A multi-slot based mobile edge computing method and system

Abstract

The application discloses a kind of based on multi-slot mobile edge computing method and system, task buffer frame stage includes: according to the distribution of base station area task request present time buffer decision;Wireless device k0 according to present time buffer decision carries out task unloading, after wireless device k0 unloads task, the result of the task is calculated, and the calculation result is stored in mobile edge computing server;Task arrives and task computing frame stage includes: base station starts to receive the task unloaded by other wireless devices under coverage, and mobile edge computing server calculates the task unloaded;Combined with the processing of corresponding wireless device local processing and task unloading, the calculation result is returned to corresponding wireless device.The application carries out calculation task buffer, task scheduling and resource configuration with base station and wireless device weighted energy consumption as target, obtains optimal task allocation, to strengthen the calculation, storage and processing of wireless access network and the like, to realize the minimization of the sum of weighted energy consumption.

Description

Technical Field

[0001] This invention relates to the technical field of mobile edge computing and caching, and in particular to a mobile edge computing method and system based on multi-time slots. Background Technology

[0002] As a key technology in the evolution of fifth-generation mobile communication systems, mobile edge computing task caching technology has emerged and received widespread attention from industry and academia. Mobile edge computing deploys computing servers at the network edge (such as cellular base stations, WiFi access points, and gateway interfaces) close to user devices, enabling the processing of end-user computing tasks at the network edge and providing cloud computing-like services. Simultaneously, mobile edge caching technology can temporarily store popular videos, applications, or computation results on mobile edge computing servers, thus saving these servers from repetitive computations and bringing users advantages such as low power consumption and low latency.

[0003] In a mobile edge computing system, a wireless base station deploying a computing server provides computing and communication services to multiple wireless devices within its coverage area. Assume each wireless device has a series of computing tasks within a given time period; different tasks have different computational requirements, and different wireless devices can perform the same tasks. At the beginning of a certain time period, the input data for these computing tasks from the multiple wireless devices is offloaded to the server. The mobile edge computing server then runs the computing tasks and returns the results directly to the wireless devices.

[0004] However, mobile edge computing servers have very limited computing and storage resources. In scenarios where a large number of wireless devices simultaneously need to request the mobile edge computing server to complete task computing, the mobile edge computing server cannot respond to all the requests from wireless devices containing a large number of repetitive task computing in a short period of time, thereby affecting the computing service quality of end users. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a mobile edge computing method based on multiple time slots.

[0006] To achieve the above objectives, the technical solution provided by this invention is as follows:

[0007] A mobile edge computing method based on multiple time slots includes a task buffer frame stage and a task arrival and task computing frame stage;

[0008] in,

[0009] The task buffer frame stage includes:

[0010] The mobile edge computing server makes current caching decisions based on the distribution of task requests in the area where the base station is located;

[0011] Wireless device k0 offloads the task based on the current cache decision. After wireless device k0 offloads the task, the mobile edge computing server starts to calculate the result of the task and stores the calculation result on the mobile edge computing server. Wireless device k0 is the wireless device with the minimum selected path loss.

[0012] The mission arrival and mission calculation frame phase includes:

[0013] The base station begins receiving tasks offloaded by other wireless devices within its coverage area. The mobile edge computing server then calculates the offloaded tasks. If the mobile edge computing server has cached the corresponding calculation results, the task calculation process is skipped, and the calculation results are directly returned to the corresponding wireless device. If the task of the corresponding wireless device is not cached, task scheduling and allocation are performed, and the corresponding wireless device's local processing and task offloading are combined for processing. After the calculation is completed, the calculation results are returned to the corresponding wireless device.

[0014] Furthermore, the mobile edge computing server makes current caching decisions based on the distribution of task requests in the area where the base station is located, including:

[0015] For scenarios where task request distribution changes significantly over a large time scale but not significantly over a short period, the branch and bound method is used to solve the mobile edge computing optimization model based on the task request distribution data of the previous time period to obtain the optimal caching decision of the previous time period, and this optimal caching decision of the previous time period is used as the current caching decision.

[0016] For scenarios where the distribution of request task types does not change significantly over a large time scale, the most frequently used tasks will be cached as the current caching decision.

[0017] Furthermore, the task scheduling and allocation includes:

[0018] For tasks with specific operating protocols, the upcoming tasks and their arrival times are predictable. Solve the mobile edge computing optimization model and allocate tasks based on the solution results.

