Task unloading method and device, equipment and storage medium

Abstract

The embodiment of the invention discloses a task unloading method and device, equipment and a storage medium, and the method comprises the steps: obtaining the historical position information of a moving object in a target region, and a global scene heat map corresponding to the target region; inputting the historical position information and the global scene heat map into a pre-trained target task unloading model to obtain a target task unloading result of the moving object; wherein the target task unloading result comprises a target edge server used for executing the data processing task of the moving object. According to the technical scheme, the problem that in the prior art, scene information of a target area and position information of a moving object in the area can be analyzed based on a target task unloading model obtained through training, and then a corresponding edge server is matched for the moving object and used for executing a data processing task of the moving object, so that the data processing efficiency is improved is solved. Task processing delay can be reduced, and the resource utilization rate is increased.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of data processing technology, and in particular to a task offloading method, apparatus, device, and storage medium. Background Art

[0002] Mobile edge computing allows edge servers to be deployed at base stations closer to mobile devices, addressing issues such as insufficient computing resources on terminal devices and extended task processing times. Task offloading involves assigning computationally intensive tasks to edge servers with sufficient computing resources, and then retrieving the completed computational results from the edge servers.

[0003] This approach significantly reduces task processing latency and mobile terminal energy consumption, thereby improving service quality and user satisfaction. More importantly, by transferring data to adjacent edge servers rather than remote cloud servers, the risk of information leakage and data tampering is effectively reduced. However, task offloading strategies significantly impact both task latency reduction and resource efficiency, making the design of effective and intelligent offloading strategies essential. Summary of the Invention

[0004] The embodiments of the present invention provide a task offloading method, apparatus, device and storage medium, which can reduce task processing delay and improve resource utilization.

[0005] In a first aspect, an embodiment of the present invention provides a task offloading method, the method comprising:

[0006] Obtain historical location information of a mobile object within a target area, and a global scene heat map corresponding to the target area; input the historical location information and the global scene heat map into a pre-trained target task offloading model to obtain a target task offloading result for the mobile object; wherein the target task offloading result includes a target edge server for executing a data processing task for the mobile object.

[0007] In a second aspect, an embodiment of the present invention provides a task offloading device, the device comprising:

[0008] A data acquisition module is used to obtain the historical location information of mobile objects in the target area and the global scene heat map corresponding to the target area; a task offloading module is used to input the historical location information and the global scene heat map into a pre-trained target task offloading model to obtain the target task offloading result of the mobile object; wherein, the target task offloading result includes a target edge server for executing the data processing task of the mobile object.

[0009] In a third aspect, an embodiment of the present invention provides a computer device, the computer device comprising:

[0010] one or more processors;

[0011] a memory for storing one or more programs;

[0012] When the one or more programs are executed by the one or more processors, the one or more processors implement the task offloading method described in any embodiment.

[0013] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the task offloading method described in any embodiment.

[0014] The technical solution provided by the embodiment of the present invention obtains the historical location information of the mobile object in the target area and the global scene heat map corresponding to the target area; the historical location information and the global scene heat map are input into a pre-trained target task offloading model to obtain the target task offloading result of the mobile object; wherein the target task offloading result includes a target edge server for executing the data processing task of the mobile object. The technical solution of the embodiment of the present invention solves the problem in the prior art that the scene information of the target area and the location information of the mobile object in the area can be analyzed based on the trained target task offloading model, and then the corresponding edge server is matched to the mobile object for executing the data processing task of the mobile object, which can reduce the task processing delay and improve resource utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of a task offloading method provided by an embodiment of the present invention;

[0016] Figure 2 This is a flow chart of another task offloading method provided by an embodiment of the present invention;

[0017] Figure 3 This is a workflow diagram for trajectory prediction provided by an embodiment of the present invention;

[0018] Figure 4 This is a workflow diagram for performing optimization problem analysis provided by an embodiment of the present invention;

[0019] Figure 5 This is a structural diagram of a task offloading device provided by an embodiment of the present invention;

[0020] Figure 6 It is a structural diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0022] Figure 1 This is a flowchart of a task offloading method provided by an embodiment of the present invention. The embodiment of the present invention can be applied to the scenario of matching the corresponding edge server for the mobile objects in the target area. The method can be executed by a task offloading device, which can be implemented by software and / or hardware.

