JP2007083800A - Passage searching method, passage searching program and passage searching system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a passage searching method and a passage searching program capable of searching a substitute passage superior next to a plurality of most superior passage information besides the most superior passage information in a public traffic network. <P>SOLUTION: The most superior passage information is generated by DIJKSTRA'S ALGORITHM by using a link table storing a starting station, an arrival station, required time and a distance up to the arrival station from the starting station and a bridge table of storing a position of a record of the link table with the starting station of the link table as the arrival station. A substitute passage is generated on the basis of a passage searched in its generating process, and the substitute passage is corrected by a GA method, and a plurality of substitutable passage superior next to the most superior passage information are searched. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a method, a program, and a system for searching for a plurality of types of optimal routes from a departure station to a destination station of a public transportation network using a computer. In particular, the best route information is obtained by the Dijkstra method and the best route is obtained. A route search method and route search program for reusing the route information used in the process of obtaining information and obtaining alternative routes of the best route information in the excellent order using a genetic algorithm (hereinafter referred to as GA). And a route search system.

  In general, the Dijkstra method has been used as a route search method for public transportation networks according to the prior art. This method selects the best route information when there are multiple selectable routes at the passing point, and when the route reaches the point where there are multiple selectable routes, the best route is selected again. This is a technique for searching for the best route information as a whole by selecting and repeating this.

Here, the public transportation network means transportation means in which departure time and arrival time including a railroad, arrival distance, section distance, and fare are determined in advance.
The best route information is, for example, the route with the shortest time in the case of a search using time as a search condition, and the route with the smallest total number of transfers in the case of using the number of transfers as a search condition. When the search condition is used, it means the route with the optimum search condition such as the route with the lowest charge.

In Patent Document 1, after searching for the best route information by the Dijkstra method, the transfer station used in the best route information is prohibited, and the route next to the best route information is sequentially searched again by the Dijkstra method. The technology is described.
Further, Patent Document 2 describes a technology related to a car navigation system that uses a genetic algorithm to obtain an optimum route and a plurality of alternative routes corresponding thereto.
Japanese Patent Laid-Open No. 11-44547 JP-A-9-178500

The above-described conventional public transport network route searching method has the following problems.
First, the Dijkstra method can obtain the best route information at high speed and is widely used. However, this method has a drawback that it cannot obtain an alternative route other than the best route information.

On the other hand, the technique disclosed in Patent Document 1 may not be able to search for an effective alternative route.
For example, according to the Dijkstra method, when searching for a route from Makurazaki to the center of Kagoshima, the best route information is (Makurazaki)-JR Ibusuki Makurazaki Line (Makurazaki-Yamakawa)-(Yamakawa)-JR Ibusuki Makurazaki Line (Yamakawa- Kagoshima Chuo)-Kagoshima Chuo is searched for a route from Makurazaki to Yamakawa, transfer at Yamakawa, and go to Kagoshima Chuo.
On the other hand, if the change in Yamakawa is prohibited in the route search according to the technique disclosed in Patent Document 1, an effective alternative route cannot be searched. In reality, however, (Makurazaki)-JR Ibusuki Makurazaki Line (Makurazaki-Yamakawa)-(Yamakawa)-JR Ibusuki Makurazaki Line (Yamakawa-Kagoshima Chuo)-(Minami Kagoshima)-Walking-(Minami Kagoshima Station)-Kagoshima City Tram 1 There are also routes to transfer to Minami Kagoshima after a transfer at Yamakawa, such as:-(Korimoto)-Kagoshima City Tram 2-(Kagoshima Chuo Ekimae)-On foot-(Kagoshima Chuo).
As described above, the technique described in Patent Document 1 has a problem that an alternative route cannot be searched when only one transfer station exists in the best route information or when there is no transfer station.

  Moreover, the technique described in Patent Document 2 is widely used in car navigation systems. However, in the road traffic network, the driver can freely select a route, whereas in the railway network, the train operation time, the destination, the transfer station, and the like must be taken into consideration. In addition, the car navigation system uses GPS to select a route based on the distance from the current position and the destination, the number of left / right turns, etc., whereas in the railway network, the route may be selected based on the required time. is there. For this reason, there exists a malfunction that the technique of patent document 2 cannot be applied to a railway network simply.

  An object of the present invention is to eliminate the above-described problems caused by the prior art, and a route search method and route capable of searching for an excellent alternative route next to a plurality of best route information in addition to the best route information in a public transport network. A search program and a route search system are provided.

