A fusion terminal interaction method and system based on flexible edge service

Abstract

The present invention discloses a fusion terminal interaction method and system based on flexible edge service. The resource utilization rate of each fusion terminal is predicted by using an LSTM network, so that task scheduling and resource allocation can be optimized based on the future resource status. By setting node removal and addition rules, the number and structure of nodes in the network are flexibly adjusted according to the resource utilization of the fusion terminal, and the task execution network of the fusion terminal is flexibly configured, thereby realizing dynamic management of the fusion terminal in the task network and ensuring the stability of the task network and the balanced utilization of resources.

Description

Technical Field

[0001] The present invention relates to the technical field of power system data processing, and in particular to a fusion terminal interaction method and system based on flexible edge services. Background Art

[0002] In the field of power systems and networks, converged terminals usually refer to intelligent terminal devices that integrate multiple functions. These terminals can interact with multiple power devices, systems and users, and realize data collection, processing, transmission and control in the process. Multi-convergence terminal interaction refers to the efficient interconnection and collaborative work of different types of power terminal devices in smart grids through a unified communication and control platform. However, there are many types of converged terminals in power systems, and different devices may use different communication protocols and data formats. How to achieve compatibility and interoperability of these heterogeneous devices is an important challenge. With the continuous development of power systems, the number and types of converged terminals will continue to increase. The system needs to have good scalability to adapt to future development needs, while maintaining sufficient flexibility in structure and function. The multi-convergence terminal interaction system needs to perform efficient resource management and optimized scheduling according to the dynamic state and task requirements of the terminal equipment to improve the overall performance of the system. Summary of the invention

[0003] In view of the above problems, the present invention proposes a fusion terminal interaction method and system based on flexible edge services to optimize task scheduling and resource allocation in smart grids, achieve efficient collaboration between multiple terminals, improve system resource utilization, reduce communication delays, ensure efficient processing and stable operation of tasks in a multi-fusion terminal network, and thus enhance the intelligence and robustness of the overall system.

[0004] A technical solution of the present invention is:

[0005] A fusion terminal interaction method based on flexible edge services, the method comprising:

[0006] Determine a neighboring fusion terminal based on the task initiating fusion terminal, and determine a team leader terminal and a team member terminal from the neighboring fusion terminal and the task initiating fusion terminal;

[0007] Determine the task network based on the team leader terminal and the team member terminals, and the team leader terminal constructs and updates the task table and the terminal information table in real time;

[0008] The group leader terminal determines a candidate extended fusion terminal based on the task network;

[0009] Predicting resources of the candidate extended fusion terminals through an LSTM network based on the terminal information table, and selecting a task extended fusion terminal from the candidate extended fusion terminals according to the predicted result;

[0010] Adding the task extension fusion terminal to the task network to form a task allocation network;

[0011] Pre-allocating the tasks based on the task allocation network and the task table;

[0012] Adjusting the task allocation network based on the result of the pre-allocation to form a task execution network;

[0013] The tasks are executed in the task execution network.

[0014] Furthermore, the task table includes all currently published tasks and the status of the tasks, and the terminal information table includes the resource status, communication status of each fusion terminal in the current task network and the number of hops of the communication path between each fusion terminal and the group leader terminal.

[0015] Further, the determining of the neighboring fusion terminal based on the task initiating fusion terminal includes:

[0016] When the number of hops of the communication path between the candidate adjacent fusion terminal and the task initiating fusion terminal is less than or equal to the set maximum number of hops When the candidate adjacent fusion terminal is the adjacent fusion terminal.

