JP5178645B2 - Multi-route search apparatus and method - Google Patents

Description

  The present invention relates to an apparatus and a method for searching for a route that satisfies a constraint when providing a plurality of routes in a space.

  The problem of searching for a route connecting two points by giving a start point and an end point in a two-dimensional plane or three-dimensional space is a well-known problem as the shortest route problem, and there are many applications in the industry. . For example, in the case of a two-dimensional plane, application to a car navigation system is well known. In the case of a three-dimensional space, there are application examples to LSI (Large Scale Integrated Circuit) wiring design, piping for electric power generation plants and general industrial plants, and design of electric cables.

  In either case, a constraint condition or an index (evaluation index) for obtaining a desired route is set. Although the constraint condition varies depending on the application target, in many cases, a portion that is not allowed to pass through a route (prohibited area) in a given two-dimensional plane or three-dimensional space is representative. In many cases, the evaluation index is set so that the length of the route is the shortest or the cost is minimum. As the size of the two-dimensional plane or three-dimensional space to be searched for routes increases and the number of routes that must be determined increases, the number of combinations increases and it becomes difficult to find a solution to the shortest route problem. .

  Conventionally, there are a backtrack method and a breadth-first search as a method for solving the shortest path problem in a two-dimensional plane and a three-dimensional space. The backtracking method is a trial-and-error search method in which a route is selected from a given starting point, and when it gets stuck, it goes back and selects another route. Therefore, there is no guarantee that the first route found is the shortest route. In the breadth-first search, search is advanced in parallel for all routes, so the first route found becomes the shortest route. In both methods, since the route from the start point to the end point is obtained one by one, the calculation amount increases.

  Therefore, a search method based on the principle of optimality proposed by Bellman has been devised. The principle of optimality is that the partial solution of the optimal solution of the overall problem matches the optimal solution of the subproblem. Dynamic programming based on this is a technique in which the target problem is divided into a plurality of subproblems and the solutions of the subproblems are stored so that the same subproblem is not repeatedly solved. When analyzing the shortest path problem with a computer, if the route up to a certain point in the middle is saved, if the saved route is used when searching for the route including that point, the calculation up to that point becomes unnecessary. , Can save the calculation amount.

  Another method using the principle of optimality is the Dijkstra method. The Dijkstra method is a method mainly used in graph theory. A graph is composed of a plurality of points and paths connecting them. In the Dijkstra method, when the shortest route is obtained as the optimum route, a point having the shortest route length among points that can be reached by one route is determined as the shortest route from the points having the shortest route length. It is.

  When obtaining a plurality of routes, the starting point and the ending point of the first route are given, and the route is determined using the above-described method. Next, the start point and end point of the second route are given, and the route is obtained in the same manner. By repeating this procedure for the required number of routes, a plurality of routes are obtained. However, in these cases, there is only one type of route, and in order to obtain a combination of a number of different routes, the above-described procedure must be repeated while changing the constraint conditions given to each route. Since conditions are set manually by the user, it is difficult to obtain a comprehensive combination of routes.

  Therefore, there is a method of obtaining a plurality of routes by parameterizing some of the constraint conditions and using a genetic algorithm.

  The genetic algorithm was described in 1975, J.A. Originated in the study of adaptation systems by Holland. This is a technique in which the framework of the algorithm is systematized by Goldberg. Specifically, by using an individual that expresses the solution of the problem as a symbol string and simulating the genetic mechanism and natural selection of an organism, optimization is performed in a discrete search space by a plurality of individuals. In addition, a new search point is generated by a crossover operation in which genetic information is exchanged between individuals. In order to determine the superiority or inferiority of each individual, an index called fitness function is set. The fitness function is usually the same as the evaluation index in the optimization problem. Unlike other methods, the genetic algorithm can be searched by a plurality of individuals, so there are many application examples to combination problems and scheduling problems. Non-patent document 1 describes systematized algorithms and simple examples. Non-patent document 2 describes application examples to the industry.

  Patent Document 1 describes a method for obtaining a plurality of route patterns by applying a genetic algorithm using a priority of a piping route direction as a parameter as a method for searching for a piping route of a power plant. .

JP 2002-288250 A

D. E. Goldberg, “Genetic Algorithm in Search, Optimization, and Machine Learning”, Addison-Wesley (1989). Nobuo Sannomiya, Hajime Kita, Hisahama Tamaki, Takashi Iwamoto, “Genetic Algorithms and Optimization”, Asakura Shoten (1998)

  When a plurality of routes are set in a two-dimensional plane or a three-dimensional space, there is a problem that the route is set largely depending on the order of route search. Taking the problem of determining two routes in a three-dimensional space as an example, the problem regarding the order of route determination will be described in detail with reference to FIGS. 2 and 3.

