JPH02105972A - Route search device - Google Patents

Abstract

PURPOSE:To shorten the search time of the shortest route by utilizing an actual distance which is stored corresponding to a tip node to calculate the evaluated value of route searching. CONSTITUTION:It is evident from a route where the shortest routes from two different start point nodes A and A' to an end node K is searched for as shown in a figure that both the shortest routes pass a node C and passage nodes following the node C are exactly the same. A search is made temporarily and then when a next search is made, a search following a search starting at the node A' is made by utilizing the route from the start point node A to the end point node K after the former search reaches the node C. Consequently, a route search can be omitted and plural target search routes can be searched for at a high speed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Purpose of the Invention (Industrial Application Field) The present invention provides a method for connecting a plurality of starting nodes to a terminal node in a network consisting of a plurality of nodes and a plurality of buses connecting these nodes. The present invention relates to a route search device that searches for the shortest route to .

(Prior Art) The problem of searching for the shortest route between two nodes on a network consisting of a plurality of nodes and a plurality of buses connecting these nodes often appears in our daily life. For example, the problem of searching for a road between two specific points on a map corresponds to this problem, where nodes correspond to intersections and paths correspond to roads between intersections. There have already been various studies on algorithms for finding the shortest route between two different nodes on a network, and the A8 algorithm is known as an efficient algorithm (for example, "Artificial Intelligence" by Wlnston, translated by Ike Nagao,
Published by Baifukan, see Section 4.1, “Basic Search”).

However, in conventional route searching devices, the starting point and ending point are fixed and cannot be changed during the search. This is a limitation in actual search services. for example,
If you depart from Tokyo and want to take a hot spring bath in the Niigata area, would you limit your travel plans to the only hot spring in Niigata prefecture? There may be cases like that, but usually it is reasonable in terms of budget and time, taking into account the schedule of traveling further beyond Niigata.
Moreover, you will be looking for a hot spring value that has a certain level. However, the algorithms of conventional route search mechanisms are powerless against such search problems that target "multiple targets."

For search problems with multiple targets, there are methods that can be relatively easily deduced from conventional techniques by those skilled in the art. it is,
For example, select all hot spring resorts that correspond to "hot spring resorts in the Niigata area" as starting point node candidates, repeat route searches between each starting point and the ending node, and select the shortest route from each of the obtained routes. It's a method. however,
In this method, the time required for route search is multiplied by the number of starting point node candidates, making it impractical.

Hereinafter, a multi-target route search device using the A1 algorithm, which is a basic search algorithm, and the A'' algorithm, which can be analogized to by those skilled in the art mentioned above, will be described.

First, the A'' algorithm will be explained with reference to Fig. 4. In Fig. 4, the circles indicated by 1 indicate nodes;
A and B shown in the vicinity of this node 1.

C and the like are node names, and correspond to, for example, cities in the automobile road map network. Further, the line connecting each node 1 indicated by 2 is a bus, and corresponds to, for example, a road in an automobile road map network. The name of bus 2 is expressed by arranging the names of the nodes on both sides of the bus. Therefore, the name of the bus between node A and node B is rABJ. Furthermore,
In order to show the distance between nodes, we assume a coordinate axis with the goal node K as the origin and node A at (-6, 33), which allows us to calculate the straight line distance from each node to the goal node K. is written near each node, and the length of each bus is written near the bus. In the figure, for example, node A is at a distance of 33.5 from node K, and the length of bus AB is 9.2.

In order to find the shortest route between two different nodes on a network, it is necessary to minimize the sum of all path lengths that make up this route. The shortest route on the network shown in FIG. 4 is determined as described below. however,
Here, as an example, we will consider the case of finding a route from node A to node K, and the search will start from node A. The route search basically extends the route from node A, and if there is a branch depending on the node, the route is extended by making multiple branches in accordance with the branch.

The tip node of the currently extended route will be referred to as the "frontier." During an exploration, there are usually multiple frontiers due to branching.

