Distributed scheduling computing method, system, medium and device for the Internet of Things

Abstract

The present application provides a distributed scheduling computing method, system, medium and device for the Internet of Things; including: based on the resource set of each subtask and computing node of the computing task, obtaining the optimal matching mode set of the computing task, so as to obtain the target mode by dynamic compression encoding and transmit it to the computing node; then calculating the marginal computing power sequence based on the target mode to obtain the weak marginal computing power sequence, and obtaining the target computing node corresponding to the weak marginal computing power sequence; finally, pressing the target computing node into the computing power pool, and obtaining the replacement node of the target computing node based on the computing power pool, so as to transfer the subtask corresponding to the weak marginal computing power sequence to the target computing node. The present application can improve the overall computing efficiency while ensuring the smooth progress of distributed computing tasks, and is particularly suitable for the computing scenarios of the Internet of Things with high real-time requirements and unstable networks; and improves the efficiency and stability of the algorithm while improving the utilization rate of the idle computing resources of the Internet of Things.

Description

Technical Field

[0001] The present application relates to the technical field of scheduling algorithms, and in particular to a distributed scheduling calculation method, system, medium and device for the Internet of Things. Background Art

[0002] With the continuous advancement of chip manufacturing technology and the continuous maturity of information technology, the Internet of Things, which consists of various sensor devices, intelligent hardware, and computing terminals, has begun to take shape, and the Internet of Things is expected to become the supporting technology for the next generation of networks. Under normal circumstances, the data collected by IoT devices needs to be calculated and processed in real time to improve the performance of the system, so that the Internet of Things can serve various complex production and manufacturing, environmental monitoring and other scenarios.

[0003] At present, the conventional method is that when an IoT node has a computing demand, it directly sends the computing task to the remote server, and the high-performance server completes the calculation and then returns the result to the IoT node. The advantage of this method is that the computing efficiency is high when the network condition is good. However, in many cases, IoT devices are often networked in the form of a local area network. Even if they can sometimes connect to a remote server, the network signal is unstable; in addition, when the system has a large computing request, it often causes network congestion, thus affecting the computing efficiency. Therefore, it is necessary to design a distributed scheduling computing algorithm based on the IoT node itself. Summary of the invention

[0004] In order to solve the above technical problems, the present application provides a distributed scheduling calculation method, system, medium and device for the Internet of Things.

[0005] Specifically, the present application provides a distributed scheduling calculation method for the Internet of Things, which completes the distributed scheduling calculation through a scheduling node, and the method includes the following steps:

[0006] S100: Based on the resource sets of each subtask of the computing task and the resource sets of the computing nodes, an optimal matching pattern set of the computing task is obtained.

[0007] S200: Dynamically compress and encode each pattern of the optimal matching pattern set to obtain a target pattern and transmit it to a computing node.

[0008] S300: Calculate the marginal computing power sequence of each computing node based on the target mode to obtain a weak marginal computing power sequence, and obtain the target computing node corresponding to the weak marginal computing power sequence.

[0009] S400: Pushing the target computing node into a computing power pool, and acquiring a replacement node for the target computing node based on the computing power pool, so as to transfer the subtask corresponding to the weak marginal computing power sequence to the target computing node.

[0010] Furthermore, in step S100, the resource set of the subtask includes the computing power resource value p required by the subtask i and storage resource value q i ; The resource set of the computing node includes the computing power resource value x owned by the computing node j and storage resource value y j .

[0011] Among them, p i is the computing power resource value required by the i∈[1,N]th subtask, q i is the storage resource value required by the i-th subtask; x j is the computing power resource value owned by computing node j, j∈[1,M], y j is the storage resource value owned by computing node j.

[0012] Furthermore, the step S100 of obtaining the optimal matching mode set of the computing task includes:

[0013] When the resource set of the computing node meets When δ>0, the optimal matching pattern set F={(x j ,y j )|j∈[1,M]and|F|=N}.

[0014] Where N is the total number of resources required by the subtasks, M is the total number of resources required by the computing nodes, and M>N; |F| is the number of elements in the optimal matching pattern set F; δ and ε are both arbitrary constants greater than zero.

