KR102374595B1 - Vehicle routing method and device using noising scheme - Google Patents

Abstract

Disclosed is a vehicle routing method and a vehicle routing apparatus using noise technique. The vehicle routing method using the noise technique executed by a disclosed processor comprises the following steps of: removing a predetermined number of nodes from a route of one or more vehicles based on randomly selected nodes and closest nodes of the selected nodes among a plurality of nodes included in the route; and adding the removed nodes to the route of the one or more vehicles. The closest node is determined by adding a noise value to a distance from a corresponding node to each of the other nodes. By applying noise to resetting of the closest node, search for a best solution of a vehicle route problem is efficiently performed.

Description

Vehicle routing method and apparatus using noise technique {VEHICLE ROUTING METHOD AND DEVICE USING NOISING SCHEME}

The following description relates to a vehicle routing method and apparatus using a noise technique, and more specifically, noise in the reconfiguration of the nearest node used in the destruction process of LNS (Large Neighborhood Search) and/or the local search process. It relates to a vehicle routing method and apparatus to apply.

The Vehicle Routing Problem (VRP) includes any problem involving vehicles visiting a customer, and can be described in a variety of ways, such as vehicle scheduling, vehicle work planning, and delivery planning. VRP can be applied to solve a variety of problems that may be common in real-world situations, including mail collection, commuter bus (eg school bus) routing, last mile delivery, and maintenance tours.

In general, a VRP may contain a set of homogeneous vehicles with a fixed capacity capable of serving a set of customers from a single depot. All vehicles must depart from the point of departure and return to the point of departure. Restrictions such as route length or time restrictions may limit the distance the vehicle travels. The routing goal is to allocate a set of deliveries to each vehicle to serve all customers, while minimizing the total distance traveled or the total time consumed by the vehicle. A standard VRP (standard VRP) can often be referred to as a Capacitated Vehicle Routing Problem (CVRP). Various types of VRP variants can exist, such as the Multi-Depot Vehicle Routing Problem (MVRP), in which vehicles depart from multiple warehouses, and the Open Vehicle Route Problem (OVRP), in which vehicles do not return to the warehouse.

A vehicle routing method using a noise technique, executed by a processor according to an embodiment, is based on a node randomly selected from among a plurality of nodes included in a path of one or more vehicles and a closest node of the corresponding node to remove a predetermined number of nodes from the corresponding path; and adding the removed node to the path of the one or more vehicles, wherein the nearest node is determined by adding a noise value to a distance from a corresponding node to each of the remaining nodes.

In the vehicle routing method according to an embodiment, the nearest node may be determined as a node corresponding to a minimum value among result values obtained by adding a noise value to distances from the corresponding node to each of the remaining nodes.

In the vehicle routing method according to an embodiment, the nearest node may be re-determined whenever the removing and the adding are repeated a predetermined number of times.

In the vehicle routing method according to an embodiment, the nearest node may be determined differently for each of the plurality of threads when the vehicle routing method is performed in parallel in each of a plurality of threads.

In the vehicle routing method according to an embodiment, the removing may include removing the selected node and the nearest node from among the plurality of nodes.

In the vehicle routing method according to an embodiment, the process of removing a node randomly selected from among the remaining non-removed nodes among the plurality of nodes and a node closest to the node may be repeated until a predetermined condition is satisfied. .

In the vehicle routing method according to an embodiment, the removing may include: removing a first node randomly selected from among the plurality of nodes and a node closest to the first node; repeating the process of removing the closest node of the removed nearest node from the corresponding path in response to the case that the closest node of the removed closest node is not the already removed node; and a third node randomly selected from among the remaining non-removed nodes among the plurality of nodes and a nearest node of the third node in response to a case in which the closest node of the removed nearest node is an already removed node. may include the step of removing

In the vehicle routing method according to an embodiment, the removing may include randomly selecting a seed node from among the plurality of nodes; removing from the reference path previous nodes or subsequent nodes of the seed node from the reference path including the seed node; A predetermined condition is satisfied in the step of setting a new seed node in response to a case where the closest node of the seed node is not a node that has already been removed, and removing previous or subsequent nodes of the new seed node from the corresponding reference path. repeat until and in response to a case in which the closest node of the seed node is a node that has already been removed, randomly selects a new seed node from among the remaining non-removed nodes among the plurality of nodes, and selects previous nodes of the selected new node or Thereafter, it may include removing the nodes from the corresponding reference path.