If the frontier has reached the terminal node K, and the extension distance of all other routes that have not reached the terminal is greater than the route that has reached the terminal, that is, if the frontier has not reached the terminal node A8 evaluation value of all routes (- "distance from the starting point node to the relevant node" + "straight line distance between the relevant node and the end, end node") evaluation value of the route that has reached the end of the force meter If it is larger than , the search ends and the route that has reached the terminal is the shortest bus sought.

In the above-described method, the amount of search processing is greatly influenced by the method of calculating the evaluation value for determining from which frontier the route should be extended, that is, the quality of the evaluation function. The A8 algorithm is known as the most excellent search algorithm, and this A'' algorithm uses the evaluation value shown in the following equation as the evaluation value of the route arriving at each frontier.

Evaluation value - "Distance via route from the starting point node to the relevant frontier (actual distance)" (10) Straight-line distance from the relevant frontier to the terminal end] Further, details of the A'' algorithm will be described below.

(1) Among all routes each having a frontier, select the route with the minimum evaluation value, let the length of this minimum route be R, and let this frontier be F.

(2) Branch and extend the route for each bus connected to node F, and add the extended light node to the frontier set.

Delete old Frontier F from the Frontier set. The length of each new route is the length of each bus plus the minimum route length R.

Also, the evaluation value is recalculated for each new frontier.

If the frontier matches the terminal node, it is included in the frontier set, but not included in the subsequent extension, and its evaluation value is stored as the evaluation value of the provisional solution (route).

(3) If there is no route that reaches the terminal node,
Return to (1). If not, proceed to (4).

(4) If the evaluation values of all other routes that have not reached the terminal node are larger than the evaluation values of the routes that have reached the terminal node, proceed to (5). If not, return to (1).

(5) End.

Note that if the route being searched merges with another route, the route with the smaller evaluation value is given priority, and the route with the larger evaluation value is discontinued. Figure 5 shows the results of actually searching for a route from node A to node K using the A'' algorithm.The thick solid line indicates the shortest route obtained as a result, and the bus shown as a solid line was expanded during the search. indicates a rejected bus, and a dashed line indicates a bus that was not expanded.

(Problem to be Solved by the Invention) In FIG. 5, the starting point node and the ending point node that are the targets of the route search are each specified as one, but the starting point node is not limited to a specific node A. Node A
There are three nodes A, A'' and A' close to each other, and a search may be requested as "around node A". In this case, it is not possible to set only the specific node A as a starting point candidate. In this case, to find the shortest route between "around node A" and the node, node A, A-A
It is necessary to perform a route search from each node of P to node K, and select the route having the shortest distance among the respective shortest routes obtained as shown in FIGS. 5 to 7, for example, as the shortest route. In such a case, with conventional route search devices, it is necessary to search for each of multiple nodes A, A-, A-- selected as being within the range specified by "around node A". The time required for this increases as the number of nodes increases, and there is a problem in that it is extremely time consuming.

The present invention has been made in view of the above, and an object thereof is to provide a route search device that quickly searches for the shortest route when there are a plurality of starting point nodes or ending point nodes.

[Structure of the Invention] (Means for Solving the Problems) In order to achieve the above-mentioned object, the route search device of the present invention uses a network consisting of a plurality of nodes and a plurality of buses connecting these nodes. connection information storage means for storing connection information; route search means for searching for a route from a first starting point node to a terminal node on the network; and a second starting point node located at a different position from the first starting point node. actual distance storage means for storing for each node an actual distance from each node belonging to the shortest route from When searching for the shortest route, when the tip node of the route being searched matches a node stored in the actual distance storage means, the actual distance stored in the actual distance storage means corresponding to the node is calculated. The gist of the present invention is to have an evaluation means used for calculating an evaluation value of route search.

(Operation) In the route search device of the present invention, the actual distance from each node to the terminal node belonging to the shortest route from the second starting point node located at a different position from the first starting point node to the terminal node is stored for each node. However, in the search for the shortest route from the first starting point node to the terminal node, if the tip note matches the stored node, the actual distance stored corresponding to the node is used to evaluate the route search. It is used for value calculation.

(Example) Examples of the present invention will be specifically described below.

