Network topology generation method and device, computer equipment and storage medium
By augmenting the historical data of drones and other devices, more sample data are generated, and network topology is trained using preset learning models, the problem of insufficient equipment data is solved and the efficiency and effect of network topology generation is improved.
Patent Information
- Application Number
- CN202411366982.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-29
- Publication Date
- 2024-11-26
AI Technical Summary
Existing network topology generation methods require a large amount of sample data, especially when the equipment data is insufficient, such as drones, which are difficult to meet the requirements of machine learning.
Data augmentation processing is performed by obtaining the historical data of the device to be trained, more sample data is generated, and inputting it into the preset learning model for training to output the network topology.
Reduce the demand for sample data volume and improve the efficiency and effectiveness of network topology generation, especially in the case of insufficient data.
Smart Images

Figure CN119030881A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of network topology technology, and in particular to a network topology generation method, device, computer equipment and storage medium. Background Art
[0002] In today's information age, network topology design plays a vital role in the fields of communication, computing, and data processing. Traditional network topology generation methods mainly rely on manual experience for network design, which is not only time-consuming and labor-intensive, but also difficult to respond quickly to changes in demand. With the advancement of machine learning technology, machine learning has gradually been applied to the generation of network topology. However, current machine learning methods require a large amount of sample data for training, and for many devices, the sample data they can provide is relatively small. For example, for drone equipment, due to legal and regulatory restrictions, the flight data is usually small and difficult to meet the requirements of machine learning. Summary of the invention
[0003] The embodiments of the present invention provide a network topology generation method, apparatus, computer equipment and storage medium, aiming to solve the problem of insufficient sample data required by current network topology generation methods.
[0004] In a first aspect, an embodiment of the present invention provides a method for generating a network topology, the method comprising: Acquire historical data of the device to be trained, and perform data enhancement processing on the historical data to obtain data to be trained; A preset learning model is obtained, and the data to be trained is used as input data of the preset learning model so that the preset learning model trains the data to be trained to output a network topology.
[0005] In a second aspect, an embodiment of the present invention further provides a network topology generation device, the device comprising: A first acquisition unit is used to acquire historical data of the device to be trained, and perform data enhancement processing on the historical data to obtain data to be trained; The first training unit is used to obtain a preset learning model, and use the data to be trained as input data of the preset learning model so that the preset learning model trains the data to be trained to output a network topology.
[0006] In a third aspect, an embodiment of the present invention further provides a computer device, which includes a memory and a processor connected to the memory, wherein a computer program is stored in the memory, and the processor implements the above method when executing the computer program.
[0007] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, wherein the storage medium stores a computer program, and the computer program can implement the above method when executed by a processor.
[0008] The embodiments of the present invention provide a network topology generation method, device, computer equipment and storage medium. The method includes: obtaining historical data of a device to be trained, and performing data enhancement processing on the historical data to obtain data to be trained; obtaining a preset learning model, and using the data to be trained as input data of the preset learning model so that the preset learning model trains the data to be trained to output a network topology. The embodiments of the present invention can obtain historical data that has been generated by the device to be trained, and perform data enhancement processing on the obtained historical data to obtain data to be trained, thereby making up for the problem of insufficient sample data, and then input the data to be trained into the preset learning model, train the data to be trained through the preset learning model, and finally output the network topology, thereby reducing the demand for sample data. BRIEF DESCRIPTION OF THE DRAWINGS
[0009] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying any creative work.
[0010] Figure 1 It is a flowchart of a network topology generation method provided by an embodiment of the present invention; Figure 2 is a schematic block diagram of a network topology generating device provided by an embodiment of the present invention; Figure 3 It is a schematic block diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0011] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0012] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, operations, elements, components and / or groups thereof.
[0013] It should also be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms of "a", "an" and "the" are intended to include plural forms. It should also be further understood that the term "and / or" used in the specification of the present invention and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes these combinations.
