Route search support device, route search support method, and route search support system

Abstract

To provide a route search support device, method, and system that support efficient search for an appropriate route that passes through transit points that should be passed through even when there are many transit points.SOLUTION: A route search support device 100 includes: a unit node creation unit configured to create unit nodes that indicate transit points that a moving body can pass through; a real node setting unit configured to set a combination of unit nodes that satisfy a first condition among the unit nodes as a real node; a dummy node setting unit configured to contract the unit nodes in the real node; a link connection unit configured to create a directed graph indicating a route that a vehicle can travel based on each set node; a graph search unit configured to search for a route a vehicle takes from a departure point to a destination point via the transit points based on the directed graph; a coefficient storage unit configured to store various calculated data; an optimal route creation processing unit configured to find an optimal solution for a travel route of each vehicle; and a result output processing unit configured to output information indicating whether or not the moving body passes through each transit point based on the calculated value.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present invention relates to a route search support device, a route search support method, and a route search support system.

Background Art

[0002] In the shortest path problem using graph theory, complex constraint conditions are often set. In such cases, patterns of candidate routes are created, and by applying a solution algorithm for the set covering problem or the set problem to the created patterns, a pattern of adoption or non-adoption of each route is determined to solve the optimization problem.

[0003] As a method for creating a route plan using the set covering problem, for example, in Patent Document 1, for a set having as elements users who are targets for creating an operation plan including carpooling, an index indicating a high possibility of satisfying subadditivity is used to order the elements of the set, and it is determined whether a combination of elements in descending order and within a predetermined number satisfies subadditivity, and a combination of elements that satisfies subadditivity is added to a subset that is a target for carpooling. Thus, as a set covering problem, a set is divided into subsets allowing duplication of elements, and an operation plan program that generates an operation plan using the divided subsets is disclosed.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, since the number of route patterns increases exponentially with the number of waypoints on the route, when the number of waypoints is large, the processing steps of the optimization problem exceed the processing capacity of the computer or the performance of the algorithm, increasing the possibility that no solution can be obtained.

[0006] The present invention has been made in view of such problems, and an object thereof is to provide a route search support device, a route search support method, and a route search support system capable of efficiently searching for an appropriate route passing through waypoints to be passed even when there are a large number of waypoints.

Means for Solving the Problems

[0007] One aspect of the present invention for solving the above problems is a device for assisting in creating a route for each of a plurality of moving bodies to move through one or more waypoints, the unit node creation process for creating unit nodes indicating waypoints that can be passed through in the process of the moving body moving from the departure place to the arrival place, and among the unit nodes, one that satisfies the first condition One unit node or two adjacent to each other One or a combination of two or more unit nodes is set as a real node in the real node setting process, and among the unit nodes constituting the real node, one that satisfies the second condition Two unit nodes or two adjacent to each other Two or a combination of two or more unit nodes is searched, and the searched unit nodes are set as dummy nodes in the dummy node setting process, and the number of times of passing through the waypoints indicated by the unit nodes other than the dummy nodes among the unit nodes constituting the real node is used as a constraint condition, and when optimizing the value of an objective function including a predetermined evaluation function related to the moving route of the moving body, an optimal route creation process for calculating a value indicating the availability of using the waypoints indicated by the unit nodes, and a result output process for outputting information indicating the passage or non-passage of each waypoint of the moving body based on the calculated value are executed, and a route search support device including a processing device.

[0008] Another aspect of the present invention for solving the above problems is a route search support system for assisting in creating a route for each of a plurality of moving bodies to move through one or more waypoints, the unit node creation process for creating unit nodes indicating waypoints that can be passed through in the process of the moving body moving from the departure place to the arrival place, and among the unit nodes, one that satisfies the first condition Three unit nodes or two adjacent to each otherThree A real node setting process for setting the above combination of unit nodes to a real node, and among the unit nodes constituting the real node, one that satisfies a second condition Four unit nodes or two adjacent to each other Four A dummy node setting process for searching for a combination of two or more unit nodes that satisfies the above conditions among the unit nodes constituting the real node and setting the searched unit nodes to dummy nodes, and using, as a constraint condition, the number of times of passage of the waypoints indicated by the unit nodes other than the dummy nodes among the unit nodes constituting the real node, and calculating a value indicating whether or not to use the waypoints indicated by the unit nodes when optimizing the value of an objective function including a predetermined evaluation function related to the movement route of the moving body. An optimal route creation process, A route creation support system including a processing device that executes a result output process for outputting information indicating whether or not each waypoint of the moving body has been passed based on the calculated value, an input process for receiving an input of waypoints that can be passed by the moving body from a user, and a processing device that executes a display process for displaying the information output in the result output process on a screen.

Advantages of the Invention

[0009] According to the present invention, even when there are a large number of waypoints, it is possible to efficiently support the search for an appropriate route passing through the waypoints to be passed. Configurations, effects, and the like other than those described above will be clarified by the description of the following embodiments.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Embodiments for Carrying Out the Invention

[0011] Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[0012] <Route Search Support System> FIG. 1 is a diagram showing an example of the configuration of a route search support system 1 according to the present embodiment. As shown in the figure, the route search support system 1 includes a route search support device 100 and information processing devices of one or more user terminals 200.

[0013] Between the route search support device 100 and each user terminal 200, for example, a wired or wireless communication network 10 such as the Internet, a LAN (Local Area Network), a WAN (Wide Area Network), or a dedicated line enables communication.

[0014] When each of a plurality of moving bodies moves from a departure place to an arrival place via one or more waypoints, the route search support device 100 supports the calculation of the optimal route (the waypoints that each moving body should pass through) that each moving body should adopt. of the waypoints) to be passed.

[0015] In this embodiment, the route search support device 100 is configured to create a delivery plan for relief supplies when a disaster occurs over a wide area across multiple cities. That is, bases where relief supplies are accumulated (supply bases) exist in various locations in each city, and each vehicle for delivering relief supplies is arranged at these supply bases. Then, each vehicle loads relief supplies at the supply base, departs from the supply base as the starting point, supplies relief supplies while stopping at each evacuation shelter, and arrives at a predetermined destination. And the route search support device 100 is configured to determine the routes of the vehicles such that the amount of relief supplies to be supplied is as fair as possible among the cities. Note that there is a constraint condition that only one vehicle can visit each evacuation shelter.

[0016] Specifically, first, the user terminal 200 requests the route search support device 100 to create a delivery plan for relief supplies. Then, the route search support device 100 creates a graph with supply bases, evacuation shelters, and destinations as nodes respectively, and performs graph search to create candidates for the routes of each vehicle. And the route search support device 100 solves an optimization problem for fair distribution of relief supplies by incorporating the created route candidates into an optimization model including a predetermined objective function, and specifies the optimal routes of each vehicle. The user terminal 200 displays the route information created by the route search support device 100. Note that the user of the user terminal 200 is, for example, a country, a local government, or an operator that manages disaster response.

[0017] In this embodiment, as an optimization model for calculating the optimal route, for example, the model disclosed in International Publication No. WO2016 / 157333 is adopted.

[0018] The applicant of the present application has developed quantum computing technology and has been attempting to solve various problems in, for example, the exhaustive search problem based on big data (including the concept of combinatorial optimization problems). For such an exhaustive search problem, generally, high expectations are placed on quantum computers. A quantum computer consists of basic elements called qubits and can simultaneously realize "0" and "1". Therefore, it is possible to simultaneously calculate all solution candidates as initial values and has the potential to realize exhaustive search. However, a quantum computer needs to maintain quantum coherence over the entire calculation time.

[0019] Among these, a method called adiabatic quantum computing has come to be noticed (Reference: E. Farhi, et al., ”A quantum adiabatic evolution algorithm applied to random instances of an NP-complete problem,” Science 292, 472 (2001).). This method attempts to obtain a solution by transforming the problem so that the ground state of a certain physical system becomes the solution and finding the ground state.

