸送サービスの予約やチケッティングに用いられる。 The passenger terminal 30 is a passenger information processing terminal (typically, a mobile information processing terminal such as a smartphone). The passenger terminal 30 is used, for example, for reservations and ticketing of transportation services.
　需給マッチングサーバ１０は、複数の交通運行管理サーバ２０と複数の旅客端末３０と通信し、移動需要と輸送サービスとのマッチングである需給マッチングを行う。 The demand-supply matching server 10 communicates with a plurality of traffic operation management servers 20 and a plurality of passenger terminals 30 to perform demand-supply matching, which is matching of travel demand and transportation services.
　図２は、需給マッチングサーバ１０のハードウェア構成例を示す。 FIG. 2 shows a hardware configuration example of the supply and demand matching server 10. FIG.
　需給マッチングサーバ１０は、Ｉ／Ｏインターフェース装置１０７と、通信インターフェース装置１０５と、プログラムやデータを記憶する記憶装置（永続記憶装置１０４及びメモリ１０３）と、それらに接続されたプロセッサ１０１とを有する。Ｉ／Ｏインターフェース装置１０７は、キーボードやマウス等の入力デバイス１０８と、液晶ディスプレイや有機ＥＬディスプレイ等の表示デバイス１０９のインターフェースである。 The supply and demand matching server 10 has an I/O interface device 107, a communication interface device 105, storage devices (permanent storage device 104 and memory 103) for storing programs and data, and a processor 101 connected to them. The I/O interface device 107 is an interface between an input device 108 such as a keyboard and a mouse, and a display device 109 such as a liquid crystal display or an organic EL display.
　図３は、需給マッチングサーバ１０の機能構成例を示す。 FIG. 3 shows an example of the functional configuration of the supply and demand matching server 10.
　需給マッチングサーバ１０は、移動需要ＤＢ１２８と、仮経路ＤＢ１２９と、確定経路ＤＢ１３０と、旅客受け入れ履歴ＤＢ１３１と、交通ネットワークＤＢ１３２とを備える。これらのＤＢは、例えば永続記憶装置１０４に格納される。 The supply and demand matching server 10 comprises a travel demand DB 128, a temporary route DB 129, a fixed route DB 130, a passenger acceptance history DB 131, and a transportation network DB 132. These DBs are stored, for example, in persistent storage 104 .
　移動需要ＤＢ１２８は、旅客の移動需要を旅客毎に表す情報を格納するデータベースである。本実施形態において、「移動需要」は、顧客の移動に関する要素（例えば、目的地、出発地及び出発時刻など）と、顧客が要求するサービスレベルとで構成される。すなわち、移動需要ＤＢ１２８において、旅客毎に、旅客の移動需要のＩＤと、移動需要を表す情報は、当該旅客の移動情報（例えば、目的地、出発地及び出発時刻などを表す情報）と、当該旅客のサービスレベル情報（輸送サービスについて顧客が要求するサービスレベルを表す情報）とを含む。「サービスレベル」は、輸送サービスに関する複数のサービス項目の各々についてのレベルである。また、本実施形態では、移動需要のＩＤは、旅客のＩＤと同義でよい。すなわち、同一の移動需要が旅客Ａ及びＢから登録された場合、旅客Ａにより登録された移動需要のＩＤと旅客Ｂにより登録された移動需要のＩＤが異なっていてよい。 The movement demand DB 128 is a database that stores information representing the movement demand of each passenger. In this embodiment, the "movement demand" consists of elements related to the customer's movement (eg, destination, departure point, departure time, etc.) and the service level requested by the customer. That is, in the movement demand DB 128, for each passenger, the ID of the movement demand of the passenger and the information representing the movement demand are divided into the movement information of the passenger (for example, information representing the destination, departure place and departure time) and the relevant Passenger service level information (information representing the service level requested by the customer for transportation services). A "service level" is a level for each of a plurality of service items relating to transportation services. Further, in this embodiment, the ID of travel demand may be synonymous with the ID of a passenger. That is, when the same travel demand is registered by passengers A and B, the ID of the travel demand registered by passenger A and the ID of the travel demand registered by passenger B may be different.
　仮経路ＤＢ１２９は、旅客の目的地から出発地までの仮の経路を旅客毎に表す情報を格納するデータベースである。例えば、仮経路ＤＢ１２９は、旅客の移動需要のＩＤ毎に、仮経路を表してよい。 The temporary route DB 129 is a database that stores information representing a temporary route from the passenger's destination to the departure point for each passenger. For example, the temporary route DB 129 may represent a temporary route for each ID of travel demand of passengers.
　確定経路ＤＢ１３０は、旅客の目的地から出発地までの確定した経路を旅客毎に表す情報を格納するデータベースである。例えば、確定経路ＤＢ１３０は、旅客の移動需要のＩＤ毎に、確定経路を表してよい。 The fixed route DB 130 is a database that stores information representing the fixed route from the passenger's destination to the departure point for each passenger. For example, the fixed route DB 130 may represent a fixed route for each ID of travel demand of a passenger.
　旅客受け入れ履歴ＤＢ１３１は、新しい経路候補の受け入れ可否に関する履歴を格納するデータベースである。当該履歴は、例えば、旅客の移動需要のＩＤ毎に、一つ以上の新しい経路候補の各々について当該新しい経路候補を受け入れるか否かを表してよい。 The passenger acceptance history DB 131 is a database that stores the history of acceptance or rejection of new route candidates. The history may indicate, for example, whether or not the new route candidate is accepted for each of one or more new route candidates for each passenger travel demand ID.
　交通ネットワークＤＢ１３２は、交通ネットワークを表す情報を格納するデータベースである。「交通ネットワーク」は、複数のノードとそれぞれノード間を結ぶ複数のリンクとで構成されたグラフ構造のネットワークである。ノードは、旅客が乗り降り可能な予め決められた車両の停止場所である。リンクは、それらの停止場所を結ぶ物理的又は論理的な路（例えば、線路、道路又は空路）である。交通ネットワークは、複数のサブ交通ネットワークで構成される。一つのサブ交通ネットワークは、複数の輸送サービス（複数の交通手段）のうちの一つの輸送サービスに対応する。交通ネットワークは、複数のサブ交通ネットワークが論理的に同一レイヤとされた一つのネットワークである。鉄道、バス及びタクシーといった複数の交通手段のうちの任意の二つ以上の交通手段の利用が可能な場所に対応したノードは、当該二つ以上の交通手段に複数のサブ交通ネットワークにおいて共通である。すなわち、複数の輸送サービスでノードが共通していれば、異なるサブ交通ネットワーク間でノードだけが重ねられ（共通となり）、リンクは重ならない。例えば、交通ネットワークでは、ノードＡ（駅Ａ）とノードＢ（駅Ｂ）間に、鉄道用のリンクとバス用のリンクといった複数のリンクが存在し得る。一方、サブ交通ネットワークでは、サブ交通ネットワークは一つの輸送サービスに対応しているため、ノード間を結ぶリンクは一つである。輸送サービス別に、サブ交通ネットワークを表す情報が、交通ネットワークＤＢ１３２に含まれていてよい。また、交通ネットワークＤＢ１３２は、ノード毎に、当該ノードに対応した場所の詳細を表す情報（例えば、名称、位置情報、輸送サービス毎に折返し可能か否か、輸送サービス毎の時刻表、混雑状況、遅延状況等）を含んでよく、また、リンク毎に、当該リンクに対応した路の詳細を表す情報（例えば、対応する輸送サービス（交通手段）、混雑状況、遅延状況等）を含んでよい。 The transportation network DB 132 is a database that stores information representing transportation networks. A "traffic network" is a graph-structured network composed of a plurality of nodes and a plurality of links connecting the nodes. A node is a pre-determined vehicle stopping location where passengers can board and disembark. A link is a physical or logical path (eg, a railroad, road, or airway) connecting those stops. A transport network consists of a plurality of sub-transport networks. One sub-transportation network corresponds to one transportation service out of a plurality of transportation services (a plurality of transportation means). A traffic network is one network in which a plurality of sub-transport networks are logically placed on the same layer. A node corresponding to a place where any two or more modes of transportation such as railways, buses, and taxis can be used is common to the two or more modes of transportation in a plurality of sub-transportation networks. . That is, if a node is common to a plurality of transport services, only nodes are overlapped (common) between different sub-transport networks, and links are not overlapped. For example, in a transportation network, there may be multiple links between node A (station A) and node B (station B), such as a rail link and a bus link. On the other hand, in the sub-transport network, since the sub-transport network corresponds to one transportation service, there is one link connecting nodes. Information representing a sub-transport network may be included in the transport network DB 132 for each transport service. In addition, the transportation network DB 132 stores, for each node, information representing the details of the location corresponding to the node (for example, name, location information, whether or not return is possible for each transportation service, timetable for each transportation service, congestion status, delay status, etc.), and for each link, information representing details of the road corresponding to the link (for example, corresponding transport service (transport means), congestion status, delay status, etc.) may be included.
　以下の説明では、一つの輸送サービスについて一つのリンクを挟む二つのノードの各々を、便宜上、「隣接ノード」と呼ぶことがある。つまり、隣接ノード間に複数のリンクがある場合、隣接ノード間に異なる複数の輸送サービスがある（一方の隣接ノードに対応した一方の場所から他方の隣接ノードに対応した他方の場所への輸送のサービスとして複数の輸送サービスがある）。 In the following explanation, each of the two nodes sandwiching one link for one transportation service may be called an "adjacent node" for convenience. That is, if there are multiple links between adjacent nodes, there are multiple different transportation services between adjacent nodes (transport from one location corresponding to one adjacent node to another location corresponding to another adjacent node). There are multiple transportation services as a service).
　また、以下の説明では、出発地（ノード）から目的地（ノード）への連続した一つ以上のリンクの集合を、「経路」と呼ぶことができる。 Also, in the following description, a set of one or more continuous links from a starting point (node) to a destination (node) can be called a "route".
　プログラムがプロセッサ１０１に実行されることで、旅客アプリケーション部１２１と、経路検索部１２２と、部分ネットワーク抽出部１２３と、需要クラスタリング部１２４と、旅客カウント部１２５と、需給最適化部１２６と、運行手配部１２７といった機能が実現される。これらの機能の少なくとも一つが、ＤＢ１２８～１３２の少なくとも一つを適宜に参照又は更新する。旅客アプリケーション部１２１が、旅客端末３０と通信する。運行手配部１２７が、交通運行管理サーバ２０と通信する。これらの機能１２１～１２７の詳細は後に説明する。 When the program is executed by the processor 101, a passenger application unit 121, a route search unit 122, a partial network extraction unit 123, a demand clustering unit 124, a passenger counting unit 125, a supply and demand optimization unit 126, an operation A function such as the ordering unit 127 is realized. At least one of these functions appropriately references or updates at least one of the DBs 128-132. A passenger application unit 121 communicates with the passenger terminal 30 . The operation arrangement unit 127 communicates with the traffic operation management server 20 . Details of these functions 121-127 will be described later.
　図４は、旅客アプリケーション処理の流れを示す。 Figure 4 shows the flow of passenger application processing.
　旅客アプリケーション部１２１は、移動情報及びサービスレベル情報の入力を旅客端末３０から受け付け、当該情報を移動需要ＤＢ１２８に登録する（Ｓ１）。例えば、旅客アプリケーション部１２１は、図５Ａに例示する輸送サービス要求ＵＩ１４０（輸送サービスの要求を入力するための画面）を旅客端末３０に表示する。輸送サービス要求ＵＩ１４０を介して、移動情報（例えば、出発地、目的地、出発時刻及び／又は到着時刻）と、サービスレベル情報（複数のサービス項目の各々についてのレベル）とが入力される。サービスレベル情報の入力はオプションでよい。 The passenger application unit 121 receives input of travel information and service level information from the passenger terminal 30, and registers the information in the travel demand DB 128 (S1). For example, the passenger application unit 121 displays a transportation service request UI 140 (screen for inputting a request for transportation service) illustrated in FIG. 5A on the passenger terminal 30 . Travel information (eg, origin, destination, departure time and/or arrival time) and service level information (level for each of a plurality of service items) are entered via the transportation service request UI 140 . Entering service level information may be optional.
　旅客アプリケーション部１２１は、旅客端末３０から入力された上述の移動情報（及びサービスレベル情報）を基に、経路検索部１２２に経路の検索指示を出し、当該指示に応答して経路検索部１２２により検索された経路である仮経路を表す情報を仮経路ＤＢ１２９に登録する（Ｓ２）。検索指示には、入力された上述の移動情報（及びサービスレベル情報）の少なくとも一部（例えば、出発地、目的地、出発時刻及び／又は到着時刻、及び、複数のサービス項目の各々についてのレベル）が検索条件として関連付けられる。経路検索部１２２は、検索指示に応答して、検索指示に関連付けられている検索条件を満たす経路を交通ネットワークＤＢ１３２から検索し、検索された経路を表す情報を返すようになっている。 The passenger application unit 121 issues a route search instruction to the route search unit 122 based on the above-described movement information (and service level information) input from the passenger terminal 30, and in response to the instruction, the route search unit 122 Information representing the searched provisional route is registered in the provisional route DB 129 (S2). The search instruction includes at least a part of the input travel information (and service level information) (for example, departure place, destination, departure time and/or arrival time, and level for each of a plurality of service items). ) is associated as a search condition. In response to the search instruction, the route search unit 122 searches the traffic network DB 132 for a route that satisfies the search conditions associated with the search instruction, and returns information representing the searched route.
　旅客アプリケーション部１２１は、経路確定の期限が過ぎたか否かを判定する（Ｓ３）。「経路確定」とは、一つ又は複数の経路候補のうちの一つの経路候補を確定経路とすることである。「一つ又は複数の経路候補」は、Ｓ２で登録された仮経路と、Ｓ５で受け入れ可とされた一つ以上の新しい経路候補とのうちの少なくとも仮経路である。「経路確定の期限」は、所定のポリシーで決められた期限でよい（例えば、Ｓ２で仮経路が登録されてから一定期間経った時点でもよいし、最後に新しい経路候補が受入れ可とされてから一定時間経った時点でもよいし、旅客に指定された期限でもよい）。 The passenger application unit 121 determines whether or not the deadline for route confirmation has passed (S3). "Route determination" means to set one route candidate out of one or more route candidates as a definite route. "One or more route candidates" is at least the provisional route selected from the provisional route registered in S2 and one or more new route candidates accepted in S5. The "deadline for route determination" may be a deadline determined by a predetermined policy (for example, it may be a time when a certain period of time has passed since the provisional route was registered in S2, or it may be when a new route candidate is finally accepted. (It may be after a certain period of time has passed since, or it may be the time limit specified by the passenger).
　Ｓ３の判定結果が偽の場合（Ｓ３：Ｎｏ）、旅客アプリケーション部１２１は、経路変更通知を受けたか否かを判定する（Ｓ４）。Ｓ４の判定結果が偽の場合（Ｓ４：Ｎｏ）、処理がＳ３に戻る。ここで、「経路変更通知」とは、仮経路（及び通知済の新しい経路候補）とは異なる新しい経路候補が見つかったことの通知である。この通知は、図６に示す需給マッチングバッチ処理において出される（具体的には、図１５のＳ６７で出る通知が、経路変更通知である）。 If the determination result of S3 is false (S3: No), the passenger application unit 121 determines whether or not a route change notification has been received (S4). If the determination result of S4 is false (S4: No), the process returns to S3. Here, the "route change notification" is a notification that a new route candidate different from the provisional route (and the notified new route candidate) has been found. This notification is issued in the demand-supply matching batch process shown in FIG. 6 (specifically, the notification issued in S67 of FIG. 15 is the route change notification).
　Ｓ４の判定結果が真の場合（Ｓ４：Ｙｅｓ）、旅客アプリケーション部１２１は、経路変更通知に示される新しい経路の候補の受け入れ可否を旅客端末３０から受け付ける（Ｓ５）。例えば、旅客アプリケーション部１２１は、図５Ｂに例示する受入れ問合せＵＩ１５０（新しい経路候補を受け入れるか否かを入力するための画面）を旅客端末３０に表示する。図５Ｂが示す例によれば、二つの新しい経路候補があり、それぞれにチェックボックス５２が用意されている。旅客は、受け入れる新しい経路を選択できる（チェックボックスにチェックを入れることができる）。旅客アプリケーション部１２１は、選択の内容（各新しい経路候補について受け入れが可能か否かを示す情報）を旅客受け入れ履歴ＤＢ１３１に登録する（Ｓ６）。その後、処理がＳ３に戻る。 If the determination result of S4 is true (S4: Yes), the passenger application unit 121 receives from the passenger terminal 30 whether or not the new route candidate indicated in the route change notification can be accepted (S5). For example, the passenger application unit 121 displays on the passenger terminal 30 an acceptance inquiry UI 150 (a screen for inputting whether or not to accept the new route candidate) illustrated in FIG. 5B. According to the example shown in FIG. 5B, there are two new route candidates, each provided with a checkbox 52 . Passengers can select new routes to accept (check boxes). The passenger application unit 121 registers the content of the selection (information indicating whether or not each new route candidate can be accepted) in the passenger acceptance history DB 131 (S6). After that, the process returns to S3.
