Power business time delay optimization method and system

Abstract

The invention provides a power business time delay optimization method and system, and relates to the technical field of power system communication. According to the power business time delay optimization method provided by the invention, the maximum tolerant time delay of each calculation task is obtained by utilizing a clustering algorithm, and the maximum tolerant time delay is used for constraining the time delay optimization method of each calculation task, so that the differentiation demand between the calculation tasks is reduced to a certain extent; a service delay optimization problem is converted into a plurality of sub-problems and solved by using a distributed algorithm based on an alternating direction multiplier method, so that the probability of high calculation complexity caused by a large number of devices is reduced, the efficiency of solving a service delay optimization model is improved to a certain extent, and the time delay of running a calculation task is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of power system communications, and in particular to a method and system for optimizing power service delay. Background Art

[0002] With the rapid development of power wireless private networks, the number of power equipment connected to the network is also increasing, and various computing-intensive and delay-sensitive power services are emerging. Compared with traditional power services, these new power services require more powerful computing power and more energy. This has brought great challenges to traditional cloud computing. The overload of the core network and the increase in task transmission delay have become problems that need to be solved urgently. Mobile edge computing (MEC) is considered to be a key technology for the next generation of wireless communications. It deploys servers with powerful computing capabilities on the edge of wireless networks (such as base stations, wireless access points, etc.) to provide users with the required services nearby. With mobile edge computing, users can offload their computing tasks to edge servers for execution, thereby reducing the computing delay of tasks, reducing users' energy consumption, and meeting users' service quality requirements. However, due to limited computing resources, too many task offloading terminals will cause computing congestion on MEC servers and increase system latency. Therefore, in order to give full play to the potential advantages of mobile edge computing in power wireless private networks, it is necessary to study efficient and reasonable resource allocation and computing offloading strategies.

[0003] In addition, due to the diversity of power business types and the uneven distribution of power equipment, base stations may be overloaded in areas with dense power business or during peak hours. Existing studies generally use queues or retransmissions to reduce the load on base stations. However, considering the differentiated needs of power business, the uncertainty of matching computing resources with business needs will greatly affect the reliability of task completion and processing delay, resulting in a decrease in service quality. Summary of the invention

[0004] In order to solve the problem in the prior art that the use of queues or retransmissions will lead to increased uncertainty in matching computing resources with business needs, the present invention proposes a method for optimizing power business delay. A clustering algorithm is used to obtain the maximum tolerable delay of each computing task. The maximum tolerable delay is used to constrain the delay optimization method for each computing task, thereby reducing the differentiated requirements between computing tasks to a certain extent. A distributed algorithm based on the alternating direction multiplier method is used to convert the business delay optimization problem into multiple sub-problems and solve them, thereby reducing the probability of high computational complexity due to a large number of devices, improving the efficiency of solving the business delay optimization model, and reducing the delay of running computing tasks.

[0005] On the one hand, the present invention provides a method for optimizing power service delay, comprising:

[0006] A clustering algorithm is used to cluster multiple computing tasks connected to edge computing terminals to obtain multiple types of computing tasks and the maximum tolerable delay of each type of computing task;

[0007] Based on a preset computing mode, and according to computing resources allocated to each type of computing task by the computing mode, determining the latency and energy consumption of executing each type of computing task by the computing mode, the computing mode including local computing, base station computing or cloud computing;

[0008] Based on the latency and energy consumption of executing each type of computing task in each computing mode, a distributed algorithm based on the alternating direction multiplier method is used to iteratively solve the constructed optimization problem and determine the business offloading decision.

[0009] Optionally, the construction of the optimization problem includes:

[0010] Based on the latency and energy consumption of executing each type of computing task in each computing mode, the total overhead of each type of computing task is determined by using the service offloading variable for correction;

[0011] Based on the total overhead of each type of computing task, the optimization goal is to minimize the total overhead of multiple types of computing tasks and make each type of computing task less than its own maximum tolerable delay. Combined with the delay constraints of each type of computing task, the energy consumption constraints of each type of computing task and the computing resource constraints of each type of computing task, an optimization problem is constructed.

[0012] Optionally, the latency and energy consumption of executing each type of computing task based on each computing mode are corrected by using the service offloading variable to determine the total overhead of each type of computing task, including:

[0013] Based on the latency of executing each type of computing task in each computing mode, the service offloading variable is used to make corrections to determine the corrected total latency of each type of computing task;

[0014] Based on the energy consumption of executing each type of computing task in each computing mode, the energy consumption is corrected by using the business offloading variable to determine the corrected total energy consumption of each type of computing task;

[0015] The total delay and total energy consumption of each type of computing task after correction are comprehensively considered to determine the total overhead of each type of computing task.

[0016] Optionally, the corrected total delay of each type of computing task satisfies the following expression:

[0017]

[0018] In the formula, is the total delay of the kth type of computing task corresponding to the mth edge computing terminal associated with the nth base station; The latency of the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing; The time delay of the mth edge computing terminal associated with the nth base station to perform the kth computing task using the base station calculation; The latency of the mth edge computing terminal associated with the nth base station to perform the kth computing task using cloud computing; The task offloading variable for the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing. When the task offloading decision for the kth computing task is local computing: The value is 1; The mth edge computing terminal associated with the nth base station uses the base station computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is the base station computing, The value is 1; The mth edge computing terminal associated with the nth base station uses cloud computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is cloud computing, The value is 1.

[0019] Optionally, the corrected total energy consumption of each type of computing task satisfies the following expression:

[0020]

[0021] In the formula, is the total energy consumption of the kth computing task after correction corresponding to the mth edge computing terminal associated with the nth base station; The transmission energy consumption of the mth edge computing terminal associated with the nth base station using local computing to perform the kth computing task; The transmission energy consumption of the mth edge computing terminal associated with the nth base station to perform the kth computing task using base station computing or cloud computing; The task offloading variable for the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing. When the task offloading decision for the kth computing task is local computing: The value is 1; The mth edge computing terminal associated with the nth base station uses the base station computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is the base station computing, The value is 1; The mth edge computing terminal associated with the nth base station uses cloud computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is cloud computing, The value is 1.

[0022] Optionally, the total cost of each type of computing task satisfies the following expression:

[0023]

[0024] In the formula, is the total overhead of the k-th computing task of the n-th base station corresponding to the m-th edge computing terminal; and are correction coefficients, which are used to indicate the sensitivity of the k-th computing task to latency and energy consumption. and is the total delay of the kth type of computing task corresponding to the mth edge computing terminal associated with the nth base station; is the total energy consumption of the kth computing task after correction corresponding to the mth edge computing terminal associated with the nth base station.