  First, as shown in FIGS. 2 and 3, the target space is divided into lattice meshes. The target space is composed of 6 × 6 × 3 grids, and the hatched portion simulates an obstacle and indicates a prohibited area where the route cannot be passed. Two paths P1 and P2 having S1 and S2 as start points and G1 and G2 as end points are obtained. The connection point is a waypoint that the path P2 always passes. In the case of piping for power generation plants, general industrial plants, etc., this corresponds to the position where piping and equipment are connected.

  FIG. 2 shows an example in which a route search is first performed for P1, and then a route search is performed for P2. In this case, P1 can select the optimum route (shortest route) in the space, but P2 does not become the optimum route because it is necessary to avoid the already secured route of P1.

  On the other hand, FIG. 3 shows an example in which a route search is first performed for P2, and then a route search is performed for P1. In this case, contrary to the case of FIG. 2, P2 is the optimum route (shortest route), but P1 is not the optimum route because it is necessary to avoid the route of P2.

  For these two examples, the set route is evaluated using the total of the length of the straight line portion and the number of curved portions (bending corners) as an evaluation index. The smaller the evaluation index value is, the shorter the path is, and the path having the smallest evaluation index value is the optimum path (shortest path). The start points S1 and S2 and the end points G1 and G2 are both located at the center of the corresponding mesh, and the length of the straight line portion is expressed by the number of grids.

  In the case of FIG. 2, the total length of the path length and the curved portion of P1 is 9, and P2 is 20. In the case of FIG. 3, P1 is 15 and P2 is 14. In both cases, the sum of the evaluation indexes of P1 and P2 is 29. As described above, P1 is the optimum route in the case of FIG. 2, and P2 is the optimum route in the case of FIG. 3, and the route searched earlier is obtained as the optimum route.

  However, even if neither P1 nor P2 is the optimum route, the total of the above-described evaluation indexes becomes the same value as when one of P1 and P2 is the optimum route (in the case of FIGS. 2 and 3). There are cases. FIG. 4 shows an example of this. In FIG. 4, the total length of the path of P1 and the curved portion is 13, and P2 is 16. Therefore, P1 and P2 in this case both have larger evaluation index values than P2 when P1 is optimal (FIG. 2) and P2 when P1 is optimal (FIG. 3). That is, in the case of FIG. 4, both P1 and P2 are not optimal paths. However, the sum of the evaluation indexes of P1 and P2 is 29, which is the same value as in the case of FIGS.

  As described above, even when none of the individual routes is optimal, there is a case where the sum of the evaluation indexes is minimized and the entire route is optimal. The route in such a case cannot be searched by the conventional method.

  Therefore, it is difficult for the techniques described in Non-Patent Documents 1 and 2 and Patent Document 1 described above to obtain a large number of different routes, particularly optimal routes, without depending on the order of route search.

  An object of the present invention is to provide a route search apparatus and method that can automatically obtain a plurality of routes without depending on the route search order when searching for a plurality of routes in a three-dimensional space. is there.

  In order to solve the above-described problems, a multi-path search apparatus and method according to the present invention are configured with the following functions.

  A first multi-path search apparatus according to the present invention is a multi-path search apparatus that searches for a plurality of paths connecting a start point and an end point in a three-dimensional space, and is an external input that captures shape data of the three-dimensional space. An interface, an initial setting unit for setting a condition necessary for searching for the plurality of routes, a route calculation unit for executing a search for the plurality of routes based on an evaluation index for obtaining an optimum route, and the initial setting unit. A search condition database for storing set search conditions; a search result database for storing the shape data and a route search result by the route calculation unit; and an external output interface for outputting the route search result to the outside. The route calculation unit is characterized in that a virtual prohibited area that cannot pass the route is provided in the three-dimensional space.

  The second multi-path search device according to the present invention is the first multi-route search device according to the present invention, wherein the route calculation unit sets a route search order, and the virtual search is performed according to the route search order. Change the position of the prohibited area.

  A third multi-path search device according to the present invention is the second multi-route search device according to the present invention, wherein the evaluation index value is determined according to the number of route patterns obtained by the plurality of route searches. to correct.

  The fourth multi-path search device according to the present invention is the second multi-route search device according to the present invention, wherein the value of the evaluation index is determined according to the degree of coincidence of the pattern of the routes obtained by searching the plurality of routes. Correct.

  A fifth multi-path search device according to the present invention is connected to an input device in the first to fourth multi-route search devices according to the present invention, and the number of the virtual prohibited areas and the At least one of the evaluation index value correction methods is input.