In order to achieve the above object, the invention according to claim 1 is a route search system for searching for route information that matches a search condition from a departure station to an arrival station,
A link table storing a departure station, an arrival station, a required time from the departure station to the arrival station, a distance from the departure station to the arrival station, and a link ID uniquely assigned to the record including these;
The record stored in the link table is used as an input, and the best route information that matches the search conditions from the departure station to the destination station using the Dijkstra method and the best route information are calculated based on the record. A route search unit for holding a plurality of passage route information,
A route generation unit that calculates a plurality of generation route information from the passage route information held by the route search unit;
And a path correction unit that corrects passage path information by a genetic algorithm using the plurality of generated path information.

The invention according to claim 2 includes, in the route search system, connection information indicating connection link ID, transfer time, connection type of a corresponding record in which the arrival station of the link table is a departure station, and these. Providing a bridge table for storing the bridge ID uniquely assigned to the record, and adding the bridge ID to the corresponding record in which the arrival station of the link table is the departure station;
When the route search unit searches for a corresponding record in which the arrival station of the link table is the departure station, the bridge ID stored in the record of the link table is read, and the record having the bridge ID is read from the bridge table. A search is performed, a connection link ID stored in the searched record is read, and a record including the connection link ID is searched from a link table.

The invention according to claim 3 is a route search method of a route search system for searching for route information that matches a search condition from a departure station to an arrival station,
A link table for storing a departure station, an arrival station, a required time from the departure station to the arrival station, a distance from the departure station to the arrival station, and a link ID uniquely assigned to a record including these, and stored in the link table The best route information that matches the search conditions from the departure station to the destination station using the Dijkstra method based on the record and a plurality of passage routes obtained by calculating the best route information A route search unit that holds information, a route generation unit that calculates a plurality of generated route information from the pass route information held by the route search unit, And a path correction unit that performs correction,
The route search unit searches for a plurality of passage route information from the departure station to the destination station using the Dijkstra method based on a plurality of records stored in the link table, and searches for the best route information and a plurality of passage route information. Hold
The route generation unit calculates a plurality of generation route information based on the passage route information,
The route correction unit obtains an alternative route for the best route information in an excellent order from routes other than the best route information.

According to a fourth aspect of the present invention, in the route search method, the connection link ID of the corresponding record in which the arrival station of the link table is the departure station, transfer time, connection information indicating a connection type, and a record including these A bridge table for storing the bridge ID uniquely assigned to the link table, and adding the bridge ID to a corresponding record in which the arrival station of the link table is a departure station,
When the route search unit searches for a corresponding record in which the arrival station of the link table is the departure station, the bridge ID stored in the record of the link table is read, and the record having the bridge ID is read from the bridge table. A search is performed, a connection link ID stored in the searched record is read, and a record including the connection link ID is searched from a link table.

The invention according to claim 5 stores the departure station, the arrival station, the required time from the departure station to the arrival station, the distance from the departure station to the arrival station, and the link ID uniquely assigned to the record including these. And the record stored in the link table as input, and using the Dijkstra method based on the record, the best route information and the best route information that match the search conditions from the departure station to the arrival station are obtained. A route search unit that holds a plurality of passage route information obtained from the calculation process, a route generation unit that calculates a plurality of generation route information from the passage route information held by the route search unit, and the plurality of generation route information A route correction unit that corrects passage route information using a genetic algorithm, and a computer that controls each of the above components, and route information that matches a search condition from the departure station to the arrival station. A route search program of a route search system for searching a,
In the computer,
The route search unit searches for a plurality of passage route information from the departure station to the destination station using the Dijkstra method based on a plurality of records stored in the link table, and searches for the best route information and a plurality of passage route information. A function to hold
A function of the route generation unit to calculate a plurality of generation route information based on the passage route information;
The route correction unit realizes a function of obtaining an alternative route of the best route information from a route other than the best route information in an excellent order.

The invention according to claim 6 includes, in the route search program, connection link ID, transfer time, connection information indicating a type of connection, and the like of a corresponding record in which the arrival station of the link table is a departure station. 6. The route search system according to claim 5, wherein a bridge table for storing a bridge ID uniquely assigned to a record is provided, and the bridge ID is added to a corresponding record in which an arrival station of the link table is a departure station. A route search program,
In the computer,
When the route search unit searches for a corresponding record in which the arrival station of the link table is the departure station, the function of reading the bridge ID stored in the record of the link table and the record having the bridge ID are bridged A function of searching from a table, a function of reading a connection link ID stored in the searched record, and a function of searching a record including the connection link ID from the link table are realized.

