JP2017058194A - Route search system, route search method and computer program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a route search system, a route search method and a computer program capable of searching an appropriate route in a manner of heuristic search under any search conditions.SOLUTION: The route search system is configured: to create an orthogonal table in which a predetermined number of test patterns are prescribed in a combination of departure point, a destination, and travelling time; to search a route from a departure point to a destination within the travelling time prescribed by the test pattern as an evaluation route in a manner of heuristic search for each test pattern prescribed in the orthogonal table; and to set a parameter to be used for heuristic search so that the evaluation route obtains a highest evaluation point.SELECTED DRAWING: Figure 6

Description

  The present invention relates to a route search system, a route search method, and a computer program for searching a route by heuristic search.

  2. Description of the Related Art In recent years, a navigation device is often mounted on a vehicle that provides vehicle travel guidance so that a driver can easily arrive at a desired destination. Here, the navigation device detects the current position of the vehicle by a GPS receiver or the like, acquires map data corresponding to the current position through a recording medium such as a DVD-ROM or HDD or a network, and displays it on a liquid crystal monitor. It is a device that can do. Further, the navigation device has a route search function for searching for an optimum route from the vehicle position to the destination when a desired destination is input, and sets the searched optimum route as a guide route, and displays it. A guide route is displayed on the screen, and when the user approaches an intersection, the user is surely guided to a desired destination by voice guidance. In recent years, some mobile phones, smartphones, tablet terminals, personal computers, and the like have functions similar to those of the navigation device.

  In the above route search function, the search cost (link cost, intersection cost) is calculated for each link and each node included in the route, and the route that minimizes the added value of the search cost is optimal. Specify as a route. As an algorithm for searching for the optimum route, a Dijkstra method, a breadth-first search, or the like is conventionally used. In these Dijkstra methods and breadth-first search, the optimum route is searched by blind search without using heuristics, and thus there is a problem that a long time is required for search processing particularly when there are a large number of links and intersections. Accordingly, heuristic search has been proposed as a route search method for performing route search processing at higher speed (Japanese Patent No. 3012096).

  Here, in the heuristic search, cost calculation is not performed for all possible routes from the departure point to the destination, but a future cost addition value is predicted using a heuristic function or the like, and a route search direction ( It is characterized in that the amount of calculation is reduced by limiting the direction in which the route is extended) or by providing priority in the search direction, thereby speeding up the route search process.

Japanese Patent No. 3012096 (page 7-10, FIG. 11, FIG. 12)

  However, in the heuristic search described in Patent Document 1, the route search processing can be speeded up. On the other hand, the search direction is limited in advance, and therefore an optimal route may not be searched unless the restriction is set appropriately. . For example, when the combination of the departure point and the destination is a specific combination, or when the traveling time is a specific time (that is, a search condition that is not compatible with the limited search direction), the quality of the searched route is degraded. There is a problem to do.

  The present invention has been made to solve the above-described conventional problems, and optimizes the parameters used for the heuristic search by performing a test pattern using an orthogonal table, and relates to the combination of the starting point and the destination and the traveling time. An object of the present invention is to provide a route search system, a route search method, and a computer program that can search for an appropriate route by a heuristic search under any search condition.

  In order to achieve the above object, the route search system according to the present invention includes an orthogonal table creating means for creating an orthogonal table in which a predetermined number of test patterns including a combination of a starting point, a destination, and a travel time are defined, For each specified test pattern, an evaluation route search means for searching a route from the departure place to the destination at the travel time specified in the test pattern by heuristic search, and a test by the evaluation route search means A route evaluation unit that evaluates an evaluation route that is each route searched for each pattern; an optimum parameter identification unit that identifies an optimum value of a parameter used for the heuristic search based on an evaluation result of the route evaluation unit; and the optimum The optimum value of the parameter specified by the parameter specifying means is used for the heuristic search. Having a parameter setting means for setting as that parameter.

  The route search method according to the present invention is a route search method for searching for a more appropriate recommended route by setting parameters used for heuristic search to optimum values. Specifically, the orthogonal table creating means creates an orthogonal table in which a predetermined number of test patterns composed of combinations of departure place, destination, and travel time are defined, and the evaluation route search means is defined in the orthogonal table. Searching for a route from the departure point to the destination at the travel time specified in the test pattern for each of the test patterns, by means of a heuristic search, and route evaluation means by the evaluation route search means A step of evaluating an evaluation route that is each route searched for each test pattern; and a step of specifying an optimum value of a parameter used for the heuristic search by an optimum parameter specifying unit based on an evaluation result of the route evaluation unit; A parameter setting means that sets the optimum value of the parameter specified by the optimum parameter specifying means in advance It has a step of setting as a parameter used for heuristic search, the.

  The computer program according to the present invention is a computer program for searching for a more appropriate recommended route by setting parameters used for heuristic search to optimum values. Specifically, the computer includes an orthogonal table creating means for creating an orthogonal table in which a predetermined number of test patterns including a combination of a starting point, a destination, and a traveling time are defined, and for each test pattern defined in the orthogonal table. , An evaluation route search means for searching a route from the departure place to the destination at the travel time specified in the test pattern by heuristic search, and each of the test patterns searched by the evaluation route search means for each test pattern A route evaluation unit that evaluates an evaluation route that is a route, an optimum parameter identification unit that identifies an optimum value of a parameter to be used for the heuristic search based on an evaluation result of the route evaluation unit, and the optimum parameter identification unit The optimum value of the parameter is set as a parameter used for the heuristic search. And parameter setting means, thereby to function.

  According to the route search system, route search method, and computer program according to the present invention having the above-described configuration, it is possible to set a parameter used for heuristic search to an optimum value by executing a test pattern using an orthogonal table. As a result, an appropriate route can be searched by heuristic search regardless of the search conditions regardless of the combination of the departure point and the destination and the travel time.

