Intelligent management scheduling method for roadside edge computing equipment

Abstract

The invention discloses an intelligent management scheduling method for roadside edge computing equipment, and is applied to the technical field of vehicle-road cooperation. Comprising the following steps: acquiring load characteristic data of all edge computing nodes; obtaining predicted load data of all edge computing nodes based on the trained neural network; screening the edge computing nodes based on the predicted load data and the computing task requirements to obtain alternative nodes; constructing a multi-objective optimization model taking minimum time delay, minimum energy consumption and maximum load balance degree as objective functions, and setting constraint conditions; and solving the multi-objective optimization model to obtain a task scheduling scheme of the edge computing node. According to the method, the load of the edge calculation nodes is predicted, the alternative nodes are screened, the multi-target optimization model is established based on the alternative nodes to solve the scheduling scheme, the calculation time delay and energy consumption of the task can be reduced as much as possible, and meanwhile the load balance degree is guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of vehicle-road collaboration technology, and more specifically to a method for intelligent management and scheduling of roadside edge computing equipment. Background Art

[0002] Manufacturing, represented by the automobile industry, is an important cornerstone of my country's economy. With the vigorous development of a new round of scientific and technological revolution and industrial transformation, automobiles are deeply integrated with technologies such as artificial intelligence, information communication, and large models. Intelligence and networking have become important directions for industrial development. The intelligent connected vehicles that have emerged will bring new impetus to the transformation and upgrading of the global automobile industry and inject powerful new momentum into the high-quality development of the industry. Edge computing also plays an important role in ensuring data security and traffic safety management. Edge computing sinks data processing and storage functions to edge devices, so that data can be processed locally, reducing dependence on the cloud. Sensitive data can be encrypted locally to reduce the risk of data being attacked or stolen by hackers during transmission. In addition, since edge computing distributes computing and storage resources on multiple edge devices, when one of the edge devices fails or the network is interrupted, other edge devices can still continue to provide services, greatly improving the reliability and stability of the system.

[0003] However, edge devices usually have limited computing and storage resources, which limits the capabilities and application scope of edge computing. The optimization of resource allocation and task scheduling of edge computing is of great significance to improving system performance and efficiency. How to effectively utilize limited resources and improve the computing and storage capabilities of edge devices is a challenge that needs to be solved. Currently, common resource allocation optimization methods and task scheduling methods have certain limitations. Therefore, how to provide an intelligent management and scheduling method for roadside edge computing equipment is a problem that technicians in this field urgently need to solve. Summary of the invention

[0004] In view of this, the present invention provides a method for intelligent management and scheduling of roadside edge computing equipment, which takes into account the predicted load and task requirements of the edge computing nodes, performs scheduling optimization with the goal of minimum delay, minimum energy consumption and maximum load balancing, and solves the problems in the prior art.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] A method for intelligent management and scheduling of roadside edge computing equipment includes the following steps:

[0007] S1. Collect load characteristic data of all edge computing nodes;

[0008] S2. Input the load characteristic data of the edge computing node and its adjacent edge computing nodes into the trained neural network to obtain the predicted load data of all edge computing nodes;

[0009] S3. Screen edge computing nodes based on predicted load data and computing task requirements to obtain candidate nodes;

[0010] S4. Construct a multi-objective optimization model with minimum delay, minimum energy consumption and maximum load balancing as objective functions;

[0011] S5. Setting constraints of the multi-objective optimization model;

[0012] S6. Solve the multi-objective optimization model and obtain the task scheduling solution for edge computing nodes.

[0013] Optionally, the load characteristic data in S1 includes: historical load data and node computing resource data, wherein the node computing resource data includes node CPU resource data, memory resource data and bandwidth resource data.

[0014] Optionally, S3 specifically includes: obtaining computing task requirements, including the latest completion time of the computing task, required computing resources, and the longest transmission distance, and excluding edge computing nodes that do not meet the computing task requirements.

[0015] Optionally, the objective function of the multi-objective optimization model in S4 is specifically:

[0016] P:min cost(X),min T(X),min LB(X)

[0017]

[0018] min cost(X) = minTP i

[0019]

[0020] In the formula, cost is the total energy consumption, T is the total time consumption, LB is the load balancing degree, X represents the sequence of edge computing nodes, N represents the total number of tasks, M represents the total number of edge computing nodes, and x i Calculate the ratio locally for task i, y i Calculate the ratio for task i node, c i Calculate the required resources for task i, f i is the computing resource of the current computing device, d i is the size of task i, s i is the communication capability, P i is the power of the current computing device, T j is the task time of the jth edge computing node, Tavg is the average task duration.

[0021] Optionally, the constraints of the multi-objective optimization model are:

[0022]

[0023] C4:f i,j >0

[0024] Where, T i The time required to complete task i is is the maximum allowed time for task i, C1 means that the time required to complete any computing task cannot exceed its maximum allowed time, cost i The energy consumption to complete task i is is the maximum allowable energy consumption of task i, C2 means that the energy consumption of completing any computing task cannot exceed its maximum allowable energy consumption, f i,j represents the computing resources of edge computing node j to execute task i, represents the maximum computing resources of edge computing node j, C3 indicates that the computing resources used by edge computing node j to execute task i cannot exceed its maximum computing resources, and C4 indicates that all edge computing nodes are assigned tasks.