[0020] Let the Hamiltonian of the physical system for which the problem is set be H^p. However, at the start of the operation, instead of setting the Hamiltonian as H^p, another Hamiltonian H^0 for which the ground state is clear and easy to prepare is used. Next, the Hamiltonian is shifted from H^0 to H^p over a sufficiently long time. If enough time is taken, the system will remain in the ground state and the ground state of the Hamiltonian H^p can be obtained. This is the principle of adiabatic quantum computing. If the calculation time is τ, the Hamiltonian becomes Equation (1), [Equation 1] JPEG0007685980000001.jpg24170 Based on the Schrödinger equation of Equation (2), time evolution is carried out to obtain a solution.

[0021] [Equation 2] JPEG0007685980000002.jpg23170 Adiabatic quantum computation is applicable to problems that require exhaustive search and reaches a solution in a one-way process. However, if the computational process needs to follow the Schrödinger equation of Equation (2), it is necessary to maintain quantum coherence, similar to a quantum computer.

[0022] However, while a quantum computer repeats gate operations for one or two qubits, adiabatic quantum computation interacts all qubits simultaneously across the entire qubit system, with a different concept of coherence.

[0023] For example, consider the gate operation on a certain qubit. At this time, if there is an interaction between that qubit and other qubits, it will cause decoherence. However, in adiabatic quantum computation, since all qubits interact simultaneously, it will not cause decoherence in cases like this example. Reflecting this difference, adiabatic quantum computation is considered to be more robust against decoherence than a quantum computer.

[0024] As described above, adiabatic quantum computation is effective for difficult problems that require exhaustive search. Then, spin is used as a variable in the operation, and the problem to be solved is set by the spin-spin interaction and the local field acting on each spin.

[0025] At time t = 0, all spins are aligned in one direction by an external magnetic field, and the external magnetic field is gradually decreased so that it becomes zero at time t = τ.

[0026] Each spin is time-evolved assuming that its direction is determined according to the effective magnetic field determined by all the actions of the external magnetic field and the spin-spin interaction at each site at time t.

[0027] At that time, by making the direction of the spin not completely aligned with the effective magnetic field but quantum mechanically corrected, the system is made to almost maintain the ground state.

[0028] In addition, a term (relaxation term) that maintains each spin in its original direction during time evolution is added to the effective magnetic field to improve the convergence of the solution. Based on the above, next, the functions of the route search support device 100 will be described.

[0029] <Route Search Support Device> FIG. 2 is a diagram showing an example of the functions provided by the route search support device 100. The route search support device 100 includes an input data storage unit 110, a graph analysis unit 120, an optimal route creation processing unit 130, a result output processing unit 140, and an output data storage unit 150.

[0030] The input data storage unit 110 stores data input via the user terminal 200.

[0031] The graph analysis unit 120 creates a graph showing candidates for the routes of each vehicle. This graph is a directed graph having the departure point, the arrival point, and the waypoints as nodes respectively, and the routes connecting these points as links.

[0032] Specifically, the graph analysis unit 120 includes a unit node creation unit 121, a real node setting unit 122, a dummy node setting unit 123, a link connection unit 124, a graph search unit 125, and a coefficient storage unit 126.

[0033] The unit node creation unit 121 creates unit nodes that are nodes indicating waypoints that can be passed through in the process of a vehicle moving from the departure point to the arrival point.

[0034] The real node setting unit 122 reduces a part of the unit nodes. That is, the real node setting unit 122 sets, as real nodes, combinations of one or two or more adjacent unit nodes that satisfy the first condition among the unit nodes.

[0035] The unit nodes that satisfy the first condition are, for example, unit nodes related to shelters that a vehicle is highly likely to pass through.

[0036] In addition, a combination of two or more adjacent unit nodes that satisfy the first condition means that the evaluation value indicating the strength of the connection between the unit nodes is equal to or greater than a predetermined value. In the present embodiment, it is assumed that the travel time or distance of the vehicle between the unit nodes is equal to or greater than a predetermined value, but it may be based on an evaluation value from another perspective (for example, an evaluation value indicating a low degree of route congestion).

[0037] In addition, the real node setting unit 122 creates real nodes that satisfy predetermined constraint conditions (for example, designation of some routes where vehicles cannot pass) commonly applied to each vehicle.

[0038] Next, the dummy node setting unit 123 reduces the unit nodes within the real nodes. That is, the dummy node setting unit 123 searches for one or two or more adjacent unit nodes that satisfy the second condition among the unit nodes constituting the real nodes, and sets the searched unit nodes as dummy nodes.

[0039] In the present embodiment, the second condition is that when a vehicle moves from a departure point to a destination via unit nodes other than the set real nodes (that is, when a different travel route is adopted), if one or two or more adjacent unit nodes that are necessary for the passage are within the unit nodes of the set real nodes, those unit nodes are set as dummy nodes.

[0040] In addition, the second condition may be that among the unit nodes constituting the set real nodes, unit nodes (for example, the travel time from other unit nodes is equal to or greater than a predetermined time or the distance is equal to or greater than a predetermined distance) whose evaluation value indicating the strength of the connection with other unit nodes is equal to or less than a predetermined value are set as dummy nodes. The second condition described here is an example, and the second condition may be any condition for designating unit nodes that are particularly necessary for the passage of the vehicle in each travel route among the unit nodes constituting the real nodes. Note that these second conditions can be applied in combination.

[0041] The link connection unit 124 creates a directed graph indicating a route along which the vehicle can move, based on each node set by the real node setting unit 122 and the dummy node setting unit 123.

[0042] Based on the directed graph created by the link connection unit 124, the graph search unit 125 searches for a route (route candidate) by which the vehicle travels from the departure point, via the waypoint, and reaches the arrival point.

[0043] The coefficient storage unit 126 stores various data such as the route (list of supply points, evacuation shelters to pass through, and arrival points) of each route candidate calculated in the process of the processing by the graph search unit 125, the distance of each route candidate, and the supply amount of relief supplies for each city by the vehicle moving along each route candidate. These data include information such as the coefficients of the optimization model created by the optimal route creation processing unit 130.

[0044] Next, the optimal route creation processing unit 130 creates an optimization model based on the data stored in the coefficient storage unit 126, and obtains an optimal solution for the movement route of each vehicle such that the supply amount of relief supplies in each city becomes fair.

[0045] The optimization model includes a constraint condition that is the number of times of passing through the waypoint indicated by the unit node other than the dummy node among the unit nodes constituting the real node, and an objective function including an evaluation function related to the movement route of the vehicle. Further, the optimization model may include different constraint conditions in the objective function according to the vehicle.

[0046] Also, in the present embodiment, it is assumed that the evaluation function is a leveling function indicating the sufficiency rate of relief supplies between cities (the ratio of the actual supply amount to the demand amount of relief supplies), but it is also possible to adopt other evaluation functions, for example, a function for reducing the moving distance.

[0047] Also, in the present embodiment, it is assumed that the objective function includes a constraint condition on the number of vehicles arranged at each departure point, but the present invention is not limited to this.

[0048] Here, in this embodiment, an Ising model is adopted as the optimization model. Specifically, the Ising model is an Ising model in which, for the objective function including as terms a constraint function that becomes minimum when the constraint conditions regarding the number of transit times of the transit points indicated by the unit nodes other than the dummy nodes among the unit nodes constituting the real nodes are satisfied, and an evaluation function regarding the movement route of the vehicle, the availability of the transit points indicated by the unit nodes is set as a spin, and the sensitivity between variables in the constraint function is set as the strength of the interaction between the spins. Details of the Ising model will be described later.

[0049] Note that the optimization model is not limited to the Ising model, and any model that can appropriately solve the combinatorial optimization problem in accordance with the route search support method of the present invention is applicable. The outline of the Ising model is disclosed in the above-mentioned international publication, and details such as its specific configuration and operation will be omitted as appropriate (the same applies hereinafter).

[0050] Next, the result output processing unit 140 displays the calculation result by the optimal route creation processing unit 130 on the screen of the user terminal 200. The output data storage unit 150 stores the calculation result by the optimal route creation processing unit 130.

