MEC task unloading method based on chance constraint

By constructing a joint optimization model and a hybrid offloading algorithm, the uncertainty and privacy protection issues of task offloading in multi-access edge computing are solved, achieving efficient, low-cost and secure task offloading decisions and optimizing resource allocation between terminals and cloud computing centers.

CN121396989APending Publication Date: 2026-01-23XIAN UNIV OF POSTS & TELECOMM
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202511783311.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-12-01
Publication Date
2026-01-23

AI Technical Summary

Technical Problem

Existing technologies in multi-access edge computing suffer from environmental uncertainties caused by dynamic changes in task characteristics and network resources, heterogeneity of multiple network accesses, and personalized privacy protection requirements. These issues lead to inaccurate offloading decisions, increased terminal pressure, and data transmission consumption.

Method used

By constructing a latency-energy consumption model, a fuzzy privacy model, and a chance-constrained model, a joint optimization model is established. An improved hybrid offloading algorithm combining adaptive genetics, Monte Carlo simulation, and neural networks is used to solve the task offloading strategy and network selection strategy, thereby optimizing the allocation of tasks across terminals, edge servers, and cloud computing centers.

Benefits of technology

It achieves the lowest latency, energy consumption, and privacy protection in multi-access edge computing, improves the efficiency and security of task offloading, reduces network transmission occupancy, and optimizes the task processing process.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121396989A_ABST
    Figure CN121396989A_ABST
Patent Text Reader

Abstract

The invention discloses an MEC task unloading method based on opportunity constraint, and relates to the technical field of cloud computing, and the method comprises the steps: building a time delay energy consumption model, a fuzzy privacy model and an opportunity constraint model through the fuzzy data input quantity and the fuzzy CPU cycle number of a fuzzy to-be-processed task; establishing a joint optimization model through a time delay energy consumption model, a fuzzy privacy model and an opportunity constraint model; solving the joint optimization model to obtain a task unloading strategy and a network selection strategy; distributing the fuzzy to-be-processed task to an unloading position according to a task unloading strategy; and performing resource transmission according to the network selection strategy. According to the method, the joint optimization model is solved through the designed unloading algorithm, and the optimal solution about the task unloading position strategy and the network data transmission selection strategy is obtained. And distributing the tasks to the corresponding unloading positions. And according to the network data transmission selection strategy, distributing task data to different networks according to the proportion so as to transmit the data.
Need to check novelty before this filing date? Find Prior Art

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cloud computing, and particularly relates to a MEC task offloading method based on an opportunity constraint. BACKGROUND

[0002] MEC, the full name of which is Multi-Access Edge Computing, is a new type of computing mode that can provide a user with three-layer resources of a cloud server, an edge server and a terminal. Multi-Access Edge Computing can increase the available resources of a terminal and improve the task processing efficiency by offloading a computing task from the terminal to an edge server and a cloud server. However, when a user decides whether a task is executed locally on a terminal or offloaded to an edge server or a cloud server for execution, an accurate decision needs to be made. An incorrect offloading decision not only cannot reduce the pressure on the terminal, but also can cause more consumption due to additional data transmission.

[0003] In the prior art, there are still some problems in dealing with complex and variable network environments and uncertain task requirements, such as environmental uncertainty caused by dynamic changes of task characteristics and network resources, heterogeneity of multi-network access and personalized privacy protection requirements. SUMMARY

[0004] The embodiment of the present application provides a MEC task offloading method based on an opportunity constraint, to solve the problems of environmental uncertainty caused by dynamic changes of task characteristics and network resources, heterogeneity of multi-network access and personalized privacy protection requirements in the prior art.

[0005] In one aspect, the embodiment of the present application provides a MEC task offloading method based on an opportunity constraint, comprising: constructing a time delay-energy consumption model, a fuzzy privacy model and an opportunity constraint model by fuzzifying the fuzzy data input quantity and the fuzzy CPU cycle number of a to-be-processed task; establishing a joint optimization model through the time delay-energy consumption model, the fuzzy privacy model and the opportunity constraint model; solving the joint optimization model to obtain a task offloading strategy and a network selection strategy; allocating the fuzzy to-be-processed task to an offloading position according to the task offloading strategy; performing resource transmission according to the network selection strategy.

[0006] In one possible implementation manner, the offloading position comprises a terminal, an edge server and a cloud computing center, and the cloud computing center is provided with a plurality of cloud servers.

[0007] In a possible implementation, the time delay and energy consumption model is used to obtain time delay and energy consumption cost of the to-be-processed task in the process of data transmission and in the process of processing at the offloading position, and the time delay and energy consumption model is used to simultaneously perform data transmission between the terminal and the edge server and between the edge server and the cloud server through multiple wireless networks.