  A sixth multi-path search device according to the present invention is connected to an image display device in the first to fifth multi-route search devices according to the present invention, and is obtained by searching for a value of the evaluation index and a plurality of the routes. At least one of the obtained route patterns is displayed on the image display device.

  According to the first multiple route search method of the present invention, a multiple route search device uses a genetic algorithm based on an evaluation index for obtaining an optimum route for a plurality of routes connecting a start point and an end point in a three-dimensional space. A method of searching for a plurality of routes to search using, a step of arranging a virtual prohibited area that cannot pass through the route in the three-dimensional space, and a placement position of the virtual prohibited area as parameters, Correcting the parameters so that the number of individuals having the same route pattern in the solution for a plurality of the obtained routes is reduced, and determining a placement of a new virtual prohibited area by the correction parameters. The route is searched for.

  A second multiple route search method according to the present invention is the first multiple route search method according to the present invention, wherein the placement position of the virtual prohibited area and the search order of the plurality of routes are represented by parameters and obtained by the search. Correcting the parameters so that the number of individuals having the same route pattern in the solution for a plurality of the routes is reduced, and determining the placement of a new virtual prohibited area by the correction parameters, Search for a route.

  A third multiple route search method according to the present invention is the second multiple route search method according to the present invention, wherein the value of the evaluation index is set according to the number of route patterns obtained by the plurality of route searches. A plurality of the routes are searched by including a correcting step.

  A fourth multiple route search method according to the present invention is the second multiple route search method according to the present invention, wherein the value of the evaluation index is determined according to the degree of coincidence of the route patterns obtained by the plurality of route searches. And searching for a plurality of the routes.

  In the route searching apparatus and method according to the present invention, when searching for a route between a predetermined start point and an end point, a result of searching for a plurality of routes can be automatically obtained without manually changing the condition setting. The trial and error work can be greatly reduced. Moreover, since an exhaustive search is possible, it is possible to prevent an oversight or a search omission.

It is a block diagram which shows the structure of the search apparatus of the multiple path | route by one Embodiment of this invention. It is a figure explaining the route search (P1 is the optimal route) by a conventional method. It is a figure explaining the route search (P2 is the optimal route) by a conventional method. It is a figure explaining the path | route (P1 and P2 are not the optimal paths | routes) which cannot be searched by the conventional method. It is a figure which shows the shape data of the apparatus arrange | positioned in the three-dimensional space used as the object of the route search apparatus by one Embodiment of this invention. It is a figure which shows the shape data of the piping arrange | positioned in the three-dimensional space used as the object of the route search apparatus by one Embodiment of this invention. It is a figure which shows the aspect of the data memorize | stored in the search result database of the route search apparatus by one Embodiment of this invention. It is a figure which shows the aspect of the data memorize | stored in the search condition database of the route search apparatus by one Embodiment of this invention. It is a figure explaining the virtual prohibition cell applied in the route calculating part of the route search apparatus by one Embodiment of this invention. It is a figure which shows an example of the gene string of the route search parameter applied with the route calculating part of the route search apparatus by one Embodiment of this invention. It is a figure explaining the calculation coefficient used for correction | amendment of the evaluation function applied with the route calculating part of the route search apparatus by one Embodiment of this invention. It is a flowchart of the process which the route calculating part of the route search apparatus by one Embodiment of this invention performs. 4 is an initial screen displayed on the image display device according to the embodiment of the present invention. It is a search condition setting screen displayed on the image display apparatus by one Embodiment of this invention. It is a display screen of search result information displayed on an image display device by one embodiment of the present invention. It is the three-dimensional route information display screen of the route obtained as a result of the search, which is displayed on the image display device according to the embodiment of the present invention.

  Hereinafter, an example of an embodiment of a multi-route search apparatus and method according to the present invention will be described with reference to the accompanying drawings. FIG. 1 is a block diagram showing the configuration of a multi-path search apparatus according to the present invention. In the present embodiment, an example in which the multi-path search device and method are applied to piping of a power generation plant or a general industrial plant will be described. In the present embodiment, the optimum route to be obtained is the shortest route having the smallest sum of the length of the straight portion of the pipe and the number of curved portions (bending angles). The multi-path search apparatus and method according to the present invention can automatically determine a large number of routes without depending on the search order when determining such an optimal route.

  The three-dimensional space 100 that is the target of route search (piping) is data stored in the storage device 110, and is created by, for example, general-purpose three-dimensional CAD. A case where a plurality of devices are arranged in the three-dimensional space 100 and these devices are connected by piping will be described. The devices are also arranged by the user using a general-purpose three-dimensional CAD, and the user also gives the starting point and the ending point of piping connecting these devices.