[0025] Optionally, the determining the latency and energy consumption of executing each type of computing task by the computing mode according to the computing resources allocated to each type of computing task by the computing mode includes:

[0026] If the computing mode is local computing, the computing resources allocated to each type of computing task by the computing mode include the computing power and energy consumption parameters of the edge computing terminal, then according to the computing power and energy consumption parameters of the edge computing terminal, the attributes of each type of computing task are comprehensively considered, and the preset first expression of local computing is used to obtain the delay and energy consumption of executing each type of computing task;

[0027] If the calculation mode is base station calculation, the calculation resources allocated to each type of calculation task by the calculation mode include the calculation capability of the base station, the energy consumption parameter, and the first uplink transmission efficiency between the edge computing terminal and the base station, then according to the calculation capability of the base station, the energy consumption parameter, and the first uplink transmission efficiency, the attributes of each type of calculation task are comprehensively considered, and the preset second expression of the base station calculation is used to determine the delay and energy consumption of executing each type of calculation task;

[0028] If the computing mode is cloud computing, the computing resources allocated to each type of computing task by the computing mode include the computing power of the cloud platform, energy consumption parameters, the first uplink transmission efficiency between the edge computing terminal and the base station, and the second uplink transmission efficiency between the base station and the cloud platform. Then, based on the computing power, energy consumption parameters, the first uplink transmission efficiency, and the second uplink transmission efficiency of the cloud platform, the attributes of each type of computing task are comprehensively considered, and a third expression calculated by the preset cloud platform is used to determine the delay and energy consumption of executing each type of computing task.

[0029] Optionally, the preset first expression of the local calculation is expressed as follows:

[0030]

[0031] In the formula, The latency of the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing; Indicates the number of CPU cycles required for the k-th computing task; The computing capacity of the mth edge computing terminal associated with the nth base station; The energy consumption of the mth edge computing terminal associated with the nth base station using local computing to perform the kth computing task; κ is the energy consumption parameter.

[0032] Optionally, the acquiring of a first uplink transmission efficiency between the edge computing terminal and the base station includes:

[0033] According to the edge computing terminal and the base station, using the properties of the communication channel between the edge computing terminal and the base station, determine the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the edge computing terminal and the base station and the receiving end;

[0034] According to the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the edge computing terminal and the base station and the receiving end, a preset first uplink transmission efficiency expression is used to determine the first uplink transmission efficiency between the edge computing terminal and the base station.

[0035] Optionally, the preset first uplink transmission efficiency expression is expressed as follows:

[0036]

[0037] in, is the first uplink transmission efficiency between the mth edge computing terminal and the nth base station; B is the spectrum bandwidth of the communication channel between the mth edge computing terminal and the nth base station; is the channel gain of the communication channel between the mth edge computing terminal and the nth base station; is the transmission power between the mth edge computing terminal and the nth base station; N0 is the additive white Gaussian noise power of the edge computing terminal corresponding to the receiving end.

[0038] Optionally, acquiring the second uplink transmission efficiency between the base station and the cloud platform includes:

[0039] According to the base station and the cloud platform, using the properties of the communication channel between the base station and the cloud platform, determine the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the base station and the cloud platform;

[0040] According to the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the base station and the cloud platform and the receiving end, a preset second uplink transmission efficiency expression is used to determine the second uplink transmission efficiency between the base station and the cloud platform.

[0041] Optionally, the preset second uplink transmission efficiency expression is expressed as follows:

[0042]

[0043] In the formula, R n is the second uplink transmission efficiency between the nth base station and the cloud platform; B n is the spectrum bandwidth of the communication channel between the nth base station and the cloud platform; H n is the channel gain of the communication channel between the nth base station and the cloud platform; P n is the transmission power between the nth base station and the cloud platform; N0 is the additive white Gaussian noise power at the receiving end.

[0044] Optionally, the clustering algorithm is used to cluster multiple computing tasks accessing the edge computing terminal to obtain multiple types of computing tasks and the maximum tolerable delay of each type of computing task, including:

[0045] Determine the requirements of each computing task based on multiple computing tasks accessing the edge computing terminal and the latency attributes of each computing task;

[0046] Based on the requirements of each computing task, determine the CPU, memory, and disk space occupied by each computing task;

[0047] Based on the CPU, memory, and disk space occupied by each computing task, matrix transformation is performed on each computing task to obtain the delay attribute matrix of each computing task.

[0048] Based on the delay attribute matrix of each computing task, a clustering algorithm is used to cluster multiple computing tasks to obtain multiple categories of computing tasks and the maximum tolerable delay of each category of computing tasks.

[0049] On the other hand, the present invention proposes a power service delay optimization system, comprising:

[0050] The task clustering module is used to cluster multiple computing tasks connected to the edge computing terminal using a clustering algorithm to obtain multiple types of computing tasks and the maximum tolerable delay of each type of computing task;

[0051] A task parameter acquisition module, used to determine the latency and energy consumption of executing each type of computing task in the computing mode based on a preset computing mode and the computing resources allocated to each type of computing task in the computing mode, wherein the computing mode includes local computing, base station computing or cloud computing;

[0052] The decision determination module is used to iteratively solve the constructed optimization problem based on the latency and energy consumption of executing each type of computing task in each computing mode, and determine the business offloading decision by adopting a distributed algorithm based on the alternating direction multiplier method.

[0053] In another aspect, the present invention further provides an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus;

[0054] The memory is used to store one or more programs;

[0055] When the one or more programs are executed by the at least one processor, the service delay optimization method described in the above technical solution is implemented.

[0056] On the other hand, the present invention also provides a readable storage medium having an execution program stored thereon, and when the execution program is executed, a service delay optimization method described in the above technical solution is implemented.

[0057] Compared with the prior art, the present invention has the following beneficial effects:

[0058] The present invention provides a method for optimizing the delay of electric power services. Each received computing task is clustered by using a clustering algorithm, and the maximum tolerable delay of each type of computing task is obtained. The maximum tolerable delay of each type of computing task is used to constrain the delay optimization method of each type of computing task. The clustering algorithm is used to reduce the difference of the same type of computing tasks of different computing tasks in the same cluster, while maintaining the differentiation between computing tasks of different clusters. Therefore, under the premise of ensuring the service quality requirements of differentiated electric power services, the business delay optimization problem is converted into multiple sub-problems and solved by using a distributed algorithm based on an alternating direction multiplier method, the computing resources of the electric power communication network are reasonably used, and the probability of high computing complexity caused by a large number of connected devices is reduced, thereby improving the efficiency of solving the business delay optimization model, achieving the purpose of reducing the delay of running computing tasks, and optimizing the energy consumption required for running computing tasks, improving the reliability of running computing tasks, and ensuring the service quality of users. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 A flow chart of a method for optimizing power service delay according to the present invention;

[0060] Figure 2 A schematic diagram of a power wireless private network cloud-edge collaborative task offloading architecture constructed for a power service latency optimization method of the present invention;

[0061] Figure 3A schematic diagram of an execution flow of a method for optimizing power service delay according to the present invention;

[0062] Figure 4 It is a structural schematic diagram of a power service delay optimization system of the present invention;

[0063] Figure 5 The figure is a schematic diagram of the structure of an electronic device of the present invention. DETAILED DESCRIPTION

[0064] Embodiment 1:

[0065] like Figure 1 As shown, the present invention proposes a method for optimizing power service delay, which includes:

[0066] S1, clustering multiple computing tasks connected to the edge computing terminal using a clustering algorithm to obtain multiple types of computing tasks and the maximum tolerable delay of each type of computing task;

[0067] S2, based on a preset computing mode, determining the latency and energy consumption of executing each type of computing task according to computing resources allocated to each type of computing task by the computing mode, wherein the computing mode includes local computing, base station computing or cloud computing;

[0068] S3, based on the latency and energy consumption of executing each type of computing task in each computing mode, adopts a distributed algorithm based on the alternating direction multiplier method to iteratively solve the constructed optimization problem and determine the business offloading decision.

