JP6867314B2 - Program and passenger travel route estimation device - Google Patents

Description

The present invention relates to a passenger movement route estimation device and the like that estimate a passenger movement route.

In order to evaluate and verify train schedules, passenger flow simulations for train schedules are being conducted using computers. In the passenger flow simulation, the movement route (boarding route) of each passenger with respect to a given train schedule is estimated. In many cases, there are multiple travel routes from the station where passengers appear (departure station) to the destination station, and among these multiple routes, one that satisfies certain conditions such as the earliest arrival at the destination station and the least number of transfers. Two routes are selected. The passengers assumed in the passenger flow simulation are given by, for example, OD data in which the number of passengers appearing for each combination (OD: Origin Destination) of the appearance station and the destination station for each time zone is determined (see, for example, Patent Document 1). ).

Japanese Unexamined Patent Publication No. 2008-62729

In recent years, a method of utilizing recorded data (hereinafter referred to as "entrance / exit data") of each passenger's transit time (entrance / exit time) that can be acquired from an automatic ticket gate has been studied. As the entry / exit data, for example, the entry / exit time data of the station ticket gate in units of 1 minute can be acquired. Therefore, it is possible to accurately estimate the movement route of the passenger as compared with the case of using the OD data which is the data for each time zone of up to about 2 hours. However, when using entry / exit data, it is necessary to handle a huge amount of data such as hundreds of thousands to millions of people, and when it comes to estimating the movement route of each passenger, the amount of calculation is enormous. Therefore, it is necessary to speed up the processing.

In recent years, the time spent at station ticket gates for shopping and eating and drinking has increased due to the enhancement of so-called station Naka facilities, and the conventional conditions such as the time to reach the destination station and the number of transfers are sufficient. In some cases, it was insufficient to estimate an appropriate passenger travel route. Furthermore, when using the above-mentioned entry / exit data, the entry time and entry time of each passenger are given, so that the passengers enter the entrance station at the given entry time and exit from the target station at the given entry time. It is necessary to estimate the travel route in consideration of the station and the length of stay during the journey. However, in the conventional method, the boarding route is not estimated in consideration of the stay in the station ticket gate.

The present invention has been made in view of the above circumstances, and an object of the present invention is to enable estimation of a passenger's travel route in consideration of staying at a station ticket gate, and to enable each passenger's travel route to be estimated. The speed of estimation is to be increased.

The first invention for solving the above problems is
Computer,
Based on a given train schedule, 1) set arrival nodes and departure nodes that represent events related to arrivals and departures at each station of each train, and connect the arrival nodes and departure nodes with edges along the operation of the train. 2) Corresponding to the arrival node and the departure node of each station of each train, an entrance node and an exit node representing an event related to the entry and exit of passengers at the station, and a station stay node of the station are set. The station stay node is connected to the corresponding arrival node or the departure node at the edge, the entrance node is connected to the corresponding departure node at the edge, and the entry node is connected to the corresponding arrival node at the edge. 3) Connecting the entrance nodes, the entry nodes, and the station stay nodes that are adjacent to each other in time at each station with an edge, and 4) setting a given cost at each edge to form a diamond network. Creation method to create,
A search means for searching the minimum cost route from a given search target entry node to a given search target entry node in the diamond network based on the cost set at the edge.
It is a program to function as.

As another invention
Based on a given train schedule, 1) set arrival nodes and departure nodes that represent events related to arrivals and departures at each station of each train, and connect the arrival nodes and departure nodes with edges along the operation of the train. 2) Corresponding to the arrival node and the departure node of each station of each train, an entrance node and an exit node representing an event related to the entry and exit of passengers at the station, and a station stay node of the station are set. The station stay node is connected to the corresponding arrival node or the departure node at the edge, the entrance node is connected to the corresponding departure node at the edge, and the entry node is connected to the corresponding arrival node at the edge. 3) Connecting the entrance nodes, the entry nodes, and the station stay nodes that are adjacent to each other in time at each station with an edge, and 4) setting a given cost at each edge to form a diamond network. How to create and how to create
A search means for searching the minimum cost route from a given search target entry node to a given search target entry node in the diamond network based on the cost set at the edge.
A passenger movement route estimation device including the above may be configured.

According to the first invention or the like, as a diamond network based on a given timetable, a station stay node is set corresponding to each train arrival node and departure node of each station, and the station stay node corresponds to the arrival node. Alternatively, connect the departure node with the edge, connect the entry node with the corresponding departure node with the edge, connect the entry node with the corresponding arrival node with the edge, and connect the entry nodes that are temporally adjacent to each other at each station, the entry nodes, and the station stay. By connecting the nodes with an edge, it is possible to create a diamond network that takes into consideration the passenger's stay at the station ticket gate. Then, by searching for the minimum cost route from the given search target entry node of the created diamond network to the given search target entry node based on the cost set at the edge, the stay in the station ticket gate was considered. It is possible to estimate the travel route of passengers. For example, by setting the cost of the edge of each station according to the ease of staying at the ticket gate of the station, it is possible to estimate a route for passengers to stay at a station where it is easy to stay. ..

