Edge computing resource collaborative allocation method and system for Internet of Things

Abstract

The invention provides an edge computing resource collaborative allocation method and system oriented to the Internet of Things, and relates to the field of data processing, and the method comprises the steps: obtaining test computing resource demand data of a plurality of Internet of Things terminals; based on the test computing resource demand data of the plurality of Internet of Things terminals, establishing an Internet of Things terminal association diagram of the plurality of Internet of Things terminals; generating an initial edge computing resource collaborative allocation scheme of the plurality of Internet of Things terminals based on the Internet of Things terminal association diagram of the plurality of Internet of Things terminals; acquiring historical computing resource demand data of a plurality of Internet of Things terminals; performing real-time optimization on the initial edge computing resource collaborative allocation scheme based on the historical computing resource demand data of the plurality of Internet of Things terminals and the Internet of Things terminal association diagram, and generating a real-time edge computing resource collaborative allocation scheme; according to the real-time edge computing resource collaborative allocation scheme, computing resource scheduling is carried out on a plurality of edge computing nodes, and the method has the advantage of realizing balanced allocation of edge computing resources.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and particularly to a method and system for collaborative allocation of edge computing resources for the Internet of Things. Background Art

[0002] Edge computing technology can make up for the uncertain latency problem caused by Internet transmission in the traditional cloud service architecture by deploying edge servers with relatively strong computing power and relatively rich resources in the access network near users at the edge of the network. Therefore, edge computing technology is regarded as a key technology in the implementation of future intelligent Internet of Things applications. However, in the actual application of edge computing, due to the relatively limited processing power and resources of edge servers, there are services that need to be allocated and scheduled for processing in actual applications.

[0003] For large-scale Internet of Things systems, the changes in the generated computing tasks and data volume are very large in terms of space and time. If a fixed Internet of Things architecture is adopted, that is, an edge node receives computing tasks and related data from a fixed number of sensors and intelligent devices, it will lead to very unbalanced loads borne by each edge node. For example, some edge nodes are idle, while the computing tasks, storage space, and communication capabilities of some edge nodes are overloaded. Therefore, how to reasonably schedule computing tasks and their related data and evenly allocate computing tasks to matching edge nodes is an urgent problem to be solved at present.

[0004] Therefore, there is a need to provide a method and system for collaborative allocation of edge computing resources for the Internet of Things to achieve balanced allocation of edge computing resources. Summary of the Invention

[0005] The present invention provides a method for collaborative allocation of edge computing resources for the Internet of Things, including: obtaining test computing resource requirement data of multiple Internet of Things terminals, where the test computing resource requirement data of the Internet of Things terminals includes the requirements for various computing resources at multiple test time points; establishing an Internet of Things terminal association graph of multiple Internet of Things terminals based on the test computing resource requirement data of the multiple Internet of Things terminals; generating an initial edge computing resource collaborative allocation plan for the multiple Internet of Things terminals based on the Internet of Things terminal association graph of the multiple Internet of Things terminals; obtaining historical computing resource requirement data of the multiple Internet of Things terminals; optimizing the initial edge computing resource collaborative allocation plan in real time based on the historical computing resource requirement data of the multiple Internet of Things terminals and the Internet of Things terminal association graph to generate a real-time edge computing resource collaborative allocation plan; and performing computing resource scheduling on multiple edge computing nodes according to the real-time edge computing resource collaborative allocation plan.

[0006] Further, based on the test computing resource requirement data of multiple Internet of Things (IoT) terminals, an IoT terminal association graph of multiple IoT terminals is established, including: for each type of computing resource, based on the test computing resource requirement data of multiple IoT terminals, calculate the requirement similarity of the computing resource requirements corresponding to any two IoT terminals; based on the requirement similarity of the computing resources corresponding to any two IoT terminals, determine the first similar IoT terminal of the computing resource corresponding to each IoT terminal; based on the first similar IoT terminal of the computing resource corresponding to each IoT terminal, establish the first IoT terminal association sub-graph of the computing resources corresponding to multiple IoT terminals; for each IoT terminal, based on the test computing resource requirement data of the IoT terminal, calculate the requirement correlation coefficient of any two types of computing resources corresponding to the IoT terminal; based on the requirement correlation coefficient of any two types of computing resources corresponding to each IoT terminal, calculate the requirement correlation similarity of any two IoT terminals; based on the requirement correlation similarity of any two IoT terminals, determine the second similar IoT terminal of each IoT terminal; based on the second similar IoT terminal of each IoT terminal, establish the second IoT terminal association sub-graph of multiple IoT terminals, where the IoT terminal association graph includes the first IoT terminal association sub-graph and the second IoT terminal association sub-graph corresponding to the computing resource requirement of each type.

[0007] Further, based on the test computing resource requirement data of multiple IoT terminals, calculating the requirement similarity of the computing resource requirements corresponding to any two IoT terminals includes: for each IoT terminal, perform time-varying filtering empirical mode decomposition on the requirements for computing resources of the IoT terminal at multiple test time points to obtain multiple target intrinsic mode functions of the computing resource corresponding to the IoT terminal, and extract the function features of each target intrinsic mode function; for any two IoT terminals, based on the function features of each target intrinsic mode function of the computing resources corresponding to the two IoT terminals, calculate the requirement similarity of the computing resource requirements corresponding to the two IoT terminals.

[0008] Further, based on the computing resource requirement association graph of multiple IoT terminals, an initial edge computing resource collaborative allocation scheme for multiple IoT terminals is generated, including: establishing a sample database, where the sample database is used to store the computing resource requirement characteristics of multiple sample IoT terminal groups, the first IoT terminal association sub-graph and the second IoT terminal association sub-graph corresponding to each type of computing resource, and the edge computing resource collaborative allocation scheme; based on the computing resource requirement association graph of multiple IoT terminals, determine the similar sample IoT terminal groups from multiple sample IoT terminal groups; based on the edge computing resource collaborative allocation scheme of the similar sample IoT terminal groups, generate an initial edge computing resource collaborative allocation scheme for multiple IoT terminals.