[0019] For tasks that are sequential and about to arrive, the arrival time is predictable, only the arrival time is unknown. Assuming that the remaining tasks will arrive in the last few time slots, the mobile edge computing optimization model is solved to obtain the number of bits for local processing and offloading. After receiving the task from the wireless device, the mobile edge computing server processes the task at the maximum CPU frequency to avoid the subsequent arrival of tasks having too many bits, which would lead to increased energy consumption.

[0020] Furthermore, the mobile edge computing optimization model considers the limitations of cache space, the relationship between task offloading and processing, and the time constraint to minimize the weighted energy consumption of the system. Its objective function is:

[0021]

[0022] Where w0 and w1 are the weighting coefficients for the mobile edge computing server and the wireless device, respectively, E mec This refers to the energy consumption of mobile edge computing servers during task arrival and task computation frame stages. This refers to the computational energy consumption of a mobile edge computing server during the task caching frame stage. E represents the energy consumption value for wireless device k0 to unload the computation task during the task buffer frame phase. k The energy consumption of wireless device k in locally processing and unloading computing tasks during the task arrival and task computing frame stages.

[0023] The constraints are as follows:

[0024]

[0025]

[0026]

[0027]

[0028]

[0029]

[0030]

[0031]

[0032] Equations (1)-(4) are constraints in the task buffer frame stage;

[0033] Equation (1) indicates that the space occupied by the cached tasks needs to be less than the maximum cache space of the mobile edge computing server; α l ∈{0,1} represents the caching decision of the mobile edge computing server, if α l =1 indicates that the mobile edge computing server has cached task l; otherwise, α l =0;

[0034] Equation (2) indicates that during the task buffer frame stage, from the first time slot to the Nth time slot... p The total number of bits of tasks offloaded in slot -1 is equal to the space occupied by the cached tasks; N represents the number of task bits that wireless device k0 offloads during time slot n in the task buffer frame period. p This represents the number of time slots in the task's buffer frame phase.

[0035] Equation (3) indicates that the number of tasks that a mobile edge computing server can handle cannot exceed the number of tasks that have already been unloaded at the current time; Let i = 1,...,N be the computational load of time slot j in the task buffering frame stage of the mobile edge computing server. p ;

[0036] Equation (4) indicates that the mobile edge computing server must complete all tasks that the wireless device k0 must unload during the task buffer frame phase;

[0037] Equations (5)-(8) are constraints during the task arrival and task calculation frame stages;

[0038] Equation (5) indicates that the number of bits processed and offloaded to the mobile edge computing server by the wireless device k during time slots 1 to n in the task arrival and task computation frame stages must be less than or equal to the task scheduling; 1 A is the indicator function; N is the number of time slots in the task arrival and task calculation frame stages; and These represent the number of bits processed and offloaded to the base station by the wireless device k in the j-th time slot during the task buffer frame stage, respectively.

[0039] Equation (6) indicates that wireless device k must complete the task scheduling through local computation or task unloading before the task arrives and the computation frame ends; k = 1, ..., K, where K is the total number of wireless devices under the base station coverage area;

[0040] Equation (7) indicates that the number of bits processed by the mobile edge computing server within time slots 1-n during the task arrival and task calculation frame phase must be less than or equal to the amount of task data that the mobile edge computing server has offloaded from all wireless devices before the nth time slot. This represents the number of bits processed by the edge computing server during the j-th time slot.

[0041] Equation (8) indicates that the mobile edge computing server needs to complete all unloading tasks during the task arrival and task calculation frame stages.

[0042] Furthermore, during the task buffer frame stage,

[0043] The energy consumption of wireless device k0 offloading computing tasks is calculated using the following formula:

[0044]

[0045] Where, σ 2The additive white Gaussian noise at the base station receiver end. Let k0 be the channel coefficient between wireless device k0 and base station during the i-th time slot. During the i-th time slot, k0 represents the system bandwidth used by the wireless device for the offloading task. This represents the number of task bits that wireless device k0 offloads during the i-th time slot.

[0046] Set the CPU speed of the mobile edge computing server in the i-th time slot to be... At this point, the energy consumption expression for the mobile edge computing server is:

[0047]

[0048] Where ζ0 is a coefficient determined by the CPU hardware architecture of the mobile edge computing server, and C0 is the number of CPU cycles required to process one bit task.