[0017] Further, the determining of a team leader terminal and a team member terminal from among the neighboring fusion terminal and the task initiating fusion terminal, and determining a task network based on the team leader terminal and the team member terminal, includes:

[0018] If there are multiple group leader terminals in the adjacent fusion terminal, the task initiating fusion terminal joins the task network of the group leader terminal with the smallest number of hops in the communication path between the task initiating fusion terminal and the initiating terminal;

[0019] If there is a group leader terminal among the adjacent fusion terminals, the task initiating fusion terminal joins the task network of the group leader terminal;

[0020] If there is no group leader terminal in the adjacent fusion terminals, and there are multiple task initiating fusion terminals, in the adjacent fusion terminals, a group leader terminal is elected by comparing resources to obtain the group leader terminal, and the remaining fusion terminals after removing the group leader terminal from the adjacent fusion terminals are group member terminals, and the task initiating fusion terminal joins the task network formed by the adjacent fusion terminals;

[0021] If there is no group leader terminal among the neighboring fusion terminals, and there is only one task initiating fusion terminal, the task initiating fusion terminal is selected as the group leader terminal.

[0022] Furthermore, the election rule of the group leader terminal obtained by performing the election of the group leader terminal through resource comparison is:

[0023] ,

[0024] in, Respectively represent the task initiation fusion terminal CPU, memory, and bandwidth utilization; Initiate a fusion terminal for the task and adjacent fusion terminals The number of hops in the communication path between them; To initiate the fusion terminal in the task Set the maximum number of hops All fusion terminals within the range; and are weight parameters, which control the influence weights of resources and paths respectively;

[0025] choose The fusion terminal with the largest value is the group leader terminal.

[0026] Further, predicting the resources of the candidate extended fusion terminals through an LSTM network based on the terminal information table, and selecting a task extended fusion terminal from the candidate extended fusion terminals according to the predicted result, includes:

[0027] The input of the LSTM network is:

[0028] ,

[0029] The size is ,in ,in , indicating time CPU, memory, and bandwidth utilization of the alternative extended converged terminals;

[0030] The output of the LSTM network is:

[0031] Indicates the predicted CPU, memory and bandwidth utilization of the candidate extended converged terminals in the next time period;

[0032] According to the prediction results, the comprehensive resource utilization of the candidate extended fusion terminal is calculated. The calculation formula is as follows:

[0033] ,

[0034] in is the number of hops of the communication path from the candidate extended fusion terminal to the group leader terminal, To set the maximum number of hops, respectively represent the CPU, memory and bandwidth utilization of the candidate extended fusion terminal in the next predicted time period, is a hyperparameter and satisfies ;

[0035] All candidate extended fusion terminals are ranked according to the calculated comprehensive resource utilization rates of the candidate extended fusion terminals, and the candidate extended fusion terminals with the highest rankings are selected as the task extended fusion terminals.

[0036] Further, the pre-allocating the tasks based on the task allocation network and the task table includes:

[0037] The unassigned tasks are counted and the total amount of resources required is calculated. Then, the tasks are pre-assigned. The pre-assignment rules are as follows:

[0038] The unassigned tasks in the task table are grouped and sorted by priority, and all fusion terminals are sorted. The sorting rules are:

[0039] ,

[0040] in, is the number of hops of the communication path from the fusion terminal to the group leader terminal, They represent the CPU, memory, and bandwidth utilization of the converged terminal respectively, and n is the number of tasks. For the task Required CPU, memory, bandwidth, and task size, is a hyperparameter;

[0041] The tasks with higher priority groups are assigned to the fusion terminals with higher rankings. Each time a task is assigned, the ranking of the fusion terminals is updated. Based on the updated ranking of the fusion terminals, the remaining unassigned tasks are assigned according to the priority group ranking until the tasks are assigned or the comprehensive resource utilization of all fusion terminals is greater than the set value.

[0042] Furthermore, adjusting the task allocation network based on the pre-allocation result to form a task execution network includes:

[0043] Will The fusion terminal without assigned tasks is removed from the task assignment network;

[0044] If high priority tasks have not been assigned, and more than 60% of the fusion terminals in the task assignment network meet , then the task allocation network is expanded to increase the number of fusion terminals of the task allocation network.

[0045] Furthermore, when executing the task in the task execution network, the priority groups of unassigned tasks in the terminal information table and the task table are updated, and unassigned tasks of corresponding priorities are assigned to the fusion terminal according to the updated terminal information table and the priority groups.

