Edge calculation method and device, equipment and medium

Abstract

The invention provides an edge calculation method and device, equipment and a medium, and the method comprises the steps: determining a corresponding resource weight according to the slice resource density of a candidate base station; determining base station density according to the base station coordinates of the different candidate base stations, and determining control parameters of tent mapping according to the base station density; performing tent mapping based on the control parameters and the random initial values of the candidate base stations to generate a chaotic sequence; calculating and determining a product value of the resource weight and a final value in the chaos sequence, and taking the product value as a score value of the candidate base station; the maximum value of the score values corresponding to the different candidate base stations is determined, the candidate base station corresponding to the maximum value serves as the initial position of the particle to be distributed, and the position of the particle to be distributed is used for indicating the task strategy. Therefore, the non-uniform initial particle swarm can be generated in combination with base station distribution and slice resource density, particle aggregation is avoided through the chaotic characteristic of tent mapping, resource dense area coverage is improved, and therefore the global search efficiency of tasks and the resource utilization rate are improved.

Description

Technical Field

[0001] This disclosure relates to the field of 5G communication technology, specifically to an edge computing method, apparatus, device, and medium. Background Technology

[0002] In 5G vehicle-to-everything (V2X) scenarios, the deployment of ultra-dense base stations and the diverse URLLC (Ultra-Reliable Low-Latency Communication) services (such as autonomous driving and vehicle-to-everything communication) place stringent demands on edge computing tasks.

[0003] Existing technologies generally employ the traditional Particle Swarm Optimization (PSO) algorithm, which iteratively optimizes the task strategy by swarming particles in the search space. However, in highly dynamic environments, traditional PSO algorithms are prone to getting trapped in local optima, leading to resource imbalances, low task completion rates, and large latency fluctuations. This makes them unsuitable for the requirements of 5G network slicing and Ultra Reliable Low Latency Communication (URLLC). Summary of the Invention This disclosure aims to at least partially address one of the technical problems in the related art.

[0004] Therefore, the purpose of this disclosure is to propose an edge computing method, apparatus, electronic device, and storage medium that can generate a non-uniform initial particle swarm by combining base station distribution and slice resource density, avoid particle aggregation by using the chaotic characteristics of tent mapping, improve coverage of resource-dense areas, and thus improve the global search efficiency and resource utilization of the task.

[0005] To achieve the above objectives, the edge computing method proposed in the first aspect of this disclosure includes: The corresponding resource weights are determined based on the slice resource density of the candidate base stations; The base station density is determined based on the base station coordinates of the different candidate base stations, and the control parameters for tent mapping are determined based on the base station density. Tent mapping is performed based on the control parameters and the random initial values ​​of the candidate base stations to generate a chaotic sequence; The product of the determined resource weight and the final value in the chaotic sequence is calculated and used as the score of the candidate base station. The maximum value of the score corresponding to different candidate base stations is determined, and the candidate base station corresponding to the maximum value is used as the initial position of the particle to be assigned, wherein the position of the particle to be assigned is used to indicate the task strategy.

[0006] To achieve the above objectives, an edge computing device according to a second aspect embodiment of this disclosure includes: The first determining module is used to determine the corresponding resource weights based on the slice resource density of the candidate base stations; The second determining module is used to determine the base station density based on the base station coordinates of different candidate base stations, and to determine the control parameters for tent mapping based on the base station density. A generation module is used to perform tent mapping based on the control parameters and the random initial values ​​of the candidate base stations to generate a chaotic sequence; The third determining module is used to calculate the product of the determined resource weight and the final value in the chaotic sequence, as the score of the candidate base station; The fourth determining module is used to determine the maximum value of the score corresponding to different candidate base stations, and to take the candidate base station corresponding to the maximum value as the initial position of the particle to be assigned, wherein the position of the particle to be assigned is used to indicate the task strategy.

[0007] An electronic device according to a third aspect of this disclosure includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the edge computing method as proposed in a first aspect of this disclosure.

[0008] The fourth aspect of this disclosure provides a non-transitory computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the edge computing method as proposed in the first aspect of this disclosure.

[0009] A fifth aspect of this disclosure provides a computer program product that, when instructions in the computer program product are executed by a processor, performs an edge computing method as proposed in a first aspect of this disclosure.

[0010] The edge computing method, apparatus, electronic device, and storage medium disclosed herein determine corresponding resource weights based on the slice resource density of candidate base stations; determine base station density based on the base station coordinates of different candidate base stations, and determine control parameters for tent mapping based on the base station density; perform tent mapping based on the control parameters and random initial values ​​of candidate base stations to generate a chaotic sequence; calculate the product of the determined resource weights and the final values ​​in the chaotic sequence as the score of the candidate base station; determine the maximum value of the corresponding score for different candidate base stations, and use the candidate base station corresponding to the maximum value as the initial position of the particles to be assigned, wherein the position of the particles to be assigned is used to indicate the task strategy. Thus, it is possible to generate a non-uniform initial particle swarm by combining base station distribution and slice resource density, avoid particle aggregation through the chaotic characteristics of tent mapping, improve coverage of resource-dense areas, and thereby improve the global search efficiency and resource utilization of the task.