[0008] In a possible implementation, the fuzzy privacy model comprises a fuzzy privacy level requirement of the fuzzy to-be-processed task and data privacy entropy quantization, the fuzzy privacy level requirement is used to constrain the degree of privacy protection of the fuzzy to-be-processed task in the transmission process, and the data privacy entropy quantization is used to obtain a leakage probability and a leakage level of data of the fuzzy to-be-processed task in the data transmission process.

[0009] In a possible implementation, the chance constraint model is used to convert between determination of the fuzzy task data amount and the number of CPU cycles and determination of the fuzzy privacy model in the form of a chance constraint.

[0010] In a possible implementation, the time delay and energy consumption model comprises a communication model and a calculation model, the communication model is used to obtain time delay cost of the fuzzy to-be-processed task, and the calculation model is used to obtain energy consumption cost of the to-be-processed task.

[0011] In a possible implementation, solving the joint optimization model to obtain a task offloading strategy and a network selection strategy comprises: Solving the joint optimization model according to the improved hybrid offloading algorithm of adaptive genetic algorithm, Monte Carlo simulation and neural network.

[0012] The MEC task offloading method based on a chance constraint in the application has the following advantages: (1) By converting the multi-access edge computing task offloading problem into a joint optimization model based on a communication model, a calculation model, a privacy entropy model and a chance constraint model, the problem that the offloading decision in the multi-access edge computing task offloading process is insufficient to meet high task processing efficiency, the local terminal processing pressure is relatively large, and time delay and energy consumption cost is relatively high in the process of additional data transmission is solved from the perspective of an uncertain environment, the perspective of time delay and energy consumption measurement and the perspective of privacy protection.

[0013] (2) The joint optimization model is solved by the designed unloading algorithm, and the optimal solution about the task unloading position strategy and the network data transmission selection strategy is obtained. According to the task unloading strategy, the task is distributed to the corresponding unloading position. According to the network data transmission selection strategy, the task data is proportionally distributed to different networks for data transmission. The minimum time delay energy cost problem in the task unloading process and the privacy protection problem in the data transmission process are solved, and the efficiency, low cost and security of task unloading are realized. At the same time, the multi-terminal transmission between the edge server and other terminals can be realized, the occupancy rate of network transmission is reduced, and the transmission speed is improved. BRIEF DESCRIPTION OF DRAWINGS

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0015] Figure 1 A flowchart of a MEC task unloading method based on opportunity constraints provided by the present application is shown in the figure. Figure 2 A flowchart of a hybrid unloading HTOA-IMN algorithm based on improved adaptive genetic, Monte Carlo simulation and neural network provided by the embodiment of the present application is shown in the figure. Figure 3 A schematic diagram of target values and task distribution ratios under different computing powers of different unloading positions using the unloading method of the present application provided by the embodiment of the present application is shown in the figure. A schematic diagram of target values and task distribution ratios under different computing powers of local terminal unloading positions using the unloading method of the present application provided by the embodiment of the present application is shown in the figure. Figure 4 A schematic diagram of target values and task distribution ratios under different computing powers of edge server unloading positions using the unloading method of the present application provided by the embodiment of the present application is shown in the figure. Figure 5 A schematic diagram of target values and task distribution ratios under different computing powers of cloud server unloading positions using the unloading method of the present application provided by the embodiment of the present application is shown in the figure. A schematic diagram of the influence of the change of the number of network accesses on the target value using the unloading method of the present application provided by the embodiment of the present application is shown in the figure. Figure 6 A schematic diagram of the influence of the change of the number of network accesses on the target value using the unloading method of the present application provided by the embodiment of the present application is shown in the figure. Figure 7 A schematic diagram of the influence of the change of the number of network accesses on the target value using the unloading method of the present application provided by the embodiment of the present application is shown in the figure. Figure 8 ​​A target value schematic diagram provided by the embodiment of the present application using the unloading method of the present application in the stage of the number of tasks being 10 to 40; Figure 9 A target value schematic diagram provided by the embodiment of the present application using the unloading method of the present application in the stage of the number of tasks being 60 to 90; Figure 10 A target value schematic diagram provided by the embodiment of the present application using the unloading method of the present application in the stage of the number of tasks being 60 to 90; Figure 11 A target value schematic diagram provided by the embodiment of the present application using the unloading method of the present application in the stage of the number of tasks being 60 to 90; Figure 12 A performance comparison line schematic diagram of the unloading method of the present application and other technologies in different iteration numbers; Figure 13 A performance comparison box plot of the unloading method of the present application and other technologies in different iteration numbers. DETAILED DESCRIPTION

[0016] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0017] Figure 1 A flow schematic diagram of a MEC task unloading method based on an opportunity constraint provided by the embodiment of the present application; the embodiment of the present application provides a MEC task unloading method based on an opportunity constraint, which comprises: A delay energy consumption model, a fuzzy privacy model and an opportunity constraint model are constructed by fuzzing the fuzzy data input quantity and the fuzzy CPU cycle number of the to-be-processed task; A joint optimization model is established through the delay energy consumption model, the fuzzy privacy model and the opportunity constraint model; A task unloading strategy and a network selection strategy are obtained by solving the joint optimization model; The fuzzy to-be-processed task is distributed to an unloading location according to the task unloading strategy; Resource transmission is performed according to the network selection strategy.