1 is a schematic configuration diagram illustrating a route search system according to the present embodiment. It is the block diagram which showed the structure of the route search system which concerns on this embodiment. It is a figure explaining the search method by a heuristic search. It is a flowchart of the search parameter setting processing program concerning this embodiment. It is a figure explaining the extraction method of the representative point. It is the figure which showed the orthogonal table. It is the figure which showed the required number of the test patterns for calculating | requiring an optimal value. It is a flowchart of the sub process program of the parameter optimization process which concerns on this embodiment. It is the figure which showed an example of the delay tolerance, the error tolerance, and the error reliability tolerance which a user desires. It is the figure which showed distribution of the travel time of the whole link row | line | column which comprises an evaluation path | route. It is the figure which showed an example of the evaluation result of a test pattern.

  Hereinafter, an embodiment embodying a route search system according to the present invention will be described in detail with reference to the drawings. First, a schematic configuration of the route search system 1 according to the present embodiment will be described with reference to FIGS. 1 and 2. FIG. 1 is a schematic configuration diagram showing a route search system 1 according to the present embodiment. FIG. 2 is a block diagram showing the configuration of the route search system 1 according to the present embodiment.

  As shown in FIG. 1, the route search system 1 according to this embodiment basically includes a route search server 3 provided in a map information center 2 and information terminals 5 possessed by a plurality of users 4. Yes. The route search server 3 and the information terminal 5 are configured to be able to send and receive electronic data to and from each other via the communication network 6. Examples of the information terminal 5 include a mobile phone, a smartphone, a tablet terminal, a personal computer, and a navigation device. In addition, the user 4 may be in a state of getting on the vehicle, or may be in a state of not getting on the vehicle.

  Here, the route search server 3 performs route search on behalf of the information terminal 5. Specifically, when a destination is set in the information terminal 5, information necessary for route search such as a departure point and a destination is transmitted from the information terminal 5 to the route search server 3. On the other hand, the route search server 3 performs route search using the map information of the route search server 3 and specifies a recommended route from the departure point to the destination. Then, the specified recommended route is transmitted to the information terminal 5 that is the transmission source. Then, the information terminal 5 sets the received recommended route as a guide route, and performs movement guidance according to the guide route. Thereby, even if the map information stored in the information terminal 5 at the time of setting the guide route is an old version of the map information, an appropriate guide route is set based on the latest version of the map information that the route search server 3 has. It becomes possible to do. In particular, in the present embodiment, when searching for a recommended route, a future cost addition value is predicted using a heuristic function or the like, the search direction of the route (direction in which the route is extended) is limited, or priority is given to the search direction. The search is performed by heuristic search. Furthermore, parameters used for heuristic search are also set. Note that the route search may be performed not on the route search server 3 but on the information terminal 5 side.

  On the other hand, the information terminal 5 is an information terminal possessed by the user 4 and having a navigation function. For example, a mobile phone, a smartphone, a tablet terminal, a personal computer, a navigation device, or the like is applicable.

  Here, the navigation function displays the map image around the current position of the user 4 based on the map data acquired from the server or stored in the memory, or displays the current position of the user 4 in the displayed map image. Or a function for performing movement guidance along a set guidance route. Note that the information terminal 5 does not have to have all the navigation functions described above, and the present invention can be configured as long as it has a function of performing movement guidance along at least a set guidance route.

  The communication network 6 includes a large number of base stations arranged in various parts of the country and a communication company that manages and controls each base station. The base station and the communication company are connected to each other by wire (optical fiber, ISDN, etc.) or wirelessly. It is configured by connecting. Here, the base station has a transceiver (transmitter / receiver) for communicating with the information terminal 5 and an antenna. The base station performs wireless communication between the communication companies, and at the end of the communication network 6, the communication between the information terminals 5 in the range (cell) within which the radio waves of the base station can reach is transmitted to the route search server 3. Have a role to relay.

  Next, the configuration of the route search server 3 in the route search system 1 will be described in more detail with reference to FIG. As shown in FIG. 2, the route search server 3 includes a server control ECU 11, a map information DB 12 as information recording means connected to the server control ECU 11, a travel time DB 13, and a server side communication device 14.

  The server control ECU 11 (electronic control unit) is an electronic control unit that performs overall control of the route search server 3. The CPU 21 as an arithmetic device and a control device, and a working memory when the CPU 21 performs various arithmetic processes. In addition to the RAM 22 to be used, a control program, a ROM 23 in which a search parameter setting processing program (see FIG. 4) to be described later is recorded, a program read from the ROM 23, a flash memory 24 to store a later-described orthogonal table, etc. A storage device is provided. The server control ECU 11 has various means as processing algorithms together with an ECU of the information terminal 5 described later. For example, the orthogonal table creation means creates an orthogonal table in which a predetermined number of test patterns including combinations of a departure place, a destination, and a travel time are defined. The evaluation route search means searches for a route from the departure point to the destination at the traveling time specified in the test pattern by heuristic search for each test pattern specified in the orthogonal table. The route evaluation means evaluates an evaluation route that is each route searched for each test pattern. The optimum parameter specifying unit specifies the optimum value of the parameter used for the heuristic search based on the evaluation result of the route evaluation unit. The parameter setting means sets the optimum value of the parameter specified by the optimum parameter specifying means as a parameter used for heuristic search.

  The map information DB 12 is registered based on input data and input operations from the outside, and the latest map information used when searching for a recommended route in response to a request from the information terminal 5 is for each area (for example, information amount). Storage means that is divided and stored in a plurality of meshes having different levels. In this embodiment, the map information of the whole country is divided into a primary mesh with a side length of 80 km, a secondary mesh with a side length of 10 km, and a tertiary mesh with a side length of 1 km. And remembered.

  The map information stored in the map information DB 12 includes, for example, link data 31 related to links, node data 32 related to node points, search data 33 used for route search processing, facility data related to facilities, and a map for displaying a map. Display data, intersection data regarding each intersection, search data for searching for points, and the like are included. In addition, various information for searching for routes using moving means such as railways, buses, ships, airplanes, etc. (railway tracks, railway timetables, bus stops, bus routes, bus timetables, ship routes, ship routes, etc.) A timetable, an airplane route, an airplane timetable, etc.) may also be stored.