The second invention is a program of the first invention.
The search means
By selecting the edge that has the minimum cost by going back from the search target participation node, the minimum cost route from the node connected to the selected edge to the search target participation node is sequentially determined. Performs a route search process to search for the minimum cost route from the search target entry node to the search target entry node.
It is a program.

According to the second invention, in the search of the minimum cost route, the search target participation is performed from the node connected to the selected edge by selecting the edge having the minimum cost by tracing back from the search target participation node. It is performed by a route search process that sequentially determines the minimum cost route to the node.

The third invention is a program of the second invention.
Based on the entry / exit data in which the entry station, entry time, entry station, and entry time of each passenger are determined, the entry node (hereinafter referred to as "specific entry node") and entry node (hereinafter "specific entry node") of each passenger in the timetable network. Specific means to identify (referred to as "specific participation node"),
To make the computer function as
The search means
For passengers with the same specific participation node, the specific participation node is set as the search target participation node, and the search result of the overlapping minimum cost route is diverted to perform the route search process omitting the duplicate search. Search for the minimum cost route from each specific entry node to the specific entry node for the same passenger with the specific entry node.
It is a program.

According to the third invention, the search for the minimum cost route traced back from the search target participation node is performed by the route search process of the second invention, so that the minimum cost route for passengers having the same specific participation node is determined. Part of the route to the specific participating node will overlap. Therefore, by diverting the duplicated search results, it is possible to reduce the amount of calculation required for estimating the passenger's movement route and speed up the processing.

The fourth invention is a program of the third invention.
The search means
For passengers with the same specific entry node, the route search process is performed in the order of the earliest set time of the specific entry node.
It is a program.

According to the fourth invention, for passengers having the same specific entry node, the route search process is performed in the order of the earliest set time of the specific entry node, so that the route search process is performed in a wide range in advance, and the route is performed later. When the search process is performed, the route portion of the overlapping minimum cost route can be in a wider range, so that it is possible to further reduce the amount of calculation required for estimating the movement route of the passenger.

The fifth invention is a program of the third or fourth invention.
The search means
From the group of passengers with the same specific participation node, the minimum cost route is searched in order from the group with the latest set time of the specific participation node.
When the specific participating node related to the group to be searched this time is included as a transit node in the minimum cost route searched in the previous group, the search result searched retroactively from the transit node can be diverted. Perform the route search process omitting the duplicate search.
It is a program.

According to the fifth invention, when the specific participating node related to the group to be searched this time is included as a transit node in the minimum cost route searched by the previous group, the minimum cost route in the current search and the previous search. Since a part of the above is duplicated, it is possible to further reduce the amount of calculation required for estimating the movement route of the passenger by diverting the previous search result.

An example of a diamond network. An example of a train schedule. A diamond network representing the train schedule shown in FIG. An example of entry / exit data. An example of grouping for entry / exit data. A search node pair list corresponding to the entry / exit data of FIG. An example of searching for the lowest cost route. Functional configuration diagram of the passenger movement route estimation device. Flowchart of passenger movement route estimation processing. An example of searching for the lowest cost route. An example of searching for the lowest cost route.

[Overview]
In this embodiment, in order to estimate the passenger flow for a given train schedule, it is assumed that a large number of passengers are based on the recorded data (entrance / exit data) of the passage time (entrance / exit time) of each passenger that can be acquired from the automatic ticket gate. Using a diamond network that models the movement of passengers on the train schedule, the movement route that specifies the boarding train and transfer station of each passenger is estimated.

[principle]
(A) Diamond network FIG. 1 is an example partially shown for explaining the outline of the diamond network. As shown in FIG. 1, the diamond network has nodes P (Pa, Pb, Pc, Pd, Pe) representing events related to trains and passengers, and edges Q (Qa, Qb, Qc, Qd, Qe) representing the movement of passengers. ) Is connected. Further, the right direction of the drawing represents the passage of time, and the edge Q is connected only in the direction along the passage of time, not in the direction against the passage of time.

Node P is the arrival node Pa, which represents the event (arrival station and arrival time) when the train arrives at the station according to the train schedule (train operation), and the event where the train according to the train schedule (train operation) departs from the station. Departure node Pb indicating (departure station and departure time), entrance node Pc indicating passenger entry into the station ticket gate corresponding to departure node Pb, and passenger outside the station ticket gate corresponding to arrival node Pa It includes a participation node Pd representing participation and a station stay node Pe indicating a passenger's stay in the station ticket gate corresponding to each of the arrival node Pa and the departure node Pb.

A reference time representing the occurrence time of the corresponding event is set for each node P. The time is indicated in the horizontal position with respect to the drawing. That is, the reference time of the arrival node Pa is the arrival time of the corresponding station of the corresponding train, the reference time of the departure node Pb is the departure time of the corresponding station of the corresponding train, and the reference time of the entrance node Pc corresponds to it. It is the same as the reference time (departure time) of the departure node Pb, the reference time of the participation node Pd is the same as the reference time (arrival time) of the corresponding arrival node Pa, and the reference time of the station stay node Pe corresponds to it. It is the same as the reference time of the incoming node Pa or the outgoing node Pb.

