Task segmentation and distribution method based on greedy algorithm

Abstract

The invention discloses a task segmentation and allocation method based on a greedy algorithm. The method comprises the following steps: S1, collecting and standardizing a resource state; s2, constructing a hardware resource pool; S3, managing tasks and sub-tasks: establishing a task list, wherein the task list comprises a main task and sub-tasks segmented by the main task; s4, task segmentation: segmenting the main task into a series of sub-task units and adding the sub-task units to a task list; s5, task allocation: based on a greedy algorithm, selecting a required'greedy 'target as sub-task allocation, and preferentially selecting the shortest total time of task execution, the least resources occupied by the task and the minimum resource cost-benefit ratio; and S7, dynamic adjustment: monitoring the state of the hardware resource and the completion condition of the task in real time, returning to the step S5 when the resource is idle or unforeseen circumstances cannot be continued, and reallocating or adjusting the task.

Description

Technical Field

[0001] The present invention belongs to the field of industrial Internet of Things, and in particular relates to a task segmentation and allocation method based on a greedy algorithm. Background Art

[0002] Efficient cross-domain and multi-factor collaboration is the basic guarantee for the joint capabilities of the system under information conditions. By building an intelligent resource group, we can meet the needs of the development of group equipment collaborative operations, greatly improve the manageability, evaluability, flexibility, and precise service level of major tasks, and realize that resources can be connected to the network at any time, stable and reliable interconnection, efficient data intercommunication, on-demand sharing, and physical resource collaborative operations. Within the resource group, some physical resources are limited by their small size and simple functions, and cannot complete major tasks, resulting in them not being utilized during the task execution process and unable to contribute capabilities to the entire group and improve the efficiency of group tasks. The increasing number of large and complex tasks has seriously aggravated the occurrence of this situation. Therefore, in a multi-resource environment, how to effectively divide and allocate tasks has become the key to improving group efficiency. Summary of the invention

[0003] 1. Technical issues to be resolved

[0004] By dividing tasks and allocating subtask units, the collaborative operation of devices in the resource group can be achieved, thus improving task efficiency. Use unified standard specifications to evaluate resource status and attributes, and at the same time break up resources to decouple the relationship between resources, forming a hardware resource pool that is easy to operate in a unified manner; provide a measurable, safe, and rapidly scalable continuous resource supply pool for task division and subtask allocation, and achieve efficient collaboration of intelligent group systems.

[0005] (II) Technical solution

[0006] A task segmentation and allocation method based on a greedy algorithm, the process is as follows:

[0007] S1 Resource status collection and specification: Based on unified standard specifications, collect the resource status and attributes of hardware entities within the group; collect data of different structures from different resources and integrate them into structured data with unified structure;

[0008] S2 Hardware Resource Pool Construction: Put the hardware entity resources and their related status and attributes into the cloud pool for unified management, which is called the hardware resource pool. All resources in the group are interconnected and interoperable; update the status and attributes of all resources in the group in real time, monitor the task execution status and real-time changes of resources, and dynamically adjust the task allocation strategy to adapt to environmental changes;

[0009] S3 task and subtask management: Create a task list, including the main task and the subtasks divided from the main task;

[0010] S4 Task division: According to the hardware resource status and attributes in the resource pool and the characteristics and requirements of the main task, the hardware entity resource capabilities and task attributes are matched to determine which resources can perform which tasks, and the main task is divided into a series of subtask units and added to the task list;

[0011] S5 task allocation: Based on the greedy algorithm, the "greedy" target is selected as the subtask allocation, and the task with the shortest total execution time, the least resources occupied by the task, and the minimum resource cost-effectiveness is given priority;

[0012] S6, Task execution: After the hardware resource receives the subtask, it starts to execute it and updates the running status and task status to the hardware resource pool in real time to ensure the smooth execution of the task;

[0013] S7, dynamic adjustment: monitor the status of hardware resources and the completion of tasks in real time. When resources become idle or unexpected circumstances prevent the process from continuing, return to step S5 to reallocate or adjust tasks.

[0014] Furthermore, the hardware entity resources described in S1 include sensor equipment, transportation equipment, computing and storage equipment, and communication network equipment.

[0015] Furthermore, the resource status and attribute parameters in S1 include but are not limited to: owner ID, whether it is occupied, occupant ID, whether it is allowed to be preempted, occupation start time, estimated occupation end time, capabilities, corresponding tasks that can be performed, and advantage tasks.