[0009] Further, based on the computational resource demand correlation graph of multiple Internet of Things (IoT) terminals, determining a similar sample IoT terminal group from multiple sample IoT terminal groups includes: for each first IoT terminal association sub-graph of the multiple IoT terminals, extracting the global features and local features of the first IoT terminal association sub-graph; extracting the global features and local node features of the second IoT terminal association sub-graph of the multiple IoT terminals; and determining a similar sample IoT terminal group from the multiple sample IoT terminal groups based on the computational resource demand features of the multiple IoT terminals, each first IoT terminal association sub-graph, and the global features and local node features of the second IoT terminal association sub-graph.

[0010] Further, based on the historical computational resource demand data and the computational resource demand correlation graph of multiple IoT terminals, performing real-time optimization on an initial edge computing resource collaborative allocation scheme to generate a real-time edge computing resource collaborative allocation scheme, including: predicting the future computational resource demand data of the multiple IoT terminals based on the historical computational resource demand data and the computational resource demand correlation graph of the multiple IoT terminals; and performing real-time optimization on the initial edge computing resource collaborative allocation scheme based on the future computational resource demand data and the computational resource demand correlation graph of the multiple IoT terminals to generate a real-time edge computing resource collaborative allocation scheme.

[0011] Further, predicting the future computational resource demand data of multiple IoT terminals based on the historical computational resource demand data and the computational resource demand correlation graph of the multiple IoT terminals includes: for each IoT terminal, predicting the initial future computational resource demand data of the IoT terminal based on the historical computational resource demand data of the IoT terminal, where the initial future computational resource demand data of the IoT terminal includes the initial demands of the IoT terminal for multiple types of computational resources at multiple future time points; for each type of computational resource, performing a first iterative correction on the initial demands of the IoT terminal for the computational resource at multiple future time points based on the initial demands of the first similar IoT terminal corresponding to the computational resource of the IoT terminal at multiple future time points to generate the corrected demands of the IoT terminal for the computational resource at multiple future time points; and for each IoT terminal, performing a second iterative correction on the corrected demands of the IoT terminal for the computational resource at multiple future time points based on the corrected demands of the second similar IoT terminal of the IoT terminal for each type of computational resource at multiple future time points to generate the future computational resource demand data of the IoT terminal.

[0012] Further, based on the future computing resource requirement data of multiple Internet of Things (IoT) terminals and the computing resource requirement correlation graph, the initial edge computing resource collaborative allocation scheme is optimized in real time to generate a real-time edge computing resource collaborative allocation scheme, including: determining the IoT terminals to be allocated and the edge computing nodes to be allocated based on the future computing resource requirement data of multiple IoT terminals and the initial edge computing resource collaborative allocation scheme; generating multiple candidate edge computing resource collaborative allocation schemes based on the IoT terminals to be allocated and the edge computing nodes to be allocated through a Monte Carlo model; establishing an optimization evaluation function; and generating a real-time edge computing resource collaborative allocation scheme based on the computing resource requirement correlation graph, the optimization evaluation function, and the multiple candidate edge computing resource collaborative allocation schemes.

[0013] Further, based on the computing resource requirement correlation graph, the optimization evaluation function, and the multiple candidate edge computing resource collaborative allocation schemes, a real-time edge computing resource collaborative allocation scheme is generated, including: calculating the optimization value of each candidate edge computing resource collaborative allocation scheme based on the computing resource requirement correlation graph and the optimization evaluation function; and generating a real-time edge computing resource collaborative allocation scheme based on the optimization value of each candidate edge computing resource collaborative allocation scheme through a genetic algorithm.

[0014] The present invention provides an edge computing resource collaborative allocation system for the Internet of Things, which is used for the edge computing resource collaborative allocation method for the Internet of Things described above, including: a requirement testing module for obtaining the test computing resource requirement data of multiple IoT terminals, where the test computing resource requirement data of the IoT terminals includes the requirements for various computing resources at multiple test time points; an association establishment module for establishing an IoT terminal association graph of multiple IoT terminals based on the test computing resource requirement data of multiple IoT terminals; an initial allocation module for generating an initial edge computing resource collaborative allocation scheme for multiple IoT terminals based on the IoT terminal association graph of multiple IoT terminals; a requirement acquisition module for obtaining the historical computing resource requirement data of multiple IoT terminals; a real-time optimization module for optimizing the initial edge computing resource collaborative allocation scheme in real time based on the historical computing resource requirement data of multiple IoT terminals and the IoT terminal association graph to generate a real-time edge computing resource collaborative allocation scheme; and a collaborative allocation module for scheduling the computing resources of multiple edge computing nodes according to the real-time edge computing resource collaborative allocation scheme.