[0025] Optionally, S6 is specifically:

[0026] S61, taking the computing devices in the computing task scheduling scheme as individuals in a population, and stratifying the population by a non-dominated sorting algorithm;

[0027] S62. For each value of the objective function, sort the individuals in the population in descending order and calculate the crowding distance of each individual;

[0028] S63, screen the next generation population, combine the offspring with the parent population, construct a new population through elite strategy and genetic methods of crossover and mutation, and keep the population number unchanged;

[0029] S64, repeat S63 until no new optimal solution set appears, and take the individuals in the first non-dominated layer as the optimal solution.

[0030] Optionally, the congestion distance is calculated as follows:

[0031]

[0032] In the formula, i d is the crowding distance of individual i, represents the objective function value of the i+1th individual, Represents the objective function value of the i-1th individual.

[0033] It can be seen from the above technical solution that compared with the prior art, the present invention discloses a method for intelligent management and scheduling of roadside edge computing equipment, which has the following beneficial effects: the present invention realizes real-time and efficient scheduling and management of roadside edge computing equipment based on node load calculation, makes full use of the system's idle computing resources, reduces computing latency, and improves the system's resource utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0035] Figure 1 It is a flow chart of the intelligent management and scheduling method of roadside edge computing equipment of the present invention;

[0036] Figure 2 The present invention is a flowchart for solving the multi-objective optimization model. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] The embodiment of the present invention discloses a method for intelligent management and scheduling of roadside edge computing equipment, such as Figure 1 As shown, the following steps are included:

[0039] S1. Collect load characteristic data of all edge computing nodes;

[0040] S2. Input the load characteristic data of the edge computing node and its adjacent edge computing nodes into the trained neural network to obtain the predicted load data of all edge computing nodes;

[0041] S3. Screen edge computing nodes based on predicted load data and computing task requirements to obtain candidate nodes;

[0042] S4. Construct a multi-objective optimization model with minimum delay, minimum energy consumption and maximum load balancing as objective functions;

[0043] S5. Setting constraints of the multi-objective optimization model;

[0044] S6. Solve the multi-objective optimization model and obtain the task scheduling solution for edge computing nodes.

[0045] Furthermore, the load characteristic data in S1 includes: historical load data and node computing resource data, wherein the node computing resource data includes node CPU resource data, memory resource data and bandwidth resource data.

[0046] In an embodiment of the present invention, the neural network model performs weighted summation on the features of the current edge computing node and the features of the adjacent edge nodes to calculate the predicted load value of the current edge node in a predetermined time period in the future. A forward propagation algorithm is used. For an edge computing node, the output of the next layer is obtained by weighted summation of the input features of the current edge computing node and the adjacent edge computing nodes of the current edge computing node. The weight is determined by the weight of the edge and the coefficient of the neural network model of each layer.

[0047] Furthermore, S3 is specifically: obtaining computing task requirements, including the latest completion time of the computing task, required computing resources and the longest transmission distance, and excluding edge computing nodes that do not meet the computing task requirements.

[0048] Furthermore, the objective function of the multi-objective optimization model in S4 is specifically:

[0049] P:min cost(X),min T(X),min LB(X)

[0050]

[0051] min cost(X) = minTP i

[0052]

[0053] In the formula, cost is the total energy consumption, T is the total time consumption, LB is the load balancing degree, X represents the sequence of edge computing nodes, N represents the total number of tasks, M represents the total number of edge computing nodes, and x i Calculate the ratio locally for task i, y i Calculate the ratio for task i node, c i Calculate the required resources for task i, f i is the computing resource of the current computing device, d i is the size of task i, s i is the communication capability, P i is the power of the current computing device, T j is the task time of the jth edge computing node, T avg is the average task duration.

[0054] Furthermore, the constraints of the multi-objective optimization model are as follows:

[0055]

[0056] C4:f i,j >0

[0057] Where, T i The time required to complete task i is is the maximum allowed time for task i, C1 means that the time required to complete any computing task cannot exceed its maximum allowed time, cost i The energy consumption to complete task i is is the maximum allowable energy consumption of task i, C2 means that the energy consumption of completing any computing task cannot exceed its maximum allowable energy consumption, f i,j represents the computing resources of edge computing node j to execute task i, represents the maximum computing resources of edge computing node j, C3 indicates that the computing resources used by edge computing node j to execute task i cannot exceed its maximum computing resources, and C4 indicates that all edge computing nodes are assigned tasks.

[0058] Further, such as Figure 2 As shown, S6 is specifically:

[0059] S61, taking the computing devices in the computing task scheduling scheme as individuals in a population, and stratifying the population by a non-dominated sorting algorithm;

[0060] S62. For each value of the objective function, sort the individuals in the population in descending order and calculate the crowding distance of each individual;

[0061] S63, screen the next generation population, combine the offspring with the parent population, construct a new population through elite strategy and genetic methods of crossover and mutation, and keep the population number unchanged;

[0062] S64, repeat S63 until no new optimal solution set appears, and take the individuals in the first non-dominated layer as the optimal solution.