[0014] See also Figure 1 , Figure 1 1 is a flow chart of a network topology generation method provided by an embodiment of the present invention. The network topology generation method of the embodiment of the present invention can be applied to a computer device to obtain training data based on historical data and generate a network topology based on the training data. Figure 1 As shown, the method includes steps S100-S110.
[0015] S100, acquiring historical data of a device to be trained, and performing data enhancement processing on the historical data to obtain data to be trained.
[0016] In the embodiment of the present invention, the equipment to be trained refers to the equipment that needs to apply the network topology, for example, the unmanned transport equipment used in the logistics system. When the unmanned transport equipment is performing logistics distribution, most of them are in urban scenes, and the urban scenes are relatively complex. The communication status, network status and data transmission status of the unmanned transport equipment will be greatly affected. The network topology is used to improve the communication efficiency of the unmanned transport equipment during the distribution process, enhance network coverage and improve network stability.
[0017] Take drones as an example. During the logistics distribution process, drones need to communicate with base stations or other drones in real time. Reasonable network topology can improve communication efficiency and coverage, ensuring that drones can communicate normally during the distribution process. At the same time, drones can also connect to terminals through wireless networks, which can improve the network coverage and stability of drones during the distribution process through network topology, and prevent drones from entering areas without network coverage.
[0018] When it is necessary to generate a network topology, the historical data of the device to be trained can be obtained. For example, the historical flight data of the drone can be obtained. The historical flight data includes flight trajectory and communication data. The flight trajectory refers to the path taken by the drone during a delivery process. The communication data refers to the communication-related data generated by the drone during this delivery process, including but not limited to location information, communication delay, signal strength, connection status, routing information, and network performance information. Location information refers to the location point of the drone during the delivery process, that is, the location coordinates. Communication delay refers to the time delay of signal transmission when the drone communicates with the ground control station or other drones. Signal strength refers to the signal strength between the drone and the communication base station or relay device. Connection status refers to the connection status information between the drone and the communication network, including whether the connection is successful, the quality of the connection, and whether there is a risk of disconnection. Routing information is used to determine the best transmission path for data packets from the source drone to the target drone. Network performance information includes indicators such as packet loss rate, network delay, and bandwidth.
[0019] After obtaining the historical data, it is necessary to perform data enhancement processing on the historical data. Taking the drone as an example, the historical flight data of the drone includes multiple different types of data. When performing data enhancement, it is necessary to first extract the data that needs to be enhanced. For example, the flight trajectory data generated by a drone in the historical flight data can be extracted, and then the flight trajectory data can be enhanced to obtain multiple flight trajectory data. Specifically, the extracted flight trajectory data can be subjected to trajectory perturbation processing or trajectory combination processing. The trajectory perturbation processing refers to confirming each position point of the flight trajectory data and adding a random displacement to each position point so that the aircraft trajectory changes to generate a new flight trajectory. Trajectory combination processing refers to extracting at least one flight trajectory data from the historical flight data in addition to the flight trajectory data that has been extracted, and combining the extracted flight trajectory data together to form a new flight trajectory. In addition, when performing data enhancement, the same data enhancement method can be used for each flight trajectory data, or different data enhancement methods can be used. For example, for each flight trajectory data, trajectory perturbation processing can be used for data enhancement, or, for each flight trajectory data, one of trajectory perturbation processing and trajectory combination processing is randomly used for data enhancement. Of course, another data enhancement method can also be used to perform secondary data enhancement processing on the newly generated data.
[0020] Through the above data enhancement processing, more sample data can be expanded on the basis of the existing historical data, and all the expanded sample data can be used as training data, thereby increasing the amount of data contained in the training data, and then ensuring that there is enough data for training to generate the network topology.
[0021] In some embodiments, such as this embodiment, step S100 may include the following steps: Acquire a historical movement trajectory of the device to be trained from the historical data, wherein the historical data includes a plurality of the historical movement trajectories, and the historical movement trajectory includes a plurality of position points; Adding random displacements to the plurality of position points of the historical movement trajectory respectively and adjusting trajectory parameters according to the added random displacements to obtain a disturbed trajectory; The disturbance trajectory is used as the data to be trained.