In the vehicle routing method according to an embodiment, the removing may be repeatedly performed until a ratio of the removed nodes among the plurality of nodes satisfies a predetermined threshold.

In the vehicle routing method according to an embodiment, the removing and the adding may be repeated until a predetermined condition is satisfied.

The vehicle routing method according to an embodiment may further include, after the adding is performed, determining whether to swap any one of the plurality of nodes with a node closest to the corresponding node. there is.

In the vehicle routing method according to an embodiment, the noise value belongs to a minimum value and a maximum value of noise determined based on a lower n% average of distances extracted from a distance matrix expressing distances between the plurality of nodes, and the n may be a real number between 0 and 100.

A vehicle routing apparatus according to an embodiment includes one or more processors, wherein the one or more processors are randomly selected from among a plurality of nodes included in a path of one or more vehicles and a closest node of the node. based on a predetermined number of nodes are removed from the corresponding path, the removed node is added to the path of the one or more vehicles, and the nearest node adds a noise value to the distance from the corresponding node to each of the remaining nodes. is determined by

According to an embodiment, by applying noise to the reconfiguration of the nearest node used in the destruction process of the LNS and/or the local search process, it is possible to effectively perform the search for the best solution to the vehicle path problem.

According to an embodiment, when performing LNS, the best solution can be effectively obtained even with a small number of random number calls by resetting the nearest node based on noise at each specific iteration.

1 is a diagram for explaining an LNS technique for solving a vehicle path problem according to an embodiment.
2 is a view for explaining a vehicle routing method according to an embodiment.
3 to 5 are diagrams for explaining examples of a destruction technique using a noise technique according to an embodiment.
6 and 7 are diagrams for explaining a process of determining an amount of noise according to an exemplary embodiment.
8 is a diagram for explaining an example in which a vehicle routing method is performed in parallel in each of a plurality of threads according to an embodiment.
9 is a view showing a vehicle routing apparatus according to an embodiment.

Specific structural or functional descriptions of the embodiments are disclosed for purposes of illustration only, and may be changed and implemented in various forms. Accordingly, the actual implementation form is not limited to the specific embodiments disclosed, and the scope of the present specification includes changes, equivalents, or substitutes included in the technical spirit described in the embodiments.

Although terms such as first or second may be used to describe various elements, these terms should be interpreted only for the purpose of distinguishing one element from another. For example, a first component may be termed a second component, and similarly, a second component may also be termed a first component.

When a component is referred to as being “connected to” another component, it may be directly connected or connected to the other component, but it should be understood that another component may exist in between.

The singular expression includes the plural expression unless the context clearly dictates otherwise. In this specification, terms such as "comprise" or "have" are intended to designate that the described feature, number, step, operation, component, part, or combination thereof exists, and includes one or more other features or numbers, It should be understood that the possibility of the presence or addition of steps, operations, components, parts or combinations thereof is not precluded in advance.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and should not be interpreted in an ideal or excessively formal meaning unless explicitly defined in the present specification. does not

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. In the description with reference to the accompanying drawings, the same components are assigned the same reference numerals regardless of the reference numerals, and overlapping descriptions thereof will be omitted.

1 is a diagram for explaining an LNS technique for solving a vehicle path problem according to an embodiment.

Referring to FIG. 1 , a routing method of an LNS technique according to an embodiment is illustrated. The first stage 110 may represent an initial sub-optimal solution of a VRP instance with a warehouse (ie, a depot) and three vehicle routes in the center. In the second step 120 , some of the nodes representing customers the vehicle should visit may be removed (ie, destroyed) from the path. The removed node may be included in another path (ie, repaired) to derive the best solution, and this process may be expressed in the third step 130 .

According to an embodiment, pointer information on the Closest Node, which is the closest node to the node removed in the destruction process described above and the local search process to be described later, is required, and the nearest node A noise technique may be applied to the setting. In other words, the nearest node of each node can be reset with some noise, and the best routing solution can be obtained more effectively through this noise technique. Hereinafter, it will be described in more detail with reference to the drawings.

2 is a view for explaining a vehicle routing method according to an embodiment.

Referring to FIG. 2 , a vehicle routing method performed by one or more processors included in a vehicle routing apparatus according to an embodiment is illustrated.