[0051] Here, FIG. 3 is a diagram showing an example of the hardware included in the route search support apparatus 100. The route search support apparatus 100 includes a processing device 101 such as a CPU (Central Processing Unit), a DSP (Digital Signal Processor), a GPU (Graphics Processing Unit), an FPGA (Field-Programmable Gate Array), a memory 102 such as a RAM (Random Access Memory) and a ROM (Read Only Memory), a storage device 103 such as an HDD (Hard Disk Drive) and an SSD (Solid State Drive), and is composed of a NIC (Network Interface Card), a wireless communication module, a USB (Universal Serial Interface) module, or a serial communication module, etc. The routing assistance device 104 includes an input device 105 such as a keyboard, a mouse, or a touch panel, and an output device 106 such as a liquid crystal monitor or an LCD (Liquid Crystal Display). The user terminal 200 also includes similar hardware.

[0052] Note that the routing assistance device 100 may be implemented not only by hardware implemented by an electronic circuit (such as a digital circuit) in the annealing method, but also by a superconducting circuit or the like. Further, the routing assistance device 100 may include hardware that realizes an Ising model by a method other than the annealing method. This method is, for example, a laser network method (optical parametric oscillation), a quantum neural network, or the like. Further, as described above, although some concepts are different, the present invention can also be realized in a quantum gate method in which calculations performed in the Ising model are replaced with gates such as Hadamard gates, rotation gates, and controlled NOT gates. It can be realized.

[0053] Each function realized by the routing assistance device 100 described above is realized by the processing device 101 reading and executing each program stored in the memory 102 or the storage device 103. Each program can be recorded and distributed on a recording medium, for example. Note that all or part of the routing assistance device 100 may be realized using virtual information processing resources provided using virtualization technology, process space separation technology, or the like, such as a virtual server provided by a cloud system. Further, all or part of the functions provided by the routing assistance device 100 may be realized by a service provided by a cloud system via an API (Application Programming Interface) or the like.

[0054] Note that adiabatic quantum computing in the Ising model is also known as quantum annealing, which is the development of the concept of classical annealing in quantum mechanics. That is, adiabatic quantum computing is originally capable of classical operation, and can also be interpreted as adding quantum mechanical effects to improve performance in terms of speed and solution accuracy. Therefore, in this embodiment, the processing device 101 of the route search support device 100 itself is classical, and by introducing parameters determined quantum mechanically in the calculation process, a classical but quantum mechanical effect-including calculation method and device is realized, and calculation using the Ising model is performed. However, a form in which the processing device 101 is configured by a quantum computer can of course also be adopted. Next, the processing performed in the route search support system 1 will be described.

[0055] <Route search support processing> FIG. 4 is a flowchart for explaining an outline of a route search support process, which is a process for searching for an optimal transportation route of relief supplies by each vehicle. The route search support process is started, for example, when an instruction for route search is input from the user terminal 200 to the route search support device 100.

[0056] First, the route search support device 100 executes a route candidate creation process s10 for creating candidates for the transportation route of relief supplies (route candidates). Then, the route search support device 100 executes an optimal route creation process s20 for creating an Ising model based on the route candidates created in the route candidate creation process s10 and calculating an optimal transportation route of relief supplies. After that, the route search support device 100 outputs information regarding the optimal movement route created in the optimal route creation process s20 and displays it on the screen of the user terminal 200 (s30).

[0057] For example, the route search support device 100 displays each movement route on a map (for real nodes and dummy nodes, the via points are displayed for each unit node constituting them), and also displays the amount of relief supplies transported by each movement route, the total amount of relief supplies supplied to each city, and the like. Details of the route candidate creation process s10 and the optimal route creation process s20 will be described below.

[0058] <Route candidate creation process> FIG. 5 is a flowchart for explaining the details of the route candidate creation process s10. First, the graph analysis unit 120 stores the bases to be set as unit nodes among the conceivable aggregation bases, destinations, and shelters (s101).

[0059] FIG. 6 is a diagram showing an example of the setting screen displayed on the user terminal 200 during the process of s101. This setting screen 500 includes a location operation screen 510, an inter-location operation screen 520, a transportation equipment operation screen 530, and an advanced screen 540.

[0060] The location operation screen 510 displays a list of each location (candidates for the departure place, shelter, and destination) such as the aggregation bases of relief materials in each city and public facilities, and accepts the selection of the location to be adopted as the unit node from the user.

[0061] The inter-location operation screen 520 displays a list of the routes connecting the locations and the distances of those routes, and accepts the selection of available routes (links capable of connecting between unit nodes) from the user.

[0062] Note that the setting screen 500 may display a predetermined map screen, and accept the selection of the location corresponding to the location operation screen 510 and the selection of the route corresponding to the inter-location operation screen 520 from the displayed map screen.

[0063] The transportation equipment operation screen 530 displays a list of vehicles and the amounts of relief materials transported by those vehicles, and accepts the selection of available vehicles from the user.

[0064] The advanced screen 540 receives settings from the user regarding the optimization model, real nodes, and dummy nodes. Specifically, the advanced screen 540 includes a variable scale magnification setting unit 541 that receives a setting of the upper limit of the number of variables available for use in the optimization model created in the optimal route creation process s20 or the upper limit of the ratio to the default value (variable scale magnification) to adjust the complexity of the optimization problem, a dummy node number setting unit 542 that receives a setting of the upper limit number of dummy nodes that can be set, and a longest real node setting unit 543 that receives a setting of the upper limit of the length of real nodes that can be reduced.

[0065] By adjusting the upper limit number of dummy nodes, it is possible to adjust the ease of passing through shelters related to dummy nodes in route search. For example, by reducing the upper limit number of dummy nodes, the route search speed can be improved. Also, by adjusting the longest real node, the route search speed can be adjusted. Further, by adjusting the variable scale magnification, it is possible to ensure that the number of variables in the objective function does not exceed the number of variables indicated by the variable scale magnification in the process in the optimal route creation process s20.

[0066] Next, as shown in FIG. 5, the graph analysis unit 120 acquires the mandatory constraint conditions commonly applied to each vehicle regarding the locations indicated by each unit node set in s101 (s102). For example, the graph analysis unit 120 acquires information on the time zones of shelters where the supply of relief supplies is impossible, or routes that vehicles cannot pass through among the routes between the supply base and shelters, routes between shelters, or routes between shelters and destinations (for example, routes that vehicles cannot pass through due to disasters) from a predetermined database.

[0067] Then, the graph analysis unit 120 creates a temporary directed graph indicating a movement route that satisfies the above constraint conditions and also satisfies a predetermined evaluation criterion (s103). The graph analysis unit 120 sets each unit node on this directed graph as a real node.

[0068] Specifically, the graph analysis unit 120 sets an evaluation criterion regarding the length of the distance of the route between locations (an equation of an evaluation criterion such that the shorter the distance, the higher the evaluation value). Then, after setting the destination and the arrival location, the graph analysis unit 120 uses an algorithm such as Dijkstra's algorithm to search for a vehicle route from the departure location to the destination that satisfies the constraint conditions set in s102 and maximizes the evaluation value according to the above evaluation criterion. Note that the graph analysis unit 120 stores in advance information (for example, information between bases) necessary for creating the evaluation criterion.

[0069] Then, the graph analysis unit 120 reduces and combines real nodes that are adjacent to each other and have a strong mutual connection (real nodes that satisfy the first condition) among the set real nodes into one real node (s104).

[0070] Specifically, the graph analysis unit 120 reduces real nodes when the evaluation value indicating the strength of the connection between adjacent real nodes is equal to or greater than a predetermined value. For example, the graph analysis unit 120 reduces and stores real nodes when the distance between adjacent real nodes is shorter than a predetermined distance. Note that this predetermined distance may be stored in the route search support device 100 in advance, or may be a statistical value (for example, an average value) of the distances used so far.

