A power distribution network edge computing device-oriented computing task optimization scheduling method

Abstract

A kind of computing task optimization scheduling method for power distribution network edge computing device, for the power distribution network edge computing device to be optimized, input device performance parameters, the type, characteristic parameters and execution logic sequence of the computing task to be handled, total number of computing task to be handled;Respectively, each task processing constraint is established and part of the constraint is linearized;For all types of computing task to be handled, computing state constraint is established, and the computing state constraint is linearized;Establish the execution logic associated constraint between all computing tasks to be handled;Establish the resource limit constraint of edge computing device;Establish the comprehensive objective function of computing task optimization scheduling to be handled;The comprehensive objective function of computing task optimization scheduling to be handled and each constraint are formed into power distribution network edge computing device computing task optimization scheduling model, and the model is solved using solver, and the optimization scheduling result is output.The present application realizes the full use of computing resource and the efficient, reliable operation of power distribution network edge computing device.

Description

Technical Field

[0001] This invention relates to a method for optimizing the scheduling of computing tasks. In particular, it relates to a method for optimizing the scheduling of computing tasks for edge computing devices in power distribution networks. Background Technology

[0002] With the rapid development of information and communication technologies and the widespread application of smart terminal devices in power distribution networks, edge computing technology has become an important means to cope with the explosive growth of data volume and increasingly diversified service scenarios in power distribution networks. Edge computing can push computing resources down to the vicinity of terminal devices, greatly improving the computing and storage capabilities at the edge of the power distribution network, as well as the ability to analyze and utilize massive data resources locally. It is an important supporting technology for the operation and control of future smart power distribution networks. With the increasing access of more and more sensing and measurement devices and user-side smart terminal devices, and the continuous enrichment of power distribution network business types, higher requirements are placed on the utilization of computing resources and business processing capabilities of edge computing in power distribution networks. Under the constraint of limited computing resources, edge computing devices must not only reliably complete different types of power grid business such as protection, control, metering, and monitoring, but also respond to various possible sudden computing task demands such as power grid fault handling and self-healing control at any time. At the same time, it is necessary to give full play to the synergistic effects between various businesses in terms of data flow and business flow, facing huge technical challenges. Therefore, it is necessary to study effective edge computing task optimization and scheduling methods, rationally schedule and allocate the computing and storage resources of edge computing devices to meet the diverse computing task needs of power distribution networks, maximize resource utilization, and improve the quality of power distribution network business services. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to overcome the shortcomings of the existing technology and provide a method for optimizing and scheduling computing tasks for distribution network edge computing devices, which can realize the full utilization of computing resources and the efficient and reliable operation of distribution network edge computing devices.

[0004] The technical solution adopted in this invention is: a method for optimizing and scheduling computing tasks for edge computing devices in power distribution networks, comprising the following steps:

[0005] 1) For the edge computing device of the distribution network to be optimized, input the device performance parameters, the type, characteristic parameters, and execution logic order of the computing tasks to be processed, and the total number of computing tasks to be processed N; where the device performance parameters include the maximum number of CPU cycles e per time period of the edge computing device. M Maximum memory capacity d M The preset utilization coefficients for computing resources are μ1 and storage resources are μ2. The types of computing tasks to be processed include deferred tasks, deferred and interruptible tasks, periodic tasks, non-deferred tasks, and continuous tasks. The characteristic parameters of the computing tasks to be processed include the total number of CPU cycles b required to complete each task.i Memory capacity required for task processing d i Task arrival time Latest completion time The duration T of a periodic task i The number of times a periodic task needs to be executed within the entire scheduling cycle, J. i Task discard weight factor L i , i represents the task number; set the total optimized scheduling duration to H;

[0006] 2) Based on the type and characteristic parameters of the edge computing task to be processed by the edge computing device input in step 1), establish processing constraints for each task, including processing constraints for delayed tasks, processing constraints for delayed and interruptible tasks, processing constraints for periodic tasks, processing constraints for non-delayed tasks, and processing constraints for continuous tasks. Linearize the processing constraints for delayed tasks, processing constraints for delayed and interruptible tasks, processing constraints for periodic tasks, and processing constraints for non-delayed tasks.