[0011] Additional aspects and advantages of this disclosure will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this disclosure. Attached Figure Description

[0012] The above and / or additional aspects and advantages of this disclosure will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, in which: Figure 1 This is a schematic flowchart of an edge computing method proposed in an embodiment of this disclosure; Figure 2 This is a schematic diagram of the 5G chaos initialization process based on tent mapping proposed in this disclosure; Figure 3 This is a flowchart illustrating an edge computing method proposed in another embodiment of this disclosure; Figure 4 This is a schematic diagram of the dynamic network slice weight adjustment process proposed in this disclosure; Figure 5 This is a flowchart illustrating an edge computing method proposed in another embodiment of this disclosure; Figure 6 This is a schematic diagram of the collaborative optimization and resource allocation process proposed in this disclosure; Figure 7 This is a schematic diagram of the edge computing system architecture proposed in this disclosure; Figure 8 This is a schematic diagram of the structure of an edge computing device according to an embodiment of the present disclosure; Figure 9 This is a block diagram of an electronic device according to an embodiment of the present application. Detailed Implementation

[0013] Embodiments of this disclosure are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are used only to explain this disclosure, and should not be construed as limiting this disclosure. Rather, embodiments of this disclosure include all variations, modifications, and equivalents falling within the spirit and scope of the appended claims.

[0014] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, data stored, data displayed, etc.) and signals involved in this disclosure are all authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0015] Figure 1 This is a flowchart illustrating an edge computing method proposed in one embodiment of this disclosure.

[0016] It should be noted that the execution subject of the edge computing method in this embodiment is an edge computing device, which can be implemented by software and / or hardware. The device can be configured in an electronic device, which may include, but is not limited to, a terminal, a server, etc., such as a mobile phone, a PDA, etc.

[0017] like Figure 1 As shown, the edge computing method includes: S101: Determine the corresponding resource weights based on the slice resource density of the candidate base stations.

[0018] In this context, a candidate base station can refer to a base station that can be assigned a task in this embodiment of the disclosure, such as a service area base station or a neighboring base station.

[0019] Among them, slice resource density can be used to indicate the comprehensive density of various resource data of candidate base stations (such as base station CPU utilization, bandwidth utilization, etc.).

[0020] Among them, resource weight can refer to the weight information corresponding to different base stations determined based on the slice resource density of candidate base stations.

[0021] Optionally, in some embodiments, when determining the corresponding resource weight based on the slice resource density of candidate base stations, the sum of the slice resource densities of all candidate base stations can be calculated; the ratio of the slice resource density of each candidate base station to the sum can then be used as the resource weight. This effectively improves the accuracy of the resulting resource weight description.

[0022] In this embodiment of the disclosure, when the corresponding resource weight is determined based on the slice resource density of the candidate base station, reliable reference information can be provided for the subsequent determination of the score value of the candidate base station, thereby providing data support for the slice resource dimension for generating the initial particle swarm.

[0023] S102: Determine the base station density based on the base station coordinates of different candidate base stations, and determine the control parameters for tent mapping based on the base station density.

[0024] Among them, base station density can be used to describe the distribution density of base stations in a specified area.

[0025] The tent map is a mathematical nonlinear dynamic map commonly used in the study of chaotic systems. It is a one-dimensional discrete map, typically used to demonstrate how simple nonlinear equations can generate complex dynamic behaviors. The control parameters of the tent map refer to the variables that affect the system's behavior; for example, they can be used to control the piecewise slope of the tent map.

[0026] In this embodiment of the disclosure, when determining the control parameters for tent mapping based on base station density, it can be based on a pre-trained machine learning model, or it can be based on a combination of numerical and graphical methods, without any limitation.

[0027] Optionally, in some embodiments, when determining the control parameters for tent mapping based on base station density, the matching result between the base station density and a preset relationship table can be determined. This preset relationship table includes multiple candidate density intervals and candidate parameter values ​​corresponding to each interval. Based on the matching result, candidate density intervals that match the base station density are determined. The candidate parameter values ​​corresponding to the matching intervals in the preset relationship table are then used as the control parameters for tent mapping. Therefore, rapid matching of base station density can be achieved based on the preset relationship table, thereby accurately obtaining the applicable control parameters.

[0028] The preset relationship table can refer to a pre-configured relationship table used to indicate the mapping relationship between base station density and control parameters.

[0029] The candidate density interval refers to the interval consisting of resource densities associated with the same candidate parameter value.

[0030] In this embodiment of the disclosure, when the base station density is determined based on the base station coordinates of different candidate base stations, and the control parameters for tent mapping are determined based on the base station density, reliable parameter support can be provided for the subsequent tent mapping process.

