Relay power and edge computing resource optimization method and system based on cloud edge collaboration

Abstract

The invention discloses a relay power and edge computing resource optimization method and system based on cloud edge collaboration, and the method comprises the steps: constructing a cloud edge collaboration resource optimization model with the purpose of minimizing the weighted sum of the total task time delay and the total task energy consumption in a relay-assisted NOMA-MEC system scene; the relay-assisted NOMA-MEC system scene comprises a cloud control center, a plurality of relays, a plurality of power areas and a plurality of edge gateways for acquiring data from the corresponding power areas; the task total time delay comprises a first time slot, a second time slot, a third time slot and a calculation total time delay of the cloud control center; the task total energy consumption comprises task transmission total energy consumption and task calculation total energy consumption; and solving the cloud edge collaborative resource optimization model based on a joint solving algorithm of block coordinate descent to obtain an optimal resource allocation scheme. According to the invention, real-time reliable transmission and processing of mass power data can be realized, and stable operation of a smart power grid is guaranteed.

Description

A Method and System for Optimizing Relay Power and Edge Computing Resources Based on Cloud-Edge Collaboration Technical Field

[0001] This invention relates to a method and system for optimizing relay power and edge computing resources based on cloud-edge collaboration, belonging to the field of smart grid control. Background Technology

[0002] With the widespread integration of distributed energy resources and the development of information and communication technologies, various intelligent data acquisition and monitoring devices have been deployed extensively in the power grid, enabling the traditional power system to gradually transform into a smart grid. As a new generation of power systems, the smart grid provides the foundation for the "intelligent" operation of the power grid through real-time collection and analysis of massive amounts of grid data, enabling it to achieve unprecedented sensing, prediction, control, and self-healing capabilities. The smart grid not only improves the reliability and flexibility of power supply but also optimizes energy efficiency management on the user side, ultimately driving society towards a low-carbon, efficient, and sustainable energy future.

[0003] However, with the continuous development of smart grids, the contradiction between the real-time and reliable processing of massive amounts of power data and the rational allocation of limited information resources is becoming increasingly prominent. On the one hand, the safe and reliable operation of smart grids cannot be separated from the real-time transmission and analysis of massive amounts of diverse and heterogeneous power data; on the other hand, computationally intensive and latency-sensitive data processing tasks (such as processing images from UAV inspections and analyzing critical power data) are growing exponentially, placing higher demands on the real-time performance and reliability of smart grid data processing.

[0004] While existing cloud-based data processing models can improve data processing performance to some extent, several problems remain: 1) Existing point-to-point communication between the cloud and the edge suffers from channel gain limitations, resulting in significant data transmission latency and energy consumption, making it difficult to meet data transmission requirements; 2) Although cloud control centers are typically equipped with high-performance cloud servers, massive and complex data processing tasks exacerbate computational latency and energy consumption on the cloud side. Therefore, relying solely on cloud-based data processing models is insufficient to meet the current data processing needs of smart grids, potentially leading to the inability to transmit and process control-required data in real time, threatening the safe and stable operation of the smart grid.

[0005] To address the aforementioned issues, collaborative relay technology and mobile edge computing technology are considered feasible solutions. Collaborative relay technology, as a novel communication method, can improve the communication quality of point-to-point communication and has been used to solve the problem of massive data transmission. Furthermore, mobile edge computing technology, as a new parallel computing paradigm, can transfer computational tasks from the remote cloud to the network edge by deploying mobile edge computing servers with cloud-like functions at the network edge, thereby effectively reducing task processing time and energy consumption. In conclusion, researching how to combine these two technologies to ensure the real-time and reliable processing of smart grid data is crucial for improving the performance of information systems, meeting the diverse needs of power services, and ensuring the safe and stable operation of the system. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and system for optimizing relay power and edge computing resources based on cloud-edge collaboration. This method can realize the real-time and reliable transmission of power data in smart grids.

[0007] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0008] On the one hand, this invention provides a method for optimizing relay power and edge computing resources based on cloud-edge collaboration, including:

[0009] In the relay-assisted NOMA-MEC system scenario, a cloud-edge collaborative resource optimization model is constructed with the goal of minimizing the weighted sum of total task latency and total task energy consumption.

[0010] The relay-assisted NOMA-MEC system scenario includes a cloud control center, multiple relays, multiple power zones, and multiple edge gateways that collect data from the corresponding power zones;

[0011] The total task latency includes the first time slot, the second time slot, the third time slot, and the total computation latency of the cloud control center; the total task energy consumption includes the total energy consumption of task transmission and the total energy consumption of task computation.

[0012] The tasks include: several local tasks to be executed on the edge gateway, several relay tasks to be executed on the relay, and several cloud tasks to be executed on the cloud control center; wherein, the local tasks and relay tasks are all unloaded by the edge gateway, the cloud tasks are unloaded by the edge gateway, and the tasks are unloaded by the edge gateway to the relay and then unloaded by the relay.

[0013] The first time slot is the maximum transmission delay of several relay tasks offloaded from the edge gateway to the relay; the second time slot is the maximum of the transmission delay of several cloud tasks offloaded from the edge gateway to the relay and the transmission delay of several cloud tasks directly offloaded from the edge gateway to the cloud control center; the third time slot is the maximum transmission delay of several cloud tasks offloaded from the relay to the cloud control center.

[0014] A joint solution algorithm based on block coordinate descent is used to solve the cloud-edge collaborative resource optimization model and obtain the optimal resource allocation scheme.

[0015] Optionally, the cloud-edge collaborative resource optimization model includes an objective function and constraints; the constraints include: task allocation constraints, relay selection constraints, relay selection factor constraints, latency constraints, time slot allocation constraints, relay resource allocation constraints, and relay reliable decoding and forwarding constraints.