[0007] 3) Based on the characteristic parameters of the edge computing device to be processed input in step 1), establish computing state constraints for all types of computing tasks to be processed, and linearize the computing state constraints.

[0008] 4) Based on the execution logic order of the edge computing tasks to be processed by the edge computing device input in step 1), establish execution logic association constraints between all the computing tasks to be processed;

[0009] 5) Based on the performance parameters of the edge computing device input in step 1), establish resource constraints for the edge computing device, including computing resource constraints and storage resource constraints;

[0010] 6) Taking into account the minimum sum of completion times of all pending computing tasks, the minimum total penalty caused by the overuse of computing and storage resources exceeding the preset available value, and the minimum loss of discarded tasks, a comprehensive objective function for optimizing the scheduling of pending computing tasks is established.

[0011] 7) Combine the objective function of the optimization scheduling of the computing tasks to be processed in step 6) with the processing constraints of each task, the computing state constraints of each task, the execution logic association constraints between all computing tasks to be processed, and the resource constraint constraints of the edge computing device formed in steps 2)-5) to form the computing task optimization scheduling model of the distribution network edge computing device. Use the CPLEX solver to solve the model and output the optimization scheduling results.

[0012] This invention provides a method for optimizing the scheduling of computing tasks for edge computing devices in power distribution networks. It addresses the problem of optimizing the scheduling of computing tasks for edge computing devices in power distribution networks by classifying power distribution network services according to different characteristics and considering the temporal logical relationships between these services. A computing task optimization scheduling model for edge computing devices in power distribution networks is established. This model comprehensively considers the constraints of each task's processing, the constraints of each task's computational state, the constraints of the execution logic relationships between tasks to be processed, and the resource constraints of the edge computing device. It also comprehensively considers the objectives of minimizing the sum of the completion times of each computing task, keeping the computational and storage resource usage below preset values, and minimizing the loss from discarded tasks. A comprehensive objective function for optimizing the scheduling of computing tasks in power distribution network edge computing devices is established, forming the computing task optimization scheduling model for edge computing devices in power distribution networks. This model can be solved to obtain the optimal scheduling strategy for computing tasks in power distribution network edge computing devices. This invention rationally allocates the computing and storage resources of the distribution network edge computing device and flexibly schedules each task to be processed in time. This can meet the various requirements of various computing tasks to be processed, improve the overall completion rate of various computing tasks to be processed, and effectively improve the efficiency of the edge computing device in processing various computing tasks. This achieves full utilization of computing resources and efficient and reliable operation of the distribution network edge computing device, greatly improving the quality of distribution network business services. Attached Figure Description

[0013] Figure 1 This is a flowchart of a computing task optimization scheduling method for edge computing devices in a power distribution network according to the present invention;

[0014] Figure 2 It shows the execution status of each pending computation task at each moment under the influence of Scheme 1 and Scheme 2;

[0015] Figure 3 This shows the usage of computing resources of the power distribution network edge computing device at different times under Scheme 1 and Scheme 2.

[0016] Figure 4 It shows the storage resource usage of the power distribution network edge computing device at various times under Scheme 1 and Scheme 2. Detailed Implementation

[0017] The following describes in detail, with reference to embodiments and accompanying drawings, a method for optimizing and scheduling computing tasks for edge computing devices in power distribution networks according to the present invention.

[0018] like Figure 1 As shown, the present invention provides a method for optimizing and scheduling computing tasks for edge computing devices in power distribution networks, comprising the following steps:

[0019] 1) For the edge computing device of the distribution network to be optimized, input the device performance parameters, the type, characteristic parameters, and execution logic order of the computing tasks to be processed, and the total number of computing tasks to be processed N; where the device performance parameters include the maximum number of CPU cycles e per time period of the edge computing device. M Maximum memory capacity d M The preset utilization coefficients for computing resources are μ1 and storage resources are μ2. The types of computing tasks to be processed include deferred tasks, deferred and interruptible tasks, periodic tasks, non-deferred tasks, and continuous tasks. The characteristic parameters of the computing tasks to be processed include the total number of CPU cycles b required to complete each task. i Memory capacity required for task processing d i Task arrival time Latest completion time The period duration Ti of the periodic task and the number of times the periodic task needs to be executed within the entire scheduling cycle J. i Task discard weight factor L i , i represents the task number; set the total optimized scheduling duration to H;

[0020] 2) Based on the task types and characteristic parameters of the edge computing device to be processed input in step 1), establish processing constraints for each task, including constraints for delayed tasks, delayed and interruptible tasks, periodic tasks, non-delayed tasks, and continuous tasks. Linearize the constraints for delayed tasks, delayed and interruptible tasks, periodic tasks, and non-delayed tasks.