[0049] Furthermore, during the task arrival and task calculation frame phase,

[0050] The energy consumption of wireless device k in processing tasks locally and offloading computing tasks is denoted as E. k ,satisfy in This refers to the energy consumption of wireless device k during the task arrival and task calculation frame phase when it locally processes the task. The energy consumption of wireless device k during data transmission in the task arrival and task computation frame phase is modeled as follows:

[0051]

[0052]

[0053] Energy consumption of mobile edge computing servers for processing tasks

[0054] To achieve the above objectives, the present invention further provides a multi-timeslot-based mobile edge computing system for implementing the above-mentioned multi-timeslot-based mobile edge computing method, comprising: a base station equipped with a mobile edge computing server and a wireless device;

[0055] in,

[0056] A base station equipped with a mobile edge computing server includes an information acquisition module, an optimization problem solving module, a first communication module, a first task processing module, and a calculation result caching module;

[0057] The information acquisition module is used to acquire task request information initiated by the wireless device;

[0058] The optimization problem-solving module is used to solve the above optimization problem in the mobile edge computing server based on the obtained task request information, obtain the optimal task scheduling decision for the wireless device, and guide the wireless device to divide the computing task.

[0059] The first communication module is used for information exchange between the mobile edge computing server and the wireless device, including sending task scheduling decision information from the optimization problem module to the wireless device, sending task calculation results, and receiving tasks that are offloaded by the wireless device to the access point.

[0060] The first task processing module is used to process computing tasks offloaded from wireless devices.

[0061] The calculation result caching module is used to store calculation results in the mobile edge computing server;

[0062] The wireless device includes a second communication module, a task segmentation module, a second task processing module, and a calculation result output module;

[0063] The second communication module is used for the wireless device to receive task processing coordination information from the mobile edge computing server, receive calculation results from the mobile edge computing server, and unload task data according to the coordination information from the mobile edge computing server.

[0064] The task splitting module is used to split a task into two sub-tasks based on coordination information from the mobile edge computing server. The first type of sub-task is processed locally, and the second type of sub-task is offloaded to the edge server.

[0065] The second task processing module is used for wireless devices to process the first type of sub-tasks;

[0066] The calculation result output module is used by the wireless device to output the calculation results of the task.

[0067] Compared with existing technologies, the principles and advantages of this solution are as follows:

[0068] This solution targets the weighted energy consumption of base stations and wireless devices by performing computational task caching, task scheduling, and resource allocation to obtain optimal task allocation. This enhances the computation, storage, and processing capabilities of the wireless access network, thereby minimizing the sum of weighted latency and energy consumption. Attached Figure Description

[0069] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the services required in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0070] Figure 1 A schematic diagram of a mobile edge caching computing system;

[0071] Figure 2 This is a schematic diagram of the system operation protocol;

[0072] Figure 3 This is a flowchart illustrating the principle of a multi-timeslot-based mobile edge computing method according to the present invention. Detailed Implementation

[0073] The present invention will be further described below with reference to specific embodiments:

[0074] The scenarios considered in this embodiment are as follows: Figure 1 As shown below, the system structure and operation are explained:

[0075] Assume there are a total of L low-latency computing tasks, and the input data size of the computing task with index 1 is D. l ,and The number of input bits for task l is D l The computational tasks of K wireless devices can be offloaded to the base station via the wireless channel. Let f be the CPU speed of the computation task of the k-th wireless device, which is a set of wireless devices. k After a task arrives at the wireless device, in order to obtain the computation result, the wireless device performs local processing through its device processing module, while simultaneously offloading part of the task to the mobile edge computing server for computation. The mobile edge computing server receives the computation task from the wireless device and performs the task computation using its processing module. Let f be the CPU speed of the mobile edge computing server. max Let D be the computing cache capacity of the mobile edge computing server. max In scenarios where the computational task can be decomposed, the wireless device first decomposes the task, performs a portion of the computation locally, and transmits the remaining portion via a wireless communication link to the base station's server for processing. If the mobile edge computing server caches the wireless device's computational task, the wireless device only needs to download the computation results directly.