[0023] like Figure 1 As shown, the task offloading method includes the following steps:

[0024] S110: Acquire historical location information of mobile objects within a target area, and a global scene heat map corresponding to the target area.

[0025] The target area may be an area where task offloading is required. The mobile object may be an object moving within the target area. Specifically, a customer within the target area may be considered a mobile object. The historical location information may be location movement information of the mobile object within the target area during a historical period. The specific duration of the historical period may be manually set. Specifically, the coordinate information of the mobile object within the historical period may be collected at preset intervals based on a preset device, and the collected coordinate information may be used as the historical location information. Furthermore, the global scene heat map may be a heat map corresponding to the target area.

[0026] S120: Input the historical location information and the global scene heat map into a pre-trained target task offloading model to obtain a target task offloading result of the mobile object.

[0027] The target task offloading model can be a model for performing task offloading. Specifically, the target task offloading model can be a pre-trained neural network model. The target task offloading result includes a target edge server for performing the data processing task for the mobile object. By inputting historical location information and a global scene heat map into the target task offloading model, the target task offloading model can analyze the historical location information and the global scene heat map to determine an edge server that can be used to process the mobile object, thereby obtaining the target task offloading result for the mobile object.

[0028] Optionally, the target task offloading model includes: a trajectory prediction submodel and a task offloading submodel, and the historical location information and the global scene heat map are input into the pre-trained target task offloading model to obtain the target task offloading result of the mobile object, including: inputting the historical location information and the global scene heat map into the trajectory prediction submodel to obtain the mobile trajectory prediction information of the mobile object; obtaining the resource parameters of multiple edge servers in the target area, and inputting the resource parameters and the mobile trajectory prediction information into the task offloading submodel to obtain the target task offloading result of the mobile object.

[0029] The trajectory prediction sub-model can be used to predict the movement trajectory of a mobile object. The task offloading sub-model can be used to analyze input data and generate a corresponding task offloading result for the mobile object. The target task carrying result can be the final determined task offloading result. Specifically, the target task carrying result can include assigning the data processing task of the mobile object to one or more edge servers for execution.

[0030] Optionally, the training process of the target task offloading model includes: obtaining a preliminary task offloading model and preset training samples; training the preliminary task offloading model based on the preset training samples and a deep reinforcement learning method to obtain a target task offloading model; wherein the reward function of the deep reinforcement learning method is to minimize the time for executing data processing tasks.

[0031] The preliminary task offloading model may be an original, untrained task offloading model. Furthermore, the preset training samples may be samples preset for training the preliminary task offloading model. The preliminary task offloading model may include labels for corresponding task offloading results, which are used to evaluate the output of the preliminary task offloading model and adjust the model based on the evaluation results.

[0032] The technical solution provided by the embodiment of the present invention obtains the historical location information of the mobile object in the target area and the global scene heat map corresponding to the target area; the historical location information and the global scene heat map are input into a pre-trained target task offloading model to obtain the target task offloading result of the mobile object; wherein the target task offloading result includes the target edge server for executing the data processing task of the mobile object. The technical solution of the embodiment of the present invention solves the problem in the prior art that the scene information of the target area and the location information of the mobile object in the area can be analyzed based on the trained target task offloading model, and then the corresponding edge server is matched to the mobile object for executing the data processing task of the mobile object, which can reduce task processing delay and improve resource utilization.

[0033] Figure 2This is another task offloading method flow chart provided by an embodiment of the present invention. The embodiment of the present invention can be applied to the scenario of matching the corresponding edge server for the mobile objects in the target area. Based on the above embodiment, this embodiment further explains how to input the historical location information and the global scene heat map into the pre-trained target task offloading model to obtain the target task offloading result of the mobile object. The device can be implemented by software and / or hardware and integrated into a computer device with application development function.

[0034] like Figure 2 As shown, the task offloading method includes the following steps:

[0035] S210: Input the historical location information into a first feature determination network to obtain historical location features.

[0036] The historical location information may be information about the movement of a mobile object within a target area during a historical period. The specific duration of the historical period may be manually set. Specifically, the coordinate information of the mobile object during the historical period may be collected at predetermined intervals based on a predetermined device, and the collected coordinate information may be used as the historical location information.