[0016] Optionally, the objective function is expressed as follows:

[0017] ;

[0018] in: It is a dimensional parameter used to unify the units of time delay and energy consumption; Represent the objective function; , All represent weighting coefficients; N represents the total number of edge gateways; n represents the number of sequences; Indicates the first time slot; Indicates the second time slot; Indicates the third time slot; This represents the computation latency of the cloud task offloaded by the nth edge gateway in the cloud control center. This represents the transmission energy consumption of the task offloaded by the nth edge gateway within period T; This represents the computing energy consumption of all tasks of the nth edge gateway within period T, including the computing energy consumption of local tasks, relay tasks, and cloud tasks of the nth edge gateway. This represents the total computational latency of the cloud control center; Indicates the total energy consumption of the task; This indicates the total energy consumption for task computation; Indicates the total energy consumption of task transmission; These are all decision variables of the objective function, used to determine the total task latency and total task energy consumption in the objective function, and are represented as follows:

[0019] ;

[0020] ;

[0021] ;

[0022] ;

[0023] ;

[0024] Where A represents the relay selection matrix; The relay selection factor indicates whether the nth edge gateway selects or does not select the mth relay for collaborative communication and task offloading; M represents the total number of relays; L represents the task matrix. , and These represent the local task, relay task, and cloud task of the nth edge gateway, respectively; P represents the relay power allocation matrix. represents the transmission power of the cloud task allocated by the m-th relay to the n-th edge gateway; f represents the relay computing resource matrix; This represents the computing resources allocated by the m-th relay to handle the relay task of the n-th edge gateway; t represents the time matrix; T represents the period, which is the upper limit of the total computational latency.

[0025] On the other hand, the present invention provides a relay power and edge computing resource optimization system based on cloud-edge collaboration, comprising:

[0026] The module is used to build a cloud-edge collaborative resource optimization model in the relay-assisted NOMA-MEC system scenario, with the goal of minimizing the weighted sum of total task latency and total task energy consumption;

[0027] The relay-assisted NOMA-MEC system scenario includes a cloud control center, multiple relays, multiple power zones, and multiple edge gateways that collect data from the corresponding power zones;

[0028] The total task latency includes the first time slot, the second time slot, the third time slot, and the total computation latency of the cloud control center; the total task energy consumption includes the total energy consumption of task transmission and the total energy consumption of task computation.

[0029] The tasks include: several local tasks to be executed on the edge gateway, several relay tasks to be executed on the relay, and several cloud tasks to be executed on the cloud control center; wherein, the local tasks and relay tasks are all unloaded by the edge gateway, the cloud tasks are unloaded by the edge gateway, and the tasks are unloaded by the edge gateway to the relay and then unloaded by the relay.

[0030] The first time slot is the maximum transmission delay of several relay tasks offloaded from the edge gateway to the relay; the second time slot is the maximum of the transmission delay of several cloud tasks offloaded from the edge gateway to the relay and the transmission delay of several cloud tasks directly offloaded from the edge gateway to the cloud control center; the third time slot is the maximum transmission delay of several cloud tasks offloaded from the relay to the cloud control center.

[0031] The solver module is used to solve the cloud-edge collaborative resource optimization model based on the block coordinate descent joint solver algorithm to obtain the optimal resource allocation scheme.

[0032] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0033] This invention proposes a cloud-edge collaborative method for optimizing relay power and edge computing resources. In a relay-assisted NOMA-MEC system scenario, a cloud-edge collaborative resource optimization model is constructed with the objective of minimizing the weighted sum of total task latency and total task energy consumption. The model is then solved using a joint solution algorithm based on block coordinate descent to obtain the optimal resource allocation scheme. Firstly, the relay-assisted NOMA-MEC system scenario includes a cloud control center, multiple relays, multiple power zones, and multiple edge gateways collecting data from the corresponding power zones, allowing for many-to-many selection of edge gateways and relays. Firstly, by combining the resource optimization method of this invention, the resource waste problem under the traditional one-to-one model is solved. Secondly, the cloud-edge collaborative resource optimization model aims to minimize the weighted sum of the total system task processing latency and total energy consumption by jointly optimizing relay power (relay allocation power matrix P), relay CPU frequency (relay computing resource matrix f), execution time (time matrix t), task offloading decision (task matrix L), and relay selection matrix A. This solves the problem of optimal allocation of communication resources under the many-to-many model, realizes the real-time reliable transmission and processing of massive power data, and ensures the stable operation of the smart grid.

[0034] The present invention proposes a method for optimizing relay power and edge computing resources based on cloud-edge collaboration. This method uses a joint solution algorithm with block coordinate descent to solve the original complex optimization problem into several sub-problems. The optimal solution of the original problem is obtained by iteratively solving the sub-problems, thus solving the problem of effectively solving non-convex optimization problems with strongly coupled variables. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly described below. Obviously, the drawings described below 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, wherein:

[0036] Figure 1 shows a flowchart of the relay power and edge computing resource optimization method based on cloud-edge collaboration in an embodiment of the present invention;

[0037] Figure 2 shows a schematic diagram of a relay-assisted NOMA-MEC system scenario in an embodiment of the present invention.

[0038] Figure 3 shows a schematic diagram of the transmission time slot model in an embodiment of the present invention;

[0039] Figure 4 shows a comparison of the total task latency under different methods in the embodiments of the present invention;

[0040] Figure 5 shows a comparison of the total energy consumption of the task under different methods in the embodiments of the present invention;

[0041] Figure 6 shows a comparison of the maximum number of edge gateways that can be served by the comparative method in the embodiment of the present invention and the method of the present invention. Detailed Implementation

[0042] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0043] The application principle of the present invention will be described in detail below with reference to the accompanying drawings.

[0044] Example 1: As shown in Figure 1, this embodiment of the invention provides a method for optimizing relay power and edge computing resources based on cloud-edge collaboration, including the following steps:

[0045] S01: In the scenario of relay-assisted NOMA-MEC system, construct a cloud-edge collaborative resource optimization model with the goal of minimizing the weighted sum of total task latency and total task energy consumption;

[0046] S02: Solve the cloud-edge collaborative resource optimization model based on the block coordinate descent joint solution algorithm to obtain the optimal resource allocation scheme.

[0047] Furthermore, the relay-assisted NOMA-MEC (Non-Orthogonal Multiple Access - MobileEdge Computing) system scenario includes a cloud control center, multiple relays, multiple power zones, and multiple edge gateways that collect data from the corresponding power zones. As shown in Figure 2, each edge gateway transmits the collected power data information (data from the power zone) to the cloud control center through the relay-assisted NOMA-MEC network, and generates the optimal resource allocation at the cloud control center based on the relay power and edge computing resource optimization method based on cloud-edge collaboration in this embodiment.

[0048] Furthermore, the total task latency includes the first time slot, the second time slot, the third time slot, and the total computation latency of the cloud control center; the total task energy consumption includes the total energy consumption of task transmission and the total energy consumption of task computation.