[0016] Furthermore, the format of S3 main tasks and subtasks adopts the 5W1H principle: Who represents the hardware resources that issue tasks or receive tasks, What represents the actions that need to be performed for the task, When represents the time when the task is issued, Where represents the work area targeted by the task, Why represents the notes of the task, and How represents the relevant attributes and status of the task.

[0017] Furthermore, the S4 hardware entity resource capabilities include resource reconnaissance capabilities, computing capabilities, transportation capabilities, memory capacity, and network bandwidth.

[0018] Furthermore, S4 task segmentation; consider the nature of the task, whether it needs to be executed in parallel or sequentially, and whether there are time and space restrictions; consider the dependencies between subtasks, some subtasks may need to be executed after other subtasks are completed, while other subtasks can be executed in parallel, and establish a suitable dependency graph to ensure that subtasks are executed in the correct order, or there is no conflict between subtasks that can be executed concurrently; the minimum unit of a subtask is determined based on the maximum amount of tasks that can be executed by the minimum resources in the hardware resource pool to execute the task.

[0019] Furthermore, the "greedy" goal described in S4 traverses all hardware resources that can execute the task from the hardware resource pool, considering whether they are currently occupied, whether they are allowed to be preempted, the estimated execution time of the task, the resource usage time, space constraints, and the resource advantage tasks, so as to allocate different most suitable subtasks to different hardware resources and achieve the optimal task allocation under the "greedy" goal.

[0020] (III) Beneficial technical effects

[0021] (1) Improving task execution efficiency and resource utilization Improving task execution efficiency: The present invention realizes the collaborative operation of devices in a resource group by dividing tasks and allocating sub-task units, which will significantly improve the efficiency of task execution.

[0022] (2) Optimize resource management: Use unified standards and specifications to evaluate resource status and attributes, and decouple the relationships between resources by breaking them up, forming a hardware resource pool that is easy to operate in a unified manner, which will improve the convenience of resource management;

[0023] (3) Stable and reliable resource supply: The present invention provides a measurable, safe, and rapidly scalable continuous resource supply pool for task segmentation and subtask allocation, which will ensure the stability and reliability of the intelligent swarm system during task execution;

[0024] (4) Achieving efficient collaboration: Through the construction of intelligent resource clusters, the present invention ensures efficient data intercommunication and on-demand resource sharing, which will enable the intelligent group system to achieve efficient collaboration when performing major tasks;

[0025] (5) Improving resource utilization: The present invention divides and allocates important tasks, effectively improving resource utilization and reducing resource waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 : A flow chart of a task segmentation and allocation method based on greedy algorithm;

[0027] Figure 2 ;Hardware resource pool construction structure diagram;

[0028] Figure 3 : Task management 5W1H principle diagram; DETAILED DESCRIPTION

[0029] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection claimed by the present invention.

[0030] Figure 1 As shown, the present invention provides a task segmentation and allocation method based on a greedy algorithm, and the specific working steps are as follows:

[0031] S1. Resource status collection and specification: Based on unified standard specifications, collect the status and attributes of hardware entity resources (such as sensor equipment, transportation equipment, computing storage equipment, communication network equipment, etc.) within the group. Status and attribute parameters include but are not limited to: owner ID, whether it is occupied, occupant ID, whether it is allowed to be preempted, occupation start time, expected occupation end time, capabilities (corresponding to the tasks that can be performed), advantageous tasks (some resources are more efficient than other resources when performing specific types of tasks, for example: CPU is more efficient when performing sequential tasks, while GPU is more suitable for parallel tasks), etc. Collect data of different structures from different resources and merge them into structured data with unified structure. The data format involves: Int, String, Float, Double, Boolean, etc. Therefore, through simple data preprocessing methods, outliers can be standardized, data formats can be unified, and resource status collection and specification can be completed. ;

[0032] S2. Hardware resource pool construction: Put the hardware entity resources and their related status and attributes into the cloud pool for unified management, which is called the hardware resource pool. At the same time, open the call interface and status attribute query interface to all devices in the group, so that all resources in the group can be interconnected and interoperable for subsequent task segmentation and subtask allocation. On this basis, the hardware resource pool updates the status and attributes of all resources in the group in real time, monitors the execution of tasks and real-time changes of resources, and dynamically adjusts the task allocation strategy to adapt to environmental changes;