[0071] FIG. 7 is a diagram for explaining an example of real node search. As shown in the figure, when the vehicle searches for a moving route from node 700, which is the departure point, to a destination node (not shown), as the adoptable moving routes after the first unit node 701, which is a real node, there are a route to the second unit node 702 and a route to the fourth unit node 704. However, since the distance from the first unit node 701 to the second unit node 702 is shorter than the distance from the first unit node 701 to the fourth unit node 704, the graph analysis unit 120 sets the route from the first unit node 701 to the second unit node 702 as the shortest route and sets the second unit node 702 as a real node. Further, the graph analysis unit 120 sets the third unit node 703, which is the next adjacent node of the second unit node 702, as a real node. Then, when the distances between the first unit node 701 and the second unit node 702 and between the second unit node 702 and the third unit node 703 are each equal to or less than a predetermined distance, the graph analysis unit 120 reduces the first unit node 701, the second unit node 702, and the third unit node 703 into one real node 705.

[0072] FIG. 8 is a diagram showing an example of the configuration of a directed graph to be created. In the directed graph shown in this figure, regarding the movement route from the unit node 811 related to the first departure place to the unit node 821 related to the first arrival place, a first movement route 831 that passes through the first unit node 801 (real node) related to the first shelter and the fourth unit node 804 (real node) related to the fourth shelter in this order is calculated. Also, regarding other movement routes from the unit node 811 related to the first departure place to the unit node 821 related to the first arrival place, a second movement route 832 that passes through the fourth unit node 804 (real node) related to the fourth shelter, the second unit node 802 (real node) related to the second shelter, the third unit node 803 (real node) related to the third shelter, the sixth unit node (real node) related to the sixth shelter, and the fifth unit node (real node) related to the fifth shelter in this order is calculated. On the other hand, regarding the movement route from the unit node 812 related to the second departure place to the unit node 822 related to the second arrival place, a third movement route 833 that passes through the third unit node 803 (real node) related to the third shelter, the sixth unit node 806 (real node) related to the sixth shelter, and the seventh unit node 807 (real node) related to the seventh shelter in this order is calculated. Furthermore, regarding the movement route from the unit node 812 related to the second departure place to the unit node 821 related to the first arrival place, a fourth movement route 834 that passes through the third unit node 803 (real node) related to the third shelter, the sixth unit node 806 (real node) related to the sixth shelter, and the fifth unit node 805 (real node) related to the fifth shelter in this order is calculated.

[0073] When the distance between the first unit node 801 (real node) related to the first evacuation site and the fourth unit node 804 (real node) related to the fourth evacuation site is equal to or less than a predetermined distance, the graph analysis unit 120 reduces these real nodes and sets them as the first reduced real node 841. Also, when the distance between the third unit node 803 (real node) related to the third evacuation site and the sixth unit node 806 (real node) related to the sixth evacuation site is equal to or less than a predetermined distance, and the distance between the sixth unit node 806 (real node) related to the sixth evacuation site and the seventh unit node 807 (real node) related to the seventh evacuation site is equal to or less than a predetermined distance, the graph analysis unit 120 reduces these three real nodes and sets them as the second reduced real node 842.

[0074] Note that when the total length of the unit nodes constituting the reduced real node exceeds the maximum real node length set by s101, the graph analysis unit 120 divides the reduced real node into two or more reduced real nodes so that the length of the unit nodes constituting each reduced real node does not exceed the maximum real node length.

[0075] Next, as shown in FIG. 5, the graph analysis unit 120 sets the unit nodes that satisfy the second condition among the unit nodes constituting each real node in each movement path as dummy nodes (s105, s106). Specifically, the graph analysis unit 120 identifies the unit nodes that are important as waypoints in other movement paths among the unit nodes within each real node in each movement path, and sets the identified unit nodes as dummy nodes (s105). Then, the graph analysis unit 120 reduces a plurality of adjacent dummy nodes among the set dummy nodes and sets them as one dummy node (s106).

[0076] For example, the graph analysis unit 120 selects a real node of a certain movement route, and among the unit nodes of the selected real node, it searches for and sets as dummy nodes the unit nodes related to the shelters that need to be passed through in other movement routes (movement routes passing through shelters indicated by unit nodes other than the unit nodes within the real node). Also, for example, the graph analysis unit 120 may set as dummy nodes the unit nodes within a real node whose preset importance is higher than a predetermined value.

[0077] Also, for example, the graph analysis unit 120 sets as dummy nodes the unit nodes among the unit nodes within a real node for which the evaluation value indicating the strength of the connection with other unit nodes is equal to or less than a predetermined value (for example, the distance from other unit nodes is equal to or more than a predetermined distance).

[0078] Next, the graph analysis unit 120 reduces a plurality of adjacent dummy nodes among the dummy nodes set as described above into one dummy node. At this time, the graph analysis unit 120 may reduce into one dummy node a combination of dummy nodes for which the evaluation value indicating the strength of the connection with each other is equal to or more than a predetermined value (for example, a combination of real nodes whose distance between adjacent real nodes is equal to or less than a predetermined distance).

[0079] Note that when the number of dummy nodes has reached or exceeded the upper limit number of dummy nodes set in s101, the graph analysis unit 120 stops setting any more dummy nodes.

[0080] Here, FIG. 9 is a diagram for explaining an example of a method for setting dummy nodes by the graph analysis unit 120 (showing the situation following FIG. 8). For example, the first reduced real node 841 set in the first movement route 831 has a first unit node 801 and a fourth unit node 804. Here, in the second movement route 832, which is another movement route, it is necessary to pass through the fourth unit node 804. Therefore, the graph analysis unit 120 sets the fourth unit node 804 in the first reduced real node 841 as a dummy node.

[0081] Also, in the second reduced real node 842 set in the third movement path 833, there are a third unit node 803, a sixth unit node 806, and a seventh unit node 807. Here, in the second movement path 832, which is another movement path, it is necessary to pass through the third unit node 803 and the sixth unit node 806. Therefore, the graph analysis unit 120 sets the third unit node 803 and the sixth unit node 806 in the second reduced real node 842 as dummy nodes. Furthermore, the graph analysis unit 120 reduces these adjacent third unit node 803 and sixth unit node 806 into one dummy node.

[0082] Furthermore, regarding the third movement path 833, in the fourth movement path 834, which is another movement path, it is necessary to pass through the sixth unit node 806. Therefore, the graph analysis unit 120 sets the sixth unit node 806 in the second reduced real node 842 as a dummy node. In this way, the dummy nodes may be set simultaneously (duplicately) in relation to a plurality of movement paths (here, the second movement path 832 and the fourth movement path 834).

[0083] As described above, the graph analysis unit 120 creates real nodes and dummy nodes whose reduction perspectives are different from each other (s104 - s106).

[0084] Next, as shown in FIG. 5, the graph analysis unit 120 creates a graph composed of nodes and links by creating links that connect the nodes based on the real nodes and dummy nodes identified above (s107).

[0085] For example, the graph analysis unit 120 sets one real node by representing it with the unit node closest to the departure point in the reduced real node (however, a real node that is not set as a dummy node). Further, the graph analysis unit 120 sets one dummy node by representing it with the unit node closest to the departure point in the dummy node. The graph analysis unit 120 creates a graph consisting of nodes and links based on the node related to the departure point, the node related to the arrival point, the set nodes, and the real nodes and dummy nodes other than the set nodes (not reduced).

[0086] At this time, the graph analysis unit 120 calculates the distances between the created links by obtaining the respective distances between the unit nodes (departure point, each evacuation site, and arrival point) set in s101 from a predetermined database.

[0087] Then, the graph analysis unit 120 searches for and identifies all movement routes (route candidates) from each departure point to each arrival point via evacuation sites based on the directed graph created in s107 (s108). Then, the graph analysis unit 120 stores various data calculated by searching for and identifying all movement routes in the database. Thus, the route candidate creation process s10 ends.

[0088] FIG. 10 is a diagram showing an example of the database stored in the graph analysis unit 120. This database is a database having data items of a plurality of movement routes 1001 (route candidates).

[0089] The data items of the movement route 1001 have each data sub-item of the supply point 1002 as the departure point, one or more evacuation sites 1003 via which relief supplies are supplied, the movement distance 1004, and the supply amount 1005 of relief supplies for each city (note that information on the city to which each evacuation site belongs can be obtained from a predetermined database). Note that in each data sub-item, 1 is set when the supply point or an evacuation site is used or passed through, and 0 is set when it is not used or passed through.