[0069] The present invention designs a service delay optimization method for electric power wireless private network. In view of the continuous emergence of new electric power services, the overload of the core network of the traditional electric power wireless private network and the increase in transmission delay, the mobile edge computing technology is introduced to build a cloud-edge collaborative computing unloading model. Under the premise of ensuring the service quality requirements of differentiated electric power services, the computing resources of the electric power communication network are reasonably utilized, and efficient resource allocation and computing unloading strategies are studied. This solution can effectively reduce the computing delay of tasks, reduce the energy consumption of terminals, and improve the service quality requirements of users.

[0070] Reference Figure 2 and Figure 3In order to facilitate the explanation of the power service delay optimization method provided by the present invention, a power wireless private network cloud-edge collaborative task offloading architecture is first constructed, which includes: N small base stations (SBS) and M edge computing terminals (TE), wherein each small base station is equipped with an MEC server, and each MEC server is used to provide computing offloading services for the edge computing terminal. The small base station can be the power wireless private network base station in the figure, and the edge computing terminal is the edge server in the figure. The cloud server is used to provide cloud computing services to the edge computing terminal.

[0071] N={1,...,n,...,N} represents the set of small base stations, M={1,...,m,...,M} represents the set of edge computing terminals, n ={1,...,m n ,...,M n} represents the set of edge computing terminals associated with the nth small base station; m n ∈M n Represents the mth edge computing terminal in the set of edge computing terminals associated with the nth small base station.

[0072] In some embodiments, S1 may be implemented as:

[0073] S101, determining the demand of each computing task based on multiple computing tasks accessed to the edge computing terminal and the latency attribute of each computing task;

[0074] S102, based on the requirements of each computing task, determining the CPU, memory, and disk space occupied by each computing task;

[0075] S103, comprehensively considering the CPU, memory and disk space occupied by each computing task, and performing matrix conversion on each computing task to obtain a delay attribute matrix of each computing task;

[0076] S104, clustering the multiple computing tasks using a clustering algorithm based on the delay attribute matrix of each computing task, to obtain multiple categories of computing tasks and the maximum tolerable delay of each category of computing tasks.

[0077] Each computing task is converted into a q-dimensional task attribute matrix, which is expressed as follows:

[0078] attr=[a i1 、a i2 , …, a iq ],

[0079] Among them, attr represents the q-dimensional task attribute matrix of the i-th computing task; ai1 、a i2 , …, a iq They respectively represent an attribute of the i-th computing task, such as throughput, security, latency, etc.

[0080] After the received computing task is transformed into a matrix, a task attribute matrix is ​​obtained, which is expressed as follows:

[0081]

[0082] In the formula, attrs represents the set of computing tasks; attr1 represents the first computing task; a 11 , …, a 1q Respectively represent an attribute of the first computing task; attr m represents the mth computing task; a m1 , …, a mq They respectively represent an attribute of the m-th computing task.

[0083] In order to reduce the dimension of the task attribute matrix, three latency-related features are selected to represent the requirements of computing tasks: CPU (Central Processing Unit), memory, and disk space size. CPU and memory are related to the processing time of computing tasks, and disk space size is related to the receiving time of computing tasks. The latency attribute matrix of the i-th computing task can be obtained, which is expressed as follows:

[0084] attr i =[cpu,memory,disk space],

[0085] In the formula, attr i represents the i-th computing task; cpu represents the CPU occupied by the i-th computing task; memory represents the memory occupied by the i-th computing task; disk space represents the disk space occupied by the i-th computing task.

[0086] According to the above three attribute characteristics, the tasks are divided into K categories using the K-means clustering algorithm to obtain the maximum tolerable delay of each category of tasks.

[0087] Exemplarily, in order to facilitate processing of the received multiple computing tasks, the received computing tasks are divided into K categories, where K={1, ..., k, ..., K} represents the divided K categories of computing tasks. Indicates the kth k Class computing tasks, in, represents the properties of the k-th computing task, represents the input data size of the k-th computing task, represents the number of CPU cycles required for the k-th computing task, represents the program size required for the k-th type of computing task, is the maximum tolerable delay of the k-th task.

[0088] The K-means clustering algorithm provided by the present invention is a widely used clustering algorithm. Compared with other clustering algorithms, the K-means clustering algorithm divides the data set into K clusters in an iterative manner, so that the similarity between data points within a cluster is high, while the similarity between data points between clusters is low.

[0089] The present invention classifies multiple computing tasks by using the K-means clustering algorithm, which can greatly reduce the differentiated requirements between the same type of computing tasks and ensure the differentiated requirements between different types of computing tasks. It is convenient to divide each computing task according to its own delay attribute matrix, and obtain the maximum tolerable delay of each type of computing task, so as to facilitate the subsequent use of the maximum tolerable delay of each type of computing task to constrain the subsequent solution process of each type of computing task.

[0090] The present invention comprehensively considers differentiated power business needs, unloading decisions and resource allocation. First, considering the needs of different tasks, each computing task is classified according to its characteristics, and the maximum tolerable delay of each type of task is obtained using the K-means clustering algorithm. Secondly, a task unloading and resource allocation scheme is proposed to minimize the total overhead of the system while meeting the maximum tolerable delay of the task, and a generation optimization problem is constructed. Since the constructed optimization problem is a large-scale mixed integer nonlinear programming problem, it is difficult to solve within a reasonable time. Therefore, a distributed algorithm based on the alternating direction multiplier method is used to obtain an approximate optimal solution to the problem.

[0091] Exemplarily, the computing modes preset in S2 include: local computing, base station computing and cloud computing, where local computing includes edge computing terminal operation. Local computing is to use edge computing terminals (edge ​​servers) to perform their tasks locally; base station computing includes base station operation connected to edge computing terminals, that is, edge computing terminals unload tasks to associated small base stations (power wireless private network base stations), which are operated by small base stations; cloud computing includes using cloud centers connected to edge computing terminals through base stations, that is, edge computing terminals unload tasks to associated small base stations, which then unload computing tasks to cloud servers, which are operated by cloud servers.

[0092] In order to obtain the latency and power consumption of each type of computing task in different computing modes, and then provide data support for the optimization of subsequent computing tasks, the present invention uses the orthogonal frequency division multiple access method based on the previously constructed power wireless private network cloud-edge collaborative task offloading architecture to obtain the latency and energy consumption of each type of computing task under local computing, base station computing and cloud computing. S2 provided by the present invention includes:

[0093] If the computing mode is local computing, the computing resources allocated to each type of computing task by the computing mode include the computing power and energy consumption parameters of the edge computing terminal, then according to the computing power and energy consumption parameters of the edge computing terminal, the attributes of each type of computing task are comprehensively considered, and the preset first expression of local computing is used to obtain the delay and energy consumption of executing each type of computing task;

[0094] If the calculation mode is base station calculation, the calculation resources allocated to each type of calculation task by the calculation mode include the calculation capability of the base station, the energy consumption parameter, and the first uplink transmission efficiency between the edge computing terminal and the base station, then according to the calculation capability of the base station, the energy consumption parameter, and the first uplink transmission efficiency, the attributes of each type of calculation task are comprehensively considered, and the preset second expression of the base station calculation is used to determine the delay and energy consumption of executing each type of calculation task;

[0095] If the computing mode is cloud computing, the computing resources allocated to each type of computing task by the computing mode include the computing power of the cloud platform, energy consumption parameters, the first uplink transmission efficiency between the edge computing terminal and the base station, and the second uplink transmission efficiency between the base station and the cloud platform. Then, based on the computing power, energy consumption parameters, the first uplink transmission efficiency, and the second uplink transmission efficiency of the cloud platform, the attributes of each type of computing task are comprehensively considered, and a third expression calculated by the preset cloud platform is used to determine the delay and energy consumption of executing each type of computing task.