[0046] Based on the above-mentioned converged terminal interaction method, the present invention also provides a converged terminal interaction system based on flexible edge services, the system comprising:

[0047] Terminal determination module: used to determine the adjacent fusion terminal based on the task initiating fusion terminal, determine the team leader terminal and the team member terminal from the adjacent fusion terminal and the task initiating fusion terminal, the team leader terminal determines the candidate extended fusion terminal based on the task network, predicts the resources of the candidate extended fusion terminal through the LSTM network based on the terminal information table, and selects the task extended fusion terminal from the candidate extended fusion terminals according to the prediction result;

[0048] Network construction module: used to determine the task network based on the team leader terminal and the team member terminal, the team leader terminal constructs and updates the task table and terminal information table in real time; adds the task extension fusion terminal to the task network to form a task allocation network, and adjusts the task allocation network based on the pre-allocation result to form a task execution network;

[0049] Task allocation module: used for pre-allocating the tasks based on the task allocation network and the task table;

[0050] Task execution module: used for executing the task in the task execution network.

[0051] The embodiment of the present invention provides a fusion terminal interaction method and system based on flexible edge service. By using the LSTM network to predict the resource utilization of each fusion terminal, task scheduling and resource allocation can be optimized based on the future resource status. By setting the node removal and addition rules, the number and structure of nodes in the network are flexibly adjusted according to the resource utilization of the fusion terminal, and the task execution network of the fusion terminal is flexibly configured, thereby realizing dynamic management of the fusion terminal in the task network and ensuring the stability of the task network and the balanced utilization of resources.

[0052] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0054] Figure 1 A specific flow chart of a fusion terminal interaction method based on flexible edge services provided by an embodiment of the present invention is shown.

[0055] Figure 2 The present invention shows an overall flow chart of a fusion terminal interaction method based on flexible edge services provided by an embodiment of the present invention.

[0056] Figure 3 A block diagram of a converged terminal interaction system based on flexible edge services provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0057] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present invention and to enable the scope of the present invention to be fully communicated to those skilled in the art.

[0058] The converged terminal can interact with multiple power equipment, systems and users, and realize data collection, processing, transmission and control in the process. Multi-convergence terminal interaction means that in the smart grid, different types of power terminal equipment can achieve efficient interconnection and collaborative work through a unified communication and control platform. In multi-convergence terminal interaction, various solutions have their own advantages. Smart grid and edge computing improve the efficiency, response speed and real-time decision-making ability of the system; distributed energy management system (DERMS) and virtual power plant (VPP) enhance the flexibility of resource scheduling and support the efficient integration of new energy; the Internet of Things and big data analysis provide comprehensive data monitoring and intelligent prediction capabilities; adaptive scheduling and optimization algorithms improve the efficiency of resource allocation and the intelligent scheduling level of the system. Although the above solutions improve the intelligence and flexibility of the system, they also have some shortcomings. Among them, the implementation cost of smart grid and edge computing is high and relies on complex infrastructure; the optimization algorithms of distributed energy management system and virtual power plant are complex, and there are challenges in coordinating decentralized resources; although the Internet of Things and big data analysis bring intelligence, data privacy and security issues are prominent; although the adaptive scheduling algorithm improves the scheduling capability, it requires a lot of computing resources and accurate data support. In general, while these solutions improve system performance, they also face problems such as implementation complexity and security risks.

[0059] Based on the above problems, the technical solution of the present invention provides a fusion terminal interaction method and system based on flexible edge services. By using an LSTM network to predict the resource utilization of each fusion terminal, task scheduling and resource allocation can be optimized based on the future resource status. By setting node removal and addition rules, the number and structure of nodes in the network can be flexibly adjusted according to the resource utilization of the fusion terminal, and the task execution network of the fusion terminal can be flexibly configured, thereby realizing dynamic management of the fusion terminal in the task network and ensuring the stability of the task network and the balanced utilization of resources.