[0015] Furthermore, the step S200 includes:

[0016] Obtain the total data volume of the subtasks corresponding to each mode in the optimal matching mode set, and the corresponding average data transmission rate between the computing node and the scheduling node, so as to calculate the average duration of data transmission between all computing nodes and the scheduling node according to the total data volume and the average data transmission rate

[0017] Use dynamic compression ratio Each pattern is dynamically compressed and encoded, and the dynamically compressed and encoded pattern is transmitted to the computing node as the target pattern; wherein,

[0018] Furthermore, the step S300 of calculating the marginal computing power sequence of each computing node based on the target mode includes:

[0019] The computing progress of each computing node is obtained at preset time intervals to obtain the computing progress sequence of the computing nodes corresponding to each target mode, and the marginal computing power sequence is calculated based on the computing progress sequence.

[0020] Furthermore, obtaining the target computing node in step S300 includes:

[0021] When the marginal computing power sequence satisfies , we obtain the weak marginal computing power sequence.

[0022] Among them, b j,k+1 represents the k+1th marginal computing power sequence of computing node j, b j,k represents the Kth marginal computing power sequence of computing node j, a j,k represents the kth calculation progress returned by calculation node j, G is the number of elements in the calculation progress sequence at the current moment, and b j,G-1 -b j,G-2 Represents the difference in marginal computing power, b j,G-1 represents the marginal computing power of computing node j when the computing progress sequence is G-1 elements at the current moment, b j,G-2 It represents the marginal computing power of computing node j when the computing progress sequence is G-2 elements at the current moment. α is any constant greater than zero, and β is any constant less than zero.

[0023] Obtain the sum of the calculation progress of the calculation nodes corresponding to the weak marginal computing power sequence, determine whether the sum of the calculation progress is greater than a default value, and determine whether the weak marginal computing power sequence is a non-negative non-decreasing sequence.

[0024] When the sum of the computing progress is greater than a default value and the weak marginal computing power sequence is a non-negative non-decreasing sequence, the computing node is used as the target computing node.

[0025] Furthermore, the step S400 of obtaining a replacement node for the target computing node includes:

[0026] The latest computing progress values ​​of all target computing nodes in the computing power pool are obtained, and the difference of the latest marginal computing power of the target computing node is calculated based on the latest computing progress value, so as to obtain a target computing node whose mean square error of the difference between it and the adjacent preset marginal computing power is the smallest as the final replacement node.

[0027] Based on the same concept, the present application also provides a system for a distributed scheduling calculation method for the Internet of Things, including:

[0028] The first acquisition module is used to acquire the optimal matching mode set of the computing task based on the resource set of each subtask of the computing task and the resource set of the computing node.

[0029] Compression module: used for dynamically compressing and encoding each mode of the optimal matching mode set to obtain the target mode and transmit it to the computing node.

[0030] The second acquisition module is used to calculate the marginal computing power sequence of each computing node based on the target mode to obtain a weak marginal computing power sequence, and obtain the target computing node corresponding to the weak marginal computing power sequence.

[0031] Migration module: used to push the target computing node into the computing power pool, and obtain a replacement node for the target computing node based on the computing power pool, so as to transfer the subtasks corresponding to the weak marginal computing power sequence to the target computing node.

[0032] Based on the same concept, the present application also provides a medium, in which a computer program is stored, wherein the computer program is configured to execute the distributed scheduling calculation method for the Internet of Things when running.

[0033] Based on the same concept, the present application also provides an electronic device, which includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the distributed scheduling computing method for the Internet of Things.

[0034] Compared with the prior art, the beneficial effects of this application are:

[0035] This application improves the overall computing efficiency by matching each subtask with the computing node in the optimal mode and dynamically compressing and transmitting data. At the same time, it solves the computing bottleneck problem in a timely manner by finding the weak marginal computing power sequence in real time and migrating the corresponding subtasks to the computing power pool to ensure the smooth progress of distributed computing tasks. It is particularly suitable for IoT computing scenarios with high real-time requirements and unstable networks.

[0036] This application distributes computing tasks to the internal nodes of the Internet of Things, effectively solving the problem of traditional algorithms' high dependence on remote high-performance servers and high requirements for network transmission. It improves the efficiency and stability of the algorithm while increasing the utilization rate of the Internet of Things's spare computing resources, thereby expanding the application prospects of Internet of Things technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 This is a flow chart of the distributed scheduling calculation method for the Internet of Things described in this application.

[0038] Figure 2 for Figure 1 A system framework diagram of the distributed scheduling calculation method for the Internet of Things.