  Here, the link data 31 includes the width of the road to which the link belongs, the gradient, the cant, the bank, the road surface state, the number of road lanes, the number of lanes that decrease, and the width. The data representing the narrowing part, the railroad crossing, etc., the data regarding the corner, the radius of curvature, the intersection, the T-junction, the entrance and the exit of the corner, the data indicating the downhill road, the uphill road, etc. Regarding road types, in addition to general roads such as national roads, prefectural roads, and narrow streets, data representing toll roads such as national highways, urban highways, general toll roads, and toll bridges are recorded.

  Further, as the node data 32, the coordinates (positions) of node points set for each predetermined distance according to the radius of curvature, etc., on the actual road branch points (including intersections, T-junctions, etc.) and each road, Node attribute indicating whether a node is a node corresponding to an intersection, etc., a connection link number list that is a list of link numbers of links connected to the node, and an adjacency that is a list of node numbers of nodes adjacent to the node via the link Data relating to the node number list, the height (altitude) of each node point, and the like are recorded.

  The amount of information of the link data 31 and the node data 32 is different for each mesh level. That is, the widest primary mesh includes only information on main roads and facilities such as national roads and expressways, while the narrowest mesh, the tertiary mesh, includes information on detailed roads and facilities such as narrow streets. Is also included.

  On the other hand, as the search data 33, various data used for route search processing for searching for a route from a departure place (for example, the current position of the vehicle) to a set destination as described later are recorded. For example, a search cost such as a cost obtained by quantifying the appropriate degree as a route for an intersection (hereinafter referred to as an intersection cost) or a cost obtained by quantifying an appropriate degree as a route for a link constituting a road (hereinafter referred to as a link cost). Cost calculation data used for calculation is stored. Further, in the present embodiment, since the route search is performed by the heuristic search as described above, various data used for the heuristic search is also stored.

  Here, the heuristic search is a search method in which a future cost addition value is predicted to limit the search direction of the route (the direction in which the route is extended) or to give priority to the search direction. Specifically, in addition to a normal cost function used in the Dijkstra method or the like, a recommended route is specified from a route from the departure point to the destination in consideration of a heuristic function for predicting the cost to the destination. For example, as shown in FIG. 3, when searching for a recommended route from the departure point to the destination, any of the X direction, the Y direction, and the Z direction from the point A in a state where the route search process has advanced to the point A Is determined as a next search target using a heuristic function. That is, the predicted cost to the destination predicted by the heuristic function when traveling in the X direction, the predicted cost to the destination predicted by the heuristic function when traveling in the Y direction, and the traveling in the Z direction The predicted cost to the destination predicted by the heuristic function is compared, and the direction of the smallest predicted cost is set as the next search target. Note that only the direction of the smallest prediction cost may be set as the search target (that is, the search direction is limited to the direction with the low prediction cost), or the search is performed in other directions after the direction of the lowest prediction cost is set as the search target. It may be a target (that is, a search is performed with priority given to a direction with a lower predicted cost).

  In this embodiment, the heuristic function is configured such that a test pattern using an orthogonal table is performed a plurality of times, and an optimum value is set in units of secondary meshes from the test result. The details of the heuristic function setting method will be described later.

  The travel time DB 13 is registered based on input data and input operations from the outside, and for each link included in the map information DB 12, the travel time required to pass the link is the time zone, day of the week, etc. It is a memory | storage means divided and memorize | stored by. The travel time is calculated by statistics of probe information collected from each vehicle traveling throughout the country. Specifically, the average value and distribution of travel times of vehicles actually traveling on the target link are classified into time zones, days of the week, etc., and stored in the travel time DB 13.

  On the other hand, the server side communication device 14 is a communication device for communicating with the information terminal 5 through the communication network 6. In addition to the information terminal 5, reception of traffic information including information such as traffic jam information, regulation information, and traffic accident information transmitted from the Internet network or a traffic information center such as a VICS (registered trademark) center or a probe center. Is also possible.

  Next, a search parameter setting processing program executed in the route search server 3 in the route search system 1 having the above configuration will be described with reference to FIG. FIG. 4 is a flowchart of the search parameter setting processing program according to this embodiment. Here, the search parameter setting processing program is executed at a predetermined time interval (for example, every 24 hours) or when the map information DB 12 or the travel time DB 13 is updated, and is a parameter used for the heuristic search as a preparation step for performing a route search. It is a program to set up. The programs shown in the flowcharts of FIGS. 4 and 8 below are stored in the RAM 22 and the ROM 23 provided in the route search server 3 and are executed by the CPU 21. In addition, it is good also as a structure which not the route search server 3 but individual information terminal 5 performs the following processes.

  Here, in the search parameter setting processing program, parameters used for heuristic search are set in units of secondary meshes constituting the nationwide map information. Therefore, the following steps (hereinafter abbreviated as S) 1 and subsequent processes are repeatedly executed for each secondary mesh constituting the map information of the whole country. The parameters used for the heuristic search may be set in an area unit other than the secondary mesh unit (for example, an administrative division unit such as a municipality or a tertiary mesh unit). In that case, the following processing after S1 is executed for each set area unit.

  First, in S1, the CPU 21 selects representative points included in a total of 25 secondary meshes of 5 × 5 centered on the secondary mesh to be processed as shown in FIG. 5 from the map information stored in the map information DB 12. Extract. In this embodiment, one representative point is extracted for each tertiary mesh. Here, since there are a total of 100 tertiary meshes (1 km square) included in one secondary mesh (10 km square), a total of 2500 representative points are extracted. The representative point may be the center point of the tertiary mesh, or may be a facility such as a landmark included in the tertiary mesh. The representative point extracted in S1 is a candidate for setting a starting point or a destination in a test pattern described later.

  Next, in S2, the CPU 21 extracts a representative time. In this embodiment, “midnight (22:00 to 3 o'clock)”, “commuting (6 o'clock to 10 o'clock)”, “noon (11 o'clock to 16 o'clock)”, “retirement (17:00 to 21:00)” A total of eight time zones obtained by further dividing the four time zones into the first half and the second half are set as representative times. The representative time extracted in S2 is a travel time setting candidate in a test pattern described later.