[0022] In an embodiment of the present invention, taking a drone as an example, the historical movement trajectory of the device to be trained, that is, the historical flight trajectory of the drone, is obtained from historical data. The historical data may include multiple historical movement trajectories, and each historical movement trajectory may be generated by the same drone or different drones.
[0023] Each historical moving trajectory includes multiple position points. The position points can be determined by the position coordinates of the drone during flight. A random displacement is added to each position point, so that the historical flight trajectory of the drone is offset from the original trajectory to form a new flight trajectory, that is, a perturbed trajectory. For example, the flight trajectory of a drone is ABCDE, and a random displacement is added to the B position point to make it offset. The offset position point is F, and the new aircraft trajectory is AFCDE. It should be noted that when adding random displacement to the position point, a random displacement can be added to each position point, or a random displacement can be added to several position points, or a random displacement can be added to one or more position points at random. When obtaining a new flight trajectory, new trajectory parameters can also be configured for the new flight trajectory, such as the speed, acceleration and other parameters of the drone, so as to ensure that the new flight trajectory conforms to the actual scenario.
[0024] In addition to adding random displacement to the position points, you can also randomly insert a new position point between two position points. For example, if the flight trajectory of a drone is ABCDE, you can insert a new position point G between B and C, and the new flight trajectory will be ABGCDE. You can insert a new position point between any two position points to increase the smoothness of the flight trajectory.
[0025] In some embodiments, such as this embodiment, step S100 may include the following steps: Acquire a plurality of historical movement trajectories of the to-be-trained devices from the historical data, wherein the historical data includes a plurality of the historical movement trajectories, and the historical movement trajectories include a plurality of position points; The acquired historical movement trajectories are randomly combined to obtain a combined trajectory, and the combined trajectory is used as the data to be trained.
[0026] In an embodiment of the present invention, each drone can generate at least one historical movement trajectory, that is, the historical movement trajectory in the historical data comes from at least one drone. Preferably, historical data of multiple different drones can be obtained to facilitate improving data diversity.
[0027] When performing data enhancement, multiple historical movement trajectories can be obtained from historical data. The multiple historical movement trajectories come from multiple different drones, and each historical movement trajectory includes multiple location points. The obtained historical movement trajectories are combined to obtain a new flight trajectory, that is, a combined trajectory. For example, two historical movement trajectories can be obtained from historical data. The two historical movement trajectories come from different drones. The two historical movement trajectories can be combined into a new flight trajectory.
[0028] The following is a comprehensive description of data enhancement. Historical data may include multiple historical movement trajectories, and multiple historical movement trajectories come from different drones. When performing data enhancement, one or more historical movement trajectories may be randomly selected from the historical data, and a random method may be used to perform data enhancement. For example, a historical movement trajectory may be obtained from the historical data, and random displacements may be added to the position points of the historical movement trajectory to obtain a perturbed trajectory, or two historical movement trajectories may be randomly obtained from the historical data, and the two historical movement trajectories may be merged to obtain a combined trajectory.
[0029] S110, obtaining a preset learning model, and using the data to be trained as input data of the preset learning model so that the preset learning model trains the data to be trained to output a network topology.
[0030] In an embodiment of the present invention, the preset learning model is a pre-configured model used to train the training data to obtain the network topology.
[0031] In some embodiments, such as this embodiment, step S110 may include the following steps: Setting a state space and an action space according to the data to be trained, wherein the state space includes state information of the device to be trained, and the action space includes action information of the device to be trained; Obtaining a preset reward function and a preset neural network model, and setting a reward strategy of the preset neural network model according to the preset reward function; The state space and the action space are used as input data of the preset neural network model so that the preset neural network model outputs the network topology.