[0096] In some embodiments, the preset first expression of the local calculation is expressed as follows:

[0097]

[0098] In the formula, The latency of the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing; Indicates the number of CPU cycles required for the k-th computing task; The computing capacity of the mth edge computing terminal associated with the nth base station; The energy consumption of the mth edge computing terminal associated with the nth base station using local computing to perform the kth computing task; k is the energy consumption parameter.

[0099] In some embodiments, the preset second expression calculated by the base station is expressed as follows:

[0100]

[0101] In the formula, The time delay of the mth edge computing terminal associated with the nth base station to perform the kth computing task using the base station calculation; Indicates the input data size of the k-th computing task; is the first uplink transmission efficiency between the mth edge computing terminal and the nth base station; Indicates the number of CPU cycles required for the k-th computing task; Allocate computing resources of the kth type of computing task to the nth base station associated with the mth edge computing terminal; The energy consumption of transmitting the kth type of computing task to the mth edge computing terminal associated with the nth base station; Transmission power between the mth edge computing terminal and the nth base station; is the transmission duration between the mth edge computing terminal and the nth base station.

[0102] In some embodiments, the acquisition of the first uplink transmission efficiency between the edge computing terminal and the base station includes:

[0103] According to the edge computing terminal and the base station, using the properties of the communication channel between the edge computing terminal and the base station, determine the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the edge computing terminal and the base station and the receiving end;

[0104] According to the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the edge computing terminal and the base station and the receiving end, a preset first uplink transmission efficiency expression is used to determine the first uplink transmission efficiency between the edge computing terminal and the base station.

[0105] In some embodiments, the preset first uplink transmission efficiency expression is expressed as follows:

[0106]

[0107] in, is the first uplink transmission efficiency between the mth edge computing terminal and the nth base station; B is the spectrum bandwidth of the communication channel between the mth edge computing terminal and the nth base station; is the channel gain of the communication channel between the mth edge computing terminal and the nth base station; is the transmission power between the mth edge computing terminal and the nth base station; N0 is the additive white Gaussian noise power of the edge computing terminal corresponding to the receiving end.

[0108] In some embodiments, the third expression of the preset cloud computing is expressed as follows:

[0109]

[0110] In the formula, The latency of the mth edge computing terminal associated with the nth base station to perform the kth computing task using cloud computing; Indicates the input data size of the k-th computing task; is the first uplink transmission efficiency between the mth edge computing terminal and the nth base station; R n is the second uplink transmission efficiency between the nth base station and the cloud platform; Indicates the number of CPU cycles required for the k-th computing task; Allocate computing resources for the kth type of computing task to the cloud platform of the nth base station associated with the mth edge computing terminal; The energy consumption of transmitting the kth type of computing task to the mth edge computing terminal associated with the nth base station; Transmission power between the mth edge computing terminal and the nth base station; is the transmission duration between the mth edge computing terminal and the nth base station.

[0111] Optionally, acquiring the second uplink transmission efficiency between the base station and the cloud platform includes:

[0112] According to the base station and the cloud platform, using the properties of the communication channel between the base station and the cloud platform, determine the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the base station and the cloud platform;

[0113] According to the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the base station and the cloud platform and the receiving end, a preset second uplink transmission efficiency expression is used to determine the second uplink transmission efficiency between the base station and the cloud platform.

[0114] Optionally, the preset second uplink transmission efficiency expression is expressed as follows:

[0115]

[0116] In the formula, R n is the second uplink transmission efficiency between the nth base station and the cloud platform; B n is the spectrum bandwidth of the communication channel between the nth base station and the cloud platform; H n is the channel gain of the communication channel between the nth base station and the cloud platform; P nis the transmission power between the nth base station and the cloud platform; N0 is the additive white Gaussian noise power at the receiving end.

[0117] After obtaining the latency and energy consumption of each type of computing task under local computing, base station computing, and cloud computing, in order to solve the service latency optimization problem, it is necessary to first construct the optimization problem.

[0118] In some embodiments, constructing the optimization problem includes:

[0119] Based on the latency and energy consumption of executing each type of computing task in each computing mode, the total overhead of each type of computing task is determined by using the service offloading variable for correction;

[0120] Based on the total overhead of each type of computing task, the optimization goal is to minimize the total overhead of multiple types of computing tasks and make each type of computing task less than its own maximum tolerable delay. Combined with the delay constraints of each type of computing task, the energy consumption constraints of each type of computing task and the computing resource constraints of each type of computing task, an optimization problem is constructed.

[0121] Exemplarily, based on the latency and energy consumption of executing each type of computing task in each computing mode, the total overhead of each type of computing task is determined by using the service offloading variable for correction, including:

[0122] Based on the latency of executing each type of computing task in each computing mode, the service offloading variable is used to make corrections to determine the corrected total latency of each type of computing task;

[0123] Based on the energy consumption of executing each type of computing task in each computing mode, the energy consumption is corrected by using the business offloading variable to determine the corrected total energy consumption of each type of computing task;

[0124] The total delay and total energy consumption of each type of computing task after correction are comprehensively considered to determine the total overhead of each type of computing task.

[0125] Exemplarily, the decision to offload a computing task is expressed as:

[0126]

[0127] In the formula, Represents the offloading decision of computing tasks; The task offloading variable representing the local computation, represents the task offloading variable calculated by the base station, represents the task offloading variable of cloud computing; when the offloading decision of the computing task is to use the edge computing terminal for local computing, then When the offloading decision of the computing task is to use small base stations for base station computing, then When the offloading decision of computing tasks is cloud computing,

[0128] In some embodiments, the modified total latency of each type of computing task satisfies the following expression:

[0129]

[0130] In the formula, is the total delay of the kth type of computing task corresponding to the mth edge computing terminal associated with the nth base station; The latency of the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing; The time delay of the mth edge computing terminal associated with the nth base station to perform the kth computing task using the base station calculation; The latency of the mth edge computing terminal associated with the nth base station to perform the kth computing task using cloud computing; The task offloading variable for the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing. When the task offloading decision for the kth computing task is local computing: The value is 1; The mth edge computing terminal associated with the nth base station uses the base station computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is the base station computing, The value is 1; The mth edge computing terminal associated with the nth base station uses cloud computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is cloud computing, The value is 1.