In step 210 , the vehicle routing apparatus may generate an initial solution including each of the plurality of nodes in the path of one or more vehicles. For example, a vehicle routing device may simply randomly generate an initial solution. A vehicle is randomly selected for each node, and in this case, a probability that a specific vehicle is selected for any node may be the same. The corresponding node may be included in the path of the selected vehicle.

And, the vehicle routing apparatus may record the nearest node of each node as a pointer. For example, if the VRP size is 1000, the nearest node of node 1 is Distance(1->2), Distance(1->3), Distance(1->4), ..., Distance(1- >1000), the node with the shortest distance from node 1 may be set as the closest node to node 1.

In step 220 , the vehicle routing device may reconfigure the nearest node of each of the plurality of nodes. All nodes can each point to the closest node to itself (ie, the nearest node). The nearest node is reset every specific iteration, and since each reset includes noise, the nodes do not always point to the same node. As will be described later in detail, in an embodiment in which routing is performed in parallel, the same node included in different threads at the same time may point to a different node as the nearest node. The closest node set in this way can be actively used in the destruction process and the local search process.

As such, the present invention may utilize a unique and effective noise method. Conventionally, noise is applied by adding a limited random number while operating the destruction or recovery heuristic. For example, when greedy repair is in progress, a random number is added to the actual distance increment, resulting in noise at the greedy insertion location for each request. This requires generating a myriad of random numbers for every customer node, every insertion point, every iteration, and every thread of work. On the other hand, the method newly proposed in this specification does not require random number generation for every iteration, which can save considerable time. In the proposed scheme, a pointer to the most relevant (ie, closest) node can be applied to each node. The pointer is initialized again when the corresponding iteration is completed, and noise may occur in the distance between nodes. A pointer reset after initialization may be maintained until the next iteration.

It may be a unique feature of the present invention to actively utilize a pointer to the nearest node from each node containing noise. The noise value may be an important parameter affecting the operation of the noise term. If the noise value is too small, no effect by noise occurs, and conversely, if the noise value is too large, for example, if the noise exceeds the distance between nodes, the nearest node decision can be made completely random. Therefore, when setting the noise value, it is also necessary to refer to the distance matrix. By sampling and averaging the distance between nodes and then multiplying by another parameter called noiseMultipier, the noise value can be determined. The noiseMultipier value can be determined empirically. Table 1 below shows an example of a pseudo code for setting a noise parameter.

Pseudo Code for Setting Noise Parameter

The closest node can be searched for each node with the noise value. Table 2 below shows an example of a pseudo code that sets a pointer to the nearest node by generating noise in the distance between two nodes.

Pseudo Code for Setting Pointers to the Closest Nodes

If the noise-based resetting of the nearest node is applied to the example of setting the nearest node of node 1 above, noise can be applied as follows: Distance(1->2) + Random [Minimum Noise~Maximum Noise] ,

Distance(1->3) + Random[Min Noise~Max Noise],

Distance(1->4) + Random[Min Noise~Max Noise],

...,

Distance(1->1000) + Random[Min Noise~Max Noise]

The vehicle routing device may reset the node having the shortest distance among them as the closest node to the node 1. Due to this, the number of random number calls is as follows.

Random Number calls to node 1: 1000 times

Random Number calls to node 2: 1000

Random Number calls to node n: 1000

Node Random Number calls to Node: 1000 times

That is, the total number of random number calls becomes one million times. The closest node once set may be reset after a certain number of iterations (eg, it may correspond to the Max Equilibrium Count of the Simulated Annealing algorithm, but may vary depending on the setting). Accordingly, the number of random number calls can be effectively reduced.

In step 230 , the vehicle routing device may perform a destruction process on the current solution based on the previously set nearest node. The vehicle routing apparatus may remove some of the plurality of nodes from a corresponding path according to a predetermined destruction size. For example, the fracture size may be selected within the range of 1 to 50% of the total number of nodes, but the embodiment is not limited thereto. The upper and lower limits of the fracture size may be tunable parameters.

Any one of the three destruction techniques described below may be selected. This selection may be performed at every iteration. If Random Related Destroy is selected in random iteration and the destruction size is 40%, Random Related Destroy can be repeatedly performed until the number of removed nodes is 40% of the total number of nodes. The three destruction techniques will be described in detail with reference to FIGS. 3 to 5 .