The present invention searches for a plurality of routes from a plurality of different start-point nodes to a destination node such that the start-point node or the end-point node is not clearly specified, but is specified with "ambiguity" such as "around node A". , when searching for the shortest route among these multiple routes, 2
The feature is that the previous search results are used for the second and subsequent searches, thereby shortening the search time for the second and subsequent searches. As shown in Figures 5 and 6 above, there are two different
As can be seen from the search for the shortest routes from the two starting point nodes A and - to the ending point node K, both of the shortest routes pass through node C, and the nodes passed after this node C are exactly the same. Once the search corresponding to Figure 5 is executed,
When executing the search in FIG. 6, once the search starting from node A' reaches node C, subsequent searches can be omitted by using the route from end node A to end node K. , eliminate waste of search,
It is thought that a search route for multiple targets can be searched at high speed.

Therefore, first, the A8 algorithm will be specifically explained in order to look at the structure for storing route information in the first search.

In the A1 algorithm, the following two preprocessing steps are required before starting the search.

(1) Input the starting point node and the ending point node to the route search mechanism.

(2) Calculate the straight-line distance from all nodes to the end node.

The length of each path is calculated in advance by the system and stored as network information. However, the straight-line distance from all the nodes to the end node mentioned above is calculated using the coordinate values of the frontier node and In some cases, it may be more efficient to calculate the straight-line distance from the coordinates of the target node.

In the A8 algorithm, information on the route that has reached a certain frontier is held, for example, as follows. In this example, the data structure is based on the computer language LISP 8.
It is expressed as an expression, that is, as a list structure. For example, (a b c) is a list consisting of elements a, b, , c.

This is because search systems are often constructed using LISP. In addition, to understand this example, LIS
No understanding of the P language specification is required.

The following is an example of a list used for searching.

((14,435,5)(E CA)) This list is a list consisting of two lists (this list is usually called a list of lists), and the root starting from node A is the node From A to node E via node C, the current frontier node is E, the route distance from the starting point node A is 14.4, and the straight line distance between the frontier node E and the terminal node is 2.
By adding 1.1 (see FIG. 4) to the above-mentioned path, it means that the evaluation value of A'' is 35.5.

Here, when the path is expanded from node E and the route is further expanded to node G, the data held above is changed as follows. This becomes the detailed information of the route from A with node G as the frontier.

((20,839,2)(G E CA)) FIG. 9 shows the full search trace results when the starting point node of the search is A and the terminal node is K. The search starting from node A extends to A'', B, and C, but among these, node C with the smallest evaluation value is the next expansion target (step 1゜2).From node C are A-, A=, E
, F can extend the route in four directions. The evaluation value for each of them is indicated by an arrow (steps 2 to 3).

However, for node A-, the path is already A.

Since the route has arrived from A-, the route that came from CA- has a large evaluation value and is immediately deleted.

Next, in step 3, the node B with the smallest evaluation value becomes the target of route expansion. Thereafter, similar processing is performed to finally arrive at the target node in step 9. The total number of steps is 9.

In contrast to the search using the A'' algorithm described above, in the present invention, when performing a route search, information on the shortest route from another starting point node to the terminal is actively utilized.Eighth
The figure is a reproduction of the network shown in FIG. 4, but after the shortest route from node A to node K has already been searched. That is, the transit nodes on the shortest route from node A to node K, that is, nodes A, C, F, and I.

Regarding J, the "actual distance" when actually passing through the node is stored, not the straight line distance from the node, and in Fig. 8, this real distance is shown with double underlining. .

As shown in FIG. 8, node A stores the actual distance.

For C, F, I, and J, the route extension from the starting point node to the terminal node is determined when the frontier of the route search arrives at the node.

Therefore, for a node that has memorized the actual distance, not only is route expansion after that node unnecessary, but if the A゜evaluation value of another route being expanded is greater than the finalized value, the finalized evaluation The route with the value is guaranteed to be the final shortest route.

FIGS. 10(a) and 10(b) are based on the network information in FIG. 8, and are based on the network information shown in FIG. 8. - Shows the trace results of searching.

In FIG. 10(a), the search starts from the starting point node A', but the evaluation of the route leading to nodes A and C is determined at node C, and the route finally reaches the goal, that is, the ending point node K. The total distance will be confirmed upon arrival. Therefore, no further searches need to be performed for nodes A and C.