[0131] In some embodiments, the corrected total energy consumption of each type of computing task satisfies the following expression:

[0132]

[0133] In the formula, is the total energy consumption of the kth computing task after correction corresponding to the mth edge computing terminal associated with the nth base station; The transmission energy consumption of the mth edge computing terminal associated with the nth base station using local computing to perform the kth computing task; The transmission energy consumption of the mth edge computing terminal associated with the nth base station to perform the kth computing task using base station computing or cloud computing; The task offloading variable for the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing. When the task offloading decision for the kth computing task is local computing: The value is 1; The mth edge computing terminal associated with the nth base station uses the base station computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is the base station computing, The value is 1; The mth edge computing terminal associated with the nth base station uses cloud computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is cloud computing, The value is 1.

[0134] In some embodiments, the total overhead of each type of computing task satisfies the following expression:

[0135]

[0136] In the formula, is the total overhead of the k-th computing task of the n-th base station corresponding to the m-th edge computing terminal; and are correction coefficients, which are used to indicate the sensitivity of the k-th computing task to latency and energy consumption. and is the total delay of the kth type of computing task corresponding to the mth edge computing terminal associated with the nth base station; is the total energy consumption of the kth computing task after correction corresponding to the mth edge computing terminal associated with the nth base station.

[0137] Based on the above, the service delay optimization problem can be expressed as:

[0138]

[0139] In the formula, and

[0140] Among them, c1 indicates that the task completion time is limited; c2 indicates that the energy consumption of each edge computing terminal to complete the task cannot exceed its energy consumption budget; c3-c6 indicate the computing resources allocated to the computing task and cannot exceed the computing resources of each device. max Indicates that cloud computing provides the largest computing resources; f max It means that the base station computing provides the largest computing resources; c7 means that the task can only be executed on the edge computing terminal, base station or cloud platform; c8 means that the offloading decision is a binary variable.

[0141] Since the optimization problem is a mixed integer nonlinear programming problem, the objective function is non-convex and involves binary variables and products between different variables. To solve this problem, this application first transforms it into a convex optimization problem, and then proposes a distributed algorithm based on the alternating direction method of multipliers (ADMM) to solve the problem.

[0142] For example, first relax the binary variable X to a continuous variable and And introduce the substitution parameter and Replace the coupling term.

[0143] Therefore, the constructed service delay optimization problem P1 can be transformed into:

[0144]

[0145] In the formula, represents the transformed f; represents the transformed F; Indicates the converted c′1 represents the transformed c1; Indicates the converted c′2 represents the transformed c2; Indicates the converted c′3 represents the transformed c3; Indicates the converted c′5 represents the transformed c5; Indicates the converted c′8 represents the transformed c8.

[0146] The transformed problem P2 is a convex optimization problem, which can be solved using the classic convex optimization method. However, when a large number of devices are involved, the computational complexity of solving the problem using the convex optimization method is very high. The distributed algorithm can greatly reduce the complexity of solving the optimization problem and improve the efficiency of solving the problem by solving the optimization problem of service latency.

[0147] Exemplarily, the present invention decomposes the original problem into a series of sub-problems by adopting a distributed algorithm based on ADMM, and these sub-problems can be solved independently on each small base station.

[0148] In order to facilitate the splitting of the problem, the present invention introduces local copies of global variables. and Represents a global variable and A local copy of

[0149]

[0150] definition

[0151] The feasible solution for the local variables of the nth base station SBS n satisfies the following expression:

[0152]

[0153] In the formula, γ n represents the feasible solution of the local variables of the nth base station SBS n; Represents a local variable; Represents a local variable; Indicates the converted Indicates the converted Indicates the converted represents the task offloading variable of the local computation after conversion, represents the task offloading variable of the base station calculation after conversion, Represents the task offloading variable of cloud computing.

[0154] Therefore, the objective function of each SBS n is defined as:

[0155]

[0156] Where, ν n represents the objective function of the nth base station SBS n, represents the total cost obtained according to the feasible solution of the local variables of the nth base station SBS n.

[0157] Further transform the P2 problem into:

[0158]

[0159] Where C9 and C10 are consistency constraints, which are used to ensure that all local variables are equal to global variables.

[0160] Since problem P3 is divisible for each small base station, the distributed algorithm of ADMM can be used to solve problem P3.

[0161] ADMM solution: First, establish the augmented Lagrangian of problem P3, which satisfies the following:

[0162]

[0163] in, and is the Lagrange multiplier; is the penalty parameter, which is a constant to control the convergence speed of ADMM.

[0164] The problem is solved using a distributed algorithm based on ADMM. The iterative optimization steps are as follows:

[0165] 1. Local variable update:

[0166] By using the augmented Lagrangian of problem P3, problem P3 is transformed into N subproblems, which can be solved using a distributed method.

[0167] The objective function and feasible solution of problem P3 are convex and satisfy the following expressions:

[0168]

[0169] Where t represents the iteration index when decomposing the objective function of problem P3.

[0170] Therefore, the primal-dual interior point algorithm can be used to solve this convex optimization problem.

[0171] 2. Global variable update:

[0172] For global variables X and Update to satisfy the following expression:

[0173]

[0174] Since in the global variables X and The quadratic regularization term is added in . The above two problems are strictly convex and unconstrained, and can be solved by setting the first-order derivative to zero.

[0175] So we can get:

[0176]

[0177] 3. Update the Lagrange multiplier to satisfy the following expression:

[0178]

[0179] 4. Stopping conditions of the constructed distributed algorithm:

[0180]

[0181] Among them, ε pri and ε dual Indicates the threshold value.

[0182] 5. Restore the binary variable X to satisfy the following expression:

[0183]

[0184] In the formula, A binary variable representing the recovery, used to determine the service offloading decision.

[0185] The distributed algorithm description based on the ADMM algorithm includes the following steps:

[0186] 1. Initialize feasible solution Lagrange multiplier (μ n[0] , τ n[0] ), algorithm stopping threshold ε pri and ε dual , penalty parameter ρ and iteration index t = 0

[0187] 2. Repeat the iteration based on the above parameters;

[0188] 3. Local variables for each small base station Make updates;

[0189] 4. For the global variable X [t+1] and Make updates;

[0190] 5. For the Lagrange multiplier {μ n[t+1]} and {τ n[t+1]} to update;

[0191] 6. When the stop condition of the distributed algorithm is met, output the global variable X [t+1] and

[0192] 7. Restore the binary variable X.

[0193] The present invention provides a method for optimizing the delay of electric power services. Each received computing task is clustered by using a clustering algorithm, and the maximum tolerable delay of each type of computing task is obtained. The maximum tolerable delay of each type of computing task is used to constrain the delay optimization method of each type of computing task. The clustering algorithm is used to reduce the difference of the same type of computing tasks of different computing tasks in the same cluster, while maintaining the differentiation between computing tasks of different clusters. Therefore, under the premise of ensuring the service quality requirements of differentiated electric power services, the business delay optimization problem is converted into multiple sub-problems and solved by using a distributed algorithm based on an alternating direction multiplier method, the computing resources of the electric power communication network are reasonably used, and the probability of high computing complexity caused by a large number of connected devices is reduced, thereby improving the efficiency of solving the business delay optimization model, achieving the purpose of reducing the delay of running computing tasks, and optimizing the energy consumption required for running computing tasks, improving the reliability of running computing tasks, and ensuring the service quality of users.

