A task decomposition and allocation method and system for industrial Internet

Abstract

The present invention relates to the field of industrial Internet technology, and specifically to a task decomposition and allocation method and system for the industrial Internet, which decomposes a production task into multiple independent single subtasks, takes the minimum computing resources required for a single subtask as the initial resource amount of the corresponding single subtask; determines the reserve resource amount of a single subtask according to the impact degree of the single subtask, the reserve resource sacrifice coefficient and the initial resource amount; and then traverses each edge server in the cluster domain, and allocates each subtask to the edge server according to the initial resource amount and reserve resource amount of each single subtask in the production task, thereby determining the number of edge servers and the configuration resources of a single edge server. The present invention effectively improves the production efficiency of industrial tasks from the perspective of network resource allocation.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial Internet, and in particular to a method and system for decomposing and allocating tasks for industrial Internet. Background Art

[0002] In the context of Industry 4.0 and intelligent manufacturing, the Industrial Internet has gradually become an important technical means to promote the transformation and upgrading of the manufacturing industry. It combines advanced technologies such as the Internet of Things, big data, cloud computing, and artificial intelligence with traditional industries to achieve equipment interconnection, data-driven and intelligent decision-making, and improve production efficiency and resource utilization.

[0003] In the current industrial intelligent manufacturing production process, a strategy of decomposing a single task into multiple subtasks for parallel processing is usually adopted to improve production efficiency and optimize resource utilization. However, with the increasing number of task splits, IoT monitoring data has also shown a significant growth trend. In the scenario where a large number of tasks are running in parallel with high load, the data transmission time will be significantly extended, and the load on computing resources will also increase. This situation may lead to a decrease in the real-time identification and processing capabilities of relevant information such as product quality at each subtask node in the industrial production process, which in turn causes task congestion. Summary of the invention

[0004] In order to solve the technical problem of limited network resource allocation after industrial task segmentation, which causes task delay and seriously affects production efficiency, the purpose of the present invention is to provide a task decomposition and allocation method and system for the industrial Internet. The technical solutions adopted are as follows:

[0005] In a first aspect, the present invention is a task decomposition and allocation method for the industrial Internet, the method comprising: decomposing a production task in the industrial Internet into multiple independent single subtasks; when the overall production efficiency of the production task is the lowest, using the minimum computing resources required for a single subtask as the initial resource amount of the corresponding single subtask; determining the reserve resource amount of the single subtask according to the influence degree of the single subtask, the reserve resource sacrifice coefficient of the single subtask and the initial resource amount of the single subtask; taking the initial resource amount of the single subtask and the reserve resource amount of the single subtask as the resource configuration requirement of the single subtask; dividing the computing network resources into multiple cluster domains according to the industrial task segmentation system; traversing each edge server in the cluster domain, and allocating each subtask to the edge server according to the resource configuration requirement of each of the single subtasks in the production task, and then determining the number of edge servers and the configuration resources of each edge server.

[0006] In view of the prior art, when industrial tasks are subdivided into various subtasks, a large amount of data requirements will be generated, involving storage and calculation, which often causes huge pressure on the bandwidth of the central server, thereby causing task delays and seriously affecting production efficiency. The present invention divides the computing network resources into multiple cluster domains according to the industrial task segmentation system, introduces a cluster domain network architecture, and aims to effectively disperse the computing resources to multiple edge servers to reduce the burden on the central server. Subsequently, a detailed evaluation is carried out for each subtask, and the minimum computing resource threshold required for the subtask, that is, the initial resource amount, is accurately determined through in-depth analysis. At the same time, according to the degree of influence of each subtask on the overall process, the corresponding reserve resource amount is reasonably allocated to ensure the flexibility of resources and the ability to cope with emergencies, thereby determining the reserve resource amount of a single subtask. On this basis, the initial resource amount and the reserve resource amount are jointly regarded as the actual total amount of resources used by a single subtask, that is, the resource configuration requirement of a single subtask. In order to further optimize the resource configuration, a traversal strategy is adopted to comprehensively investigate the number of edge server nodes, and the results of each traversal are strictly evaluated for the optimization degree of computing resources. Finally, the number of nodes with the peak value of computing resource optimization is selected as the ideal node configuration for the current industrial task. Through the above measures, from the perspective of network resource allocation, the production efficiency of industrial tasks has been effectively improved, laying a solid foundation for the sustainable development and competitiveness improvement of the enterprise.