  According to the route search method, route search program, and route search system of the present invention, the departure station, the arrival station, the link table that stores the required time and distance from the departure station to the arrival station, and the departure station of the link table as the arrival station. Prepare the bridge table that stores the position of the record of the link table in advance, read these tables, generate the best route information by Dijkstra method, and based on the passing route information searched in the generation process Since the generation route information is generated and the generation route information is corrected by the GA method, only the best route information can be derived by the Dijkstra method, whereas in the present invention, the best route information is obtained. A number of excellent alternative routes can then be searched.

Hereinafter, an embodiment of a route search method, a route search program, and a route search system according to the present invention will be described in detail with reference to the drawings.
FIG. 1 is an overall configuration diagram of a route search system according to an embodiment of the present invention.
FIG. 2 is a flowchart of processing in the embodiment of the present invention.
FIG. 3 is an example of a link table and a bridge table.
FIG. 4 is an example of a route map, a link for each stop station, and a link for each transfer station.
FIG. 5 is a flowchart of the route search program.
FIG. 6 is an example of a route map and an example of passage route information.
FIG. 7 is a flowchart of the route generation program.
FIG. 8 is an example of a route generated by the route generation program.
FIG. 9 is a flowchart of a route correction program based on a genetic algorithm.
FIG. 10 is an example of route division by the route correction program.
FIG. 11 shows a synthesis example of routes divided by the route correction program.
FIG. 12 is an example of the result corrected by the path correction program.

<Description of configuration>
A route search system in which a route search method and a route search program according to an embodiment of the present invention are implemented will be described with reference to FIG.
The system of this embodiment includes a route search system 100, a computer 30 held by a user, and a public communication line network 20 that connects the route search system 100 and the computer 30 held by the user.

  The route search system 100 includes a server 10 on which a program operates, a route search unit 11 that searches for a route by the Dijkstra method, a route generation unit 12 that generates an alternative route from the route searched by the route search unit 11, A route correction unit 13 that corrects the route generated by the route generation unit 12 using a genetic algorithm, a link table 14 that stores the required time and distance from the departure station to the arrival station, and the arrival station in the link table 14 A bridge table 15 that stores connection information that is information relating to the ID and transfer time and train connection of the record, a passing route information table 16 that stores a passing route used by the route search unit 11, and a passing route information table 16 is a generation path information table for storing generation path information generated from Provided with a table 17.

The route search unit 11 obtains the best route information by the Dijkstra method, and stores all the routes used in the process in the passing route information table 16.
The route generation unit 12 derives generated route information that can be substituted for the best route information from the passage route information stored in the passage route information table 16 and stores it in the generated route information table 17.
The route correction unit 13 corrects the generated route information stored in the generated route information table 17 and derives a plurality of excellent alternative routes following the best route information.

<Description of data>
A data configuration example of the link table 14 will be described with reference to FIG.
The link table 14 includes a link ID 301 uniquely defined in the record, a departure station 302 representing a departure station name, an arrival station 303 representing an arrival station name, a train 304 representing a type of train such as limited express and express, and a bridge table 15. A start bridge ID 305 indicating the bridge ID 308, a required time 306 representing the time taken from the departure station to the arrival station, and a distance 307 representing a distance from the departure station to the arrival station. The link ID 301 is “1”, the departure station 302 is “n (1)”, the arrival station 303 is “n (2)”, the start bridge ID 305 is “1”, the required time 306 is “x (1)”, and the distance 307 Is “d (1)”.
Here, the link means data that holds a departure station, an arrival station, a train, a required time from the departure station to the arrival station, a distance from the departure station to the arrival station, and the like. The link table 14 holds links between all stations.

A data configuration example of the bridge table 15 will be described with reference to FIG.
The bridge table 15 includes a bridge ID 308 uniquely defined in the record, a connection link ID 309 indicating a link ID of a record whose departure station is an arrival station of a certain record in the link table 14, and a transfer time indicating a time required for transfer. 310 and connection information 311 representing information related to train connection such as transfer, direct connection, or return. Specific examples of data include a bridge ID 308 of “1”, a connection link ID 309 of “112”, and a transfer time of 310. “32” and the connection information is “transfer”.