[0194] There are three computing modes for the computing tasks in the method provided by the present invention. By calculating all the delays and energy consumption of the equipment and establishing a mathematical model for the problem of minimizing the total overhead of all equipment, the optimal solution for unloading decision and resource allocation in the system is solved under the premise of ensuring the differentiated power business service quality requirements, including:

[0195] Establish a cloud-edge collaborative system. The cloud-edge collaborative system corresponds to the constructed power wireless private network cloud-edge collaborative task offloading architecture, based on the computing tasks in the area and the base stations and edge computing terminals that provide edge offloading services;

[0196] Construct a task classification model to classify the received tasks using a clustering algorithm, so as to facilitate the division of each computing task according to its own delay attribute matrix and obtain the maximum tolerable delay of each type of computing task, so as to facilitate the subsequent use of the maximum tolerable delay of each type of computing task to constrain the subsequent solution process of each type of computing task;

[0197] Construct a communication model of the cloud-edge collaborative system, including the communication path between the base station associated with each edge computing terminal and the cloud server, which is used to provide the transmission path and corresponding transmission efficiency for the edge offloading service of the edge computing terminal, so as to facilitate the subsequent transmission energy consumption of the cloud server;

[0198] Construct a computing model for the cloud-edge collaborative system. The computing model of the cloud-edge collaborative system includes computing resources that can be provided by edge computing terminals, computing resources provided by base stations, and computing resources provided by cloud platforms.

[0199] Construct problem models, including constructing optimization problems based on the latency and energy consumption of computing tasks under different computing tasks, to provide task offloading decisions for computing tasks;

[0200] Transform the constructed problem into a convex optimization problem, so as to facilitate the subsequent solution;

[0201] Decomposing the constructed problem, and decomposing the transformed problem by using a distributed algorithm based on the alternating direction multiplier method, so as to improve the efficiency of obtaining the calculation results;

[0202] Propose solutions to the constructed problems, solve the decomposed problems, and use the solution results to obtain solutions to the constructed problems, thereby achieving task offloading decisions for power business.

[0203] Embodiment 2:

[0204] Reference Figure 4 The present invention based on the same inventive concept also provides a power service delay optimization system, including:

[0205] The task clustering module is used to cluster multiple computing tasks connected to the edge computing terminal using a clustering algorithm to obtain multiple types of computing tasks and the maximum tolerable delay of each type of computing task;

[0206] A task parameter acquisition module, used to determine the latency and energy consumption of executing each type of computing task in the computing mode based on a preset computing mode and the computing resources allocated to each type of computing task in the computing mode, wherein the computing mode includes local computing, base station computing or cloud computing;

[0207] The decision determination module is used to iteratively solve the constructed optimization problem based on the latency and energy consumption of executing each type of computing task in each computing mode, and determine the business offloading decision by adopting a distributed algorithm based on the alternating direction multiplier method.

[0208] Optionally, it also includes optimization model building modules, including:

[0209] A computing task total cost determination unit, used to determine the total cost of each type of computing task based on the latency and energy consumption of executing each type of computing task in each computing mode, and to make corrections using the service offloading variables;

[0210] The optimization problem construction unit is used to minimize the total overhead of multiple types of computing tasks and make each type of computing task less than its own maximum tolerable delay based on the total overhead of each type of computing task, and to construct the optimization problem by combining the delay constraints of each type of computing task, the energy consumption constraints of each type of computing task and the computing resource constraints of each type of computing task.

[0211] Optionally, the computing task total cost determination unit includes:

[0212] A delay determination subunit is used to determine the total delay of each type of computing task after the delay of executing each type of computing task in each computing mode, and to make corrections using the service offloading variables;

[0213] An energy consumption determination subunit, used to determine the total energy consumption of each type of computing task after performing each type of computing mode based on the energy consumption of each type of computing task, and to make corrections using the business offloading variables;

[0214] The total cost synthesis subunit is used to synthesize the corrected total delay and total energy consumption of each type of computing task to determine the total cost of each type of computing task.

[0215] Optionally, the corrected total delay of each type of computing task satisfies the following expression:

[0216]

[0217] In the formula, is the total delay of the kth type of computing task corresponding to the mth edge computing terminal associated with the nth base station; The latency of the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing; The time delay of the mth edge computing terminal associated with the nth base station to perform the kth computing task using the base station calculation; The latency of the mth edge computing terminal associated with the nth base station to perform the kth computing task using cloud computing; The task offloading variable for the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing. When the task offloading decision for the kth computing task is local computing: The value is 1; The mth edge computing terminal associated with the nth base station uses the base station computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is the base station computing, The value is 1; The mth edge computing terminal associated with the nth base station uses cloud computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is cloud computing, The value is 1.

[0218] Optionally, the corrected total energy consumption of each type of computing task satisfies the following expression:

[0219]

[0220] In the formula, is the total energy consumption of the kth computing task after correction corresponding to the mth edge computing terminal associated with the nth base station; The transmission energy consumption of the mth edge computing terminal associated with the nth base station using local computing to perform the kth computing task; The transmission energy consumption of the mth edge computing terminal associated with the nth base station to perform the kth computing task using base station computing or cloud computing; The task offloading variable for the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing. When the task offloading decision for the kth computing task is local computing: The value is 1; The mth edge computing terminal associated with the nth base station uses the base station computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is the base station computing, The value is 1; The mth edge computing terminal associated with the nth base station uses cloud computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is cloud computing, The value is 1.

[0221] Optionally, the total cost of each type of computing task satisfies the following expression:

[0222]

[0223] In the formula, is the total overhead of the k-th computing task of the n-th base station corresponding to the m-th edge computing terminal; and are correction coefficients, which are used to indicate the sensitivity of the k-th computing task to latency and energy consumption. and is the total delay of the kth type of computing task corresponding to the mth edge computing terminal associated with the nth base station; is the total energy consumption of the kth computing task after correction corresponding to the mth edge computing terminal associated with the nth base station.

[0224] Optionally, the task parameter acquisition module includes:

[0225] A first parameter acquisition unit is used for, if the computing mode is local computing, the computing resources allocated to each type of computing task by the computing mode include the computing power and energy consumption parameters of the edge computing terminal, then according to the computing power and energy consumption parameters of the edge computing terminal, comprehensively considering the attributes of each type of computing task, and using a preset first expression of local computing to obtain the delay and energy consumption of executing each type of computing task;

[0226] A second parameter acquisition unit is used for, if the computing mode is base station computing, the computing resources allocated to each type of computing task by the computing mode include the computing power of the base station, the energy consumption parameter, and the first uplink transmission efficiency between the edge computing terminal and the base station, then according to the computing power of the base station, the energy consumption parameter, and the first uplink transmission efficiency, the attributes of each type of computing task are comprehensively considered, and a preset second expression of base station calculation is used to determine the delay and energy consumption of executing each type of computing task;

[0227] The third parameter acquisition unit is used for, if the computing mode is cloud computing, the computing resources allocated to each type of computing task by the computing mode include the computing power of the cloud platform, energy consumption parameters, the first uplink transmission efficiency between the edge computing terminal and the base station, and the second uplink transmission efficiency between the base station and the cloud platform. Then, according to the computing power of the cloud platform, energy consumption parameters, the first uplink transmission efficiency, and the second uplink transmission efficiency, the attributes of each type of computing task are comprehensively considered, and the third expression calculated by the preset cloud platform is used to determine the delay and energy consumption of executing each type of computing task.