The edge Q is a directed edge from the earlier node P to the later node P, and is a train edge Qa connecting the arrival node Pa and the departure node Pb of each train along the train schedule (train operation). At each station, the station stay node Pe and the corresponding arrival node Pa or the departure node Pb are connected, or the entrance node Pc and the corresponding departure node Pb are connected, or the entry node and the corresponding arrival node Pa are connected. The boarding / alighting edge Qb that connects the two, and the station stay edge Qe that indicates that the passenger is staying in the ticket gate of the station, which is the edge that connects the station stay nodes Pe that have the reference time before and after at each station, and the reference at each station. The edge that connects the entry nodes Pc before and after the time, and the entry edge Qc that indicates that the passenger who entered is staying in the ticket gate of the station, and the edge that connects the entry nodes Pd whose reference time is before and after at each station. The entry edge Qd, which indicates that the passenger who got off at the destination station is staying in the ticket gate of the station, is included.

A cost is set for each edge Q. The cost is a value indicating the difficulty of selecting the edge Q as a route, and the smaller the cost, the easier it is to be selected as a route. On the contrary, instead of the cost, a value that is easier to be selected may be defined and used, but in that case, the value obtained by reversing the magnitude relation of the value (for example, the reciprocal) corresponds to the cost. Therefore, it can be said that the value is equal to the cost. In the present embodiment, basically, the time required for the transition between the nodes P, that is, the value corresponding to the time difference between the reference time of the start node of the edge Q and the reference time of the end node is set as the cost of the edge Q. Will be done. The cost of each edge Q can be set so that a route passing through the edge can be easily selected in addition to the time difference. For example, in order to select a route that stays at the station as much as possible, the costs of the station stay edge Qe, the entrance edge Qc, and the entry edge Qd should be reduced, and conversely, the route that is on the train as much as possible. Is set to reduce the cost of the train edge Qa so that is selected. In addition, the costs of the station stay edge Qe, the entrance edge Qc, and the entry edge Qd are set in consideration of the ease of staying at the relevant station in addition to the time difference. This "ease of stay" means that passengers tend to stay for purposes other than changing trains, for example, the cost of each edge Q at a station equipped with station Naka facilities, other It can be set to be smaller than each edge Q at the station. As a result, it is possible to estimate a travel route for staying at a specific station where it is easy to stay.

(B) Estimating the movement route As a specific example, the estimation of the passenger's movement route with respect to the train schedule shown in FIG. 2 will be described. The train schedule in FIG. 2 is composed of three trains (1 to 3 trains) heading from station C to station A via station B. For this train schedule, it is assumed that a passenger enters C station or B station and exits from A station.

FIG. 3 is a diagram network for the train diagram of FIG. Since it is assumed that passengers enter C station or B station and exit from A station, in the timetable network of FIG. 3, the arrival node Pa, entry node Pd, station stay node Pe, and station B station of C station are assumed. Participation node Pd, departure node Pb of station A, entry node Pc, and station stay node Pe are omitted. That is, the departure nodes Pb5, Pb4, Pb3 of the 1st to 3rd trains at the C station, the arrival nodes Pa5, Pa4 and the departure nodes Pb2, Pb1 of the 1st and 2nd trains at the B station, and the 1st to 3rd trains at the A station, respectively. Arrival nodes Pa3, Pa1, Pa2, station stay nodes Pe4, Pe2, Pe3, Pe1 corresponding to the arrival nodes Pa5, Pa4 and departure nodes Pb2, Pb1 of the 1st and 2nd trains at station B, and 1 at station C. ~ 3 entrance nodes Pc5, Pc4, Pc3 corresponding to the departure nodes Pb5, Pb4, Pb3 of each of the 3 trains, and 2 entrance nodes Pc2, Pc1 corresponding to the departure nodes Pb2, Pb1 of the 1st and 2nd trains at B station And, it is a diamond network having three entry nodes Pd3, Pd1, Pd2 corresponding to the arrival nodes Pa3, Pa1, Pa2 of each of the 1st to 3rd trains at the A station.

FIG. 4 is an example of entry / exit data 306. The entry / exit data 306 is composed of a plurality of details in which one statement corresponds to one railroad use from entry to exit of one passenger, and each statement includes an entry station, an entry time, an entry station, and , Has a data item of entry time. The entry / exit data 306 shown in FIG. 4 is composed of seven items, and six passengers (detail IDs "1" to "4", "6") who enter from station C and exit from station A. , "7") and one passenger entering from station B and exiting from station A (corresponding to "5" in detail ID), and seven passengers are assumed.