[0018] The unloading location comprises a terminal, an edge server and a cloud computing center, and the cloud computing center is provided with a plurality of cloud servers.

[0019] The time delay and energy consumption model is used to obtain the time delay and energy consumption cost of the to-be-processed task in the process of data transmission and in the process of processing at the offloading position, and the time delay and energy consumption model is used to simultaneously perform data transmission between the terminal and the edge server and between the edge server and the cloud server through multiple wireless networks.

[0020] The fuzzy privacy model includes a fuzzy privacy level requirement of the fuzzy to-be-processed task and data privacy entropy quantization, the fuzzy privacy level requirement is used to constrain the degree of privacy protection of the fuzzy to-be-processed task in the transmission process, and the data privacy entropy quantization is used to obtain a leakage probability and a leakage level of data of the fuzzy to-be-processed task in the data transmission process.

[0021] The chance constraint model is used to convert between the determination of the fuzzy task data quantity and the number of CPU cycles and the determination of the fuzzy privacy model in the form of a chance constraint.

[0022] The time delay and energy consumption model includes a communication model and a calculation model, the communication model is used to obtain the time delay cost of the fuzzy to-be-processed task, and the calculation model is used to obtain the energy consumption cost of the to-be-processed task.

[0023] Solving the joint optimization model to obtain the task offloading strategy and the network selection strategy includes: Solving the joint optimization model according to the improved hybrid offloading algorithm of adaptive genetic, Monte Carlo simulation and neural network.

[0024] For example, the offloading position of the present application includes a terminal, an edge server and a cloud computing center, which are connected through l, s, c respectively. The terminal and the edge server can be connected in parallel through multiple wireless networks, and the edge server and the cloud computing center can be connected through wireless or wired networks.

[0025] The terminal can be an intelligent monitoring device, an intelligent robot or other industrial equipment, the wireless network can be a wireless wide area network, a wireless local area network, a wireless metropolitan area network and a wireless personal local area network, and the protocol of network transmission can be Bluetooth, 5G, WIFI, ZigBee and the like. The edge server is regarded as an edge computing center closer to the terminal side. The cloud computing center layer includes multiple cloud servers or high-performance servers.

[0026] The information of the to-be-processed task includes a fuzzy input data quantity of the to-be-processed task and a fuzzy number of CPU cycles required for processing the to-be-processed task.

[0027] In the embodiment of the present application, the information of the to-be-processed task is as follows.

[0028]

[0029] In the formula, denotes the task to be processed, and the subscript n denotes the task label in the set, is the fuzzy input data volume of the task to be processed, is the number of fuzzy CPU cycles required to process the task to be processed, where , is the lower bound of the pessimistic estimate, , is the upper bound of the optimistic estimate, and constitute the actual value confidence interval. The parameterized model allows the decision maker to reflect the differentiated assessment of the task attributes by adjusting the interval boundary parameters, both of which are fuzzy numbers.

[0030] A joint optimization model is established through the delay-energy consumption model, the fuzzy privacy model and the chance constraint model.

[0031] In the embodiments of the present application, the delay-energy consumption model can support the use of multiple wireless networks to simultaneously transmit task data to the offloading location, which includes the terminal, the edge server and the cloud computing center server.

[0032] Therefore, the communication model supports the use of multiple wireless networks to simultaneously transmit data between the terminal and the edge server, and between the edge server and the cloud server.

[0033] The construction process of the delay-energy consumption model is as follows: When the task to be processed is offloaded to the edge server, the data uplink transmission delay cost is calculated as: .

[0034] wherein, denotes the offloading from the local terminal to the edge terminal, denotes the maximum delay of the terminal in the multi-network set ; denotes the data transmission allocation strategy of the nth task; denotes the task data transmission proportion of the nth task under the mth wireless network, multiplied by the fuzzy total data volume of the nth task to obtain , as the fuzzy data transmission volume of the task under the mth network; the denominator comes from the Shannon formula, wherein , , and respectively denote the bandwidth, transmission power, channel gain and noise power spectral density of the mth wireless network.

[0035] At the same time, the data transmission energy consumption during the offloading of the task to the edge server is the cumulative sum of energy consumption of all wireless networks, which can be expressed as .

[0036] When the task to be processed is offloaded to the cloud server, the task data must first be transmitted to the access point (co-located with the edge server) through the wireless network, and then transmitted to the cloud through the fiber backbone network. The data transmission delay of the task offloaded to the cloud is expressed as: .