[0007] Furthermore, the method for obtaining the minimum computing resources required for the single subtask is as follows: by analyzing the historical production data and resource usage of the single subtask, the computing resources of the single subtask and the production efficiency of the single subtask are obtained; according to the fitting relationship between the computing resources and production efficiency of the single subtask, the minimum computing resources required for the single subtask are determined. The computing resources of the single subtask can be displayed by the resource load per unit time.

[0008] Furthermore, the production data includes: production task data, equipment monitoring data, production environment data, material data and order data.

[0009] Furthermore, by analyzing the number of associated tasks of the single subtask and the situation of the tasks that affect it, and the usage ratio of quality inspection computing resources of the single subtask, the impact degree of the single subtask is obtained.

[0010] Furthermore, the number of associated tasks of the single subtask and the impact task status are determined according to the number of tasks of the level to which the single subtask belongs and the number of associated tasks of the single subtask.

[0011] Furthermore, the usage ratio of quality inspection computing resources of the single subtask is determined based on the number and types of computing resources of the single subtask, the required number of each computing resource of the single subtask at the quality inspection level, and the initial resource amount of the single subtask.

[0012] Furthermore, the reserve resource sacrifice coefficient of the single subtask is determined according to the computing resource usage, the required daily output and the current daily output of the single subtask.

[0013] Furthermore, each edge server in the cluster domain is traversed, and each subtask is allocated to an edge server according to the resource configuration requirements of each single subtask in the production task, so as to determine the number of edge servers and the configuration resources of a single edge server, specifically including: traversing each edge server in the cluster domain, and using a greedy algorithm to allocate an edge server to each subtask according to the resource configuration requirements of each single subtask in the production task, and after each subtask is allocated, traversing the allocated resources of each edge server in the cluster domain again, and giving priority to allocating new subtasks to edge servers that currently occupy the least resources.

[0014] Furthermore, after all subtasks are allocated, by analyzing the ratio of the highest efficiency and delay of the current production task, when the ratio is the highest, the number of edge servers and the configuration resources of a single edge server are optimally allocated.

[0015] In a second aspect, the present invention is a task decomposition and allocation system for the industrial Internet, the system comprising: a task decomposition unit: used to decompose the production tasks in the industrial Internet into multiple independent single subtasks; an initial resource determination unit: used to use the minimum computing resources required for a single subtask as the initial resource amount of the corresponding single subtask when the overall production efficiency of the production task is the lowest; a reserve resource determination unit: used to determine the reserve resource amount of a single subtask according to the degree of influence of the single subtask, the reserve resource sacrifice coefficient of the single subtask and the initial resource amount of the single subtask; a task allocation unit: used to take the initial resource amount of the single subtask and the reserve resource amount of the single subtask as the resource configuration requirement of the single subtask; used to divide the computing network resources into multiple cluster domains according to the industrial task segmentation system; traverse each edge server in the cluster domain, and allocate each subtask to the edge server according to the resource configuration requirement of each single subtask in the production task, and then determine the number of edge servers and the configuration resources of each edge server.

[0016] The present invention has the following beneficial effects:

[0017] The present invention divides the computing network resources into multiple cluster domains according to the industrial task segmentation system, introduces the cluster domain network architecture, and aims to effectively disperse the computing resources to multiple edge servers to reduce the burden on the central server. Subsequently, a detailed evaluation is implemented for each single subtask, and the minimum computing resource threshold required for the subtask, that is, the initial resource amount, is accurately determined through in-depth analysis. At the same time, according to the degree of influence of each subtask on the overall process, the corresponding reserve resource amount is reasonably allocated to ensure the flexibility of resources and the ability to cope with emergencies, thereby determining the reserve resource amount of a single subtask. On this basis, the initial resource amount and the reserve resource amount are jointly regarded as the actual total amount of resources used by a single subtask, that is, the resource configuration requirement of a single subtask. In order to further optimize the resource configuration, a traversal strategy is adopted to comprehensively investigate the number of edge server nodes, and a strict computing resource optimization degree evaluation is performed on the results of each traversal. Thereby determining the number of edge servers and the configuration resources of a single edge server. Through the above measures, from the perspective of network resource allocation, the production efficiency of industrial tasks is effectively improved, laying a solid foundation for the sustainable development and competitiveness improvement of enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 A method for decomposing and allocating tasks for an industrial Internet provided by an embodiment of the present invention;