For each passenger defined in the entry / exit data 306, the entry node Pc and the entry node Pd are specified. That is, the entry node Pc immediately after the entry time at the entrance station of a certain passenger is specified as the entry node Pc of the passenger, and the entry node Pd immediately before the entry time at the entry station is specified as the entry node Pd of the passenger. .. For example, for a passenger whose detail ID is "1", since "C station" is entered at "5:50", the entrance node Pc is "entrance node Pc5" whose reference time is "6:00". Therefore, since the player participates in "A station" at "7:40", the participation node Pd becomes the "participation node Pd3" whose reference time is "7:30".

Then, for each of these passengers, the passenger is searched for the minimum cost route that is the route from the specified entry node Pc to the exit node Pd in the diamond network and that minimizes the total cost of the edge Q passing through. Estimate the movement route of.

(C) Search for the minimum cost route The number of passengers assumed by the entry / exit data 306 is as small as 7 in the example of FIG. 4 for the sake of simplification of explanation, but it is large in the actual passenger flow estimation. If we try to search for the minimum cost route for each of these many passengers, the amount of calculation will be enormous. Therefore, in the present embodiment, in order to reduce the amount of calculation, the search for the minimum cost route is performed as follows.

First, as shown in FIG. 5, passengers are grouped by the participation node Pd and sorted in the order of the latest reference time of the participation node Pd. Further, for each group, the passengers are sorted in order from the earliest reference time of the entrance node Pc. That is, each passenger of the entry / exit data 306 shown in FIG. 4 participates including three passengers (corresponding to the detailed IDs “3”, “5”, “7”) in the order of the latest reference time of the entry node Pd. A group of participating nodes Pd2 including a group of node Pd1 and three passengers (corresponding to detail IDs "2", "4", "6") and one passenger (corresponding to detail ID "1") ) Is included in the group of the entry node Pd3, and the passengers are further sorted in each group in the order of the earliest reference time of the entry node Pc.

Next, based on the result of this grouping, the combination of the entry node Pc and the exit node Pd of each passenger is set as the search node pair to be searched for the minimum cost route, and the search node pair list 310 listing these without duplication is provided. create.

FIG. 6 is a search node pair list 310 created for the entry / exit data 306 of FIG. In the search node pair list 310, the search node pairs are listed in order from the group with the latest reference time of the entry node Pd, and in each group, the reference time of the entry node Pc is listed in the order of earliest. In the search node pair list 310 of FIG. 6, the six search node pairs are in the order of three search node pairs including the participation node Pd1, two search node pairs including the participation node Pd2, and one search node pair including the participation node Pd3. Enumerated. In the search node pair list 310, the “search flag” associated with each search node pair is a flag indicating whether the minimum cost route for the search node pair has been searched.

By creating the search node pair list 310, for a plurality of passengers who have the same combination of the entry node Pc and the entry node Pd, it is sufficient to search the minimum cost route once for the combination, so that the minimum cost for the same combination is the minimum cost. It is possible to prevent the route from being searched multiple times.

Then, the search node pair is selected as the search target in the order listed in the search node pair list 310 created in this way, and the minimum cost route from the entry node Pc to the entry node Pd is searched. In the search node pair list 310, the search node pairs are listed in order from the group with the latest reference time of the entry node Pd, and each group is listed in order from the earliest reference time of the entry node Pc. Among the search node pairs with the latest reference time, the search node pair with the earliest reference time of the entrance node Pc is the first search target. That is, in the example of FIG. 6, the entry node Pc4 and the search node pair of the entry node Pd1 are the targets of the first search.

FIG. 7 is an explanatory diagram of the search for the minimum cost route. FIG. 7 shows a case where the search for the minimum cost route from the entry node Pc4 to the entry node Pd1 is performed. The search for the minimum cost route is performed by the route search method based on Dijkstra's algorithm. That is, the node that minimizes the cost of the route from the entry node Pd1 that is the end point of the route to the entry node Pd1 by tracing the edge Q in the opposite direction is sequentially determined, and the entry node Pc4 that is the start point of the route is determined. This is a route search method that ends the search when the node P reaches, and in the process of this search, the minimum cost of the route to the participating node Pd1 is obtained (hereinafter, referred to as “searched”) for each node P. Information on the cost and the next edge Q of the route (hereinafter referred to as "node search information") is stored.

FIG. 7 shows a state at the time when the search is completed, and node search information is shown for each node P that has been searched. That is, as the minimum cost route from the entry node Pc4 to the entry node Pd1, the route from the entry node Pc4 to the entry node Pd1 via the entry node Pc3, the departure node Pb3, the arrival node Pa2, and the entry node Pd2 is searched. It is shown that

A feature of this route search method is the minimum cost from another node P to the entry node Pd1 which is the end point in the search process of the minimum cost route from the entry node Pc4 which is the start point of the desired route to the entry node Pd1 which is the end point. The route is also required. In the example of FIG. 7, in the process of searching for the minimum cost route from the entry node Pc4 to the entry node Pd1, the minimum cost route from each of the other entry nodes Pc to the entry node Pd1 is also obtained. That is, in the search node pair list 310 of FIG. 6, the search node pair of the entry node Pc2 and the entry node Pd1 and the search node pair of the entry node Pc3 and the entry node Pd1 are also listed in the same group of the entry node Pd1. , The minimum cost route has been sought.