[0032] In an embodiment of the present invention, the state space may include the current location information of the device to be trained, the communication connection status, the current network topology, etc. For example, the state space may include the location, speed, direction, connection status with the base station, and communication links with other drones of the drone.
[0033] The action space can include topology connection operations, such as adding new device connections, deleting existing device connections, adjusting connection weights, etc. These actions are designed to optimize the network topology and improve network performance.
[0034] The preset reward function is used to evaluate whether the actions taken in different states are effective and give corresponding rewards or penalties. The preset reward function can be set according to the performance to be achieved by the network topology, such as communication coverage, network stability, data transmission efficiency, etc. For example, if the network topology needs to have better communication coverage, the preset reward function can be defined as the weighted sum of the increase in communication coverage, the reduction in latency, and the improvement in network stability. It should be noted that the performance indicators of the network topology can be multiple, for example, communication coverage and data transmission efficiency can be included at the same time.
[0035] During the initial training, the state space and action space can be set according to the data to be trained. The state space includes the above-mentioned state information, and the action space includes the above-mentioned action information. At the same time, a preset reward function and a preset neural network model can be obtained. The preset neural network model can be a DQN (Deep Q-Network) deep Q network algorithm and a DDPG (Deep Deterministic Policy Gradient) deep deterministic policy gradient algorithm. After configuring the state space, action space, and preset reward function, the intelligent agent is trained through the preset neural network model and the corresponding results are generated. Then, new data to be trained are obtained, and the state space and action space are adjusted according to the new data to be trained until a network topology that meets the requirements is generated.
[0036] In some embodiments, such as this embodiment, the network topology generation may further include the following steps: Acquire the trained data generated by the preset learning model during the training process; All the acquired trained data are put into the experience pool for storage.
[0037] In an embodiment of the present invention, the experience pool is used to store trained data, which can be abstractly represented as the current state of the agent, the action performed in the current state, the reward obtained after performing the action (determined by a preset reward function), and the new state after performing the action.
[0038] In some embodiments, such as this embodiment, the network topology generation may further include the following steps: Randomly extract at least one of the trained data from the experience pool and use it as repeated training data; The repeated training data is used as input data of the preset learning model so that the preset learning model performs intensive training on the network topology to output the network topology.
[0039] In the embodiment of the present invention, during training, in addition to training with the data to be trained, at least one trained data can be randomly extracted from the experience pool and used as repeated training data, and then the repeated training data is used as input data of the preset learning model for training. When randomly extracting trained data from the reward pool, it can be extracted once after several trainings, which can break the correlation between the data and improve the stability and efficiency of training.
[0040] In some embodiments, such as this embodiment, the network topology generation may further include the following steps: Get the preset diffusion model; The network topology is used as input data of the preset diffusion model so that the preset diffusion model trains the network topology to optimize the network topology.
[0041] In the embodiment of the present invention, the diffusion model is a probabilistic model, which can explore and optimize the network topology at different time steps and states by simulating the changes of random variables (such as the addition of Gaussian white noise). It ensures that the network topology gradually tends to the optimal under the condition of small sample size data. In addition, the communication connection between drones can be optimized by simulating the random changes of drones in different positions and states, thereby improving the coverage and stability of the network.
[0042] Figure 2 FIG. 2 is a schematic block diagram of a network topology generating device 200 provided in an embodiment of the present invention. Figure 2 As shown, corresponding to the above network topology generation method, the present invention also provides a network topology generation device 200. The network topology generation device 200 includes a unit for executing the above network topology generation method. Figure 2 The network topology generating device 200 includes a first acquiring unit 201 and a first training unit 202 .
[0043] The first acquisition unit 201 is used to acquire historical data of the device to be trained, and perform data enhancement processing on the historical data to obtain data to be trained; The first training unit 202 is used to obtain a preset learning model, and use the data to be trained as input data of the preset learning model so that the preset learning model trains the data to be trained to output a network topology.