[0037] wherein, represents that the task is offloaded to the cloud by the local terminal; represents the round-trip propagation delay of the fiber backbone network. Considering that the fiber transmission has higher energy efficiency than the wireless transmission, the negligible energy consumption is ignored, and the data transmission energy consumption is expressed as .

[0038] In addition to the delay and energy consumption cost of the above task offloading process, the corresponding computing delay and energy consumption of the task processed by the different offloading locations are calculated as follows: , , wherein, represents the CPU energy consumption coefficient, represents the three offloading locations, l : local, s : edge server, c : cloud server, represents the computing power of the offloading location.

[0039] The above delay and energy consumption of the task to be processed offloaded to each offloading location based on the delay and energy consumption model can obtain the total delay and total energy consumption cost of each offloading location, for example: the total delay cost of the task offloaded to the edge server by the local terminal and the total energy consumption cost can be calculated as and respectively. In addition, the minimum task processing delay and energy consumption can also be obtained.

[0040] The embodiment of the application further provides a fuzzy privacy model, including the fuzzy privacy level requirement of the task and the data privacy entropy quantization.

[0041] In order to evaluate the channel security, the transmitted data is divided into several blocks with a length of bit. Let represent the energy consumption of the mth wireless network for the task The number of data blocks transmitted. The calculation formula is , in Let represent the upper bound function. Let represent the probability of leakage of k data blocks on the m-th wireless network. Assume that privacy leakage events follow a Poisson distribution. The calculation formula is , in Represents the average privacy leakage on the m-th wireless network, where... It is the base of the natural logarithm. Through the m-th wireless network and the task... Privacy entropy associated with data transmission Calculated using the following formula: .

[0042] Fuzzy task privacy level requirements are used to constrain the degree of privacy protection during task transmission. The requirements for data privacy and security protection during transmission are divided into three levels: "low," "medium," and "high." Different security levels correspond to different fuzzy entropy thresholds. , and By calculating the privacy entropy threshold and tasks Privacy entropy The difference quantifies user satisfaction with privacy protection, and is expressed as... Negative values ​​indicate that the user's privacy and security level meets the threshold requirements, while positive values ​​indicate insufficient privacy protection.

[0043] The fuzzy task privacy level requirement is used to constrain the degree of privacy protection for the task during transmission; the data privacy entropy is used to obtain the probability and level of data leakage of the task to be processed during data transmission. Different privacy protection levels, different specific locations, and different multi-network transmission strategies set by the user will result in different probabilities of information leakage, thus affecting the final offloading strategy.

[0044] The joint optimization model consists of an objective function and multiple constraints. The objective function aims to minimize task processing time and minimize task processing energy consumption, and is expressed as: .

[0045] in and Representing tasks Different uninstall locations i ( i e l, s, clatency and energy cost of the task, denote the total number of tasks. denote the offloading decision at the corresponding offloading location, i denote the total number of tasks. N denote the total number of tasks.

[0046] The constraints of the joint optimization model are as follows: (c1) , which strictly controls the user privacy level within a predefined threshold, thereby ensuring the security of user privacy.

[0047] (c2) , which mathematically imposes a restriction on the offloading strategy, ensuring that each task is executed only once in the mobile terminal, edge, or cloud.

[0048] (c3) , which imposes a mathematical restriction on the allocation strategy, such that no data transmission occurs when the task is executed locally, and all data transmissions are completed by wireless network collaboration when the task is offloaded.

[0049] (c4) , which represents the offloading decision of different offloading locations is a binary value, either 0 or 1.

[0050] (c5) , which limits the network transmission data allocation strategy of the nth task in the mth wireless network to any value between 0 and 1.

[0051] The embodiments of the present application also provide a chance constraint model, which includes the fuzzification of task data volume and CPU cycle number, the determination of objective function, and the conversion of fuzzy privacy model chance constraint form.

[0052] The of the chance constraint is a possibility measure function, whose value quantifies the degree of possibility that the input fuzzy condition is satisfied. Under the fuzzy chance constraint, the optimization objective is transformed into minimizing At the same time, the objective function based on the fuzzy chance constraint and the privacy entropy constraint are respectively transformed into , .

[0053] wherein and are preset confidence levels, respectively measuring the credibility of the satisfaction of the privacy constraint and the objective function constraint. Subsequently, the joint optimization model can be transformed into the following form: .

[0054] An optimal solution is admissible if the probability of satisfying the constraints is at least a pre-specified confidence threshold, while allowing occasional violations of the constraints. By transforming the fuzzy chance constraints, the solution process is converted to finding the minimum objective function subject to a confidence level of satisfying the constraints.