[0049] The first time slot is the maximum transmission delay of several relay tasks offloaded from the edge gateway to the relay; the second time slot is the maximum of the transmission delay of several cloud tasks offloaded from the edge gateway to the relay and the transmission delay of several cloud tasks directly offloaded from the edge gateway to the cloud control center; the third time slot is the maximum transmission delay of several cloud tasks offloaded from the relay to the cloud control center.

[0050] Furthermore, the tasks include: several local tasks to be executed on the edge gateway, several relay tasks to be executed on the relay, and several cloud tasks to be executed on the cloud control center; wherein, the local tasks and relay tasks are all unloaded by the edge gateway, the cloud tasks are unloaded by the edge gateway, and the tasks are unloaded by the edge gateway to the relay and then unloaded by the relay.

[0051] In this embodiment, the specific steps in step S01 of constructing the cloud-edge collaborative resource optimization model with the objective of minimizing the weighted sum of total task latency and total task energy consumption in the relay-assisted NOMA-MEC system scenario are as follows:

[0052] S011: Model the communication model, offloading model, and computation model of the relay-assisted NOMA-MEC system;

[0053] (1) Communication model:

[0054] (1) Transmission time slot model:

[0055] Assuming the first [connection] in the smart grid corresponding to the power area Edge gateway in a cycle The total number of tasks that need to be transmitted and processed is , It can be divided into three parts, namely: the first Local tasks of the edge gateway Relay mission and cloud mission And it is executed locally, on a relay, and in the cloud control center respectively. Therefore satisfy

[0056] (1)

[0057] in, express A collection of edge gateways.

[0058] To implement the resource optimization method described in this embodiment, a time-slot transmission model as shown in Figure 3 is used, where the period... It is divided into four time slots (first time slot) Second time slot Third time slot and the fourth time slot This is used for the transmission and processing of tasks. The fourth time slot represents the computational latency of the cloud control center for all cloud-layer tasks.

[0059] The specific collaboration process among the edge gateway, relay, and control center in each time slot is described below:

[0060] First time slot The nth edge gateway will need to perform relay tasks in the relay. The task is unloaded to the appropriate relay, and then the relay receiving the task can... These tasks are executed within the time slot. In the second time slot... and the third time slot , No. Each edge gateway, in collaboration with the relay, will perform cloud-layer tasks that need to be executed in the cloud control center. Offloaded to the cloud. Because the relay uses a DF (Decode-and-forward) strategy for transmission, in the second time slot, the... Each edge gateway will first handle cloud tasks. Simultaneously transmitted to the relay and cloud control centers. In the third time slot, the relay decodes the cloud layer task. And then resend it to the cloud control center. After receiving the offloading task from the edge gateway and relay, the cloud control center, in the fourth time slot... Begin executing these tasks. Because the edge gateway can process and transmit tasks simultaneously, local tasks... The execution time is the entire cycle. .

[0061] (2) Relay selection model

[0062] To improve the utilization efficiency of relay computing and communication resources, this embodiment assumes that each relay can provide services to multiple users, but each user can only select one relay. Therefore, a relay selection matrix is ​​introduced. Correspondingly, Satisfy the following constraints

[0063] (2)

[0064] (3)

[0065] in, As a relay selection factor, Explanation of edge gateway Select relay Perform collaborative communication and task unloading; otherwise... . express A set of cooperative relays.

[0066] (3) Relay interference model based on NOMA

[0067] In the third time slot, the relay will receive the cloud mission from the second time slot. The data is resent to the cloud control center. Since each relay can provide auxiliary transmission for multiple edge gateways simultaneously, internal interference exists when the same relay transmits tasks from different gateways. Furthermore, considering that relays use a common channel for transmission, inter-group interference also exists between different relays. To address the intra-group interference problem, NOMA technology is employed. Because the distance between the same relay and the cloud control center is equidistant, the cloud control center can determine the signal strength by the power allocated to different tasks by the relay. This embodiment assumes uplink NOMA transmission; the cloud control center prioritizes decoding strong signals, and the decoded signal is removed from the original signal. Therefore, subsequently decoded weak signals will not be interfered with by strong signals.

[0068] To avoid loss of generality, assume that the first option is chosen. The number of edge gateways for each relay is The power allocated to each cloud task by the relay is arranged in descending order. Cloud mission The interference signal based on NOMA technology can be written as

[0069] (4)

[0070] in, This indicates inter-group interference. This indicates interference within the group; This represents the set of edge gateways that select the m-th relay for communication; This represents the set of edge gateways that select the j-th relay for communication; This represents the relay selection factor that determines whether the h-th edge gateway selects or does not select the j-th relay for collaborative communication and task offloading. This represents the relay selection factor that determines whether the k-th edge gateway selects or does not select the m-th relay for collaborative communication and task offloading. Indicates the first The transmission power of each relay allocated to the cloud task of the kth edge gateway; This represents the transmission power of the cloud task allocated by the j-th relay to the h-th edge gateway; This represents the channel gain between the j-th relay and the control center; j, h, and k are sequence numbers.

[0071] (ii) Unloading Model:

[0072] (1) Local-relay unloading:

[0073] The nth edge gateway first relays the task in the first time slot. Unloading the m-th relay selected for it, according to Shannon's theorem, the transmission rate of this process is... It can be represented as:

[0074] (5)

[0075] in, This indicates that the nth edge gateway will relay the task. and cloud mission Unload the transmission rate of the m-th relay selected for it; Indicates channel bandwidth; This represents the transmission power of the nth edge gateway; This represents the channel gain between the nth edge gateway and the mth relay; The variance of the Gaussian noise in the channel is represented;

[0076] Therefore, the actual transmission delay of this process It can be represented as

[0077] (6)

[0078] Because all edge gateways need to complete the offloading of relay tasks within the first time slot, the first time slot Need to meet

[0079] (7)

[0080] Accordingly, the nth edge gateway will Transmission power consumption for transmission to the m-th relay Recorded as

[0081] (8)

[0082] (2) Local-to-cloud offloading based on collaborative relay:

[0083] This process mainly takes place in the second and third time slots. In the second time slot, the nth edge gateway simultaneously loads cloud tasks. Transmitted to the m-th relay and the cloud control center. The transmission rate from the n-th edge gateway to the cloud control center. It can be written as

[0084] (9)

[0085] in, This indicates that the nth edge gateway will handle cloud tasks. The transmission rate offloaded to the cloud control center; This represents the channel gain between the nth edge gateway and the cloud control center.