That is, for a group of participating nodes Pd listed in the search node pair list 310, the search range in the diamond network is the widest by first searching for the search node pair having the earliest reference time of the entry node Pc. Since it can be a range, it is more likely that the minimum cost route is being sought for other search node pairs in the same group during the search process. As a result, it is possible to search for the minimum cost route without searching for the overlapping route by diverting the search result of the overlapping minimum cost route for the passengers having the same participating node Pd.

[Functional configuration]
FIG. 8 is a functional configuration diagram of the passenger movement route estimation device 1. According to FIG. 8, the passenger movement route estimation device 1 includes an operation input unit 102, a display unit 104, a sound output unit 106, a communication unit 108, a processing unit 200, and a storage unit 300, and is a kind of type. It is configured as a computer system.

The operation input unit 102 is an input device realized by, for example, a keyboard, a mouse, a touch panel, various switches, etc., and outputs an operation signal corresponding to the operation input to the processing unit 200. The display unit 104 is a display device realized by, for example, a liquid crystal display or an organic EL display, and performs various displays based on a display signal from the processing unit 200. The sound output unit 106 is an audio output device realized by, for example, a speaker or the like, and outputs various audio based on an audio signal from the processing unit 200. The communication unit 108 is a communication device realized by, for example, a wireless communication module, a jack of a wired communication cable, a control circuit, or the like, and performs data communication with an external device.

The processing unit 200 is realized by an arithmetic unit such as a CPU (Central Processing Unit), and estimates a passenger's movement route along a given train schedule according to a program or data stored in the storage unit 300. As a functional unit that realizes the estimation of the passenger's movement route, the processing unit 200 moves with the diamond network creation unit 202, the entry / exit node identification unit 204, the search node pair list creation unit 206, and the minimum cost route search unit 208. It has a route estimation unit 210 and performs passenger movement route estimation processing (see FIG. 9) according to the passenger movement route estimation program 302. These functional parts may be processing blocks realized as software by executing a program, or are realized by hardware circuits such as ASIC (Application Specific Integrated Circuit) and FPGA (Field Programmable Gate Array). It may be a circuit block. In the present embodiment, it will be described as a processing block realized as software by executing the passenger movement route estimation program 302.

The timetable network creation unit 202 corresponds to the creation means, and based on a given train timetable, 1) sets arrival nodes and departure nodes that represent events related to arrivals and departures at each station of each train, and follows the operation of the train. Connecting the arrival node and departure node with an edge, 2) Corresponding to the arrival node and departure node of each station of each train, the entrance node and exit node representing the event related to the entry and exit of passengers at the station, and the station Station stay node is set, the station stay node is connected to the corresponding arrival node or departure node at the edge, the entrance node is connected to the corresponding departure node at the edge, and the entry node is connected to the corresponding arrival node. 3) Connecting entry nodes, exit nodes, and station stay nodes that are adjacent in time at each station with an edge, and 4) Setting a given cost for each edge. Create a diamond network.

That is, a diamond network representing a given diamond stored as train diamond data 304 is created. Specifically, the arrival node Pa and the departure node Pb representing the arrival / departure event at each station of each train are set, and the arrival node Pa and the departure node Pb are connected by the train edge Qa along with the operation of the train. In addition, the entry node Pd corresponding to the arrival node Pa and indicating the passenger's participation outside the station ticket gate and the entry node Pc corresponding to the departure node Pb indicating the passenger's entry into the station ticket gate are set, and the arrival node Pa is set. And the departure node Pb are connected to the corresponding entry node or entry node Pc by the boarding / alighting edge Qb. Further, at each station, the entry nodes Pd whose reference time is before and after are connected by the entry edge Qd, and the entry nodes Pc whose reference time is before and after are connected by the entry edge Qc. Further, the station stay node Pe corresponding to each of the arrival node Pa and the departure node Pb is set, and the station stay node Pe is connected to the corresponding arrival node Pa or the departure node Pb. Further, at each station, the station stay nodes Pes whose reference times are before and after are connected by the station stay edge Qe. Then, at each edge Q, the time difference between the reference times of the start node and the end node, the relative difference between the ease of getting on the train and the ease of staying at the station, and the relative ease of staying with other stations. Set the cost according to the above (see Fig. 1). The created diamond network is stored as diamond network data 308.

The entry / exit node identification unit 204 corresponds to a specific means, and is based on the entry / exit data in which the entry station, entry time, entry station, and entry time of each passenger are determined, and the entry node of each passenger in the diamond network (hereinafter, "" Specify the entry node (hereinafter referred to as the "specific entry node") and the entry node (hereinafter referred to as the "specific entry node").

That is, the entry node Pc and the entry node Pd are specified for each passenger defined by the entry / exit data 306. Specifically, the entry node Pc immediately after the entry time at the entrance station of a certain passenger is specified as the entry node Pc of the passenger, and the entry node Pd immediately before the entry time at the entry station is the entry node Pd of the passenger. (See FIGS. 4 and 5).