[0021] The aforementioned time-delayable task processing constraint is expressed as follows:

[0022]

[0023]

[0024]

[0025]

[0026]

[0027]

[0028]

[0029]

[0030]

[0031] In the formula, Ω1∈i, where Ω1 represents the set of time-delayable task numbers; t represents the optimization time period and This indicates the working status of task Ω1 during time period t. Indicates the execution of task calculations. This indicates that task calculations will not be performed. This indicates the working status of task Ω1 in time period t-1. Indicates the execution of task calculations. This indicates that task calculations will not be performed. The arrival time of task Ω1; This is the latest completion time for task Ω1; As an auxiliary variable; This represents the number of CPU cycles allocated to task Ω1 by the edge computing device during time period t; This indicates the total number of CPU cycles required for task Ω1 to complete processing.

[0032] The aforementioned delayable and interruptible task processing constraints are expressed as follows:

[0033]

[0034]

[0035]

[0036]

[0037]

[0038]

[0039]

[0040] In the formula, Ω2∈i, where Ω2 represents the set of task numbers for the delayable and interruptible class; t represents the optimization time period and This indicates the working status of task Ω2 during time period t. Indicates the execution of task calculations. This indicates that task calculations will not be performed. This indicates the working status of task Ω2 in time period t-1. Indicates the execution of task calculations. This indicates that task calculations will not be performed. Let i be the arrival time of task i; The latest completion time for task Ω2; As an auxiliary variable; This represents the number of CPU cycles allocated to task Ω2 by the edge computing device during time period t; This indicates the total number of CPU cycles required for task Ω2 to complete processing.

[0041] The periodic task processing constraints are expressed as follows:

[0042]

[0043]

[0044]

[0045] In the formula, Ω3∈i, where Ω3 represents the set of periodic task numbers; t represents the optimization period and Let Ω3 be the number of times it needs to be executed during the entire scheduling cycle, and Represents the set of all positive integers; The arrival time of task Ω3; The duration of the periodic task Ω3; This indicates the working status of task Ω3 during time period t. Indicates the execution of task calculations. This indicates that task calculations will not be performed. This represents the number of CPU cycles allocated to task i by the edge computing device during time period t; This indicates the total number of CPU cycles required to complete task Ω3.

[0046] The non-delayable task processing constraint is expressed as follows:

[0047]

[0048]

[0049]

[0050]

[0051]

[0052]

[0053]

[0054]

[0055]

[0056]

[0057] In the formula, Ω4∈i, where Ω4 represents the set of non-delayable task numbers; t represents the optimization time period and This indicates the working status of task Ω4 during time period t. Indicates the execution of task calculations. This indicates that task calculations will not be performed. This indicates the working status of task Ω4 in time period t-1. Indicates the execution of task calculations. This indicates that task calculations will not be performed. The arrival time of mission Ω4; The latest completion time for task Ω4; As an auxiliary variable; This represents the number of CPU cycles allocated to task Ω4 by the edge computing device during time period t; This indicates the total number of CPU cycles required to complete task Ω4. This indicates that task Ω4 is in Work status during a specific time period; Indicates the execution of task calculations. This indicates that task calculation will not be performed.

[0058] The continuous task processing constraints are expressed as follows:

[0059]

[0060]

[0061] In the formula, Ω5∈i, where Ω5 represents the set of continuous task numbers; t represents the optimization period and t∈[0,H]; This indicates the working status of task Ω5 during time period t. Indicates the execution of task calculations. This indicates that task calculations will not be performed. This represents the number of CPU cycles allocated to task Ω5 by the edge computing device at time t; This indicates the total number of CPU cycles required to complete task Ω5.