[0086] In addition, to ensure that both the m-th relay and the cloud control center can receive the complete task offloading in the second time slot, the cloud task... The transmission delay in the second time slot, i.e., the nth edge gateway will transmit the cloud task The maximum value of the transmission delay between the m-th relay and the cloud control center selected for unloading. , can be represented as

[0087] (10)

[0088] Specifically, the channel gain model can be expressed as: ,in Indicates path loss. The distance between the sender and the receiver. This represents the path loss exponent. To improve communication quality, relays are typically located between the edge gateway and the cloud control center; that is, the distance between the relay and the edge gateway is shorter than the distance between the cloud control center and the edge gateway, thus incurring channel gain. Furthermore, it is assumed that the nth edge gateway uses the same power. send Therefore, we can obtain Therefore, the nth edge gateway will handle cloud tasks. The maximum value of the transmission delay between the m-th relay selected for unloading and the cloud control center is:

[0089] (11)

[0090] Because all edge gateways need to offload cloud tasks during the second time slot, the second time slot The following conditions must be met:

[0091] (12)

[0092] Based on the NOMA-based relay interference model, the transmission rate from the third time slot relay to the cloud control center can be expressed as:

[0093] (13)

[0094] in, This indicates that the m-th relay will carry the cloud task of the n-th edge gateway. The transmission rate offloaded to the cloud control center; This represents the channel gain between the m-th relay and the control center; This represents the m-th relay transmission cloud task. Interference signals,

[0095] Furthermore, assuming the cloud control center uses maximal ratio combining (MRC) to receive signals and its actual rate equals the achievable rate, the actual rate at the cloud control center can be expressed as:

[0096] (14)

[0097] To ensure reliable decoding and forwarding by the relay, the transmission capacity from the relay to the cloud control center cannot exceed the transmission capacity from the edge gateway to the relay.

[0098] (15)

[0099] Based on the above analysis, for:

[0100] (16)

[0101] The task transmission delay in the third time slot, i.e., the cloud task of the nth edge gateway being transmitted by the mth relay. Transmission latency from offloading to the cloud control center , can be represented as:

[0102] (17)

[0103] Because all relays need to offload cloud tasks within the third time slot, the third time slot... Need to meet

[0104] (18)

[0105] The transmission energy consumption of the second and third time slots are as follows:

[0106] (19)

[0107] (20)

[0108] in, This represents the energy consumption of the nth edge gateway transmitting cloud tasks to the mth relay and the cloud control center. This indicates that the m-th relay will carry the cloud task of the n-th edge gateway. Energy consumption of transmission when offloading to the cloud control center;

[0109] The nth edge gateway throughout the entire cycle Internal task transmission power consumption It can be represented as:

[0110] (twenty one)

[0111] (III) Calculation Model:

[0112] (1) Local computation at the edge gateway:

[0113] The nth edge gateway, having an MEC server, can process some tasks locally, thus implementing edge computing functionality. The local computing latency model for the nth edge gateway, i.e., the computation latency of the local MEC server for the local tasks of the nth edge gateway, is as follows:

[0114] (twenty two)

[0115] in, (cycle / bit) is the number of CPU cycles required for an edge gateway to process 1 bit of a task. (cycle / s) represents the CPU frequency of the MEC server equipped with the nth edge gateway.

[0116] The energy consumption model for local computing, i.e., the computing energy consumption of the local MEC server executing the local task of the nth edge gateway, is...

[0117] (twenty three)

[0118] in, This is a coefficient related to the CPU hardware architecture of the local MEC server.

[0119] (2) Collaborative computation in relay:

[0120] Because relays can allocate appropriate computing resources to each offloading task based on the computing needs of each edge gateway. Relay tasks The computation time, i.e., the computation latency of the relay MEC server for the relay task of the nth edge gateway, is...

[0121] (twenty four)

[0122] in, (cycle / bit) is the number of CPU cycles required to relay a 1-bit task. (cycle / s) represents the computing resources allocated by the m-th relay to handle the relay task of the n-th edge gateway.

[0123] The m-th relay is executed. The computational energy consumption can be expressed as

[0124] (25)

[0125] in, The coefficients are related to the CPU hardware architecture of the relay MEC server.

[0126] (3) Remote computing in the cloud:

[0127] The cloud control center is equipped with high-performance MEC servers capable of handling relatively complex tasks, assuming the cloud processes all remote offloading tasks at its maximum frequency. Cloud Tasks The computation delay is:

[0128] (26)

[0129] in, (cycle / bit) is the number of CPU cycles required to process 1 bit of task in the cloud control center. (cycle / s) represents the maximum computing frequency of the cloud control center.

[0130] The cloud control center executes cloud-level tasks. The computational energy consumption can be expressed as:

[0131] (27)

[0132] in, The coefficients are related to the CPU hardware architecture of the cloud control center.

[0133] Based on the above analysis, the energy consumption of the task computation of the nth edge gateway is...

[0134] (28)

[0135] S012: By combining the communication model, offloading model, and computation model, the resource optimization problem is modeled to obtain the cloud-edge collaborative resource optimization model.

[0136] Specifically, based on the communication model, offloading model, and computation model, the following decision variables are defined: , , , and .

[0137] Where A represents the relay selection matrix; The relay selection factor indicates whether the nth edge gateway chooses or does not choose the mth relay for collaborative communication and task offloading; M represents the total number of relays; N represents the total number of edge gateways; L represents the task matrix. , and These represent the local task, relay task, and cloud task of the nth edge gateway, respectively; P represents the relay power allocation matrix. represents the transmission power of the cloud task allocated by the m-th relay to the n-th edge gateway; f represents the relay computing resource matrix; This represents the computing resources allocated by the m-th relay to handle the relay task of the n-th edge gateway; t represents the time matrix; T represents the period, which is the upper limit of the total computational latency.

[0138] The objective function, which aims to minimize the weighted sum of the total task latency and total task energy consumption, can be expressed as: (29)

[0139] Wherein: the first part of the objective function represents the total time required for all edge gateway tasks to be executed from the start to the completion of all tasks, i.e., the total task latency; the second part represents the total energy consumption of all edge gateway tasks from the start to the completion of all tasks, i.e., the total task energy consumption.