[0031] S103: Perform tent mapping based on control parameters and random initial values ​​of candidate base stations to generate chaotic sequences.

[0032] The initial value can be randomly selected between 0 and 1.

[0033] Chaotic sequences refer to sequences obtained by performing tent mapping based on control parameters and random initial values ​​of candidate base stations.

[0034] In this embodiment of the disclosure, when tent mapping is performed based on control parameters and random initial values ​​of candidate base stations to generate a chaotic sequence, reliable data support can be provided for the subsequent calculation of the score value of the candidate base stations.

[0035] S104: Calculate the product of the determined resource weight and the final value in the chaotic sequence, and use it as the score of the candidate base station.

[0036] The final value in the chaotic sequence refers to the value obtained in the last iteration of the tent mapping process.

[0037] In this embodiment of the disclosure, when calculating the product of the determined resource weight and the final value in the chaotic sequence as the score of the candidate base station, the resource information of the base station can be effectively combined in the process of calculating the score of the candidate base station, thereby effectively improving the descriptive accuracy of the obtained score.

[0038] S105: Determine the maximum score value corresponding to different candidate base stations, and use the candidate base station corresponding to the maximum score value as the initial position of the particle to be assigned, wherein the position of the particle to be assigned is used to indicate the task strategy.

[0039] In this embodiment of the disclosure, when the maximum score value corresponding to different candidate base stations is determined, the candidate base station corresponding to the maximum score value is used as the initial position of the particle to be assigned. In this embodiment, resource weights are determined based on the slice resource density of candidate base stations; base station density is determined based on the base station coordinates of different candidate base stations, and control parameters for tent mapping are determined based on the base station density; tent mapping is performed based on the control parameters and random initial values ​​of candidate base stations to generate a chaotic sequence; the product of the determined resource weights and the final values ​​in the chaotic sequence is calculated as the score of the candidate base station; the maximum score of different candidate base stations is determined, and the candidate base station corresponding to the maximum score is used as the initial position of the particles to be assigned, where the position of the particles to be assigned is used to indicate the task strategy. Thus, a non-uniform initial particle swarm can be generated by combining base station distribution and slice resource density, and particle aggregation can be avoided through the chaotic characteristics of tent mapping, improving coverage of resource-dense areas, thereby improving the global search efficiency and resource utilization of the task.

[0040] In summary, this disclosure enables 5G-Tent chaotic initialization: by generating a non-uniform initial particle swarm based on base station distribution and slice resource density through an edge computing platform, global search efficiency is improved. For example, such as Figure 2 As shown, Figure 2 This is a schematic diagram of a 5G chaos initialization process based on tent mapping proposed in this disclosure, wherein the specific steps include: Step 1: Input Data Preparation Base station distribution data: Each base station The coordinate information (latitude and longitude or grid coordinates) originates from base station equipment or network management systems in the 5G network, and is actively reported by the base station or collected uniformly by the network management platform.

[0041] The resource density matrix, ρ(x,y), represents the slice resource density (such as CPU and bandwidth) at coordinates (x,y). It is generated by the edge computing platform, integrating multi-dimensional resource data (such as base station CPU utilization and bandwidth occupancy), and is periodically updated by the real-time monitoring module. For example, the edge server collects real-time resource status from base stations and network slices through the southbound interface.

[0042] Scene parameters: a: Tent mapping control parameters; k: Number of iterations (usually taken as 100≤k≤1000); M: Particle swarm size (e.g., M=50).

[0043] Step 2: Calculate resource weights The weights are calculated based on the resource density of base station i:

[0044] Step 3: Generate a Tent chaotic sequence 1. Initialization parameters: The value of a is dynamically adjusted according to the base station density: If the base station density is >100 base stations / km²: a=0.5 (uniform distribution); if 50 / km²≤base station density≤100 / km²: a=0.4 (biased towards high-density areas); if the base station density is <50: a=0.3 (concentrated in key areas).

[0045] initial value ∈(0,1) is randomly generated.

[0046] 2. Iteratively generate sequences: The mathematical expression for the Tent mapping is:

[0047] Where x is the input value, and a∈(0,1) is the control parameter, a controls the segmented slope of the Tent mapping, affecting the distribution characteristics of the chaotic sequence. When a=0.5, the mapping exhibits symmetry, generating a uniformly distributed chaotic sequence; when a<0.5, the mapping is biased to the right (high-density region), used to concentrate particles in key areas (such as high-resource-density base stations); by adjusting the value of a, the base station density features are encoded into the chaotic sequence generation process, ensuring that the particle initialization stage covers the entire region while focusing on key areas.

[0048] For each base station i, from Start iterating k times:

[0049] Get the final value .