[0228] Optionally, the preset first expression of the local calculation is expressed as follows:

[0229]

[0230] In the formula, The latency of the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing; Indicates the number of CPU cycles required for the k-th computing task; The computing capacity of the mth edge computing terminal associated with the nth base station; The energy consumption of the mth edge computing terminal associated with the nth base station using local computing to perform the kth computing task; κ is the energy consumption parameter.

[0231] Optionally, the second parameter obtaining unit includes:

[0232] A first channel determination subunit is used to determine the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the edge computing terminal and the base station according to the edge computing terminal and the base station, using the properties of the communication channel between the edge computing terminal and the base station;

[0233] The first efficiency determination subunit is used to determine the first uplink transmission efficiency of the edge computing terminal and the base station based on the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the edge computing terminal and the base station, using a preset first uplink transmission efficiency expression.

[0234] Optionally, the preset first uplink transmission efficiency expression is expressed as follows:

[0235]

[0236] in, is the first uplink transmission efficiency between the mth edge computing terminal and the nth base station; B is the spectrum bandwidth of the communication channel between the mth edge computing terminal and the nth base station; is the channel gain of the communication channel between the mth edge computing terminal and the nth base station; is the transmission power between the mth edge computing terminal and the nth base station; N0 is the additive white Gaussian noise power of the edge computing terminal corresponding to the receiving end.

[0237] Optionally, the third parameter acquisition unit includes:

[0238] A second channel determination subunit is used to determine the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the base station and the cloud platform according to the base station and the cloud platform and by using the properties of the communication channel between the base station and the cloud platform;

[0239] The second efficiency determination subunit is used to determine the second uplink transmission efficiency between the base station and the cloud platform based on the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the base station and the cloud platform, using a preset second uplink transmission efficiency expression.

[0240] Optionally, the preset second uplink transmission efficiency expression is expressed as follows:

[0241]

[0242] In the formula, R n is the second uplink transmission efficiency between the nth base station and the cloud platform; B n is the spectrum bandwidth of the communication channel between the nth base station and the cloud platform; H n is the channel gain of the communication channel between the nth base station and the cloud platform; P n is the transmission power between the nth base station and the cloud platform; N0 is the additive white Gaussian noise power at the receiving end.

[0243] Optionally, the task clustering module includes:

[0244] A task demand determination unit, configured to determine the demand of each computing task based on multiple computing tasks accessing the edge computing terminal and the respective latency attributes of each computing task;

[0245] A task attribute determination unit is used to determine the CPU, memory and disk space occupied by each computing task based on the requirements of each computing task;

[0246] The matrix conversion unit is used to comprehensively consider the CPU, memory, and disk space occupied by each computing task, and perform matrix conversion on each computing task to obtain the delay attribute matrix of each computing task;

[0247] The task clustering unit is used to cluster multiple computing tasks based on the delay attribute matrix of each computing task and adopt a clustering algorithm to obtain multiple categories of computing tasks and the maximum tolerable delay of each category of computing tasks.

[0248] In the power business delay optimization system provided by the present invention, differentiated power business needs, unloading decisions and resource allocation are comprehensively considered. First, considering the needs of different tasks, each computing task is classified according to its characteristics, and the maximum tolerable delay of each type of task is obtained using the K-means clustering algorithm. Secondly, a task unloading and resource allocation scheme is proposed to minimize the total overhead of the system while meeting the maximum tolerable delay of the task, and a generation optimization problem is constructed. Since the constructed optimization problem is a large-scale mixed integer nonlinear programming problem, it is difficult to solve within a reasonable time. Therefore, a distributed algorithm based on the alternating direction multiplier method is used to obtain an approximate optimal solution to the proposed problem.

[0249] Embodiment 3:

[0250] like Figure 5 As shown, the present invention also provides an electronic device, which may be a computer device, a single-chip device, an intelligent mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, the processor, and the transceiver component are connected via a bus; the memory may be used to store an execution program, and an exemplary execution program may include instructions; the processor is used to execute the instructions stored in the memory. The memory may also be used to store data, which may be called and / or modified when the instructions are executed.

[0251] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), or application-specific integrated circuits.

[0252] (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, and are suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the storage medium to implement the corresponding method flow or corresponding function, so as to implement the steps of a power business delay optimization method in the above embodiment.

[0253] Embodiment 4:

[0254] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in an electronic device for storing programs and data. It can be understood that the storage medium here can include both built-in storage media in electronic devices and, of course, extended storage media supported by electronic devices. The storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by a processor are also stored in the storage space, and these instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor loads and executes one or more instructions stored in the storage medium, which can implement the steps of a method for optimizing power service delay in the above embodiment.

[0255] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented 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.

[0256] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0257] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0258] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0259] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention to be approved.

Claims

1. A method for optimizing power service delay, characterized in that: include: A clustering algorithm is used to cluster multiple computing tasks connected to edge computing terminals to obtain multiple types of computing tasks and the maximum tolerable delay of each type of computing task; Based on a preset computing mode, and according to computing resources allocated to each type of computing task by the computing mode, determining the latency and energy consumption of executing each type of computing task by the computing mode, the computing mode including local computing, base station computing or cloud computing; Based on the latency and energy consumption of executing each type of computing task in each computing mode, a distributed algorithm based on the alternating direction multiplier method is used to iteratively solve the constructed optimization problem and determine the business offloading decision.

2. The method according to claim 1, characterized in that The construction of the optimization problem includes: Based on the latency and energy consumption of executing each type of computing task in each computing mode, the total overhead of each type of computing task is determined by using the service offloading variable for correction; Based on the total overhead of each type of computing task, the optimization goal is to minimize the total overhead of multiple types of computing tasks and make each type of computing task less than its own maximum tolerable delay. Combined with the delay constraints of each type of computing task, the energy consumption constraints of each type of computing task and the computing resource constraints of each type of computing task, an optimization problem is constructed.

3. The method according to claim 2, characterized in that The latency and energy consumption of executing each type of computing task based on each computing mode are corrected by using the service offloading variable to determine the total overhead of each type of computing task, including: Based on the latency of executing each type of computing task in each computing mode, the service offloading variable is used to make corrections to determine the corrected total latency of each type of computing task; Based on the energy consumption of executing each type of computing task in each computing mode, the energy consumption is corrected by using the business offloading variable to determine the corrected total energy consumption of each type of computing task; The total delay and total energy consumption of each type of computing task after correction are comprehensively considered to determine the total overhead of each type of computing task.

4. The method according to claim 3, characterized in that The corrected total delay of each type of computing task satisfies the following expression: In the formula, is the total delay of the kth type of computing task corresponding to the mth edge computing terminal associated with the nth base station; The latency of the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing; The time delay of the mth edge computing terminal associated with the nth base station to perform the kth computing task using the base station calculation; The latency of the mth edge computing terminal associated with the nth base station to perform the kth computing task using cloud computing; The task offloading variable for the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing. When the task offloading decision for the kth computing task is local computing: The value is 1; The mth edge computing terminal associated with the nth base station uses the base station computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is the base station computing, The value is 1; The mth edge computing terminal associated with the nth base station uses cloud computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is cloud computing, The value is 1.