[0140] In the formula, It is a dimensional parameter used to unify the units of time delay and energy consumption; Represent the objective function; , All represent weighting coefficients, satisfying Furthermore, it can be adjusted according to different scenario requirements; N represents the total number of edge gateways; n represents the number of sequences; Indicates the first time slot; Indicates the second time slot; Indicates the third time slot; This represents the computation latency of the cloud task offloaded by the nth edge gateway in the cloud control center. This represents the transmission energy consumption of the task offloaded by the nth edge gateway within period T; This represents the computing energy consumption of all tasks of the nth edge gateway within period T, including the computing energy consumption of local tasks, relay tasks, and cloud tasks of the nth edge gateway. This represents the total computational latency of the cloud control center; Indicates the total energy consumption of the task; This indicates the total energy consumption for task computation; Indicates the total energy consumption of task transmission;

[0141] According to the transmission time slot model, the local, relay, and cloud control centers must all begin execution upon receiving the task and within the specified period. The task is to be completed within a certain timeframe; therefore, the time constraint for task execution is as follows:

[0142] (30)

[0143] (31)

[0144] (32)

[0145] (33)

[0146] (34)

[0147] (35)

[0148] The implicit task partitioning of C4-C6 needs to meet the following requirements: conditions.

[0149] According to formulas (7), (12), and (18), to ensure that all tasks can be unloaded within each time slot, the following time slot allocation constraints apply:

[0150] (36)

[0151] (37)

[0152] (38)

[0153] In this embodiment, each relay can be selected by multiple edge gateways and provided with collaborative services. Considering that each relay has limited communication and computing resources, the resource allocation of each relay needs to meet relay resource allocation constraints:

[0154] (39)

[0155] (40)

[0156] To ensure reliable decoding and forwarding of relays, according to formula (15), there are constraints on reliable decoding and forwarding of relays:

[0157] (41)

[0158] Therefore, the joint optimization problem, i.e., the cloud-edge collaborative resource optimization model, can be expressed as:

[0159]

[0160]

[0161] In this embodiment, the specific process of solving the cloud-edge collaborative resource optimization model based on the block coordinate descent joint solution algorithm in step S02 includes:

[0162] S021: Determine the relay selection matrix based on the greedy algorithm, and update the cloud-edge collaborative resource optimization model accordingly to obtain the updated cloud-edge collaborative resource optimization model;

[0163] Specifically, considering that the transmission performance of a relay is related to the channel gain, and the channel gain is affected by the distance, a greedy algorithm is used to determine the relay selection matrix. Specifically, each edge gateway selects the nearest relay as the task offloading relay to obtain the maximum channel gain and improve communication quality.

[0164] Among relay selection matrix Once confirmed, select the relay. A collection of edge gateways for collaborative communication This was also subsequently determined. Cloud mission. The interference signal is thus updated to Therefore, the joint optimization problem, namely the updated cloud-edge collaborative resource optimization model, is... It can be written in the following form:

[0165]

[0166]

[0167]

[0168]

[0169] No longer contains variables However, it is still a non-convex optimization problem with strongly coupled variables.

[0170] S022: The updated cloud-edge collaborative resource optimization model is decomposed into four sub-problems using a joint solution algorithm based on block coordinate descent, and the four sub-problems are solved alternately to obtain the optimal resource allocation scheme.

[0171] Specifically, Block Coordinate Descent (BCD) is an efficient iterative algorithm for solving optimization problems with strongly coupled variables. Its core idea is to decompose a high-dimensional, complex optimization problem into a series of small, easily tractable subproblems, and then gradually approximate the optimal solution of the original problem by optimizing the variable blocks in turn. This embodiment proposes a joint solution algorithm based on BCD. First, the BCD algorithm is used to decompose the optimization problem... The problem is decomposed into four sub-problems (solving problems for variables P, f, L, and t), and then the optimization problem is obtained by solving these sub-problems alternately. The optimal solution;

[0172] (1) Variables Optimization solution

[0173] By using the BCD method, first fix the variables. , and Optimization problem It is then transformed into a variable Sub-optimization problem

[0174]

[0175]

[0176] According to constraints Therefore, to meet the task offloading requirements of the relay-cloud control center, the relay's transmission power should meet the following requirements.

[0177] (42)

[0178] Sub-optimization problem The optimal power allocation is derived from Theorem 1.

[0179] Lemma 1: Function It is a concave function that is monotonically increasing.

[0180] Proof: Function first derivative It is easy to prove by differentiation. ,therefore, ,function Monotonically increasing. Function The second derivative is ,make ,right Find the first derivative and let It can be obtained According to monotonicity, when hour Get the maximum value ,therefore Monotonically decreasing, and with a maximum value Based on the above derivation, we can obtain... Therefore, the function It is a monotonically increasing concave function.

[0181] Theorem 1: When the variable , and Fixed and satisfying constraints and , Sub-optimization problem The optimal solution.

[0182] Proof: Let Because of the function Dividing variables Apart from the other parameters, all other parameters are constants, therefore the function can be... Simplified to ( ),in , , According to the conclusion of Lemma 1, The function is a monotonically increasing concave function, and its optimal solution is obtained at the lower bound of the variable. Furthermore, considering the constraints... , And formula (42), when and At that time, sub-optimization problem The optimal solution is .

[0183] (2) Variables Optimization solution

[0184] Similarly, when the variable is fixed... , and Optimization problem It is then transformed into a variable Sub-optimization problem

[0185]

[0186]

[0187] Sub-optimization problem The optimal solution is derived from Theorem 2.

[0188] Theorem 2: When the variable , and Fixed and satisfying constraints , Sub-optimization problem The optimal solution.

[0189] Proof: It is easy to see from the objective function that the sub-optimization problem It's about variables. A monotonically increasing function, combined with constraints and ,when Take the lower bound and satisfy The objective function value is minimized. Therefore, Sub-optimization problem The optimal solution.

[0190] According to Theorem 1 and Theorem 2, to guarantee the suboptimal problem and The optimal solution is obtained in each iteration, and the solution is obtained with respect to the variables. and When dealing with sub-optimization problems, the following constraints need to be added.

[0191] (43)

[0192] (44)

[0193] (45)

[0194] (3) Variables Optimization solution

[0195] When fixed variables , and Optimization problem It is then transformed into a variable Sub-optimization problem

[0196]

[0197]

[0198] Lemma 2: Suboptimal Problems It's about variables. Convex optimization.