The search node pair list creation unit 206 groups each passenger defined by the entry / exit data 306 by the specified entry node Pd, sorts the passengers in the order of the latest reference time of the entry node Pd, and for each group, the entry node Pc. Sort in ascending order of reference time. Next, based on the result of this grouping, the combination of the entry node Pc and the entry node Pd of each passenger is used as a search node pair, and a search node pair list 310 in which these are listed without duplication is created (FIGS. 4 to 6). reference).

The minimum cost route search unit 208 corresponds to the search means, and sets the minimum cost route from a given search target entry node to a given search target entry node in the diamond network based on the cost set at the edge. Explore. In addition, by selecting the edge with the lowest cost by going back from the search target entry node, the minimum cost route from the node connected to the selected edge to the search target entry node is sequentially determined. The route search process is performed to search for the minimum cost route from the search target entry node to the search target entry node, and for passengers with the same specific entry node, the specific entry node is set as the search target entry node, and the minimum duplication By diverting the search result of the cost route, the route search process omitting the duplicate search is performed, and the minimum cost route from each specific entry node to the specific entry node for the same passenger is searched for by the specific entry node. In addition, for passengers with the same specific entry node, the route search process is performed in the order of the earliest set time of the specific entry node. In addition, among the passenger groups with the same specific participation node, the minimum cost route is searched in order from the group with the latest setting time of the specific participation node, and the specific participation node related to the group to be searched this time is the minimum searched in the previous group. When it is included as a transit node in the cost route, the route search process that omits the duplicate search is performed by diverting the search result that has been searched retroactively from the transit node.

That is, the search node pairs listed in the search node pair list 310 are selected as search targets in the order listed, and the minimum cost route from the entry node Pc of the search node pair to the entry node Pd in the search node pair is searched for in the diamond network. The search for the minimum cost route is performed by the route search method described with reference to FIG. 7, and for each node P searched in the search process, the node search information between the minimum cost and the next edge of the route is supported. Attach and remember.

When the minimum cost route for the search node pair to be searched is searched, the search flag of the search node pair to be searched is set to "1" in the search node pair list 310, and "search completed" is set. Further, for each of the other search node pairs in the same group of participating node Pd as the search node pair to be searched, it is determined whether the entry node Pc has been searched in the search process, and the search node pair in which the entry node Pc has been searched is determined. The minimum cost route for the search node pair is obtained from the node search information stored in association with the entry node Pc, and the search flag is set to "1" to set "search completed".

Next, in the search node pair list 310, if there is a search node pair having a search flag of "0" in the same group of participating node Pd as the search node pair to be searched, the search node pair is set as the next search target in the order of enumeration, and the minimum cost route is set. Search for. At this time, since the participating node Pd, which is the end point of the desired route, does not change, the node search information of each node P stored in the previous search is used as it is, and the search target is searched so as to inherit the previous search result. The minimum cost route is searched until the entry node Pc of the node pair is reached.

Then, if the search flags of all the search node pairs in the group of the participating node Pd that is the same as the pair to be searched are set to "1", the search node pair of the next group is set as the next search target in the search node pair list 310 in the order of enumeration. , Similarly, search for the minimum cost route. In this case, since the participating node Pd, which is the end point of the desired route, changes, the minimum cost of each searched node pair that clears all the stored node search information of each node P and starts the search for the minimum cost route. The route is stored as the minimum cost route search result data 312.

The movement route estimation unit 210 estimates the movement route of each passenger defined by the entry / exit data 306 based on the searched minimum cost route.

That is, for each passenger, boarding from the minimum cost route searched for the search node pair corresponding to the combination of the entry node Pc and the entry node Pd to the entry at the entry station after entering the entry station based on the train schedule. Estimate the travel route of the passenger, such as a train, transfer station, or stay station. The estimated travel route of each passenger is stored as the travel route estimation result data 314.

The storage unit 300 stores programs and data for realizing various functions for the processing unit 200 to collectively control the device, and is used as a work area of the processing unit 200. The processing unit 200 follows various programs. The executed calculation result and the input data from the operation input unit 102 and the communication unit 108 are temporarily stored. In the present embodiment, the storage unit 300 contains the passenger movement route estimation program 302, the train diagram data 304, the entry / exit data 306, the diagram network data 308, the search node pair list 310, and the minimum cost route search result data 312. And the movement route estimation result data 314 are stored.

[Processing flow]
FIG. 9 is a flowchart of the passenger movement route estimation process.
This process is realized by the processing unit 200 executing the passenger movement route estimation program 302.

In the passenger movement route estimation process, first, the diamond network creation unit 202 creates a diamond network based on a given train schedule (step S1). Next, the entry / exit node identification unit 204 identifies the entry node Pc and the entry node Pd in the diamond network for each passenger of the entry / exit data 306 (step S3).