[0076] like Figure 3 As shown in the figure, the mobile edge computing method based on multi-time slots described in this embodiment includes a task buffer frame stage and a task arrival and task computing frame stage, and its application is as follows: Figure 2 The protocol shown;

[0077] in,

[0078] The task buffer frame stage includes:

[0079] S1. The mobile edge computing server determines the current caching decision based on the distribution of task requests in the area where the base station is located. The process includes:

[0080] For scenarios where the distribution of task requests changes significantly over a large time scale but not much over a short period, the branch and bound method is used to solve the mobile edge computing optimization model based on the task request distribution data of the previous time period to obtain the optimal caching decision of the previous time period, and this optimal caching decision of the previous time period is used as the current caching decision.

[0081] For scenarios where the distribution of request task types does not change significantly over a large time scale, the most frequently used tasks will be cached as the current caching decision.

[0082] S2. Wireless device k0 offloads the task based on the current cache decision. After wireless device k0 offloads the task, the mobile edge computing server starts to calculate the result of the task and stores the calculation result in the mobile edge computing server. Wireless device k0 is the wireless device with the minimum selected path loss.

[0083] The mission arrival and mission calculation frame phase includes:

[0084] S3. The base station begins receiving tasks offloaded by other wireless devices within its coverage area. The mobile edge computing server calculates the offloaded tasks. If the mobile edge computing server has cached the corresponding calculation results, the task calculation process is skipped, and the calculation results are directly returned to the corresponding wireless device. If the task of the corresponding wireless device is not cached, task scheduling and allocation are performed, and the corresponding wireless device's local processing and task offloading are combined for processing. After the calculation is completed, the calculation results are returned to the corresponding wireless device.

[0085] Specifically, regarding task scheduling, in real-world scenarios, at the start of time slot n, the edge server needs to predict the task distribution for time slot n+1 and subsequent time slots. Therefore, how to schedule computing resources between the mobile edge computing server and wireless devices to minimize system energy consumption is a challenge. To address this, this embodiment proposes different solutions based on the characteristics of the tasks.

[0086] For tasks with specific operating protocols, such as precision tasks, where the arrival time of the task is predictable, the mobile edge computing optimization model is solved, and task scheduling and allocation are performed based on the solution results.

[0087] For tasks with a sequential relationship (such as video streams or pipelines) that are about to arrive, the arrival time is predictable, only the arrival time is unknown. It is assumed that the remaining tasks will arrive in sequence in the last few time slots. Then, the mobile edge computing optimization model is solved to obtain the number of bits for local processing and offloading. After receiving the task from the wireless device, the mobile edge computing server processes the task at the maximum CPU frequency to avoid the subsequent arrival of tasks having too many bits, which would lead to increased energy consumption.

[0088] In the above, the mobile edge computing optimization model considers the constraints of cache space limitations, the relationship between task offloading and processing, and the time limit to minimize the weighted energy consumption of the system. Its objective function is:

[0089]

[0090] Where w0 and w1 are the weighting coefficients for the mobile edge computing server and the wireless device, respectively, E mec This refers to the energy consumption of mobile edge computing servers during task arrival and task computation frame stages. This refers to the computational energy consumption of a mobile edge computing server during the task caching frame stage. E represents the energy consumption value for wireless device k0 to unload the computation task during the task buffer frame phase. k The energy consumption of wireless device k in locally processing and unloading computing tasks during the task arrival and task computing frame stages.

[0091] The constraints are as follows:

[0092]

[0093]

[0094]

[0095]

[0096]

[0097]

[0098]

[0099]

[0100] Equations (1)-(4) are constraints in the task buffer frame stage;

[0101] Equation (1) indicates that the space occupied by the cached tasks needs to be less than the maximum cache space of the mobile edge computing server; α l∈{0,1} represents the caching decision of the mobile edge computing server, if α l =1 indicates that the mobile edge computing server has cached task l; otherwise, α l =0;

[0102] Equation (2) indicates that during the task buffer frame stage, from the first time slot to the Nth time slot... p The total number of bits of tasks offloaded in slot -1 is equal to the space occupied by the cached tasks; N represents the number of task bits that wireless device k0 offloads during time slot n in the task buffer frame period. p Let N be the number of time slots in the task buffer frame stage. Since data transmission takes time, after wireless device k0 unloads the task in time slot n, the mobile edge computing server can only start processing in time slot n+1. Therefore, the left-hand side of the above constraints is accumulated to N. p -1 instead of N p .