5. The method according to claim 3, characterized in that The total energy consumption of each type of computing task after correction satisfies the following expression: In the formula, is the total energy consumption of the kth computing task after correction corresponding to the mth edge computing terminal associated with the nth base station; The transmission energy consumption of the mth edge computing terminal associated with the nth base station using local computing to perform the kth computing task; The transmission energy consumption of the mth edge computing terminal associated with the nth base station to perform the kth computing task using base station computing or cloud computing; The task offloading variable for the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing. When the task offloading decision for the kth computing task is local computing: The value is 1; The mth edge computing terminal associated with the nth base station uses the base station computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is the base station computing, The value is 1; The mth edge computing terminal associated with the nth base station uses cloud computing to perform the task offloading variable of the kth computing task. When the task offloading decision of the kth computing task is cloud computing, The value is 1.

6. The method according to claim 3, characterized in that The total cost of each type of computing task satisfies the following expression: In the formula, is the total overhead of the k-th computing task of the n-th base station corresponding to the m-th edge computing terminal; and are correction coefficients, which are used to indicate the sensitivity of the k-th computing task to latency and energy consumption. and is the total delay of the kth type of computing task corresponding to the mth edge computing terminal associated with the nth base station; is the total energy consumption of the kth computing task after correction corresponding to the mth edge computing terminal associated with the nth base station.

7. The method according to claim 1, characterized in that The step of allocating computing resources to each type of computing task according to the computing mode and determining the latency and energy consumption of executing each type of computing task by the computing mode includes: If the computing mode is local computing, the computing resources allocated to each type of computing task by the computing mode include the computing power and energy consumption parameters of the edge computing terminal, then according to the computing power and energy consumption parameters of the edge computing terminal, the attributes of each type of computing task are comprehensively considered, and the preset first expression of local computing is used to obtain the delay and energy consumption of executing each type of computing task; If the calculation mode is base station calculation, the calculation resources allocated to each type of calculation task by the calculation mode include the calculation capability of the base station, the energy consumption parameter, and the first uplink transmission efficiency between the edge computing terminal and the base station, then according to the calculation capability of the base station, the energy consumption parameter, and the first uplink transmission efficiency, the attributes of each type of calculation task are comprehensively considered, and the preset second expression of the base station calculation is used to determine the delay and energy consumption of executing each type of calculation task; If the computing mode is cloud computing, the computing resources allocated to each type of computing task by the computing mode include the computing power of the cloud platform, energy consumption parameters, the first uplink transmission efficiency between the edge computing terminal and the base station, and the second uplink transmission efficiency between the base station and the cloud platform. Then, based on the computing power, energy consumption parameters, the first uplink transmission efficiency, and the second uplink transmission efficiency of the cloud platform, the attributes of each type of computing task are comprehensively considered, and a third expression calculated by the preset cloud platform is used to determine the delay and energy consumption of executing each type of computing task.

8. The method according to claim 7, characterized in that The first expression of the preset local calculation is expressed as follows: In the formula, The latency of the mth edge computing terminal associated with the nth base station to perform the kth computing task using local computing; Indicates the number of CPU cycles required for the k-th computing task; The computing capacity of the mth edge computing terminal associated with the nth base station; The energy consumption of the mth edge computing terminal associated with the nth base station using local computing to perform the kth computing task; κ is the energy consumption parameter.

9. The method according to claim 7, characterized in that The acquiring of a first uplink transmission efficiency between the edge computing terminal and the base station includes: According to the edge computing terminal and the base station, using the properties of the communication channel between the edge computing terminal and the base station, determine the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the edge computing terminal and the base station and the receiving end; According to the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the edge computing terminal and the base station and the receiving end, a preset first uplink transmission efficiency expression is used to determine the first uplink transmission efficiency between the edge computing terminal and the base station.

10. The method according to claim 9, characterized in that The preset first uplink transmission efficiency expression is expressed as follows: in, is the first uplink transmission efficiency between the mth edge computing terminal and the nth base station; B is the spectrum bandwidth of the communication channel between the mth edge computing terminal and the nth base station; is the channel gain of the communication channel between the mth edge computing terminal and the nth base station; is the transmission power between the mth edge computing terminal and the nth base station; N0 is the additive white Gaussian noise power of the edge computing terminal corresponding to the receiving end.

11. The method according to claim 7, characterized in that The acquiring of the second uplink transmission efficiency between the base station and the cloud platform includes: According to the base station and the cloud platform, using the properties of the communication channel between the base station and the cloud platform, determine the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the base station and the cloud platform; According to the spectrum bandwidth, channel gain, transmission power and additive white Gaussian noise power of the communication channel between the base station and the cloud platform and the receiving end, a preset second uplink transmission efficiency expression is used to determine the second uplink transmission efficiency between the base station and the cloud platform.

12. The method according to claim 11, characterized in that The preset second uplink transmission efficiency expression is expressed as follows: In the formula, R n is the second uplink transmission efficiency between the nth base station and the cloud platform; B n is the spectrum bandwidth of the communication channel between the nth base station and the cloud platform; H n is the channel gain of the communication channel between the nth base station and the cloud platform; P n is the transmission power between the nth base station and the cloud platform; N0 is the additive white Gaussian noise power at the receiving end.

13. The method according to claim 1, characterized in that The clustering algorithm is used to cluster multiple computing tasks accessing the edge computing terminal to obtain multiple types of computing tasks and the maximum tolerable delay of each type of computing task, including: Determine the requirements of each computing task based on multiple computing tasks accessing the edge computing terminal and the latency attributes of each computing task; Based on the requirements of each computing task, determine the CPU, memory, and disk space occupied by each computing task; Based on the CPU, memory, and disk space occupied by each computing task, matrix transformation is performed on each computing task to obtain the delay attribute matrix of each computing task. Based on the delay attribute matrix of each computing task, a clustering algorithm is used to cluster multiple computing tasks to obtain multiple categories of computing tasks and the maximum tolerable delay of each category of computing tasks.

14. A power service delay optimization system, characterized in that: include: The task clustering module is used to cluster multiple computing tasks connected to the edge computing terminal using a clustering algorithm to obtain multiple types of computing tasks and the maximum tolerable delay of each type of computing task; A task parameter acquisition module, used to determine the latency and energy consumption of executing each type of computing task in the computing mode according to the computing resources allocated to each type of computing task in the computing mode based on a preset computing mode, wherein the computing mode includes local computing, base station computing or cloud computing; The decision determination module is used to determine the service offloading decision by iteratively solving the constructed optimization problem based on the latency and energy consumption of executing each type of computing task in each computing mode and adopting a distributed algorithm based on the alternating direction multiplier method.

15. An electronic device, characterized in that: include: at least one processor and memory; The memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, a service delay optimization method as described in any one of claims 1 to 13 is implemented.

16. A readable storage medium, characterized in that: An execution program is stored thereon, and when the execution program is executed, a service delay optimization method as described in any one of claims 1 to 13 is implemented.

Images