[0020] Figure 2 A task decomposition and allocation system for the industrial Internet provided by an embodiment of the present invention;

[0021] Figure 3 A schematic diagram of the relationship between production efficiency and computing resources shown in an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the task decomposition and allocation method and system of the industrial Internet proposed by the present invention, its specific implementation method, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0023] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0024] The specific scheme of the task decomposition and allocation method and system for the industrial Internet provided by the present invention is described in detail below with reference to the accompanying drawings.

[0025] See also Figure 1 , which shows a flow chart of a task decomposition and allocation method for the industrial Internet provided by an embodiment of the present invention.

[0026] The method of the embodiment of the present invention includes:

[0027] First, the production tasks in the industrial Internet are decomposed into multiple independent subtasks. For an entire industrial production task, such as car manufacturing, engine manufacturing can be regarded as a subtask under the entire industrial production task of "car manufacturing". When the overall production efficiency of the production task is the lowest, the minimum computing resources required for a single subtask are used as the initial resource amount of the corresponding single subtask. Among them, the production efficiency is the time to produce a finished product.

[0028] Specifically, by analyzing the historical production data and resource usage of a single subtask, the computing resources and production efficiency of a single subtask are obtained; based on the fitting relationship between the computing resources and production efficiency of a single subtask, the minimum computing resources required for a single subtask are determined. The computing resources of a single subtask can be displayed by the resource load per unit time. The production data includes: production task data, equipment monitoring data, production environment data, material data, and order data.

[0029] Secondly, the reserve resource quantity of a single subtask is determined based on the impact degree of a single subtask, the reserve resource sacrifice coefficient of a single subtask, and the initial resource quantity of a single subtask. Specifically, the impact degree of a single subtask is obtained by analyzing the number of associated tasks of a single subtask, the impact task situation, and the proportion of quality inspection computing resources used by a single subtask. Among them, the number of associated tasks and the impact task situation of a single subtask are determined based on the number of tasks at the level to which the single subtask belongs and the number of associated tasks of the single subtask. Then, the proportion of quality inspection computing resources used by a single subtask is determined based on the number of types of computing resources of a single subtask, the number of requirements of each computing resource of a single subtask at the quality inspection level, and the initial resource quantity of a single subtask. Finally, the reserve resource sacrifice coefficient of a single subtask is determined based on the computing resource usage, required daily output, and current daily output of a single subtask.

[0030] Finally, the initial resource amount of a single subtask and the reserve resource amount of a single subtask are summed up as the resource configuration requirement of a single subtask. According to the industrial task segmentation system, the computing network resources are divided into multiple cluster domains; each edge server in the cluster domain is traversed, and each subtask is assigned to an edge server according to the resource configuration requirement of each individual subtask in the production task, thereby determining the number of edge servers and the configuration resources of a single edge server. Specifically, each edge server in the cluster domain is traversed, and each subtask is assigned to an edge server according to the resource configuration requirement of each individual subtask in the production task, thereby determining the number of edge servers and the configuration resources of a single edge server, specifically including: traversing each edge server in the cluster domain, according to the resource configuration requirement of each individual subtask in the production task, using a greedy algorithm to assign an edge server to each subtask, after each subtask is assigned, traversing the allocated resources of each edge server in the cluster domain again, and giving priority to assigning new subtasks to the edge servers with the least current resource occupancy. After all subtasks are allocated, the ratio of the highest efficiency and delay of the current production task is analyzed. When the ratio is the highest, the number of edge servers and the configuration resources of a single edge server are optimally allocated.