[0044] In some embodiments, such as this embodiment, the first acquiring unit 201 further includes a second acquiring unit, a first adding unit, and a first setting unit.
[0045] The second acquisition unit is used to acquire the historical movement trajectory of the device to be trained from the historical data, wherein the historical data includes a plurality of the historical movement trajectories, and the historical movement trajectory includes a plurality of position points; The first adding unit is used to add random displacements to the plurality of position points of the historical movement trajectory respectively and adjust trajectory parameters according to the added random displacements to obtain a disturbed trajectory; The first setting unit is used to use the disturbance trajectory as the data to be trained.
[0046] In some embodiments, such as this embodiment, the first acquiring unit 201 further includes a third acquiring unit and a first combining unit.
[0047] The third acquisition unit is used to acquire a plurality of historical movement trajectories of the to-be-trained devices from the historical data, wherein the historical data includes a plurality of the historical movement trajectories, and the historical movement trajectories include a plurality of position points; The first combining unit is used for randomly combining the acquired historical movement trajectories to obtain a combined trajectory, and using the combined trajectory as the to-be-trained data.
[0048] In some embodiments, such as this embodiment, the first training unit 202 further includes a second setting unit, a fourth acquiring unit and a second training unit.
[0049] The second setting unit is used to set a state space and an action space according to the data to be trained, wherein the state space includes the state information of the device to be trained, and the action space includes the action information of the device to be trained; The fourth acquisition unit is used to acquire a preset reward function and a preset neural network model, and set a reward strategy of the preset neural network model according to the preset reward function; The second training unit is used to use the state space and the action space as input data of the preset neural network model so that the preset neural network model outputs the network topology.
[0050] In some embodiments, such as this embodiment, the network topology generating device 200 further includes a fifth acquiring unit and a first storage unit.
[0051] The fifth acquisition unit is used to acquire the trained data generated by the preset learning model during the training process; The first storage unit is used to store all the acquired trained data into an experience pool.
[0052] In some embodiments, such as this embodiment, the network topology generating device 200 further includes a first extraction unit and a third training unit.
[0053] The first extraction unit is used to randomly extract at least one of the trained data from the experience pool and use it as repeated training data; The third training unit is used to use the repeated training data as input data of the preset learning model so that the preset learning model performs intensive training on the network topology to output the network topology.
[0054] In some embodiments, such as this embodiment, the network topology generating device 200 further includes a sixth acquisition unit and a fourth training unit.
[0055] Wherein, the sixth acquisition unit is used to acquire a preset diffusion model; The fourth training unit is used to use the network topology as input data of the preset diffusion model so that the preset diffusion model trains the network topology to optimize the network topology.
[0056] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the above-mentioned network topology generation device and each unit can refer to the corresponding description in the aforementioned method embodiment, and for the convenience and brevity of description, it will not be repeated here.
[0057] The network topology generating device can be implemented in the form of a computer program. Figure 3 Runs on the computer device shown.
[0058] See also Figure 3 The computer device 300 includes a processor 302 , a memory and an interface 305 connected via a system bus 301 , wherein the memory may include a non-volatile storage medium 303 and an internal memory 304 .
[0059] The non-volatile storage medium 303 may store an operating system 3031 and a computer program 3032. When the computer program 3032 is executed, the processor 302 may execute a network topology generation method.
[0060] The processor 302 is used to provide computing and control capabilities to support the operation of the entire computer device 300 .
[0061] The internal memory 304 provides an environment for the operation of the computer program 3032 in the non-volatile storage medium 303. When the computer program 3032 is executed by the processor 302, the processor 302 can execute a network topology generation method.
[0062] The interface 305 is used to communicate with other devices. Those skilled in the art will appreciate that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device 300 to which the solution of the present application is applied. The specific computer device 300 may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0063] It should be understood that in the embodiment of the present application, the processor 302 may be a central processing unit (CPU), and the processor 302 may also be other general-purpose processors, digital signal processors (FSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.