Subsequently, the search node pair list creation unit 206 groups the passengers by the entry node Pd (step S5), and lists the search node pairs that are a combination of the entry node Pc and the entry node Pd of each passenger without duplication. The node pair list 310 is created (step S7). In the search node pair list 310, the groups are listed in the order of the latest reference time of the participating nodes, and in each group, the search node pairs are listed in the order of the earliest reference time of the entry node Pc.

Subsequently, the minimum cost route search unit 208 selects and selects the group of the participating nodes having the latest reference time from the unsearched (search flag is “0”) search node pair in the search node pair list 310 (step S9). Among the groups, the unsearched search node pair having the earliest reference time of the entry node Pc is selected as the search target (step S11). Then, the minimum cost route from the entry node Pc to the entry node Pd is searched for the search node pair to be searched (step S13). When the minimum cost route for the search node pair to be searched is searched, the search flag of the search node pair to be searched is set to "1" in the search node pair list 310, and "search completed" is set (step S15). Further, for each of the other search node pairs in the selected group, it is determined whether the entry node Pc has become "searched" in the search process, and the search node pair for which the entry node Pc has been searched is assigned to the entry node Pc. The minimum cost route for the search node pair is obtained from the node search information stored in association with each other, and the search flag is set to "1" to set "search completed" (step S17).

Subsequently, it is determined whether all the search node pairs in the selected group have "search completed", and if there is an unsearched search node pair (step S19: NO), the process returns to step S11, and the unsearched search is performed again. Select the search target from the node pairs and perform the same processing.

When all the search node pairs in the selected group are "search completed" (step S19: YES), the node search information stored in association with each node P of the diamond network is cleared (step S21), and then. , It is determined whether all the search node pairs in the search node pair list 310 are "search completed", and if there is an unsearched search node pair (step S23: NO), the process returns to step S9 and the next group is selected. Perform the same process.

When all the search node pairs are "search completed" (step S23: YES), the movement route estimation unit 210 determines the movement route from the entrance station to the exit station for each passenger according to the searched minimum cost route. Estimate (step S25). When the above processing is performed, the passenger movement route estimation processing is completed.

[Action effect]
As described above, according to the present embodiment, as a diamond network based on a given timetable, a station stay node Pe is set corresponding to each of the arrival node Pa and the departure node Pb of each train at each station, and the station stay node is set. Pe is connected to the corresponding arrival node Pa or departure node Pb at the edge, the entrance node Pc is connected to the corresponding departure node Pb at the edge, and the entry node Pd is connected to the corresponding arrival node Pa at the edge, in terms of time at each station. By connecting adjacent entry nodes Pc, entry nodes Pd, and station stay nodes Pe with an edge, it is possible to create a diamond network that takes into consideration the passenger's stay at the station ticket gate. Then, by searching for the minimum cost route from the given search target entry node of the created diamond network to the given search target entry node based on the cost set at the edge, the stay in the station ticket gate was considered. It is possible to estimate the travel route of passengers. For example, by setting the cost of the edge of each station according to the ease of staying at the ticket gate of the station, it is possible to estimate a route for passengers to stay at a station where it is easy to stay. ..

The search for the minimum cost route is to select the edge with the lowest cost by going back from the search target participation node, and the minimum cost route from the node connected to the selected edge to the search target participation node. Is performed by a route search process that sequentially determines. Therefore, the minimum cost route for passengers having the same specific participation node overlaps a part of the route portion leading to the specific participation node. Therefore, in the process of searching for the minimum cost route, it is assumed that the node for which the minimum cost route to the search target participating node has been determined has been searched, and the information related to the minimum cost and the minimum cost route (node search information) is stored in association with the node. By doing so, it is possible to divert the duplicated search results, reduce the amount of calculation required for estimating the travel route of the passenger, and speed up the processing.

[Modification example]
It should be noted that the applicable embodiments of the present invention are not limited to the above-described embodiments, and of course, they can be appropriately changed without departing from the spirit of the present invention.

(A) Diversion of the previous search result When searching for the minimum cost route for a search node pair whose entry node is different from the search node pair (combination of the entry node and the entry node) of the search target immediately before, the search immediately before the search. The result may be diverted. When the reference time of the entry node of the search node pair of the search target this time is earlier than the reference time of the entry node of the search node pair of the search target immediately before, the search result immediately before can be diverted.

Specifically, as shown in FIG. 10, in the process of searching for the minimum cost route, the participating nodes (in FIG. 10, the participating node Pd1 in FIG. 10) are excluded from the node search information of each node that has been searched. In FIG. 10, among the participating nodes Pd2 and Pd3), the participating node Pd that arrives earliest in the minimum cost route to the end point is included and stored as the earliest transit node. On the other hand, this is an example of searching for the minimum cost route from the entry node Pc4 to the entry node Pd1.