[0199] Proof: It is easy to see that when the variable , and When fixed, sub-optimization problem The objective function is about the variable Convex function, constraint It's about variables. The convex set. Let , ,because It is about If a function is convex, then its nonnegative weighted sum is... It remains a convex function. Furthermore, due to the convex function's... It is a convex set, therefore it is constrained. For about variables Convex set. Similarly, constraint. Also about variables Convex set. Therefore, sub-optimization problem It's about variables. Convex optimization.

[0200] (4) Variables Optimization solution

[0201] When fixed variables , and Optimization problem It is then transformed into a variable Sub-optimization problem

[0202]

[0203]

[0204] Lemma 3: Suboptimal Problems It's about variables. Convex optimization.

[0205] Proof: It is easy to prove that when the variable , and When fixed, sub-optimization problem The objective function is about the variable Convex function, constraint It's about variables. A convex set. Let the function , , The second derivative is Therefore, the function It is a convex function. The convex function's... It is a convex set, therefore it is constrained. For about variables A convex set. Similarly, constraints exist The scope is also about variables Convex set. Therefore, sub-optimization problem It's about variables. Convex optimization.

[0206] Based on the principles of the BCD algorithm, the optimization subproblem is solved through alternating iterations. When the number of iterations satisfies The algorithm terminates, and the final variable... , , and It can converge to the original optimization problem The optimal solution;

[0207] in, Indicates the maximum number of iterations;

[0208] In summary, this embodiment introduces a method for optimizing relay power and edge computing resources based on cloud-edge collaboration. By jointly optimizing the communication and computing resources of the smart grid information system, a cloud-edge collaborative resource optimization model is constructed with the goal of minimizing the weighted sum of total task latency and total task energy consumption. The cloud-edge collaborative resource optimization model is solved using a joint solution algorithm based on block coordinate descent to obtain the optimal resource allocation scheme, thereby enabling real-time and reliable transmission and processing of massive amounts of power data and ensuring the stable operation of the smart grid.

[0209] Example 2 builds upon the cloud-edge collaborative relay power and edge computing resource optimization method introduced in Example 1. This example conducts the following experimental verification:

[0210] Experiment Introduction: This experiment simulates the joint resource optimization of a relay-assisted edge computing network based on cloud-edge collaboration, as shown in Figure 2. The scenario includes a cloud control center, M repeaters, and N power zones. The cloud control center, based on data collected from the edge gateways and relay resource information, uses the joint resource optimization method for the relay-assisted edge computing network based on cloud-edge collaboration described in Example 1 to jointly optimize relay power, relay CPU frequency, execution time, and task offloading decisions.

[0211] Experimental Analysis:

[0212] As shown in Figure 4, the system latency of different solutions tends to increase with the increase in the number of edge gateways. This is because the increase in edge gateways leads to a relative decrease in the resources allocated to each gateway, thus affecting the task processing time.

[0213] Furthermore, Figure 4 visually illustrates the lowest system latency achieved by the method proposed in this invention. A detailed analysis follows: Compared to a relay-less cloud-edge collaboration scheme, the proposed scheme reduces task offloading latency by introducing a relay and optimizing its transmission power. Compared to a cloud-edge collaboration scheme where the relay only participates in communication, the relay in the proposed scheme can undertake certain task processing functions, reducing the transmission pressure of offloading tasks to the cloud to a certain extent, thereby further reducing system latency. Cloud computing with relays lacks cloud-edge collaboration, increasing the pressure of task offloading and leading to increased system latency.

[0214] Figure 5 illustrates the change in system energy consumption with the number of edge gateways under different schemes. It can be clearly seen that the system energy consumption of different schemes increases with the increase in the number of edge gateways, and the scheme proposed in this invention has the lowest energy consumption. Due to the poor performance of the local MEC server, the task processing energy consumption of the local-only scheme is the highest. Although the MEC server equipped in the cloud control center has high performance, the task offloading latency is too long, resulting in high energy consumption. The scheme using a cloud-edge collaborative architecture can reduce cloud transmission energy consumption while reducing local computing energy consumption. Compared to the cloud-edge collaborative scheme where the relay only participates in communication, the scheme proposed in this invention can offload some local tasks to the relay for processing and optimize the allocation of computing resources, thereby reducing the system's transmission energy consumption. The results in Figure 4 show that the introduction of relays can reduce transmission latency, and with the optimized allocation of relay transmission power, the energy consumption of the scheme with relay collaboration is lower than that without relays.

[0215] Figure 6 shows the maximum number of edge gateways the system can serve under different numbers of relays. In the comparative method, due to the use of a one-to-one model, the maximum number of edge gateways that can be served is equal to the number of relays. As can be seen from the figure, when the number of relays is the same, each relay in the model proposed in this invention can provide services to multiple edge gateways, and its maximum number of edge gateways that can be served is greater than that of the comparative method. Therefore, it can be proven that the model proposed in this invention can improve the utilization rate of system resources.

[0216] Based on the above experimental analysis, it can be concluded that, compared with other methods, the method proposed in this invention can improve the utilization rate of network resources while reducing the overall latency and energy consumption of the system, thereby providing strong support for the stable operation of the smart grid.

[0217] Example 3: This embodiment of the invention introduces a relay power and edge computing resource optimization system based on cloud-edge collaboration, including;

[0218] The module is used to build a cloud-edge collaborative resource optimization model in the relay-assisted NOMA-MEC system scenario, with the goal of minimizing the weighted sum of total task latency and total task energy consumption;

[0219] The relay-assisted NOMA-MEC system scenario includes a cloud control center, multiple relays, multiple power zones, and multiple edge gateways that collect data from the corresponding power zones;

[0220] The total task latency includes the first time slot, the second time slot, the third time slot, and the total computation latency of the cloud control center; the total task energy consumption includes the total energy consumption of task transmission and the total energy consumption of task computation.

[0221] The tasks include: several local tasks to be executed on the edge gateway, several relay tasks to be executed on the relay, and several cloud tasks to be executed on the cloud control center; wherein, the local tasks and relay tasks are all unloaded by the edge gateway, the cloud tasks are unloaded by the edge gateway, and the tasks are unloaded by the edge gateway to the relay and then unloaded by the relay.

[0222] The first time slot is the maximum transmission delay of several relay tasks offloaded from the edge gateway to the relay; the second time slot is the maximum of the transmission delay of several cloud tasks offloaded from the edge gateway to the relay and the transmission delay of several cloud tasks directly offloaded from the edge gateway to the cloud control center; the third time slot is the maximum transmission delay of several cloud tasks offloaded from the relay to the cloud control center.