  A system that processes data in the three-dimensional space 100 stored in the storage device 110 includes a route search device 200, an input device 900, a support tool 910, and an image display device 950.

  The route search apparatus 200 takes in the shape data 1 characterizing the three-dimensional shape of the three-dimensional space 100 from the storage device 110 via the external input interface 210. The shape data 1 is data related to the shape of equipment and piping in the three-dimensional space 100. Further, the route search apparatus 200 transmits the search result 5 to the support tool 910 via the external output interface 260.

  FIG. 5 shows a data format related to the device among the shape data 1. As shown in FIG. 5, the data regarding the device of the shape data 1 includes the ID of the target device, the coordinates of the minimum cuboid or the center V of the cube V including the target device, and the minimum of the cuboid or cube V. Includes maximum point coordinates.

  FIG. 6 shows the data format related to the shape of the piping in the shape data 1. The ID, name, type, diameter, wall thickness, heat insulation material thickness, and clearance with other equipment or piping are described. Furthermore, the coordinates of the start point and end point of the pipe (route) designated by the user are also included.

  The shape data 1 captured by the external input interface 210 is transmitted as input data 2 to the initial setting unit 220 and the search result database (search result DB) 240.

  The initial setting unit 220 sets a specified value or default value from the user as the calculation input value 3 necessary for the route search in the route calculation unit 230, and outputs it to the route calculation unit 230. The set operation input value 3 is stored in a search condition database (search condition DB) 250. The calculation input value 3 will be described in detail in the description of the table of the search condition database 250.

  The route calculation unit 230 performs a route search based on the start / end point data 6 of the route among the input data 2 stored in the search result database 240 and the calculated input value 3 from the initial setting unit 220. The route calculation result 4 is transmitted to the external output interface 260 and the search result database 240. Detailed processing regarding the calculation of the route will be described in detail later.

  The search result database 240 receives the input data 2 from the external input interface 210 and the route calculation result 4 from the route calculation unit 230 and stores them in a predetermined format. In addition, the start point / end point data 6 of the route necessary for processing is transmitted to the route calculation unit 230. The format of data stored in the search result database 240 will be described in detail later.

  The external output interface 260 receives the route calculation result 4 from the route calculation unit 230 and transmits it to the support tool 910 as the search result 5. In addition, the search result 5 is transmitted to the storage device 110 to be stored so that it can be visualized together with the three-dimensional space 100.

  The user uses the input device 900 including the keyboard 901 and the mouse 902 and the support tool 910 connected to the image display device 950 to generate a new shape or already place the three-dimensional space 100. It is possible to correct and delete the shape data. In addition, the user uses the input device 900, the support tool 910, and the image display device 950 to set search conditions in the initial setting unit 220 and access information stored in the search condition database 250 and the search result database 240. You can do it.

  The support tool 910 includes an external input interface 920, a data transmission / reception processing unit 930, and an external output interface 940.

  The input signal 31 generated by the input device 900 is taken into the support tool 910 via the external input interface 920 of the support tool 910 and transmitted to the data transmission / reception processing unit 930 as the input signal 32.

  The data transmission / reception processing unit 930 acquires the search information 10 from the route search device 200 according to the information of the input signal 32, and refers to or sets the setting value of the initial setting unit 220 of the route search device 200. In addition, the data transmission / reception processing unit 930 transmits the search information 10 and the search result 5 as the output signal 33 to the external output interface 940.

  The output signal 33 transmitted to the external output interface 940 is transmitted as an output signal 34 to the image display device 950 and displayed on the image display device 950.

  In the present embodiment, all the databases are arranged inside the route search apparatus 200, but these can also be arranged outside the route search apparatus 200. Further, in the present embodiment, the storage device 110 is arranged outside the route search device 200, but may be arranged inside.

  Hereinafter, information stored in the database and functions of the route calculation unit 230 will be described.

  Information stored in the search result database 240 will be described with reference to FIG. FIG. 7 is a diagram for explaining an aspect of information stored in the search result database 240.

  The search result database 240 stores the input data 2 from the external input interface 210 and the route calculation result 4 from the route calculation unit 230. From the input data 2, information on the shape of the pipe (piping ID, type, caliber, wall thickness, heat insulation material thickness, and name) is obtained from the route calculation result 4, and the route information (number of vertices of the route, coordinates of each vertex, And the angle of the curved portion of the route) are reflected.