In step 240 , the vehicle routing device may restore the removed nodes by including them in the existing path. For example, recovery techniques include Greedy Repair and Regret Repair. Greedy recovery is a technique for retrieving a position with the lowest insertion cost for each request. For each request, the lowest insertion position can be retrieved by comparing two states. One of the two states may be a current state, and the other may be a state in which a node is to be inserted. Regret recovery can be similar to greedy recovery in that requests are inserted at the cheapest (greedy) location, except that it considers the order in which repairs are requested before other nodes. Greedy recovery is performed based on the lowest insertion cost, whereas regret recovery can be performed based on the regret value. In this case, regret value can be very similar to opportunity cost.

In step 250, the vehicle routing device may perform a local search for the restored solution. The local search is not always performed, and may or may not be performed according to embodiments.

The above-described noise technique may be applied to the local search as well. After the solution is destroyed and repaired, several additional permutations can be created for further refinement before performing the next destruction process. The local retrieval method described in Table 3 below can use a swap operation to retrieve local enhancements. In the solution, the destruction size of a randomly selected node, and the distance difference when the selected node is swapped to the nearest node of that node, can be calculated. If it is confirmed that swapping improves the solution, the two nodes can actually be swapped. Such a local search is not always performed in all iterations, and, for example, a probability of being performed may be set to 50%.

Pseudo Code for the Additional Local Improvement Method

In other words, a local search requires little effort after recovery after destruction is performed, that is, after a certain % (eg 40%) of nodes have been excluded from the solution, and all nodes that have been excluded are included in the solution again. Quickly review to see if a better solution can be found only with the help of a method, and if improvement is possible, it may be applied to the solution. Local search can basically apply the swap method. Swap is to exchange the positions of two nodes. For example, the path of vehicle 1 is [Node 0 -> Node 1 -> Node 2 -> Node 0], and the path of vehicle 2 is [Node 0 -> Node 3]. When -> node 4 -> node 0], if node 1 and node 3 are swapped, the path of vehicle 1 is [Node 0 -> Node 3 -> Node 2 -> Node 0], and the path of vehicle 2 is [ node 0 -> node 1 -> node 4 -> node 0]. In this case, to which node to apply swap may be randomly selected, and to which node to swap with which node to swap may be determined by a node closest to the node.

For example, the path of vehicle 1 is [Node 0 -> Node 1 -> Node 2 -> Node 0], and the path of Vehicle 2 is [Node 0 -> Node 3 -> Node 4 -> Node 0]. In , when node 2 is selected as a swap target, at this time, node 2 is exchanged with the closest node of node 2, and if the exchange result is judged to be a better solution, it can be applied to the solution.

This local search method can be much faster and more efficient than considering the number of all swap cases. Furthermore, the number of nodes to be swapped may be determined based on the destruction size. For example, if the total number of nodes is 100 and the destruction size is 40%, the swap target node selection and swap attempt can be performed 40 times.

In step 260, the vehicle routing device may determine whether to stop the operation of the vehicle routing method. The vehicle routing device may determine whether to stop based on various criteria. For example, the vehicle routing device may determine whether to stop based on whether the performance of the current solution satisfies a predetermined condition. Alternatively, the vehicle routing device may determine whether to stop based on whether the current solution is improved over the previous solution. This description does not limit the exemplary embodiment of the criterion for determining whether to stop, and in addition, various criterion for determining whether to stop may be applied without limitation.

If it is determined that the vehicle routing device does not stop the operation, step 270 is subsequently performed. On the contrary, if it is determined that the vehicle routing device stops the operation, the vehicle routing method shown in FIG. 2 may be terminated.

In step 270 , the vehicle routing device may determine whether the number of iterations corresponds to a predetermined number. If the number of iterations does not correspond to a predetermined number, step 230 is performed without resetting the nearest node. can be performed. This effectively achieves the best solution by resetting the nearest node at every specific iteration.

3 to 5 are diagrams for explaining examples of a destruction technique using a noise technique according to an embodiment.

Referring to FIG. 3 , a method of operating Random Related Destroy according to an embodiment is illustrated. Random related destruction may also be referred to as Distance Random Destroy.

In step 310, the vehicle routing apparatus may randomly select a node that has not been removed from among the plurality of nodes. In step 320 , the vehicle routing device may remove the selected node and a node closest to the node. If the closest node to the node has already been removed, only the selected node can be removed. In step 330 , the vehicle routing device may determine whether a fracture magnitude has been satisfied. For example, if the destruction size is 40%, it may be determined whether the number of nodes removed in step 320 is 40% or more of the total nodes. If the fracture size has not been satisfied, step 310 may then be performed. Conversely, if the fracture size is satisfied, the main destruction operation is terminated, and the recovery process may be subsequently performed.