[0223] The solver module is used to solve the cloud-edge collaborative resource optimization model based on the block coordinate descent joint solver algorithm to obtain the optimal resource allocation scheme.

[0224] The specific functions of each module described above are explained in the relevant content of the method in Embodiment 1, and will not be repeated here.

[0225] Example 4 This example also introduces a readable storage medium on which a computer program is stored. When the computer program is executed by a processor, it implements the relay power and edge computing resource optimization method based on cloud-edge collaboration in Example 1.

[0226] Example 5 This example also introduces a computer program product, including a computer program that, when executed by a processor, implements the relay power and edge computing resource optimization method based on cloud-edge collaboration in Example 1.

[0227] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0228] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or one or more blocks of the block diagrams.

[0229] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0230] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for optimizing relay power and edge computing resources based on cloud-edge collaboration, characterized in that, include: In the relay-assisted NOMA-MEC system scenario, a cloud-edge collaborative resource optimization model is constructed with the goal of minimizing the weighted sum of total task latency and total task energy consumption. The relay-assisted NOMA-MEC system scenario includes a cloud control center, multiple relays, multiple power zones, and multiple edge gateways that collect data from the corresponding power zones; the total task latency includes the first time slot, the second time slot, the third time slot, and the total computation latency of the cloud control center; The total energy consumption of the task includes the total energy consumption of task transmission and the total energy consumption of task computation; the task includes: several local tasks offloaded to the edge gateway for execution, several relay tasks offloaded to the relay for execution, and several cloud tasks offloaded to the cloud control center for execution; wherein, local tasks and relay tasks are all offloaded by the edge gateway, cloud tasks are offloaded by the edge gateway, and tasks are offloaded from the edge gateway to the relay and then offloaded by the relay; wherein, the first time slot is the maximum transmission delay of several relay tasks offloaded from the edge gateway to the relay; the second time slot is the maximum of the transmission delay of several cloud tasks offloaded from the edge gateway to the relay and the transmission delay of several cloud tasks directly offloaded from the edge gateway to the cloud control center; the third time slot is the maximum transmission delay of several cloud tasks offloaded from the relay to the cloud control center; the cloud-edge collaborative resource optimization model is solved based on the block coordinate descent joint solution algorithm to obtain the optimal resource allocation scheme.

2. The method for optimizing relay power and edge computing resources based on cloud-edge collaboration according to claim 1, characterized in that, The cloud-edge collaborative resource optimization model includes an objective function and constraints; the constraints include: task allocation constraints, relay selection constraints, relay selection factor constraints, latency constraints, time slot allocation constraints, relay resource allocation constraints, and relay reliable decoding and forwarding constraints.

3. The method for optimizing relay power and edge computing resources based on cloud-edge collaboration according to claim 2, characterized in that, The objective function is expressed as follows: ;in: It is a dimensional parameter used to unify the units of time delay and energy consumption; Represent the objective function; ， All represent weighting coefficients; N represents the total number of edge gateways; n represents the number of sequences; Indicates the first time slot; Indicates the second time slot; Indicates the third time slot; This represents the computation latency of the cloud task offloaded by the nth edge gateway in the cloud control center. This represents the transmission energy consumption of the task offloaded by the nth edge gateway within period T; This represents the computing energy consumption of all tasks of the nth edge gateway within period T, including the computing energy consumption of local tasks, relay tasks, and cloud tasks of the nth edge gateway. This represents the total computational latency of the cloud control center; Indicates the total energy consumption of the task; This indicates the total energy consumption for task computation; Indicates the total energy consumption of task transmission; These are all decision variables of the objective function, used to determine the total task latency and total task energy consumption in the objective function, and are represented as follows: ； ； ； ； Where A represents the relay selection matrix; The relay selection factor indicates whether the nth edge gateway selects or does not select the mth relay for collaborative communication and task offloading; M represents the total number of relays; L represents the task matrix. 、 and These represent the local task, relay task, and cloud task of the nth edge gateway, respectively; P represents the relay power allocation matrix. represents the transmission power of the cloud task allocated by the m-th relay to the n-th edge gateway; f represents the relay computing resource matrix; This represents the computing resources allocated by the m-th relay to handle the relay task of the n-th edge gateway; t represents the time matrix; T represents the period, which is the upper limit of the total computational latency.

4. The method for optimizing relay power and edge computing resources based on cloud-edge collaboration according to claim 3, characterized in that, The calculation process for the total task latency, including the first time slot, the second time slot, the third time slot, and the total computation latency of the cloud control center, is as follows: First time slot: ； ； ;in: This indicates that the nth edge gateway will relay the task. The transmission delay of the m-th relay selected for unloading; This indicates that the nth edge gateway will relay the task. and cloud mission Unload the transmission rate of the m-th relay selected for it; Indicates channel bandwidth; This represents the transmission power of the nth edge gateway; This represents the channel gain between the nth edge gateway and the mth relay; The variance of the channel's Gaussian noise is represented; second time slot: ； ;in: This indicates that the nth edge gateway will handle cloud tasks. The maximum value of the transmission delay between the m-th relay and the cloud control center selected for unloading; This indicates that the nth edge gateway will handle cloud tasks. The transmission rate offloaded to the cloud control center; This represents the channel gain between the nth edge gateway and the cloud control center; the third time slot: ； ； ;in: This indicates that the m-th relay will carry the cloud task of the n-th edge gateway. Transmission latency from offloading to the cloud control center; This indicates that the m-th relay will carry the cloud task of the n-th edge gateway. The transmission rate offloaded to the cloud control center; This represents the channel gain between the m-th relay and the control center; This represents the m-th relay transmission cloud task. The interference signal is expressed by the following formula: ;in, This represents the set of edge gateways that select the m-th relay for communication; This represents the set of edge gateways that select the j-th relay for communication; This represents the relay selection factor that determines whether the h-th edge gateway selects or does not select the j-th relay for collaborative communication and task offloading. This represents the relay selection factor that determines whether the k-th edge gateway selects or does not select the m-th relay for collaborative communication and task offloading. Indicates the first The transmission power of each relay allocated to the cloud task of the kth edge gateway; This represents the transmission power of the cloud task allocated by the j-th relay to the h-th edge gateway; Let represent the channel gain between the j-th relay and the control center; j, h, and k are sequence numbers; the total computational delay of the cloud control center is obtained by summing the computational delays of the cloud control center for the cloud tasks of all edge gateways, where the computational delay of the cloud control center for the cloud tasks of the n-th edge gateway is denoted by . The calculation formula is expressed as follows: ;in, This indicates the number of CPU cycles required to process a 1-bit task in the cloud control center. This indicates the maximum computing frequency of the cloud control center.