[0062] The linearization of constraints on delayed task processing, delayed and interruptible task processing, periodic task processing, and non-delayable task processing is specifically performed as follows:

[0063] Using auxiliary variables Replace the nonlinear constraints in formulas (7), (14), (17), and (27) of the original delayed task, delayed interruptible task, periodic task, and non-delayable task processing constraints. The terms are then supplemented with formulas (32) to (35) as auxiliary constraints after the replaced formulas (7), (14), (17), and (27):

[0064]

[0065]

[0066]

[0067]

[0068] In the formula, This indicates the task number of the four types of tasks that need to be linearized; t represents the optimization period. Indicates task Working status during time period t Indicates the execution of task calculations. This indicates that task calculations will not be performed. This indicates that the edge computing device is assigned to a task during time period t. CPU cycle count; Indicates task The total number of CPU cycles required to complete the processing.

[0069] 3) Based on the characteristic parameters of the edge computing task to be processed input in step 1), establish computing state constraints for all types of computing tasks to be processed, and linearize the computing state constraints; wherein,

[0070] The computational state constraints include:

[0071] Task discard state constraints:

[0072]

[0073]

[0074] Task i completion time constraint:

[0075]

[0076] Task i start execution time constraint:

[0077]

[0078] Among them, the auxiliary variable p i,t and q i,t It can be determined by the following formula:

[0079]

[0080]

[0081] In the formula, i represents the task number; t represents the optimization period and t∈[0,H]; l i Indicates the discard state of task i, l i =1 indicates that the task is abandoned, l i =0 indicates that the task is not abandoned; s i,t This represents the working status of task i during time period t, s i,t =1 indicates that the task calculation is performed, s i,t =0 indicates that the task calculation is not performed; r i,t This represents the storage state of task i in time period t, r i,t =1 indicates that storage resources are required, r i,t =0 indicates that no storage resources are needed; ω i,t As an auxiliary variable and ω i,t =s i,t e i,t .

[0082] The linearization of the computational state constraints is specifically expressed as follows:

[0083]

[0084]

[0085]

[0086]

[0087] In the formula, i represents the task number; t represents the optimization period and t∈[0,H]; ω i,t As an auxiliary variable, and ω i,t =s i, t e i,t ;s i,t This represents the working status of task i during time period t, s i,t =1 indicates that the task calculation is performed, s i,t =0 indicates that the task calculation is not performed; e i,t This represents the number of CPU cycles allocated to task i by the edge computing device during time period t; M indicates a value greater than 10. 5 A constant; R represents a value less than 10. -5 The constant of p; i,t q i,t It is an auxiliary variable.

[0088] 4) Based on the execution logic order of the edge computing tasks to be processed by the edge computing device input in step 1), establish execution logic association constraints between all the computing tasks to be processed;

[0089] The execution logic association constraints among all the pending computational tasks are represented as follows:

[0090]

[0091] z n,m -l n =0 (47)

[0092] z n,m ≤l n (48)

[0093] z n,m ≤l m (49)

[0094] z n,m ≥l n +l m -1 (50)

[0095] In the formula, n and m represent the numbers of the two tasks, where n is the number of the task to be executed first and m is the number of the task to be executed later. Let n be the execution completion time; The start time of task m; l n l m Let z represent the discard states of tasks n and m, respectively. n,m As an auxiliary variable, and z n,m =l n l m .

[0096] 5) Based on the performance parameters of the edge computing device input in step 1), establish resource constraints for the edge computing device, including computing resource constraints and storage resource constraints; wherein,

[0097] The aforementioned computing resource constraints are specifically in the following form:

[0098]

[0099] The aforementioned storage resource constraints are specifically in the following form:

[0100]

[0101]

[0102] In the formula, i represents the task number; t represents the optimization time period and t∈[0,H]; s i,tThis represents the working status of task i during time period t, s i,t =1 indicates that the task calculation is performed, s i,t =0 indicates that the task calculation is not performed; e i,t This represents the number of CPU cycles allocated to task i by the edge computing device during time period t; r i,t This represents the storage state of task i in time period t, r i,t =1 indicates that storage resources are required, r i,t =0 indicates that no storage resources are needed; ω i,t As an auxiliary variable, and ω i,t =s i,t e i,t ; The completion time of task i.