  For example, assume that a route having a start point and an end point as shown in FIG. 7 is obtained as the first route pattern. At this time, the pattern number is stored as 1 and the number of vertices is stored as 4. The values of the input data 2 are stored in the ID, pipe type, diameter, wall thickness, heat insulation material thickness, and pipe name. Next, the vertex coordinates (0, 0, 0), (x2, y2, z2), (x3, y3, z3), and (100, 500, 200) of the route are stored for each point. Finally, the angle of the music part is stored. Since the start point and the end point are end parts and there is no music part, the angle of the music part is zero. For the points other than the end portion, for example, when the angle of the curved portion is 90 degrees, a value of 90 is stored. In the case of FIG. 7, 0, a, b, and 0 are stored. In this format, all pipe information and route information are stored.

  Next, information stored in the search condition database 250 will be described with reference to FIG. FIG. 8 is a diagram for explaining the mode of information stored in the search condition database 250.

  The search condition database 250 stores the operation input value 3 output from the initial setting unit 220. The operation input value 3 includes an ID for distinguishing the target pipe, and the number of generations, number of individuals, selection operator, crossover operator, crossover probability, and mutation rate that are parameters required for the genetic algorithm. . Furthermore, the number of virtual prohibition cells, a correction function, and a weight coefficient are stored. The meaning of each item will be described in the description of the function of the route calculation unit 230.

  The calculated input value 3 includes a list C of routes used for determining the order of route search, and this list C is also stored in the search condition database 250. The route list C will also be described in the description of the function of the route calculation unit 230.

  Next, the operation of the route calculation unit 230 will be described. The route calculation unit 230 searches for the route (shortest route) having the smallest value of the evaluation index (the sum of the length of the straight line portion of the route and the number of curved portions) as the optimum route.

  There is a problem that a large number of various routes cannot be generated as a problem of the route search in the conventional method. Therefore, in the present invention, a virtual obstacle is arranged in the target space, and a number of different paths are generated by changing the arrangement position of the obstacle in various ways.

  FIG. 9 shows an example in which a virtual obstacle is arranged in the target space and a route between the start point and the end point is obtained. As a method for obtaining a path between two points, a dynamic programming method which is a conventional method is used. However, since there is an obstacle, a route that avoids this is generated. This virtual obstacle is called a virtual prohibited cell. In other words, the virtual prohibited cell is a virtual prohibited area that cannot be routed. Therefore, it is possible to solve the problem of the route search by the conventional method by performing the optimization calculation using the placement position of the virtual prohibition cell and the route search order as parameters.

  This parameter is obtained using a calculation method of a genetic algorithm. In the genetic algorithm, it is necessary to convert the target parameter into the state of the gene string. This is called coding. The state before being converted into the gene sequence state is called a phenotype.

  FIG. 10 shows an example in which a parameter for route search in the present invention is coded in a gene string. FIG. 10 shows an initial setting in which three virtual prohibited cells are arranged in the space for two routes. The head part of the gene string indicates the search order of the route, and the remaining part indicates the arrangement position of the virtual prohibition cell.

  The route search order is determined using the route list C described above as follows. The route list C is determined by the number of routes drawn in the space. For example, the route list is C = (p, q). This list C represents that two routes are drawn in the space, the first in the list is the route p, and the second is the route q. When the start point of the route p is Sp and the end point is Gp, the route p means a route candidate that connects the start point Sp and the end point Gp (if there are a plurality of route candidates, all of them are included). Similarly, the route q means a route candidate connecting the start point Sq and the end point Gq.

  Since the first gene string in FIG. 10 is “1”, the first path p in the list C is set as the first search candidate. After that, since the route p is selected, the list is updated with C = (q) (the second route q is moved up to the first). Since the second gene string in FIG. 10 is “1”, the first route q in the list C is also set as the second search candidate. The route search order is determined by the above method.

  As for the placement position of the virtual prohibition cell, coordinates in three directions of x, y, and z are required, and therefore, three gene strings are used for one cell. Each numerical value of the gene string represents the number of lattice points on the x, y, and z axes when the target space is divided into meshes.

  Using the list C and the gene sequence shown in FIG. 10, a virtual prohibited cell in which a route p (a route candidate between the start point Sp and the end point Gp) and a route q (a route candidate between the start point Sq and the end point Gq) are arranged And how many routes it is. This route search is repeated by changing the placement position of the virtual prohibited cell.

  With the settings so far, it has become possible to search for parameters using a genetic algorithm. However, as it is, there is no mechanism for arranging the virtual prohibition cell at a position where the route pattern increases in order to generate a large number of different routes, so that a result depending on the initial value is obtained. Therefore, in the present invention, a technique for correcting an evaluation function necessary for a genetic algorithm is introduced.