[0050] Step 4: Particle Position Assignment Weighted selection of base stations: For each particle m, calculate the weighted score of all base stations:

[0051] The meanings of each symbol in the above formula are as follows: : This refers to the resource weight (calculated from the resource density matrix ρ(x,y)). : where the superscript i represents the i-th base station (corresponding to the i-th base station in the base station set B); the subscript k represents the Tent chaotic sequence value after k iterations. The overall meaning is the final value of the chaotic sequence generated by base station i through Tent chaotic mapping after k iterations.

[0052] : is the weighted score of base station i.

[0053] Assigning the optimal base station: The initial position of particle m is selected as the base station with the highest score.

[0054] The meanings of each symbol in the above formula are as follows: B: Set of base station coordinates; Resource weight of base station i; : The Tent mapping value after n iterations; : The initial position of the m-th particle.

[0055] During initialization, attention needs to be paid to parameter adjustment and scenario adaptation. The business scenario data source is obtained through the edge computing platform's network slice management system, which retrieves the currently active slice type (URLLC / eMBB / mMTC). For example, when an autonomous vehicle is driving near a service area, the edge computing platform receives its location and task type data, triggering the initialization parameter configuration of k=100.

[0056] The dynamic adjustment rules for parameter a are shown in Table 1: Table 1

[0057] The choice of iteration number k: For applications requiring high real-time performance (such as autonomous driving): k=100 (fast convergence); For complex scenarios (such as multiple slices coexisting): k=500 (fully explore the search space).

[0058] The following example illustrates a real-world scenario: an autonomous vehicle driving near a highway service area. Input data: Service area base station coordinates =(200,300), ρ=0.8; Base stations on ordinary road sections =(150,250), ρ=0.3; a=0.5 (High base station density near the service area).

[0059] Execution process: Calculate the weights: =0.8 / (0.8+0.3)=0.73, =0.27; Tent mapping: Assumption =0.3, after iteration k=100 times, =0.6, =0.4; Score Calculation: =0.73 × 0.6 = 0.44; =0.27 × 0.4 = 0.11; After comparison, the initial position of particle m was chosen as follows: This ensures priority utilization of the high resource density of the service area. The example here is for simplification only. In real-world scenarios, the resource density matrix ρ(x,y) is a set of multiple resource density values ​​that reflect various network characteristics such as CPU and bandwidth, thus providing a more comprehensive reflection of 5G network characteristics and improving the global search efficiency and resource utilization of tasks.

[0060] This application can adaptively adjust according to the scenario, that is, by dynamically adjusting a and k to adapt to different base station density scenarios such as cities and suburbs; at the same time, it adopts efficient global search, avoids particle aggregation through the chaotic characteristics of Tent mapping, and improves coverage in resource-intensive areas. In 5G vehicle networking scenarios, it can significantly improve the global search efficiency and resource utilization of the task, and solve the limitations of traditional methods.

[0061] Figure 3 This is a flowchart illustrating an edge computing method proposed in another embodiment of this disclosure.

[0062] like Figure 3 As shown, the edge computing method includes: S301: Determine the priority description value and slice load for each network slice.

[0063] The priority description value can be used to describe the priority information corresponding to different network slices.

[0064] Understandably, in 5G network slicing, different slices (such as enhanced mobile broadband (eMBB), massive machine-type communications (mMTC), and ultra-reliable low-latency communications (URLLC)) have different service levels and resource requirements. The network management system or slice orchestrator assigns a "priority description value" to each slice. This value may be a number, a level (such as high, medium, low), or a weight, used to represent information such as service importance, resource preemption rights, or scheduling order.

[0065] Slice load can be used to describe the real-time load of the corresponding network slice, such as CPU utilization and bandwidth utilization.

[0066] Optionally, in some embodiments, the index values ​​corresponding to different load types of network slices can be obtained, and the load weights corresponding to different load types can be determined. The index values ​​are then weighted and summed based on the load weights to obtain the slice load of the network slice. Thus, weighted fusion processing of index values ​​corresponding to different load types can be achieved based on the load weights, thereby effectively improving the comprehensive description effect of the obtained slice load.

[0067] Different load types, such as CPU utilization and bandwidth usage, are not restricted.

[0068] The load weight can be flexibly configured according to the load type. For example, the CPU weight factor can be 0.6 and the bandwidth weight factor can be 0.4. There are no restrictions on this.

[0069] In this embodiment of the disclosure, when the priority description value and slice load corresponding to each network slice are determined, multi-dimensional data support can be provided for the subsequent calculation of the slice weight of the network slice.

[0070] S302: Normalize the priority description values ​​of different network slices to obtain the first description value corresponding to the network slice.

[0071] The first description value can refer to the description value obtained by normalizing the priority description values ​​of different network slices.

[0072] In this embodiment of the disclosure, when the priority description values ​​of different network slices are normalized to obtain the first description value corresponding to the network slice, a fair and unified comparison and decision-making benchmark can be achieved.

[0073] S303: Normalize the slice load of different network slices to obtain the second description value corresponding to the network slice.