[0063] In an embodiment of the present invention, the non-dominated sorting algorithm specifically includes:

[0064] Let i = 1, for all j = 1, 2, ..., n, and j ≠ i, compare x i and x j The dominance and non-domination relationship between them is such that if there is no individual x j Better than individual x i , then x iMark as non-dominated individuals, let i = i + 1, and repeat the search for all non-dominated individuals; the set of non-dominated individuals obtained through the above steps is the first-level non-dominated layer of the population, then ignore the marked non-dominated individuals, repeat the process of finding non-dominated individuals again, and get the second-level non-dominated layer, and so on, until the entire population is stratified.

[0065] After the non-dominated sorting, each individual has a non-dominated level, that is, the level of the non-dominated layer. After the offspring and parent populations are combined in S63, the individuals in the combined population are sorted from low to high according to the non-dominated level and put into a new collection until the population number exceeds the original population number. Individuals are selected according to the crowding distance to keep the population number unchanged.

[0066] Furthermore, the calculation of the congestion distance is as follows:

[0067]

[0068] In the formula, i d is the crowding distance of individual i, represents the objective function value of the i+1th individual, Represents the objective function value of the i-1th individual.

[0069] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0070] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for intelligent management and scheduling of roadside edge computing equipment, characterized in that: The following steps are involved: S1. Collect load characteristic data of all edge computing nodes; S2. Input the load characteristic data of the edge computing node and its adjacent edge computing nodes into the trained neural network to obtain the predicted load data of all edge computing nodes; S3. Screen edge computing nodes based on predicted load data and computing task requirements to obtain candidate nodes; S4. Construct a multi-objective optimization model with minimum delay, minimum energy consumption and maximum load balancing as objective functions; S5. Setting constraints of the multi-objective optimization model; S6. Solve the multi-objective optimization model and obtain the task scheduling solution for edge computing nodes.

2. A method for intelligent management and scheduling of roadside edge computing equipment according to claim 1, characterized in that: The load characteristic data in S1 includes: historical load data and node computing resource data, wherein the node computing resource data includes node CPU resource data, memory resource data and bandwidth resource data.

3. A method for intelligent management and scheduling of roadside edge computing equipment according to claim 1, characterized in that: S3 is specifically: obtaining computing task requirements, including the latest completion time of the computing task, required computing resources and the longest transmission distance, and excluding edge computing nodes that do not meet the computing task requirements.

4. A method for intelligent management and scheduling of roadside edge computing equipment according to claim 3, characterized in that: The objective function of the multi-objective optimization model in S4 is specifically: P:min cost(X),min T(X),min LB(X) min cost(X)=minTP i In the formula, cost is the total energy consumption, T is the total time consumption, LB is the load balancing degree, X represents the sequence of edge computing nodes, N represents the total number of tasks, M represents the total number of edge computing nodes, and x i Calculate the ratio locally for task i, y i Calculate the ratio for task i node, c i Calculate the required resources for task i, f i is the computing resource of the current computing device, d i is the size of task i, s i is the communication capability, P i is the power of the current computing device, T j is the task time of the jth edge computing node, T avg is the average task duration.

5. A method for intelligent management and scheduling of roadside edge computing equipment according to claim 4, characterized in that: The constraints of the multi-objective optimization model are as follows: Where, T i The time required to complete task i is is the maximum allowed time for task i, C1 means that the time required to complete any computing task cannot exceed its maximum allowed time, cost i The energy consumption to complete task i is is the maximum allowable energy consumption of task i, C2 means that the energy consumption of completing any computing task cannot exceed its maximum allowable energy consumption, f i,j represents the computing resources of edge computing node j to execute task i, represents the maximum computing resources of edge computing node j, C3 indicates that the computing resources used by edge computing node j to execute task i cannot exceed its maximum computing resources, and C4 indicates that all edge computing nodes are assigned tasks.

6. A method for intelligent management and scheduling of roadside edge computing equipment according to claim 5, characterized in that: S6 is specifically: S61, taking the computing devices in the computing task scheduling scheme as individuals in a population, and stratifying the population by a non-dominated sorting algorithm; S62. For each value of the objective function, sort the individuals in the population in descending order and calculate the crowding distance of each individual; S63, screen the next generation population, combine the offspring with the parent population, construct a new population through elite strategy and genetic methods of crossover and mutation, and keep the population number unchanged; S64, repeat S63 until no new optimal solution set appears, and take the individuals in the first non-dominated layer as the optimal solution.

7. A method for intelligent management and scheduling of roadside edge computing equipment according to claim 6, characterized in that: The calculation of the crowding distance is as follows: In the formula, i d is the crowding distance of individual i, represents the objective function value of the i+1th individual, Represents the objective function value of the i-1th individual.

Images