  An object of the present invention is to obtain a large number of shortest paths (optimal paths) that minimize the total of the length of the path and the number of curved portions. Therefore, the evaluation index (the sum of the length of the straight line portion of the route and the number of curved portions) is expressed by the evaluation function f, and a plurality of routes (shortest route) that minimize the evaluation function f are obtained. The evaluation function f can be expressed by the following equation (1).

  i is a subscript representing the i-th route, I is a set of routes, xi is the i-th pipe length, yi is the number of i-th curved portions, and α and β are weighting factors. This weight coefficient is specified by the user based on the cost and specifications of the target piping.

  Based on this evaluation function f, iterative calculation is performed to obtain the number of patterns of the obtained route and the number of individuals belonging to the pattern, and according to the number of individuals belonging to the pattern, the equations (2) and (3 ) Is used to correct the evaluation function f.

  di represents the number of individuals in the i-th pattern, and N represents the total number of individuals belonging to the obtained route. fs indicates a corrected evaluation function.

  The di in the equation (2) will be described with reference to FIG. In FIG. 11, the total number of individuals N is 10, among which the number of individuals having the first piping pattern as the solution is 5, the number of individuals having the second piping pattern as the solution is 3, and the third piping. The number of individuals with the pattern as a solution is two. In this case, the values of di are d1 = 5, d2 = 3, and d3 = 2, respectively.

  In equation (2), the larger the value of di, the smaller the value of s (di). If the value of s (di) is small, the value of the evaluation function fs is corrected so as to be large from the equation (3). That is, the evaluation function fs for an individual group having many of the same path pattern is large, and the evaluation function fs for an individual group having a small number of the same path pattern is small. In the case of FIG. 11, the individual evaluation function fs having the third piping pattern as a solution is smaller than the solid evaluation function fs having the first piping pattern as a solution.

  Moreover, you may use Formula (4) instead of Formula (2). Equation (4) shows the difference in path between the j-th pipe and the k-th pipe. That is, the reciprocal of the value of Equation (4) indicates the degree of matching of the route pattern. When the difference between the two piping paths is small, the value of equation (4), that is, the value of s (di) becomes small. That is, the value of the evaluation function fs increases. The degree of matching of the route pattern (the reciprocal of s (di)) also increases. Conversely, if the difference between the two piping paths increases, that is, if the two piping paths differ greatly, the value of s (di) increases and the value of the evaluation function fs decreases. The degree of matching of the route pattern is also reduced.

  In the initial setting unit 220 of the route search apparatus 200 according to the present invention, when using the equation (2) as the correction function information (see FIG. 8) stored in the search condition database 250, the correction function is “1” and the equation ( When 4) is used, the correction function is “2”.

  FIG. 12 is a flowchart of the process of the route calculation unit 230 based on the above-described method. Note that the parameters can be set by the user using the support tool 910 for parameters necessary for the execution of this flowchart. The setting of this value will be described later.

  In FIG. 12, in step 301, the initial setting necessary for the route search calculation is read from the initial setting unit 220. This initial setting includes the above-described route list C.

  Next, in Step 302, N individuals are randomly generated, and an initial individual population P (t) is generated with t = 0. The generation is represented by t and the final generation is represented by T.

  In step 303, the generated individual population P (t) is converted into a phenotype. The conversion to the phenotype is based on a conventionally known general genetic algorithm method.

  In step 304, the fitness of each individual is calculated according to a general genetic algorithm technique. In genetic algorithms, individuals with higher fitness are more likely to be handed over to the next generation. Therefore, the fitness function and the evaluation index are usually made equivalent as a problem of maximizing the evaluation index. In the present invention, since it is a problem of minimizing the evaluation index (evaluation function), in this calculation, a value obtained by subtracting the evaluation index (a value obtained by changing the negative sign of the evaluation index) is used as the fitness function. Be consistent with the method.

  In step 305, based on the obtained fitness, a selection operator is applied to generate a new solid population P ′ (t). The selection operator is an operation for selecting an individual to be left in the next generation. Several methods have been devised for the selection operator. For example, a roulette selection method for selecting a next generation individual with a probability according to the fitness of the individual, a ranking method for arranging solids in order of high fitness, and selecting a next generation individual according to this order, N pieces There is a tournament method or the like in which s individuals are grouped from individuals and the one having the highest fitness among them is selected, and then s copies are made to be the next generation individuals.

  In the initial setting unit 220 of the route search apparatus 200 according to the present invention, “1” is used when the roulette selection method is used as the selection operator information (see FIG. 8) stored in the search condition database 250, and the ranking method is used. Is “2”, and “3” when the tournament method is used. It is also possible to combine elite preservation methods that preferentially leave the individual with the highest fitness among all the individuals.