[0090] <Optimal Route Creation Process> Next, the optimal route creation process s20 will be described. First, based on the database created in the route candidate creation process s10, the optimal route creation processing unit 130 creates an Ising model for obtaining the optimal movement route (waypoint) of each vehicle when leveling the supply of relief materials in each city as an Ising model.

[0091] (Regarding the Ising model) Here, a classical algorithm for the route search support device 100 to obtain the ground state as the solution of the Ising model and the configuration of the Ising model for realizing it will be described.

[0092] The route search support device 100 has N variables sj z (j = 1, 2, …, N) taking values in the range of -1 ≤ sj z ≤ 1, and sets an Ising model having a local field gj and an interaction Jij (i, j = 1, 2, …, N) between variables.

[0093] Specifically, the route search support device 100 discretely calculates from t = t0 (t0 = 0) to tm (tm = τ) by dividing the time into m parts, and when obtaining the variable Sj Z (tk) at each time tk, uses the value of the variable Sj z (tk - 1) (i = 1, 2,.., N) at the previous time tk - 1 and the coefficients 9pina or 9pinb of the relaxation term to obtain Bj z (tk) = {ΣiJijSi z (tk - 1) + gj + sgn(sj z (tk - 1)) · 9pina} · tk / τ or Bj z (tk) = {ΣiJiJSj z (tk - 1) + gj + 9pinb.Sj Z (tk - 1)} · tk / τ, and determines the function f so that the value range of the above variable Sj z (tk) is -1 ≤ sj z (tk) ≤ 1, and Sj z (tk) = f(Bj z(tk), tk), and as the time step progresses from t = t0 to t = tm, the above variable Sj z is brought closer to -1 or 1, and finally sj z If < 0, then Sj zd = -1, and if Sj z > 0, then Sj zd is defined as 1 to determine the solution.

[0094] The coefficient gpinb is, for example, a value from 50% to 200% of the average value of |Jij|. Also, regarding the local field gj of the problem setting, the correction term δgj' can be added only to gj' for a certain site j', and only the magnitude of gj' can be increased for that site j'. Further, the correction term δgj' is, for example, a value from 10% to 100% of the average value of |Jij|.

[0095] Subsequently, starting from the quantum mechanical description and transitioning to the classical form, the basic principle of annealing realized by the path search support device 100 is described.

[0096] The problem of searching for the ground state of the Ising spin Hamiltonian given by Equation (3) includes a classification problem called NP-hard and is known to be a useful problem (Reference: F. Barahona, ”On the computational complexity of Ising spin glass models,” J. Phys. A: Math. Gen. 15, 3241 (1982).).

[0097] [Equation 3] JPEG0007685980000003.jpg21170 Jij and gj are problem setting parameters, and σ^ Z is the z-component of the Pauli spin matrix and takes eigenvalues of ±1. i, j represent the spin sites. An Ising spin is a variable that can only take values of ±1, and in Equation (3), since the eigenvalues of σ^ z are ±1, it forms an Ising spin system.

[0098] The Ising spin in formula (3) does not necessarily have to be a literal spin; anything physical is fine as long as the Hamiltonian is described by formula (3).

[0099] For example, it is also possible to associate the availability of each base point of each vehicle described later with ±1, or to associate the high and low of a logic circuit with ±1. It is also possible to associate the vertical polarization and horizontal polarization of light with ±1, or to associate the phases of 0 and π with ±1.

[0100] In the method exemplified here, similar to adiabatic quantum computing, the operation system is prepared in the ground state of the Hamiltonian given by formula (4) at time t = 0.

[0101] [Formula 4] JPEG0007685980000004.jpg23170 γ is a proportionality constant determined by the magnitude of the external field uniformly applied to all sites j, and σ^j x is the x-component of the Pauli spin matrix. If the operation system is the spin itself, the external field means a magnetic field.

[0102] Formula (4) corresponds to applying a transverse magnetic field, and the case where all spins point in the x direction (γ > 0) is the ground state. The Hamiltonian of the problem setting was defined as an Ising spin system with only the z-component, but the x-component of the spin appears in formula (4). Therefore, the spin in the operation process is not Ising but vectorial (Bloch vector). Starting from the Hamiltonian of formula (4) at t = 0, the Hamiltonian is gradually changed as time t progresses, and finally, the Hamiltonian described by formula (3) is obtained with its ground state as the solution.

[0103] [Formula 5] JPEG0007685980000005.jpg14170 Here, $\boldsymbol{\sigma}$ represents the three components of the Pauli spin matrix as a vector. The ground state is the case where the spin points in the direction of the magnetic field, and with $\langle\cdot\rangle$ being the quantum mechanical expectation value, we can write $\langle\boldsymbol{\sigma}\rangle=\mathbf{B} / |\mathbf{B}|$. Since we try to maintain the ground state at all times during the adiabatic process, the direction of the spin always follows the direction of the magnetic field.

[0104] The above discussion can be extended to a multi-spin system. At $t = 0$, the Hamiltonian is given by Eq. (4). This means that a uniform magnetic field $\mathbf{B}_j$ X $=\gamma$ is applied to all spins. For $t>0$, the x-component of the magnetic field gradually weakens and $\mathbf{B}_j$ X $=\gamma(1 - t / \tau)$. Regarding the z-component, due to the spin-spin interaction, the effective magnetic field is given by Eq. (6).

[0105] [Eq. 6] JPEG0007685980000006.jpg26170 The direction of the spin can be defined by $\langle\boldsymbol{\sigma}$ z $\rangle / \langle\boldsymbol{\sigma}$ X $\rangle$. So, if the direction of the spin follows the effective magnetic field, the direction of the spin is determined by Eq. (7).

[0106] [Eq. 7] JPEG0007685980000007.jpg21170 Although Eq. (7) is a quantum mechanical description but taking the expectation value, unlike Eqs. (1)-(6), it is a relational expression regarding classical quantities.

[0107] In the classical system, there is no quantum mechanical nonlocal correlation (quantum entanglement), so the direction of the spin should be completely determined by the local field for each site, and Eq. (7) determines the behavior of the classical spin system. In the quantum system, due to the nonlocal correlation, Eq. (7) will be deformed. Regarding this, we will discuss it later. Here, for the purpose of describing the basic form of the invention, we describe the classical system determined by Eq. (7).

[0108] Figure 11 shows a timing chart for obtaining the ground state of the spin system. Since the description in this figure is about classical quantities, the spin of site j is represented by sj instead of σ^j. Accordingly, the effective magnetic field Bj in this figure is a classical quantity. At t = 0, an effective magnetic field Bj in the right direction is applied to all sites, and all spins Sj are initialized in the right direction.

[0109] As time t elapses, a magnetic field in the z-axis direction and spin-spin interaction are gradually applied, and finally the spin becomes in the +z direction or -z direction, and the z-component of spin Sj is sj z = +1 or -1. Although it is ideal for time t to be continuous, it can also be made discrete in the actual calculation process to improve convenience. The discrete case will be described below.

[0110] The spin illustrated here is a vector spin because not only the z-component but also the x-component is added. The behavior as a vector can also be understood from Figure 11. The y-component has not appeared so far because the y-component of the external field does not exist because the direction of the external field is taken in the xz plane, and thus <σ^ Y >= 0.

[0111] As the spin of the calculation system, a three-dimensional vector of magnitude 1 (this is called the Bloch vector and the state can be described by a point on the spherical surface) is assumed, but in the way of taking the axis in the example shown in Figure 11, only two dimensions need to be considered (the state can be described by a point on the circle).

[0112] Also, since γ is constant, Bj x (t)>0 (γ>0) or Bj x (t)<0 (γ<0) holds. In this case, the two-dimensional spin vector can be described only by a semi-circle, and if Sj is specified in [-1,1], the two-dimensional spin vector is determined by one variable of Sj z and thus, in the example here, although the spin is a two-dimensional vector, it can also be expressed as a one-dimensional continuous variable with a value range of [-1,1]. z

[0113] ​In the timing chart of FIG. 11, the effective magnetic field is obtained for each site at time t = tk, and the direction of the spin at t = tk is obtained by using this value in Equation (8).