[0031] The embodiment of the present invention is aimed at the existing technology that when industrial tasks are subdivided into various subtasks, a large amount of data requirements will be generated, involving storage and calculation, which often causes huge pressure on the bandwidth of the central server, thereby causing task delays and seriously affecting production efficiency. The embodiment of the present invention divides the computing network resources into multiple cluster domains according to the industrial task segmentation system, introduces a cluster domain network architecture, and aims to effectively disperse the computing resources to multiple edge servers to reduce the burden on the central server. Subsequently, a detailed evaluation is implemented for each subtask, and the minimum computing resource threshold required for the subtask, that is, the initial resource amount, is accurately determined through in-depth analysis. At the same time, according to the degree of influence of each subtask on the overall process, the corresponding reserve resource amount is reasonably allocated to ensure the flexibility of resources and the ability to cope with emergencies, thereby determining the reserve resource amount of a single subtask. On this basis, the initial resource amount and the reserve resource amount are jointly regarded as the actual total amount of resources used by a single subtask, that is, the cumulative sum of the two is used as the resource configuration requirement of a single subtask. In order to further optimize resource allocation, a traversal strategy is adopted to comprehensively investigate the number of edge server nodes, and a strict computing resource optimization evaluation is performed on the results of each traversal. Finally, the number of nodes that reaches the peak of computing resource optimization is selected as the ideal node configuration for the current industrial task. Through the above measures, from the perspective of network resource allocation, the production efficiency of industrial tasks is effectively improved, laying a solid foundation for the sustainable development and competitiveness improvement of enterprises.

[0032] In another possible embodiment, a method for decomposing and allocating tasks for an industrial Internet is provided, comprising the following steps:

[0033] Step 1: Collect data related to industrial production. In the current era of Industry 4.0, we are committed to improving production efficiency. To this end, a single production task is subdivided into multiple subtasks, and these subtasks are processed in parallel to achieve efficiency improvement.

[0034] Collect relevant manufacturing industry data, including production task data, equipment monitoring data, production environment data, material data, order data and corresponding historical data stored in the corresponding database. At the same time, continuously evaluate and analyze the corresponding equipment operation status and production task progress in the current real-time production process.

[0035] Step 2: In the industrial production process, in order to ensure a significant improvement in overall productivity and resource utilization, it is crucial to scientifically and reasonably split and allocate production task indicators. Although complex industrial production tasks are broken down into multiple single-point tasks that can be executed separately and in parallel, the amount of industrial IoT data for related monitoring and analysis also increases with the increase in segmented tasks, and the unified central data processing method will cause greater network congestion in this scenario. In this regard, some computing resources are delegated to each subtask processing environment, and for some task data with high real-time processing requirements, resources can be directly processed in the corresponding peripheral nodes. In addition, after a single industrial production task is divided, it is also necessary to maintain the same production efficiency of the corresponding delegated tasks so that its industrial production chain can continue to work. In this regard, the current task allocation situation can be evaluated by analyzing the highest production efficiency that can be achieved under the most effective resource allocation under the current fixed total computing resources.

[0036] 1. Set up the industrial production network architecture. In the processing of industrial production tasks, a single task can be split into multiple parallel tasks to improve production efficiency. However, although this method can increase production, it also comes with challenges. After the task is split, parallel execution will trigger a large number of computing tasks and data storage requirements. When these tasks are concentrated and transmitted to the central server through the network, it will bring heavy load pressure to the server, thereby affecting the response speed, and may eventually lead to reduced production efficiency.

[0037] In a highly flexible industrial environment, whenever a new device is connected to the network or an existing device needs to be upgraded or new functions are deployed, the control system needs to be independently configured to meet the needs of network protocol processing and computing. To solve this problem, a cluster domain network architecture is introduced, which divides the computing network into multiple cluster domains based on the industrial task segmentation system. Under this architecture, when regulating resources for industrial tasks, resources can be accurately allocated and released directly for a single cluster domain network.

[0038] In the case of limited resources, appropriate network resources are reasonably allocated to each subtask according to the specific decomposition degree of the current task to ensure that the overall task can maintain high production efficiency. This move aims to achieve dual optimization of production efficiency and resource utilization through refined resource management.

[0039] 2. Analyze the initial resource allocation of a single subtask. The overall industrial production task is broken down into multiple independent subtasks. In the production process of each subtask, huge computing and storage requirements will be generated. These requirements cover multiple links such as real-time data flow processing, logic control execution, and product quality monitoring, which together consume computing resources and storage resources. For these subtasks, different levels of initial resources need to be allocated to them so that they can be put into production operation.