[0037] The first feature determination network may be a network for extracting displacement features of a mobile object. The historical position features may be features representing the positional movement of the mobile object over a historical period. Specifically, historical position information may be input into the first feature determination network so that the network extracts features corresponding to the historical position information, thereby obtaining the historical position features.

[0038] Optionally, the first feature determination network includes a graph convolutional network and a gated linear unit. Inputting the historical location information into the first feature determination network to obtain the historical location feature includes: inputting the historical location information into the graph convolutional network to obtain the first feature; and inputting the first feature into the gated linear unit to obtain the historical location feature. The first intermediate feature may be an intermediate parameter in the process of extracting the historical location feature from the historical location information.

[0039] S220: Input the global scene heat map into the second feature determination network to obtain global scene features.

[0040] The global scene heat map may be a heat map corresponding to the target area. The second feature determination network may be a network for extracting scene features within the target area. The global scene features may be scene features related to the target area. Specifically, the global scene heat map of the target area may be input into the second feature determination network so that the network extracts scene features corresponding to the global scene heat map, thereby obtaining the global scene features.

[0041] Optionally, the second feature determination network includes a convolutional neural network and a gated linear unit. Inputting the global scene heat map into the second feature determination network to obtain the global scene feature includes: inputting historical location information into the convolutional neural network to obtain a second feature; and inputting the second feature into the gated linear unit to obtain a historical location feature. The second intermediate feature may be an intermediate feature in the process of extracting the global scene feature from the global scene heat map.

[0042] S230: Input historical location features and global scene features into a feature analysis network to obtain movement trajectory prediction information.

[0043] The predicted trajectory information can be the predicted trajectory of the mobile object within the target area. By determining the predicted trajectory information of the mobile object, it is convenient to match the data processing tasks of the mobile object to the appropriate edge server based on the predicted trajectory information, thereby reducing data processing delay time.

[0044] S240: Input resource parameters and movement trajectory prediction information into the optimization problem construction network to obtain the target optimization problem.

[0045] The resource parameters may be used to represent parameters of available computing resources of the edge server. For example, the resource parameters may include base station communication bandwidth, maximum computing resources of the edge server, and other parameters.

[0046] The target optimization problem can be the optimization problem that ultimately needs to be solved. Specifically, the optimization objective of the target optimization problem can be minimizing data processing delay time. The variables that can be optimized include the offloading method of data processing tasks and the resource allocation method corresponding to the data processing tasks. The offloading method of data processing tasks includes assigning data processing tasks to one or more edge servers for processing. The resource allocation method includes how much computing resources the edge server allocates to execute the data processing tasks for sending mobile objects.

[0047] S250: Input the target optimization problem into the optimization problem solving network to obtain the target task offloading result of the mobile object.

[0048] The target task carrying result may be the final determined task offloading result. Specifically, the target optimization problem may be input into an optimization problem solving network, which analyzes the target optimization problem and uses the answer to the target optimization problem as the target task offloading result.

[0049] For example, the specific process of deep reinforcement learning to solve optimization problems can be described as follows:

[0050] 1. Initialize the neural network parameters;

[0051] 2. Create multiple parallel training threads, each of which independently runs an agent to interact with the environment and uses the actor and critic network to approximate the strategy and value;

[0052] 3. Each thread selects an action based on the current policy network, executes the action, observes the new state and reward, and constructs a reward function based on the objective function of the optimization problem. Finally, the action, state, and reward information are stored in the experience replay buffer.

[0053] 4. When a thread reaches the set number of iterations, the thread samples the data in the experience replay buffer and performs gradient updates by calculating the advantage function;

[0054] 5. After each thread performs a certain number of gradient updates, it passes the updated parameters to the main thread for overall parameter update.

[0055] 6. Repeat the above steps until the predetermined number of training rounds is reached or the termination condition is met.

[0056] Furthermore, in order to better understand the technical solution provided by the present invention, a specific embodiment is introduced below:

[0057] Figure 3 This is a workflow diagram for trajectory prediction provided by an embodiment of the present invention. For example, Figure 3 As shown in Figure 2, the workflow for trajectory prediction includes the following steps:

[0058] On the one hand, a high-dimensional vector of the mobile user's relevant location is embedded, which is used as the input for spatiotemporal correlation modeling and then processed based on a graph convolutional network and a gated linear unit to obtain historical location features. On the other hand, the global scene heat map of the target area can be input into the convolutional network and the gated linear unit to obtain the global scene features. Finally, the historical location features and the global scene features are convolved to obtain the predicted trajectory of the mobile user.