[0103] Equation (3) indicates that the number of tasks that a mobile edge computing server can handle cannot exceed the number of tasks that have already been unloaded at the current time; Let i = 1,...,N be the computational load of time slot j in the task buffering frame stage of the mobile edge computing server. p ;

[0104] Equation (4) indicates that the mobile edge computing server must complete all tasks that the wireless device k0 must unload during the task buffer frame phase;

[0105] Equations (5)-(8) are the constraints during the task arrival and task calculation frame stages;

[0106] During the task arrival and task computation frame phase, the base station first receives the task from the wireless device. It is assumed that wireless device k will only receive tasks from the set [of data / processes] in each time slot. If a task is randomly requested in a frame, then the sequence of tasks requested in that frame is as follows: in Because the task request comes from the set The samples are randomly selected from the middle, therefore s k There may be duplicate elements in S. k,n For tasks requested by a wireless device in and before time slot n, the expression is: The task in the first time slot is generally not repeated, i.e., S. k,1 ={s k,1};S k,n Still satisfied Next, we will explain the task processing on the wireless device side, including local processing and task offloading.

[0107] Equation (5) indicates that the number of bits processed and offloaded to the mobile edge computing server by the wireless device k during time slots 1 to n in the task arrival and task computation frame stages must be less than or equal to the task scheduling; 1 A is the indicator function; N is the number of time slots in the task arrival and task calculation frame stages; and These represent the number of bits processed and offloaded to the base station by the wireless device k in the j-th time slot during the task buffer frame stage, respectively.

[0108] Equation (6) indicates that wireless device k must complete the task scheduling through local computation or task unloading before the task arrives and the computation frame ends; k = 1, ..., K, where K is the total number of wireless devices under the base station coverage area;

[0109] Equation (7) indicates that the number of bits processed by the mobile edge computing server within time slots 1-n during the task arrival and task calculation frame phase must be less than or equal to the amount of task data that the mobile edge computing server has offloaded from all wireless devices before the nth time slot. This represents the number of bits processed by the edge computing server during the j-th time slot.

[0110] Equation (8) indicates that the mobile edge computing server needs to complete all unloading tasks during the task arrival and task calculation frame stages.

[0111] Specifically, during the task buffer frame stage,

[0112] The energy consumption of wireless device k0 offloading computing tasks is calculated using the following formula:

[0113]

[0114] Where, σ 2 The additive white Gaussian noise at the base station receiver end. Let k0 be the channel coefficient between wireless device k0 and base station during the i-th time slot. During the i-th time slot, k0 represents the system bandwidth used by the wireless device for the offloading task. This represents the number of task bits that wireless device k0 offloads during the i-th time slot.

[0115] Set the CPU speed of the mobile edge computing server in the i-th time slot to be... At this point, the energy consumption expression for the mobile edge computing server is:

[0116]

[0117] Where ζ0 is a coefficient determined by the CPU hardware architecture of the mobile edge computing server, and C0 is the number of CPU cycles required to process one bit task.

[0118] Specifically, during the task arrival and task computation frame stage,

[0119] The energy consumption of wireless device k in processing tasks locally and offloading computing tasks is denoted as E. k ,satisfy in This refers to the energy consumption of wireless device k during the task arrival and task calculation frame phase when it locally processes the task. The energy consumption of wireless device k during data transmission in the task arrival and task computation frame phase is modeled as follows:

[0120]

[0121]

[0122] Energy consumption of mobile edge computing servers for processing tasks

[0123] Finally, this embodiment also includes a multi-timeslot-based mobile edge computing system for implementing the above-mentioned multi-timeslot-based mobile edge computing method, which includes: a base station equipped with a mobile edge computing server and wireless devices;

[0124] in,

[0125] A base station equipped with a mobile edge computing server includes an information acquisition module, an optimization problem solving module, a first communication module, a first task processing module, and a calculation result caching module;

[0126] The information acquisition module is used to acquire task request information initiated by the wireless device;

[0127] The optimization problem-solving module is used to solve the above optimization problem in the mobile edge computing server based on the obtained task request information, obtain the optimal task scheduling decision for the wireless device, and guide the wireless device to divide the computing task.

[0128] The first communication module is used for information exchange between the mobile edge computing server and the wireless device, including sending task scheduling decision information from the optimization problem module to the wireless device, sending task calculation results, and receiving tasks that are offloaded by the wireless device to the access point.

[0129] The first task processing module is used to process computing tasks offloaded from wireless devices.