[0015] Compared with the prior art, the edge computing resource collaborative allocation method and system provided by the present invention have at least the following beneficial effects: 1. By obtaining and analyzing the test computing resource requirement data of multiple IoT terminals, the actual resource requirements of IoT terminals can be predicted and met more accurately. This helps to avoid over-allocation or under-allocation of resources, thereby improving the overall utilization rate of edge computing resources. Generate an initial resource allocation plan based on the IoT terminal association graph and optimize the plan in real time to match the historical data. This real-time nature enables edge computing to respond more quickly to changes in the resource requirements of IoT terminals, improving the response speed and flexibility of edge computing. By comprehensively considering the test and historical computing resource requirement data of IoT terminals, a more scientific and reasonable resource allocation strategy can be formulated. This helps to balance the resource requirements among different IoT terminals and achieve fair and efficient allocation of resources. Optimizing the resource allocation plan in real time can ensure the stable operation of edge computing nodes even when the computing resource requirements fluctuate. This helps to reduce the risk of edge computing crashes or performance degradation caused by insufficient or excessive resources. 2. By performing time-varying filtering empirical mode decomposition on the computing resource requirements of IoT terminals at multiple test time points and extracting function features, the demand similarity of any two IoT terminals for the same computing resource can be accurately calculated. This helps to identify terminals with similar resource requirements, thereby more precisely matching and allocating resources. Calculating the demand correlation coefficient of IoT terminals for any two types of computing resources can reveal the correlation between the terminals' different resource requirements. This helps to consider the diversified requirements of terminals during resource allocation and achieve comprehensive optimization of resources. By constructing the first IoT terminal association sub-graph and the second IoT terminal association sub-graph, the correlation of IoT terminals in computing resource requirements can be comprehensively reflected. This provides a strong basis for collaborative resource allocation and helps to achieve optimized configuration and efficient utilization of resources. Using the historical data and experience in the sample database, an initial collaborative edge computing resource allocation plan for multiple IoT terminals can be quickly generated. This method not only improves the efficiency of resource allocation but also reduces the cost of manual intervention.

[0016] 3. Real-time optimization based on the future computing resource demand data of IoT terminals can more accurately predict and meet the future resource demands of terminals. This helps avoid over-allocation or under-allocation of resources and ensures the efficient use of resources. Generating multiple candidate edge computing resource collaborative allocation schemes through the Monte Carlo model provides the system with multiple possible resource allocation strategies. This helps the system quickly adjust the allocation scheme when the resource demand changes to adapt to the new demand. By establishing an optimization evaluation function, each candidate resource allocation scheme can be evaluated, and thus the optimal allocation scheme can be selected. This helps ensure the rationality and efficiency of resource allocation. Using a genetic algorithm to optimize the candidate edge computing resource collaborative allocation scheme can find a resource allocation strategy closer to the optimal solution. As an optimization algorithm that simulates the biological evolution process, the genetic algorithm has the characteristics of strong global search ability and strong adaptability, and can find a better solution in complex resource allocation problems. It can optimize the resource allocation scheme in real time and has a powerful dynamic adjustment ability. This helps the system quickly respond when facing changes in the resource demands of IoT terminals and ensures the stability and performance of the system. Through more precise real-time resource allocation and efficient optimization algorithms, the utilization rate of edge computing resources can be significantly improved. This helps reduce resource waste and lower operating costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] This specification will be further described in the form of exemplary embodiments, and these exemplary embodiments will be described in detail through the drawings. These embodiments are not restrictive. In these embodiments, the same numbers represent the same structures, where: Figure 1 is a schematic flowchart of a method for collaborative allocation of edge computing resources for the Internet of Things according to some embodiments of this specification; Figure 2 is a schematic diagram of a first IoT terminal association subgraph according to some embodiments of this specification; Figure 3 is a schematic diagram of the modules of a system for collaborative allocation of edge computing resources for the Internet of Things according to some embodiments of this specification. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] To more clearly illustrate the technical solutions of the embodiments of this specification, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some examples or embodiments of this specification. For those of ordinary skill in the art, without creative efforts, this specification can also be applied to other similar scenarios according to these drawings. Unless obvious from the language context or otherwise stated, the same reference numerals in the figures represent the same structures or operations.

[0019] Figure 1It is a schematic flowchart of an edge computing resource collaborative allocation method for the Internet of Things shown in some embodiments of this specification. As Figure 1 shown, the edge computing resource collaborative allocation method for the Internet of Things may include the following processes.

[0020] S101. Obtain the test computing resource requirement data of multiple Internet of Things terminals.

[0021] Among them, the test computing resource requirement data of the Internet of Things terminals includes the requirements for various computing resources (such as CPU, memory, storage, bandwidth, GPU, etc.) at multiple test time points.

[0022] It can be understood that when, in the same edge computing node, the data collected by multiple Internet of Things terminals connected to it has a processing preference for applications of the same type of computing resource, one type of resource in this edge computing node will be rapidly consumed, while there are a large number of idle and unoccupied other types of computing resources. However, due to the exhaustion of this type of computing resource, it cannot be utilized by other applications, resulting in the inability to utilize the idle other types of resources as well, causing a low resource utilization rate.

[0023] S102. Based on the test computing resource requirement data of multiple Internet of Things terminals, establish an Internet of Things terminal association graph for multiple Internet of Things terminals.

[0024] In some embodiments, S102 specifically includes: For each type of computing resource, based on the test computing resource requirement data of multiple Internet of Things terminals, calculate the requirement similarity of the corresponding computing resource requirements of any two Internet of Things terminals. Based on the requirement similarity of the corresponding computing resource requirements of any two Internet of Things terminals, determine the first similar Internet of Things terminal corresponding to each Internet of Things terminal for the corresponding computing resource. Based on the first similar Internet of Things terminal corresponding to each Internet of Things terminal for the corresponding computing resource, establish a first Internet of Things terminal association sub-graph for the corresponding computing resource of multiple Internet of Things terminals; For each Internet of Things terminal, based on the test computing resource requirement data of the Internet of Things terminal, calculate the requirement correlation coefficient between any two corresponding computing resources of the Internet of Things terminal; Based on the requirement correlation coefficient between any two corresponding computing resources of each Internet of Things terminal, calculate the requirement correlation similarity between any two Internet of Things terminals; Based on the requirement correlation similarity between any two Internet of Things terminals, determine the second similar Internet of Things terminal corresponding to each Internet of Things terminal. Based on the second similar Internet of Things terminal corresponding to each Internet of Things terminal, establish a second Internet of Things terminal association sub-graph for multiple Internet of Things terminals, where the Internet of Things terminal association graph includes the first Internet of Things terminal association sub-graph and the second Internet of Things terminal association sub-graph corresponding to the requirement of each type of computing resource.