[0103] 6) Taking into account the objectives of minimizing the sum of completion times of all pending computational tasks, minimizing the total penalty caused by the consumption of computing and storage resources exceeding the preset available value, and minimizing the loss from discarding tasks, a comprehensive objective function for optimizing the scheduling of pending computational tasks is established; including:

[0104] Objective 1 is the sum of the completion times T of all pending computational tasks. delay The smallest, specifically expressed as:

[0105]

[0106] Objective 2 is to calculate the total penalty E caused by the computational resource usage exceeding the preset available value. C If the minimum value is achieved, then objective 2 can be expressed in the following form:

[0107]

[0108]

[0109]

[0110]

[0111] k2≤e M k1 (59)

[0112]

[0113]

[0114] k2≥0 (62)

[0115] Objective 3 is to maximize the total penalty E caused by exceeding the preset available storage resource usage. D If the minimum is reached, then the specific expression of objective 3 is as follows:

[0116]

[0117]

[0118]

[0119]

[0120] k4≤d M k3 (67)

[0121]

[0122]

[0123] k4≥0 (70)

[0124] Objective 4 is to abandon the mission and incur a loss of E. A The smallest, specifically expressed as:

[0125]

[0126] Combining objectives 1 to 4, the comprehensive objective function f for optimizing the scheduling of the computational tasks to be processed is:

[0127] minf=γ1T delay +γ2E C +γ3E D +γ4E A (72)

[0128] In the formula, i represents the task number, and t represents the optimization period and t∈[0,H]; The completion time of task i; k1, k2, k3, and k4 are intermediate variables; k1, k2, k3, and k4 are auxiliary variables. when When k1 = 0, otherwise k1 = 1. When k3 = 0, otherwise k3 = 1; γ1, γ2, γ3, and γ4 represent weighting coefficients, selected according to the importance attached to each objective; ω i,t As an auxiliary variable, and ω i,t =s i,t e i,t ;r i,t This represents the storage state of task i in time period t, r i,t =1 indicates that storage resources are required, r i,t =0 indicates that no storage resources are needed; i Indicates the task discard state, l i =1 indicates that task i, l is discarded. i =0 indicates that task i is not abandoned; Li This indicates that task i discards the weight factor; M indicates that the value is greater than 10. 5 The constant.

[0129] 7) Combine the objective function of the optimization scheduling of the computing tasks to be processed in step 6) with the processing constraints of each task, the computing state constraints of each task, the execution logic association constraints between all computing tasks to be processed, and the resource constraint constraints of the edge computing device formed in steps 2)-5) to form the computing task optimization scheduling model of the distribution network edge computing device. Use the CPLEX solver to solve the model and output the optimization scheduling results.

[0130] To fully verify the advancement of the computational task optimization scheduling method for distribution network edge computing devices proposed in this invention, this embodiment uses 25 tasks of 5 categories with execution logical order as shown in Tables 1 and 2, and compares and analyzes them using the following two schemes:

[0131] Option 1: Do not flexibly schedule the computing tasks of the edge computing device in the power distribution network; the CPU of the edge computing device always works at the maximum frequency.

[0132] Option 2: Optimize scheduling by using the computing task optimization scheduling method for edge computing devices in the power distribution network according to the present invention.

[0133] Table 3 compares the optimization results of Scheme 1 and Scheme 2. The execution status of each computational task under Scheme 1 and Scheme 2 at each time step is shown in Table 3. Figure 2 The resource usage of the distribution network edge computing device at various times under Scheme 1 and Scheme 2 is shown in the figure. Figure 3 The storage resource usage of the distribution network edge computing device at various times under Scheme 1 and Scheme 2 is shown in the figure. Figure 4 .

[0134] The computer hardware environment for performing the optimized calculations was an Intel(R) Core(TM) i7-10700 CPU with a clock speed of 2.90GHz and 16.0GB of memory; the software environment was a Windows 11 operating system.