[0064] It is understood by those skilled in the art that all or part of the processes in the method for implementing the above embodiment can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a storage medium, which is a computer-readable storage medium. The computer program is executed by at least one processor in the computer system to implement the process steps of the embodiment of the above method.
[0065] Therefore, the present invention also provides a storage medium. The storage medium may be a computer-readable storage medium. The storage medium stores a computer program. When the computer program is executed by a processor, any embodiment of the above-mentioned network topology generation method is implemented.
[0066] The storage medium may be a USB flash drive, a mobile hard disk, a read-only memory (ROM), a magnetic disk, or an optical disk, or any other computer-readable storage medium that can store program codes.
[0067] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.
[0068] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of each unit is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.
[0069] The steps in the method of the embodiment of the present invention can be adjusted in order, combined and deleted according to actual needs. The units in the device of the embodiment of the present invention can be combined, divided and deleted according to actual needs. In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0070] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for a computer device to execute all or part of the steps of the method described in each embodiment of the present invention.
[0071] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0072] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
[0073] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.
Claims
1. A network topology generation method, characterized in that: The method comprises: Acquire historical data of the device to be trained, and perform data enhancement processing on the historical data to obtain data to be trained; A preset learning model is obtained, and the data to be trained is used as input data of the preset learning model so that the preset learning model trains the data to be trained to output a network topology.
2. The method according to claim 1, characterized in that The method further comprises: Acquire the trained data generated by the preset learning model during the training process; All the acquired trained data are put into the experience pool for storage.
3. The method according to claim 2, characterized in that The method further comprises: Randomly extract at least one of the trained data from the experience pool and use it as repeated training data; The repeated training data is used as input data of the preset learning model so that the preset learning model performs intensive training on the network topology to output the network topology.
4. The method according to claim 1, characterized in that The method further comprises: Get the preset diffusion model; The network topology is used as input data of the preset diffusion model so that the preset diffusion model trains the network topology to optimize the network topology.
5. The method according to claim 1, characterized in that The step of performing data enhancement processing on the historical data to obtain data to be trained includes: Acquire a historical movement trajectory of the device to be trained from the historical data, wherein the historical data includes a plurality of the historical movement trajectories, and the historical movement trajectory includes a plurality of position points; Adding random displacements to the plurality of position points of the historical movement trajectory respectively and adjusting trajectory parameters according to the added random displacements to obtain a disturbed trajectory; The disturbance trajectory is used as the data to be trained.
6. The method according to claim 1, characterized in that The step of performing data enhancement processing on the historical data to obtain data to be trained includes: Acquire a plurality of historical movement trajectories of the to-be-trained devices from the historical data, wherein the historical data includes a plurality of the historical movement trajectories, and the historical movement trajectories include a plurality of position points; The acquired historical movement trajectories are randomly combined to obtain a combined trajectory, and the combined trajectory is used as the data to be trained.
7. The method according to claim 1, characterized in that The step of using the data to be trained as input data of the preset learning model so that the preset learning model trains the data to be trained to output a network topology includes: Setting a state space and an action space according to the data to be trained, wherein the state space includes state information of the device to be trained, and the action space includes action information of the device to be trained; Obtaining a preset reward function and a preset neural network model, and setting a reward strategy of the preset neural network model according to the preset reward function; The state space and the action space are used as input data of the preset neural network model so that the preset neural network model outputs the network topology.
8. A network topology generation device, characterized in that: The device comprises: A first acquisition unit is used to acquire historical data of the device to be trained, and perform data enhancement processing on the historical data to obtain data to be trained; The first training unit is used to obtain a preset learning model, and use the data to be trained as input data of the preset learning model so that the preset learning model trains the data to be trained to output a network topology.
9. A computer device, characterized in that: The computer device comprises a memory and a processor connected to the memory; the memory is used to store a computer program; the processor is used to run the computer program stored in the memory to execute the steps of the method according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 can be implemented.