[0055] The model transformation converts the original uncertain optimization problem into a deterministic equivalent form with provable confidence level guarantee. The transformed model preserves the physical interpretation and structural properties of the original task offloading problem. It converts the original uncertain optimization problem involving complex membership function processing into an equivalent deterministic problem. The transformation provides a mathematical formula basis for replacing complex fuzzy analysis derivation with numerical calculation, thereby reducing the difficulty of solving the task offloading problem.

[0056] In order to obtain the optimal task offloading and data allocation strategy from the transformed optimization model, a hybrid task offloading algorithm based on improved adaptive genetic algorithm, Monte Carlo simulation and neural network is designed and proposed.

[0057] The adaptive genetic operator with dynamically adjusted evolutionary parameters and an improved repair mechanism are combined to handle infeasible solutions while preserving their genetic information. In order to further improve the running efficiency of the task offloading algorithm, neural networks are used instead of computationally intensive Monte Carlo simulations. In the offline phase, samples are generated by Monte Carlo simulation to train the neural network, thereby accurately approximating the input-output mapping relationship of the fuzzy chance constraint. In the online phase, the trained neural network is integrated into the optimization process of the algorithm to estimate the fuzzy chance constraint value in real time.

[0058] As shown in Figure 2 , the hybrid task offloading algorithm based on improved adaptive genetic algorithm, Monte Carlo simulation and neural network has six main steps to solve the joint optimization model, which can be summarized as follows: Initialize the random population, calculate the individual fitness and algorithm parameters.

[0059] According to the selection strategy set by the algorithm, select part of the individuals, and perform mutual crossover and mutation between individuals through adaptive parameters, and repair the gene fragments of the offspring chromosomes that violate the constraints.

[0060] Perform greedy genetic operations on the non-gene fragments of the offspring chromosomes.

[0061] Re-evaluate the offspring chromosomes using the trained neural network.

[0062] Update the population optimal solution, and generate new individuals to join the population at random iteration intervals.

[0063] Repeat the above steps until the iteration meets the termination criteria, and output the global optimal solution and fitness value.

[0064] A population of individuals is represented as where . represents the offloading location selection of an individual, 0, 1, 2 represent local, edge and cloud respectively. represents the data distribution ratio over multiple wireless networks.

[0065] The computation of individual fitness relies on a fitness function, denoted as , where is a penalty coefficient that balances the objective optimization and the degree of constraint satisfaction by scaling the size of the penalty for violating constraints, as the penalty function of a chromosome that violates the privacy entropy constraint, is defined as , When the privacy entropy condition of the fuzzy chance constraint is satisfied, the penalty function returns a value of zero, imposing no penalty on the chromosome. Otherwise, the fitness of the chromosome is gradually reduced as the fuzzy chance constraint value decreases.

[0066] In the process of computing the fitness value, Monte Carlo sampling is combined with fuzzy simulation to obtain accurate chance constraint evaluation values. The proposed double simulation method minimizes the systematic error achieved by traditional single methods, while enhancing the consistency of the results of the uncertainty spectrum.

[0067] To calculate the fuzzy chance constraint value, the initialization process needs to assign a preliminary probability estimate to the fuzzy chance constraint. It can start from a conservative (deliberately as small as possible) initial estimate. In the iterative process, fuzzy-guided Monte Carlo sampling generates input parameters that reflect the system's uncertainty. Under simulated conditions, the membership function is used to calculate the probability of the fuzzy chance constraint.

[0068] Although the Monte Carlo-based fuzzy simulation effectively solves the problem of calculating the fuzzy chance constraint, it requires independent calculations for each sample and a large number of iterations, which leads to a daunting computational cost. To overcome this efficiency bottleneck, a neural network is integrated to estimate the fuzzy chance constraint value. In order to enable the trained neural network to accurately approximate the target function value and the privacy entropy constraint, a training data set is first generated through Monte Carlo-based fuzzy simulation. This data set serves as the basis for supervised learning, ensuring that the neural network captures the underlying input-output relationship of the fuzzy objective function and the privacy entropy constraint. The trained neural network is integrated into the iterative process of the genetic algorithm, replacing the original Monte Carlo-based fuzzy simulation process. By providing real-time estimates of the fuzzy chance constraint value, the optimization speed is significantly accelerated while maintaining the quality of the solution.

[0069] Subsequently, the crossover mutation of individual gene segments uses an adaptive strategy, which is expressed as follows: .

[0070] Fmaxrepresents the maximum fitness value in the population, Fmeanrepresents the average fitness value of the population in each generation, Fmaxrepresents the larger fitness value of the two individuals to be crossed, Frepresents the fitness value of the individual to be mutated, and Frandrepresents the random number between 0 and 1, respectively, where , , and are four constants. In order to avoid the rapid convergence of the population to a local optimal solution in the early stage of evolution, the crossover and mutation probabilities of the better individual are not tending to a minimum value after calculation by the above formula, but are raised to and , which can ensure that the population will not mature prematurely to a certain extent.