[0074] The second descriptive value refers to the descriptive value obtained by normalizing the slice load of different network slices.

[0075] In this embodiment of the disclosure, when the slice load of different network slices is normalized to obtain the second description value corresponding to the network slice, the slice load of different network slices can be converted to the same measurement dimension, thereby effectively improving the practicality of the obtained second description value.

[0076] S304: The first and second descriptive values ​​of the network slice are weighted and summed based on a preset priority weight factor to obtain the reference weight.

[0077] The priority weight factor can refer to the weight that the first descriptive value occupies in the weighted processing of the second descriptive value.

[0078] The reference weight refers to the fusion weight obtained by weighting the first descriptive value and the second descriptive value.

[0079] In this embodiment of the disclosure, when the first description value and the second description value of the network slice are weighted and summed based on a preset priority weight factor to obtain a reference weight, the first description value and the second description value can be effectively fused based on the priority weight factor, thereby effectively improving the description effect of the obtained reference weight.

[0080] S305: Determine the product of the reference weight and the preset base weight as the slice weight of the network slice.

[0081] The preset base weight can refer to a pre-configured value used to adjust the reference weight. For example, it can be set to 1, or it can be flexibly adjusted according to the application scenario without any restrictions.

[0082] In this embodiment of the disclosure, when the product of the reference weight and the preset base weight is determined as the slice weight of the network slice, the reference weight can be adjusted a second time based on the preset base weight, thereby ensuring the reliability of the obtained slice weight.

[0083] S306: Determine the resource allocation strategy for network slices based on slice weights and service types.

[0084] In this embodiment, the priority description value and slice load corresponding to each network slice are determined; the priority description values ​​of different network slices are normalized to obtain a first description value corresponding to the network slice; the slice load of different network slices is normalized to obtain a second description value corresponding to the network slice; the first and second description values ​​of the network slice are weighted and summed based on a preset priority weight factor to obtain a reference weight; the product of the reference weight and the preset base weight is determined as the slice weight of the network slice; and the resource allocation strategy of the network slice is determined according to the slice weight and the service type. Thus, the inertial weight and the slice load factor can be dynamically coupled, and by monitoring the network status in real time, calculating the dynamic slice weight based on the slice load and priority, and selecting a scheduling strategy according to the service type, efficient resource allocation can be achieved.

[0085] Understandably, in 5G networks, the real-time resource requirements of different network slices (such as URLLC, eMBB, and mMTC) vary significantly. Traditional fixed-weight allocation leads to low resource utilization and large latency fluctuations. This solution proposes a Dynamic Weighted Slicing Algorithm (DWSA) based on slice load and priority, applied to edge computing platforms. By monitoring network status in real time and dynamically adjusting slice weights, it achieves efficient resource allocation.

[0086] In this application, dynamic weight calculation is used to determine the weight of slice i. Calculated using the following formula:

[0087] in: Base weights (preset values, such as...) ); Priority weighting factor ( ,For example (Indicates priority is dominant). The priority of slice i (e.g.) , , ); Real-time load of slice i (e.g., CPU utilization, bandwidth usage).

[0088] In this application, slice load calculation includes the real-time load of slice i. Calculated using weighted averages of the following indicators:

[0089] in: CPU weighting factor (e.g.) ); CPU utilization of slice i ( ); : Bandwidth utilization of slice i ( ).

[0090] like Figure 4 As shown, Figure 4 This is a schematic diagram of the dynamic network slice weight adjustment process proposed in this disclosure.

[0091] The above system parameters α, β, It is configured by network operator policies or dynamically optimized through machine learning models.

[0092] Step 1: Real-time monitoring and data collection Input data: Slice priority list: ; Real-time resource usage data: ; System parameters: .

[0093] Output: Real-time load of each slice .

[0094] Step 2: Calculate slice weights 1. Calculate the load weighting value: For each slice i, calculate :

[0095] 2. Normalized priority and load:

[0096] 3. Dynamic weight calculation:

[0097] Step 3: Resource Allocation and Scheduling Selecting a scheduling algorithm: Choose a scheduling strategy based on the service type. URLLC slices: use priority scheduling, by Allocate resources; eMBB slicing: using proportional fair scheduling (PF), according to Allocate bandwidth; mMTC slicing: uses Max-Min fair scheduling to guarantee the minimum resource requirements.

[0098] Execution allocation: Dynamically adjust the resource quota of slices through the SDN controller.

[0099] Step 4: Feedback and Adjustments Data is collected in real time and weights are updated dynamically. For details, please refer to the third step, Collaborative Optimization and Resource Allocation. It will not be repeated here.

[0100] During the design process, attention should be paid to parameter adjustment and scene adaptation. The rules for dynamic parameter adjustment are shown in Table 2. Table 2

[0101] The following example illustrates a real-world scenario where URLLC (Autonomous Driving) and eMBB (Engineering-by-Broadband) slices coexist in 5G vehicle-to-everything (V2X) networks. Input data: Slice priority: , ; Real-time load: , ; , ; Parameters: α=0.7, β=0.6.