　Ｓ３の判定結果が真の場合（Ｓ３：Ｙｅｓ）、旅客アプリケーション部１２１は、確定経路を決定し、仮経路ＤＢ１２９から、Ｓ２で登録した仮経路の情報を削除し、確定経路を表す情報を確定経路ＤＢ１３０に登録する（Ｓ７）。「確定経路」は、上述したように、一つ又は複数の経路候補のうちの一つの経路候補であるが、具体的には、例えば、下記でよい。
・一つ以上の受入れ可の新しい経路候補を表す情報が旅客受け入れ履歴ＤＢ１３１に登録されている場合、当該一つ以上の受入れ可の新しい経路候補から得られた一つの新しい経路候補が、確定経路となる。この新しい経路候補は、最後に受入れ可とされた新しい経路候補でもよいし、ランダムに選択された新しい経路候補でもよいし、旅客から選択された経路候補でもよい。
・受入れ可の新しい経路候補が一つも無い場合、仮経路が、確定経路となる。 If the determination result of S3 is true (S3: Yes), the passenger application unit 121 determines the confirmed route, deletes the information on the temporary route registered in S2 from the temporary route DB 129, and confirms the information representing the confirmed route. Register in the route DB 130 (S7). A "fixed route" is one route candidate out of one or a plurality of route candidates, as described above.
・When information representing one or more acceptable new route candidates is registered in the passenger acceptance history DB 131, one new route candidate obtained from the one or more acceptable new route candidates is a confirmed route. becomes. This new route candidate may be the last accepted new route candidate, a randomly selected new route candidate, or a route candidate selected from passengers.
- If there is no new acceptable route candidate, the tentative route becomes the final route.
　Ｓ７の後、旅客アプリケーション部１２１は、確定経路に従い移動するための輸送サービスのデジタルチケットを旅客端末３０に発行する（Ｓ８）。発行されたデジタルチケットを表示した画面であるチケットＵＩ１６０が、図５Ｃに例示の通り旅客端末３０に表示される。チケットＵＩ１６０には、確定経路「Ａ→Ｃ：鉄道、Ｃ→Ｅ：デマンドバス」（Ａ地点からＣ地点へは鉄道で移動し、Ｃ地点からＥ地点まではデマンドバスで移動する経路）を表す情報が表示される。また、チケットＵＩ１６０には、確定経路のデジタルチケットのオブジェクト１６３が表示される。オブジェクト１６３が指定されると、チケットの内容の詳細が表示される。 After S7, the passenger application unit 121 issues to the passenger terminal 30 a digital ticket for transportation services for traveling along the fixed route (S8). A ticket UI 160, which is a screen displaying the issued digital ticket, is displayed on the passenger terminal 30 as illustrated in FIG. 5C. The ticket UI 160 shows a fixed route "A→C: railroad, C→E: demand bus" (route from point A to point C by rail, and from point C to point E by demand bus). Information is displayed. The ticket UI 160 also displays a digital ticket object 163 of the fixed route. When object 163 is specified, details of the contents of the ticket are displayed.
　図４に示した処理によれば、「経路確定の期限」になるまでの間（Ｓ３：Ｙｅｓとなるまでの間）、図６に示すバッチ処理において「新しい経路」が見つかった場合には、Ｓ４：ＹｅｓとなりＳ５が行われる。一方、「経路確定の期限」になるまでの間、図６に示すバッチ処理が行われない（例えば、「経路確定の期限」までの期間が短い（例えば、リアルタイムでの応答が求められている））又はバッチ処理が行われても「新しい経路」が見つからなかった場合には、Ｓ７で仮経路が確定経路となる。 According to the process shown in FIG. 4, if a "new route" is found in the batch process shown in FIG. S4: Yes and S5 is performed. On the other hand, the batch processing shown in FIG. 6 is not performed until the "deadline for route confirmation" is reached (for example, the period until the "deadline for route confirmation" is short (for example, a real-time response is required). )) or if the "new route" is not found even after the batch processing is performed, the temporary route becomes the fixed route in S7.
　図６は、需給マッチングバッチ処理の流れを示す。この処理は、定期的に行われる。 Fig. 6 shows the flow of supply and demand matching batch processing. This process is performed periodically.
　部分交通ネットワークの抽出処理が行われる（Ｓ１１）。次に、需要クラスタ生成処理が行われる（Ｓ１２）。最後に、需給最適化処理が行われる（Ｓ１３）。以下、これらＳ１１～Ｓ１３の処理を詳細に説明する。 A partial traffic network extraction process is performed (S11). Next, demand cluster generation processing is performed (S12). Finally, supply and demand optimization processing is performed (S13). The processing of these S11 to S13 will be described in detail below.
　なお、部分交通ネットワークにおいて、連続した一つ以上のリンクの集合を、「区間」と呼ぶことができる。「区間」は、経路の一部であることもあり得る。区間を構成する「連続した一つ以上のリンク」では、同一の隣接ノード間に存在するリンクは一つであり、且つ、区間が二つ以上のリンクの場合、リンクとリンクがノードを介して繋がっている。本実施形態では、「区間」は、典型的には、一つのリンク、又は、連続した二つのリンク（同一ノードに繋がった二つのリンク）でよい。 In addition, in a partial transportation network, a set of one or more continuous links can be called a "section". A "segment" can be part of a route. In the "one or more continuous links" that make up the section, there is only one link between the same adjacent nodes, and if the section consists of two or more links, the link and the link are connected through the node It is connected. In this embodiment, a "section" may typically be one link or two consecutive links (two links connected to the same node).
　また、以下の説明では、部分交通ネットワークにおいて、連続した一つ以上の区間を「パス」と呼ぶことができる。「パス」は、経路の一部であることもあり得る。 Also, in the following description, one or more consecutive sections in a partial traffic network can be called a "path". A "path" can be part of a route.
　＜図６のＳ１１：部分交通ネットワークの抽出処理＞ <S11 in Fig. 6: Extraction processing of partial traffic network>
　図７は、部分交通ネットワークの抽出処理（図６のＳ１１）の流れを示す。 FIG. 7 shows the flow of the partial traffic network extraction process (S11 in FIG. 6).
　部分ネットワーク抽出部１２３は、要求サービスレベルが満たされる旅客の数を輸送サービス（サブ交通ネットワーク）のリンク毎に計算する（Ｓ２１）。具体的には、例えば、下記が行われる。
・部分ネットワーク抽出部１２３が、移動需要ＤＢ１２８及び確定経路１３０を参照し、経路が確定していない移動需要を特定する。
・部分ネットワーク抽出部１２３が、経路が確定していない移動需要毎に、取り得る経路（仮経路や新経路（新しい経路候補））を、仮経路ＤＢ１２９及び旅客受け入れ履歴ＤＢ１３１から特定する。
・部分ネットワーク抽出部１２３が、特定された取り得る経路毎に、当該経路に対応した移動需要と、交通ネットワークＤＢ１３２（例えば、交通ネットワークの構成、及び、リンク毎の情報）とから、リンク毎に、要求サービスレベルが満たされる旅客の数を、輸送サービス（サブ交通ネットワーク）のリンク毎に計算する。 The partial network extraction unit 123 calculates the number of passengers satisfying the required service level for each link of the transport service (sub-transport network) (S21). Specifically, for example, the following is performed.
- The partial network extraction unit 123 refers to the movement demand DB 128 and the fixed route 130, and identifies a movement demand for which the route is not fixed.
- The partial network extraction unit 123 identifies possible routes (provisional routes and new routes (new route candidates)) from the provisional route DB 129 and the passenger acceptance history DB 131 for each travel demand whose route is not fixed.
・The partial network extraction unit 123 extracts, for each possible route that has been specified, from the travel demand corresponding to the route and the traffic network DB 132 (for example, the configuration of the traffic network and information for each link), for each link , the number of passengers for which the required service level is met is calculated for each link of the transport service (sub-transport network).
　部分ネットワーク抽出部１２３は、輸送サービスの運用効率を輸送サービスのリンク毎に計算する（Ｓ２２）。例えば、運用効率は、コストに対する予測収入の割合でよい。具体的には、例えば、部分ネットワーク抽出部１２３は、交通ネットワークＤＢ１３２及び確定経路１３０を基に、輸送サービスのリンク毎に、時間帯における旅客数と車両運行数とを特定し、当該旅客数及び車両運行数とを基に、運用効率を計算してよい。なお、輸送サービスのリンク毎に、時間帯における旅客数の特定は、各交通運行管理サーバから入手可能な過去の運行管理情報を基に、時間帯における旅客数を予測することでもよい。 The partial network extraction unit 123 calculates the operation efficiency of the transportation service for each link of the transportation service (S22). For example, operational efficiency may be the ratio of expected revenue to cost. Specifically, for example, the partial network extraction unit 123 identifies the number of passengers and the number of vehicle operations in a time period for each link of the transportation service based on the transportation network DB 132 and the fixed route 130, Operational efficiency may be calculated based on the number of vehicles in operation. Note that the number of passengers in a time slot for each transport service link may be determined by predicting the number of passengers in a time slot based on past operation management information available from each traffic operation management server.
　部分ネットワーク抽出部１２３は、Ｓ２１で算出された旅客数（サービスレベルが満たされる旅客の数）が絶対的に又は相対的に少ないリンク、及び／又は、Ｓ２２で算出された運用効率が絶対的に又は相対的に低いリンクをリストアップする（Ｓ２３）。Ｓ２１で算出された旅客数が少ないリンクとは、具体的には、当該旅客数が閾値θｐａｓｓｅｎｇｅｒを下回ったリンクである。Ｓ２２で算出された運用効率が低いリンクとは、当該運用効率が閾値θｔｒａｎｓｐｏｒｔａｔｉｏｎを下回ったリンクである。θｔｒａｎｓｐｏｒｔａｔｉｏｎは、予測収入がコストより小さくなる輸送サービスの数の閾値とされてもよい。 The partial network extraction unit 123 extracts a link with an absolute or relatively small number of passengers (the number of passengers satisfying the service level) calculated in S21, and/or an absolute Or list relatively low links (S23). Specifically, the link with the small number of passengers calculated in S21 is the link with the number of passengers below the threshold θ passenger . A link with a low operational efficiency calculated in S22 is a link whose operational efficiency is below the threshold θ transportation . θ transportation may be thresholded to the number of transportation services for which the expected revenue is less than the cost.
　部分ネットワーク抽出部１２３は、Ｓ２３でリストアップされたリンクの情報を基に部分交通ネットワーク生成処理を行う（Ｓ２４）。以下、Ｓ２３でリストアップされた各リンクを、「問題リンク」と総称することがある。「問題リンク」は、サービスレベルが満たされる旅客の数が少ないリンク、又は、運用効率が低いリンクである。 The partial network extraction unit 123 performs partial traffic network generation processing based on the link information listed in S23 (S24). Hereinafter, each link listed in S23 may be collectively referred to as "problem link". A "problem link" is a link with a low number of passengers meeting its service level or a link with low operational efficiency.
　図８は、部分交通ネットワーク生成処理（図７のＳ２４）の流れを示す。 FIG. 8 shows the flow of the partial traffic network generation process (S24 in FIG. 7).
　部分ネットワーク抽出部１２３は、仮の部分交通ネットワークを構成する（Ｓ３１）。仮の部分交通ネットワークは、リストアップされた問題リンク毎にその近傍のリンクがつなげられたネットワークである。「近傍のリンク」は、問題リンクを挟む両端のノードのいずれかに繋がっているノードでよい。問題リンクの近傍のリンクは、別の問題リンクであることもあれば、問題リンクではないリンクであることもある。Ｓ３１では、例えば図９の上側に示すように、太線で表現された問題リンクに近傍リンクがつなげられることで、図９の下側に示すような部分交通ネットワークが構成される。なお、図９の下側に例示の部分交通ネットワークの端点のノードは、ノードＡ及びＥである。破線矢印は、部分交通ネットワークの外から部分交通ネットワークに入る旅客や、部分交通ネットワークからその外に出る旅客を意味する。 The partial network extraction unit 123 constructs a temporary partial traffic network (S31). A temporary partial traffic network is a network in which links in the vicinity of each listed problem link are connected. A "neighboring link" may be a node connected to either of the nodes on either end of the problem link. A link near the problem link may be another problem link or a non-problem link. In S31, for example, as shown in the upper part of FIG. 9, a partial traffic network as shown in the lower part of FIG. 9 is constructed by connecting neighboring links to the problem link represented by the thick line. Nodes A and E are end points of the partial traffic network illustrated in the lower part of FIG. Dashed arrows denote passengers entering or exiting a partial transportation network from outside the partial transportation network.
　部分ネットワーク抽出部１２３は、仮の部分交通ネットワークの端点のノードが運用制約を満たすように仮の部分交通ネットワークを広げる（Ｓ３２）。「端点のノードが運用制約を満たす」とは、端点のノードが、当該ノードに関わる全ての輸送サービスについて、図１１に例示の運用制約テーブル１７０が表す運用制約を満たすことである。運用制約テーブル１７０は、例えば交通ネットワークＤＢ１３２に含まれる。運用制約テーブル１７０は、ノードと輸送サービスの組毎に、ノードでの折り返し可否を表す。図１１が示す例によれば、鉄道や路線バスは、折り返し可能な場所が限られ、臨時便の区間の設定に制約があることが考慮されている。また、デマンド交通やタクシーなどの輸送サービスについては、全ての場所が折り返し可能として管理される。 The partial network extraction unit 123 expands the temporary partial traffic network so that the nodes at the endpoints of the temporary partial traffic network satisfy the operational constraints (S32). “End point node satisfies operational constraints” means that the terminal node satisfies the operational constraints represented by the operational constraint table 170 illustrated in FIG. 11 for all transport services related to the node. The operational constraint table 170 is included in the traffic network DB 132, for example. The operation constraint table 170 indicates whether or not a return can be made at a node for each pair of a node and transportation service. According to the example shown in FIG. 11, it is taken into consideration that railroads and fixed-route buses have limited places where they can turn back, and there are restrictions on the setting of sections for temporary services. In addition, transportation services such as demand transportation and taxis are managed so that all locations can be turned back.
　なお、運用制約は緩くされてもよい。例えば、「端点のノードで折り返しできない輸送サービスの数又は割合がα以下」が、運用制約であってもよい。すなわち、「α」は、端点のノードに繋がるリンクの数（輸送サービスの数）でもよいし、端点のノードに繋がるリンクの数に対する、折り返しできない輸送サービスの数の割合、でもよい。本実施形態では、通常、「α」の値は０である。すなわち、本実施形態では、原則として、各輸送サービスについて、部分交通ネットワークの端点ノードは、折り返し可能なノードである。 It should be noted that operational restrictions may be relaxed. For example, "the number or ratio of transport services that cannot be returned at the endpoint node is less than or equal to α" may be an operational constraint. That is, “α” may be the number of links (the number of transport services) connected to the endpoint node, or the ratio of the number of transport services that cannot be returned to the number of links connected to the endpoint node. In this embodiment, the value of "α" is normally 0. That is, in this embodiment, in principle, for each transportation service, the end point node of the partial transportation network is a node that can be turned back.
　部分ネットワーク抽出部１２３は、重複するリンクを持つ複数の部分交通ネットワークがある場合（つまり、一部が互いに重複した複数の部分交通ネットワークがある場合）、それらの部分交通ネットワークを結合させる（Ｓ３３）。これにより、それら複数の部分交通ネットワークは一つの部分交通ネットワークとされる。そのような複数の部分交通ネットワークが無い場合、Ｓ３３はスキップされる。 If there are multiple partial traffic networks with overlapping links (that is, there are multiple partial traffic networks that partially overlap each other), the partial network extraction unit 123 combines these partial traffic networks (S33). . As a result, these partial traffic networks are made into one partial traffic network. If there are no such multiple partial traffic networks, S33 is skipped.
　その後、部分ネットワーク抽出部１２３は、部分交通ネットワーク毎に、部分交通ネットワーク再構成処理を行う（Ｓ３４）。この結果、再構成された部分交通ネットワークがあれば（Ｓ３５：Ｙｅｓ）、処理が、Ｓ３３に戻る。 After that, the partial network extraction unit 123 performs partial traffic network reconfiguration processing for each partial traffic network (S34). As a result, if there is a reconfigured partial traffic network (S35: Yes), the process returns to S33.
　図１０は、部分交通ネットワーク再構成処理（図８のＳ３４）の流れを示す。図９の説明では、一つの部分交通ネットワークを例に取る（図９の説明において「対象ネットワーク」）。 FIG. 10 shows the flow of the partial traffic network reconfiguration process (S34 in FIG. 8). In the explanation of FIG. 9, one partial traffic network is taken as an example (“target network” in the explanation of FIG. 9).
　部分ネットワーク抽出部１２３は、対象ネットワークにおける問題リンクを最も多く通る経路について代替パス数を計算する（Ｓ４１）。ここで、図９の例によれば、経路がノードＢ及びＤを経由し、問題リンクがリンクＴ５の場合、代替パス数は“３”、すなわち、代替パスは、リンクＴ４及びＴ７の組、リンクＴ４及びＴ８の組、及び、リンクＴ６である。 The partial network extraction unit 123 calculates the number of alternative paths for the route that passes through the most problematic links in the target network (S41). Here, according to the example of FIG. 9, when the route passes through nodes B and D and the problem link is link T5, the number of alternative paths is "3", that is, the alternative paths are the pair of links T4 and T7, A pair of links T4 and T8 and a link T6.