Here, the existence significance of the bridge table 15 will be described.
When searching for a link to be connected from a certain link in the route search, it takes time because the data amount of the link table 14 is enormous when searching from the entire link table. For example, in the link with the link ID “1” in FIG. 3A, the arrival station is n (2), so the link connected to this is the link starting from n (2), but n (2) It is not known where in the link table 14 there is a link starting from. For this reason, it is necessary to search the entire table for a link having n (2) as a departure station. However, since the link table 14 has an enormous number of records, it takes time to search from there. Therefore, data storing link information to be connected to a certain link is prepared. This data is called a bridge. The start ID of the bridge is stored in the link. The bridge stores a link ID to be connected next, a transfer time for connection, and connection information.

  With this configuration, for example, in the link whose link ID 301 in the link table 14 in FIG. 3A is “1”, the start bridge ID 305 of this link is “1”. The bridge whose bridge ID 308 in the bridge table 15 is “1” is referred to. Since the connection link ID 309 of the bridge whose bridge ID 308 is “1” is “112”, it can be seen that the link ID 301 of the link connected from the link ID 301 is “112”. In addition, since there are a plurality of links connecting from the link having the link ID 301 of “1”, specifically, the departure station n (2) of this link as a departure station, there are as many bridges as there are links. Since the start bridge ID 305 of the link whose link ID 301 is “2” is “51”, the links connected from the link whose link ID 301 is “1” are those whose bridge ID 308 of the bridge table 15 is “1” to “50”. It can be seen that there is connection information in the bridge. By referring to all these bridges, all the links to be connected can be acquired.

Further, by repeatedly acquiring the link ID 301 to be connected next based on the acquired start bridge ID 305 of the link, the network of the route can be searched.
By using this bridge, the link to be connected next can be acquired easily and at high speed. Also, by providing the bridge with link-to-link connection information in advance, it is possible to omit determination processing such as transfer or direct connection from the route search processing, thereby improving route search processing efficiency.

A data configuration example of the link table 14 will be described with reference to FIG.
When stations and routes as shown in FIG. 4 (A) are arranged, links are not set for each stop station as shown in FIG. 4 (B), but can be changed as shown in FIG. 4 (C). Set a link to As a result, link data can be reduced, the amount of memory used can be reduced, and the processing efficiency due to the phenomenon to be searched can be improved. If the departure station or destination station is not a transferable station, it is assumed that the departure station or destination station is a transfer station and a link to the nearest transferable station is generated to create a link to the nearest transferable station. Route search can be performed.

A data configuration example of the passage route information table 16 will be described with reference to FIG.
FIG. 6A shows an example of a route network. For example, the route with the link ID “5453” starts from station B, and the immediately preceding route has the link ID “777”, the “1777”, and the “2245”. FIG. 6B shows an example of the passage route information table 16 in this case. The data configuration example of the passage route information table 15 includes, for each link, a previous link ID 601 that is a previous route, a previous storage position 602, and a route evaluation value 603. The route evaluation value 603 includes a required time 604. , The number of transfers 605 and the distance 606. When the link ID is “4553”, the specific link data 601 is “1667”, the previous storage position 602 is “3”, and the required time 604 is “78”, the number of transfers 605 is “2”, and the distance 606 is “114.3 km”.

  A data configuration example of the generation path information table 17 will be described with reference to FIG. The generated route information table 17 includes a route 801 indicating the generated generated route information, a required time 802 indicating the required time from the departure station to the destination station, a number of transfers 803 indicating the number of transfers, and from the departure station to the destination station. The distance 804 that indicates the distance of the route 801, the specific example of the data, the route 801 is “SABG”, the required time 802 is “25”, the number of transfers 803 is “1”, the distance 804 is “23.8 km”.

<Description of operation>
The outline of the processing flow of the embodiment of the present invention will be described with reference to FIG.
First, data necessary for route search is read [step 201].
Next, search conditions are set [step 202]. Here, the search condition refers to, for example, a search type that is a matter to be prioritized in a search, such as departure station, destination station, train type used, priority on required time or priority on number of transfers.

Next, the route search unit 11 performs route search by the Dijkstra method [step 203]. In the Dijkstra method, the best route information is obtained, and all the passage route information of the link table used for obtaining the best route information is held in the passage route information table 16 at the same time.
Next, a plurality of generation route information is generated by the route generation unit 12 from the all-pass information [step 204].