[0040] Analyze the minimum initial resource allocation value of a single subtask. In industrial production activities, many processes require real-time data processing and rapid response. For example, real-time monitoring of production lines, control of automation equipment, and operation of quality inspection systems. If computing resources are insufficiently allocated, it may cause system response delays and untimely execution of control instructions, thus affecting the stability of the production process. For this, it is necessary to analyze the impact of the computing resources of the current single production task subtask on production efficiency. By obtaining the historical production data and resource occupancy of the corresponding single subtask for analysis, the resource load per unit time and the production efficiency of the subtask are obtained.

[0041] Fit the computing resources and production efficiency of the corresponding single production subtask to construct a rectangular coordinate system, such as Figure 3 As shown in the figure, the horizontal axis is computing resources and the vertical axis is production efficiency. It can be found that as computing resources increase, its production efficiency is 0 at the beginning. When the computing resources are sufficient to produce one product, the production efficiency begins to gradually increase until its production speed reaches the threshold of its production line. As computing resources increase, the increase in production efficiency slows down, as shown in the figure. Figure 3 The diagram of the relationship between production efficiency and computing resources is shown. The minimum computing resources for the task operation in a single task stage are obtained, that is, the initial resource amount of a single subtask. In other words, the computing resources that can maintain the minimum production of a single subtask are used as the initial resource amount of a single subtask.

[0042] The consistency of the production process in the current industrial environment will greatly improve production efficiency. For this reason, the production efficiency of each subtask should be consistent, that is, the product efficiency of the subtask should just meet the production needs of the next stage. From a global perspective, the resource allocation of a single subtask is calculated when the overall production efficiency is the lowest. The minimum computing resources calculated by a single subtask in the above process are used as the initial resources of a single subtask, recorded as .

[0043] 3. Allocation of reserve resources for a single subtask. In the above process, the minimum resource allocation required for a single subtask in production activities, that is, the initial resource amount, has been clearly defined. When only these resources are allocated to the subtask, it is observed that in its production activities, the resource usage ratio corresponding to the subtask is significantly higher. This excessive load has a significant impact on the product quality inspection during its production process. In view of the differences in the degree of influence of different production subtasks on the overall production process, the embodiment of the present invention combines the degree of influence of a single subtask and its resource usage in the quality inspection process to accurately calculate and determine the reserve resource amount for the corresponding subtask, so as to ensure the smooth progress of production activities and the stable and reliable product quality.

[0044] Calculate the impact of a single subtask. For the allocation of the amount of preparatory resources for a single subtask, it is necessary to consider the impact of the current subtask. After splitting, an industrial task can be divided into multi-level and multi-batch subtasks, but the importance and impact of different subtasks in their differentiation structure are different. For this, it is necessary to further analyze the importance and impact of a single subtask in its position. The evaluation is performed by analyzing the number of subordinate related tasks of the current subtask, the impact of the tasks, and the current proportion of quality inspection computing resources used by the subtask.

[0045]

[0046] In the above formula, Indicates the influence degree of the i-th subtask. Indicates the number of tasks at the level to which the i-th subtask belongs. Indicates the number of associated tasks of the i-th subtask, that is, the sum of the number of downstream tasks and the number of upstream tasks of the current subtask. Indicates the types and quantity of computing resources, namely CPU, GPU, memory, hard disk and other resources. Indicates the required quantity of the mth computing resource of the ith subtask at the quality inspection level. It represents the initial resource requirement of the mth computing resource of the i-th subtask, that is, the number of allocated mth computing resources in the initial resources of the i-th subtask. It means traversing all current computing resource types, and the type of computing resources is the number of traversals. Represents the normalization function.

[0047] It means that the relevant size is obtained by analyzing the number of associated tasks and the impact tasks of the corresponding single subtask. When the number of corresponding subtasks at the same level is smaller and the number of related tasks is larger, the impact of the current subtask is considered to be higher. Indicates the usage ratio of quality inspection computing resources of the current subtask. The larger the value, the greater the usage ratio of computing resources, that is, the subtask needs to consume more resources to detect quality, and the higher the impact. The sum indicates traversing all current computing resource types, and the type of computing resources is the number of traversals.