　部分ネットワーク抽出部１２３は、Ｓ４１で算出された代替パス数がβｍａｘ以下か否かを判定する（Ｓ４２）。βｍａｘは、部分交通ネットワーク内の代替パス数の上限である。この値を設定することで、部分交通ネットワークの大きさを制限し、以って、後述の需給最適化処理（図１５）において組合せ爆発を防ぐことができる。 The partial network extraction unit 123 determines whether or not the number of alternative paths calculated in S41 is equal to or less than β max (S42). β max is the upper bound on the number of alternative paths in the partial traffic network. By setting this value, it is possible to limit the size of the partial traffic network, thereby preventing combinatorial explosion in the later-described supply and demand optimization process (FIG. 15).
　Ｓ４２の判定結果が真の場合（Ｓ４２：Ｙｅｓ）、部分ネットワーク抽出部１２３は、Ｓ４１で算出された代替パス数がβｍｉｎ以上であるか否かを判定する（Ｓ４３）。βｍｉｎは、部分交通ネットワーク内の代替パス数の下限である。この値を設定することで、問題最適化のための代替パスが設定できないことを防ぐことができる。Ｓ４３の判定結果が真の場合（Ｓ４３：Ｙｅｓ）、処理が終了する。Ｓ４３の判定結果が偽の場合（Ｓ４３：Ｎｏ）、部分ネットワーク抽出部１２３は、対象ネットワークの両端のノードから最も近傍の折り返し可能ノードまで対象ネットワークを拡大させる（Ｓ４４）。拡大後のネットワークについては、代替パス数がβｍｉｎ以上となることが期待される。 If the determination result of S42 is true (S42: Yes), the partial network extraction unit 123 determines whether or not the number of alternative paths calculated in S41 is equal to or greater than β min (S43). β min is the lower bound on the number of alternative paths in the partial traffic network. Setting this value prevents the failure to set alternate paths for problem optimization. If the determination result of S43 is true (S43: Yes), the process ends. If the determination result of S43 is false (S43: No), the partial network extraction unit 123 expands the target network from the nodes at both ends of the target network to the nearest loopable node (S44). For the expanded network, it is expected that the number of alternative paths will be greater than or equal to β min .
　Ｓ４２の判定結果が偽の場合（Ｓ４２：Ｎｏ）、部分ネットワーク抽出部１２３は、対象ネットワーク内に運用制約を満たすノードがあるか否かを判定する（Ｓ４５）。Ｓ４５の判定結果が偽の場合（Ｓ４５：Ｎｏ）、処理が終了する。なお、「運用制約を満たすノード」の有無は、図１１に示した運用制約テーブル１７０から特定される。 If the determination result in S42 is false (S42: No), the partial network extraction unit 123 determines whether or not there is a node that satisfies the operational constraints in the target network (S45). If the determination result of S45 is false (S45: No), the process ends. It should be noted that the presence or absence of "nodes satisfying operational constraints" is identified from the operational constraint table 170 shown in FIG.
　Ｓ４５の判定結果が真の場合（Ｓ４５：Ｙｅｓ）、部分ネットワーク抽出部１２３は、運用制約を満たすノードを境に対象ネットワークを分割しても代替パス数がβｍｉｎ以上か否かを判定する（Ｓ４６）。Ｓ４６の判定結果が偽の場合（Ｓ４６：Ｎｏ）、処理が終了する。Ｓ４６の判定結果が真の場合（Ｓ４６：Ｙｅｓ）、部分ネットワーク抽出部１２３は、対象ネットワークを、運用制約を満たすノードを境に分割する（Ｓ４７）。 If the determination result of S45 is true (S45: Yes), the partial network extraction unit 123 determines whether or not the number of alternative paths is β min or more even if the target network is divided at the nodes that satisfy the operational constraints ( S46). If the determination result of S46 is false (S46: No), the process ends. If the determination result of S46 is true (S46: Yes), the partial network extraction unit 123 divides the target network along the nodes that satisfy the operational constraints (S47).
　以上が、部分交通ネットワークの抽出処理（図６のＳ１１）の説明である。この処理において抽出された部分交通ネットワークを表す情報は、交通ネットワークＤＢ１３２に格納される。また、ここで抽出された部分交通ネットワークは、上述したように、図７のＳ２３でリストアップされた問題リンクを基に構築されたネットワークである。複数の部分交通ネットワークが抽出された場合、各部分交通ネットワークの情報が交通ネットワークＤＢ１３２に格納される。 The above is the description of the partial traffic network extraction process (S11 in FIG. 6). Information representing the partial traffic network extracted in this process is stored in the traffic network DB 132 . Also, the partial traffic network extracted here is a network constructed based on the problem links listed in S23 of FIG. 7, as described above. When multiple partial traffic networks are extracted, information on each partial traffic network is stored in the traffic network DB 132 .
　部分交通ネットワークの抽出処理に関する総括として、例えば下記の通りである。 An example of a summary of the partial traffic network extraction process is as follows.
　部分交通ネットワークは、一つ以上の問題リンクの影響を受け得る範囲としてのネットワークであり交通ネットワークの一部としてのグラフである。このような部分交通ネットワークがない場合、交通ネットワークのいずれかのリンクが問題リンクとなる度に多くの経路を探索しなければならなくなる。これにより、経路数が莫大な数となり、この後の需給最適化において解が得られない（又は、解を得るのに非常に時間がかかってしまう）おそれがある。問題リンクが関係する部分としての部分交通ネットワークを抽出することで、最適化問題を小さくして解を得やすくする（解き易くする）ことができる。言い換えると、本実施形態では、部分交通ネットワーク以外の範囲は最適化問題の対象外であるため、問題リンクが無い範囲についての処理が避けられ、以って、計算に要する時間及びリソースを節約することができる。 A partial traffic network is a network as a range that can be affected by one or more problem links, and a graph as a part of the traffic network. Without such a partial traffic network, many routes would have to be searched each time any link in the traffic network becomes a problem link. As a result, the number of routes becomes enormous, and there is a risk that a solution cannot be obtained in subsequent demand/supply optimization (or it will take a very long time to obtain a solution). By extracting the partial traffic network as the part related to the problem link, the optimization problem can be made smaller and the solution can be obtained (easily solved). In other words, in this embodiment, areas other than partial traffic networks are excluded from the optimization problem, thus avoiding processing areas without problem links, thus saving computational time and resources. be able to.
　＜図６のＳ１２：需要クラスタ生成処理＞ <S12 in Fig. 6: Demand cluster generation processing>
　図１２は、需要クラスタ生成処理（図６のＳ１２）の流れを示す。 FIG. 12 shows the flow of the demand cluster generation process (S12 in FIG. 6).
　需要クラスタリング部１２４は、移動需要ＤＢ１２８、交通ネットワークＤＢ１３２及び仮経路ＤＢ１２９を参照し、部分交通ネットワークを通る仮経路に対応した移動需要（旅客）を特定する（Ｓ５１）。一つの部分交通ネットワークについて、当該部分交通ネットワークを通る仮経路として、一つ以上の仮経路があり得る。 The demand clustering unit 124 refers to the movement demand DB 128, the transportation network DB 132, and the temporary route DB 129, and identifies the movement demand (passengers) corresponding to the temporary route passing through the partial transportation network (S51). For one partial traffic network, there can be one or more temporary routes as temporary routes passing through the partial traffic network.
　需要クラスタリング部１２４は、交通ネットワークＤＢ１３２を基に、部分交通ネットワークのうち仮経路に含まれるノード毎に、時刻を特定する（Ｓ５２）。例えば、交通ネットワークＤＢ１３２が、輸送サービス毎に、各ノードでの到着時刻や出発時刻や、車両の平均速度等を表す情報である時刻管理情報を含む。この時刻管理情報を基に、Ｓ５２における時刻特定が行われてよい。 Based on the transportation network DB 132, the demand clustering unit 124 identifies the time for each node included in the tentative route in the partial transportation network (S52). For example, the traffic network DB 132 includes time management information, which is information representing the arrival time and departure time at each node, the average speed of vehicles, etc., for each transportation service. Based on this time management information, time specification in S52 may be performed.
　需要クラスタリング部１２４は、移動需要ＤＢ１２８から、Ｓ５１で特定された移動需要（旅客）について、サービスレベルを特定する（Ｓ５３）。 The demand clustering unit 124 identifies the service level for the movement demand (passengers) identified in S51 from the movement demand DB 128 (S53).
　需要クラスタリング部１２４は、Ｓ５３で特定されたサービスレベルとＳ５２で特定されたノード毎の時刻から、Ｓ５１で特定された移動需要をクラスタリングする（Ｓ５４）。 The demand clustering unit 124 clusters the travel demand identified in S51 from the service level identified in S53 and the time for each node identified in S52 (S54).
　需要クラスタリング部１２４は、移動需要のクラスタ毎の代表値を算出する（Ｓ５５）。 The demand clustering unit 124 calculates a representative value for each cluster of travel demand (S55).
　以上の通り、クラスタは、同一又は類似の移動需要の集合である。本実施形態では、旅客毎に移動需要が存在するため、クラスタは、移動需要が同一又は類似である旅客の集合と言うこともできる。抽出された部分交通ネットワークを経路（典型的には仮経路）が通る旅客のみについてクラスタが生成されるので、最適化問題を小さくすることができる。 As described above, a cluster is a set of identical or similar travel demands. In this embodiment, each passenger has a travel demand, so a cluster can also be said to be a set of passengers having the same or similar travel demands. Since clusters are generated only for passengers whose routes (typically tentative routes) pass through the extracted partial traffic network, the optimization problem can be reduced.
　また、移動需要のクラスタリングでは、必ずしも出発地と目的地が使用されないでもよい。例えば、部分交通ネットワーク内に出発地か目的地（或いはその両方）が存在する移動需要は、出発地及び目的地の少なくとも一つがクラスタリングに使用されてもよい。一方、出発地と目的地の少なくとも一つが部分交通ネットワークの外にある場合、当該外にある出発地及び／又は目的地は、常にクラスタリングに使用されないでもよいし、或いは、当該外にある出発地及び／又は目的地と部分交通ネットワークの端点のノードとの間の距離を基に、当該外にある出発地及び／又は目的地がクラスタリングに使用されるか否かが決定されてよい。一つのクラスタは、例えば、部分交通ネットワーク内に出発地及び／又は目的地がある、部分交通ネットワーク外の出発地及び／又は目的地の遠さ、及び、要求サービスレベル（例えば、混雑度合、遅延度合、着席確実性といった移動嗜好）といった移動需要要素が同一又は類似の旅客の集合である。 Also, in the clustering of travel demand, the origin and destination may not necessarily be used. For example, at least one of the origin and destination may be used for clustering of travel demands that have an origin or destination (or both) within a partial transportation network. On the other hand, if at least one of the origin and destination is outside the partial transportation network, the origin and/or destination outside may not always be used for clustering, or the origin and destination outside may not be used for clustering. And/or based on the distance between the destination and the endpoint node of the partial traffic network, it may be determined whether the outlying origin and/or destination are used for clustering. One cluster is, for example, the origin and/or destination within the partial transportation network, the distance of the departure and/or destination outside the partial transportation network, and the required service level (e.g., degree of congestion, delay It is a group of passengers with the same or similar movement demand factors such as degree of movement, movement preference such as seating certainty.
　以下、図１２に示した需要クラスタ生成処理をより詳細に説明する。 The demand cluster generation processing shown in FIG. 12 will be described in more detail below.
　図１３は、移動需要テーブル１８０の構成例を示す。 FIG. 13 shows a configuration example of the movement demand table 180.
　Ｓ５４の移動需要のクラスタリングでは、移動需要テーブル１８０が参照される。移動需要テーブル１８０は、移動需要ＤＢ１２８に格納される。移動需要テーブル１８０は、移動需要毎にレコードを有する。各レコードは、ＩＤ１８１と、出発地１８２と、出発時刻１８３と、目的地１８４と、到着時刻１８５と、要求サービスレベル１８６といった情報を有する。要求サービスレベル１８６は、要求されるサービスレベルに関し複数のサービス項目の各々についてのレベルを表す情報、例えば、許容混雑１８７と、許容遅延１８８と、着席保障１８９といった情報である。一つの移動需要を例に取る（図１３の説明において「対象移動需要」）。 The travel demand table 180 is referred to in the travel demand clustering in S54. A movement demand table 180 is stored in the movement demand DB 128 . The movement demand table 180 has a record for each movement demand. Each record has information such as ID 181 , origin 182 , departure time 183 , destination 184 , arrival time 185 and requested service level 186 . The requested service level 186 is information representing the level of each of a plurality of service items with respect to the requested service level, such as allowable congestion 187, allowable delay 188, and guaranteed seating 189. FIG. Take one movement demand as an example (“target movement demand” in the description of FIG. 13).
　ＩＤ１８１は、対象移動需要のＩＤを表す。出発地１８２は、対象移動需要における出発地（例えばノードのＩＤ）を表す。出発時刻１８３は、対象移動需要における出発時刻（旅客が出発地を出発する時刻）を表す。目的地１８４は、対象移動需要における目的地（例えばノードのＩＤ）を表す。到着時刻１８５は、対象移動需要における到着時刻（旅客が目的地に到着する時刻）を表す。 The ID 181 represents the ID of the target travel demand. The departure point 182 represents the departure point (for example, the ID of the node) in the target travel demand. The departure time 183 represents the departure time (time at which the passenger departs from the departure point) in the target travel demand. The destination 184 represents the destination (eg ID of the node) in the target travel demand. The arrival time 185 represents the arrival time (time at which the passenger arrives at the destination) in the target travel demand.
　サービス項目として、例えば、許容混雑、許容遅延及び着席保証がある。許容混雑１８７は、旅客が鉄道等の混雑として許容できるレベル（例えば、１５０％までの混雑を許容できる場合、“１５０”）を表す。許容遅延１８８は、旅客が鉄道等の輸送サービスの遅延時間として許容できるレベル（例えば、１０分まで遅延が許容できる場合、“１０”）を表す。着席保障１８９は、旅客が鉄道等の輸送サービスに対して着席のなしを許容できるか否かを要求するレベル（例えば、旅客が輸送サービスに対して、着席なしを許容できる場合、“なし”）を表す。 Service items include, for example, allowable congestion, allowable delays, and seat guarantees. The permissible congestion 187 represents a permissible level of congestion on a railroad or the like for passengers (for example, "150" when a congestion of up to 150% is permissible). The permissible delay 188 represents a level (eg, "10" when a delay of up to 10 minutes is permissible) that a passenger can tolerate as a delay time of transportation services such as railroads. The seating guarantee 189 is a level requesting whether or not a passenger can accept no seat for transportation services such as railways (for example, if a passenger can accept no seating for a transportation service, "None"). represents
　Ｓ５４では、例えば、需要クラスタリング部１２４は、Ｓ５３で特定されたサービスレベルとＳ５２で特定されたノード毎の時刻が類似する移動需要の集合をクラスタとする。具体的には、例えば、需要クラスタリング部１２４は、移動需要毎に、複数種類の情報１８２～１８９の各々の特徴量を算出し、各種情報の特徴量が類似する移動需要の集合をクラスタとする。Ｓ５４において、各クラスタを表す需要クラスタリングテーブルが構築される。 In S54, for example, the demand clustering unit 124 clusters a set of travel demands in which the service level identified in S53 and the time of each node identified in S52 are similar. Specifically, for example, the demand clustering unit 124 calculates the feature amount of each of the plurality of types of information 182 to 189 for each travel demand, and clusters a set of travel demands having similar feature amounts of various types of information. . At S54, a demand clustering table representing each cluster is constructed.
　図１４は、需要クラスタリングテーブル１９０の構成例を示す。なお、図１４の説明において、「統計値」は、例えば平均値であるが、最大値、最小値又は中央値等でもよい。 FIG. 14 shows a configuration example of the demand clustering table 190. In the description of FIG. 14, the "statistical value" is, for example, the average value, but may be the maximum value, minimum value, median value, or the like.
　需要クラスタリングテーブル１９０は、クラスタ毎にレコードを有する。各レコードは、クラスタ番号１９１と、クラスタ所属数１９２、出発地１９３と、出発時刻１９４と、目的地１９５と、到着時刻１９６と、要求サービスレベル１９７といった情報を有する。要求サービスレベル１９７は、許容混雑１９８と、許容遅延１９９と、着席保障２００等を含む。一つのクラスタを例に取る（図１４の説明において「対象クラスタ」）。 The demand clustering table 190 has a record for each cluster. Each record has information such as a cluster number 191, a cluster affiliation number 192, a departure point 193, a departure time 194, a destination 195, an arrival time 196, and a requested service level 197. Required service level 197 includes acceptable congestion 198, acceptable delay 199, seating guarantee 200, and the like. Take one cluster as an example (“target cluster” in the description of FIG. 14).
　クラスタ番号１９１は、対象クラスタの識別番号を表す。クラスタ所属数１９２は、対象クラスタに所属する移動需要の数を表す。 The cluster number 191 represents the identification number of the target cluster. The cluster belonging number 192 represents the number of travel demands belonging to the target cluster.