[0025] In some embodiments, based on the test computing resource requirement data of multiple Internet of Things (IoT) terminals, calculating the requirement similarity of the corresponding computing resources between any two IoT terminals includes: For each IoT terminal, perform time-varying filtering empirical mode decomposition on the computing resource requirements of the IoT terminal at multiple test time points to obtain multiple target intrinsic mode functions of the computing resources corresponding to the IoT terminal, and extract the function characteristics of each target intrinsic mode function; For any two IoT terminals, based on the function characteristics of each target intrinsic mode function of the computing resources corresponding to the two IoT terminals, calculate the requirement similarity of the corresponding computing resources between the two IoT terminals.

[0026] Specifically, the multiple target intrinsic mode functions can be the first few intrinsic mode functions obtained according to the decomposition order. For example, the first three intrinsic mode functions. The function characteristics of the target intrinsic mode functions can at least include variance, amplitude, phase, instantaneous frequency, spectral characteristics, bandwidth, etc.

[0027] The requirement similarity of the corresponding computing resources between two IoT terminals can be calculated according to the following formula:

[0028] Where, is the requirement similarity of the e-th type of computing resource requirement corresponding to the i-th IoT terminal and the j-th IoT terminal, N is the total number of target intrinsic mode functions of the computing resources corresponding to an IoT terminal, is the function characteristic of the n-th target intrinsic mode function of the e-th type of computing resource requirement corresponding to the i-th IoT terminal, is the function characteristic of the n-th target intrinsic mode function of the e-th type of computing resource requirement corresponding to the j-th IoT terminal, is the cosine similarity between the function characteristic of the n-th target intrinsic mode function of the e-th type of computing resource requirement corresponding to the i-th IoT terminal and the function characteristic of the n-th target intrinsic mode function of the e-th type of computing resource requirement corresponding to the j-th IoT terminal.

[0029] Figure 2 is a schematic diagram of the first IoT terminal association subgraph shown in some embodiments of this specification. As Figure 2 shown, the first IoT terminal association subgraph includes nodes representing multiple IoT terminals. When the requirement similarity between two IoT terminals is greater than the requirement similarity threshold, the two nodes representing the two IoT terminals in the first IoT terminal association subgraph are connected by an edge, and the length of the edge represents the requirement similarity between the two IoT terminals. The greater the requirement similarity, the shorter the edge.

[0030] For each Internet of Things (IoT) terminal, based on the test computing resource requirement data of the IoT terminal, calculate the Pearson correlation coefficient of any two types of computing resources corresponding to the IoT terminal as the requirement correlation coefficient.

[0031] The requirement correlation similarity between two IoT terminals can be calculated according to the following formula:

[0032] where, is the requirement correlation similarity between the i-th IoT terminal and the j-th IoT terminal, is the requirement correlation coefficient of any e-th and f-th types of computing resources corresponding to the i-th IoT terminal, is the requirement correlation coefficient of any e-th and f-th types of computing resources corresponding to the j-th IoT terminal, and E is the total number of types of computing resources.

[0033] The second IoT terminal association subgraph includes nodes representing multiple IoT terminals. When the requirement correlation similarity between two IoT terminals is greater than the requirement correlation similarity threshold, two nodes representing the two IoT terminals in the first IoT terminal association subgraph are connected by an edge, and the length of the edge represents the requirement correlation similarity between the two IoT terminals. The greater the requirement correlation similarity, the shorter the edge.

[0034] S103. Generate an initial edge computing resource collaborative allocation plan for multiple IoT terminals based on the IoT terminal association graph of the multiple IoT terminals.

[0035] In some embodiments, S103 specifically includes: Establish a sample database, where the sample database is used to store the computing resource requirement characteristics of multiple sample IoT terminal groups, the first IoT terminal association subgraph and the second IoT terminal association subgraph corresponding to each type of computing resource, and the edge computing resource collaborative allocation plan. The edge computing resource collaborative allocation plan of the sample IoT terminal group can be determined through manual or experimental data; Based on the computing resource requirement association graph of the multiple IoT terminals, determine similar sample IoT terminal groups from the multiple sample IoT terminal groups; Generate an initial edge computing resource collaborative allocation plan for the multiple IoT terminals based on the edge computing resource collaborative allocation plan of the similar sample IoT terminal groups.

[0036] In some embodiments, based on the computing resource requirement association graph of the multiple IoT terminals, determining similar sample IoT terminal groups from the multiple sample IoT terminal groups includes: For each first Internet of Things (IoT) terminal associated subgraph among multiple IoT terminals, extract the global features (e.g., size, density, global clustering coefficient, etc.) and local features (e.g., degree of each node, mean of the edges of each node, local clustering coefficient, etc.) of the first IoT terminal associated subgraph; Extract the global features (e.g., size, density, global clustering coefficient, etc.) and local node features (e.g., degree of each node, mean of the edges of each node, local clustering coefficient, etc.) of the second IoT terminal associated subgraph of multiple IoT terminals; Based on the computing resource requirement characteristics of multiple IoT terminals, and the global features and local node features of each first IoT terminal associated subgraph and the second IoT terminal associated subgraph, determine a similar sample IoT terminal group from multiple sample IoT terminal groups.