[0102] Execution process: (1) Calculate the load: ; ; (2) Normalization: ; ; (3) Weight calculation: ; ; (4) Resource allocation: URLLC allocates 53% of CPU and bandwidth resources, and eMBB allocates 47%, to ensure low latency for autonomous driving tasks.

[0103] This application features strong dynamic adaptability: by adjusting the α and β parameters, it can flexibly adapt to different service scenarios such as URLLC and eMBB; it achieves resource balance by combining priority and load to avoid resource allocation imbalance caused by a single indicator; it guarantees low latency, with a second-level feedback mechanism ensuring that the latency of URLLC slices is <10 ms; and it also has high isolation, reducing resource conflicts between slices by more than 80% through dynamic weight adjustment. This solution can significantly improve resource utilization, reduce latency fluctuations, and meet the requirements of different services in 5G network slicing scenarios.

[0104] Figure 5 This is a flowchart illustrating an edge computing method proposed in another embodiment of this disclosure.

[0105] like Figure 5 As shown, the edge computing method includes: S501: Collect real-time performance data during task execution based on task strategy. The real-time performance data includes task latency and resource utilization. Task latency is the sum of transmission latency and processing latency. Resource utilization is determined by combining the single utilization rate corresponding to different resource types.

[0106] In this embodiment of the disclosure, when real-time performance data is collected during the execution of a task based on a task strategy, reliable data support can be provided for the subsequent calculation of particle fitness.

[0107] S502: Input the task latency, resource utilization, slice weight, and preset fixed constraint penalty weight into the cooperative fitness function to calculate the particle fitness. The fixed constraint penalty weight is used to weight the constraint violation penalty term in the cooperative fitness function. The constraint violation penalty term is triggered when the slice resource threshold is violated.

[0108] The cooperative fitness function can refer to a pre-configured function used to calculate particle fitness.

[0109] In other words, in this embodiment of the present disclosure, after collecting real-time performance data, task latency, resource utilization, slice weight, and preset fixed constraint penalty weight can be input into the cooperative fitness function to calculate the particle fitness, thereby providing a reliable basis for determining the optimal particle in the future.

[0110] S503: Determine the optimal particle from multiple particles based on particle fitness.

[0111] In other words, in the embodiments of this disclosure, after calculating the particle fitness, the optimal particle can be determined from multiple particles based on the particle fitness.

[0112] In this embodiment, real-time performance data is collected during task execution based on a task strategy. This real-time performance data includes task latency and resource utilization. Task latency is the sum of transmission latency and processing latency, and resource utilization is determined by combining the individual utilization rates corresponding to different resource types. Task latency, resource utilization, slice weights, and preset fixed constraint penalty weights are input into a cooperative fitness function to calculate particle fitness. The fixed constraint penalty weights are used to weight the constraint violation penalty terms in the cooperative fitness function, which are triggered when a slice resource threshold is violated. Based on the particle fitness, the optimal particle is determined from multiple particles. Thus, 5G-Tent chaotic initialization and dynamic network slice weight adjustment can be combined. The dynamic network slice weight adjustment structure provides a real-time target for particle swarm optimization, and the 5G-Tent chaotic initialization optimization results provide resource feedback for dynamic weight adjustment. This enables dynamic balancing of task and resource allocation in a rapidly changing 5G network to meet the needs of multiple services.

[0113] In summary, of the above embodiments, The first step (5G-Tent chaotic initialization) in the aforementioned scheme serves to generate an initial particle swarm through chaotic initialization. The particle positions represent possible task strategies (such as which base station the task is assigned to), providing diverse initial task strategies that cover differences in base station distribution and resource density.

[0114] The second step in the aforementioned scheme (dynamic network slice weight adjustment) serves to determine the slice resource allocation strategy (such as the resource ratio of URLLC and eMBB slices) through dynamic weight adjustment. It dynamically adjusts the slice resource allocation weights according to the real-time load, providing dynamic constraints for particle swarm optimization.

[0115] The goal of collaborative optimization and resource allocation is to combine 5G-Tent chaotic initialization with dynamic network slice weight adjustment to form a closed-loop optimization system. The dynamic network slice weight adjustment structure provides a real-time target for particle swarm optimization; the optimization results of 5G-Tent chaotic initialization provide resource feedback for dynamic weight adjustment. The two form a closed-loop optimization. Through this design, the system can dynamically balance task and resource allocation in a rapidly changing 5G network to meet the needs of multiple services.