　出発地１９３は、対象クラスタに属する移動需要における出発地（例えば、図１３の出発地１８２の統計値）を表す。出発時刻１９４は、対象クラスタに属する移動需要における出発時刻（例えば、図１３の出発地の統計値）を表す。目的地１９５は、対象クラスタに属する移動需要における目的地（例えば、図１３の目的地１８４の統計値）を表す。到着時刻１９６は、対象クラスタに属する移動需要における到着時刻（例えば、図１３の到着時刻１８５の統計値）を表す。要求サービスレベル１９７の許容混雑１９８は、対象クラスタに属する移動需要における許容混雑（例えば、図１３の許容混雑１８７の統計値）を表す。要求サービスレベル１９７の許容遅延１９９は、対象クラスタに属する移動需要における許容遅延（例えば、図１３の許容遅延１８８の統計値）を表す。要求サービスレベル１９７の着席保障２００は、対象クラスタに属する移動需要における着席保障（例えば、図１３の着席保障１８９の統計値）を表す。 The departure point 193 represents the departure point (statistical value of the departure point 182 in FIG. 13, for example) in the travel demand belonging to the target cluster. Departure time 194 represents the departure time (for example, the statistical value of the departure point in FIG. 13) in travel demand belonging to the target cluster. The destination 195 represents the destination (statistical value of the destination 184 in FIG. 13, for example) in the travel demand belonging to the target cluster. The arrival time 196 represents the arrival time (statistical value of the arrival time 185 in FIG. 13, for example) of the travel demand belonging to the target cluster. Allowable congestion 198 of requested service level 197 represents the allowable congestion (for example, the statistics of allowable congestion 187 in FIG. 13) for travel demands belonging to the target cluster. Acceptable delay 199 of requested service level 197 represents the acceptable delay (eg, statistics of acceptable delay 188 in FIG. 13) for travel demands belonging to the target cluster. Seating guarantee 200 of requested service level 197 represents the seating guarantee (eg, statistics of seating guarantee 189 in FIG. 13) for travel demand belonging to the target cluster.
　図１２のＳ５５では、例えば、需要クラスタリング部１２４は、クラスタ毎に、当該クラスタに対応した複数種類の情報１９２～２００を基に、代表値を算出する。 In S55 of FIG. 12, for example, the demand clustering unit 124 calculates a representative value for each cluster based on multiple types of information 192 to 200 corresponding to the cluster.
　以上が、需要クラスタ生成処理（図６のＳ１２）の説明である。この処理において、部分交通ネットワークを通る仮経路を基に、複数のクラスタ（移動需要（旅客）のクラスタ）が生成される。 The above is the description of the demand cluster generation process (S12 in FIG. 6). In this process, a plurality of clusters (movement demand (passenger) clusters) are generated based on the provisional route through the partial transportation network.
　なお、各クラスタについて算出された代表値は、輸送サービス割り当て対象の移動需要の特徴として利用される。すなわち、後の需給最適化の中で、移動需要は群（クラスタ）としてみなされ、故に、代表値を用いて最適化問題が解かれる。例えば、クラストの代表値を基に、当該クラスタに輸送サービスとしてバス又は鉄道が適しているといった判断が可能である。 　The representative value calculated for each cluster is used as a feature of the movement demand to which the transport service is allocated. That is, in the subsequent supply and demand optimization, the mobile demands are considered as clusters, and thus the representative values are used to solve the optimization problem. For example, based on the representative value of the cluster, it is possible to determine that a bus or railroad is suitable as a transport service for the cluster.
　＜図６のＳ１３：需給最適化処理＞ <S13 in Fig. 6: supply and demand optimization processing>
　図１５は、需給最適化処理（図６のＳ１３）の流れを示す。 FIG. 15 shows the flow of the supply and demand optimization process (S13 in FIG. 6).
　需給最適化部１２６は、交通ネットワークＤＢ１３２及び需要クラスタリングテーブル１９０を参照し、クラスタ毎に、当該クラスタに属する移動需要（情報１９３～２００）を満たす区間を探索し、見つかった区間について、利用する交通手段の組合せ候補を列挙する（Ｓ６１）。Ｓ６１では、需給最適化部１２６は、まだ手配されていない臨時交通を含む区間も候補として探索する。臨時交通の情報（例えば、臨時交通としての輸送サービスの種類と経由するノードとを表す情報）は、例えば交通ネットワークＤＢ１３２に格納されている。 The supply and demand optimization unit 126 refers to the traffic network DB 132 and the demand clustering table 190, searches for a section that satisfies the movement demand (information 193 to 200) belonging to the cluster for each cluster, and The candidate combinations of means are listed (S61). In S61, the demand/supply optimization unit 126 also searches for sections including temporary traffic that has not yet been arranged as candidates. Information on temporary traffic (for example, information representing types of transportation services as temporary traffic and nodes to be routed through) is stored in, for example, the traffic network DB 132 .
　次に、需給最適化部１２６は、旅客カウント部１２５にクラスタ内旅客カウント処理を実施させる（Ｓ６２）。Ｓ６２では、旅客カウント部１２５は、区間ごとに受け入れてくれる旅客が異なるという仮定の元、クラスタ及び区間の組毎に、当該クラスタのうち当該区間を通る旅客の割合である期待値を計算する。 Next, the supply and demand optimization unit 126 causes the passenger count unit 125 to perform intra-cluster passenger count processing (S62). In S62, the passenger count unit 125 calculates an expected value, which is the ratio of passengers passing through the section in the cluster, for each set of cluster and section, assuming that different passengers are accepted for each section.
　次に、需給最適化部１２６は、各クラスタがどの区間を用いるか、区間を組み換えながら、移動需要（要求サービスレベル）と運用効率を満たすような多目的最適化問題を解き、算出された一つ以上の解候補（解）のうちの一つを選択する（Ｓ６３）。一つの解候補は、例えば、図１９に示す通りである。 Next, the supply and demand optimization unit 126 solves a multi-objective optimization problem that satisfies the movement demand (required service level) and operational efficiency while recombining the sections to determine which section each cluster uses, and the calculated one One of the solution candidates (solutions) is selected (S63). One solution candidate is, for example, as shown in FIG.
　次に、需給最適化部１２６は、例えば交通ネットワークＤＢ１３２を基に、選択された解候補について輸送サービスの追加の手配が必要か否かを判定する（Ｓ６４）。例えば、解候補の区間経路に臨時便が含まれている場合は、追加手配が必要と判定される（つまり、Ｓ６４の判定結果が真となる）。すなわち、各クラスタの区間経路の候補には、定期便として提供される輸送サービスの他に、不定期運行の輸送サービスが存在することがあり、不定期運行の輸送サービスが存在する場合、追加手配が必要と判定される。 Next, the demand/supply optimization unit 126 determines whether or not additional transportation service arrangements are necessary for the selected candidate solution based on, for example, the transportation network DB 132 (S64). For example, if the section route of the solution candidate includes a temporary flight, it is determined that additional arrangements are necessary (that is, the determination result of S64 is true). In other words, in addition to regular transport services, there may be non-regular transport services in candidate section routes for each cluster. is deemed necessary.
　Ｓ６４の判定結果が真の場合（Ｓ６４：Ｙｅｓ）、需給最適化部１２６は、選択された解候補について追加の手配が必要な輸送サービスを、当該輸送サービスを提供し得る全ての交通事業者の交通運行管理サーバ２０に提示することを、運行手配部１２７に実施させる（Ｓ６５）。運行手配部１２７が、全ての交通事業者の交通運行管理サーバ２０に、追加の手配を依頼する。具体的には、例えば、選択された解候補に含まれる或るクラスタの区間経路に、臨時バスや臨時列車が含まれていれば（Ｓ６４：Ｙｅｓ）、Ｓ６５において、その臨時バスや臨時列車を担当する交通事業者に追加手配が依頼される。 If the determination result of S64 is true (S64: Yes), the demand/supply optimization unit 126 selects a transportation service that requires additional The operation arrangement unit 127 is caused to present the information to the traffic operation management server 20 (S65). The operation arrangement unit 127 requests additional arrangements from the traffic operation management servers 20 of all transportation companies. Specifically, for example, if a section route of a certain cluster included in the selected solution candidate includes a special bus or a special train (S64: Yes), in S65, the special bus or special train is Additional arrangements will be requested from the transportation company in charge.
　次に、需給最適化部１２６は、Ｓ６４での提示がされた全ての交通事業者が当該提示の受け入れをしたか否かを判定する（Ｓ６６）。Ｓ６６の判定結果が偽の場合（Ｓ６６：Ｎｏ）、次の解候補を選出し（Ｓ６７）、処理がＳ６５に戻る。 Next, the demand/supply optimization unit 126 determines whether or not all transportation operators presented in S64 have accepted the presentation (S66). If the determination result of S66 is false (S66: No), the next solution candidate is selected (S67), and the process returns to S65.
　Ｓ６６の判定結果が真の場合（Ｓ６６：Ｙｅｓ）、需給最適化部１２６は、選択された解候補を、決定された解とし、当該解について、経路変更通知を旅客アプリケーション部１２１に送る（Ｓ６８）。典型的には、決定された解に従う少なくとも一部の区間を、部分交通ネットワークを通る仮経路（例えば、特に、部分交通ネットワークのうちの最も多くの問題リンクを通る仮経路）を含んでいることはない。Ｓ６８では、当該仮経路に対応する旅客について、決定された解に従う区画を含んだ経路である新経路を表す経路変更通知が出力される。経路変更通知の出力は、旅客受け入れ履歴ＤＢ１３１に、経路変更通知が表す新経路に関する情報を含んだレコードを追加することを含んでよい。需給最適化部１２６は、臨時交通を手配済みとして交通ネットワークＤＢ１３２に登録する（Ｓ６９）。 If the determination result of S66 is true (S66: Yes), the supply and demand optimization unit 126 regards the selected solution candidate as the determined solution, and sends a route change notification to the passenger application unit 121 for this solution (S68 ). Typically, at least a portion of the section following the determined solution includes a provisional route through the partial traffic network (e.g., in particular a provisional route through the most problematic links of the partial traffic network). no. At S68, a route change notification representing a new route, which is a route including sections according to the determined solution, is output for the passenger corresponding to the tentative route. Outputting the route change notification may include adding to the passenger acceptance history DB 131 a record containing information about the new route represented by the route change notification. The demand/supply optimization unit 126 registers the temporary transportation as arranged in the transportation network DB 132 (S69).
　Ｓ６４の判定結果が真の場合（Ｓ６４：Ｙｅｓ）、需給最適化部１２６は、選択された解候補を、決定された解とし、当該解について、経路変更通知を旅客アプリケーション部１２１に送る（Ｓ７０）。 If the determination result of S64 is true (S64: Yes), the supply and demand optimization unit 126 regards the selected solution candidate as the determined solution, and sends a route change notification to the passenger application unit 121 for this solution (S70 ).
　図１６は、クラスタ内旅客カウント処理（Ｓ６２）の流れを示す。 FIG. 16 shows the flow of the intra-cluster passenger count process (S62).
　旅客カウント部１２５は、移動需要ＤＢ１２８及び需要クラスタリングテーブル１９０を参照し、移動需要（旅客）毎に、移動需要が属するクラスタの近傍のクラスタを列挙する（Ｓ７１）。「移動需要が属するクラスタの近傍のクラスタ」とは、例えば、移動需要が属するクラスタの代表値と一定範囲内にある代表値を持つクラスタである。旅客カウント部１２５は、移動需要（旅客）毎に、旅客受け入れモデルの入力となる特徴量を算出する（Ｓ７２）。ここでの「特徴量」は、移動需要ＤＢ１２８及び及び需要クラスタリングテーブル１９０を基に特定された特徴量であって、旅客とクラスタと区間（Ｓ６１で見つかった区間）との組である第１種の組毎の特徴量である。「旅客受け入れモデル」とは、第１種の組毎の特徴量を入力とし第１種の組毎の所属確率（図２１参照）を出力とするモデルである。モデルは、典型的には、機械学習モデル（例えばニューラルネットワーク）である。 The passenger counting unit 125 refers to the travel demand DB 128 and the demand clustering table 190, and lists clusters near the cluster to which the travel demand belongs for each travel demand (passenger) (S71). A "cluster near the cluster to which the travel demand belongs" is, for example, a cluster having a representative value within a certain range from the representative value of the cluster to which the travel demand belongs. The passenger count unit 125 calculates a feature quantity that is input to the passenger acceptance model for each movement demand (passenger) (S72). The "feature amount" here is a feature amount specified based on the travel demand DB 128 and the demand clustering table 190, and is a set of passengers, clusters, and sections (sections found in S61). is a feature quantity for each set of . The "passenger acceptance model" is a model that inputs feature values for each type 1 pair and outputs a belonging probability (see FIG. 21) for each type 1 pair. The model is typically a machine learning model (eg neural network).
　次に、旅客カウント部１２５は、Ｓ７２で算出された第１種の組毎の特徴量を旅客受け入れモデルに入力することで、第１種の組毎に所属確率（図２１参照）を算出し、第１種の組毎の所属確率を基に、クラスタと区間との組である第２種の組毎に期待値（確率）（図１８Ａ参照）を算出する（Ｓ７３）。旅客カウント部１２５は、クラスタ毎に、期待値（確率）が最も高い区間を特定し、当該クラスタのクラスタ所属数１９２（図１４参照）と期待値（確率）とを基に、旅客数（予測）を算出する（Ｓ７４）。 Next, the passenger count unit 125 inputs the feature quantity for each type 1 group calculated in S72 into the passenger acceptance model, thereby calculating the belonging probability (see FIG. 21) for each type 1 group. , an expected value (probability) (see FIG. 18A) is calculated for each type 2 pair, which is a pair of a cluster and an interval, based on the membership probability for each type 1 pair (S73). The passenger count unit 125 identifies the section with the highest expected value (probability) for each cluster, and calculates the number of passengers (predicted ) is calculated (S74).
　図１７は、旅客受け入れ履歴ＤＢ１３１の構成例を示す。 FIG. 17 shows a configuration example of the passenger acceptance history DB 131.
　旅客受け入れ履歴ＤＢ１３１は、旅客及び新経路の組毎にレコードを有する。各レコードは、レコード番号１３１ａと、クラスタ代表値との距離１３１ｂと、乗り換え回数１３１ｃと、部分交通ネットワーク外までの距離１３１ｄと、混雑率１３１ｅと、着席保障１３１ｆと、受け入れ可否１３１ｇといった情報を有する。一の旅客及び一の新経路新経路を例に取る（図１７の説明において「対象旅客」及び「対象新経路」）。 The passenger acceptance history DB 131 has a record for each pair of passenger and new route. Each record has information such as record number 131a, distance 131b to cluster representative value, number of transfers 131c, distance to outside of partial traffic network 131d, congestion rate 131e, guaranteed seating 131f, and acceptability 131g. . Take one passenger and one new route new route as an example (“target passenger” and “target new route” in the description of FIG. 17).
　レコード番号１３１ａは、対象旅客と対象新経路との組のレコードの識別番号を表す。クラスタ代表値との距離１３１ｂは、対象旅客のレコードを基に算出された特徴量と対象旅客が属するクラスタの代表値との距離を表す。乗り換え回数１３１ｃは、対象新経路における乗り換え回数を表す。部分交通ネットワーク外までの距離１３１ｄは、対象旅客の出発地又は目的地と部分交通ネットワークとの距離を表す。混雑率１３１ｅは、対象新経路における混雑率を表す。着席保障１３１ｆは、対象新経路における着席保障を表す。受け入れ可否１３１ｇは、対象新経路を対象旅客が受け入れ可としたか否かを表す（“１”が受入れ可を意味する）。 The record number 131a represents the identification number of the record of the target passenger and target new route. The distance 131b to the cluster representative value represents the distance between the characteristic value calculated based on the record of the target passenger and the representative value of the cluster to which the target passenger belongs. The number of transfers 131c represents the number of transfers on the target new route. The distance 131d to the outside of the partial transportation network represents the distance between the departure point or destination of the target passenger and the partial transportation network. The congestion rate 131e represents the congestion rate on the target new route. The seating guarantee 131f represents the seating guarantee on the target new route. The acceptability 131g indicates whether or not the target new route has been accepted by the target passenger ("1" means that the target new route is acceptable).
　旅客カウント部１２５は、旅客受け入れ履歴ＤＢ１３１に格納されたデータを基に、上述の旅客受け入れモデルを構築又は更新といった学習を行ってよい。例えば、旅客受け入れ履歴ＤＢ１３１の各レコードが、旅客、クラスタ及び区間を表す情報を含み、旅客受け入れモデルは、旅客と、クラスタと、区間と、要求サービスレベルと、受け入れ可否との組毎に、特徴量を算出し、それらの特徴量を入力とし、旅客とクラスタと区間との組である第１種の組毎の所属確率を出力としてよい。旅客カウント部１２５は、学習済の旅客受け入れモデルに対して、第１種の組毎の特徴量を入力することで、第１種の組毎の所属確率を求めることができる。 The passenger counting unit 125 may perform learning such as building or updating the passenger acceptance model described above based on the data stored in the passenger acceptance history DB 131. For example, each record of the passenger acceptance history DB 131 includes information representing passengers, clusters, and sections, and the passenger acceptance model is characterized by each set of passenger, cluster, section, required service level, and acceptability. Quantities may be calculated, their feature quantities may be used as inputs, and belonging probabilities for each set of the first type, which is a set of passengers, clusters, and sections, may be output. The passenger counting unit 125 can obtain the belonging probability for each type 1 pair by inputting the feature amount for each type 1 pair to the trained passenger acceptance model.