5. The method for optimizing relay power and edge computing resources based on cloud-edge collaboration according to claim 4, characterized in that, The calculation process for the total energy consumption of task transmission and the total energy consumption of task computation is as follows: Total energy consumption of task transmission: The total energy consumption of task transmission is obtained by adding the energy consumption of task transmission of all edge gateways; the energy consumption of task transmission of the nth edge gateway... The formula is expressed as follows: ； ； ； ;in, This represents the transmission energy consumption when the nth edge gateway transmits the relay task to the mth relay. This represents the energy consumption of the nth edge gateway transmitting cloud tasks to the mth relay and the cloud control center. This indicates that the m-th relay will carry the cloud task of the n-th edge gateway. Energy consumption for offloading to the cloud control center; Total energy consumption for task computation: The total energy consumption for task computation is obtained by adding the energy consumption of all edge gateways; Energy consumption of the nth edge gateway for task computation. The formula is expressed as follows: ； ； ； ;in, This represents the computational energy consumption of the local MEC server executing the local task of the nth edge gateway; This indicates that the m-th relay is performing a relay task. The computational energy consumption; This indicates that the cloud control center is executing cloud-level tasks. The computational energy consumption; This represents a coefficient related to the CPU hardware architecture of the local MEC server. This represents the number of CPU cycles required for the edge gateway to process a 1-bit task. This represents the CPU frequency of the MEC server equipped with the nth edge gateway; This indicates the number of CPU cycles required to relay a 1-bit task. This represents the computing resources allocated by the m-th relay to the relay task that processes the n-th edge gateway; This represents a coefficient related to the CPU hardware architecture of the relay MEC server; This indicates the number of CPU cycles required to process a 1-bit task in the cloud control center. This represents a coefficient related to the CPU hardware architecture of the cloud control center.

6. The method for optimizing relay power and edge computing resources based on cloud-edge collaboration according to claim 5, characterized in that, The task allocation constraints are expressed as follows: ;in, express A collection of edge gateways; This represents the total number of tasks for the nth edge gateway; the relay selection constraints are as follows: The relay factor selection constraint is expressed as follows: ;in, Let M represent the set of M relays; This indicates that the nth edge gateway selects the mth relay for collaborative communication and task offloading; This indicates that the nth edge gateway does not select the mth relay for collaborative communication and task offloading; the latency constraint is expressed as follows: ； ； ； ； ； ;in, This represents the computation latency of the local MEC server for the local task of the nth edge gateway; This represents the computation latency of the relay MEC server for the relay task of the nth edge gateway; the time slot allocation constraint is expressed as follows: ； ； The relay resource allocation constraints are expressed as follows: ； ;in, This represents the maximum computing resource limit allocated to relay tasks for the m-th relay. Indicates the first The upper limit of transmission power allocated to each relay for cloud tasks; the relay reliable decoding and forwarding constraints are expressed as follows: 。 7. The method for optimizing relay power and edge computing resources based on cloud-edge collaboration according to claim 6, characterized in that, The joint solution algorithm based on block coordinate descent solves the cloud-edge collaborative resource optimization model to obtain the optimal resource allocation scheme. This includes: determining the relay selection matrix based on a greedy algorithm and updating the cloud-edge collaborative resource optimization model accordingly to obtain the updated cloud-edge collaborative resource optimization model; decomposing the updated cloud-edge collaborative resource optimization model into four sub-problems using the joint solution algorithm based on block coordinate descent, and solving the four sub-problems alternately to obtain the optimal resource allocation scheme.

8. The method for optimizing relay power and edge computing resources based on cloud-edge collaboration according to claim 7, characterized in that, The updated cloud-edge collaborative resource optimization model is expressed as follows: Updated objective function: ;in, This represents the updated objective function; since the relay selection matrix is ​​determined, the cloud task... The interference signal is updated as follows: ,thereby In and The constraints are updated based on the updated interference signal; the constraints are also updated based on the determined relay selection matrix as follows: ； ； ;in, It indicates that one is bound by something.

9. The method for optimizing relay power and edge computing resources based on cloud-edge collaboration according to claim 8, characterized in that, The four sub-problems include: the problem of solving for variable P, the problem of solving for variable f, the problem of solving for variable L, and the problem of solving for variable t; the problem of solving for variable P is expressed as follows: ； The problem of solving for variable f is expressed as follows: ； The problem of solving for variable L is expressed as follows: ； ； ； ； The problem of solving for variable t is expressed as follows: ； 。 10. A relay power and edge computing resource optimization system based on cloud-edge collaboration, characterized in that, include: The module is used to build a cloud-edge collaborative resource optimization model in the relay-assisted NOMA-MEC system scenario, with the goal of minimizing the weighted sum of total task latency and total task energy consumption; The relay-assisted NOMA-MEC system scenario includes a cloud control center, multiple relays, multiple power zones, and multiple edge gateways that collect data from the corresponding power zones; the total task latency includes the first time slot, the second time slot, the third time slot, and the total computation latency of the cloud control center; The total energy consumption of the task includes the total energy consumption of task transmission and the total energy consumption of task computation; the task includes: several local tasks offloaded to the edge gateway for execution, several relay tasks offloaded to the relay for execution, and several cloud tasks offloaded to the cloud control center for execution; wherein, local tasks and relay tasks are all offloaded by the edge gateway, cloud tasks are offloaded by the edge gateway, and tasks are offloaded from the edge gateway to the relay and then offloaded by the relay; wherein, the first time slot is the maximum transmission delay of several relay tasks offloaded from the edge gateway to the relay; the second time slot is the maximum of the transmission delay of several cloud tasks offloaded from the edge gateway to the relay and the transmission delay of several cloud tasks directly offloaded from the edge gateway to the cloud control center; the third time slot is the maximum transmission delay of several cloud tasks offloaded from the relay to the cloud control center; the solution module is used to solve the cloud-edge collaborative resource optimization model based on the block coordinate descent joint solution algorithm to obtain the optimal resource allocation scheme.

Images