[0060] The specific implementation includes the following:

[0061] Embodiment 1:

[0062] refer to Figure 1 , 2 , this embodiment provides a converged terminal interaction method based on flexible edge services, the method comprising:

[0063] S1: determining a neighboring fusion terminal based on the task initiating fusion terminal, and determining a team leader terminal and a team member terminal from the neighboring fusion terminal and the task initiating fusion terminal;

[0064] S2: Determine a task network based on the team leader terminal and the team member terminals, and the team leader terminal constructs and updates a task table and a terminal information table in real time;

[0065] S3: The group leader terminal determines a candidate extended fusion terminal based on the task network;

[0066] S4: predicting resources of the candidate extended fusion terminals through an LSTM network based on the terminal information table, and selecting a task extended fusion terminal from the candidate extended fusion terminals according to the predicted result;

[0067] S5: adding the task extension fusion terminal to the task network to form a task distribution network;

[0068] S6: pre-allocate the tasks based on the task allocation network and the task table;

[0069] S7: adjusting the task allocation network based on the pre-allocation result to form a task execution network;

[0070] S8: Execute the task in the task execution network.

[0071] The fusion terminal in this embodiment specifically refers to the fusion terminal used in the power system. The above-mentioned fusion terminal can be a network device with data processing function, which has corresponding sensors and communication links for collecting, processing and transmitting data of the power system. Specifically, it can be a server, PC or handheld terminal and other intelligent devices, and no specific restrictions are made in this embodiment.

[0072] For step S1, in this step, the adjacent fusion terminal is determined based on the task initiating fusion terminal, which can be specifically implemented by the following method:

[0073] When the number of hops of the communication path between the candidate adjacent fusion terminal and the task initiating fusion terminal is less than or equal to the set maximum number of hops When the candidate adjacent fusion terminal is the adjacent fusion terminal. Before initiating a task, each fusion terminal will first detect the surrounding adjacent fusion terminals as candidate adjacent fusion terminals. The definition of adjacent fusion terminals is based on the number of hops in the communication path, and a maximum number of hops is set. Only the fusion terminals within this range are considered as adjacent fusion terminals. Generally, the maximum number of hops is set Is 2 or 3.

[0074] The task initiating fusion terminal refers to the fusion terminal that first receives or initiates the task. For example, the power grid flow analysis task is generally initiated by the main control PC in the dispatching hall after receiving the dispatcher's instructions. The power grid flow analysis involves the entire power grid system and requires a large amount of data processing. Moreover, the above data processing can be performed in segments. It is slow to process with only one main control PC, and it is not timely for the processing of emergency tasks. Therefore, it is obvious that the efficiency of task processing can be improved by processing this task together with other fusion terminals near the main control PC. However, the number of fusion terminals in the entire power dispatching system is extremely large, and the division of tasks is different. How to select a suitable fusion terminal for processing? This embodiment selects a suitable fusion terminal for processing when the number of hops in the communication path between the alternative adjacent fusion terminal and the task initiating fusion terminal is less than or equal to the set maximum number of hops. When the alternative adjacent fusion terminal is used as the adjacent fusion terminal, combined with the subsequent task allocation procedure, the number and structure of nodes in the network can be flexibly adjusted according to the resource utilization of the fusion terminal, and the task execution network of the fusion terminal can be flexibly configured, thereby realizing dynamic management of the fusion terminal in the task network and ensuring the stability of the task network and the balanced utilization of resources.

[0075] In this step, the leader terminal and the member terminal are determined from the adjacent fusion terminal and the task initiating fusion terminal, and the task network is determined based on the leader terminal and the member terminal. The determination method may include the following four situations and determination methods according to the number of leader terminals in the adjacent fusion terminal and the task initiating fusion terminal:

[0076] Scenario 1: if there are multiple group leader terminals among the adjacent fusion terminals, the task initiating fusion terminal joins the task network of the group leader terminal whose communication path has the smallest number of hops with the initiating terminal;

[0077] Scenario 2: If there is a group leader terminal among the adjacent fusion terminals, the task initiating fusion terminal joins the task network of the group leader terminal;

[0078] Scenario 3: If there is no group leader terminal in the adjacent fusion terminals, and there are multiple task initiating fusion terminals, the group leader terminal is elected by resource comparison in the adjacent fusion terminals, and the remaining fusion terminals after removing the group leader terminal from the adjacent fusion terminals are group member terminals, and the task initiating fusion terminal joins the task network formed by the adjacent fusion terminals;

[0079] Scenario 4: if there is no group leader terminal among the neighboring fusion terminals and there is only one task initiating fusion terminal, the task initiating fusion terminal is selected as the group leader terminal.