  Subsequently, in S3, the CPU 21 creates an orthogonal table based on the representative point extracted in S1 and the representative time extracted in S2. Based on the created orthogonal table, a predetermined number of test patterns including combinations of the departure point, the destination, and the travel time are created. The orthogonal table may be created in advance on the apparatus side, and the created orthogonal table may be acquired from the flash memory 24 or the like in S3 to create a test pattern.

Here, FIG. 6 is a diagram showing an example of the orthogonal table created in S3. The orthogonal table defines the test conditions for each factor by the required number of test patterns. In the orthogonal table, the test conditions are defined so that there is no bias in the factors regardless of the number of test patterns.
For example, in the present embodiment, the following (A) to (H) are defined as factors.
(A) Column number of the tertiary mesh starting from the representative point (the secondary mesh is the secondary mesh to be processed)
(B) Row number of the tertiary mesh starting from the representative point (the secondary mesh is the secondary mesh to be processed)
(C) Secondary mesh column number with the representative point as the destination (D) Secondary mesh row number with the representative point as the destination (E) Tertiary mesh column number with the representative point as the destination (F ) Third mesh line number (G) travel time zone (“0: Midnight (22:00 to 3 o'clock)”, “1: Commuting (6 o'clock to 10 o'clock)”, “ 2: Daytime (11: 00-16: 00) ”,“ 3: Retirement (17: 00-21: 00) ”)
(H) Detailed classification of travel time zone ("0: first half", "1: second half")

  For (A) to (F), as shown in FIG. 5, the row number of the mesh located most northwest is 0, the column number is 0, the column number is increased in the east direction, and the row number is set in the south direction. increase. Therefore, for example, when A = 2, B = 2, C = 4, D = 0, E = 0, and F = 9 in the orthogonal table, the starting point is the second column and the second row included in the secondary mesh to be processed. It becomes the representative point of the tertiary mesh, and the destination is the representative point of the tertiary mesh in the 0th column and the 9th row included in the secondary mesh in the 4th column and the 0th row.

  In addition, the number of test patterns defined in the orthogonal table indicates that the evaluation result of the path searched by the heuristic search (hereinafter referred to as an evaluation path) as described later in the step of specifying the optimum value of the parameter used for the heuristic search. The number of test patterns required for convergence.

Specifically, the number of test patterns necessary for the evaluation result of the evaluation path to converge is calculated by the following method.
First, a heuristic function is randomly set to a predetermined value, and it is determined whether the evaluation path searched using the set heuristic function (which is unchanged during the test) satisfies both the allowable delay range and the allowable error range. To do. The allowable delay range is an allowable range (for example, 10 minutes) that is delayed with respect to the route that reaches the destination the earliest, and the allowable error range is an allowable range that causes an error in the predicted arrival time (for example, 15 minutes before and after). The allowable delay range and the allowable error range may be set as appropriate on the server side, or may be set based on a questionnaire result obtained in advance for grasping the user's preference. Then, it is determined for each test pattern whether or not the evaluation path satisfies the allowable delay range and the allowable error range, and the ratio of the evaluation path that satisfies both the allowable delay range and the allowable error range (hereinafter referred to as non-defective condition achievement). calculate. For example, the example shown in FIG. 7 shows the transition of the non-defective condition achievement with respect to the number of tests for each set value when the heuristic function is set to three randomly selected set values (for example, X1, X2, and X3). It is a figure. In the results shown in FIG. 7, when the number of tests exceeds 2000, the non-defective condition achievement degree converges to almost one value regardless of the heuristic function. That is, in the example shown in FIG. 7, it can be seen that the number of test patterns necessary for the evaluation result of the evaluation path to converge is 2000 times. Therefore, 2000 test patterns (combinations of numbers of factors (A) to (H)) are defined in the orthogonal table. Note that the number of test patterns specified in the orthogonal table may be more than 2000, but it is desirable that the number is as small as possible in order to reduce the processing load related to parameter optimization.

  Thereafter, in S4, the CPU 21 executes a parameter optimization process (FIG. 8) described later. In the parameter optimization process, the evaluation path is searched based on the test pattern created in S3, and the optimum value of the “parameter used for heuristic search” in which the evaluation result of the evaluation path is the highest using the optimization algorithm. Is the process of identifying The “parameter used for heuristic search” specified in S4 may be a heuristic function itself or a coefficient included in the heuristic function.

  Next, in S5, the CPU 21 sets the optimum value specified in S4 as a parameter used for heuristic search, and stores it in the map information DB 12.

  Thereafter, when the route search server 3 receives the route search request transmitted from the information terminal 5 on which the operation for performing the route search has been performed, the route search server 3 departs using the optimum parameter set in S5. Route search processing from the ground to the destination is performed by heuristic search. Here, in the heuristic search, in addition to the normal cost function used in the Dijkstra method etc., the future cost addition value is predicted using the heuristic function, thereby limiting the search direction of the route (direction of extending the route) In this search method, priority is set in the search direction. Then, route information regarding the specified recommended route is transmitted to the information terminal 5 that is the transmission source of the route search request. As a result, the recommended route is guided to the user via the display or the like in the information terminal 5. Further, a recommended route guided based on the subsequent user operation is set as a guide route for the navigation function, and movement guidance based on the set guide route is performed.

  Note that the route search server 3 may distribute the optimum value parameter set in S5 to the information terminal 5 so that the route search processing from the departure point to the destination is performed on the information terminal 5 side.

  Next, sub-processing of the parameter optimization processing executed in S4 will be described with reference to FIG. FIG. 8 is a flowchart of a sub-processing program for parameter optimization processing.

  In S11, the CPU 21 initializes a temperature parameter T used in the annealing method. The temperature parameter T indicates the degree of progress of the optimization process, and defines the probability of updating the parameter even when the score does not improve in the annealing method as will be described later. The initial value is, for example, 100 Degree (100% probability). The temperature parameter T is stored in a memory such as the RAM 22.