[0033] S3. Task and subtask management: Create a task list, including the main task and the subtasks divided from the main task, to facilitate recording and unified management. The format adopts the 5W1H principle (Who, What, When, Where, Why, How). Who represents the hardware resources that issue the task or the hardware resources that receive the task, What represents the action that needs to be performed for the task (such as reconnaissance, calculation, transportation, etc.), When represents the time when the task is issued, Where represents the operating area for the task (if the task has no operating area, leave it blank), Why represents the notes for the task, and How represents the relevant attributes and status of the task (task priority, resource requirement type, task complexity, etc.);

[0034] S4. Task segmentation: According to the hardware resource status and attributes in the resource pool and the characteristics and requirements of the main task, match the hardware entity resource capabilities and task attributes, including considering the resource reconnaissance capability, computing power, transportation capacity, memory capacity, network bandwidth and other aspects to determine which resources can perform which tasks. At the same time, consider the nature of the task, whether it needs to be executed in parallel or sequentially, whether there are time and space restrictions, etc.; consider the dependencies between subtasks. Some subtasks may need to be executed after other subtasks are completed, while other subtasks can be executed in parallel. Establish a suitable dependency graph to ensure that subtasks are executed in the correct order, or that there are no conflicts between subtasks that can be executed concurrently. Based on the matching results of resource and task attributes, the main task is divided into a series of subtask units and added to the task list to provide a basis for subsequent task allocation. The minimum unit of this segmentation is determined based on the maximum amount of tasks that can be executed by the minimum resource in the hardware resource pool to execute the task;

[0035] S5, Task allocation: Based on the greedy algorithm, select the "greedy" goal as the criterion for subtask allocation, such as: the shortest total task execution time, the least resources occupied by the task, the minimum resource cost-effectiveness, the task priority, etc. Specifically, select the "greedy" goal, traverse all hardware resources that can execute the task from the hardware resource pool, consider whether they are occupied, whether they are allowed to be preempted, the expected execution time of the task, the resource usage time, space constraints, the resource advantage tasks, etc., so as to achieve the allocation of different most suitable subtasks for different hardware resources and achieve the optimal task allocation under the "greedy" goal;

[0036] S6, Task execution: After the hardware resource receives the subtask, it starts to execute it and updates the running status and task status to the hardware resource pool in real time to ensure the smooth execution of the task;

[0037] S7, dynamic adjustment: real-time monitoring of the status of hardware resources (e.g., whether they are running, temperature, anomalies and errors during execution, etc.), task completion status, etc. When resources are idle or unexpected circumstances prevent the task from continuing, return to step 5 to reallocate or adjust the task;

[0038] Physical equipment collaboration: Build a collaborative contract based on the service combination, issue the contract strategy to the resource cloud pool, pull the equipment resource capabilities required for a task from the resource pool, and form a super-equipment virtual terminal that performs tasks across equipment. As the subject of task execution, it achieves consensus on elements and makes collaborative work tamper-proof, thereby achieving cross-domain collaboration and high integration of various domains.

[0039] like Figure 2 As shown in the figure, the structure diagram of the hardware resource pool is as follows:

[0040] Based on unified standard specifications, the status and attributes of hardware entity resources within the group are collected. Attributes include: device ID, owner ID, basic capabilities, advantage tasks, whether there are time and space restrictions; status includes: whether occupied, occupant ID, whether preemption is allowed, occupation start time, and estimated occupation end time; if necessary, you can continue to add. Collect data of different structures from different resources and merge them into structured data with a unified structure. The data format involves: Int, String, Float, Double, Boolean, etc.; therefore, through simple data preprocessing methods, outliers can be standardized, data formats can be unified, and resource status collection and specification can be completed.

[0041] Put the hardware entity resources and their related status and attributes into the cloud pool for unified management, which is called the hardware resource pool. At the same time, the call interface and status attribute query interface are opened to all devices in the group, so that all resources in the group can be interconnected and interoperable for subsequent task segmentation and subtask allocation. On this basis, the hardware resource pool updates the status and attributes of all resources in the group in real time, monitors the execution of tasks and real-time changes of resources, and dynamically adjusts the task allocation strategy to adapt to environmental changes.