[0048] Calculate the reserve resource sacrifice coefficient for a single subtask. In the actual production process, when the order pressure increases, it is necessary to sacrifice a part of the reserve resources as computing resources to provide production tasks. At the same time, when the active sacrifice of some reserve resources can greatly improve its production efficiency, it is also necessary to optimize it to a certain extent. According to the above analysis, the reserve resource sacrifice coefficient of a single subtask is obtained by obtaining the load pressure of the current industrial task and the stage of the current production efficiency.

[0049]

[0050] In the above formula, Represents the reserve resource sacrifice coefficient of the i-th subtask. Indicates the computing resource usage of the current computing subtask (i.e., the i-th subtask). It represents the derivative of the computing resource and production efficiency function corresponding to the current computing subtask. When the value is larger, it means that sacrificing some of the reserve resources can achieve better computing conditions. The computing resource and production efficiency function can be obtained based on the historical data of industrial production tasks, and the corresponding relationship between different computing resources and production efficiency under historical data is used to fit this function. Represents the daily output required. Indicates the current daily output. The larger the value, the less actual output will be, and some reserve resources need to be sacrificed to improve production efficiency.

[0051] The amount of prepared resources for a single subtask. Based on the above analysis, the impact of a single subtask and its prepared resource sacrifice coefficient are obtained. Based on this, the initial minimum computing resources of the subtask are divided and allocated to obtain the actual computing situation of the single subtask.

[0052]

[0053] In the above formula, Indicates the amount of prepared resources for the i-th subtask. The minimum computing resources calculated for a single subtask, that is, the initial resource amount. Indicates the impact of the ith subtask. When the impact of a subtask is higher, more reserve resources need to be allocated to it to ensure production. Indicates the reserve resource sacrifice coefficient of the i-th subtask. When the reserve resource sacrifice coefficient of the calculation subtask is higher, it means that in order to increase the production capacity in the current environment, some reserve resources need to be sacrificed. Finally, the reserve resource amount of a single subtask is calculated.

[0054] 4. Analyze the number of edge servers and configure resources. In the above process, the initial resource amount and reserved resource amount of a single subtask are obtained. In order to reduce the bandwidth burden of the central computing center, the computing resources of the central computing center are allocated to multiple edge servers, and one edge server is responsible for the computing tasks of multiple subtasks. To this end, it is necessary to determine the number of edge servers and the computing resource allocation amount of a single edge server.

[0055] Since the allocation and management of resources is carried out by centralized scheduling of central computing resources, if too many edge servers are allocated, although the task load of a single computing subtask can be greatly reduced, and the resource situation of a single subtask can be more conveniently regulated, when too many edge servers are allocated, the computing resources of the central server will spend more on the initial resource configuration of the edge server, and the computing resources actually used to solve the subtask will be reduced, thereby reducing production efficiency to a certain extent. When fewer edge servers are allocated, the load of a single edge server is relatively large, which in turn causes computing delays and other situations that affect production efficiency. An appropriate number of edge servers can not only reasonably use the resources of the central server, but also generate a certain degree of production efficiency to a great extent, while the resource utilization rate of the central server can reach the highest.

[0056] Assume that the number of edge servers is , and traverse it from 1 to the largest subtask. At the same time, assign subtasks to it after each traversal. According to the initial resource demand of the corresponding subtask, the allocation is made, and the greedy algorithm is used to allocate each subtask, giving priority to the edge server with the current occupied resources and the smallest resource.

[0057] After each traversal, the subtasks will be allocated in an orderly manner. Specifically, the sum of the initial resources and the reserve resources of each subtask is regarded as the amount of resources required to be allocated. During the allocation process, a subtask will first be randomly assigned to an edge server that is not currently hosting any tasks. Then, when allocating the second subtask, the system will comprehensively review the allocated resources of each edge server and give priority to allocating the new subtask to the edge server with the least current resource usage. This strategy ensures that all subtasks are efficiently and evenly allocated to the corresponding edge servers.

[0058] Analyze the situation after allocation, by analyzing the current maximum efficiency and delay, when the ratio is the highest, it is the optimal allocation quantity.

[0059]

[0060] In the above formula, Indicates the resource allocation optimization degree of the current subtask. Represents the maximum schedulable resource of the central server. Indicates the assigned number of edge servers. Indicates the initial resource configuration quantity for each edge server. Indicates the amount of resources currently available for computing. The larger the value, the more resources can be dynamically allocated, and the higher the actual work efficiency.