[0135] As shown in Table 3, the task completion rate is higher and the overall completion time is earlier under Scheme 2. Figure 3 It can be seen that, under Scheme 2, the edge computing device for the power distribution network has a certain margin in computing resources to cope with emergencies, thus improving the reliability of the edge computing device. Figure 4It can be seen that Scheme 2 allows for sufficient margin in computing resources to cope with emergencies, improving the reliability of the edge computing device. However, Scheme 1 has periods exceeding the maximum storage capacity of the edge computing device, causing it to malfunction. These results demonstrate that by fully exploring the temporal and spatial transfer potential of edge computing tasks in the distribution network, a reasonable allocation of computing and storage resources can be achieved, effectively improving the efficiency of the edge computing device in processing pending tasks and ensuring its efficient and reliable operation.

[0136] Table 1. Number, type, and characteristic parameters of each computational task to be processed.

[0137]

[0138]

[0139] Table 2 shows the execution logic order of each computational task to be processed.

[0140] 4 5 5 6

[0141] Table 3 Comparison of optimization results between Scheme 1 and Scheme 2

[0142]

[0143]

Claims

1. A method for optimizing and scheduling computing tasks for edge computing devices in power distribution networks, characterized in that, Includes the following steps: 1) For the edge computing device of the distribution network to be optimized, input the device performance parameters, the type, characteristic parameters and execution logic order of the computing tasks to be processed, and the total number of computing tasks to be processed N; The device performance parameters include the maximum number of CPU cycles per time period e of the edge computing device. M Maximum memory capacity d M The preset utilization coefficients for computing resources are μ1 and storage resources are μ2. The types of computing tasks to be processed include deferred tasks, deferred and interruptible tasks, periodic tasks, non-deferred tasks, and continuous tasks. The characteristic parameters of the computing tasks to be processed include the total number of CPU cycles b required to complete each task. i Memory capacity required for task processing d i Task arrival time Latest completion time The duration T of a periodic task i The number of times a periodic task needs to be executed within the entire scheduling cycle, J. i Task discard weight factor L i , i represents the task number; set the total optimized scheduling duration to H; 2) Based on the type and characteristic parameters of the edge computing task to be processed by the edge computing device input in step 1), establish processing constraints for each task, including processing constraints for delayed tasks, processing constraints for delayed and interruptible tasks, processing constraints for periodic tasks, processing constraints for non-delayed tasks, and processing constraints for continuous tasks. Linearize the processing constraints for delayed tasks, processing constraints for delayed and interruptible tasks, processing constraints for periodic tasks, and processing constraints for non-delayed tasks. 3) Based on the characteristic parameters of the edge computing device to be processed input in step 1), establish computing state constraints for all types of computing tasks to be processed, and linearize the computing state constraints. 4) Based on the execution logic order of the edge computing tasks to be processed by the edge computing device input in step 1), establish execution logic association constraints between all the computing tasks to be processed; 5) Based on the performance parameters of the edge computing device input in step 1), establish resource constraints for the edge computing device, including computing resource constraints and storage resource constraints; 6) Taking into account the minimum sum of completion times of all pending computing tasks, the minimum total penalty caused by the overuse of computing and storage resources exceeding the preset available value, and the minimum loss of discarded tasks, a comprehensive objective function for optimizing the scheduling of pending computing tasks is established. 7) Combine the objective function of the optimization scheduling of the computing tasks to be processed in step 6) with the processing constraints of each task, the computing state constraints of each task, the execution logic association constraints between all computing tasks to be processed, and the resource constraint constraints of the edge computing device formed in steps 2)-5) to form the computing task optimization scheduling model of the distribution network edge computing device. Use the CPLEX solver to solve the model and output the optimization scheduling results.

2. The method for optimizing and scheduling computing tasks for edge computing devices in a power distribution network according to claim 1, characterized in that, The time-delayable task processing constraint mentioned in step 2) is expressed as follows: In the formula, Ω1∈i, where Ω1 represents the set of time-delayable task numbers; t represents the optimization time period and This indicates the working status of task Ω1 during time period t. Indicates the execution of task calculations. This indicates that task calculations will not be performed. This indicates the working status of task Ω1 in time period t-1. Indicates the execution of task calculations. This indicates that task calculations will not be performed. The arrival time of task Ω1; This is the latest completion time for task Ω1; As an auxiliary variable; This represents the number of CPU cycles allocated to task Ω1 by the edge computing device during time period t; This indicates the total number of CPU cycles required to complete the processing of task Ω1; The aforementioned delayable and interruptible task processing constraints are expressed as follows: In the formula, Ω2∈i, where Ω2 represents the set of task numbers for the delayable and interruptible class; t represents the optimization time period and This indicates the working status of task Ω2 during time period t. Indicates the execution of task calculations. This indicates that task calculations will not be performed. This indicates the working status of task Ω2 in time period t-1. Indicates the execution of task calculations. This indicates that task calculations will not be performed. Let i be the arrival time of task i; The latest completion time for task Ω2; As an auxiliary variable; This represents the number of CPU cycles allocated to task Ω2 by the edge computing device during time period t; This indicates the total number of CPU cycles required for task Ω2 to complete processing.