[0039] Figure 3 for Figure 2 Schematic diagram of the overall topological structure of the system. DETAILED DESCRIPTION

[0040] The following is a further detailed description of a distributed scheduling calculation method, system, medium and device for the Internet of Things of the present application in conjunction with specific embodiments and drawings.

[0041] See also Figure 1 The present application provides a distributed scheduling calculation method for the Internet of Things, which completes the distributed scheduling calculation through a scheduling node. The method includes the following steps:

[0042] S100: Based on the resource sets of each subtask of the computing task and the resource sets of the computing nodes, an optimal matching pattern set of the computing task is obtained.

[0043] Furthermore, in step S100, the resource set of the subtask includes the computing power resource value p required by the subtask i and storage resource value q i ; The resource set of the computing node includes the computing power resource value x owned by the computing node j and storage resource value y j .

[0044] Among them, p i is the computing power resource value required by the i∈[1,N]th subtask, q i is the storage resource value required by the i-th subtask; x j is the computing power resource value owned by computing node j, j∈[1,M], y j is the storage resource value owned by computing node j.

[0045] In this embodiment, let the resource set required by each subtask of a complete computing task be {(p i ,q i )|i∈[1,N]}.

[0046] The total number of nodes that can act as computing nodes in the IoT system is M>N, and the resource sets owned by these computing nodes are the full set E={(x j ,y j )|j∈[1,M]}.

[0047] Among them, x j is the computing power resource value owned by computing node j, j∈[1,M], y j is the storage resource value owned by computing node j.

[0048] Furthermore, the step S100 of obtaining the optimal matching mode set of the computing task includes:

[0049] When the resource set of the computing node meets When δ>0, the optimal matching pattern set F={(x j ,y j )|j∈[1,M]and|F|=N}.

[0050] Where N is the total number of resources required by the subtasks, M is the total number of resources required by the computing nodes, and M>N; |F| is the number of elements in the optimal matching pattern set F; δ and ε are both arbitrary constants greater than zero.

[0051] In this embodiment, the scheduling node solves the subset F of the full set E = {(x j ,y j )|j∈[1,M] and |F|=N}, |F| is the number of elements in subset F, so that subset F satisfies:

[0052]

[0053] The subset F that meets the above conditions is the optimal matching pattern set for the computing task. And the pattern set that meets the above conditions can minimize the difference in the time required for the calculation of each subtask, thereby minimizing the overall computing time.

[0054] After the optimal matching pattern set is obtained, step S200 may be executed.

[0055] S200: Dynamically compress and encode each pattern of the optimal matching pattern set to obtain a target pattern and transmit it to a computing node.

[0056] Furthermore, the step S200 includes:

[0057] Obtain the total data volume of the subtasks corresponding to each mode in the optimal matching mode set, and the corresponding average data transmission rate between the computing node and the scheduling node, so as to calculate the average duration of data transmission between all computing nodes and the scheduling node according to the total data volume and the average data transmission rate

[0058] Use dynamic compression ratio Each pattern is dynamically compressed and encoded, and the dynamically compressed and encoded pattern is transmitted to the computing node as the target pattern; wherein,

[0059] Among them, the use of the above compression algorithm can enable each computing node to start computing as synchronously as possible, reducing the overall computing efficiency caused by a subtask taking too long.

[0060] After the target mode is acquired and the corresponding transmission is completed, step S300 may be executed.

[0061] S300: Calculate the marginal computing power sequence of each computing node based on the target mode to obtain a weak marginal computing power sequence, and obtain the target computing node corresponding to the weak marginal computing power sequence.

[0062] Furthermore, the step S300 of calculating the marginal computing power sequence of each computing node based on the target mode includes:

[0063] The computing progress of each computing node is obtained at preset time intervals to obtain the computing progress sequence of the computing nodes corresponding to each target mode, and the marginal computing power sequence is calculated based on the computing progress sequence.

[0064] In this embodiment, the scheduling node obtains the computing progress of each computing node at a preset time interval, and records the target mode (x j ,y j ) The calculation progress sequence returned by the corresponding computing node j is {a j,k |k∈[1,G]},a j,k represents the kth calculation progress returned by computing node j, G is the number of elements in the calculation progress sequence at the current moment, and the scheduling node then calculates its marginal computing power sequence {b j,k =a j,k+1 -a j,k |k∈[1,G-1]}.