On the other hand, regarding the route extended to node A'', the information stored in node A'- is not the actual distance to node K, but the estimated distance, which is actually a straight line distance, so this route itself is The search cannot be completed. However, when comparing the evaluation values of the three frontier nodes, the evaluation value of the already determined node C is the smallest, so there is no need to execute the search any further. It can be seen that the shortest route from node A' to node K starts from node A' and reaches node C, and thereafter matches the shortest route from node A to node K. The search from node A- requires nine steps, but the search from node A- requires only two steps.

On the other hand, FIG. 10(b) shows the result of searching for the shortest route from the starting point node A= to the terminal node K. In this case, the effect of the present invention cannot necessarily be said to be sufficient because the node A' is located quite far from the starting point node A of the already searched route. However, despite this, this search process is completed in 6 steps. As described above, according to the present invention, when searching from a plurality of starting point nodes, the search time for the second and subsequent nodes can be significantly reduced.

FIG. 1 is a block diagram showing the configuration of a route searching device according to an embodiment of the present invention.

The route search device 10 shown in FIG. 1 is comprised of a working memory section 11 that stores network information and a search tube 12. The working memory section 11 and the working memory section 11 are connected via a working memory access means 40 and a data transfer means 41, and the working memory access means 40 is a search section]2 that executes data reading from the working memory section 11. The data transfer means 41 forms a path for transferring the resulting data from the working memory section 11 to the search section 12.

The working memory section 11 stores network connection information using a storage structure called "r frame". A frame is a means for storing a certain object (substance) and data related to it as a unit, and this frame is indicated by reference numeral 21.31 in FIG.

The frame 21 is a frame in which a node and information related to this node are grouped together. For example, the frame on the left side has the name "C" and is a means for expressing information regarding the node "C". Frames have what are called "slots" for storing related information, specifically, in the case of frame rcJ, rpath (path)
", r disjan (distance)", "route (
There are four slots: "root)" and "real (real or true)", and their "slot values" are rACCF.
CE A-Cl, r27. IJ, rN I LJ
, r ((8°9 36)(CA=”)J. This shows that j pat coming out from node C
h (path)" has paths named rAC, CF, CE, A-, and CJ, and furthermore, the distance from the origin (
distance) is 27.1, and the route information used in the A'' algorithm is currently ((8,93
6) (CA PA)). The slot "realJ" is a slot that is particularly characteristic of the present invention. That is, this slot is the slot "realJ".
This is a slot for displaying that the distance to the goal stored in aneeJ is not an estimated distance but an "actual distance." Here, if it is "estimated distance", r
If N I LJ is displayed and it is "actual distance",
ITJ is displayed. In FIG. 8, for nodes on the shortest route from node A to node K, this rre
The value of the alJ slot becomes rTJ.

Broadly speaking, the network information stored in the working memory unit 11 includes two types of frames: a frame group 21 that stores node information, and a frame 31 that stores path information. The frame group 21 that stores node information is said to form one "class" as a whole, and the name of the "class" is "node." A frame 20 represents this class, and a pointer means 22 is provided between the class frame 20 and the concrete frame 21.
is provided so that the concrete frame can be traced from the class frame. The same applies to the path information, and the frames 31 that store the path information collectively form the class "path", and a pointer means 32 is provided between the frame 30 representing the class and the concrete frame 31. In each slot of the pass frame 31,
Information about the length of each path and the name of the node to which the path is connected are stored.

Next, the search unit 12 will be explained. The search unit 12 includes a frontier list storage means 53, a frontier selection means 52, a route development means 50, and a control means 54 for controlling these. 55, 56.58.61
are data transfer means, respectively. Further, 59 is a control means activation signal, and 60 is control information.

Each component will be explained.

The frontier list storage means 53 stores the starting point as a frontier name list without making any particular distinction. Here, the following list is stored.

(A-CE G) This frontier list is changed as appropriate as the search progresses. For example, if H is added as a new frontier as a result of expanding G with (A″ CEG), removing G to be expanded and adding H to the list will result in a new list as shown below. Become.