  Next, in step 306, two pairs are randomly created from the new individual population P ′ (t) generated in step 305, and a crossover operator is applied. Part of the genotype is exchanged to generate a new population P ″ (t). The probability applied here is called the crossover rate. A plurality of methods have been devised for this crossover operator. A symbol string of the same length as a solid consisting of 0 and 1, a one-point crossing that exchanges each other's symbol strings at two points, and a two-point crossing that borders two points. For example, there is a uniform crossover in which a generated template is generated, and a symbol string at the same position as that of 1 in the template is exchanged.

  In the initial setting unit 220 of the route search apparatus 200 according to the present invention, as crossover operator information (see FIG. 8) stored in the search condition database 250, one-point crossover is “1”, two-point crossover is “2”, The uniform crossover is “3”. The search condition database 250 also stores the crossover rate (see FIG. 8).

  In step 307, a mutation operator is applied to the individual population P ″ (t) obtained by the crossover to finally generate a next generation individual population P (t + 1). Mutation is an operation that changes a part of an individual represented by a genotype to another symbol string with a certain probability. For example, genotypes using 0 and 1 refer to changing 0 to 1 or 1 to 0. The probability used here is called the mutation rate. Information on the mutation rate is stored in the search condition database 250 (see FIG. 8).

  If it is determined in step 308 that the generation t has not reached the final generation T (t <T), the process proceeds to step 309. When the generation t reaches the final generation T (t = T), the process ends.

  In step 309, the generation t is changed to the next generation t + 1. The operations after step 303 are repeated until the generation t reaches the final generation T.

  Next, a method in which the user uses the support tool 910 to display information on the search result 5, the initial setting unit 220, the search result database 240, and the search condition database 250 on the image display device 950 will be described.

  FIGS. 13 to 16 are examples of screens displayed on the image display device 950. The user uses the keyboard 901 and the mouse 902 to execute operations such as inputting parameter values in the fields displayed on these screens.

  FIG. 13 shows an initial screen displayed on the image display device 950. The user selects a necessary button from the search condition setting button 951, the search result information display button 952, and the route information display button 953, moves the cursor 954 using the mouse 902, and clicks the mouse 902. I press the button.

  FIG. 14 is a diagram for explaining a search condition setting screen. When the search condition setting button 951 is clicked in FIG. 13, the screen shown in FIG. 14 is displayed.

  In the setting information column 971, ID, pipe name, number of generations, number of individuals, selection operator, crossover operator, crossover rate, mutation rate, number of virtual forbidden cells, correction function, and weighting factor are displayed. For the selection operator, crossover operator, and correction function, preset items are selected from the pull-down menu. Also, if either the ID or the pipe name is entered in the box on the screen and the search button 972 is clicked, only the search result can be displayed. This is a convenient function when the present apparatus is applied to a large-scale plant having a large number of target pipes.

  By clicking the transfer button 973, the information displayed in the setting information column 971 is set in the initial setting unit 220 and the search condition database 250 in the route search device 200.

  If a cancel button 974 is clicked, the information input in the setting information field 971 is cancelled. Further, by clicking the return button 975, it is possible to return to the initial screen shown in FIG.

  FIG. 15 is a display screen of search result information, in which information stored in the search result database 240 is displayed on the image display device 950. In FIG. 13, when the search result information display button 952 is clicked, the screen shown in FIG. 15 is displayed.

  The user can search for a pipe to be displayed on the image display device 950 by inputting an ID or a pipe name and clicking a search button 985. In addition, by clicking the display button 983, a change for each generation number of evaluation index values (evaluation function values), which is one piece of search information for the piping to be displayed, can be displayed in the display column 981. At this time, the minimum value, average value, and maximum value of the evaluation index value are also displayed. It should be noted that a plurality of pipes can be specified for search and display.

  In addition, the user can display the change of the obtained number of path patterns for each generation number in the display field 982. The display format can be selected by a radio button 986 at the bottom of the display field 982. For example, when displaying the number of patterns of all the pipes to be searched, the total number display portion of the radio buttons 986 is clicked. Further, when displaying individual patterns of piping, display or non-display of each piping is selected from the pull-down menu 987. Thereafter, by clicking the display button 983, the change in the number of path patterns depending on the number of generations can be displayed in the display field 982 in accordance with the selected display format.

  The user can return to the initial screen of FIG. 13 by clicking the return button 984.