[0061] Indicates The number of resource processing per second by the edge servers. Indicates the current The bandwidth of the edge servers. Indicates The transmission distance of computing resources of edge servers. The latency of a single edge server is obtained by analyzing the transmission speed and transmission distance of a single edge server. The delay of all edge servers is averaged to represent the actual work delay. When the corresponding delay is high, the actual task completion efficiency will be reduced. The sum represents the cumulative average of the current estimated edge server delays.

[0062] The number of edge servers traversed each time is assigned a corresponding subtask. Then the number of computable resources is used to represent the efficiency of the current industrial task. The final result is expressed as the resource allocation optimization degree of the current industrial task. Select the one with the largest T The value is used as the optimal edge server allocation number to allocate it. Its single edge server allocates it according to the initial resource amount of a single subtask based on the greedy algorithm. Finally, its industrial tasks can be optimized from the network computing resource level.

[0063] Based on the same inventive concept as the above method embodiment, the embodiment of the present invention also provides a task decomposition and allocation system for the industrial Internet, such as Figure 2 As shown, the task decomposition and allocation system specifically includes:

[0064] Task decomposition unit: used to decompose the production tasks in the Industrial Internet into multiple independent subtasks;

[0065] Initial resource determination unit: used to use the minimum computing resources required for a single subtask as the initial resource amount of the corresponding single subtask when the overall production efficiency of the production task is the lowest;

[0066] A reserve resource determination unit: used to determine the reserve resource amount of a single subtask according to the impact degree of the single subtask, the reserve resource amount sacrifice coefficient of the single subtask and the initial resource amount of the single subtask;

[0067] Task allocation unit: used to sum up the initial resource amount of a single subtask and the reserve resource amount of a single subtask as the resource allocation requirement of a single subtask; used to divide the computing network resources into multiple cluster domains according to the industrial task segmentation system; traverse each edge server in the cluster domain, and allocate each subtask to the edge server according to the resource allocation requirement of each single subtask in the production task, and then determine the number of edge servers and the configuration resources of a single edge server.

[0068] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

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

Claims

1. A task decomposition and allocation method for industrial Internet, characterized in that: The method comprises: Decompose the production tasks in the Industrial Internet into multiple independent subtasks; When the overall production efficiency of the production task is the lowest, the minimum computing resources required for a single subtask are used as the initial resource amount of the corresponding single subtask; Determining the reserve resource amount of the single subtask according to the impact degree of the single subtask, the reserve resource amount sacrifice coefficient of the single subtask and the initial resource amount of the single subtask; The initial resource amount of the single subtask and the prepared resource amount of the single subtask are summed up as the resource allocation requirement of the single subtask; Divide computing network resources into multiple cluster domains according to the industrial task segmentation system; traverse each edge server in the cluster domain, and assign each subtask to an edge server according to the resource configuration requirements of each of the individual subtasks in the production task, thereby determining the number of edge servers and the configuration resources of a single edge server; By analyzing the number of associated tasks of the single subtask, the impact of the tasks, and the usage ratio of the quality inspection computing resources of the single subtask, the impact degree of the single subtask is obtained: In the above formula, Indicates the influence of the ith subtask, Indicates the number of tasks at the level to which the i-th subtask belongs, Indicates the number of associated tasks of the i-th subtask, Indicates the number of types of computing resources, represents the required quantity of the mth computing resource of the ith subtask at the quality inspection level, represents the initial resource requirement of the mth computing resource of the ith subtask, represents the normalization function; According to the computing resource usage, required daily output and current daily output of the single subtask, the reserve resource sacrifice coefficient of the single subtask is determined: In the above formula, represents the reserve resource sacrifice coefficient of the i-th subtask, Indicates the computing resource usage of the i-th subtask, It represents the derivative of the computing resource and production efficiency function corresponding to the ith subtask. The computing resource and production efficiency function is obtained according to the historical data of industrial production tasks. The corresponding relationship between different usage computing resources and production efficiency under historical data is used to fit this function. represents the daily output demand, Indicates the current daily output; Determine the amount of reserve resources for a single subtask: In the above formula, Indicates the amount of prepared resources for the i-th subtask.