[0059] Figure 4 This is a workflow diagram for analyzing optimization problems provided by an embodiment of the present invention. Figure 4 As shown in Figure 2, the workflow for analyzing an optimization problem includes the following steps:

[0060] The system model of the edge server is input (including parameters such as base station communication bandwidth, mobile user location and transmission power, task volume, and maximum computing resources of the edge server). A mixed integer optimization problem is constructed with offloading decision and resource allocation as optimization variables and latency minimization as the goal. The optimization problem is then solved using the deep reinforcement learning A3C algorithm, which is used to obtain the optimal task offloading decision variables. Finally, the user can decide whether to process the task locally or offload it to the edge server based on the offloading strategy.

[0061] The technical solution provided by the embodiment of the present invention obtains historical location features by inputting historical location information into a first feature determination network; inputting a global scene heat map into a second feature determination network to obtain global scene features; inputting historical location features and global scene features into a feature analysis network to obtain movement trajectory prediction information; inputting resource parameters and movement trajectory prediction information into an optimization problem construction network to obtain a target optimization problem; and inputting the target optimization problem into an optimization problem solving network to obtain a target task unloading result for a mobile object. The technical solution of the embodiment of the present invention solves the problem in the prior art that the scene information of the target area and the location information of the mobile objects in the area can be analyzed based on the target task unloading model obtained through training, and then the corresponding edge server is matched to the mobile object to perform the data processing task of the mobile object, which can reduce task processing delay and improve resource utilization.

[0062] Figure 5 This is a structural diagram of a task offloading device provided by an embodiment of the present invention. The embodiment of the present invention can be applied to the scenario of matching mobile objects in the target area with corresponding edge servers. The device can be implemented by software and / or hardware and integrated into a computer device with application development capabilities.

[0063] like Figure 5 As shown, the task offloading device includes: a data acquisition module 310 and a task offloading module 320.

[0064] Among them, the data acquisition module 310 is used to obtain the historical location information of the mobile object in the target area, and the global scene heat map corresponding to the target area; the task offloading module 320 is used to input the historical location information and the global scene heat map into a pre-trained target task offloading model to obtain the target task offloading result of the mobile object; wherein, the target task offloading result includes a target edge server for executing the data processing task of the mobile object.

[0065] The technical solution provided by the embodiment of the present invention obtains the historical location information of the mobile object in the target area and the global scene heat map corresponding to the target area; the historical location information and the global scene heat map are input into a pre-trained target task offloading model to obtain the target task offloading result of the mobile object; wherein the target task offloading result includes a target edge server for executing the data processing task of the mobile object. The technical solution of the embodiment of the present invention solves the problem in the prior art that the scene information of the target area and the location information of the mobile object in the area can be analyzed based on the trained target task offloading model, and then the corresponding edge server is matched to the mobile object for executing the data processing task of the mobile object, which can reduce the task processing delay and improve resource utilization.

[0066] In an optional embodiment, the target task offloading model includes: a trajectory prediction submodel and a task offloading submodel, and the task offloading module 320 is specifically used to: input the historical location information and the global scene heat map into the trajectory prediction submodel to obtain the movement trajectory prediction information of the mobile object; obtain the resource parameters of multiple edge servers in the target area, input the resource parameters and the movement trajectory prediction information into the task offloading submodel, and obtain the target task offloading result of the mobile object.

[0067] In an optional embodiment, the trajectory prediction sub-model includes: a first feature determination network, a second feature determination network and a feature analysis network, and the task offloading module 320 includes: a trajectory prediction unit, which is used to: input the historical position information into the first feature determination network to obtain historical position features; input the global scene heat map into the second feature determination network to obtain global scene features; input the historical position features and the global scene features into the feature analysis network to obtain mobile trajectory prediction information.