Such random related destruction may be a technique for removing nodes with similar properties together. This may enable exchanges between nodes that are in some way similar. In random-associated destruction, relevance can be measured based on the distance between nodes. Therefore, the term 'related' may be replaced with 'nearest' in the random related destruction and the sequence related destruction to be described in FIG. 4 . A pointer to the nearest node may not actually point to the nearest node due to the noise described above. Also, for the extension of random related destruction, more attributes (eg, adjacent time windows) may be considered. Table 4 below shows examples of pseudo codes of random related destruction.

Pseudo Code for Random Related Destroy Heuristic Method

4 , a method of operation of sequence-related destruction according to an embodiment is illustrated. Sequence-related destruction may also be referred to as Distance Sequence Destroy.

In step 410, the vehicle routing apparatus may randomly select a node that is not removed from among the plurality of nodes. In step 420 , the vehicle routing device may remove the selected node and a node closest to the node. In step 430 , the vehicle routing device may determine whether the vehicle routing device has satisfied the destruction magnitude. For example, if the destruction size is 40%, it may be determined whether the number of nodes removed in step 420 is 40% or more of the total nodes. If the fracture size has not been satisfied, step 440 may then be performed. Conversely, if the fracture size is satisfied, the main destruction operation is terminated, and the recovery process may be subsequently performed.

In step 440 , the vehicle routing device may determine whether the closest node of the removed closest node is already a removed node. If the closest node of the removed nearest node is not the already removed node, in step 450 the vehicle routing device removes the closest node of the removed closest node, and step 430 may be performed subsequently. . In other words, as long as the closest node of the last removed nearest node is a node that has already been removed or the destruction size is not satisfied, the closest nodes may be sequentially removed. If in step 440 the closest node of the last removed nearest node is already a removed node, step 410 may then be performed.

Such sequence-related destruction can also remove the nearest node that is still included in the current solution. This sequential method may select the closest node of the previously referenced nearest node as the next node. Through this, some clusters of closely located nodes may be sequentially removed. This chain of removals can continue if the next nearest node has not yet been excluded from the current solution. When the removal chain is broken, a new node can be randomly chosen to start a new sequence. 5 shows an example of a pseudo code of sequence related destruction.

Pseudo Code for Sequence Related Destroy Heuristic Method

Referring to FIG. 5 , an operation method of Sequence Cluster Destroy according to an embodiment is illustrated. Sequence cluster destruction may also be referred to as cluster distance destruction.

In step 510, the vehicle routing apparatus may select a node that has not been removed from among the plurality of nodes as a seed node. In step 520 , the vehicle routing apparatus may remove previous nodes or subsequent nodes of the seed node from the reference path in the reference path including the seed node. The vehicle routing device may select either “before” or “after” as the direction of removal, where each probability may be 0.5.

For example, node 3 is selected as the seed node, the path of the vehicle including the seed node is [Node 0 -> Node 1 -> Node 2 -> Node 3 -> Node 4 -> Node 5 -> Node 0] , if 'after' is selected as the direction, node 4 and node 5 are removed from the corresponding path, so that the vehicle path becomes [Node 0 -> Node 1 -> Node 2 -> Node 3 -> Node 0]. . In this case, whether to remove the seed node together may be determined according to an embodiment.

In step 530 , the vehicle routing device may determine whether a fracture magnitude has been satisfied. For example, if the destruction size is 40%, it may be determined whether the number of nodes removed in step 520 is 40% or more of the total nodes. If the fracture size has not been satisfied, step 540 may then be performed. Conversely, if the fracture size is satisfied, the main destruction operation is terminated, and the recovery process may be subsequently performed.

In step 540 , the vehicle routing device may determine whether the closest node of the seed node is a node that has already been removed. If the closest node of the seed node is not a node that has already been removed, in step 550 , the vehicle routing apparatus sets the closest node of the seed node as a new seed node, and step 520 may be subsequently performed. Conversely, if the closest node of the seed node has already been removed, step 510 may then be performed.

In this sequence cluster destruction, if the closest node of the current seed node still exists in the current solution, the closest node of the current seed node may be selected as the next seed node. Table 6 below shows an example of a pseudo code of sequence cluster destruction.