[0116] like Figure 6 As shown, Figure 6 This is a schematic diagram of the collaborative optimization and resource allocation process proposed in this disclosure, wherein, In this application, the specific formula for the collaborative fitness function is as follows:

[0117] The symbols and calculation logic in the above formulas are shown in Table 3: Table 3

[0118] The following explains the collaborative workflow in a connected vehicle scenario, using the actual scenario from the previous step as an example: The dynamic weight allocation is as follows:

[0119]

[0120]

[0121] Step 2: Assess particle fitness Particle A (task assigned to service area base station): Assuming, ,

[0122]

[0123] Particle B (task assigned to a nearby base station): Assuming, ,

[0124]

[0125] Step 3: Select the optimal particle Result: Particle A has a lower fitness ( < The system prioritizes service area base stations, with an autonomous driving task latency of 3ms, meeting the requirements. In addition, it is necessary to monitor base station overload risks; if the CPU load of a service area base station rises to 92%, a threshold will be triggered. The value will be greater than zero, after... After magnification, The value will increase significantly beyond At this point, a response is needed to remove the particles selected for that base station and regenerate particles based on the resource density of neighboring base stations. Simultaneously, by dynamically adjusting α, β, The value is optimized by improving the scheduling strategies of URLLC and eMBB slicing to avoid overload and partially offload tasks, thereby continuously optimizing the slicing scheduling strategy.

[0126] The core mechanisms of this application are summarized in Table 4: Table 4

[0127] The practical significance of dynamic weights includes: Delay item is dominant ( High): When URLLC tasks surge (e.g., dense lane changes by vehicles), weights shift towards latency, forcing particles to choose low-latency paths. Even when the service area base station load is 90%, latency is still prioritized, sacrificing some resource utilization. Resource item dominant ( High): When eMBB bandwidth demand surges (e.g., multiple vehicles uploading video), weights are shifted towards resources, guiding particles to select high-bandwidth base stations. When the bandwidth of neighboring base stations reaches 80%, video tasks are offloaded to avoid service area overload; By dynamically balancing latency, resource utilization, and constraint compliance, multi-service collaborative optimization is achieved. In the Internet of Vehicles (IoV), URLLC tasks are strictly time-limited, while eMBB tasks guarantee bandwidth. The system automatically balances priorities and real-time load, avoiding delays caused by manual parameter tuning.

[0128] In summary, the edge computing method proposed in this application can be implemented through a three-stage optimized architecture, such as... Figure 7 As shown, Figure 7 This is a schematic diagram of an edge computing system architecture based on the present disclosure, wherein, 1. 5G-Tent Chaotic Initialization: A non-uniform initial particle swarm is generated based on the distribution of base stations and the density of slice resources through the edge computing platform, thereby improving the efficiency of global search.

[0129] 2. Dynamic network slice weight adjustment: By dynamically coupling the edge computing platform's inertial weight with the slice load factor, the collaborative scheduling capability of edge servers is enhanced.

[0130] 3. Collaborative optimization and resource allocation: The 5G-Tent chaotic initialization and dynamic network slice weight adjustment are combined to form a closed-loop optimization system, ensuring that task strategies and resource allocation strategies are mutually compatible.

[0131] Figure 8 This is a schematic diagram of the structure of an edge computing device proposed in an embodiment of the present disclosure.

[0132] like Figure 8 As shown, the edge computing device 80 includes: The first determining module 801 is used to determine the corresponding resource weights based on the slice resource density of the candidate base stations; The second determining module 802 is used to determine the base station density based on the base station coordinates of different candidate base stations, and to determine the control parameters for tent mapping based on the base station density. The generation module 803 is used to perform tent mapping based on control parameters and random initial values ​​of candidate base stations to generate a chaotic sequence; The third determining module 804 is used to calculate the product of the determined resource weight and the final value in the chaotic sequence, as the score of the candidate base station; The fourth determining module 805 is used to determine the maximum value of the corresponding score value of different candidate base stations, and to take the candidate base station corresponding to the maximum value as the initial position of the particle to be assigned, wherein the position of the particle to be assigned is used to indicate the task strategy.

[0133] It should be noted that the foregoing explanation of the edge computing method also applies to the edge computing device of this embodiment, and will not be repeated here.

[0134] In this embodiment, resource weights are determined based on the slice resource density of candidate base stations; base station density is determined based on the base station coordinates of different candidate base stations, and control parameters for tent mapping are determined based on the base station density; tent mapping is performed based on the control parameters and random initial values ​​of candidate base stations to generate a chaotic sequence; the product of the determined resource weights and the final values ​​in the chaotic sequence is calculated as the score of the candidate base station; the maximum score of different candidate base stations is determined, and the candidate base station corresponding to the maximum score is used as the initial position of the particles to be assigned, where the position of the particles to be assigned is used to indicate the task strategy. Thus, a non-uniform initial particle swarm can be generated by combining base station distribution and slice resource density, and particle aggregation can be avoided through the chaotic characteristics of tent mapping, improving coverage of resource-dense areas, thereby improving the global search efficiency and resource utilization of the task.

[0135] According to embodiments of this application, this application also provides an electronic device and a readable storage medium.