  Next, in S <b> 12, the CPU 21 defines “parameter P used for heuristic search” to be optimized, and substitutes an initial value for the value of parameter P. The parameter P may be a heuristic function itself or a coefficient included in the heuristic function. The type and number of parameters P can be set as appropriate.

  Subsequently, in S13, the CPU 21 randomly changes the parameter P defined in S12 to a new value. When there are a plurality of parameters P, the parameter P to be changed to a new value can be arbitrarily selected. The number of parameters to be changed may be one or two or more. Further, the value of the new parameter P after the change may be a value for increasing the current parameter P or a value for decreasing the current parameter P. Furthermore, it can be set as appropriate how much the current parameter P is displaced.

  Thereafter, the processing after S14 is executed using the value of the parameter P after the change at S13. However, at the time of the first execution, the process of S14 is executed by using the parameter P of the initial value substituted in S12 without executing the process of S13.

  The subsequent processes of S14 to S18 are executed for all the test patterns created in S3 (a combination of the departure point, the destination, and the time). For example, when the number of created test patterns is 2000 patterns, the process is sequentially executed up to pattern 2000 in the order of pattern 1, pattern 2, pattern 3,... (That is, the processes of S14 to S18 are repeated 2000 times). Will be). And after performing the process of S14-S18 for all the test patterns, it transfers to S19.

  First, in S14, the CPU 21 acquires a test pattern to be tested this time from the test pattern created in S3.

  Next, in S15, the CPU 21 sets a starting point and a destination which are search conditions for the current test based on the test pattern acquired in S14. Note that the test pattern created based on the orthogonal table defines the combinations of the numbers of the factors (A) to (H) as described above (see FIG. 6), and in particular, the factors (A) to (F). The starting point and the destination are set from the representative points extracted in S1.

  Subsequently, in S16, the CPU 21 sets a traveling time which is a search condition for the current test based on the test pattern acquired in S14. The test pattern created based on the orthogonal table defines the combinations of the numbers of the factors (A) to (H) as described above (see FIG. 6), and in particular the factors (G) and (H). The travel time is set from the representative time extracted in S2.

  Next, in S17, the CPU 21 uses the search conditions set in S15 and S16 and the parameter P defined in S13 to search for a route from the departure point to the destination at the travel time corresponding to the search condition. The route search processing to be performed is performed by heuristic search, and the searched recommended (optimum) route is specified as the evaluation route. Here, in the heuristic search, in addition to the normal cost function used in the Dijkstra method etc., the future cost addition value is predicted using the heuristic function, thereby limiting the search direction of the route (direction of extending the route) In this search method, priority is set in the search direction. For example, as shown in FIG. 3, when searching for a recommended route from the departure point to the destination, any of the X direction, the Y direction, and the Z direction from the point A in a state where the route search process has advanced to the point A Is determined as a next search target using a heuristic function. That is, the predicted cost to the destination predicted by the heuristic function when traveling in the X direction, the predicted cost to the destination predicted by the heuristic function when traveling in the Y direction, and the traveling in the Z direction The predicted cost to the destination predicted by the heuristic function is compared, and the direction of the smallest predicted cost is set as the next search target. Note that only the direction of the smallest prediction cost may be set as the search target (that is, the search direction is limited to the direction with the low prediction cost), or the search is performed in other directions after the direction of the lowest prediction cost is set as the search target. It may be a target (that is, a search is performed with priority given to a direction with a lower predicted cost). The heuristic function used for the search is specified by the parameter P. That is, if the parameter P changes, the heuristic function also changes, and a different route may be searched even under the same search condition.

  Subsequently, in S18, the CPU 21 determines whether or not the predetermined delay allowable range and error allowable range are satisfied for the evaluation route specified by the route search processing by the heuristic search in S17.

Hereinafter, the determination method of S18 will be described in more detail.
First, the route search server 3 acquires a questionnaire result for grasping a user's preference performed in advance. Here, as a questionnaire to be implemented, at least the following (A) to (C) are asked to the user.
(A) The range that is allowed to be delayed with respect to the route that reaches the destination earliest with respect to the route searched by the route search processing (that is, the extent to which the searched route is delayed relative to the fastest route) It indicates whether or not the user thinks that this is acceptable, and is hereinafter referred to as a delay tolerance).
(B) For a route searched by the route search process, a range that allows an error to occur in the predicted arrival time at the predicted destination (that is, the predicted arrival time in the searched route is the actual arrival time). This indicates how far the user thinks it is possible to deviate from the time, and is hereinafter referred to as an error tolerance.
(C) For a route searched by route search processing, a range that allows a probability that an arrival time at an actual destination is within an error allowable range (that is, an expected arrival time predicted in the searched route is error-allowable). This indicates the probability that the user wants to fit within the range, and is hereinafter referred to as an error reliability tolerance range).

  Here, FIG. 9 is a diagram showing an example of the acquired questionnaire results of the user. As shown in FIG. 9, the questionnaire to the user is divided into a plurality of parts based on the length of travel time. That is, when searching for a particularly short route with a travel time of about 10 minutes, searching for a route with a travel time of about 30 minutes, searching for a route with a travel time of about 60 minutes, and a travel time of 120 In the case of searching for a particularly long route of about a minute, answers of a delay tolerance range, an error tolerance range, and an error reliability tolerance range desired by the user are obtained. For example, in the example shown in FIG. 9, when searching for a route having a travel time of about 30 minutes, the route searched by the route search process is delayed if it is about 7 minutes with respect to the route that reaches the destination earliest. The user thinks that it is okay, and the user thinks that the estimated arrival time may be shifted if it is up to about 5 minutes from the actual time, and the estimated arrival time is within an allowable error range (that is, up to a delay of 5 minutes). ) Indicates that the user wants to be within a probability of 95% or more.