Next, a plurality of generated route information generated by the route generating unit 12 is corrected by the GA [Step 205]. In GA, a single route, that is, a route from the departure station to the destination station, is used as one gene, a new route is generated by one-point crossover, and a plurality of routes following the best route information are generated from the generated routes. An alternative route is selected [step 206].
Here, “one-point intersection” means that two routes are exchanged at one point in the middle of the route.

  As described above, the outline of the processing of the embodiment of the present invention is as follows. First, the route search unit 11 obtains the best route information and the passing route information used when obtaining the best route information by the Dijkstra method, and then the route. Generating route information is generated from the passing route information used when the generating unit 12 obtains the best route information. Finally, the route correcting unit 13 performs route correction by GA, and the next to the best route information An alternative route is sought.

Next, detailed operation of the route search 203 by the Dijkstra method which is the route search unit 11 will be described with reference to FIG. First, variables and values used in the route search unit 11 will be described.
N: A variable for storing a search target designated station.
L (i): A group of links with the designated station as the departure station.
E: A variable for storing passage route information. The current link ID, the immediately preceding link ID, the immediately preceding passing route information storage position, and the route evaluation value are stored.
T (x, y): A two-dimensional array that stores passage route information when the link passes. x is a link ID, and y is a storage position.
S (j): An array for storing passage route information for a route search target.
D (k): A passage route information array. Hold destination station arrival route information.

Next, the processing flow will be described.
The route search unit 11 first sets a departure station [step S501].
Next, the route search unit 11 assigns a link group having N as a departure station to L [step S502].
Next, the route search part 11 determines whether a link group exists [step S503]. If the link group does not exist, the process is terminated. If the link group exists, a variable E1 for holding the passage route information to the previous link and a variable E2 for holding the passage progress information to the current link are prepared. The process proceeds to step S504. The initial values of E1 and E2 are 0 (empty). The subscript of the link group is set to i and i = 1 is set [step 504].

Next, the route search unit 11 determines whether L (i) exists [step S505]. If it exists, the process proceeds to step S506. If not, go to step 516.
The route evaluation value of the passage route information E2 of the current link L (i) is obtained by adding the evaluation value of the link of L (i) to E1 [step S506].
Next, the route search unit 11 determines whether L (i) has passed or not passed [step S507]. If L (i) has not passed, the passage route information about L (i) is added to S [step S508] and E2 is added to T (L (i)) [step S511]. If L (i) passes, the process proceeds to step S509.

  When L (i) has already passed, the route search unit 11 determines whether E2 is superior to the best route information evaluation value of T (L (i)) [step S509]. If it is superior to the best one, the passage route information for L (i) of S is rewritten to E2 [step S510], and E2 is added to T (L (i)) [step S511].

Conventionally, in the Dijkstra method, only the best route information is held in order to obtain only the best route information, and if there is more excellent route route information, the original value is rewritten. However, in this embodiment, in order to use when creating a later route, non-best route route information is also retained. For this reason, E2 is not rewritten but added to T (L (i)).
When E2 is not superior to the best route information evaluation value of T (L (i)), the route search unit 11 determines whether L (i) includes a loop section [step S512]. If not included, the process proceeds to S511. If included, the process proceeds to step S515.
Here, “including the loop section” means that there are overlapping train stations in the route. Since redundant sections are redundant sections and become impractical paths, routes including loop sections are discarded.

Following step S511, the route search unit 11 determines whether the arrival station of L (i) is the destination station [step S513]. If it is not the destination station, the process proceeds to step S515. If it is the destination station, E2 is added to D [step S514], and the process proceeds to step S515.
In step S515, 1 is added to i, and the process returns to step S505.
Further, the route search unit 11 proceeds to step S516 when L (i) no longer exists in step S505, that is, when all the processing of links having N as a departure station is completed. In step S516, it is determined whether S is empty. If it is empty, the process ends. If it is not empty, in order to obtain the passage route information to be searched next from S, the best passage route information of S is substituted into E1 [step S517], and the value set in E1 becomes the search target. Therefore, the set value for E1 is deleted from S [step S518].
Next, the route search unit 11 substitutes the arrival station of the current link of E1 for N [step S519].