　図１８Ａは、期待値管理テーブル２１０の構成例を示す。図１８Ｂは、部分交通ネットワークの例を示す。 18A shows a configuration example of the expected value management table 210. FIG. FIG. 18B shows an example of a partial traffic network.
　期待値管理テーブル２１０は、第２種の組（クラスタと区間との組）毎にレコードを有する。各レコードは、クラスタ番号２１１と、区間２１２と、期待値（確率）２１３といった情報を有する。 The expected value management table 210 has a record for each type 2 set (a set of a cluster and an interval). Each record has information such as a cluster number 211 , an interval 212 and an expected value (probability) 213 .
　クラスタ番号２１１は、クラスタの識別番号を表す。区間２１２は、区間を構成する一つの以上のリンクの並びを表す。期待値（確率）２１３は、当該クラスタについて区間を通る旅客の割合を表す。 The cluster number 211 represents the identification number of the cluster. A section 212 represents a sequence of one or more links forming the section. The expected value (probability) 213 represents the ratio of passengers passing through the section for the cluster.
　本実施形態では、期待値（確率）２１３が採用される。この場合、図１９に例示の区間管理テーブル２２０が構築され得る。区間管理テーブル２２０は、区間毎にレコードを有し、各レコードが、区間２２１と、割り当てクラスタ２２２と、コスト２２３と、運賃２２４と、定員２２５と、旅客数（予測）２２６といった情報を有する。旅客数（実測）２２７と、収益２２８と、混雑率２２９といった情報は、本実施形態の期待される効果の一例の説明に用意された情報であり、区間管理テーブル２２０に含まれている必要は無い。 In this embodiment, an expected value (probability) 213 is adopted. In this case, the interval management table 220 illustrated in FIG. 19 can be constructed. The section management table 220 has a record for each section, and each record has information such as section 221 , allocation cluster 222 , cost 223 , fare 224 , passenger capacity 225 , and number of passengers (estimate) 226 . Information such as the number of passengers (measured) 227, revenue 228, and congestion rate 229 is information prepared for explaining an example of the expected effects of this embodiment, and does not need to be included in the section management table 220. None.
　区間２２１は、区間を構成するリンクの並びを表す。割り当てクラスタ２２２は、区間に割り当てられたクラスタの識別番号を表す。コスト２２３は、区間に沿った運行にかかるコストを表す。運賃２２４は、区間の運賃を表す。定員２２５は、区間に対応した輸送サービスの定員を表す。旅客数（予測）２２６は、区間を通る旅客数の予測値を表す。旅客数（実測）２２７は、区間を通る旅客数の期待される実測値を表す。収益２２８は、区間についての期待される収益を表し、典型的には、旅客数（実測）２２７と運賃２２４との積からコスト２２３を減じた値である。混雑率２２９は、区間での期待される混雑率を表し、典型的には、旅客数（実測）２２７を定員２２５で除算することにより得られた値である。なお、コスト２２３と、運賃２２４と、定員２２５といった情報は、期待値管理テーブル２１０に代えて、区間毎に交通ネットワークＤＢ１３２に含まれていてもよい。また、定員２２５としての値は、実際の定員としての値でもよいし、移動需要ＤＢ１２８に登録された旅客の数に応じて実際の定員が調整された後の値でもよい。 A section 221 represents a row of links forming the section. Assigned cluster 222 represents the identification number of the cluster assigned to the interval. Cost 223 represents the cost of traveling along the section. The fare 224 represents the fare for the section. The capacity 225 represents the capacity of the transportation service corresponding to the section. The number of passengers (forecast) 226 represents the predicted number of passengers passing through the section. The number of passengers (actual) 227 represents the expected number of passengers passing through the leg. Revenue 228 represents the expected revenue for the leg and is typically the number of passengers (actual) 227 multiplied by the fare 224 minus the cost 223 . The congestion rate 229 represents the expected congestion rate in the section and is typically a value obtained by dividing the number of passengers (measured) 227 by the capacity 225 . Information such as the cost 223 , the fare 224 , and the capacity 225 may be included in the transportation network DB 132 for each section instead of the expected value management table 210 . Also, the value as the passenger capacity 225 may be a value as an actual passenger capacity, or may be a value after the actual passenger capacity is adjusted according to the number of passengers registered in the movement demand DB 128 .
　図１８Ａ及び図１９が示す例によれば、第１種の組（旅客とクラスタと区間との組）毎に所属確率２５４が算出される。つまり、旅客がいずれのクラスタにも属し得るとして旅客がいずれのクラスタに属した場合にいずれの区間をどのぐらいの確率的で通るかが算出される。第２種の組毎に期待値（確率）２１３が採用され、区間毎に期待値（確率）２１３に応じたクラスタが割り当てられる。区間毎に、旅客数（２２６）は、当該期待値（確率）２１３と、当該区間に割り当てられたクラスタに所属する旅客の数とから算出される。結果として、最適解とされた組合せが輸送サービスの運用効率や旅客の要求サービスレベルを満たさなくなる（現実と乖離した最適解が求められてしまう）可能性が低減される。具体的には、例えば、運用効率の一要素である収益２２８が負の値（不採算を意味する値）になったり、サービスレベルの一要素である混雑率２２９が“１００％”を超える値になったりする可能性が低減される。 According to the example shown in FIGS. 18A and 19, the belonging probability 254 is calculated for each type 1 pair (a pair of passenger, cluster, and section). In other words, it is calculated what probability the passenger passes through which section when the passenger belongs to any cluster, assuming that the passenger can belong to any cluster. An expected value (probability) 213 is adopted for each type 2 pair, and a cluster corresponding to the expected value (probability) 213 is assigned to each section. For each section, the number of passengers (226) is calculated from the expected value (probability) 213 and the number of passengers belonging to the cluster assigned to the section. As a result, it is possible to reduce the possibility that the optimum solution combination will not satisfy the operational efficiency of transportation services and the service level requested by passengers (an optimum solution that deviates from reality will be sought). Specifically, for example, the profit 228, which is one element of operational efficiency, becomes a negative value (value meaning unprofitable), or the congestion rate 229, which is one element of the service level, exceeds "100%". The possibility of becoming
　一比較例によれば、クラスタと旅客との関係は固定的である。この場合、本実施形態に比べて、最適解とされた組合せが輸送サービスの運用効率やサービスレベルを満たさなくなる可能性が高い。例えば、一比較例では、図２０が示す通り、運用効率の一要素である収益が“－５００”のように悪化したり、サービスレベルの一要素である混雑率が“１２０％”のように悪化したりし得る。 According to a comparative example, the relationship between clusters and passengers is fixed. In this case, compared to the present embodiment, there is a high possibility that the optimum solution combination will not satisfy the operational efficiency and service level of the transportation service. For example, in a comparative example, as shown in FIG. 20, the revenue, which is one element of operational efficiency, deteriorates to "-500", and the congestion rate, which is one element of service level, deteriorates to "120%". can get worse.
　図１８Ａの期待値（確率）２１３の計算方法は、例えば次の通りである。 A method of calculating the expected value (probability) 213 in FIG. 18A is, for example, as follows.
　図２１に例示の所属確率管理テーブル２５０が構築される。所属確率管理テーブル２５０は、第１種の組（旅客とクラスタと区間の組）毎にレコードを有し、各レコードが、旅客番号２５１と、クラスタ番号２５２と、区間２５３と、所属確率２５４とから構成される。 A affiliation probability management table 250 illustrated in FIG. 21 is constructed. The belonging probability management table 250 has a record for each type 1 pair (a pair of a passenger, a cluster, and a section), and each record contains a passenger number 251, a cluster number 252, a section 253, and a belonging probability 254. consists of
　旅客番号２５１は、旅客の識別番号を表す。クラスタ番号２５２は、クラスタの識別番号を表す。区間２５３は、区間を構成するリンクの並びを表す。所属確率２５４は、旅客がクラスタに所属する場合に区間を受け入れる（通る）確率を表す。この確率は、旅客受け入れモデル（旅客受け入れ履歴ＤＢ１３１の学習により構築されたモデル）を基に得られた値である。 The passenger number 251 represents the identification number of the passenger. Cluster number 252 represents the identification number of the cluster. A section 253 represents a row of links forming the section. The belonging probability 254 represents the probability of accepting (passing) the segment when the passenger belongs to the cluster. This probability is a value obtained based on a passenger acceptance model (a model constructed by learning the passenger acceptance history DB 131).
　この所属確率管理テーブル２５０を基に、図１８Ａに例示した期待値管理テーブル２１０が構築される。第２種の組（クラスタと区間との組）毎に、期待値（確率）２１３は、当該第２種の組に対応した全旅客の所属確率２５４の和である。例えば、クラスタＣ１の区間Ｔ１の期待値（確率）２１３は、図２１が示す例によれば、０．８１＋０．７５＋０．９３＋・・・であり、結果として、図１８Ａが示すように、期待値（確率）２１３は、“６０”となる。 Based on this affiliation probability management table 250, the expected value management table 210 illustrated in FIG. 18A is constructed. The expected value (probability) 213 for each type 2 pair (a cluster and section pair) is the sum of the belonging probabilities 254 of all passengers corresponding to the type 2 pair. For example, the expected value (probability) 213 of the interval T1 of the cluster C1 is 0.81 + 0.75 + 0.93 + . . . according to the example shown in FIG. (Probability) 213 is "60".
　図２２は、需給マッチングの一例を模式的に示す。 Fig. 22 schematically shows an example of supply and demand matching.
　二次元直交座標系のグラフがある。縦軸（第１の軸の例）が、サービスレベル評価関数ｗ０ｆ（Ｒ，Ｓ）であり、横軸（第１の軸と直交する第２の軸の例）が、運用効率評価関数ｗ１ｇ（Ｒ，Ｓ）である。「Ｒ」は、各クラスタの区間（利用する交通手段の組み合わせを含み得る区間）である。「Ｓ」は、各クラスタの要求サービスレベルである。 I have a graph in a two-dimensional Cartesian coordinate system. The vertical axis (example of the first axis) is the service level evaluation function w 0 f (R, S), and the horizontal axis (example of the second axis orthogonal to the first axis) is the operational efficiency evaluation function. w 1 g(R,S). “R” is the section of each cluster (the section that can include a combination of transportation means to be used). "S" is the requested service level of each cluster.
　サービスレベル及び運用効率に最適な需給マッチング解を需給最適化部１２６が多目的最適化問題を解いて探索した場合、探索結果として最適解としてＲ＊ １が得られる。解Ｒ＊ １で需給マッチングが成立しない場合、次の候補として、２番目に最適な解Ｒ＊ ２が選出される。この際、パレート最適な解の候補が複数存在する場合、サービスレベルが優先されてもよい。 When the supply and demand optimization unit 126 solves the multi-objective optimization problem and searches for the optimum supply and demand matching solution for the service level and operational efficiency, R * 1 is obtained as the optimum solution as a result of the search. If supply and demand matching is not established with the solution R * 1 , the second optimal solution R * 2 is selected as the next candidate. At this time, if there are multiple Pareto-optimal solution candidates, the service level may be prioritized.
　需給最適化部１２６が行う処理の具体例は、次の通りである。 A specific example of the processing performed by the supply and demand optimization unit 126 is as follows.
　需要クラスタ生成処理において、Ｃ個のクラスタが生成されたとする。各クラスタｃのサービルレベルと移動需要を、数１とする。
Figure JPOXMLDOC01-appb-M000001
 Assume that C clusters are generated in the demand cluster generation process. The service level and travel demand of each cluster c are represented by Equation (1).
Figure JPOXMLDOC01-appb-M000001
　この場合、全てのクラスタの需要Ｓを、数２のように表現できる。
Figure JPOXMLDOC01-appb-M000002
 In this case, the demand S of all clusters can be expressed as Equation 2.
Figure JPOXMLDOC01-appb-M000002
　区間検索にかけるクラスタｃの区間候補の一覧をｒｃとする。クラスタ毎に、Ｍｃ個の区間が得られるとする。各ｒｃ，ｊには、クラスタ毎の区間が含まれる。ｒｃを、数３のように表現できる。
Figure JPOXMLDOC01-appb-M000003
 Let rc be a list of section candidates of cluster c to be subjected to section search. Suppose that M c intervals are obtained for each cluster. Each r c,j contains an interval for each cluster. r c can be expressed as in Equation 3.
Figure JPOXMLDOC01-appb-M000003
　需給最適化部１２６は、各クラスタが取るべき区間の組み合わせとして、各クラスタが選択するパスＲを生成する。Ｒを、数４のように表現することができる。
Figure JPOXMLDOC01-appb-M000004
 The supply and demand optimization unit 126 generates a path R selected by each cluster as a combination of sections to be taken by each cluster. R can be expressed as in Equation 4.
Figure JPOXMLDOC01-appb-M000004
　需給最適化部１２６は、パスＲ内の各クラスタの区間を変えることで、数５に従って最小化問題を解く。
Figure JPOXMLDOC01-appb-M000005
 The supply and demand optimization unit 126 solves the minimization problem according to Equation 5 by changing the section of each cluster in the path R.
Figure JPOXMLDOC01-appb-M000005
　ｆｉは、サービスレベルを満たしていない場合に増加するコスト関数である。例えば、人流に関するシミュレーションの結果からサービスレベルの基準を超えた混雑率の区間数を設定することに加えて、事業者の運行コストも設定する。なお、ｗｉｆｉで構成される最小化問題は線形である。ｗｉｆｉを状況に応じて追加、削除してもよいし、ｗｉｆｉを各々改良してもよい（シミュレーションの精度向上、重みの更新など）。また、Ｒ＊は、上位３つ程度の候補を出しておいてもよい。この際、Ｒ＊に新たな手配が必要な交通（臨時便やデマンド交通）を抽出し、抽出した交通の対象事業者に手配を通知することができる。また、対象事業者全ての承認が得られない場合は、次のＲ＊の候補の処理に移動することができる。 f i is a cost function that increases if the service level is not met. For example, in addition to setting the number of sections with a congestion rate exceeding the service level standard based on the result of a simulation on the flow of people, the operator's operating cost is also set. Note that the minimization problem composed of w i f i is linear. Depending on the situation, w i fi may be added or deleted, and each w i fi may be improved (simulation accuracy improvement, weight update, etc.). In addition, about the top three candidates for R * may be presented. At this time, it is possible to extract traffic (temporary flights and demand traffic) requiring new arrangements for R * , and notify the target operators of the extracted traffic of the arrangements. Also, if the approval of all the target business operators cannot be obtained, it is possible to move to the processing of the next candidate for R * .
　運行手配部１２７は、各交通運行管理サーバ２０と情報の送受信を行い、各交通事業者が提供する輸送サービスに用いる各交通手段の運行を手配する処理を実行する。 The operation arrangement unit 127 transmits and receives information to and from each traffic operation management server 20, and executes processing for arranging operation of each means of transportation used for transportation services provided by each transportation operator.
　以上の実施形態の具体例として、下記が考えられる。 The following is conceivable as a specific example of the above embodiment.
　第１の例として、通学需要とイベント需要とが重なる例が考えられる。この場合、遅刻できない学生のクラスタと、イベントに早く行きたい参加者（学生）のクラスタと、イベント会場まで座っていきたい参加者（学生）のクラスタが生成される。移動需要のサービスレベルに応じて、学校に向けにデマンドバスを追加手配したり、学校向けにスクールバスのイベント会場への延長運行を手配したり、イベント会場へのタクシーを手配したりすることができる。また、タクシー、一部のデマンドバス等ノードが一定でない場合、経路に属する近隣のバス停等、既存のノードを活用したり、一定距離ごとに仮想的な停留所をノードに設定したりすることができる。交通手段としてシェアサイクルを対象とする場合、自転車置き場をノードに設定することができる。 As a first example, an example where demand for commuting to school and demand for events overlap can be considered. In this case, a cluster of students who cannot be late, a cluster of participants (students) who want to go to the event early, and a cluster of participants (students) who want to sit down until the event venue are generated. Depending on the service level of transportation demand, we can make additional arrangements for demand buses for schools, arrange for extended school bus services to event venues for schools, and arrange taxis to event venues. can. In addition, when nodes such as taxis and some demand buses are not fixed, existing nodes such as nearby bus stops belonging to the route can be used, and virtual stops can be set as nodes at fixed distance intervals. . When sharing bicycles as a means of transportation, a bicycle parking lot can be set as a node.
　第２の例として、都市の鉄道などで輸送障害が発生した例が考えられる。この場合、着席を優先する買い物客のクラスタと、時間に或る程度余裕がある観光客のクラスタと、時間厳守のビジネス客のクラスタとが生成される。移動需要のサービスレベルに応じて、臨時バスの運行を手配したり、新交通システムの増便を手配したりすることができる。 As a second example, we can think of an example of a transport disruption occurring on an urban railroad. In this case, a cluster of shoppers preferring to be seated, a cluster of tourists with some extra time, and a cluster of punctual business travelers are generated. Depending on the service level of travel demand, it is possible to arrange for special bus services or to increase the number of new transportation systems.