Next, it is assumed that the minimum cost route from the entry node Pc5 to the entry node Pd2 is searched for this diamond network as shown in FIG. In this case, among the search results immediately before shown in FIG. 10, the entry node that is the node before the reference time of the entry node Pd2 that is the end point of the route to be searched this time and the node that has passed through the node search information at the earliest is the end point. The stored node search information is diverted for the node that is the participating node Pd before the reference time of Pd2. For other nodes, the stored node search information is deleted without being diverted, and the node is not searched. That is, as shown in FIG. 11, the arrival node Pa1 whose reference time is later than the entry node Pd2 which is the end point of this search, the station stay nodes Pe1 and Pe2 whose entry node Pd1 is the earliest transit node of the node search information, and the departure node. Nodes Pb1, arrival node Pa4, entrance nodes Pc1, Pc2, etc. are not diverted and are not searched. The entry node Pd3, the arrival node Pa2, Pa3, the departure node Pb2, Pb3, the entry node Pc3, Pc4, etc., whose reference time is before the reference time of the entry node Pd2 and whose transit node is any of the entry nodes Pd2 and Pd3, are , The target of diversion of node search information.

Then, for each node to be diverted, the minimum cost of the stored node search information is set to the minimum cost from the entry node Pd2 which is the end point of this search to the entry node Pd1 which is the end point of the immediately preceding search. Update to the subtracted value. When the end point of the route to be searched this time has already been searched, the minimum cost of the node search information at the end point corresponds to the minimum cost to reach the end point of the immediately preceding search. In this way, the previous search result is diverted, and the current search is performed in a state where some nodes have already been searched and the node search information is stored.

1 Passenger movement route estimation device 200 Processing unit 202 Diamond network creation unit, 204 Entry / exit node identification unit 206 Search node pair list creation unit, 208 Minimum cost route search unit 210 Movement route estimation unit 300 Storage unit 302 Passenger movement route estimation program 304 Train Diamond data, 306 entry / exit data 308 Diamond network data, 310 Search node pair list 312 Minimum cost route search result data, 314 Travel route estimation result data P node Pa arrival node, Pb departure node, Pc entry node, Pd entry node Pe station stay Node Q edge Qa train edge, Qb boarding / alighting edge, Qc entry edge, Qd entry edge Qe station stay edge

Claims (6)

Computer,
Based on a given train schedule, 1) set arrival nodes and departure nodes that represent events related to arrivals and departures at each station of each train, and connect the arrival nodes and departure nodes with edges along the operation of the train. 2) Corresponding to the arrival node and the departure node of each station of each train, an entrance node and an exit node representing an event related to the entry and exit of passengers at the station, and a station stay node of the station are set. The station stay node is connected to the corresponding arrival node or the departure node at the edge, the entrance node is connected to the corresponding departure node at the edge, and the entry node is connected to the corresponding arrival node at the edge. 3) Connecting the entrance nodes, the entry nodes, and the station stay nodes that are adjacent to each other in time at each station with an edge, and 4) setting a given cost at each edge to form a diamond network. Creation method to create,
A search means for searching the minimum cost route from a given search target entry node to a given search target entry node in the diamond network based on the cost set at the edge.
A program to function as. The search means
By selecting the edge that has the minimum cost by going back from the search target participation node, the minimum cost route from the node connected to the selected edge to the search target participation node is sequentially determined. Performs a route search process to search for the minimum cost route from the search target entry node to the search target entry node.
The program according to claim 1. Based on the entry / exit data in which the entry station, entry time, entry station, and entry time of each passenger are determined, the entry node (hereinafter referred to as "specific entry node") and entry node (hereinafter "specific entry node") of each passenger in the timetable network. Specific means to identify (referred to as "specific participation node"),
To make the computer function as
The search means
For passengers with the same specific participation node, the specific participation node is set as the search target participation node, and the search result of the overlapping minimum cost route is diverted to perform the route search process omitting the duplicate search. Search for the minimum cost route from each specific entry node to the specific entry node for the same passenger with the specific entry node.
The program according to claim 2. The search means
For passengers with the same specific entry node, the route search process is performed in the order of the earliest set time of the specific entry node.
The program according to claim 3. The search means
From the group of passengers with the same specific participation node, the minimum cost route is searched in order from the group with the latest set time of the specific participation node.
When the specific participating node related to the group to be searched this time is included as a transit node in the minimum cost route searched in the previous group, the search result searched retroactively from the transit node can be diverted. Perform the route search process omitting the duplicate search.
The program according to claim 3 or 4. Based on a given train schedule, 1) set arrival nodes and departure nodes that represent events related to arrivals and departures at each station of each train, and connect the arrival nodes and departure nodes with edges along the operation of the train. 2) Corresponding to the arrival node and the departure node of each station of each train, an entrance node and an exit node representing an event related to the entry and exit of passengers at the station, and a station stay node of the station are set. The station stay node is connected to the corresponding arrival node or the departure node at the edge, the entrance node is connected to the corresponding departure node at the edge, and the entry node is connected to the corresponding arrival node at the edge. 3) Connecting the entrance nodes, the entry nodes, and the station stay nodes that are adjacent to each other in time at each station with an edge, and 4) setting a given cost at each edge to form a diamond network. How to create and how to create
A search means for searching the minimum cost route from a given search target entry node to a given search target entry node in the diamond network based on the cost set at the edge.
Passenger movement route estimation device equipped with.

Images