[0071] Since each chromosome encodes the offloading decision and allocation ratio of tasks, applying standard crossover and mutation operations to the entire chromosome will significantly impair the exploration ability of the offspring population in the solution space. Therefore, we simultaneously perform crossover and mutation operations on all gene segments. The system model imposes a constraint that the sum of all task allocation ratios must equal 1. However, after random crossover and mutation operations, the resulting child chromosomes may violate this constraint, thereby hindering the generation of feasible solutions within the search space. To restore feasibility, all allocation ratios are scaled proportionally to normalize and equalize them to 1 while maintaining their original relative proportions.

[0072] Individual offloading strategy is a single discrete value, which fundamentally limits the effectiveness of traditional genetic crossover. To avoid the degeneration of crossover into a value exchange that disrupts the pattern, a greedy-based genetic operation is proposed to replace the traditional crossover operation, with the following specific steps: Step 1: Sequentially modify the offloading position of the individual using a greedy strategy Step 2: Recalculate and compare the fitness values under different offloading positions Step 3: Only when the fitness value is improved, the modification will be retained; otherwise, the original position will be retained.

[0073] While greedy genetic operations effectively accelerate convergence, their strong goal-driven nature can lead to chromosomes getting trapped in local optima. Adaptive mutation is employed as a complementary mechanism, randomly perturbing the offspring position with probability after genetic operations, thus achieving a balance between exploitation and exploration.

[0074] After determining the task offloading strategy and network allocation strategy, the proportion of tasks to be processed in the offloading location is determined according to the offloading strategy, and the proportion of task data transmitted simultaneously on each network is determined according to the network allocation strategy.

[0075] In this embodiment of the application, the parameter settings of the multi-access edge computing system are shown in Table 1: Table 1 Parameter Table of Multi-Access Edge Computing System

[0076] like Figures 3 to 13 As shown, the task unloading method provided in this application is used to perform task unloading, and the results and effects are as follows.

[0077] like Figure 3 As shown, in order to evaluate the impact of the changes in computing power at each proposed unloading location on the objective function during the unloading process, the changes in the objective value and the task unloading distribution at different unloading locations under different computing power are studied by adjusting the percentage of computing power at each unloading location. Figure 3 This demonstrates the impact of changes in computing power on the target value. Figure 4 , 5 Figures 6 and 6 respectively demonstrate the task offloading and allocation of local terminals, edge servers, and cloud computing centers under conditions of changing computing power.

[0078] Figure 3 The change in computing power at each location corresponds to a broken line, showing how the target value of the local terminal changes with computing power ( Local computing power Figure 3 The local terminal (referred to as the "local terminal") gradually increases and then monotonically decreases. And... (Edge server computing power, Figure 3 (referred to as edge end) and (Computing power of cloud computing centers) Figure 3 When the computing power (hereinafter referred to as cloud computing power) reaches the 20% to 100% stage, the target values ​​of edge servers and cloud computing centers show a trend of first decreasing and then slowly rising to a stable level. This phenomenon is because when computing power is too low, higher latency accounts for the majority of the target value; while when computing power exceeds an appropriate level, the energy consumption caused by high computing power also increases, thus affecting the target value. In the stage exceeding 100% (i.e., basic computing power), the target values ​​of edge servers and cloud servers show slight fluctuations, reflecting the complex relationship between latency and energy consumption.Figure 4 As computing power increases, local terminals are able to complete tasks with lower target values, thus gradually acquiring more tasks. The task ratio increased from an initial 16% to 60%, and even when computing power reached 140%-160%, all tasks remained executed locally. Clearly, the increase in computing power brought significant benefits to the local machine. Figure 5 In the process, the number of tasks on edge servers showed a trend of first increasing and then decreasing, while the number of tasks on local terminals fluctuated slightly. At 40% computing power, edge servers could handle nearly 60% of the tasks, while the number of tasks in the cloud computing center accounted for only 20%, mainly due to privacy and security constraints and computing power limitations. As computing power continued to increase, the energy consumption of edge servers increased faster than their latency reduction, leading to a rise in target values. This caused tasks to tend to be processed on local terminals and in the cloud computing center. Ultimately, when computing power reached 160%, only 3% of tasks were offloaded to edge servers. Figure 6 It exhibits a similar trend to edge servers, which will not be elaborated here.

[0079] It can be seen that continuously increasing computing power has a positive effect because local computing eliminates the overhead of the unloading process. However, for edge servers and cloud computing centers, maintaining computing power at an appropriate level is crucial to optimize task unloading efficiency and system target values.