3. The method for optimizing and scheduling computing tasks for edge computing devices in a power distribution network according to claim 1, characterized in that, The periodic task processing constraint mentioned in step 2) is expressed as follows: In the formula, Ω3∈i, where Ω3 represents the set of periodic task numbers; t represents the optimization period and Let Ω3 be the number of times it needs to be executed during the entire scheduling cycle, and Represents the set of all positive integers; The arrival time of task Ω3; The duration of the periodic task Ω3; This indicates the working status of task Ω3 during time period t. Indicates the execution of task calculations. This indicates that task calculations will not be performed. This represents the number of CPU cycles allocated to task i by the edge computing device during time period t; This indicates the total number of CPU cycles required to complete task Ω3.

4. The method for optimizing and scheduling computing tasks for edge computing devices in a power distribution network according to claim 1, characterized in that, The non-delayable task processing constraint mentioned in step 2) is expressed as follows: In the formula, Ω4∈i, where Ω4 represents the set of non-delayable task numbers; t represents the optimization time period and This indicates the working status of task Ω4 during time period t. Indicates the execution of task calculations. This indicates that task calculations will not be performed. This indicates the working status of task Ω4 in time period t-1. Indicates the execution of task calculations. This indicates that task calculations will not be performed. The arrival time of mission Ω4; The latest completion time for task Ω4; As an auxiliary variable; This represents the number of CPU cycles allocated to task Ω4 by the edge computing device during time period t; This indicates the total number of CPU cycles required to complete task Ω4. This indicates that task Ω4 is in Work status during a specific time period; Indicates the execution of task calculations. This indicates that task calculation will not be performed.

5. The method for optimizing and scheduling computing tasks for edge computing devices in a power distribution network according to claim 1, characterized in that, The continuous task processing constraint mentioned in step 2) is expressed as follows: In the formula, Ω5∈i, where Ω5 represents the set of continuous task numbers; t represents the optimization period and t∈[0,H]; This indicates the working status of task Ω5 during time period t. Indicates the execution of task calculations. This indicates that task calculations will not be performed. This represents the number of CPU cycles allocated to task Ω5 by the edge computing device at time t; This indicates the total number of CPU cycles required to complete task Ω5.

6. The method for optimizing and scheduling computing tasks for edge computing devices in a power distribution network according to claim 1, characterized in that, Step 2) involves linearizing the constraints on delayed task processing, delayed and interruptible task processing, periodic task processing, and non-delayable task processing. The specific method is as follows: Using auxiliary variables Replace the nonlinear constraints in formulas (7), (14), (17), and (27) of the original delayed task, delayed interruptible task, periodic task, and non-delayable task processing constraints. The terms are then supplemented with formulas (32) to (35) as auxiliary constraints after the replaced formulas (7), (14), (17), and (27): In the formula, This indicates the task number of the four types of tasks that need to be linearized; t represents the optimization period. Indicates task Working status during time period t Indicates the execution of task calculations. This indicates that task calculations will not be performed. This indicates that the edge computing device is assigned to a task during time period t. CPU cycle count; Indicates task The total number of CPU cycles required to complete the processing.