Next, the route search unit 11 determines whether N is the destination station [step S520]. If N is the destination station, E1 is added to D and the process returns to step S516. If N is not the destination station, the process returns to step S520, and a link group having N as the departure station is acquired. Here, when acquiring a link group having N as a departure station, a bridge can be acquired based on the bridge start ID of the link, and information relating to transfer and a link to be connected next can be acquired. By preparing data in advance in the bridge table 15 in this way, it is possible to omit the search for the next connected link, the determination and acquisition of information related to the transfer information, reduce the search processing, and the search time. Improvements can be made.
The route search unit 11 ends the process when S becomes empty [step S516].

  As a result of the above processing, each link is in a state of holding the passage route information. For example, in the network as shown in FIG. 6A, the link with the link ID 4553 includes the passage route information when coming from the link ID 1667, the passage route information when coming from the link ID 777, and the passage when coming from the link ID 2245. As the route information, when the route search is completed, the network and the generated route information table 17 in which each link dries the passing route information generated during the route search are generated. Using this passage route information, a route from the departure station to the destination station is created.

A method of creating a route based on the passing route information obtained by the route searching unit 11 by the route generating unit 12 will be described with reference to FIG.
The route from the destination station to the destination station is acquired by using the passing route information group D that reaches the destination station acquired by the route search unit 11 to follow the destination station.
First, the route generation unit 12 substitutes 1 for k [step S701].
Next, the path generation unit 12 determines whether D (k) exists [step S702]. When D (k) does not exist, it is determined whether an existing route is generated or a new route is generated [step S714]. When a new route is generated, the process returns to step S701 to generate an existing route. The process ends. If D (k) exists, the link possessed by D (k) is acquired [step S703], and the process proceeds to step S740.

Next, the route generation unit 12 determines whether or not the link acquired in step S703 has reached the departure station. If not reached, the immediately preceding passage route information is acquired [step S705].
Next, the route generation unit 12 determines whether or not the passage route information acquired in step S705 has been used [step S706]. If not used, it is used as it is, and the process proceeds to step S709. If it has been used, it is determined whether there is other unused passage route information in the passage route information acquired in step S705 [step S708]. If there is unused passage route information, the best passage route information is acquired from the unused passage route information [step S708] and the process proceeds to step S709.

Next, the route generation unit 12 compares the passing route information acquired in step S705 or the passing route information acquired in step S708 with the passing route information acquired so far, and determines whether the route includes a loop section. [Steps S709, S710]. If the loop section is not included, the passage route information obtained in step S705 or the link of the passage route information obtained in step S708 is obtained [step S711] and the process returns to step S704.
If the link has reached the departure station in step S704, and if step S710 includes a loop section, the route generation unit 12 sets all the passage route information acquired for route generation to be used [step S712] 1 is added to k [step S713], and the process returns to step S702.
The route information generated as described above is referred to as generated route information.

  FIG. 8 shows an example of the generation route information table 17 generated by the route generation unit 12. In FIG. 8, the departure station is S and the destination station is G. The generated route information generated by the route generating unit 12 is forcibly connected avoiding the used passing route information and is low in accuracy, but after this, the route generated using the GA is corrected. .

Next, an example of a processing flow of the route correction unit 13 using GA will be described with reference to FIG.
First, the route correction unit 13 acquires the generated route information generated by the route generation unit 12 and generates a ranking. Here, the ranking is a table in which a certain number of generation path information is retained in the order of evaluation while eliminating duplication. When the route correction unit 13 inserts new generation route information into the ranking, for example, the generation route information is superior to the eleventh generation route information of the current ranking and not superior to the tenth generation route information. Insert between 10th and 11th positions. All the generation path information of 11th and lower ranks are moved down by one, the lowest generation path information is deleted, and the number of generation path information held in the ranking is kept constant.

After the route is inserted into the ranking, it is attached to all the generated route information generated by the route generation unit 12 and divided at each stop station to generate a set of the first half route and the second half route. For example, as shown in FIG. 10, when there is a route A-B-C-D-E, (A-B) (B-C-D-E), (A-B-C) (C-D-- E), (A-B-C-D), and (D-E).
Next, the route correction unit 13 summarizes the divided first half route and the second half route for each divided route [step S902], and determines whether there is a stop station on the route [step S903]. If there is no stop station, the process ends. When there is a stop station, the route correction unit 13 tries to evaluate all connection patterns for each stop station, and acquires an excellent route with a high evaluation. Specifically, first, one stop station is selected, and the presence or absence of the first half route is determined [step 905]. When the first half route does not exist, the next stop station is acquired [step S904], and the process returns to step S903. If the first half path exists, the presence / absence of the second half path is determined [step S907]. If the second half route also exists, the first half route and the second half route are connected to create a route from the departure station to the destination station [step S908].