[0042] like Figure 3 The 5W1H principle diagram for task management is as follows:

[0043] Create a task list, including the main task and the subtasks divided from the main task, to facilitate recording and unified management. The format adopts the 5W1H principle (Who, What, When, Where, Why, How). Who represents the hardware resources that issue the task or the hardware resources that receive the task, What represents the action that needs to be performed for the task (such as reconnaissance, calculation, transportation, etc.), When represents the time when the task is issued, Where represents the work area for the task (if the task has no work area, leave it blank), Why represents the notes for the task, and How represents the relevant attributes and status of the task (task priority, resource requirement type, task complexity, etc.). Figure 2 As shown, it indicates that task S1001 is sent from device E1 to E2, the content is the reconnaissance area Region, the sending time is 12:01:30 on December 6, 2023, and the task priority is 5.

[0044] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A task segmentation and allocation method based on a greedy algorithm, characterized in that: The process is as follows: S1 Resource status collection and specification: Based on unified standard specifications, collect the resource status and attributes of hardware entities within the group; collect data of different structures from different resources and integrate them into structured data with unified structure; S2 Hardware Resource Pool Construction: Put the hardware entity resources and their related status and attributes into the cloud pool for unified management, which is called the hardware resource pool. All resources in the group are interconnected and interoperable; update the status and attributes of all resources in the group in real time, monitor the task execution status and real-time changes of resources, and dynamically adjust the task allocation strategy to adapt to environmental changes; S3 task and subtask management: Create a task list, including the main task and the subtasks divided from the main task; S4 Task segmentation: According to the hardware resource status and attributes in the resource pool and the characteristics and requirements of the main task, the hardware entity resource capabilities and task attributes are matched to determine which resources can perform which tasks, and the main task is segmented into a series of subtask units and added to the task list; S5 task allocation: Based on the greedy algorithm, the "greedy" target is selected as the subtask allocation, and the task with the shortest total execution time, the least resources occupied by the task, and the minimum resource cost-effectiveness is given priority; S6, Task execution: After the hardware resource receives the subtask, it starts to execute it and updates the running status and task status to the hardware resource pool in real time to ensure the smooth execution of the task; S7, dynamic adjustment: monitor the status of hardware resources and the completion of tasks in real time. When resources become idle or unexpected circumstances prevent the process from continuing, return to step S5 to reallocate or adjust tasks.

2. The method according to claim 1, characterized in that The hardware entity resources described in S1 include sensor equipment, transportation equipment, computing and storage equipment, and communication network equipment.

3. The method according to claim 1, characterized in that The parameters of resource status and attributes described in S1 include but are not limited to: owner ID, whether occupied, occupant ID, whether preemption is allowed, occupation start time, estimated occupation end time, capabilities, corresponding executable tasks, and advantage tasks.

4. The method according to claim 2, characterized in that The format of S3 main tasks and subtasks adopts the 5W1H principle: Who indicates the hardware resources that issue tasks or receive tasks, What indicates the actions that need to be performed for the task, When indicates the time when the task is issued, Where indicates the work area targeted by the task, Why indicates the notes of the task, and How indicates the relevant attributes and status of the task.

5. The method according to claim 2, characterized in that S4 hardware entity resource capabilities include resource reconnaissance capability, computing power, transportation capacity, memory capacity, and network bandwidth.

6. The method according to claim 2, characterized in that S4 task segmentation; consider the nature of the task, whether it needs to be executed in parallel or sequentially, and whether there are time and space restrictions; consider the dependencies between subtasks, some subtasks may need to be executed after other subtasks are completed, while other subtasks can be executed in parallel, and establish a suitable dependency graph to ensure that subtasks are executed in the correct order, or there is no conflict between subtasks that can be executed concurrently; the minimum unit of a subtask is determined based on the maximum amount of tasks that can be executed by the minimum resource in the hardware resource pool to execute the task.

7. The method according to claim 2, characterized in that The "greedy" goal described in S4 is to traverse all hardware resources that can execute the task from the hardware resource pool, considering whether they are currently occupied, whether they are allowed to be preempted, the expected execution time of the task, the resource usage time, space constraints, and the resource advantage tasks, so as to allocate different most suitable subtasks to different hardware resources and achieve the optimal task allocation under the "greedy" goal.

Images