　以上、一実施形態を説明したが、これは本発明の説明のための例示であって、本発明の範囲をこの実施形態にのみ限定する趣旨ではない。本発明は、他の種々の形態でも実施することが可能である。 Although one embodiment has been described above, this is an example for explaining the present invention, and is not intended to limit the scope of the present invention only to this embodiment. The present invention can also be implemented in various other forms.
　例えば、上述の説明を、以下のように総括することができる。以下の総括は、上述の説明の補足説明及び変形例の説明を含んでもよい。 For example, the above explanation can be summarized as follows. The following summary may include additional explanations and explanations of variations of the above discussion.
　旅客アプリケーション部１２１が、対象旅客（いずれかの旅客）の対象旅客端末３０（対象旅客の旅客端末）から移動需要を受け付け、当該移動需要を表す情報を移動需要ＤＢ１２８（移動需要情報の一例）に含める。部分ネットワーク抽出部１２２が、移動需要ＤＢ１２８及び交通ネットワークＤＢ１３２（交通ネットワーク情報の一例）を基に、それぞれが運用効率と要求サービスレベルの少なくとも一つが満たされないリンクである一つ以上の問題リンクを特定し、当該一つ以上の問題リンクを含む部分交通ネットワークを抽出する。需要クラスタリング部１２４が、仮経路ＤＢ１２９（旅客毎の仮経路を表す情報である仮経路情報の一例）及び移動需要ＤＢ１２８を基に、部分交通ネットワークを通る仮経路に対応した旅客の移動需要のクラスタを生成する。需給最適化部１２６が、生成された各クラスタと交通ネットワークＤＢ１３２とを基に、移動需要を満たし運用効率を維持する解を決定する需給最適化を行う。決定された解は、部分交通ネットワークにおける区間毎のクラスタ割当てである。決定された解が、対象旅客について仮経路に含まれない区間を含む場合、旅客アプリケーション部１２１が、対象旅客の仮経路と、当該仮経路に含まれない区間を含んだ新経路とのいずれかを確定経路とし、対象旅客端末３０からの移動需要に対する応答として、対象旅客の確定経路に関する情報を対象旅客端末３０に送信する。 The passenger application unit 121 receives travel demand from the target passenger terminal 30 (passenger terminal of the target passenger) of the target passenger (one of the passengers), and stores information representing the travel demand in the travel demand DB 128 (an example of travel demand information). include. The partial network extraction unit 122 identifies one or more problem links that are links that do not meet at least one of operational efficiency and required service level, based on the movement demand DB 128 and the transportation network DB 132 (an example of transportation network information). and extract a partial traffic network containing the one or more problematic links. Based on the temporary route DB 129 (an example of temporary route information representing a temporary route for each passenger) and the movement demand DB 128, the demand clustering unit 124 clusters passenger movement demand corresponding to the temporary route through the partial transportation network. to generate The demand/supply optimization unit 126 performs demand/supply optimization to determine a solution that satisfies travel demand and maintains operational efficiency, based on the generated clusters and the transportation network DB 132 . The determined solution is the cluster assignment for each leg in the partial traffic network. If the determined solution includes a section not included in the provisional route for the target passenger, the passenger application unit 121 selects either the provisional route for the target passenger or a new route that includes the section not included in the provisional route. is the definite route, and information on the definite route of the target passenger is transmitted to the target passenger terminal 30 as a response to the movement demand from the target passenger terminal 30 .
　これにより、旅客の多種多様な移動需要を満たすことと輸送サービスの運用効率を維持することとを両立した需給マッチングを実現することができる。また、移動需要を個々に満たそうとすると情報処理としてタクシーや自転車など小さなモビリティ（交通手段）が割り当てられがちになるおそれがあるが、移動需要のクラスタを基に解を出すため、旅客数が比較的多い（効率の良い）バスや鉄道などの大規模な交通手段の割り当てが可能となる。これは、輸送サービス（交通手段）の割り当て特有の技術的効果の一つである。 As a result, it is possible to realize demand-supply matching that satisfies the diverse travel demands of passengers and maintains the operational efficiency of transportation services. In addition, if we try to meet travel demand individually, there is a risk that information processing will tend to allocate small forms of mobility (means of transportation) such as taxis and bicycles. It becomes possible to allocate large-scale means of transportation such as relatively many (highly efficient) buses and railroads. This is one of the technical effects inherent in the allocation of transport services (means of transportation).
　抽出された部分交通ネットワークは、端点のノードを折り返し可能とする輸送サービスの数又は割合が所定値以下であるネットワークである。このため、部分交通ネットワークにおける区間（例えば問題リンクを含む区間）について単位時間当たりの運行数に変更があっても、折り返し可能な端点のノードでの車両の折り返しが可能である。結果として、部分交通ネットワークの外に問題リンクの影響を及ぼすことを無くす又は小さくすることができる。これは、輸送サービス（交通手段）の割り当て特有の技術的効果の一つである。 The extracted partial transportation network is a network in which the number or ratio of transportation services that allow return nodes at the end points is less than or equal to a predetermined value. Therefore, even if there is a change in the number of operations per unit time in the section (for example, the section containing the problematic link) in the partial traffic network, it is possible for the vehicle to turn back at the end point node where it is possible to turn back. As a result, the impact of the problem link outside the partial traffic network can be eliminated or reduced. This is one of the technical effects inherent in the allocation of transport services (means of transport).
　部分ネットワーク抽出部１２２が、問題リンクに隣接するリンクを問題リンクにつなぐことを、端点のノードを折り返し可能とする輸送サービスの数又は割合が所定値以下であることが満たされる範囲で行うことで、一つ以上の部分交通ネットワークを生成してよい。部分ネットワーク抽出部１２２が、互いに少なくとも一部が重複する二つ以上の部分交通ネットワークが生成された場合、当該二つ以上の部分交通ネットワークを一つの部分交通ネットワークとしてよい。これにより、互いに少なくとも一部が重複するにも関わらず二つ以上の部分交通ネットワークのそれぞれの最適化問題の解の整合性が取れないといったことが生じ得ることを避けることができる。 The partial network extraction unit 122 connects the link adjacent to the problem link to the problem link within a range where the number or ratio of transport services that can return the end node node is a predetermined value or less. , may generate one or more partial traffic networks. When the partial network extraction unit 122 generates two or more partial traffic networks that at least partially overlap each other, the two or more partial traffic networks may be regarded as one partial traffic network. As a result, it is possible to avoid the possibility that the solutions of the optimization problems of two or more partial traffic networks do not match even though they overlap each other at least partially.
　抽出された部分交通ネットワークは、問題リンクを最も多く通る仮経路である該当仮経路の代替パス数が下限値以上であるネットワークでよい。代替パスは、該当仮経路における問題リンク毎に、当該問題リンクを迂回しそれぞれ問題リンクではない一つ以上のリンクでよい。部分ネットワーク抽出部１２２は、生成された部分交通ネットワークにおける問題リンクを最も多く通る仮経路の代替パス数が前記下限値未満の場合、端点のノードを折り返し可能とする輸送サービスの数又は割合が所定値以下であることが満たされる範囲で当該部分交通ネットワークを拡大してよい。これにより、問題最適化のための代替パスが設定できないことを防ぐことができる。 The extracted partial traffic network may be a network in which the number of alternative paths for the relevant temporary route, which is the temporary route that passes through the problem links the most, is equal to or greater than the lower limit. The alternative path may be one or more links that bypass the problem link and are not the problem link for each problem link in the relevant tentative route. When the number of alternative paths of the provisional route that passes the most problem links in the generated partial traffic network is less than the lower limit, the partial network extraction unit 122 determines the number or ratio of transportation services that can return the node at the end point. The partial traffic network may be expanded to the extent that it is equal to or less than the value. This prevents the inability to set alternative paths for problem optimization.
　抽出された部分交通ネットワークは、問題リンクを最も多く通る仮経路である該当仮経路の代替パス数が上限値以下であるネットワークでよい。部分ネットワーク抽出部１２２は、生成された部分交通ネットワークにおける問題リンクを最も多く通る仮経路の代替パス数が上限値を超えている場合、下記が満たされているか否かを判定し、当該判定の結果が真の場合に、生成された部分交通ネットワークを、該当中間ノードを境に分割してよい。これにより、部分交通ネットワークの大きさを制限し、以って、需給最適化において組合せ爆発を防ぐことができる。
・当該部分交通ネットワークに、端点のノード以外のノードであって、折り返し可能とする輸送サービスの数又は割合が所定値以下であるノードである該当中間ノードがある。
・当該該当中間のノードを境に当該部分交通ネットワークを複数の部分交通ネットワークに分割したとしても当該複数の部分交通ネットワークがそれぞれ該当仮経路の代替パス数が下限値以上であることを満たす。 The extracted partial traffic network may be a network in which the number of alternative paths of the relevant provisional route, which is the provisional route passing through the problem links the most, is equal to or less than the upper limit value. When the number of alternative paths of the provisional route that passes the most problematic links in the generated partial traffic network exceeds the upper limit, the partial network extraction unit 122 determines whether the following conditions are satisfied, and If the result is true, the generated partial traffic network may be divided along the corresponding intermediate node. This limits the size of the partial traffic network and thus prevents combinatorial explosion in supply and demand optimization.
- The partial transportation network has an intermediate node, which is a node other than the end point node and the number or ratio of transport services that can be returned is less than or equal to a predetermined value.
- Even if the partial traffic network is divided into a plurality of partial traffic networks with the relevant intermediate node as a boundary, each of the plurality of partial traffic networks satisfies that the number of alternative paths for the corresponding temporary route is equal to or greater than the lower limit value.
　需給最適化処理において、旅客カウント部１２５は、第１種の組（旅客と、クラスタと、部分交通ネットワークにおける区間との組）毎に、旅客がクラスタに所属する場合に区間を通る確率である所属確率を算出してよい。クラスタと、前記抽出された部分交通ネットワークにおける区間との組である第２種の組毎に、当該第２種の組について得られた全旅客の所属確率を基に、当該クラスタのうち当該区間を通る旅客の割合である期待値を算出してよい。需要最適化部１２６は、第２種の組毎の期待値を基に、解を決定してよい。すなわち、旅客が生成されたクラスタに必ず所属するとは限らず別のクラスタに所属する可能性を基に、第１種の組毎の所属確率が算出され、第１種の組毎の所属確率を基に、第２種の組毎の期待値が算出される。これにより、最適化問題の解が適切である可能性の向上が期待される。 In the demand-supply optimization process, the passenger counting unit 125 calculates the probability that a passenger will pass through a section for each type 1 group (a group of a passenger, a cluster, and a section in the partial transportation network) when the passenger belongs to the cluster. Membership probabilities may be calculated. For each type 2 set, which is a set of a cluster and a section in the extracted partial transportation network, based on the belonging probability of all passengers obtained for the type 2 set, the section in the cluster An expected value, which is the percentage of passengers passing through The demand optimization unit 126 may determine the solution based on the expected value for each type 2 set. That is, based on the possibility that the passenger does not necessarily belong to the generated cluster and belongs to another cluster, the probability of belonging to each type 1 group is calculated, and the probability of belonging to each type 1 group is calculated. Based on this, the expected value for each type 2 pair is calculated. This is expected to improve the probability that the solution of the optimization problem is appropriate.
　旅客アプリケーション部１２１が、新経路（対象旅客が属するクラスタに割り当てられ対象旅客の仮経路に含まれていない上記決定された解に従う区間を含んだ経路）を対象旅客端末３０に通知し、当該新経路に関する情報を旅客受け入れ履歴ＤＢ１３１（旅客受け入れ履歴情報の一例）に含めてよい。旅客アプリケーション部１２１が、新経路の受け入れ可否の回答を対象旅客端末３０から受け、当該回答を、上記通知された新経路について旅客受け入れ履歴ＤＢ１３１に含めてよい。第１種の組毎の所属確率の算出は、旅客受け入れ履歴ＤＢ１３１を基に学習された機械学習モデルを用いて行われてよい。旅客受け入れ履歴ＤＢ１３１は、新経路に関する情報や新経路が受け入れられたか否かを表す情報を含む。すなわち、旅客が属するクラスタが割り当てられた区間を含んだ新経路が旅客に提示されてもその経路が旅客に拒否された場合には拒否がされたことを表す情報が旅客受け入れ履歴ＤＢ１３１に含まれる。言い換えれば、新経路を受け入れるか否かは旅客の裁量に委ねられており、その旅客の裁量の結果が旅客受け入れ履歴ＤＢ１３１に反映される。このような旅客受け入れ履歴ＤＢ１３１を基に学習された機械学習モデルを基に、第１種の組毎の所属確率が算出される。これにより、最適化問題の解が適切である可能性の向上が期待される。 The passenger application unit 121 notifies the target passenger terminal 30 of the new route (route including the section that is assigned to the cluster to which the target passenger belongs and is not included in the tentative route of the target passenger and that follows the above-determined solution). Information about the route may be included in the passenger acceptance history DB 131 (an example of passenger acceptance history information). The passenger application unit 121 may receive a response as to whether or not the new route can be accepted from the target passenger terminal 30, and include the response in the passenger acceptance history DB 131 for the notified new route. The calculation of the belonging probability for each type 1 group may be performed using a machine learning model learned based on the passenger acceptance history DB 131 . The passenger acceptance history DB 131 includes information about new routes and information indicating whether or not the new routes have been accepted. In other words, if a passenger is presented with a new route that includes a section to which the cluster to which the passenger belongs is rejected, the passenger acceptance history DB 131 contains information indicating that the route has been rejected. . In other words, whether or not to accept the new route is left to the passenger's discretion, and the result of the passenger's discretion is reflected in the passenger acceptance history DB 131 . Based on the machine learning model learned based on such passenger acceptance history DB 131, the belonging probability for each type 1 group is calculated. This is expected to improve the probability that the solution of the optimization problem is appropriate.
　需要最適化部１２６は、第２種の組毎に予測された旅客数と、交通ネットワーク情報とを基に、一つ以上の解を算出し、当該一つ以上の解のうち未選択の一つの解を選択し、交通ネットワーク情報を基に、選択された解が、輸送サービスの追加の手配を必要とする解か否かを判定してよい。この判定の結果が偽の場合、需要最適化部１２６は、選択された解を、決定された解としてよい。一方、この判定の結果が真の場合、運行手配部１２７は、追加の手配が必要な輸送サービス毎に当該輸送サービスに対応した一つ以上の交通運行管理サーバに追加の手配を依頼し、依頼がされた全ての交通運行管理サーバから承認の回答を受けた場合、需要最適化部１２６は、選択された解を、決定された解としてよい。これにより、新経路の適切な通知（適切な経路変更通知）を出すことができる。 The demand optimization unit 126 calculates one or more solutions based on the number of passengers predicted for each type 2 group and the traffic network information, and selects one unselected one of the one or more solutions One solution may be selected, and based on transportation network information, it may be determined whether the selected solution is one that requires additional arrangements for transportation services. If the result of this determination is false, the demand optimization unit 126 may set the selected solution as the determined solution. On the other hand, if the result of this determination is true, the operation arrangement unit 127 requests additional arrangements to one or more traffic operation management servers corresponding to each transportation service for which additional arrangements are required. When receiving approval responses from all the traffic management servers to which the request has been made, the demand optimization unit 126 may set the selected solution as the determined solution. This makes it possible to issue an appropriate notification of the new route (appropriate route change notification).
１０：需給マッチングサーバ 10: Supply and demand matching server