[0114] [Equation 8] JPEG0007685980000008.jpg17170 Since Equation (8) is a rewritten form of Equation (7) in terms of classical quantities, it does not have the symbol <·>. Next, the effective magnetic field at t = tk + l is obtained by using the spin value at t = tk. Specifically writing the effective magnetic field at each time gives Equations (9) and (10).

[0115] [Equation 9] JPEG0007685980000009.jpg27151 [Equation 10] JPEG0007685980000010.jpg26151 Hereinafter, following the procedure schematically shown in the timing chart of FIG. 11, the spin and the effective magnetic field are alternately obtained.

[0116] In the classical system, the magnitude of the spin vector is 1. In this case, each component of the spin vector is given by tanθ = Bj z (tk) / Bj x (tk) using the mediating variable θ defined by Sj z (tk)=sinθ, Sj x (tk)=COSθ.

[0117] Rewriting this gives Sj z (tk)=sin(arctan(Bj z (tk) / Bj x (tk))), Sj x (tk)=cos(arctan(Bj z (tk) / Bj x (tk))).

[0118] As is clear from Equation (9), the variable of Bj x (tk) is only tk, and τ and γ are constants. Therefore, Sj z(tk) = sin(arctan(Bj z (tk) / Bj x (tk))) and Sjx(tk) = cos(arctan(Bj z (tk) / Bj x (tk))) are functions of Bj z (tk) and tk, and can be expressed in a generalized form such as Sj z (tk) = f1(Bj z (tk), tk) and Sj x (tk) = f2(Bj z (tk), tk).

[0119] Since the spin is described as a two - dimensional vector, two components of Sj z (tk) and sj x (tk) appear. However, if Bj z (tk) is determined based on Equation (10), then Sj x (tK) is not necessary.

[0120] This corresponds to the fact that the spin state can be described only by Sj z (tk) with a value range of [-1, 1]. The final solution Sj zd should be Sj zd = -1 or 1. If Sj z (τ)>0, then Sj zd = 1; if Sj z (τ)<0, then Sj zd = -1.

[0121] Figure 12 shows a flowchart summarizing the solution algorithm in the above - mentioned Ising model. Here, tm = τ. Each step s1~s9 in the flowchart of this figure corresponds to the processing at a certain time in the timing chart of Figure 11 from time t = 0 to t = τ. That is, steps s2, s4, and s6 in the flowchart correspond to the above - mentioned Equations (9) and (10) at t = t1, tk + l, tm respectively. The final solution is at step s8, if sj z <0, then Sj zd = -1; if Sj z >0, then Sj zdIt is determined by setting it to 1 (s9).

[0122] So far, it has been shown how to solve the edging model, which is a specific problem, when it is expressed by Equation (3). Next, a specific example will be given and explained regarding how the edging model is expressed by Equation (3) including the local field gj and the variable - to - variable interaction Jij (i, j = 1, 2, …, N).

[0123] The edging model, for example, regarding an objective function that includes, as terms, an evaluation function related to the route of a vehicle and a function for constraints that becomes minimum when a predetermined constraint condition is satisfied, sets the presence or absence of use of each passing point of each vehicle as a spin, and sets the sensitivity between variables in the constraint function as the strength of the interaction between spins.

[0124] In this case, it is assumed that the local field gj is set as the influence degree that the value of the variable in, for example, the evaluation value of the above - mentioned evaluation function and the function for constraints that becomes minimum when a predetermined constraint condition is satisfied, gives to the objective function.

[0125] Through the above considerations, the variable - to - variable interaction Jij and the local field gj (regarding each term of the above - mentioned objective function) are specifically set, and the route of each vehicle is specified through the search for the ground state of the edging model expressed by Equation (3), that is, the search for the ground state where the above - mentioned objective function becomes minimum.

[0126] Note that what is calculated by the edging model and the annealing method is only "minimizing the objective function". Therefore, if there are constraint conditions that need to be satisfied when minimizing the objective function, they need to be incorporated into the objective function in some form.

[0127] For example, [Equation 11] JPEG0007685980000011.jpg16151 Consider the constraint condition of. If this constraint condition is converted into a "function that becomes minimum when the constraint condition is satisfied", it becomes the following equation.

[0128] [Equation 12] JPEG0007685980000012.jpg19151 Since the squared part is always a positive value, this equation reaches its minimum value only when the content inside the square is equal to 0. Since the content inside becomes 0 only when ΣXi - A = 0, solving the optimization problem for which this function reaches its minimum will automatically yield a solution that satisfies ΣXi = A.

[0129] Also, for example, in the above-mentioned annealing method, since it is necessary to incorporate the items to be constrained into the objective function, the objective function and the constraints will be treated with the same importance.

[0130] For example, assume there is an optimization problem as follows.

[0131] [Equation 13] JPEG0007685980000013.jpg47170 When this is changed to the formulation for annealing, it becomes as follows.

[0132] [Equation 14] JPEG0007685980000014.jpg20170 Here, P and Q are constants and are factors that determine which term to preferentially minimize. For example, when minimizing all three terms equally (i.e., solving the problem without bias in the strength of the constraints), values are set to balance between the terms, such as making P and Q equal.

[0133] On the other hand, for a problem setting where "the constraint of the third term is not as important as the constraint of the third term", a desired solution can be obtained by making the value of P, which is the coefficient of the term to be emphasized, larger than the value of Q.

[0134] According to the annealing method in this way, priorities can be assigned to the constraint conditions, and for the constraint conditions that are not highly emphasized, settings such as "satisfy as much as possible" can be made. Note that various settings related to the icing model shall be appropriately carried out in a form conforming to existing general concepts according to the conditions and information of this embodiment.

[0135] Based on the above, the icing model created by the route search support device 100 of this embodiment will be described.

[0136] First, the optimization problem to be solved by the route search support device 100 in the optimal route creation process s20 is to optimize the movement routes (waypoints) of each vehicle in order to equalize the supply rate Y of relief supplies in city c. c This optimization problem is generally represented by the following formula (15) of the objective function and formulas (16) and (17) of the constraint conditions (Yayoi Takahashi, Akiko Masaki, Taro Arai, Takahiro Mashima, Kazuo Ono, "Effectiveness of a Quantum-like Computer for Distribution Planning in Disasters", The 2022 Autumn Conference of the Japan Operations Research Society, 1-E-7).

[0137] [Equation 15] TIFF0007685980000015.tif45170

[0138] [Equation 16] TIFF0007685980000016.tif12170

[0139] [Equation 17] TIFF0007685980000017.tif12170

[0140] Here, the supply rate Y in formula (15) c is represented by the following formula (18).

[0141] [Equation 18] TIFF0007685980000018.tif12170

[0142] In the above formulas, N​city The number of local governments is \(n\), \(M\) is the set of local governments, and \(W\) c supplied is the amount of relief supplies already supplied in city \(c\), \(W\) is the demand (required amount) of relief supplies in city \(c\), \(W\) c demand is the supply amount of relief supplies in city \(c\) in route candidate \(j\) from the departure point to the arrival point, \(x\) c j is the binary value indicating route candidate \(j\) (specifically, the binary value indicating whether to adopt (1) or not to adopt (0)), \(D_j\) is the moving distance between bases in route candidate \(j\), \(J\) is the set of route candidates, \(P\) j is the supply base \(k\) (departure point) in route candidate \(j\), \(N\) is the number of vehicles at supply base \(k\), \(\varLambda\) is the set of vehicles, \(A\) kj is the evacuation shelter \(i\) that can be passed through in route candidate \(j\), \(V\) is the set of evacuation shelters, \(\alpha\) k truck is the weight value indicating the degree of emphasis on the ratio of the supply amount to the demand amount of relief supplies (to what extent the demand for relief supplies should be satisfied), \(\alpha\) ij is the weight value indicating the degree of emphasis on shortening the moving distance of vehicles. 1 2 city c

[0143] Note that \(N\) city , \(M\), \(W\) c supplied , \(W\) c demand , \(J\), \(N\) k truck , \(\varLambda\), \(V\), \(\alpha\) 1 , and \(\alpha\) 2 may be stored in advance by the route search support device 100, or the device may receive input of values from the user. Also, the route search support device 100 obtains \(W\) c j , \(D_j\), \(P\) kj , and \(A\) ij from the database created in route candidate creation process \(s1\) 0. Here, when the above optimization problem is converted into an integer model, it becomes as shown in the following formula (19). ​

[0144] [Formula 19] TIFF0007685980000019.tif35170

[0145] The first term on the right side of formula (19) is a term indicating the deviation of the ratio (sufficiency rate) of the supply volume to the demand volume of relief supplies between cities. When the Hamiltonian H is in the ground state (when the objective function reaches the minimum value), the value of this term is minimized, that is, the deviation is minimized. When the Hamiltonian H is in the ground state (when the objective function reaches the minimum value), the value of this term is minimized, that is, the deviation is minimized.