[0130] The calculation result caching module is used to store calculation results in the mobile edge computing server;

[0131] The wireless device includes a second communication module, a task segmentation module, a second task processing module, and a calculation result output module;

[0132] The second communication module is used for the wireless device to receive task processing coordination information from the mobile edge computing server, receive calculation results from the mobile edge computing server, and unload task data according to the coordination information from the mobile edge computing server.

[0133] The task splitting module is used to split a task into two sub-tasks based on coordination information from the mobile edge computing server. The first type of sub-task is processed locally, and the second type of sub-task is offloaded to the edge server.

[0134] The second task processing module is used for wireless devices to process the first type of sub-tasks;

[0135] The calculation result output module is used by the wireless device to output the calculation results of the task.

[0136] The above-described embodiments are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Therefore, any changes made in accordance with the shape and principle of the present invention should be covered within the protection scope of the present invention.

Claims

1. A mobile edge computing method based on multi-time slots, characterized in that, This includes the task buffer frame stage and the task arrival and task calculation frame stage; in, The task buffer frame stage includes: The mobile edge computing server makes current caching decisions based on the distribution of task requests in the area where the base station is located; wireless devices Task offloading is performed based on current cache decisions on wireless devices. After unloading the task, the mobile edge computing server begins calculating the result of the task and stores the result on the mobile edge computing server; wireless devices The wireless device with the least path loss among the selected paths; The mission arrival and mission calculation frame phase includes: The base station begins receiving tasks offloaded by other wireless devices within its coverage area. The mobile edge computing server then calculates the offloaded tasks. If the mobile edge computing server has cached the corresponding calculation results, the task calculation process is skipped, and the calculation results are directly returned to the corresponding wireless device. If the task of the corresponding wireless device is not cached, task scheduling is performed, and the tasks are allocated to the wireless device and the mobile edge computing server respectively. The tasks are processed by combining the local processing of the corresponding wireless device and task offloading. After the calculation is completed, the calculation results are returned to the corresponding wireless device. The mobile edge computing server makes current caching decisions based on the distribution of task requests in the area where the base station is located, including: For scenarios where the distribution of task requests changes significantly over a large time scale but not much over a short period, the branch and bound method is used to solve the mobile edge computing optimization model based on the task request distribution data of the previous time period to obtain the optimal caching decision of the previous time period, and this optimal caching decision of the previous time period is used as the current caching decision. For scenarios where the distribution of request task types does not change significantly over a large time scale, the most frequently used tasks will be cached as the current caching decision. The task scheduling and allocation includes: For tasks with specific operating protocols, the upcoming tasks and their arrival times are predictable. Solve the mobile edge computing optimization model and allocate tasks based on the solution results. For tasks that are sequential and about to arrive, the arrival time is predictable, only the arrival time is unknown. Assuming that the remaining tasks will arrive in sequence in the last few time slots, the mobile edge computing optimization model is solved to obtain the number of bits for local processing and offloading. After receiving the task from the wireless device, the mobile edge computing server processes the task at the maximum CPU frequency to avoid the subsequent arrival of tasks having too many bits, which would lead to increased energy consumption. The mobile edge computing optimization model considers the limitations of cache space, the relationship between task offloading and processing, and the time constraint to minimize the weighted energy consumption of the system. Its objective function is: ； in, and These are the weighting coefficients for mobile edge computing servers and wireless devices, respectively. This refers to the energy consumption of mobile edge computing servers during task arrival and task computation frame stages. This refers to the computational energy consumption of a mobile edge computing server during the task caching frame stage. For wireless devices The energy consumption value of unloading the computation task during the task buffer frame stage. For wireless devices Energy consumption for local processing and unloading of computing tasks during the task arrival and task computation frame phases; The constraints are as follows: （1） （2） （3） （4） （5） （6） （7） （8） Equations (1)-(4) are constraints in the task buffer frame stage; Equation (1) indicates that the space occupied by the cached tasks needs to be less than the maximum cache space of the mobile edge computing server; This indicates the caching decision of the mobile edge computing server, if This indicates that the mobile edge computing server has cached the task. ,otherwise ; For the serial number The size of the input data for the computation task; The computing cache capacity for the mobile edge computing server; To reduce the number of low-latency computing tasks; Equation (2) indicates that during the task buffer frame stage, from the first time slot to the second time slot... The total number of bits for tasks offloaded from a time slot is equal to the space occupied by the cached tasks; For wireless devices Time slots during the task cache frame period Number of bits of tasks unloaded at any time This represents the number of time slots in the task's buffer frame phase. Equation (3) indicates that the number of tasks that a mobile edge computing server can handle cannot exceed the number of tasks that have already been unloaded at the current time; For mobile edge computing servers, time slots in the task caching frame stage The computational load of the task ; Equation (4) indicates that the mobile edge computing server must complete the task buffering frame phase within the wireless device. All tasks were uninstalled; Equations (5)-(8) are the constraints during the task arrival and task calculation frame stages; Equation (5) represents the wireless device The number of bits processed and offloaded to the mobile edge computing server on the wireless device side within time slots 1 to n during the task arrival and task computation frame phase needs to be less than or equal to the task scheduling. is the indicator function; N is the number of time slots in the task arrival and task calculation frame stages; and They represent wireless devices. In the task buffer frame stage The number of bits processed and offloaded to the base station in each time slot at the wireless device end; For wireless device k in time slot and the tasks previously requested; Equation (6) represents the wireless device Task scheduling must be completed through local computation or task unloading before the task arrives and the computation frame ends. =1,…,K, where K is the total number of wireless devices within the base station's coverage area; Equation (7) indicates that the number of bits processed by the mobile edge computing server within time slots 1-n during the task arrival and task calculation frame phase must be less than or equal to the number of bits processed by the mobile edge computing server in the first time slot. The amount of task data that was offloaded from all wireless devices in the previous time slot; Indicates the first The number of bits processed by the mobile edge computing server per time slot; Equation (8) indicates that the mobile edge computing server needs to complete all unloading tasks during the task arrival and task calculation frame stages.