  Next, based on the acquired questionnaire result, the delay tolerance range, the error tolerance range, and the error reliability tolerance range that the user desires in the conditions (starting place, destination) of the current test pattern are specified. For example, if the route from the current departure point to the destination is a route with a travel time of around 30 minutes in the state where the questionnaire results shown in FIG. 9 are acquired, the delay allowable range is 7 minutes and the error allowable range is 5 The error reliability tolerance range is 95%. The allowable delay range, the allowable error range, and the allowable error reliability range are not acquired from the user questionnaire results, but are appropriately set on the route search server 3 side (for example, the allowable delay range is less than 60 minutes travel time). 10 minutes, error tolerance 5 minutes, error reliability tolerance 95%, travel time 60 minutes or more, delay tolerance 20 minutes, error tolerance 10 minutes, error reliability tolerance 90%) You may comprise so that it may acquire.

  Subsequently, the CPU 21 estimates a required time t1 required for the vehicle to travel on the evaluation route specified by the route search processing by the heuristic search in S17. Specifically, a total value obtained by adding the average values μ of travel times of all the links constituting the specified evaluation route is set as the required time t1. In addition, the average value of the travel time of each national link is calculated in advance based on the probe information and stored in the travel time DB 13. Next, the required time t2 required for the vehicle to travel on the route that reaches the earliest from the departure point to the destination (hereinafter referred to as the shortest route) is also estimated. Note that the shortest route can be specified by route search considering only that the total travel time of the link is shortened. Then, a total value obtained by adding the average travel times μ of all the links constituting the shortest route is defined as a required time t2. If the difference between the required time t1 and the required time t2 is within the allowable delay range, the CPU 21 determines that the allowable delay range is satisfied.

  On the other hand, the CPU 21 determines whether or not the error tolerance range is satisfied, that is, the probability within the error tolerance range, from the travel time distribution of the entire link train constituting the evaluation route specified by the route search processing by the heuristic search of S17. It is determined whether or not the probability that the variable is included is within the error reliability tolerance range. When it is determined that the error tolerance is satisfied with the probability within the error reliability tolerance, that is, the probability that the random variable within the error tolerance is included is within the error reliability tolerance, the error tolerance is satisfied. It is determined that Here, FIG. 10 is a diagram showing the travel time distribution of the entire link train constituting the evaluation route. The travel time distribution of the entire link train is obtained from the travel time distribution of each link constituting the link train by a known convolution integral. Further, the travel time distribution of each link nationwide is stored in the travel time DB 13. Even if the travel time distribution for each link is not a normal distribution, the travel time distribution of the entire link string basically converges to a normal distribution as shown in FIG. Therefore, the CPU 21 first specifies the upper limit value (μt + D) or lower limit value (μt−D) of the random variable included in the error reliability tolerance range (for example, 95%) centered on the average value μt. And, if the difference D between the specified upper limit value or lower limit value and the average value (if the normal distribution, the difference between the upper limit value and the lower limit value is the same value) is within the error tolerance range, It is determined that the allowable error range is satisfied. In addition, when the ratio of the random variable included in the error allowable range (μt−D to μt + D) is specified and the specified random variable ratio is in the error reliability allowable range, the error allowable range is satisfied. You may judge.

  Then, the processes of S14 to S18 are performed for each of the plurality of test patterns, and the determination result as to whether or not the preset allowable delay range and allowable error range are satisfied is stored in the memory for each test pattern. For example, FIG. 11 is a diagram showing an example of the test pattern evaluation result. In the example shown in FIG. 11, the patterns 1, 2, and 5 satisfy both the allowable delay range and the allowable error range. On the other hand, the pattern 3 does not satisfy both the allowable delay range and the allowable error range, and the pattern 4 satisfies the allowable error range but does not satisfy the allowable delay range.

  Thereafter, in S19, based on the determination result in S18, the CPU 21 scores the ratio of the test patterns that satisfy both the preset delay allowable range and error allowable range among all the test patterns created in S3. Calculate as For example, when 1270 patterns are satisfied among 2000 patterns, 63.5% is calculated as a score. The score may be a ratio of test patterns that satisfy at least one of the test patterns that satisfies both the allowable delay range and the allowable error range. The larger the score calculated in S19, the shorter the required time to the destination and the shorter the difference between the predicted arrival time and the actual arrival time, the more heuristic search is possible.

  Thereafter, in S20, the CPU 21 makes the score calculated in S19 larger than the score calculated before changing the parameter P in S13 (that is, the score calculated in the process of S19 executed last time). It is determined whether or not. When the evaluation path is specified at the first execution, that is, when the initial value of the parameter P is specified in S18, the process returns to S13 without performing the processes after S20.

  And when it determines with the score having become larger than the value calculated last time (S20: YES), it transfers to S21. On the other hand, when it is determined that the score is equal to or smaller than the previously calculated value (S20: NO), the process proceeds to S22.

  In S21, the CPU 21 recognizes that the value of the new parameter P changed in S13 is more appropriate than before the change, and updates the parameter P to the value after the change. Thereafter, the process proceeds to S23.

  On the other hand, in S22, the CPU 21 determines that the value of the new parameter P changed in S13 is more inappropriate than before the change. Then, the CPU 21 updates the parameter P to the changed value determined to be inappropriate with a probability based on the current value of the temperature parameter T. For example, in this embodiment, the temperature parameter T is changed in units of 1 degree from 100 degrees to 0 degrees. Then, the parameter of S22 is updated with a probability of 100% if it is 100 degrees and 50% if it is 50 degrees. If it is determined not to update, the value of the parameter P changed in S13 is returned to the value before the change. Thereafter, the process proceeds to S23.

  In S23, the CPU 21 reads the current temperature parameter T from the memory and updates it to a new value. In this embodiment, the initial value is set to 100 degrees, and the value is subtracted by 1 degree.

  Next, in S24, the CPU 21 reads the current temperature parameter T from the memory, and determines whether or not the temperature parameter T is 0 degrees.

  If it is determined that the temperature parameter T has become 0 degrees (S24: YES), the process proceeds to S25. On the other hand, when it is determined that the temperature parameter T is not 0 degree (S24: NO), the process returns to S13. Thereafter, the parameter P is randomly changed to a new value again, and the processes after S14 are executed. And as a result of repeating the process of S13-S23 until the temperature parameter T becomes 0 degree | times, the optimal value of the parameter P is finally specified. Note that the optimum value of the specified parameter P is that the route searched by the heuristic search in the route search process has a high score (evaluation), that is, the required time to the destination is short regardless of the search condition, and the arrival prediction This is a “parameter used for heuristic search” for a route with little difference between time and actual arrival time.