[0065] Among them, the marginal computing power sequence reflects the stability of the computing power of the computing node over time.

[0066] Furthermore, obtaining the target computing node in step S300 includes:

[0067] When the marginal computing power sequence satisfies , we obtain the weak marginal computing power sequence.

[0068] Among them, b j,k+1 represents the k+1th marginal computing power sequence of computing node j, b j,k represents the Kth marginal computing power sequence of computing node j, a j,k represents the kth calculation progress returned by calculation node j, G is the number of elements in the calculation progress sequence at the current moment, and b j,G-1 -b j,G-2 Represents the difference in marginal computing power, b j,G-1 represents the marginal computing power of computing node j when the computing progress sequence is G-1 elements at the current moment, b j,G-2 It represents the marginal computing power of computing node j when the computing progress sequence is G-2 elements at the current moment. α is any constant greater than zero, and β is any constant less than zero.

[0069] In this embodiment, if the target mode (xj ,y j ) corresponds to the marginal computing power sequence {b j,k =a j,k+1 -a j,k |k∈[1,G-1]} is a decreasing sequence, and the sum of its computing progress sequence is less than the default value and the difference between the last two marginal computing powers is less than another default value, that is, the subtasks of computing node j that meet the above conditions need to be migrated.

[0070] The marginal computing power sequence that meets this condition is called a weak marginal computing power sequence. The computing power of the corresponding computing node j becomes worse and worse over time, and it is unable to complete the assigned subtasks on time.

[0071] Obtain the sum of the calculation progress of the calculation nodes corresponding to the weak marginal computing power sequence, determine whether the sum of the calculation progress is greater than a default value, and determine whether the weak marginal computing power sequence is a non-negative non-decreasing sequence.

[0072] When the sum of the computing progress is greater than a default value and the weak marginal computing power sequence is a non-negative non-decreasing sequence, the computing node is used as the target computing node.

[0073] In this embodiment, if the sum of the computation progress of computing node j is greater than the default value, that is, And its marginal computing power sequence is a non-negative non-decreasing sequence, that is, (b j,k =a j,k+1 -a j,k |b j,k ≥0,b j,k+1 ≥ b j,k ,k∈[1,G-1]}, the scheduling node takes computing node j as the target computing node, and γ is any constant greater than zero.

[0074] After the target computing node is acquired, step S400 may be executed.

[0075] S400: Pushing the target computing node into a computing power pool, and acquiring a replacement node for the target computing node based on the computing power pool, so as to transfer the subtask corresponding to the weak marginal computing power sequence to the target computing node.

[0076] Furthermore, the step S400 of obtaining a replacement node for the target computing node includes:

[0077] The latest computing progress values ​​of all target computing nodes in the computing power pool are obtained, and the difference of the latest marginal computing power of the target computing node is calculated based on the latest computing progress value, so as to obtain a target computing node whose mean square error of the difference between it and the adjacent preset marginal computing power is the smallest as the final replacement node.

[0078] In this embodiment, when the next preset time interval arrives, the scheduling node receives the latest calculation progress value a of all computing nodes j in the computing power pool. j,G+1 .

[0079] Then, based on a j,G+1 calculate the latest marginal computing power difference b of computing node j j,G -b j,G-1 .

[0080] Then select a computing node j with the smallest mean square error of the marginal computing power differences {b j,G-u -b j,G-u-1 |k ∈ [1, H]} from among its neighboring ones as the final replacement node.

[0081] Here, H is an integer greater than 1 and less than G. In the actual development process, a fixed value will be selected according to requirements (such as accuracy requirements).

[0082] The scheduling node then transfers the corresponding computing task to computing node j.

[0083] Above, it should be noted that α, β, γ, δ, and ε all belong to constants (constants of the real number type). The meanings of these several constants are different, and the specific meaning or function is to limit the range of the calculation results of the left part of the inequalities where these constants are located.

[0084] Please refer to Figure 2 , this application also provides a system for a distributed scheduling calculation method for the Internet of Things, including:

[0085] The first acquisition module: used to obtain the optimal matching mode set of the computing task based on the resource sets of the subtasks of the computing task and the resource set of the computing nodes.

[0086] The compression module: used to perform dynamic compression encoding on each mode of the optimal matching mode set to obtain the target mode and transmit it to the computing nodes.