Pseudo Code for Sequence Cluster Destroy Heuristic Method

6 and 7 are diagrams for explaining a process of determining an amount of noise according to an exemplary embodiment.

Referring to FIG. 6 , an exemplary node distribution and distance matrix for determining an amount of noise is shown in accordance with one embodiment. In the distance matrix, the distance between two nodes corresponding to the source and destination is expressed as a matrix.

By utilizing the noise-applied nearest node in the aforementioned destruction process and local search process, it can contribute to effective search for the best solution. The amount of noise applied to the nearest node may be expressed as a parameter. In this case, an important value may be set first, and this value may be an important parameter affecting the operation of the noise term. If the amount of noise is too small, noise is meaningless, conversely, noise exceeding the distance between nodes is meaningless because it makes the selection purely random. To obtain meaningful noise, the pivotal value must be set appropriately.

In the node distribution example shown in the upper part of FIG. 6 , the average value of the distances between nodes densely close to each other is meaningful, but the amount of noise determined based on the average value including the distances between distant nodes is too large. Therefore, it is necessary to focus only on d[i,j] when the distance between nodes is close in accordance with the reason for the need for noise and its intention.

In other words, it is meaningless to extract some element d[i,j] from the distance matrix and then determine the amount of noise by using the average value of the extracted elements as it is, and among the extracted elements, the lower n% (here, n is The amount of noise should be determined based on the average value for a real number between 0 and 100). As a specific example, the amount of noise may be determined by multiplying an average value for the lower n% by an adjustment parameter. As such, a significant amount of noise can be effectively derived through the new method focusing on the lower n% of the extracted elements. In FIG. 7, various examples focusing on the lower n% described above are shown.

8 is a diagram for explaining an example in which a vehicle routing method is performed in parallel in each of a plurality of threads according to an embodiment.

Referring to FIG. 8 , the aforementioned destruction, recovery, and local search are performed multiple times in each of a plurality of threads according to an embodiment, and the best solution of each thread may be shared with each other in a specific iteration.

As the vehicle routing device consists of one master device and several slave devices, the LNS can be spatially parallelized. Each slave device can perform the best solution search by randomly combining the removal and recovery processes. After a certain run time, each slave device can pass the best solution to the master device for synchronization. The master device may select the best solution from among the solutions received from each slave device and share it with a plurality of slave devices. Each of the plurality of slave devices can randomly combine the removal and recovery processes in the shared solution to perform the best solution search. This parallel structure can diversify solution search and effectively improve solution quality.

9 is a view showing a vehicle routing apparatus according to an embodiment.

Referring to FIG. 9 , the vehicle routing apparatus 900 according to an embodiment includes a memory 910 and a processor 920 . The memory 910 and the processor 920 may communicate with each other through a bus 930 .

The memory 910 may include computer-readable instructions. The processor 920 may perform the aforementioned operations as an instruction stored in the memory 910 is executed in the processor 920 . The memory 910 may be a volatile memory or a non-volatile memory.

The processor 920 is a device that executes instructions or programs, or controls the vehicle routing device 900 , for example, a central processing unit (CPU), a graphic processing unit (GPU), or a neural processing unit (NPU). and the like. The processor 920 removes a predetermined number of nodes from the path based on a node randomly selected from among a plurality of nodes included in the path of one or more vehicles and a closest node of the node, and removes the removed node from the path. By adding a node to the path of one or more vehicles, the nearest node may be determined by adding a noise value to the distance from the corresponding node to each of the remaining nodes.

Vehicle routing device 900 is a variety of computing devices, such as smart phones, tablets, laptops, personal computers, smart watches, smart glasses, various wearable devices such as smart clothes, smart speakers, smart TVs, smart refrigerators, various home appliances, smart cars, It may be implemented as a part of various computing devices, such as a smart kiosk, an Internet of Things (IoT) device, a walking assist device (WAD), a drone, or a robot.

In addition, the above-described operation can be processed with respect to the vehicle routing device 900 .

The embodiments described above may be implemented by a hardware component, a software component, and/or a combination of the hardware component and the software component. For example, the apparatus, methods and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate (FPGA) array), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions, may be implemented using a general purpose computer or special purpose computer. The processing device may execute an operating system (OS) and a software application running on the operating system. A processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For convenience of understanding, although one processing device is sometimes described as being used, one of ordinary skill in the art will recognize that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that can include For example, the processing device may include a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as parallel processors.