[0037] Specifically, in the global features, the size represents the number of nodes in the IoT terminal associated subgraph, which can reflect the complexity and information volume of the IoT terminal associated subgraph. The density represents the ratio of the number of edges to the number of nodes in the IoT terminal associated subgraph. An IoT terminal associated subgraph with a high density may indicate a strong mutual association between nodes. The global clustering coefficient can measure the connection tightness between the neighbor nodes of all nodes in the IoT terminal associated subgraph. The number of all triples (structures composed of three nodes) in the IoT terminal associated subgraph can be counted, and the number of closed triples (i.e., all three nodes are interconnected) among them can be counted. The global clustering coefficient is equal to the number of closed triples divided by the total number of triples.

[0038] In the local node features, the degree of each node represents the number of edges connecting the node to other nodes, which can reflect the activity level of the node in the IoT terminal associated subgraph. The mean of the edges of each node can be the mean of the lengths of the edges connecting the node to other nodes. The local clustering coefficient can measure the connection tightness between the neighbor nodes of a single node in the IoT terminal associated subgraph. For a node, find all its neighbor nodes and count the number of edges actually existing among these neighbor nodes. The local clustering coefficient is equal to the number of actually existing edges divided by the maximum number of edges that may exist among the neighbor nodes. For example, if there is an edge between node A and node B, then node A is a neighbor node of node B, and at the same time node B is also a neighbor node of node A. The number of neighbor nodes of node A, k = 4. The maximum number of edges that may exist among the neighbor nodes of node A is 4(4 - 1) / 2 = 6. If there are edges connecting node B and node C, node C and node D, and node D and node E respectively, then the number of actually existing edges is 3. Therefore, the local clustering coefficient of node A is 3 / 6 = 0.5.

[0039] The computational resource requirement characteristics of the Internet of Things terminals may include the mean and variance of the requirements of each Internet of Things terminal for each computational resource, which can be calculated based on the requirements of the Internet of Things terminals for computational resources at multiple test time points.

[0040] Quantify the characteristics of each first Internet of Things terminal associated subgraph and the characteristics of the second Internet of Things terminal associated subgraph that are extracted, to form a feature vector, which includes the characteristics of each first Internet of Things terminal associated subgraph and the global characteristics, local characteristics, and computational resource requirement characteristics of the second Internet of Things terminal associated subgraph.

[0041] For each sample Internet of Things terminal group, the characteristics extracted from each first Internet of Things terminal associated subgraph and the second Internet of Things terminal associated subgraph of the sample Internet of Things terminal group can be quantified to form the feature vector of the sample Internet of Things terminal group.

[0042] The Euclidean distance between the feature vectors of multiple Internet of Things terminals and the feature vector of any one sample Internet of Things terminal group can be calculated, and the sample Internet of Things terminal group with the Euclidean distance less than the Euclidean distance threshold is used as the similar sample Internet of Things terminal group.

[0043] Based on the feature vectors of multiple Internet of Things terminals, the feature vectors of the similar sample Internet of Things terminal groups, and the edge computing resource collaborative allocation scheme of the similar sample Internet of Things terminal groups, an initial edge computing resource collaborative allocation scheme for multiple Internet of Things terminals can be generated through a scheme generation model, where the scheme generation model can be a Feedforward Neural Network model.

[0044] S104. Obtain the historical computational resource requirement data of multiple Internet of Things terminals.

[0045] Specifically, the historical computational resource requirement data of the Internet of Things terminals may include the requirements of the Internet of Things terminals for various computational resources (such as CPU, memory, storage, bandwidth, GPU, etc.) at multiple historical time points in the current period (such as one day, one week, etc.).

[0046] S105. Based on the historical computational resource requirement data of multiple Internet of Things terminals and the Internet of Things terminal association graph, perform real-time optimization on the initial edge computing resource collaborative allocation scheme to generate a real-time edge computing resource collaborative allocation scheme.

[0047] In some embodiments, S105 specifically includes: Based on the historical computational resource requirement data of multiple Internet of Things terminals and the computational resource requirement association graph, predict the future computational resource requirement data of multiple Internet of Things terminals; Based on the future computing resource demand data of multiple Internet of Things (IoT) terminals and the computing resource demand correlation graph, the initial edge computing resource collaborative allocation scheme is optimized in real time to generate a real-time edge computing resource collaborative allocation scheme.

[0048] In some embodiments, based on the historical computing resource demand data of multiple IoT terminals and the computing resource demand correlation graph, the future computing resource demand data of multiple IoT terminals is predicted, including: For each IoT terminal, based on the historical computing resource demand data of the IoT terminal, the initial future computing resource demand data of the IoT terminal is predicted, where the initial future computing resource demand data of the IoT terminal includes the initial demands of the IoT terminal for multiple types of computing resources at multiple future time points; For each type of computing resource, based on the initial demands of the first similar IoT terminals corresponding to the computing resource of the IoT terminal for the computing resource at multiple future time points, the initial demands of the IoT terminal for the computing resource at multiple future time points are iteratively corrected for the first time to generate the corrected demands of the IoT terminal for the computing resource at multiple future time points; For each IoT terminal, based on the corrected demands of the second similar IoT terminals of the IoT terminal for each type of computing resource at multiple future time points, the corrected demands of the IoT terminal for the computing resource at multiple future time points are iteratively corrected for the second time to generate the future computing resource demand data of the IoT terminal.

[0049] Specifically, for each IoT terminal, a demand prediction model corresponding to the IoT terminal can be established, and through the demand prediction model corresponding to the IoT terminal, based on the historical computing resource demand data of the IoT terminal, the initial future computing resource demand data of the IoT terminal is predicted, where the demand prediction model can be a long short-term memory network model.