[0087] The second acquisition module: used to calculate the marginal computing power sequence of each computing node based on the target mode to obtain the weak marginal computing power sequence, and obtain the target computing node corresponding to the weak marginal computing power sequence.

[0088] The migration module: used to push the target computing node into the computing power pool, and obtain the replacement node of the target computing node based on the computing power pool, so as to transfer the subtasks corresponding to the weak marginal computing power sequence to the target computing node.

[0089] This application also provides a medium, in which a computer program is stored, and wherein the computer program is set to execute the distributed scheduling calculation method for the Internet of Things when running.

[0090] The present application also provides an electronic device, which includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the distributed scheduling computing method for the Internet of Things.

[0091] In summary, the present application provides a distributed scheduling computing method, system, medium and device for the Internet of Things; distributed scheduling computing is completed through scheduling nodes, including: based on the resource set of each subtask of the computing task and the resource set of the computing node, the optimal matching pattern set of the computing task is obtained; then each mode of the optimal matching pattern set is dynamically compressed and encoded to obtain the target mode and transmit it to the computing node; then the marginal computing power sequence of each computing node is calculated based on the target mode to obtain a weak marginal computing power sequence, and the target computing node corresponding to the weak marginal computing power sequence is obtained; finally, the target computing node is pushed into the computing power pool, and the replacement node of the target computing node is obtained based on the computing power pool to transfer the subtask corresponding to the weak marginal computing power sequence to the target computing node.

[0092] This application improves the overall computing efficiency by matching each subtask with the computing node in the optimal mode and dynamically compressing and transmitting data. At the same time, it solves the computing bottleneck problem in a timely manner by finding the weak marginal computing power sequence in real time and migrating the corresponding subtask to the computing power pool to ensure the smooth progress of distributed computing tasks. It is particularly suitable for IoT computing scenarios with high real-time requirements and unstable networks. This application distributes computing tasks to the internal nodes of the IoT, effectively solving the problem that traditional algorithms have high dependence on remote high-performance servers and high requirements for network transmission. It improves the efficiency and stability of the algorithm while improving the utilization rate of the idle computing resources of the IoT, thereby expanding the application prospects of IoT technology.

[0093] Although example embodiments have been described herein with reference to the accompanying drawings, it should be understood that the above example embodiments are merely exemplary and are not intended to limit the scope of the present application to this. Those of ordinary skill in the art may make various changes and modifications therein without departing from the scope and spirit of the present application. All these changes and modifications are intended to be included within the scope of the present application as required by the appended claims.

[0094] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0095] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed.

[0096] The various component embodiments of the present application can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or digital signal processor (DSP) can be used in practice to implement some or all functions of some modules according to the embodiments of the present application. The application can also be implemented as a device program (e.g., computer program and computer program product) for executing a part or all of the methods described herein. Such a program implementing the present application can be stored on a computer-readable medium, or can have the form of one or more signals. Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0097] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0098] Although the description of the present application is carried out in conjunction with the above specific embodiments, it is obvious that those skilled in the art can make many substitutions, modifications and changes based on the above content. Therefore, all such substitutions, improvements and changes are included in the spirit and scope of the appended claims.

Claims

1. A distributed scheduling calculation method for the Internet of Things, It is characterized in that The distributed scheduling calculation is completed by scheduling nodes, and the method includes the following steps: S100: based on resource sets of each subtask of the computing task and resource sets of computing nodes, obtaining an optimal matching mode set of the computing task; S200: Dynamically compress and encode each pattern of the optimal matching pattern set to obtain a target pattern and transmit it to a computing node; S300: Calculate the marginal computing power sequence of each computing node based on the target mode to obtain a weak marginal computing power sequence, and obtain a target computing node corresponding to the weak marginal computing power sequence; S400: Pushing the target computing node into a computing power pool, and acquiring a replacement node for the target computing node based on the computing power pool, so as to transfer the subtask corresponding to the weak marginal computing power sequence to the target computing node.

2. According to the distributed scheduling calculation method for the Internet of Things according to claim 1, It is characterized in that In step S100, the resource set of the subtask includes the computing power resource value p required by the subtask. i and storage resource value q i ; The resource set of the computing node includes the computing power resource value x owned by the computing node j and storage resource value y j ; Among them, p i is the computing power resource value required by the i∈[1,N]th subtask, q i is the storage resource value required by the i-th subtask; x j is the computing power resource value owned by computing node j, j∈[1,M], y j is the storage resource value owned by computing node j.