Claims (12)

  1. 　複数の旅客端末と通信するインターフェース装置と、
    　旅客毎の移動需要を表す情報である移動需要情報と、旅客毎の仮経路を表す情報である仮経路情報と、交通ネットワークを表す情報である交通ネットワーク情報とを記憶する記憶装置と、
    　前記インターフェース装置及び記憶装置に接続されたプロセッサと
    を備え、
    　旅客端末は、旅客の情報処理端末であり、
    　移動需要は、出発地と、目的地と、旅客の輸送に関する一つ又は複数の項目の各々について要求するレベルである要求サービスレベルとを含み、
    　仮経路は、出発地から目的地までの仮の経路であり、
    　前記交通ネットワークは、複数のノードと複数のリンクとで構成されたグラフ構造のネットワークであり、
    　ノードは、旅客が乗車及び降車のうちの少なくとも一つを行い得る場所に対応し、
    　リンクは、輸送サービスに対応し、
    　同一のノード間に、一つ以上の輸送サービスに対応した一つ以上のリンクが存在し、
    　前記交通ネットワーク情報は、輸送サービス毎に、輸送サービスの定員と輸送サービスのサービスレベルに関する情報とを含み、
    　前記プロセッサは、
    　　（Ａ）対象旅客の対象旅客端末から移動需要を受け付け、当該移動需要を表す情報を前記移動需要情報に含め、
    　　　　前記対象旅客は、いずれかの旅客であり、
    　　　　前記対象旅客端末は、前記対象旅客の旅客端末であり、
    　　（Ｂ）前記移動需要情報及び前記交通ネットワーク情報を基に、それぞれが運用効率と要求サービスレベルの少なくとも一つが満たされないリンクである一つ以上の問題リンクを特定し、
    　　（Ｃ）当該一つ以上の問題リンクを含み前記交通ネットワークの一部分としてのグラフである部分交通ネットワークを抽出し、
    　　（Ｄ）前記仮経路情報及び前記移動需要情報を基に、前記部分交通ネットワークを通る仮経路に対応した旅客の移動需要のクラスタを生成し、
    　　（Ｅ）生成された各クラスタと前記交通ネットワーク情報とを基に、移動需要を満たし運用効率を維持する解を決定する需給最適化を行い、
    　　　　前記決定された解は、前記部分交通ネットワークにおける区間毎のクラスタ割当てであり、
    　　　　前記部分交通ネットワークにおいて、区間は、ノード間が異なる一つ以上のリンクであり、
    　　（Ｆ）前記決定された解が、前記対象旅客について仮経路に含まれない区間を含む場合、前記対象旅客の仮経路と、当該仮経路に含まれない区間を含んだ新経路とのいずれかを確定経路とし、
    　　（Ｇ）前記対象旅客端末からの移動需要に対する応答として、前記対象旅客の確定経路に関する情報を前記対象旅客端末に送信する、
    需給マッチング装置。     an interface device that communicates with a plurality of passenger terminals;
    a storage device for storing movement demand information that is information representing movement demand for each passenger, temporary route information that is information representing a temporary route for each passenger, and transportation network information that is information representing a transportation network;
    a processor connected to the interface device and the storage device;
    The passenger terminal is a passenger information processing terminal,
    the travel demand includes a point of origin, a destination and a requested service level, which is the level requested for each of the one or more items relating to the carriage of the passenger;
    The provisional route is a provisional route from the departure point to the destination,
    The traffic network is a graph-structured network composed of a plurality of nodes and a plurality of links,
    a node corresponds to a location where passengers may board and/or disembark;
    The link corresponds to the transport service,
    There are one or more links corresponding to one or more transport services between the same nodes,
    The transportation network information includes information on the capacity of the transportation service and the service level of the transportation service for each transportation service,
    The processor
    (A) Receive travel demand from the target passenger terminal of the target passenger, include information representing the travel demand in the travel demand information,
    Said Eligible Passenger is any Passenger,
    The target passenger terminal is the passenger terminal of the target passenger,
    (B) based on the movement demand information and the transportation network information, identifying one or more problematic links, each link failing to satisfy at least one of operational efficiency and required service level;
    (C) extracting a partial traffic network that includes the one or more problem links and is a graph as a part of the traffic network;
    (D) based on the temporary route information and the movement demand information, generate clusters of passenger movement demand corresponding to the temporary route through the partial transportation network;
    (E) based on the generated clusters and the traffic network information, supply and demand optimization is performed to determine a solution that satisfies travel demand and maintains operational efficiency;
    the determined solution is a cluster assignment for each leg in the partial traffic network;
    In the partial traffic network, a section is one or more links with different nodes,
    (F) If the determined solution includes a section not included in the provisional route for the target passenger, either the provisional route for the target passenger or a new route including the section not included in the provisional route. is the definite path, and
    (G) transmitting information regarding the confirmed route of the target passenger to the target passenger terminal in response to a travel demand from the target passenger terminal;
    Supply and demand matching device.
  2. 　前記抽出された部分交通ネットワークは、端点のノードを折り返し可能とする輸送サービスの数又は割合が所定値以下であるネットワークである、
    請求項１に記載の需給マッチング装置。     The extracted partial transportation network is a network in which the number or ratio of transportation services that allow returning end nodes is a predetermined value or less.
    The supply and demand matching device according to claim 1.
  3. 　（Ｃ）において、前記プロセッサは、
    　　問題リンクに隣接するリンクを問題リンクにつなぐことを、端点のノードを折り返し可能とする輸送サービスの数又は割合が所定値以下であることが満たされる範囲で行うことで、一つ以上の部分交通ネットワークを生成し、
    　　互いに少なくとも一部が重複する二つ以上の部分交通ネットワークが生成された場合、当該二つ以上の部分交通ネットワークを一つの部分交通ネットワークとする、
    請求項２に記載の需給マッチング装置。     In (C), the processor
    One or more partial traffics are generated by connecting links adjacent to the problem link to the problem link within a range where the number or ratio of transport services that allow return nodes at the end points to be returned is equal to or less than a predetermined value. generate a network,
    When two or more partial traffic networks that at least partially overlap with each other are generated, the two or more partial traffic networks are regarded as one partial traffic network;
    The supply and demand matching device according to claim 2.
  4. 　前記抽出された部分交通ネットワークは、問題リンクを最も多く通る仮経路である該当仮経路の代替パス数が下限値以上であるネットワークであり、
    　前記代替パスは、前記該当仮経路における問題リンク毎に、当該問題リンクを迂回しそれぞれ問題リンクではない一つ以上のリンクである、
    請求項２に記載の需給マッチング装置。     The extracted partial traffic network is a network in which the number of alternative paths of the relevant temporary route, which is the temporary route that passes through the problem links the most, is equal to or greater than the lower limit;
    The alternative path is one or more links that bypass the problematic link and are not the problematic link, for each problematic link in the relevant temporary route.
    The supply and demand matching device according to claim 2.
  5. 　（Ｃ）において、前記プロセッサは、生成された部分交通ネットワークにおける問題リンクを最も多く通る仮経路の代替パス数が前記下限値未満の場合、端点のノードを折り返し可能とする輸送サービスの数又は割合が所定値以下であることが満たされる範囲で当該部分交通ネットワークを拡大する、
    請求項４に記載の需給マッチング装置。     In (C), if the number of alternative paths of the provisional route that passes the most problematic links in the generated partial traffic network is less than the lower limit, the processor determines the number or ratio of transport services that can turn back the end node. expands the partial traffic network to the extent that is less than or equal to a predetermined value;
    The supply and demand matching device according to claim 4.
  6. 　前記抽出された部分交通ネットワークは、問題リンクを最も多く通る仮経路である該当仮経路の代替パス数が上限値以下であるネットワークであり、
    　前記代替パスは、前記該当仮経路における問題リンク毎に、当該問題リンクを迂回しそれぞれ問題リンクではない一つ以上のリンクである、
    請求項２に記載の需給マッチング装置。     The extracted partial traffic network is a network in which the number of alternative paths of the relevant temporary route, which is the temporary route that passes through the problem links the most, is equal to or less than the upper limit;
    The alternative path is one or more links that bypass the problematic link and are not the problematic link, for each problematic link in the relevant temporary route.
    The supply and demand matching device according to claim 2.
  7. 　（Ｃ）において、前記プロセッサは、
    　　生成された部分交通ネットワークにおける問題リンクを最も多く通る仮経路の代替パス数が前記上限値を超えている場合、下記が満たされているか否かを判定し、
    　　　　・当該部分交通ネットワークに、端点のノード以外のノードであって、折り返し可能とする輸送サービスの数又は割合が所定値以下であるノードである該当中間ノードがある、
    　　　　・当該該当中間のノードを境に当該部分交通ネットワークを複数の部分交通ネットワークに分割したとしても当該複数の部分交通ネットワークがそれぞれ該当仮経路の代替パス数が下限値以上であることを満たす、
    　　前記判定の結果が真の場合に、前記生成された部分交通ネットワークを、前記該当中間ノードを境に分割する、
    請求項２に記載の需給マッチング装置。     In (C), the processor
    If the number of alternative paths of the provisional route that passes the most problematic links in the generated partial traffic network exceeds the upper limit value, determining whether the following is satisfied;
    - The partial transportation network has a corresponding intermediate node which is a node other than the terminal node and for which the number or ratio of transport services that can be turned back is a predetermined value or less.
    ・Even if the partial transportation network is divided into a plurality of partial transportation networks with the relevant intermediate node as a boundary, each of the plurality of partial transportation networks satisfies that the number of alternative paths for the corresponding provisional route is equal to or greater than the lower limit value.
    If the result of the determination is true, dividing the generated partial traffic network along the corresponding intermediate node as a boundary;
    The supply and demand matching device according to claim 2.
  8. 　（Ｅ）において、前記プロセッサは、
    　　（ｅ１）旅客と、クラスタと、前記抽出された部分交通ネットワークにおける区間との組である第１種の組毎に、当該旅客が当該クラスタに所属する場合に当該区間を通る確率である所属確率を算出し、
    　　（ｅ２）クラスタと、前記抽出された部分交通ネットワークにおける区間との組である第２種の組毎に、当該第２種の組について得られた全旅客の所属確率を基に、当該クラスタのうち当該区間を通る旅客の割合である期待値を算出し、
    　　（ｅ３）第２種の組毎の期待値を基に、解を決定する、
    請求項１に記載の需給マッチング装置。     In (E), the processor
    (e1) Belonging probability, which is the probability that the passenger passes through the section when the passenger belongs to the cluster, for each set of the first type, which is a set of a passenger, a cluster, and a section in the extracted partial transportation network. to calculate
    (e2) For each type 2 set, which is a set of a cluster and a section in the extracted partial transportation network, based on the probability of all passengers belonging to the type 2 set, the cluster Calculate the expected value, which is the ratio of passengers passing through the section,
    (e3) determining the solution based on the expected value for each set of the second type;
    The supply and demand matching device according to claim 1.
  9. 　前記記憶装置が、経路に関する情報と旅客による受け入れ可否との履歴を表す情報である旅客受け入れ履歴情報を記憶し、
    　前記プロセッサが、
    　　前記対象旅客が属するクラスタに割り当てられ前記対象旅客の仮経路に含まれていない前記決定された解に従う区間を含んだ経路である新経路を、当該新経路を前記対象旅客端末に通知し、
    　　当該新経路に関する情報を前記旅客受け入れ履歴情報に含め、
    　　前記新経路の受け入れ可否の回答を前記対象旅客端末から受け、当該回答を、前記通知された新経路について前記旅客受け入れ履歴情報に含め、
    　（ｅ１）の算出は、前記旅客受け入れ履歴情報を基に学習された機械学習モデルを用いて行われる、
    請求項８に記載の需給マッチング装置。     The storage device stores passenger acceptance history information, which is information representing a history of information on routes and acceptance or rejection of acceptance by passengers,
    the processor
    Notifying the target passenger terminal of a new route that is a route including a section according to the determined solution that is assigned to the cluster to which the target passenger belongs and that is not included in the tentative route of the target passenger;
    Include information about the new route in the passenger acceptance history information,
    receiving a response from the target passenger terminal as to whether or not the new route can be accepted, and including the response in the passenger acceptance history information for the notified new route;
    The calculation of (e1) is performed using a machine learning model learned based on the passenger acceptance history information,
    The supply and demand matching device according to claim 8.
  10. 　（ｅ３）において、前記プロセッサは、
    　　（ｅ３ａ）第２種の組毎に予測された旅客数と、前記交通ネットワーク情報とを基に、一つ以上の解を算出し、
    　　（ｅ３ｂ）当該一つ以上の解のうち未選択の一つの解を選択し、
    　　（ｅ３ｃ）前記交通ネットワーク情報を基に、（ｅ３ｂ）で選択された解が、輸送サービスの追加の手配を必要とする解か否かを判定し、
    　　（ｅ３ｄ）（ｅ３ｃ）の判定の結果が偽の場合、（ｅ３ｂ）で選択された解を、決定された解とする、
    請求項８に記載の需給マッチング装置。     In (e3), the processor
    (e3a) calculating one or more solutions based on the number of passengers predicted for each type 2 set and the transportation network information;
    (e3b) selecting one unselected solution from the one or more solutions;
    (e3c) Based on the transportation network information, determine whether the solution selected in (e3b) is a solution that requires additional arrangements for transportation services;
    (e3d) if the result of the determination in (e3c) is false, let the solution selected in (e3b) be the determined solution;
    The supply and demand matching device according to claim 8.
  11. 　前記インターフェース装置は、複数の交通運行管理サーバと通信可能であり、
    　前記交通運行管理サーバは、交通事業者が提供する輸送サービスにおける交通手段の運行を管理するサーバであり、
    　（ｅ３）において、前記プロセッサは、
    　　（ｅ３ａ）第２種の組毎に予測された旅客数と、前記交通ネットワーク情報とを基に、一つ以上の解を算出し、
    　　（ｅ３ｂ）当該一つ以上の解のうち未選択の一つの解を選択し、
    　　（ｅ３ｃ）前記交通ネットワーク情報を基に、当該選択された解が、輸送サービスの追加の手配を必要とする解か否かを判定し、
    　　（ｅ３ｄ）（ｅ３ｃ）の判定の結果が真の場合、追加の手配が必要な輸送サービス毎に当該輸送サービスに対応した一つ以上の交通運行管理サーバに追加の手配を依頼し、
    　　（ｅ３ｅ）（ｅ３ｄ）で依頼がされた全ての交通運行管理サーバから承認の回答を受けた場合、（ｅ３ｂ）で選択された解を、決定された解とし、
    　　（ｅ３ｆ）（ｅ３ｄ）で依頼がされた少なくとも一つの交通運行管理サーバから承認の回答が無かったと判断した場合、（ｅ３ｂ）を行う、
    請求項８に記載の需給マッチング装置。     The interface device is capable of communicating with a plurality of traffic operation management servers,
    The traffic operation management server is a server that manages the operation of transportation means in transportation services provided by transportation operators,
    In (e3), the processor
    (e3a) calculating one or more solutions based on the number of passengers predicted for each type 2 set and the transportation network information;
    (e3b) selecting one unselected solution from the one or more solutions;
    (e3c) determining whether the selected solution requires additional arrangements for transportation services based on the transportation network information;
    (e3d) if the result of determination in (e3c) is true, for each transportation service that requires additional arrangements, request additional arrangements from one or more traffic operation management servers corresponding to the transportation service;
    (e3e) If approval responses have been received from all traffic operation management servers requested in (e3d), the solution selected in (e3b) is taken as the determined solution,
    (e3f) If it is determined that at least one traffic management server requested in (e3d) has not responded with approval, perform (e3b);
    The supply and demand matching device according to claim 8.
  12. 　（Ａ）コンピュータが、対象旅客の対象旅客端末から移動需要を受け付け、当該移動需要を表す情報を移動需要情報に含め、
    　　前記移動需要情報は、旅客毎の移動需要を表す情報であり、
    　　前記対象旅客は、いずれかの旅客であり、
    　　前記対象旅客端末は、前記対象旅客の旅客端末であり、
    　　旅客端末は、旅客の情報処理端末であり、
    　　移動需要は、出発地と、目的地と、旅客の輸送に関する一つ又は複数の項目の各々について要求するレベルである要求サービスレベルとを含み、
    　（Ｂ）コンピュータが、前記移動需要情報及び交通ネットワーク情報を基に、それぞれが運用効率と要求サービスレベルの少なくとも一つが満たされないリンクである一つ以上の問題リンクを特定し、
    　　前記交通ネットワーク情報は、交通ネットワークを表す情報であり、
    　　前記交通ネットワークは、複数のノードと複数のリンクとで構成されたグラフ構造のネットワークであり、
    　　ノードは、旅客が乗車及び降車のうちの少なくとも一つを行い得る場所に対応し、
    　　リンクは、輸送サービスに対応し、
    　　同一のノード間に、一つ以上の輸送サービスに対応した一つ以上のリンクが存在し、
    　　前記交通ネットワーク情報は、輸送サービス毎に、輸送サービスの定員と輸送サービスのサービスレベルに関する情報とを含み、
    　（Ｃ）コンピュータが、当該一つ以上の問題リンクを含み前記交通ネットワークの一部分としてのグラフである部分交通ネットワークを抽出し、
    　（Ｄ）コンピュータが、仮経路情報及び前記移動需要情報を基に、前記部分交通ネットワークを通る仮経路に対応した旅客の移動需要のクラスタを生成し、
    　　前記仮経路情報は、旅客毎の仮経路を表す情報であり、
    　　仮経路は、出発地から目的地までの仮の経路であり、
    　（Ｅ）コンピュータが、生成された各クラスタと前記交通ネットワーク情報とを基に、移動需要を満たし運用効率を維持する解を決定する需給最適化を行い、
    　　前記決定された解は、前記部分交通ネットワークにおける区間毎のクラスタ割当てであり、
    　　前記部分交通ネットワークにおいて、区間は、ノード間が異なる一つ以上のリンクであり、
    　（Ｆ）コンピュータが、前記決定された解が、前記対象旅客について仮経路に含まれない区間を含む場合、前記対象旅客の仮経路と、当該仮経路に含まれない区間を含んだ新経路とのいずれかを確定経路とし、
    　（Ｇ）コンピュータが、前記対象旅客端末からの移動需要に対する応答として、前記対象旅客の確定経路に関する情報を前記対象旅客端末に送信する、
    需給マッチング方法。     (A) The computer receives travel demand from the subject passenger terminal of the subject passenger, includes information representing the travel demand in the travel demand information,
    The travel demand information is information representing travel demand for each passenger,
    Said Eligible Passenger is any Passenger,
    The target passenger terminal is the passenger terminal of the target passenger,
    The passenger terminal is a passenger information processing terminal,
    the travel demand includes a point of origin, a destination and a requested service level, which is the level requested for each of the one or more items relating to the carriage of the passenger;
    (B) the computer identifies one or more problematic links, each link failing to meet at least one of operational efficiency and required service level, based on the movement demand information and the transportation network information;
    The transportation network information is information representing a transportation network,
    The traffic network is a graph-structured network composed of a plurality of nodes and a plurality of links,
    a node corresponds to a location where passengers may board and/or disembark;
    The link corresponds to the transport service,
    There are one or more links corresponding to one or more transport services between the same nodes,
    The transportation network information includes information on the capacity of the transportation service and the service level of the transportation service for each transportation service,
    (C) a computer extracts a partial traffic network that is a graph as part of said traffic network containing said one or more problem links;
    (D) the computer generates clusters of passenger movement demand corresponding to the temporary route through the partial transportation network based on the temporary route information and the movement demand information;
    The provisional route information is information representing a provisional route for each passenger,
    The provisional route is a provisional route from the departure point to the destination,
    (E) The computer performs supply and demand optimization to determine a solution that satisfies travel demand and maintains operational efficiency based on each generated cluster and the transportation network information;
    the determined solution is a cluster assignment for each leg in the partial traffic network;
    In the partial traffic network, a section is one or more links with different nodes,
    (F) If the determined solution includes a section not included in the provisional route for the target passenger, the computer determines the provisional route for the target passenger and a new route including the section not included in the provisional route. as a definite route,
    (G) a computer, in response to a travel demand from said target passenger terminal, transmitting information regarding said target passenger's confirmed route to said target passenger terminal;
    Supply and demand matching method.
PCT/JP2022/042536 2021-12-20 2022-11-16 Supply–demand matching device and method WO2023119965A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2021-205920 2021-12-20
JP2021205920A JP2023091265A (en) 2021-12-20 2021-12-20 Supply/demand matching apparatus and method