[0050] The initial demands of the IoT terminal for the computing resource at multiple future time points can be iteratively corrected for the first time according to the following process: S1051. Establish a global correction model corresponding to the computing resource, where the global correction model can be a long short-term memory network model; S1052. Through the global correction model corresponding to the computing resource, based on the initial demands of the first similar IoT terminals corresponding to the computing resource of each IoT terminal for the computing resource at multiple future time points, the initial demands of each IoT terminal for the computing resource at multiple future time points are corrected for the first round to generate the demands of each IoT terminal for the computing resource at multiple future time points after the first-round correction; S1053. Calculate the global correction difference based on the difference between the initial demand for computing resources of each IoT terminal at multiple future time points and the demand for computing resources of each IoT terminal at multiple future time points after the first round of correction. S1054. Determine whether the global correction difference is less than the global correction difference threshold. If so, complete the first iterative correction of the corresponding computing resources, perform the first iterative correction of the next type of computing resources, and execute S1051. If not, execute S1055. S1055. Based on the demand for computing resources of the first similar IoT terminals corresponding to each IoT terminal at multiple future time points after the (t - 1)-th round of correction through the global correction model corresponding to the computing resources, perform the t-th round of correction on the demand for computing resources of each IoT terminal at multiple future time points after the (t - 1)-th round of correction, and generate the demand for computing resources of each IoT terminal at multiple future time points after the t-th round of correction. S1056. Calculate the global correction difference based on the difference between the demand for computing resources of each IoT terminal at multiple future time points after the t-th round of correction and the demand for computing resources of each IoT terminal at multiple future time points after the (t - 1)-th round of correction, and execute S1054.

[0051] For each IoT terminal, establish a corresponding associated correction model. Based on the correction demand for each type of computing resources of the second similar IoT terminal of the IoT terminal at multiple future time points through the associated correction model corresponding to each IoT terminal, correct the correction demand for computing resources of the IoT terminal at multiple future time points. Among them, the associated correction model can be a short-term memory network model.

[0052] In some embodiments, based on the future computing resource demand data of multiple IoT terminals and the computing resource demand association graph, perform real-time optimization on the initial edge computing resource collaborative allocation scheme to generate a real-time edge computing resource collaborative allocation scheme, which specifically includes: Based on the future computing resource demand data of multiple IoT terminals and the initial edge computing resource collaborative allocation scheme, determine the IoT terminals to be allocated and the edge computing nodes to be allocated. Through the Monte Carlo model, based on the IoT terminals to be allocated and the edge computing nodes to be allocated, generate multiple candidate edge computing resource collaborative allocation schemes. Establish an optimization evaluation function. Based on the computing resource demand association graph, the optimization evaluation function, and multiple candidate edge computing resource collaborative allocation schemes, generate a real-time edge computing resource collaborative allocation scheme.

[0053] Specifically, for each edge computing node, based on the initial collaborative allocation scheme of edge computing resources and the future computing resource requirement data of multiple Internet of Things (IoT) terminals, it can be determined whether a certain computing resource overload occurs in the edge computing node. If so, the IoT terminal with the smallest future computing resource requirement data in this edge computing node is used as the IoT terminal to be allocated. Based on the initial collaborative allocation scheme of edge computing resources and the future computing resource requirement data of multiple IoT terminals, it can be determined whether a certain computing resource is idle in the edge computing node. If so, this edge computing node is used as the edge computing node to be allocated.

[0054] Initial parameters of Monte Carlo simulation can be set, such as the number of simulations, the randomness range of resource allocation, etc. According to the probability model, random sampling is performed on IoT terminals and edge computing nodes to determine which node each terminal is allocated to and how much resource is allocated. The result of each random sampling is used as a candidate collaborative allocation scheme of edge computing resources. This process is repeated multiple times to generate multiple candidate schemes.

[0055] The optimization evaluation function can be related to multiple factors. For example, the average completion processing time or the maximum completion time of the data collected by IoT terminals, the utilization rates of resources such as computing, storage, and network of edge computing nodes, and the load balance among edge computing nodes.

[0056] The optimization evaluation function can also be related to the demand similarity of computing resources corresponding to any two IoT terminals and the demand correlation similarity of any two IoT terminals. For example, in a candidate collaborative allocation scheme of edge computing resources, the smaller the average value of the demand similarity of computing resources corresponding to any two IoT terminals allocated to each edge computing node and the smaller the demand correlation similarity, the larger the optimization value of the candidate collaborative allocation scheme of edge computing resources.

[0057] In some embodiments, based on the computing resource requirement association graph, the optimization evaluation function, and multiple candidate collaborative allocation schemes of edge computing resources, a real-time collaborative allocation scheme of edge computing resources is generated, including: Based on the computing resource requirement association graph and the optimization evaluation function, calculate the optimization value of each candidate collaborative allocation scheme of edge computing resources; Based on the optimization value of each candidate collaborative allocation scheme of edge computing resources, generate a real-time collaborative allocation scheme of edge computing resources through a genetic algorithm.

[0058] Specifically, each candidate edge computing resource collaborative allocation scheme is encoded as an individual in the genetic algorithm, usually using binary encoding or real-number encoding. According to the optimization value of each candidate edge computing resource collaborative allocation scheme, methods such as roulette wheel selection and tournament selection are used to select a certain number of excellent individuals from multiple candidate edge computing resource collaborative allocation schemes as the parent generation. Cross operations are performed on the selected parent individuals to generate new offspring individuals. The cross operation simulates gene exchange in the biological reproduction process and can generate new edge computing resource collaborative allocation schemes. Mutation operations are performed on the offspring individuals to randomly change some of their gene values to increase the diversity of the population. The mutation operation helps to jump out of the local optimal solution and explore a broader solution space. The selection, cross, and mutation operations are repeated continuously to generate new populations. As the iteration progresses, the individuals in the population gradually approach the optimal solution. Termination conditions are set, such as reaching the maximum number of iterations or the fitness value reaching a certain threshold. When the termination conditions are met, the iteration is stopped, and the optimal individual is output as the real-time edge computing resource collaborative allocation scheme.