7. The method for optimizing and scheduling computing tasks for edge computing devices in a power distribution network according to claim 1, characterized in that, The computational state constraints mentioned in step 3) include: Task discard state constraints: Task i completion time constraint: Task i start execution time constraint: Among them, the auxiliary variable p i，t and q i，t It can be determined by the following formula: In the formula, i represents the task number; t represents the optimization period and t∈[0,H]; l i Indicates the discard state of task i, l i =1 indicates that the task is abandoned, l i =0 indicates that the task is not abandoned; s i，t This represents the working status of task i during time period t, s i，t =1 indicates that the task calculation is performed, s i，t =0 indicates that the task calculation is not performed; r i,t This represents the storage state of task i in time period t, r i,t =1 indicates that storage resources are required, r i,t =0 indicates that no storage resources are needed; ω i，t As an auxiliary variable and ω i，t =s i，t e i，t ; The linearization of the computational state constraints is specifically expressed as follows: In the formula, i represents the task number; t represents the optimization period and t∈[0,H]; ω i,t As an auxiliary variable, and ω i,t =s i,t e i,t ;s i,t This represents the working status of task i during time period t, s i,t =1 indicates that the task calculation is performed, s i,t =0 indicates that the task calculation is not performed; e i,t This represents the number of CPU cycles allocated to task i by the edge computing device during time period t; M indicates a value greater than 10. 5 A constant; R represents a value less than 10. -5 The constant of p; i,t q i,t It is an auxiliary variable.

8. The method for optimizing and scheduling computing tasks for edge computing devices in a power distribution network according to claim 1, characterized in that, The execution logic association constraints among all the computational tasks to be processed mentioned in step 4) are represented as follows: z n,m -l n ＝0 (47) z n,m ≤l n (48) z n,m ≤l m (49) z n,m ≥l n +l m -1 (50) In the formula, n and m represent the numbers of the two tasks, where n is the number of the task to be executed first and m is the number of the task to be executed later. Let n be the execution completion time; The start time of task m; l n l m Let z represent the discard states of tasks n and m, respectively. n,m As an auxiliary variable, and z n,m =l n l m .

9. The method for optimizing and scheduling computing tasks for edge computing devices in a power distribution network according to claim 1, characterized in that, As described in step 5): Computational resource constraints, specifically in the form of: Storage resource constraints, specifically in the form of: In the formula, i represents the task number; t represents the optimization time period and t∈[0,H]; s i,t This represents the working status of task i during time period t, s i,t =1 indicates that the task calculation is performed, s i，t =0 indicates that the task calculation is not performed; e i，t This represents the number of CPU cycles allocated to task i by the edge computing device during time period t; r i,t This represents the storage state of task i in time period t, r i,t =1 indicates that storage resources are required, r i,t =0 indicates that no storage resources are needed; ω i,t As an auxiliary variable, and ω i,t =s i,t e i,t ; The completion time of task i.

10. A method for optimizing and scheduling computing tasks for edge computing devices in a power distribution network according to claim 1, characterized in that, Step 6) includes: Objective 1 is the sum of the completion times T of all pending computational tasks. delay The smallest, specifically expressed as: Objective 2 is to calculate the total penalty E caused by the computational resource usage exceeding the preset available value. C If the minimum value is achieved, then objective 2 can be expressed in the following form: k2≤e M k1 (59) k2≥0 (62) Objective 3 is to maximize the total penalty E caused by exceeding the preset available storage resource usage. D If the minimum is reached, then the specific expression of objective 3 is as follows: k4≤d M k3 (67) k4≥0 (70) Objective 4 is to abandon the mission and incur a loss of E. A The smallest, specifically expressed as: Combining objectives 1 to 4, the comprehensive objective function f for optimizing the scheduling of the computational tasks to be processed is: minf=γ1T delay +γ2E C +γ3E D +γ4E A (72) In the formula, i represents the task number, and t represents the optimization period and t∈[0,H]; The completion time of task i; k1, k2, k3, and k4 are intermediate variables; k1, k2, k3, and k4 are auxiliary variables. when When k1 = 0, otherwise k1 = 1. When k3 = 0, otherwise k3 = 1; γ1, γ2, γ3, and γ4 represent weighting coefficients, selected according to the importance attached to each objective; ω i,t As an auxiliary variable, and ω i,t =s i,t e i,t ;r i,t This represents the storage state of task i in time period t, r i,t =1 indicates that storage resources are required, r i,t =0 indicates that no storage resources are needed; i Indicates the task discard state, l i =1 indicates that task i, l is discarded. i =0 indicates that task i is not abandoned; L i This indicates that task i discards the weight factor; M indicates that the value is greater than 10. 5 The constant.

Images