[0080] To evaluate the multi-network transmission communication model during the unloading process. Figure 7 The impact of changes in the number of network access points on the target value was compared, among which... Figure 8 and 9 They were shown respectively Figure 7 The details at different stages are as follows. It can be observed that as the number of tasks increases, the target value also rises, a trend consistent with real-world application scenarios. When the number of tasks is low, the system load is small, and the target value changes relatively smoothly. However, as the number of tasks increases, the number of tasks the system needs to process surges, and the target value inevitably rises.

[0081] It is worth noting that, in Figure 8In the phase of 10 to 40 tasks, the target value does not follow the monotonic decreasing rule of the number of access networks, and the following sequence appears: Net=3<Net=4<Net=5<Net=2. This indicates that, under the same number of tasks, the increase in the number of access networks does not always bring a better target value. This non-monotonic phenomenon can be attributed to multiple factors, such as the differences between networks, the privacy and security requirements, and the trade-off between energy consumption and latency, etc. In the experimental design, the bandwidth of multiple networks fluctuates, leading to unstable channel quality between different networks. When the number of tasks is small, although the number of network access is increased, the unstable fluctuations in bandwidth quality can lead to an increase in energy consumption. In addition, in a fluctuating network environment, maintaining the security of data privacy becomes more complex, which can lead to more tasks being selected for local computing, thereby increasing the system burden and additional overhead. Although theoretically, multiple network access helps to share the task load, too many network accesses can lead to higher energy consumption, especially in the case of poor channel quality, the additional overhead can exceed the benefits brought by transmission latency.

[0082] As shown in Figure 9 , with the further increase in the number of tasks (60-90 phase), the relationship between the target values of different network numbers tends to be stable, indicating that under a certain task size, more network access means smaller transmission overhead. At this time, the number of access networks starts to show a positive effect on the target value, showing that in a high-load environment, a reasonable number of networks can effectively reduce the system burden.

[0083] To analyze the influence of the degree of privacy and security requirements on the system target value and task offloading decision, the privacy requirements are changed to observe the influence on the above indicators under different numbers of tasks. As shown in Figure 10 , in Figure 10 , the influence of low, medium, and high privacy and security requirements on the system target value is shown, Figure 11 shows the influence of different privacy and security requirements on the proportion of task offloading at the terminal, edge server, and cloud service center.

[0084] With the increase in the number of tasks, the system target value under different privacy requirements shows an overall increasing trend. This indicates that under the same number of tasks, the higher the privacy and security requirements, the higher the system target value. This is because higher privacy requirements will limit more task offloading to the cloud server, causing tasks to remain in local computing, thereby increasing the overall overhead of the system. At the same time, since the cloud server is far away, it is more likely to cause privacy leakage during transmission, so stricter privacy requirements prompt the system to reduce its dependence on the cloud server and increase the proportion of local intelligent terminal processing tasks.

[0085] When the privacy requirement is low, most tasks can meet the transmission privacy requirement, and tend to be offloaded to edge servers and cloud servers with stronger computing power. Since edge servers have a distance advantage compared to cloud service centers, tasks are mainly offloaded to edge servers. When the privacy requirement is improved, the proportion of tasks left in the local terminal increases significantly, even exceeding the number of tasks offloaded to cloud service centers. With the increase in the number of tasks, the proportion of tasks offloaded to cloud service centers gradually decreases, reflecting that the increasing emphasis on privacy security increases the demand for local computing. In the context of high privacy requirements, almost all tasks do not choose to be offloaded to cloud service centers, and edge servers may also face the risk of attacks. In this case, most tasks choose to stay in the local for computing, although this will result in longer delays and higher system objectives, but it meets the requirements of privacy security.

[0086] It can be seen that the improvement of the privacy requirement significantly affects the task offloading proportion and the system objective value. When the privacy requirement is low, tasks tend to be offloaded to servers with stronger computing power, while high privacy requirements promote more tasks to stay in local computing, resulting in an increase in the system objective value.

[0087] Figure 12 、 13 The performance comparison diagram of using the task offloading method provided by the embodiments of the present application and other technologies under different iteration numbers is provided. The algorithm (HTOA-IMN) proposed in the present application is compared and evaluated with existing new benchmark algorithms: (1) task offloading algorithm based on whale optimization algorithm (TOA-WOA), (2) task offloading algorithm based on bat algorithm (TOA-BA), (3) task offloading algorithm based on genetic algorithm (TOA-GA), (4) task offloading algorithm based on simulated annealing algorithm (TOA-SA), and the results are shown in Figures 12-13 .