[0068] In an optional embodiment, the first feature determination network includes: a graph convolutional network and a gated linear unit, and the trajectory prediction unit includes: a position feature extraction subunit, which is used to: input the historical position information into the graph convolutional network to obtain a first feature; input the first feature into the gated linear unit to obtain the historical position feature.

[0069] In an optional embodiment, the second feature determination network includes: a convolutional neural network and a gated linear unit, and the trajectory prediction unit also includes: a scene feature extraction subunit, which is used to: input the historical position information into the convolutional neural network to obtain the second feature; input the second feature into the gated linear unit to obtain the historical position feature.

[0070] In an optional embodiment, the task offloading sub-model includes: an optimization problem construction network and an optimization problem answering network, and the task offloading module 320 also includes: an optimization problem analysis unit, which is used to: input the resource parameters and the movement trajectory prediction information into the optimization problem construction network to obtain the target optimization problem; input the target optimization problem into the optimization problem answering network to obtain the target task offloading result of the mobile object.

[0071] In an optional embodiment, the task offloading device also includes: a model training module, used to: obtain a preliminary task offloading model and preset training samples; train the preliminary task offloading model based on the preset training samples and a deep reinforcement learning method to obtain the target task offloading model, wherein the reward function of the deep reinforcement learning method is to minimize the time to execute the data processing task.

[0072] The task offloading device provided by the embodiment of the present invention can execute the task offloading method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0073] Figure 6 A schematic structural diagram of a computer device provided in an embodiment of the present invention. Figure 6 A block diagram of an exemplary computer device 12 suitable for use in implementing embodiments of the present invention is shown. Figure 6 The computer device 12 shown is only an example and should not limit the functionality and scope of use of the embodiments of the present invention. The computer device 12 can be any terminal device with computing capabilities and can be configured in the task offloading device.

[0074] like Figure 6 As shown, computer device 12 is implemented as a general-purpose computing device. Components of computer device 12 may include, but are not limited to, one or more processors or processing units 16, system memory 28, and a bus 18 that connects various system components (including system memory 28 and processing unit 16).

[0075] The bus 18 may be one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or a local bus using any of a variety of bus architectures. Examples of these architectures include, but are not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MAC) bus, an Enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.

[0076] The computer device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the computer device 12, including volatile and non-volatile media, removable and non-removable media.

[0077] System memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache 32. Computer device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 34 may be configured to read and write non-removable, non-volatile magnetic media ( Figure 6 Not shown, often called a "hard drive"). Although Figure 6 Not shown, a magnetic disk drive for reading and writing to a removable non-volatile magnetic disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 via one or more data media interfaces. System memory 28 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of various embodiments of the present invention.

[0078] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in system memory 28. Such program modules 42 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may include an implementation of a network environment. Program modules 42 generally perform the functions and / or methods of the embodiments described herein.

[0079] The computer device 12 may also communicate with one or more external devices 14 (e.g., a keyboard, a pointing device, a display 24, etc.), one or more devices that enable a user to interact with the computer device 12, and / or any device that enables the computer device 12 to communicate with one or more other computing devices (e.g., a network card, a modem, etc.). Such communication may occur through an input / output (I / O) interface 22. Furthermore, the computer device 12 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 20. Figure 6 As shown, the network adapter 20 communicates with the other modules of the computer device 12 via the bus 18. Figure 6Not shown, other hardware and / or software modules may be used in conjunction with computer device 12, including but not limited to microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0080] The processing unit 16 executes various functional applications and data processing by running programs stored in the system memory 28, for example, implementing the task offloading method provided in an embodiment of the present invention, which includes:

[0081] Obtain historical location information of a mobile object within a target area, and a global scene heat map corresponding to the target area; input the historical location information and the global scene heat map into a pre-trained target task offloading model to obtain a target task offloading result for the mobile object; wherein the target task offloading result includes a target edge server for executing a data processing task for the mobile object.

[0082] This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the method for offloading a task provided in any embodiment of the present invention is implemented, including:

[0083] Obtain historical location information of a mobile object within a target area, and a global scene heat map corresponding to the target area; input the historical location information and the global scene heat map into a pre-trained target task offloading model to obtain a target task offloading result for the mobile object; wherein the target task offloading result includes a target edge server for executing a data processing task for the mobile object.

[0084] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to: an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device.