(A″ CE H) The frontier list storage means 53 stores the new frontiers (usually there are multiple) sent from the data transfer means 55 and the new frontiers sent from the frontier selection means 52 via the data transfer means 61. It has a function to modify the frontier list as described above based on the frontier information to be developed.

The frontier selection means 52 reads out the frontier list stored in the frontier list storage means 53 via the data transfer means 58, and reads out the evaluation value of each frontier node written in this frontier list. The working memory section 11 is accessed by the access means 40. and,
For each starting point, select the frontier node with the smallest evaluation value. Therefore, the resulting number of frontier nodes is one.

The route expansion means 50 receives the node name to be expanded from the frontier selection means 52 using the data transfer means 56. Then, all routes arriving at this node are expanded, and the new frontier obtained as a result is transferred to the frontier list storage means 53 by the transfer means 55.
to be notified.

By using the frontier list storage means 53, frontier selection means 52, and route expansion means 50 described above, one frontier node to be expanded is selected from the frontier list for each starting point, and information related to this selected node ( That is, the name of the node that can be reached by one path from the node, the length of each bus, the straight line distance from those nodes to the terminal node, etc.) are read from the working memory unit 11, the route is developed, and a new An operation can be performed to write route information to a flotier node. However, what must be noted here is that if the node that stores the "actual distance" is a frontier node, no further expansion is required.

The control means 54 holds the starting point and end point for the search, and sequentially operates the frontier list storage means 53, the frontier selection means 52, and the route expansion means 50 to control the entire search.

Next, a more detailed configuration of each element constituting the embodiment shown in FIG. 1 will be described.

First, FIG. 2(a) is a block diagram showing the configuration of the working memory section 11. As shown in FIG. As shown in the figure, the working memory section 11 includes a storage cell plane 104 and an address decoder 10 for energizing the storage cell plane 104.
3. Address register 100, read data register 101 that stores information read from memory cells, write data register 10 that stores information written to memory cells.
2, and a control section 111 that performs overall control. In the figure, 106 is an address information transfer section, 105 is a cell activation path, 107 is a write information transfer path,
108 is a read information transfer path, 109 is an address information transfer path, and 110 is a control signal transfer path.

Frame information on the working memory is stored in the X memory cell brain 104 in FIG. 2(a);
), the frame name and address are associated in advance by the compiler. This correspondence is possible using a general compilation method. In Figure 2(b), the frame (its name is "C")
corresponds to address 0, and the symbol name rCJ of that frame is stored in the storage cell at address 0,
By reading cells one after another in the direction of address addition from the address, slot r dlstanJ (actually stores address 12 of the compilation result), slot r
There is a real J, and its value is ``27゜1''.
, rNl, and LJ. In addition, r/E
NDJ indicates that the slot sequence ends there.

The operation shown in FIG. 2 will be explained as an operation of reading out a certain slot. However, it should be noted that the read frame is an rCJ frame with address 0, and the slot name is handled in advance as the address of the compilation result (assuming address 13) within the system. First, the control section 11
1, a control signal 110 is sent to the slot value readout operation. Read address 0 is first placed in address register 100. Also, the address corresponding to the slot name is stored in the read data register 101 for comparison.
is stored in The address register 100 adds 1 to this address and sends this information to the address decoder 103.
Light water is supplied to the water via the transfer path 106. Address decoder 1
03 energizes this No. 1 memory cell and reads cell information into the read data register 101. Here, the content of this read data register 101 is address number 12, not address number 13 as expected. Therefore, in the address register 100, an address is created by further adding 2. When this address No. 3 is accessed, the target address 13 is certainly stored, so the address register 100 is further incremented by 1 and the target information (N I L) is read from the address No. 4 in the data register 101.
Transfer to. The read operation of the slot value is as explained above, but the write operation is also similar, and after setting the write data in the write data register 102, the data is transferred to the memory cell with the target slot value activated. The only difference is that the data is transferred.

FIG. 3 is a block diagram showing the detailed configuration of the search unit 12. The search unit 12 is comprised of the frontier list storage means 53, the frontier selection means 52, the route development means 50, and the control means 54, as described above.