  Here, as described above, the number of test patterns created in S3 is the number of test patterns necessary for the evaluation result of the evaluation path to converge (FIG. 7). In addition, for each test pattern defined by the orthogonal table, various search conditions starting from the secondary mesh to be processed are defined without bias. Therefore, the parameter P finally specified by executing all the test patterns created in S3 is acceptable regardless of the search conditions such as the departure place, the destination, and the travel time. It becomes the optimum value.

  After that, in S25, the CPU 21 sets the parameter P of the optimum value finally specified as a “parameter used for heuristic search” as a result of executing the processing of S11 to S24. Then, the route search server 3 performs the subsequent route search processing using the parameter P of the optimum value set as described above.

  As described above in detail, in the route search system 1 according to the present embodiment, the route search method by the route search system 1 and the computer program executed by the route search system 1, the combination of the departure point, the destination, and the travel time is used. An orthogonal table defining a predetermined number of test patterns is created (S3), and for each test pattern specified in the orthogonal table, a heuristic search is made for a route from the departure point to the destination at the travel time specified in the test pattern. (S17), and the parameters used for heuristic search are set so that the evaluation path becomes the highest evaluation (S13 to S25). Therefore, the test pattern using the orthogonal table is implemented to perform the heuristic search. It is possible to set the parameters to be used to optimum values. As a result, an appropriate route can be searched by heuristic search regardless of the search conditions regardless of the combination of the departure point and the destination and the travel time.

Note that the present invention is not limited to the above-described embodiment, and various improvements and modifications can be made without departing from the scope of the present invention.
For example, in the present embodiment, the parameter P for calculating the optimum value is the heuristic function itself or a coefficient included in the heuristic function, but other parameters may be used as long as they are parameters used for heuristic search.

  In this embodiment, the annealing method is used as an algorithm for deriving the optimum value of the parameter P. However, other algorithms such as mountain climbing and tabu search may be used.

  In this embodiment, the delay allowable range, the error allowable range, and the error reliability allowable range are set based on the result of the questionnaire conducted for the user. Allowable range is 10 minutes, error tolerance range is 5 minutes, error reliability tolerance range is 95%, travel tolerance is 20 minutes for travel time of 60 minutes or more, error tolerance range is 10 minutes, error reliability tolerance range is 90% ) May be set.

  In the present embodiment, probe information is used as a method for obtaining the average value and distribution of travel times for each link in the whole country, but VICS information and other traffic information may be used. Furthermore, it may be calculated from the past travel history of the host vehicle that performs the route search.

  In this embodiment, the evaluation path evaluation method is configured to calculate, as a score, the ratio of test patterns that satisfy both the allowable delay range and the allowable error range. However, the evaluation path evaluation method is limited to the above method. There is nothing. For example, for each test pattern specified in the orthogonal table, the CPU 21 searches for a route from the departure point to the destination at the travel time specified in the test pattern by a normal Dijkstra method, and the searched route is a reference route. As specified. The reference route is an optimal route (correct route) for reaching the destination from the departure point at the travel time specified in the test pattern. Next, the CPU 21 calculates, as a score, the rate at which the evaluation path matches the reference path for each test pattern. Then, the parameter whose evaluation route is closest to the reference route (the parameter for which the highest score is calculated) may be specified as the optimum value of the parameter used for the heuristic search.

  In the present embodiment, the execution body of the search parameter setting processing program shown in FIG. 4 is the route search server 3, but the information terminal 5 may be configured to execute a part or all of it.

  Moreover, although the embodiment which actualized the route search system according to the present invention has been described above, the route search system can also have the following configuration, and in that case, the following effects can be obtained.

For example, the first configuration is as follows.
An orthogonal table creating means (21) for creating an orthogonal table in which a predetermined number of test patterns comprising a combination of a starting point, a destination, and a traveling time are defined, and a test pattern is defined for each test pattern defined in the orthogonal table. Evaluation route search means (21) for searching a route from the departure point to the destination at the travel time determined by heuristic search, and each route searched for each test pattern by the evaluation route search means. A route evaluation means (21) for evaluating an evaluation route, an optimum parameter specification means (21) for specifying an optimum value of a parameter used for the heuristic search based on an evaluation result of the route evaluation means, and the optimum parameter specification means The optimum value of the identified parameter is set as a parameter used for the heuristic search Has a parameter setting unit that, the.
According to the route search system having the above configuration, it is possible to set a parameter used for heuristic search to an optimum value by executing a test pattern using an orthogonal table. As a result, an appropriate route can be searched by heuristic search regardless of the search conditions regardless of the combination of the departure point and the destination and the travel time.

The second configuration is as follows.
The optimum parameter specifying means (21) uses an optimization algorithm to specify the parameter having the highest evaluation result of the evaluation path as the optimum value of the parameter used for the heuristic search.
According to the route search system having the above-described configuration, the parameters used for the heuristic search for the route searched by the heuristic search to be the highest evaluation route regardless of the combination of the departure place and the destination and the travel time, It can be set using an optimization algorithm.

The third configuration is as follows.
The predetermined number is the number of test patterns necessary for the evaluation result by the route evaluation means (21) to converge.
According to the route search system having the above configuration, since the number of test patterns is the number of test patterns necessary for the evaluation result of the evaluation route to converge, it is finally specified by executing all the test patterns. The parameters can be set to optimum values that are acceptable regardless of the search conditions such as the departure place, the destination, and the travel time.

The fourth configuration is as follows.
The parameter specified by the optimum parameter specifying means (21) is a heuristic function used for the heuristic search.
According to the route search system having the above configuration, it is possible to set a heuristic function to an optimum value by executing a test pattern using an orthogonal table. As a result, it is possible to accurately predict the cost to the destination by heuristic search.