[0085] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0086] Program code embodied on a computer-readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0087] Computer program code for performing the operations of the present invention can be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as C, Java, Smalltalk, C++, C#, and Python, and also conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0088] Those skilled in the art will appreciate that the modules or steps of the present invention described above can be implemented using a general-purpose computing device. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Alternatively, they can be implemented using program code executable by a computer device, which can then be stored in a storage device and executed by the computing device. Alternatively, they can be fabricated into separate integrated circuit modules, or multiple modules or steps can be fabricated into a single integrated circuit module. Thus, the present invention is not limited to any specific combination of hardware and software.

[0089] Note that the above are only preferred embodiments of the present invention and the technical principles employed. Those skilled in the art will appreciate that the present invention is not limited to the specific embodiments herein, and that various obvious changes, readjustments, and substitutions are possible for those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments and may include many other equivalent embodiments without departing from the scope of the present invention. The scope of the present invention is determined by the scope of the appended claims.

Claims

1. A task offloading method, characterized in that: include: Obtain historical location information of mobile objects within a target area, and a global scene heat map corresponding to the target area; Inputting the historical location information and the global scene heat map into a pre-trained target task offloading model to obtain a target task offloading result of the mobile object; The target task offloading result includes a target edge server for executing the data processing task of the mobile object.

2. The method according to claim 1, characterized in that The target task offloading model includes: a trajectory prediction submodel and a task offloading submodel. The historical location information and the global scene heat map are input into the pre-trained target task offloading model to obtain the target task offloading result of the mobile object, including: Inputting the historical position information and the global scene heat map into the trajectory prediction sub-model to obtain the movement trajectory prediction information of the moving object; Resource parameters of multiple edge servers in a target area are obtained, and the resource parameters and the movement trajectory prediction information are input into a task offloading sub-model to obtain a target task offloading result of the mobile object.

3. The method according to claim 2, characterized in that The trajectory prediction sub-model includes: a first feature determination network, a second feature determination network, and a feature analysis network. Inputting the historical location information and the global scene heat map into the trajectory prediction sub-model to obtain the movement trajectory prediction information of the mobile object includes: Inputting the historical location information into the first feature determination network to obtain historical location features; Inputting the global scene heat map into the second feature determination network to obtain global scene features; The historical location features and the global scene features are input into the feature analysis network to obtain movement trajectory prediction information.

4. The method according to claim 3, characterized in that The first feature determination network includes: a graph convolutional network and a gated linear unit, and the historical location information is input into the first feature determination network to obtain the historical location feature, including: Inputting the historical location information into the graph convolutional network to obtain a first feature; The first feature is input into the gated linear unit to obtain the historical position feature.

5. The method according to claim 3, characterized in that The second feature determination network includes: a convolutional neural network and a gated linear unit, and the global scene heat map is input into the second feature determination network to obtain the global scene feature, including: Inputting the historical location information into the convolutional neural network to obtain a second feature; The second feature is input into the gated linear unit to obtain the historical position feature.

6. The method according to claim 2, characterized in that The task offloading sub-model includes: an optimization problem construction network and an optimization problem answering network. Inputting the resource parameters and the movement trajectory prediction information into the task offloading sub-model to obtain the target task offloading result of the mobile object includes: Inputting the resource parameters and the movement trajectory prediction information into the optimization problem construction network to obtain a target optimization problem; The target optimization problem is input into the optimization problem solving network to obtain the target task offloading result of the mobile object.

7. The method according to claim 1, characterized in that The training process of the target task offloading model includes: Obtain a preliminary task offloading model and preset training samples; Training the preliminary task offloading model based on the preset training samples and deep reinforcement learning to obtain the target task offloading model; The reward function of the deep reinforcement learning method is to minimize the time of executing the data processing task.

8. A task offloading device, characterized in that: The device comprises: A data acquisition module is used to obtain historical location information of mobile objects in a target area and a global scene heat map corresponding to the target area; A task offloading module is used to input the historical location information and the global scene heat map into a pre-trained target task offloading model to obtain a target task offloading result of the mobile object; The target task offloading result includes a target edge server for executing the data processing task of the mobile object.

9. A computer device, characterized in that: The computer device comprises: one or more processors; a memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the task offloading method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the task offloading method according to any one of claims 1 to 7 is implemented.

Images