[0146] The second term on the right side of formula (19) is a term indicating the total shortage amount of relief supplies in the shelters. When the Hamiltonian H is in the ground state, the value of this term is minimized.

[0147] The third term on the right side of formula (19) is a term indicating the moving distance of all vehicles. When the Hamiltonian H is in the ground state, the value of this term is minimized.

[0148] The fourth term on the right side of formula (19) is a constraint function indicating the constraint on the number of vehicles arranged at the departure place (material aggregation base).

[0149] The fifth term on the right side of formula (19) is a constraint function indicating the constraint on the number of vehicles passing through each shelter i (in this embodiment, only once). For this shelter i, the shelters related to the dummy nodes among the unit nodes constituting the real nodes are not targeted, and only the shelters indicated by the unit nodes that are not dummy nodes among the unit nodes constituting the real nodes are targeted. That is, for the shelters related to the dummy nodes, each vehicle can pass through without restriction. For this shelter i, the shelters related to the dummy nodes among the unit nodes constituting the real nodes are not targeted, and only the shelters indicated by the unit nodes that are not dummy nodes among the unit nodes constituting the real nodes are targeted. That is, for the shelters related to the dummy nodes, each vehicle can pass through without restriction. For this shelter i, the shelters related to the dummy nodes among the unit nodes constituting the real nodes are not targeted, and only the shelters indicated by the unit nodes that are not dummy nodes among the unit nodes constituting the real nodes are targeted. That is, for the shelters related to the dummy nodes, each vehicle can pass through without restriction.

[0150] In formula (19), x, which is a binary value indicating the availability of the route candidate j, j corresponds to the spin in the Ising model. corresponds to the spin in the Ising model.

[0151] The optimal route creation processing unit 130 identifies the movement routes of each vehicle (determines whether to use route candidate j) by solving the above icing model. Also, the optimal route creation processing unit 130 calculates the amount of relief supplies supplied to each evacuation site and each city by each movement route.

[0152] As described above, the route search support device 100 of the present embodiment sets (contracts) a combination of one or two or more adjacent unit nodes that satisfy the first condition among the unit nodes indicating the possible waypoints in the process of the vehicle moving from the departure point to the arrival point as real nodes, searches for one or two or more real nodes that satisfy the second condition among the real nodes, sets (contracts) the searched real nodes as dummy nodes, uses the number of times of passing through the waypoints indicated by the unit nodes other than the dummy nodes among the unit nodes of the real nodes as a constraint condition, and calculates a value indicating whether to use the waypoints indicated by the unit nodes when optimizing the value of the objective function (leveling function) including the evaluation function (leveling function) related to the movement route of the vehicle, and outputs information indicating whether each waypoint is passed through.

[0153] That is, in using the objective function, the route search support device 100 of the present embodiment sets the reduced real nodes and the reduced dummy nodes in the real nodes based on the first condition and the second condition. Then, the route search support device 100 calculates whether each waypoint of the vehicle is passed through by the objective function with the number of times of passing through the waypoints indicated by the unit nodes other than the dummy nodes as a constraint condition.

[0154] In this way, the route search support device 100 of the present embodiment preliminarily narrows down the candidates for waypoints input to the objective variable for obtaining the optimal route of the vehicle by the real nodes obtained by reducing the unit nodes, while setting the constraints regarding the number of times of passing through the waypoints only for the unit nodes other than the dummy nodes, and supports efficiently searching for an appropriate route that passes through the waypoints to be passed even when there are a large number of waypoints for the vehicle.

[0155] Here, the route search support device 100 of the present embodiment sets a combination of unit nodes whose evaluation value indicating the strength of the connection between the unit nodes is equal to or greater than a predetermined value as real nodes with respect to the first condition.

[0156] Thereby, it is possible to extract combinations of waypoints that the moving body is highly likely to pass through.

[0157] In addition, the route search support device 100 of the present embodiment sets a combination of unit nodes whose travel time or distance between the unit nodes is equal to or greater than a predetermined value as real nodes with respect to the first condition.

[0158] Thereby, it is possible to extract combinations of waypoints that can shorten the travel time of the moving body.

[0159] Furthermore, the route search support device 100 of the present embodiment searches for unit nodes that need to pass through unit nodes other than the unit nodes constituting the real node from the unit nodes in the real node with respect to the second condition, and sets the searched unit nodes as dummy nodes.

[0160] Thereby, it is possible to explicitly set, as passable nodes, waypoints that the vehicle is highly likely to pass through when searching for each travel route, so that route search can be performed to optimize each travel route.

[0161] In addition, the route search support device 100 of the present embodiment searches for unit nodes (for example, unit nodes with a long distance from other unit nodes) whose evaluation value indicating the strength of the connection with other unit nodes is equal to or less than a predetermined value from the unit nodes in the real node with respect to the second condition, and sets the searched unit nodes as dummy nodes.

[0162] Thereby, it is possible to explicitly set, as passable nodes, waypoints that the vehicle is highly likely to pass through when searching for each travel route, so that route search can be performed to optimize each travel route.

[0163] In addition, the route search support device 100 of the present embodiment receives from the user via the setting screen 500 the setting of the upper limit (real node maximum length) of the connection between unit nodes in a real node by a combination of two or more unit nodes, and sets a combination of unit nodes that does not exceed the upper limit in the real node.

[0164] This prevents the number of constituent unit nodes of the real node from increasing and the candidates for the optimal route from being overly limited, and enables the optimal route of the vehicle to be calculated more reliably.

[0165] In addition, the route search support device 100 of the present embodiment receives from the user via the setting screen 500 the setting of the upper limit of the number of dummy nodes, and sets the dummy nodes so as not to exceed the upper limit.

[0166] This prevents overly complex routes from being searched and enables the optimal route of the moving body to be calculated more reliably.

[0167] In addition, the route search support device 100 of the present embodiment creates real nodes that satisfy constraint conditions (for example, impassability of routes between predetermined waypoints, time zones when entry into waypoints is possible) commonly applied to each vehicle, and sets additional different constraint conditions (for example, constraints on the number of vehicles at each departure point) according to the vehicle in the objective function.

[0168] In this way, for constraint conditions with a high load commonly applied to each vehicle, they are incorporated into the route candidate creation process s10 (processing related to real nodes) with a relatively small processing load as the number of variables increases. On the other hand, constraint conditions with different strengths according to the moving body are incorporated into the optimal route creation process s20 (processing of the objective function) with a large processing load as the number of variables increases. By performing such distribution of constraint conditions, the processing can be performed at high speed as a whole.

[0169] In addition, when the constraint conditions regarding the number of times of passing through the waypoints indicated by the unit nodes other than the dummy nodes among the unit nodes constituting the real nodes are satisfied, the route search support device 100 according to the present embodiment calculates the presence or absence of passing through each waypoint of the vehicle by using an Ising model in which the presence or absence of using the waypoint indicated by the unit node is a spin, and the sensitivity between variables in the constraint function is set as the strength of the interaction between the spins for the objective function including the constraint function that becomes minimum and the evaluation function regarding the movement route of the vehicle.

[0170] Thereby, it is possible to quickly search for an optimal route having a complicated movement route and waypoints.

[0171] In addition, the route search support device 100 according to the present embodiment receives from the user a designation of a node that is a candidate for the unit node via the setting screen 500.