3. According to the distributed scheduling calculation method for the Internet of Things according to claim 2, It is characterized in that The step S100 of obtaining the optimal matching mode set of the computing task includes: When the resource set of the computing node meets When the optimal matching pattern set F = (x j ,y j )|j∈[1,M] and |F|=N}; Where N is the total number of resources required by the subtasks, M is the total number of resources required by the computing nodes, and M>N; |F| is the number of elements in the optimal matching pattern set F; δ and ε are both arbitrary constants greater than zero.

4. The distributed scheduling calculation method for the Internet of Things according to claim 3, It is characterized in that The step S200 includes: Obtain the total data volume of the subtasks corresponding to each mode in the optimal matching mode set, and the corresponding average data transmission rate between the computing node and the scheduling node, so as to calculate the average duration of data transmission between all computing nodes and the scheduling node according to the total data volume and the average data transmission rate Use dynamic compression ratio Dynamically compress and encode each pattern, and transmit the dynamically compressed and encoded pattern as the target pattern to the computing node; in, Mode(x j ,y j ) corresponds to a total amount of data for the subtask z j The average data transmission rate between the computing node j and the scheduling node corresponding to this mode is c j .

5. The distributed scheduling calculation method for the Internet of Things according to claim 4, It is characterized in that The step S300 of calculating the marginal computing power sequence of each computing node based on the target mode includes: The computing progress of each computing node is obtained at preset time intervals to obtain the computing progress sequence of the computing nodes corresponding to each target mode, and the marginal computing power sequence is calculated based on the computing progress sequence.

6. The distributed scheduling calculation method for the Internet of Things according to claim 5, It is characterized in that The step S300 of obtaining the target computing node includes: When the marginal computing power sequence satisfies When , we obtain the weak marginal computing power sequence; Among them, b j,k+1 represents the k+1th marginal computing power sequence of computing node j, b j,k represents the Kth marginal computing power sequence of computing node j, a j,k represents the kth calculation progress returned by calculation node j, G is the number of elements in the calculation progress sequence at the current moment, and b j,G-1 -b j,G-2 Represents the difference in marginal computing power, b j,G-1 represents the marginal computing power of computing node j when the computing progress sequence is G-1 elements at the current moment, b j,G-2 It represents the marginal computing power of computing node j when the computing progress sequence is G-2 elements at the current moment, α is any constant greater than zero, and β is any constant less than zero; Obtain the sum of the calculation progress of the calculation nodes corresponding to the weak marginal computing power sequence, determine whether the sum of the calculation progress is greater than a default value, and determine whether the weak marginal computing power sequence is a non-negative non-decreasing sequence; When the sum of the computing progress is greater than a default value and the weak marginal computing power sequence is a non-negative non-decreasing sequence, the computing node is used as the target computing node.

7. The distributed scheduling calculation method for the Internet of Things according to claim 6, It is characterized in that The step S400 of obtaining a replacement node for the target computing node includes: The latest computing progress values ​​of all target computing nodes in the computing power pool are obtained, and the difference of the latest marginal computing power of the target computing node is calculated based on the latest computing progress value, so as to obtain a target computing node whose mean square error of the difference between it and the adjacent preset marginal computing power is the smallest as the final replacement node.

8. A system using the distributed scheduling calculation method for the Internet of Things as described in any one of claims 1 to 7, It is characterized in that include: A first acquisition module: used for acquiring an optimal matching mode set of the computing task based on a resource set of each subtask of the computing task and a resource set of a computing node; Compression module: used for dynamically compressing and encoding each mode of the optimal matching mode set to obtain a target mode and transmit it to a computing node; The second acquisition module is used to calculate the marginal computing power sequence of each computing node based on the target mode to obtain a weak marginal computing power sequence, and obtain the target computing node corresponding to the weak marginal computing power sequence; Migration module: used to push the target computing node into the computing power pool, and obtain a replacement node for the target computing node based on the computing power pool, so as to transfer the subtasks corresponding to the weak marginal computing power sequence to the target computing node.

9. A medium, It is characterized in that The medium stores a computer program, wherein the computer program is configured to execute the distributed scheduling computing method for the Internet of Things as described in any one of claims 1 to 7 when running.

10. An electronic device, It is characterized in that The electronic device includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the distributed scheduling computing method for the Internet of Things as described in any one of claims 1-7.

Images