The frontier list storage means 53 includes a register 41 that stores names of newly added/deleted frontier nodes.
0, a memory 414 for storing the frontier list, a marsher 412 for merging the contents of the additional frontier list node and the frontier list, a duplication checker 413 for checking and removing duplication of frontier nodes, and a read for reading and holding the contents of the frontier list. It is composed of a data buffer 411. The contents of the frontier list are transferred by transfer path 515 to marsher 412, where new frontier nodes are added and old frontier nodes are removed based on the data sent from transfer path 514. The merge result is transfer path 51
6 to the duplicate checker 413, where duplicates of frontier nodes are removed. The new frontier list is stored again in the frontier list storage memory 414 via the transfer path 517. The contents of frontier list storage memory 414 are sent to read data buffer 411 via transfer path 511 and used by the shortest route detection mechanism.

The frontier selection means 52 accesses frames corresponding to each node in the frontier list via the transfer path 508, and reads out all the routes being developed.
05, a route memory 406 that stores read route information, a sorter 407 that sorts the contents of the route memory based on evaluation values, and a register 409 that stores expanded routes. The route information read from the working memory is sent to the access mechanism 4 via the transfer path 507.
05 and is input to the route memory 4 via the transfer path 509.
06. From among these, the sorter 407 uses the transfer path 510 to find the route with the smallest evaluation value. The result is stored in the expansion route register 409 via the transfer path 408.

The route development means 50 reads the name of the frontier node to be developed from the route information to be developed, accesses the working memory, and reads out the bus connected to the node, and a register that stores the read bus information. 403, an expander 401 that performs route expansion, and a write data register 400 that writes the generated route into a working memory. Note that the newly generated frontier node name is transferred to the frontier list storage means 53 via the frontier node modification instruction bus 55. Access mechanism 40
4 specifies the node name and accesses the working memory section 11 via the transfer path 504. As a result, the bus information is read from the slot via the transfer path 503 and stored in the bus information register 403 via the transfer path 506. The route expander 401 reads out this bus information via the transfer path 505, generates new route information, and supplies it to the write data register 400 via the transfer path 501. The contents of write data register 400 are transfer path 5
02 to the working memory section 11.

It should be noted that the present invention is not limited to the embodiments described above, and it goes without saying that it can be constructed using, for example, a memory and a microcomputer.

[Effects of the Invention] As explained above, according to the present invention, the actual distance from each node belonging to the shortest route from the second starting point node located at a different position from the first starting point node to the ending node to the ending node. is stored for each node, and in the search for the shortest route from the first starting point node to the terminal node, if the tip node matches the stored node, the implementation stored corresponding to the node is stored. Since the distance is used to calculate the evaluation value of the route search, the second and subsequent duplicate searches can be omitted, and the time required to search for the shortest route can be shortened.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a route search device according to an embodiment of the present invention, FIG. 2 is a block diagram showing the configuration of a working memory section used in the route search device of FIG. 1, and FIG. The figure is a block diagram showing the configuration of the search unit used in the route search device of FIG. 1, and FIG. 4 is a block diagram showing an example of a network consisting of multiple nodes and multiple buses connecting these nodes. Figures 5 to 7 are diagrams showing route search results from different starting points in the network of Figure 4, Figure 8 is a diagram showing network information based on the spirit of the present invention, and Figure 9 is a diagram showing search trace results. FIG. 10 is a diagram showing search trace results of the present invention. 10...Route search device 11...Working memory unit 12...Search unit 21.31...Frame 5
0...Route expansion means 52...Frontier selection means 53...Frontier list storage means 54...Control means

Claims (1)

[Claims] connection information storage means for storing connection information of a network consisting of a plurality of nodes and a plurality of paths connecting these nodes; and a route for searching for a route from a first starting point node to a terminal node on the network. a search means; and an actual distance memory for storing, for each node, an actual distance from each node belonging to the shortest route from a second starting point node located at a different position from the first starting point node to the aforementioned terminal node. and when the route search means searches for the shortest route from the first starting point node to the terminal node, the tip node of the route being searched matches the node stored in the actual distance storage means. and evaluation means for using the actual distance stored in the actual distance storage means corresponding to the node in calculating the evaluation value of the route search.