[0088] With the increase of the number of iterations, the HTOA-IMN algorithm exhibits relatively rapid convergence, and it begins to stabilize at about 40 iterations, with the target value being significantly superior to other algorithms. This indicates that the HTOA-IMN algorithm can effectively search the solution space and quickly find a better solution during optimization. In comparison, the convergence speed of the TOA-SA and TOA-WOA algorithms also performs well, both reaching the convergence state at 60 iterations. However, although their convergence speed is faster, their final target value is still lower than that of HTOA-IMN, showing that HTOA-IMN performs better in specific problem contexts. This may be related to the design and mechanism of the HTOA-IMN algorithm, making it more advantageous in dealing with complex optimization tasks. The TOA-GA and TOA-BA algorithms show positive trends during the convergence process, but their convergence speed is relatively slow and they fail to reach the performance level of other algorithms within the same number of iterations. This indicates that although TOA-GA and TOA-BA can find feasible solutions, their optimization efficiency is lower and they may need more iterations to achieve the ideal target value under certain parameter settings. In summary, the convergence characteristics of each algorithm differ significantly, with the HTOA-IMN algorithm having a significant advantage in optimization performance, while TOA-SA and TOA-WOA, although having certain convergence capabilities, still cannot match HTOA-IMN.

[0089] As can be seen from Figures 12-13 the target value of the algorithm HTOA-IMN proposed in the present application is the smallest, and the convergence speed is also faster, and the performance is better than other algorithms existing in the prior art.

[0090] Although the present application provides method operation steps as described in the embodiments or flowcharts, more or fewer operation steps can be included based on routine or non-creative labor. The order of steps listed in the embodiments is only one of the many step execution orders, and does not represent the only execution order. In actual device or client product execution, the method order shown in the embodiments or the drawings can be executed in sequence or in parallel (for example, in a parallel processor or multi-thread processing environment).

[0091] The various embodiments in the specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. The entire or part of the present application can be used in a variety of general or special computer system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, mobile communication terminals, multi-processor systems, microprocessor-based systems, programmable electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like.

[0092] While the preferred embodiments of the application have been described, additional variations and modifications can be made to these embodiments by those skilled in the art once they have the benefit of the present disclosure without departing from the spirit and scope of the application. Accordingly, it is intended that the appended claims include all such modifications and variations as fall within the scope of the present application.

[0093] It is apparent that those skilled in the art can make various changes and modifications to the application without departing from the spirit and scope of the application. Thus, it is intended that the application cover all modifications and variations of this application provided they come within the scope of the appended claims and their equivalents.

Claims

1. A method for offloading MEC tasks based on opportunity constraints, characterized in that, include: A latency-energy consumption model, a fuzzy privacy model, and a chance-constraint model are constructed by using the fuzzy data input amount and the number of fuzzy CPU cycles of the fuzzy task to be processed. A joint optimization model is established using the aforementioned time-delay energy consumption model, fuzzy privacy model, and chance-constraint model. Solving the joint optimization model yields the task offloading strategy and the network selection strategy; The fuzzy pending tasks are assigned to the uninstallation location according to the task uninstallation strategy; Resource transfer is performed according to the network selection strategy.

2. The MEC task offloading method based on opportunity constraints according to claim 1, characterized in that, The unloading locations include terminals, edge servers, and cloud computing centers, with the cloud computing centers equipped with multiple cloud servers.

3. The MEC task offloading method based on opportunity constraints according to claim 2, characterized in that, The latency and energy consumption model is used to obtain the latency and energy consumption cost of the task to be processed during data transmission and processing at the offloading location. The latency and energy consumption model enables data transmission between the terminal and the edge server, as well as between the edge server and the cloud server, through multiple wireless networks simultaneously.

4. The MEC task offloading method based on opportunity constraints according to claim 1, characterized in that, The fuzzy privacy model includes fuzzy privacy level requirements for the fuzzy task to be processed and data privacy entropy quantization. The fuzzy privacy level requirements are used to constrain the degree of privacy protection of the fuzzy task to be processed during transmission, and the data privacy entropy quantization is used to obtain the probability and level of data leakage of the fuzzy task to be processed during data transmission.

5. The MEC task offloading method based on opportunity constraints according to claim 1, characterized in that, The opportunity constraint model is used for the conversion between the determination of the amount of fuzzy task data and the number of CPU cycles, and the determination of the opportunity constraint form of the fuzzy privacy model.

6. The MEC task offloading method based on opportunity constraints according to claim 1, characterized in that, The latency and energy consumption model includes a communication model and a computation model. The communication model is used to obtain the latency cost of the fuzzy task to be processed, and the computation model obtains the energy consumption cost of the task to be processed.

7. The MEC task offloading method based on opportunity constraints according to claim 1, characterized in that, Solving the joint optimization model yields the task offloading strategy and network selection strategy, including: The joint optimization model is solved using a hybrid unloading algorithm combining improved adaptive genetics, Monte Carlo simulation, and neural networks.

Citation Information

Patent Citations

  • Task unloading method for uncertain MEC environment

    CN118509920A

  • MEC task unloading method and device combining resource allocation and privacy protection

    CN118921700A

  • Reinforcement learning-based edge computing privacy task unloading method, system and equipment

    CN120469733A