[0059] S106. According to the real-time edge computing resource collaborative allocation scheme, perform computing resource scheduling on multiple edge computing nodes.

[0060] Specifically, according to the real-time edge computing resource collaborative allocation scheme, reserve the required computing resources for the corresponding Internet of Things terminals on each edge computing node. After the resource reservation is completed, dynamically allocate computing resources to the Internet of Things terminals according to the actual needs to ensure the smooth progress of their computing tasks.

[0061] Figure 3 is a schematic diagram of the modules of the edge computing resource collaborative allocation system for the Internet of Things shown in some embodiments of this specification. As Figure 3 shown, the edge computing resource collaborative allocation system for the Internet of Things may include a demand testing module, an association establishment module, an initial allocation module, a demand acquisition module, a real-time optimization module, and a collaborative allocation module.

[0062] The demand testing module is used to obtain the test computing resource demand data of multiple Internet of Things terminals. Among them, the test computing resource demand data of the Internet of Things terminals includes the demands for various computing resources at multiple test time points; The association establishment module is used to establish an Internet of Things terminal association graph of multiple Internet of Things terminals based on the test computing resource demand data of the multiple Internet of Things terminals; The initial allocation module is used to generate an initial edge computing resource collaborative allocation scheme for multiple Internet of Things terminals based on the Internet of Things terminal association graph of the multiple Internet of Things terminals; The demand acquisition module is used to obtain the historical computing resource demand data of multiple Internet of Things terminals; A real-time optimization module for optimizing the initial edge computing resource collaborative allocation scheme in real time based on the historical computing resource demand data of multiple Internet of Things (IoT) terminals and the IoT terminal association graph, and generating a real-time edge computing resource collaborative allocation scheme; A collaborative allocation module for scheduling computing resources for multiple edge computing nodes according to the real-time edge computing resource collaborative allocation scheme.

[0063] The edge computing resource collaborative allocation system for the Internet of Things can be used to execute the edge computing resource collaborative allocation method for the Internet of Things, which will not be elaborated here.

[0064] Finally, it should be understood that the embodiments described in this specification are only used to illustrate the principles of the embodiments of this specification. Other deformations may also fall within the scope of this specification. Therefore, by way of example and not limitation, alternative configurations of the embodiments of this specification can be regarded as consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly introduced and described in this specification.

Claims

1. An edge computing resource collaborative allocation method for the Internet of Things, characterized in that Including: Obtain the test computing resource requirement data of multiple Internet of Things (IoT) terminals, where the test computing resource requirement data of the IoT terminals includes the requirements for various computing resources at multiple test time points; Based on the test computing resource requirement data of multiple IoT terminals, establish an IoT terminal association graph of multiple IoT terminals; Based on the IoT terminal association graph of multiple IoT terminals, generate an initial edge computing resource collaborative allocation scheme for multiple IoT terminals; Obtain the historical computing resource requirement data of multiple IoT terminals; Based on the historical computing resource requirement data of multiple IoT terminals and the IoT terminal association graph, perform real-time optimization on the initial edge computing resource collaborative allocation scheme to generate a real-time edge computing resource collaborative allocation scheme; According to the real-time edge computing resource collaborative allocation scheme, perform computing resource scheduling on multiple edge computing nodes.

2. The edge computing resource collaborative allocation method for the Internet of Things according to claim 1, characterized in that Based on the test computing resource requirement data of multiple IoT terminals, establish an IoT terminal association graph of multiple IoT terminals, including: For each type of computing resource, based on the test computing resource requirement data of multiple IoT terminals, calculate the demand similarity of the corresponding computing resource requirements of any two IoT terminals. Based on the demand similarity of the corresponding computing resources of any two IoT terminals, determine the first similar IoT terminal corresponding to each IoT terminal's computing resource. Based on the first similar IoT terminal corresponding to each IoT terminal's computing resource, establish the first IoT terminal association sub-graph corresponding to the computing resources of multiple IoT terminals; For each IoT terminal, based on the test computing resource requirement data of the IoT terminal, calculate the demand correlation coefficient of any two corresponding computing resources of the IoT terminal; Based on the demand correlation coefficient of any two corresponding computing resources of each IoT terminal, calculate the demand correlation similarity of any two IoT terminals; Based on the demand correlation similarity of any two IoT terminals, determine the second similar IoT terminal corresponding to each IoT terminal. Based on the second similar IoT terminal corresponding to each IoT terminal, establish the second IoT terminal association sub-graph of multiple IoT terminals, where the IoT terminal association graph includes the first IoT terminal association sub-graph and the second IoT terminal association sub-graph corresponding to the requirements of each type of computing resource.

3. The edge computing resource collaborative allocation method for the Internet of Things according to claim 2, characterized in that Based on the test computing resource requirement data of multiple IoT terminals, calculate the demand similarity of the corresponding computing resource requirements of any two IoT terminals, including: For each IoT terminal, perform time-varying filtering empirical mode decomposition on the requirements for computing resources at multiple test time points of the IoT terminal to obtain multiple target intrinsic mode functions corresponding to the computing resources of the IoT terminal, and extract the function characteristics of each target intrinsic mode function; For any two IoT terminals, based on the function characteristics of each target intrinsic mode function corresponding to the computing resources of the two IoT terminals, calculate the demand similarity of the corresponding computing resource requirements of any two IoT terminals.