2. The mobile edge computing method based on multi-time slots according to claim 1, characterized in that, During the task buffer frame stage Computing wireless devices The energy consumption of unloading computing tasks is calculated using the following formula: ； in, The additive white Gaussian noise at the base station receiver end. In the first wireless devices during each time slot Channel coefficients between the base station and the base station For the first wireless devices during each time slot The system bandwidth used for unloading tasks. For the first wireless devices during each time slot Number of bits for unloaded tasks; Set the CPU speed of the mobile edge computing server in the i-th time slot to be... At this point, the energy consumption expression for the mobile edge computing server is: ； in, The coefficient is determined by the CPU hardware architecture of the mobile edge computing server. The number of CPU cycles required to process a one-bit task.

3. The mobile edge computing method based on multi-time slots according to claim 2, characterized in that, During the task arrival and task calculation frame phase wireless devices The energy consumption for processing tasks locally and unloading computing tasks is denoted as . ,satisfy ,in During the mission arrival and mission calculation frame phase, the wireless device Energy consumption for local processing tasks. During the mission arrival and mission calculation frame phase, the wireless device Energy consumption for data transmission Energy consumption of mobile edge computing servers processing tasks: 。 4. A mobile edge computing system based on multi-timeslots, characterized in that, The method for implementing the multi-timeslot-based mobile edge computing method according to any one of claims 1-3 includes: a base station equipped with a mobile edge computing server and a wireless device; in, A base station equipped with a mobile edge computing server includes an information acquisition module, an optimization problem solving module, a first communication module, a first task processing module, and a calculation result caching module; The information acquisition module is used to acquire task request information initiated by the wireless device; The optimization problem-solving module is used to solve the above optimization problem in the mobile edge computing server based on the obtained task request information, obtain the optimal task scheduling decision for the wireless device, and guide the wireless device to divide the computing task. The first communication module is used for information exchange between the mobile edge computing server and the wireless device, including sending task scheduling decision information from the optimization problem module to the wireless device, sending task calculation results, and receiving tasks that are offloaded by the wireless device to the access point. The first task processing module is used to process computing tasks offloaded from wireless devices. The calculation result caching module is used to store calculation results in the mobile edge computing server; The wireless device includes a second communication module, a task segmentation module, a second task processing module, and a calculation result output module; The second communication module is used for the wireless device to receive task processing coordination information from the mobile edge computing server, receive calculation results from the mobile edge computing server, and unload task data according to the coordination information from the mobile edge computing server. The task splitting module is used to split a task into two sub-tasks based on coordination information from the mobile edge computing server. The first type of sub-task is processed locally, and the second type of sub-task is offloaded to the edge server. The second task processing module is used for wireless devices to process the first type of sub-tasks; The calculation result output module is used by the wireless device to output the calculation results of the task.

Images