[0080] In scenario 3, the election rule for the leader terminal is obtained by selecting the leader terminal through resource comparison:

[0081] ,

[0082] in, Respectively represent the task initiation fusion terminal CPU, memory, and bandwidth utilization; Initiate a fusion terminal for the task and adjacent fusion terminals The number of hops in the communication path between them; To initiate the fusion terminal in the task Set the maximum number of hops All fusion terminals within the range; and is a weight parameter, which controls the influence weight of resources and paths respectively; the final selection The fusion terminal with the largest value is the group leader terminal.

[0083] After the team leader terminal and the team member terminal are determined in step S1, step S2 is executed to determine the task network based on the team leader terminal and the team member terminal. The team leader terminal constructs and updates the task table and the terminal information table in real time.

[0084] The task table includes all currently published tasks and the status of the tasks, including the task initiation terminal, task initiation time, task priority, resources required for the task (such as CPU, memory, bandwidth) and the task running terminal. When a new task is published or the status of an existing task changes, the team leader terminal updates the task table in real time to ensure the accuracy of task scheduling. Tasks in the power system can be divided into the following three priorities, specifically:

[0085] Highest priority: real-time data acquisition, load monitoring, state estimation, fault detection and location;

[0086] Medium priority: predictive analysis, optimized scheduling;

[0087] Lower priority: maintenance and inspection, data transfer.

[0088] The terminal information table includes the resource status, communication status, and hop count of the communication path between each fusion terminal and the group leader terminal in the current task network. When the group leader terminal communicates with other fusion terminals, the number of routing nodes in its communication path is recorded as the communication hop count. The group leader terminal periodically communicates to obtain the above information and updates the terminal information table in real time. At the same time, the old terminal information table is stored in a first-in-first-out (FIFO) manner.

[0089] During the task execution, if the status of a fusion terminal changes significantly (such as resource exhaustion, communication failure), the team leader terminal should update the terminal information table in time, including recalculating the communication distance, removing the fusion terminal that is no longer available, or adding a new fusion terminal to keep the task going smoothly.

[0090] The team leader terminal needs to maintain communication with all fusion terminals in the mission network and continuously monitor their status changes. Any status update should be immediately reflected in the terminal information table. The fusion terminals in the mission network should be updated at regular intervals. Communicate with the team leader terminal, including the resource status and task status of the converged terminal. The team leader terminal sends Time to perform a task scheduling (i.e. the subsequent steps), where and If a fusion terminal fails to communicate with the group leader terminal for three consecutive times, the fusion terminal is considered to be faulty, the group leader terminal disconnects it, removes it from the terminal information table, and sets the unfinished tasks in the task table whose task running terminal is the fusion terminal to be assigned, with the priority set to the highest priority.

[0091] When a converged terminal needs to obtain data from other converged terminals, it first determines which converged terminals the required data is located on through the IoT management platform, and then communicates with these converged terminals to obtain the data. The fusion terminals within the range are added to the task network of the team leader terminal.

[0092] After determining the task network, step S3 is executed, and the team leader terminal determines the candidate extended fusion terminal based on the task network. The team leader terminal expands to the surrounding fusion terminals through the task network topology. During the initial expansion, the number of expanded fusion terminals is set to twice the number of fusion terminals in the task network as the candidate extended fusion terminals.

[0093] After obtaining the candidate extended fusion terminals, the team leader terminal selects the fusion terminals according to the following priority rules and reserves half of them to establish connections as task extended fusion terminals. Computing resource priority: Prioritize fusion terminals with higher computing resources (CPU, memory, bandwidth). Communication path priority: Prioritize fusion terminals with smaller communication path weights with the team leader terminal.