4. The edge computing resource collaborative allocation method for the Internet of Things according to claim 2, characterized in that Based on the computing resource requirement association graph of multiple IoT terminals, generate an initial edge computing resource collaborative allocation scheme for multiple IoT terminals, including: Build a sample database, where the sample database is used to store the computing resource requirement characteristics of multiple sample IoT terminal groups, the first IoT terminal association subgraph and the second IoT terminal association subgraph corresponding to each type of computing resource, and the edge computing resource collaborative allocation scheme; Based on the computing resource requirement association graph of multiple IoT terminals, determine similar sample IoT terminal groups from multiple sample IoT terminal groups; Based on the edge computing resource collaborative allocation scheme of the similar sample IoT terminal groups, generate an initial edge computing resource collaborative allocation scheme for multiple IoT terminals.

5. The edge computing resource collaborative allocation method for the Internet of Things according to claim 4, characterized in that Based on the computing resource requirement association graph of multiple IoT terminals, determining similar sample IoT terminal groups from multiple sample IoT terminal groups includes: For each first IoT terminal association subgraph of multiple IoT terminals, extract the global features and local features of the first IoT terminal association subgraph; Extract the global features and local node features of the second IoT terminal association subgraph of multiple IoT terminals; Based on the computing resource requirement characteristics of multiple IoT terminals, the global features and local node features of each first IoT terminal association subgraph and the second IoT terminal association subgraph, determine similar sample IoT terminal groups from multiple sample IoT terminal groups.

6. The edge computing resource collaborative allocation method for the Internet of Things according to claim 2, wherein Based on the historical computing resource requirement data and the computing resource requirement association graph of multiple IoT terminals, perform real-time optimization on the initial edge computing resource collaborative allocation scheme to generate a real-time edge computing resource collaborative allocation scheme, including: Based on the historical computing resource requirement data and the computing resource requirement association graph of multiple IoT terminals, predict the future computing resource requirement data of multiple IoT terminals; Based on the future computing resource requirement data and the computing resource requirement association graph of multiple IoT terminals, perform real-time optimization on the initial edge computing resource collaborative allocation scheme to generate a real-time edge computing resource collaborative allocation scheme.

7. The edge computing resource collaborative allocation method for the Internet of Things according to claim 6, wherein Based on the historical computing resource requirement data and the computing resource requirement association graph of multiple IoT terminals, predicting the future computing resource requirement data of multiple IoT terminals includes: For each IoT terminal, based on the historical computing resource requirement data of the IoT terminal, predict the initial future computing resource requirement data of the IoT terminal, where the initial future computing resource requirement data of the IoT terminal includes the initial requirements of the IoT terminal for various computing resources at multiple future time points; For each type of computing resource, based on the initial requirements of the first similar IoT terminal corresponding to the computing resource for the computing resource at multiple future time points, perform the first iterative correction on the initial requirements of the IoT terminal for the computing resource at multiple future time points to generate the corrected requirements of the IoT terminal for the computing resource at multiple future time points; For each IoT terminal, based on the corrected requirements of the second similar IoT terminal of the IoT terminal for each type of computing resource at multiple future time points, perform the second iterative correction on the corrected requirements of the IoT terminal for the computing resource at multiple future time points to generate the future computing resource requirement data of the IoT terminal.

8. The method for collaborative allocation of edge computing resources for the Internet of Things according to claim 7, characterized in that Based on the future computing resource requirement data of multiple Internet of Things (IoT) terminals and the computing resource requirement correlation graph, the initial edge computing resource collaborative allocation scheme is optimized in real time to generate a real-time edge computing resource collaborative allocation scheme, including: Based on the future computing resource requirement data of multiple IoT terminals and the initial edge computing resource collaborative allocation scheme, determine the IoT terminals to be allocated and the edge computing nodes to be allocated; Through the Monte Carlo model, based on the IoT terminals to be allocated and the edge computing nodes to be allocated, generate multiple candidate edge computing resource collaborative allocation schemes; Establish an optimization evaluation function; Based on the computing resource requirement correlation graph, the optimization evaluation function, and multiple candidate edge computing resource collaborative allocation schemes, generate a real-time edge computing resource collaborative allocation scheme.

9. The method for collaborative allocation of edge computing resources for the Internet of Things according to claim 8, wherein Based on the computing resource requirement correlation graph, the optimization evaluation function, and multiple candidate edge computing resource collaborative allocation schemes, generate a real-time edge computing resource collaborative allocation scheme, including: Based on the computing resource requirement correlation graph and the optimization evaluation function, calculate the optimization value of each candidate edge computing resource collaborative allocation scheme; Through the genetic algorithm, based on the optimization value of each candidate edge computing resource collaborative allocation scheme, generate a real-time edge computing resource collaborative allocation scheme.

10. An edge computing resource collaborative allocation system for the Internet of Things, characterized in that, For implementing the edge computing resource collaborative allocation method for the Internet of Things according to any one of claims 1-9, including: A requirement testing module, configured to obtain the test computing resource requirement data of multiple IoT terminals, where the test computing resource requirement data of the IoT terminals includes the requirements for multiple computing resources at multiple test time points; An association establishment module, configured to establish an IoT terminal correlation graph of multiple IoT terminals based on the test computing resource requirement data of multiple IoT terminals; An initial allocation module, configured to generate an initial edge computing resource collaborative allocation scheme for multiple IoT terminals based on the IoT terminal correlation graph of multiple IoT terminals; A requirement acquisition module, configured to obtain the historical computing resource requirement data of multiple IoT terminals; A real-time optimization module, configured to optimize the initial edge computing resource collaborative allocation scheme in real time based on the historical computing resource requirement data of multiple IoT terminals and the IoT terminal correlation graph, and generate a real-time edge computing resource collaborative allocation scheme; A collaborative allocation module, configured to perform computing resource scheduling on multiple edge computing nodes according to the real-time edge computing resource collaborative allocation scheme.

Images