  FIG. 16 is a route information display screen in which the route obtained as a result of the search is displayed on the image display device 950 in a three-dimensional space. In FIG. 13, by clicking the route information display button 953, the route information display screen shown in FIG. 16 is displayed.

  The user searches for a pipe to be displayed by inputting an ID or a pipe name and clicking a search button 991, and then selects a route pattern to be displayed. Whether to display all patterns of the target piping or not is selected by a radio button 992, and when specifying the target piping individually, it is selected from a pull-down menu 993. When the route to be displayed is determined, when a display button 994 is clicked, a top view, a front view, a side view, and an isometric view (isometric view) of the selected route are displayed.

  Further, by clicking a return button 995, it is possible to return to the initial screen of FIG.

  In addition to the information described above, arbitrary information stored in the search result database 240 and the search condition database 250 in the route search device 200 can be displayed on the image display device 950 in a predetermined manner. .

  As mentioned above, the example which applied the search apparatus and method of the multiple paths | route by this invention to piping, such as a plant for electric power generation and a general industrial plant has been demonstrated. According to the present invention, a number of different piping paths can be obtained by a single search, so that a multifaceted design is possible. Since there is no repetitive work as in the conventional method, a complicated operation is unnecessary, and it is possible to take time for more detailed design examination.

  Note that the application destination of the present invention is not limited to piping for power generation plants, general industrial plants, and the like, and if it is a route search problem in a three-dimensional space, it depends on the scale of the target space and the length of the route. It can be applied to a wide range of fields.

  DESCRIPTION OF SYMBOLS 100 ... Three-dimensional space, 110 ... Memory | storage device, 200 ... Path search device, 210 ... External input interface, 220 ... Initial setting part, 230 ... Path | route calculating part, 240 ... Search result database, 250 ... Search condition database, 260 ... External Output interface, 900 ... input device, 901 ... keyboard, 902 ... mouse, 910 ... support tool, 920 ... external input interface, 930 ... data transmission / reception processing unit, 940 ... external output interface, 950 ... image display device.

Claims (8)

A search device for a plurality of routes for searching a plurality of routes connecting a start point and an end point in a three-dimensional space,
An external input interface for capturing the shape data of the three-dimensional space;
An initial setting unit that sets conditions necessary for searching for a plurality of the routes;
A route calculation unit that performs a search for a plurality of the routes based on an evaluation index for obtaining an optimum route;
A search condition database for storing search conditions set by the initial setting unit;
A search result database for storing the shape data and a route search result by the route calculation unit;
An external output interface for outputting the route search result to the outside,
The route calculation unit provides a virtual prohibited area through which the route cannot pass through in the three-dimensional space, and represents an arrangement position of the virtual prohibited area and a search order of the plurality of routes as parameters, and is obtained by searching. the parameter is corrected so that the number of solutions of the same path pattern is reduced to the plurality of paths is, the search for multipath characterized that you determine the placement of the new virtual prohibited zone by the correction parameter apparatus. The multi-path search device according to claim 1 ,
A multi-route search device that corrects the value of the evaluation index according to the number of route patterns obtained by searching the plurality of routes. The multi-path search device according to claim 1 ,
A multi-route search device that corrects the value of the evaluation index in accordance with the degree of matching of the pattern of routes obtained by searching for a plurality of routes. The multi-path search device according to any one of claims 1 to 3 ,
Connected to the input device,
A multi-path search device for inputting at least one of the number of the virtual prohibited areas and the method of correcting the value of the evaluation index via the input device. In the multi-path search device according to any one of claims 1 to 4 ,
Connected to the image display device,
A multi-route search device that displays at least one of the evaluation index value and a plurality of route patterns obtained by searching the route on the image display device. A multi-route search method for searching for a plurality of routes connecting a start point and an end point in a three-dimensional space using a genetic algorithm based on an evaluation index for obtaining an optimal route by a multi-path search device,
Placing a virtual forbidden area that cannot pass through the path in the three-dimensional space;
The placement position of the virtual prohibited area and the search order of the plurality of routes are represented by parameters, and the parameters are corrected so that the number of individuals having the same route pattern for the plurality of routes obtained by the search is reduced. and, determining a placement of a new virtual prohibited zone by the correction parameter,
And searching for a plurality of the routes. The multi-path search method according to claim 6 ,
Correcting the value of the evaluation index according to the number of route patterns obtained by searching for a plurality of the routes;
A search method for a plurality of routes, which includes searching for a plurality of the routes. The multi-path search method according to claim 6 ,
A step of correcting the value of the evaluation index according to the degree of coincidence of route patterns obtained by searching for a plurality of routes,
A search method for a plurality of routes, which includes searching for a plurality of the routes.

Images