[0172] Thereby, the user can narrow down the route to be used for the route search and perform the necessary route search.

[0173] In addition, the route search support device 100 according to the present embodiment receives from the user a setting of the upper limit of the number of variables of the objective function via the setting screen 500, and sets the variables of the objective function so as not to exceed the upper limit.

[0174] Thereby, the user can perform the route search while taking into account the balance between the accuracy of the route search and the processing speed of the route search.

[0175] In addition, the route search support system 1 according to the present embodiment includes a user terminal 200 that receives an input of waypoints through which the vehicle can pass from the user and displays the result output by the optimization model on the screen.

[0176] Thereby, the user can confirm the result while performing the route search at an appropriate timing.

[0177] The present invention is not limited to the above-described embodiment, and within the scope not departing from the gist thereof, The present invention can be implemented using any components. The above-described embodiments and modifications are merely examples, and the present invention is not limited to these contents as long as the features of the invention are not impaired. In addition, although various embodiments and modifications have been described above, the present invention is not limited to these contents. Other aspects conceivable within the scope of the technical idea of ​​the present invention are also included in the scope of the present invention.

[0178] For example, part of the hardware included in each device of this embodiment may be provided in another device.

[0179] Furthermore, each program of each device may be provided in another device, a program may consist of multiple programs, or multiple programs may be integrated into one program.

[0180] In this embodiment, the delivery of relief supplies by vehicles has been described as a target for solving the optimization problem, but the present invention can be applied to other general route search problems. For example, an optimization problem in marine transportation may be solved by solving the following problems: transportation demand, port (node) locations, routes (links) between ports that can be served, and CO 2 The emission regulation amount etc. are input and the optimized freight, ship and fuel between regions are calculated. Law, CO 2 Optimization model that outputs emissions and transportation costs (Shinnosuke Wanaka, Kazuo Hiegata, Yuji Horii, Effect of GHG Emission Reduction in International Maritime Transport Using Network Optimization Model The present invention can be applied to the evaluation of the optimization model. This allows us to handle optimization problems involving 100 times more ports. [Explanation of symbols]

[0181] 1 Route search support system, 100 Route search support device, 120 Graph analysis unit, 130 Optimal route creation processing unit, 140 result output processing unit

Claims

1. An apparatus for assisting in creating a route through which each of a plurality of moving objects moves via one or more waypoints, a unit node creation process for creating a unit node indicating a waypoint that can be passed through in the process of the moving object moving from a departure point to an arrival point, a real node setting process for setting, as a real node, one unit node that satisfies a first condition or a combination of two or more adjacent unit nodes among the unit nodes, a dummy node setting process for searching for one unit node that satisfies a second condition or a combination of two or more adjacent unit nodes among the unit nodes constituting the real node, and setting the searched unit node as a dummy node, an optimal route creation process for calculating a value indicating the availability of use of the waypoint indicated by the unit node when optimizing the value of an objective function including a predetermined evaluation function regarding the moving route of the moving object, with the number of passages through the waypoint indicated by the unit node other than the dummy node among the unit nodes constituting the real node as a constraint condition, a processing device that executes a result output process for outputting information indicating the passage or non-passage of each waypoint of the moving object based on the calculated value, a route search support device.

2. In the real node setting process, the processing device sets, as a real node, a unit node for which an evaluation value indicating the strength of the connection between the unit nodes of the combination is equal to or greater than a predetermined value as a combination of unit nodes that satisfy the first condition. The route search support device according to claim 1.

3. In the real node setting process, the processing device sets, as a real node, a unit node for which the length of the moving time or the length of the distance between the unit nodes of the combination is equal to or greater than a predetermined value as a combination of unit nodes that satisfy the first condition. The route search support device according to claim 2.

4. In the dummy node setting process, regarding the second condition, the processing device searches for one unit node or two or more adjacent unit nodes that need to be passed through in order to pass through a unit node other than the unit nodes constituting the real node from the unit nodes constituting the real node, and sets the searched unit node as a dummy node. The route search support device according to claim 1.

5. In the dummy node setting process, the processing device searches for unit nodes in the real node for which the evaluation value indicating the strength of the connection with other unit nodes is equal to or less than a predetermined value with respect to the second condition, and sets the searched unit nodes as dummy nodes. The route search support device according to claim 1.

6. The processing device executes a setting process for receiving from the user a setting of an upper limit on the connection between the unit nodes in the real node by a combination of two or more unit nodes, and in the real node setting process, sets a combination of two or more adjacent unit nodes that does not exceed the upper limit as a real node. The route search support device according to claim 1.

7. The processing device executes a setting process for receiving from the user a setting of an upper limit on the number of dummy nodes, and in the dummy node setting process, sets dummy nodes so as not to exceed the upper limit. The route search support device according to claim 1.

8. The processing device creates a real node that satisfies a predetermined constraint condition commonly applied to each moving body in the real node setting process, and in the optimal route creation process, sets different constraint conditions according to the moving body in the objective function. The route search support device according to claim 1.

9. The processing device regarding the objective function including, as terms, a constraint condition function that becomes minimum when a constraint condition regarding the number of times of passing through a waypoint indicated by a unit node other than the dummy node among the unit nodes constituting the real node in the optimal route creation process is satisfied, and an evaluation function regarding the moving route of the moving body, spins the availability of the waypoint indicated by the unit node, and calculates an Ising model in which the sensitivity between variables in the constraint condition function is set as the strength of the interaction between the spins, and in the result output process, outputs information indicating the passage or non-passage of each waypoint of the moving body to an output device based on the result of the calculation. The route search support device according to claim 1.

10. The processing device executes a setting process for receiving from the user a designation of a node that is a candidate for the unit node, the route search support device according to claim 1.

11. The processing device executes a setting process for receiving from the user a setting of an upper limit on the number of variables of the objective function, and in the optimal route creation, sets the variables of the objective function so as not to exceed the upper limit. The route search support device according to claim 1.

12. A method for supporting the creation of a route through which each of a plurality of moving objects moves through one or more waypoints, wherein an information processing device performs a unit node creation process for creating a unit node indicating a waypoint that can be passed through in the process of the moving object moving from a departure point to an arrival point, a real node setting process for setting, as a real node, one unit node that satisfies a first condition or a combination of two or more adjacent unit nodes among the unit nodes, a dummy node setting process for searching for one unit node that satisfies a second condition or a combination of two or more adjacent unit nodes among the unit nodes constituting the real node, and setting the searched unit node as a dummy node, an optimal route creation unit process for calculating a value indicating the availability of use of the waypoint indicated by the unit node when optimizing the value of an objective function including a predetermined evaluation function related to the moving route of the moving object, with the number of times of passing through the waypoint indicated by the unit node other than the dummy node among the unit nodes constituting the real node as a constraint condition, and a result output process for outputting information indicating the passage or non-passage of each waypoint of the moving object based on the calculated value. Route search support method.

13. A route search support system for supporting the creation of a route through which each of a plurality of moving objects moves through one or more waypoints, including a unit node creation process for creating a unit node indicating a waypoint that can be passed through in the process of the moving object moving from a departure point to an arrival point, a real node setting process for setting, as a real node, one unit node that satisfies a first condition or a combination of two or more adjacent unit nodes among the unit nodes, a dummy node setting process for searching for one unit node that satisfies a second condition or a combination of two or more adjacent unit nodes among the unit nodes constituting the real node, and setting the searched unit node as a dummy node, an optimal route creation process for calculating a value indicating the availability of use of the waypoint indicated by the unit node when optimizing the value of an objective function including a predetermined evaluation function related to the moving route of the moving object, with the number of times of passing through the waypoint indicated by the unit node other than the dummy node among the unit nodes constituting the real node as a constraint condition, and a route creation support device including a processing device that performs a result output process for outputting information indicating the passage or non-passage of each waypoint of the moving object based on the calculated value. An input process that receives from a user an input of a waypoint through which the mobile object can pass, and a display process that displays on a screen the information output by the result output process. An information processing apparatus comprising a processing apparatus that executes the processes. A route search support system configured to include the information processing apparatus.

Images