  The generated route is inserted into the ranking [Step S909], and it is determined whether the insertion is successful [Step S910]. If the insertion is successful, the route is separately held as a new route [step S911]. If the insertion fails, it means that the evaluation value is lower than the lowest ranking of the current ranking or that the ranking already has the same route, so this route is deleted.

FIG. 11 shows a case of division at station B. It shows that the first half route and the second half route are connected in combination, evaluated, and attempted to be inserted into the ranking. Successful insertion in the ranking means that the route is retained in the ranking, and failure means that the route is not retained in the ranking because it is not superior to the lowest ranking.
Next, the route correction unit 13 acquires the next second half route [step S912], and returns to step S907.
Next, when the latter half route does not exist in step 907, the route correction unit 13 acquires the next first half route [step S906], returns to step S905, and when the first half route does not exist in step S905, the next stop The station is acquired [step S904], the process returns to step S903, and the process ends when there is no stop station in step S903.

  FIG. 12 shows a processing result by the route correction unit 13. As an example, a case where the departure station is S and the destination station is G is shown. As a result, the routes generated by the route search unit 11 and the route generation unit 12 by the Dijkstra method are corrected, and are arranged in the order of excellent route evaluation values.

Specific examples of the route search method described above will be given. The route evaluation uses a required time, the number of transfers, and a distance, and performs a required time search, a transfer search, and a distance search. The route evaluation method in the nuclear search is as follows.
(Time required search)
Route search is performed based on the required time. In this case, a shorter route is a better route. If the required time is the same, the number of transfers is compared, and the one with the smaller number of transfers is given priority. When the required time and the number of transfers are the same, priority is given to those with a short distance.
(Search for the number of transfers)
Route search is performed based on the number of transfers. In this case, a route with a smaller number of transfers is an excellent route. If the number of transfers is the same, the required times are compared, and the shorter required time is given priority. If the number of transfers and the required time are the same, priority is given to those with a short distance.

(Distance search)
Route search based on distance. In this case, a shorter route is an excellent route. Railway routes often calculate fares based on distance, and it cannot be generally stated, but if the distance is short, the amount is often cheap. In addition, there is a rule in the JR fare rule that a fare is calculated at the lowest fare regardless of how a train is taken within a certain section. In this case, in order to find the cheapest route, the shortest distance route is obtained. There is a need. If the distance is the same, compare the required time, and give priority to the shorter required time. If the distance and required time are the same, priority is given to the one with the fewest number of transfers.

The results obtained between (Mizoguchi) and (Tokyo) are shown below. [Table 1] is the required time search result, [Table 2] is the transfer number search result, [Table 3] is the distance search result,
All show the top five routes. In this route search, the number of rankings is set to 200.

  As described above, a practical effective alternative route that is close to the first best evaluation route is derived after the second route. It can also be seen that various results are derived by each search method. In each table, only the top 5 routes are displayed, but in reality, only the number of rankings, that is, 200 routes in this case, are stored. In the results of the distance search in [Table 3], only similar routes are derived. For example, routes in the ranking are grouped under the condition that “when the route having the same transfer pattern is considered” Various patterns of routes can be derived by deriving routes from the top of the group.

  In the above example, simple time search, number-of-transfers search, and distance search are shown. For example, one point for one minute of time, 10 points for one transfer, and so on. If converted and used as a route evaluation value, it is possible to provide a route search system capable of flexible route search using a plurality of conditions by route correction according to the evaluation value by GA.

1 is an overall configuration diagram of a route search system in an embodiment of the present invention. It is a flowchart of the process in embodiment of this invention. It is an example of a link table and a bridge table. It is an example of a route map, the link of a stop station unit, and the link of a transfer station unit. It is a flowchart of a route search program. It is an example of a route map and an example of passage route information. It is a flowchart of a route generation program. It is an example of the production | generation route information produced | generated by the route production | generation program. It is a flowchart of the path | route correction program by a genetic algorithm. It is an example of route division by a route correction program. It is a synthesis example of a route divided by a route correction program. It is an example of the result corrected by the route correction program.

Explanation of symbols

10: Server, 11: Dijkstra method route creation unit, 12: GA route creation unit, 13: Link table, 14: Bridge table, 15: Passage route information table, 16: Generated route information table, 20: Public communication network 30: User terminal.