[0094] The specific sorting method is the above step S4, predicting the resources of the candidate extended fusion terminals through the LSTM network based on the terminal information table, and selecting the task extended fusion terminal from the candidate extended fusion terminals according to the predicted result, including:

[0095] The LSTM network of this embodiment includes: an input layer, an LSTM layer, and a first fully connected layer. The neural units of the previous LSTM layer are fully connected with each neuron in the first fully connected layer. A Dropout1 layer is added on top of the first fully connected layer to Some neurons are discarded, and the remaining neurons are The probability of is retained, a fully connected layer is added above the Dropout1 layer, and a Dropout2 layer is added above the second fully connected layer, with probability Some neurons are discarded, and the remaining neurons are The probability of retaining is retained, and a fully connected layer with 10 neurons is added above the Dropout2 layer to ensure that the output result is a continuous prediction value. A regression layer is added above the fourth fully connected layer. The regression layer is the output layer at the end, which outputs the CPU, memory, and bandwidth utilization of the fusion terminal in the next time period.

[0096] The input of the LSTM network is:

[0097] ,

[0098] The size is ,in ,in , indicating time CPU, memory, and bandwidth utilization of the alternative extended converged terminals;

[0099] The output of the LSTM network is:

[0100] Indicates the predicted CPU, memory and bandwidth utilization of the candidate extended converged terminals in the next time period;

[0101] According to the prediction results, the comprehensive resource utilization of the candidate extended fusion terminal is calculated. The calculation formula is as follows:

[0102] ,

[0103] in is the number of hops of the communication path from the candidate extended fusion terminal to the group leader terminal, To set the maximum number of hops, respectively represent the CPU, memory and bandwidth utilization of the candidate extended fusion terminal in the next predicted time period, is a hyperparameter and satisfies ;

[0104] All candidate extended fusion terminals are ranked according to the calculated comprehensive resource utilization rates of the candidate extended fusion terminals, and the candidate extended fusion terminals with the highest rankings are selected as the task extended fusion terminals.

[0105] After the task extension fusion terminal is acquired, step S5 is executed: the task extension fusion terminal is added to the task network to form a task distribution network.

[0106] After the task allocation network is obtained, step S6 is executed to pre-allocate tasks based on the task allocation network and the task table, including:

[0107] The unassigned tasks are counted and the total amount of resources required is calculated. Then, the tasks are pre-assigned. The pre-assignment rules are as follows:

[0108] The unassigned tasks in the task table are grouped by priority, and all terminals are sorted. The sorting rules are:

[0109] ,

[0110] in, is the number of hops of the communication path from the fusion terminal to the group leader terminal, They represent the CPU, memory, and bandwidth utilization of the converged terminal respectively, and n is the number of tasks. For the task Required CPU, memory, bandwidth, and task size, is a hyperparameter;

[0111] The tasks with higher priority groups are assigned to the fusion terminals with higher rankings. Each time a task is assigned, the ranking of the fusion terminals is updated. Based on the updated ranking of the fusion terminals, the remaining tasks are assigned according to the priority group ranking until the tasks are assigned or the comprehensive resource utilization of all fusion terminals is greater than the set value.

[0112] In the specific task allocation process, it can be known from the terminal information table that the task allocation of some fusion terminals is unreasonable, or the task is too heavy, or the task is too light, so step S7 is also performed: adjusting the task allocation network based on the result of the pre-allocation to form a task execution network, specifically including:

[0113] Will If the high-priority tasks have not been assigned, and more than 60% of the fusion terminals in the task allocation network meet the requirements, , then the task allocation network is expanded to increase the number of fusion terminals of the task allocation network.

[0114] Finally, step S8 is performed: executing the task in the task execution network. When executing the task in the task execution network, the priority groups of the unassigned tasks in the terminal information table and the task table are updated, and the unassigned tasks of corresponding priorities are assigned to the fusion terminal according to the updated terminal information table and the priority groups.

[0115] According to the above task allocation results, the tasks are distributed to each fusion terminal for execution. If all the tasks in the task list have been completed and If no new tasks are added to the task list within a certain period of time, the task network will be disbanded.