Publications (1)

Publication Number Publication Date
WO2023119965A1 true WO2023119965A1 (en) 2023-06-29

Family

ID=86902133

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2022/042536 WO2023119965A1 (en) 2021-12-20 2022-11-16 Supply–demand matching device and method

Country Status (2)

Country Link
JP (1) JP2023091265A (en)
WO (1) WO2023119965A1 (en)

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018097047A1 (en) * 2016-11-25 2018-05-31 日本電気株式会社 Traffic control system, traffic information output device, traffic control method, and recording medium

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018097047A1 (en) * 2016-11-25 2018-05-31 日本電気株式会社 Traffic control system, traffic information output device, traffic control method, and recording medium

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
UEHARA KAZUKI, AKAMINE YUHEI, TOMA NARUAKI, NEROME MOEKO, ENDO SATOSHI: "Proposal of cooperative transportation system with demand bus and large vehicle", TRANSACTIONS OF INFORMATION PROCESSING SOCIETY OF JAPAN, INFORMATION PROCESSING SOCIETY OF JAPAN, JP, vol. 56, no. 1, 1 January 2015 (2015-01-01), JP , pages 46 - 56, XP093073317, ISSN: 1882-7764 *

Also Published As

Publication number Publication date
JP2023091265A (en) 2023-06-30

Similar Documents

Publication Publication Date Title
US20220172175A1 (en) Transportation system and method for allocating frequencies of transit services therein
Tong et al. Customized bus service design for jointly optimizing passenger-to-vehicle assignment and vehicle routing
CN109409599B (en) Customized bus line opening optimization method based on real-time requirements
Ibarra-Rojas et al. Planning, operation, and control of bus transport systems: A literature review
Liao et al. An object-oriented evaluation framework for dynamic vehicle routing problems under real-time information
CN110162931B (en) Large-scale road network rapid simulation system for urban rail transit
Liu et al. Dynamic shared autonomous taxi system considering on-time arrival reliability
Chen et al. Customized bus route design with pickup and delivery and time windows: Model, case study and comparative analysis
Lin et al. Integrating traffic routing optimization and train formation plan using simulated annealing algorithm
Caros et al. Day-to-day market evaluation of modular autonomous vehicle fleet operations with en-route transfers
CN112561249B (en) Real-time demand-oriented city customized bus scheduling method
Dou et al. Customized bus service design for uncertain commuting travel demand
Luo et al. Optimal design of intermodal mobility networks under uncertainty: Connecting micromobility with mobility-on-demand transit
Zheng et al. Route design model of multiple feeder bus service based on existing bus lines
Wang et al. Assessing the potential of the strategic formation of urban platoons for shared automated vehicle fleets
Zhou et al. A scalable vehicle assignment and routing strategy for real-time on-demand ridesharing considering endogenous congestion
Poulhès et al. Single vehicle network versus dispatcher: user assignment in an agent-based model
Liu et al. An agent-based simulation model for shared autonomous taxi system
Elting et al. Potential of shared taxi services in rural areas–a case study
Wu et al. Customized bus scheme design of large transport terminals with jointly optimization of departure time, vehicle allocation and routing
Tao et al. Heuristic algorithms for the dynamic taxipooling problem based on intelligent transportation system technologies
Lotfi et al. Ride matching and vehicle routing for on-demand mobility services
Bobyl et al. Rail passenger hubs
WO2023119965A1 (en) Supply–demand matching device and method
Suh et al. A large neighborhood search heuristic to establish an optimal ad-hoc hubbing strategy in the wake of a large-scale airport outage

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22910684

Country of ref document: EP

Kind code of ref document: A1