2. The task decomposition and allocation method of the industrial Internet according to claim 1 is characterized in that: The method for obtaining the minimum computing resources required for the single subtask is as follows: By analyzing the historical production data and resource usage of the single subtask, the computing resources of the single subtask and the production efficiency of the single subtask are obtained; The minimum computing resources required for the single subtask are determined according to the fitting relationship between the computing resources and the production efficiency of the single subtask.

3. The task decomposition and allocation method of the industrial Internet according to claim 2 is characterized in that: The production data includes: production task data, equipment monitoring data, production environment data, material data and order data.

4. The task decomposition and allocation method of the industrial Internet according to claim 1 is characterized in that: The number of associated tasks of the single subtask and the impact task status are determined according to the number of tasks of the level to which the single subtask belongs and the number of associated tasks of the single subtask.

5. The task decomposition and allocation method of the industrial Internet according to claim 1 is characterized in that: The usage ratio of the quality inspection computing resources of the single subtask is determined based on the number of types of computing resources of the single subtask, the required number of each computing resource of the single subtask at the quality inspection level, and the initial resource amount of the single subtask.

6. The task decomposition and allocation method of the industrial Internet according to claim 1 is characterized in that: Traversing each edge server in the cluster domain, allocating each subtask to an edge server according to the resource configuration requirements of each single subtask in the production task, and then determining the number of edge servers and the configuration resources of a single edge server, specifically including: Traverse each edge server in the cluster domain, and use a greedy algorithm to assign an edge server to each subtask according to the resource configuration requirements of each single subtask in the production task. After each subtask is assigned, traverse the allocated resources of each edge server in the cluster domain again, and give priority to assigning new subtasks to the edge servers with the least current resource occupancy.

7. The task decomposition and allocation method of the industrial Internet according to claim 6 is characterized in that: After all subtasks are allocated, the ratio of the highest efficiency and delay of the current production task is analyzed. When the ratio is the highest, the number of edge servers and the configuration resources of a single edge server are optimally allocated.

8. A task decomposition and allocation system for the industrial Internet, characterized in that: The system comprises: Task decomposition unit: used to decompose the production tasks in the Industrial Internet into multiple independent subtasks; An initial resource determination unit: configured to use the minimum computing resources required for a single subtask as the initial resource amount of the corresponding single subtask when the overall production efficiency of the production task is the lowest; A reserve resource determination unit: used to determine the reserve resource amount of a single subtask according to the impact degree of the single subtask, the reserve resource amount sacrifice coefficient of the single subtask and the initial resource amount of the single subtask; Task allocation unit: used to sum the initial resource amount of the single subtask and the reserve resource amount of the single subtask as the resource configuration requirement of the single subtask; used to divide the computing network resources into multiple cluster domains according to the industrial task segmentation system; traverse each edge server in the cluster domain, and according to the resource configuration requirement of each single subtask in the production task, allocate each subtask to the edge server, and then determine the number of edge servers and the configuration resources of a single edge server; By analyzing the number of associated tasks of the single subtask, the impact of the tasks, and the usage ratio of the quality inspection computing resources of the single subtask, the impact degree of the single subtask is obtained: In the above formula, Indicates the influence of the ith subtask, Indicates the number of tasks at the level to which the i-th subtask belongs, Indicates the number of associated tasks of the i-th subtask, Indicates the number of types of computing resources, represents the required quantity of the mth computing resource of the ith subtask at the quality inspection level, represents the initial resource requirement of the mth computing resource of the ith subtask, represents the normalization function; According to the computing resource usage, required daily output and current daily output of the single subtask, the reserve resource sacrifice coefficient of the single subtask is determined: In the above formula, represents the reserve resource sacrifice coefficient of the i-th subtask, Indicates the computing resource usage of the i-th subtask, It represents the derivative of the computing resource and production efficiency function corresponding to the ith subtask. The computing resource and production efficiency function is obtained according to the historical data of industrial production tasks. The corresponding relationship between different usage computing resources and production efficiency under historical data is used to fit this function. represents the daily output demand, Indicates the current daily output; Determine the amount of reserve resources for a single subtask: In the above formula, Indicates the amount of prepared resources for the i-th subtask.

Images