The fifth configuration is as follows.
The parameter specified by the optimum parameter specifying means (21) is a parameter that determines the limit of the search direction or the priority of the search direction when searching for a route by the heuristic search.
According to the route search system having the above-described configuration, when performing a heuristic search, it is possible to more appropriately set a search direction limit or search direction priority. Therefore, it is possible to maintain the quality of the searched route while speeding up the route search process.

The sixth configuration is as follows.
The test pattern defines a starting point mesh that is a mesh including the starting point and a destination mesh that is a mesh including the destination. The starting point is a representative point of the starting point mesh, and the destination is A representative point of the destination mesh.
According to the route search system having the above configuration, the test pattern for the entire area can be set without excessively increasing the number of test patterns by defining the starting point and the destination in the test pattern in mesh units. Is possible. As a result, it is possible to set a parameter that enables a suitable route to be searched by a heuristic search regardless of how the departure point and the destination are set.

The seventh configuration is as follows.
Delay tolerance range setting means (21) for setting a tolerance range for delaying the route that reaches the destination earliest, and error tolerance range setting for setting an tolerance range in which an error occurs in the predicted arrival time. Means (21), wherein the path evaluation means (21) evaluates whether or not the evaluation path satisfies the allowable range of delay and the allowable range of error, and the optimum parameter specifying means (21 ) Specifies the parameter that the evaluation path satisfies the allowable range of the delay and the allowable range of the error as the optimum value of the parameter used for the heuristic search.
According to the route search system having the above configuration, in addition to the earlier arrival time at the destination, the priority is given to the fact that the error between the predicted arrival time and the actual value becomes smaller. It is possible to set parameters used for heuristic search for searching a route to the ground.

The eighth configuration is as follows.
For each test pattern specified in the orthogonal table, there is a reference route search means (21) for searching for a reference route from the departure place to the destination at the travel time specified in the test pattern, The route evaluation means (21) evaluates whether or not the evaluation route matches the reference route, and the optimum parameter specifying means (21) determines the parameter that the evaluation route is closest to the reference route as the heuristic. It is specified as the optimum value of the parameter used for search.
According to the route search system having the above-described configuration, the parameters used for the heuristic search for the route searched by the heuristic search to be closest to the reference route regardless of the combination of the departure point and the destination and the travel time are set. Is possible.

1 route search system 2 map information center 3 route search server 4 user 5 information terminal 11 server control ECU
21 CPU
22 RAM
23 ROM

Claims (10)

An orthogonal table creating means for creating an orthogonal table defining a predetermined number of test patterns comprising a combination of a starting point, a destination, and a traveling time;
For each test pattern specified in the orthogonal table, an evaluation route search means for searching a route from the departure place to the destination at the travel time specified in the test pattern by heuristic search;
Route evaluation means for evaluating an evaluation route which is each route searched for each test pattern by the evaluation route search means;
An optimum parameter specifying means for specifying an optimum value of a parameter used for the heuristic search based on an evaluation result of the route evaluation means;
And a parameter setting unit that sets an optimum value of the parameter specified by the optimal parameter specifying unit as a parameter used for the heuristic search.   2. The route search system according to claim 1, wherein the optimum parameter specifying unit specifies, as an optimum value of a parameter used for the heuristic search, a parameter having the highest evaluation result of the evaluation route using an optimization algorithm.   The route search system according to claim 1 or 2, wherein the predetermined number is a number of test patterns necessary for the evaluation result by the route evaluation means to converge.   The route search system according to any one of claims 1 to 3, wherein the parameter specified by the optimum parameter specifying means is a heuristic function used for the heuristic search.   5. The parameter specified by the optimum parameter specifying unit is a parameter for determining a limit of a search direction or a priority of a search direction when searching for a route by the heuristic search. Route search system. The test pattern defines a starting point mesh that is a mesh including the starting point and a destination mesh that is a mesh including the destination, respectively.
The starting point is a representative point of the starting point mesh,
The route search system according to claim 1, wherein the destination is a representative point of the destination mesh. An allowable delay range setting means for setting an allowable range of delay for the route that reaches the destination earliest;
Error tolerance range setting means for setting an tolerance range in which an error occurs in the predicted arrival time to the destination, and
The path evaluation means evaluates whether the evaluation path satisfies the delay tolerance and the error tolerance;
7. The optimum parameter specifying unit specifies a parameter that the evaluation path satisfies most of the allowable range of delay and the allowable range of error as an optimal value of a parameter used for the heuristic search. The route search system described in 1. For each test pattern specified in the orthogonal table, there is a reference route search means for searching a reference route from the departure place to the destination at the travel time specified in the test pattern,
The route evaluation means evaluates whether the evaluation route matches the reference route,
The route search system according to any one of claims 1 to 6, wherein the optimum parameter specifying unit specifies a parameter that the evaluation route is closest to the reference route as an optimum value of a parameter used for the heuristic search. An orthogonal table creating means creating an orthogonal table defining a predetermined number of test patterns comprising a combination of a starting point, a destination, and a running time;
An evaluation route search means, for each test pattern specified in the orthogonal table, searching for a route from the starting point to the destination at the travel time specified in the test pattern by heuristic search;
A path evaluation unit that evaluates an evaluation path that is each path searched for for each test pattern by the evaluation path search unit;
An optimal parameter specifying unit specifying an optimal value of a parameter used for the heuristic search based on an evaluation result of the route evaluation unit;
A route searching method comprising: a parameter setting unit that sets an optimum value of the parameter specified by the optimum parameter specifying unit as a parameter used for the heuristic search. Computer
An orthogonal table creating means for creating an orthogonal table defining a predetermined number of test patterns comprising a combination of a starting point, a destination, and a traveling time;
For each test pattern specified in the orthogonal table, an evaluation route search means for searching a route from the departure place to the destination at the travel time specified in the test pattern by heuristic search;
Route evaluation means for evaluating an evaluation route which is each route searched for each test pattern by the evaluation route search means;
An optimum parameter specifying means for specifying an optimum value of a parameter used for the heuristic search based on an evaluation result of the route evaluation means;
Parameter setting means for setting an optimum value of the parameter specified by the optimum parameter specifying means as a parameter used for the heuristic search;
Computer program to make it function.

Images