Software may comprise a computer program, code, instructions, or a combination of one or more thereof, which configures a processing device to operate as desired or is independently or collectively processed You can command the device. The software and/or data may be any kind of machine, component, physical device, virtual equipment, computer storage medium or apparatus, to be interpreted by or to provide instructions or data to the processing device. , or may be permanently or temporarily embody in a transmitted signal wave. The software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored in a computer-readable recording medium.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination, and the program instructions recorded on the medium are specially designed and configured for the embodiment, or may be known and available to those skilled in the art of computer software. may be Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floppy disks. - includes magneto-optical media, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

The hardware devices described above may be configured to operate as one or a plurality of software modules to perform the operations of the embodiments, and vice versa.

As described above, although the embodiments have been described with reference to the limited drawings, a person skilled in the art may apply various technical modifications and variations based thereon. For example, the described techniques are performed in an order different from the described method, and/or the described components of the system, structure, apparatus, circuit, etc. are combined or combined in a different form than the described method, or other components Or substituted or substituted by equivalents may achieve an appropriate result.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (13)

In a vehicle routing method using a noise technique, executed by a processor,
removing a predetermined number of nodes from a corresponding path based on a node randomly selected from among a plurality of nodes included in a path of one or more vehicles and a closest node of the corresponding node; and
adding the removed node to the path of the one or more vehicles;
including,
The nearest node is determined by adding a noise value to the distance from the corresponding node to each of the remaining nodes.
Vehicle routing method.
According to claim 1,
The nearest node is determined as a node corresponding to the minimum value among the results of adding a noise value to the distances from the corresponding node to each of the remaining nodes,
Vehicle routing method.
According to claim 1,
The nearest node is re-determined every time the removing and adding are repeated a predetermined number of times,
Vehicle routing method.
According to claim 1,
The nearest node is determined differently for each of the plurality of threads when the vehicle routing method is performed in parallel on each of the plurality of threads,
Vehicle routing method.
According to claim 1,
The removing step
removing the selected node and the nearest node from among the plurality of nodes
Vehicle routing method.
According to claim 1,
The removing step
repeating the process of removing a node randomly selected from among the remaining nodes that have not been removed among the plurality of nodes and a node closest to the node until a predetermined condition is satisfied,
Vehicle routing method.
According to claim 1,
The removing step
removing a first node randomly selected from among the plurality of nodes and a node closest to the first node;
repeating the process of removing the closest node of the removed closest node from the corresponding path in response to the case that the closest node of the removed nearest node is not the already removed node; and
In response to a case in which the closest node of the removed nearest node is an already removed node, a third node randomly selected from among the remaining non-removed nodes among the plurality of nodes and the nearest node of the third node are selected. steps to remove
containing,
Vehicle routing method.
According to claim 1,
The removing step
randomly selecting a seed node from among the plurality of nodes;
removing nodes before or after the seed node from the reference path in the reference path including the seed node;
A predetermined condition is satisfied in the step of setting a new seed node in response to a case where the closest node of the seed node is not a node that has already been removed, and removing previous or subsequent nodes of the new seed node from the corresponding reference path. repeat until and
In response to a case in which the closest node of the seed node is a node that has already been removed, a new seed node is randomly selected from among the remaining nodes that have not been removed from among the plurality of nodes, and previous or subsequent nodes of the selected new node are selected. removing nodes from the reference path
containing,
Vehicle routing method.
According to claim 1,
The removing step is repeatedly performed until a ratio of the removed nodes among the plurality of nodes satisfies a predetermined threshold,
Vehicle routing method.
According to claim 1,
The removing and adding are repeated until a predetermined condition is satisfied,
Vehicle routing method.
According to claim 1,
After the adding is performed, determining whether to swap any one of the plurality of nodes with the nearest node of the corresponding node
further comprising
Vehicle routing method.
According to claim 1,
The noise value is
It belongs between the minimum and maximum values of noise determined based on the average of the lower n% of the distances extracted from the distance matrix expressing the distances between the plurality of nodes, wherein n is a real number between 0 and 100,
Vehicle routing method.
one or more processors;
the one or more processors
A predetermined number of nodes are removed from the path based on a node randomly selected from among a plurality of nodes included in the path of one or more vehicles and a closest node of the node,
add the removed node to the path of the one or more vehicles;
The nearest node is determined by adding a noise value to the distance from the corresponding node to each of the remaining nodes.
vehicle routing device.

Images