[0136] Figure 9 This is a block diagram of an electronic device according to an embodiment of the present application.

[0137] like Figure 9 As shown, the electronic device includes: The memory 901, the processor 902, and the computer instructions stored in the memory 901 and executable on the processor 902.

[0138] When processor 902 executes instructions, it implements the edge computing method provided in the above embodiments.

[0139] Furthermore, electronic devices also include: Communication interface 903 is used for communication between memory 901 and processor 902.

[0140] Memory 901 is used to store computer instructions that can be executed on processor 902.

[0141] The memory 901 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0142] The processor 902 is used to implement the edge computing method of the above embodiments when executing the program.

[0143] If the memory 901, processor 902, and communication interface 903 are implemented independently, then the communication interface 903, memory 901, and processor 902 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 9 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0144] Optionally, in a specific implementation, if the memory 901, processor 902, and communication interface 903 are integrated on a single chip, then the memory 901, processor 902, and communication interface 903 can communicate with each other through an internal interface.

[0145] The processor 902 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0146] This application also proposes a computer program product that implements the edge computing method of the embodiments of this application when the instruction processor in the computer program product is executed.

[0147] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0148] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0149] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0150] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0151] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0152] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0153] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0154] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. An edge computing method, characterized in that, include: The corresponding resource weights are determined based on the slice resource density of the candidate base stations; The base station density is determined based on the base station coordinates of the different candidate base stations, and the control parameters for tent mapping are determined based on the base station density. Tent mapping is performed based on the control parameters and the random initial values ​​of the candidate base stations to generate a chaotic sequence; The product of the determined resource weight and the final value in the chaotic sequence is calculated and used as the score of the candidate base station. The maximum value of the score corresponding to different candidate base stations is determined, and the candidate base station corresponding to the maximum value is used as the initial position of the particle to be assigned, wherein the position of the particle to be assigned is used to indicate the task strategy.

2. The method as described in claim 1, characterized in that, The step of determining the corresponding resource weight based on the slice resource density of the candidate base station includes: Calculate and determine the sum of the slice resource densities corresponding to all the candidate base stations; The ratio of the slice resource density corresponding to the candidate base station to the sum value is determined as the resource weight.

3. The method as described in claim 1, characterized in that, The control parameters for determining tent mapping based on the base station density include: Determine the matching result between the base station density and the preset relationship table, wherein the preset relationship table includes multiple candidate density intervals and candidate parameter values ​​corresponding to each candidate density interval; Based on the matching results, the candidate density interval that matches the base station density is determined; The candidate parameter value corresponding to the candidate density interval that matches the preset relationship table is used as the control parameter for tent mapping.

4. The method as described in claim 1, characterized in that, The method further includes: Determine the priority description value and slice load for each network slice; The priority description values ​​of different network slices are normalized to obtain a first description value corresponding to the network slice; The slice loads of different network slices are normalized to obtain a second description value corresponding to the network slice; The first and second description values ​​of the network slice are weighted and summed based on a preset priority weight factor to obtain a reference weight. The product of the reference weight and the preset base weight is determined as the slice weight of the network slice; Based on the slice weight and service type, the resource allocation strategy for the network slice is determined.

5. The method as described in claim 4, characterized in that, in, The slice load is obtained based on the following method: Obtain the index values ​​corresponding to the network slices and different load types, and determine the load weights corresponding to different load types; The index values ​​are weighted and summed based on the load weights to obtain the slice load of the network slice.

6. The method as described in claim 4, characterized in that, The method further includes: Real-time performance data is collected during the execution of tasks based on the task strategy. The real-time performance data includes task latency and resource utilization. The task latency is the sum of transmission latency and processing latency. The resource utilization is determined by combining the single utilization rate corresponding to different resource types. The task latency, resource utilization, slice weight, and preset fixed constraint penalty weight are input into the cooperative fitness function to calculate the particle fitness. The fixed constraint penalty weight is used to weight the constraint violation penalty term in the cooperative fitness function. The constraint violation penalty term is triggered when the slice resource threshold is violated. Based on the particle fitness, the optimal particle is determined from multiple particles.

7. An edge computing device, characterized in that, include: The first determining module is used to determine the corresponding resource weights based on the slice resource density of the candidate base stations; The second determining module is used to determine the base station density based on the base station coordinates of different candidate base stations, and to determine the control parameters for tent mapping based on the base station density. A generation module is used to perform tent mapping based on the control parameters and the random initial values ​​of the candidate base stations to generate a chaotic sequence; The third determining module is used to calculate the product of the determined resource weight and the final value in the chaotic sequence, as the score of the candidate base station; The fourth determining module is used to determine the maximum value of the score corresponding to different candidate base stations, and to take the candidate base station corresponding to the maximum value as the initial position of the particle to be assigned, wherein the position of the particle to be assigned is used to indicate the task strategy.

8. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.

9. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, in, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-6.

10. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the steps of the method according to any one of claims 1-6.

Images