Claims (6)

A route search system that searches for route information that matches a search condition from a departure station to an arrival station,
A link table storing a departure station, an arrival station, a required time from the departure station to the arrival station, a distance from the departure station to the arrival station, and a link ID uniquely assigned to the record including these;
The record stored in the link table is used as an input, and the best route information that matches the search conditions from the departure station to the destination station using the Dijkstra method and the best route information are calculated based on the record. A route search unit for holding a plurality of passage route information,
A route generation unit that calculates a plurality of generation route information from the passage route information held by the route search unit;
A route search system comprising: a route correction unit that corrects passage route information by a genetic algorithm using the plurality of generated route information. A bridge table for storing a connection link ID of a corresponding record in which the arrival station of the link table is a departure station, a transfer time, connection information indicating a connection type, and a bridge ID uniquely assigned to a record including these And adding the bridge ID to the corresponding record in which the arrival station of the link table is the departure station,
When the route search unit searches for a corresponding record in which the arrival station of the link table is the departure station, the bridge ID stored in the record of the link table is read, and the record having the bridge ID is read from the bridge table. The route search system according to claim 1, wherein a search is performed, a connection link ID stored in the searched record is read, and a record including the connection link ID is searched from a link table. A route search method of a route search system that searches for route information that matches a search condition from a departure station to an arrival station,
A link table for storing a departure station, an arrival station, a required time from the departure station to the arrival station, a distance from the departure station to the arrival station, and a link ID uniquely assigned to a record including these, and stored in the link table The best route information that matches the search conditions from the departure station to the destination station using the Dijkstra method based on the record and a plurality of passage routes obtained by calculating the best route information A route search unit that holds information, a route generation unit that calculates a plurality of generated route information from the pass route information held by the route search unit, And a path correction unit that performs correction,
The route search unit searches for a plurality of passage route information from the departure station to the destination station using the Dijkstra method based on a plurality of records stored in the link table, and searches for the best route information and a plurality of passage route information. Hold
The route generation unit calculates a plurality of generation route information based on the passage route information,
The route search method of the route search system, wherein the route correction unit obtains an alternative route of the best route information from a route other than the best route information in an excellent order. A bridge table for storing a connection link ID of a corresponding record in which the arrival station of the link table is a departure station, a transfer time, connection information indicating a connection type, and a bridge ID uniquely assigned to a record including these And adding the bridge ID to the corresponding record in which the arrival station of the link table is the departure station,
When the route search unit searches for a corresponding record in which the arrival station of the link table is the departure station, the bridge ID stored in the record of the link table is read, and the record having the bridge ID is read from the bridge table. 4. The route search method according to claim 3, wherein a search is performed, a connection link ID stored in the searched record is read, and a record including the connection link ID is searched from a link table. A link table for storing a departure station, an arrival station, a required time from the departure station to the arrival station, a distance from the departure station to the arrival station, and a link ID uniquely assigned to a record including these, and stored in the link table The best route information that matches the search conditions from the departure station to the destination station using the Dijkstra method based on the record and a plurality of passage routes obtained by calculating the best route information A route search unit that holds information, a route generation unit that calculates a plurality of generated route information from the pass route information held by the route search unit, A route search system that includes a route correction unit that performs correction, and a computer that controls each of the components, and searches for route information that matches a search condition from the departure station to the arrival station. A route search program,
In the computer,
The route search unit searches for a plurality of passage route information from the departure station to the destination station using the Dijkstra method based on a plurality of records stored in the link table, and searches for the best route information and a plurality of passage route information. A function to hold
A function of the route generation unit to calculate a plurality of generation route information based on the passage route information;
A route search program for a route search system, wherein the route correction unit realizes a function of obtaining an alternative route of the best route information in an excellent order from routes other than the best route information. A bridge table for storing a connection link ID of a corresponding record in which the arrival station of the link table is a departure station, a transfer time, connection information indicating a connection type, and a bridge ID uniquely assigned to a record including these A route search program for a route search system according to claim 5, wherein the bridge ID is added to a corresponding record in which the arrival station of the link table is a departure station.
In the computer,
When the route search unit searches for a corresponding record in which the arrival station of the link table is the departure station, the function of reading the bridge ID stored in the record of the link table and the record having the bridge ID are bridged 6. A function for retrieving from a table, a function for reading a connection link ID stored in the retrieved record, and a function for retrieving a record including the connection link ID from the link table are realized. The described route search program.

Images