[0116] Embodiment 2:

[0117] refer to Figure 3 Based on the above-mentioned fusion terminal interaction method based on flexible edge service, the present invention also provides a fusion terminal interaction system based on flexible edge service, the system comprising:

[0118] Terminal determination module: used to determine the adjacent fusion terminal based on the task initiating fusion terminal, determine the team leader terminal and the team member terminal from the adjacent fusion terminal and the task initiating fusion terminal, the team leader terminal determines the candidate extended fusion terminal based on the task network, predicts the resources of the candidate extended fusion terminal through the LSTM network based on the terminal information table, and selects the task extended fusion terminal from the candidate extended fusion terminals according to the prediction result;

[0119] Network construction module: used to determine the task network based on the team leader terminal and the team member terminal, the team leader terminal constructs and updates the task table and terminal information table in real time; adds the task extension fusion terminal to the task network to form a task allocation network, and adjusts the task allocation network based on the pre-allocation result to form a task execution network;

[0120] Task allocation module: used for pre-allocating the tasks based on the task allocation network and the task table;

[0121] Task execution module: used for executing the task in the task execution network.

[0122] The specific implementation method of this embodiment is the same as that of Embodiment 1, which will not be repeated here. Please refer to the description of Embodiment 1 for details.

[0123] The method and system of the above-mentioned embodiment are adopted, based on the intelligent fusion terminal, and the LSTM network is used to predict the resource utilization of each fusion terminal, so that task scheduling and resource allocation can be optimized based on the future resource status. By introducing the number of communication path hops as a key factor and comprehensively considering the resource utilization, the present invention effectively reduces the communication delay and resource waste during the task execution process, thereby improving the overall performance of the system. By setting the removal and addition rules of the nodes, the dynamic management of the fusion terminals in the task network is realized. According to the resource utilization of the fusion terminal, the number and structure of nodes in the network are flexibly adjusted to ensure the stability of the task network and the balanced utilization of resources.

[0124] The present invention can effectively solve common task scheduling, resource management, network expansion and node management problems in multi-convergent terminal collaborative task networks, and improve the resource utilization, task processing efficiency and network stability of the system. This not only helps to optimize system performance in practical applications, but also provides an intelligent and dynamic solution for future multi-convergent terminal collaborative systems.

[0125] Those skilled in the art will appreciate that although some embodiments herein include certain features included in other embodiments but not other features, the combination of features from different embodiments is meant to be within the scope of the present invention and to form different embodiments.

[0126] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A fusion terminal interaction method based on flexible edge services, characterized in that: The method comprises: Determine a neighboring fusion terminal based on the task initiating fusion terminal, and determine a team leader terminal and a team member terminal from the neighboring fusion terminal and the task initiating fusion terminal; Determine the task network based on the team leader terminal and the team member terminals, and the team leader terminal constructs and updates the task table and the terminal information table in real time; The group leader terminal determines a candidate extended fusion terminal based on the task network; Predicting resources of the candidate extended fusion terminals through an LSTM network based on the terminal information table, and selecting a task extended fusion terminal from the candidate extended fusion terminals according to the predicted result; Adding the task extension fusion terminal to the task network to form a task allocation network; Pre-allocating the tasks based on the task allocation network and the task table; Adjusting the task allocation network based on the result of the pre-allocation to form a task execution network; executing the task in the task execution network; The predicting the resources of the candidate extended fusion terminals through the LSTM network based on the terminal information table, and selecting the task extended fusion terminal from the candidate extended fusion terminals according to the predicted result, comprises: The input of the LSTM network is: ， The size is ,in ,in , indicating time CPU, memory, and bandwidth utilization of the alternative extended converged terminals; The output of the LSTM network is: Indicates the predicted CPU, memory and bandwidth utilization of the candidate extended converged terminals in the next time period; According to the prediction results, the comprehensive resource utilization of the candidate extended fusion terminal is calculated. The calculation formula is as follows: ， in is the number of hops of the communication path from the candidate extended fusion terminal to the group leader terminal, To set the maximum number of hops, respectively represent the CPU, memory and bandwidth utilization of the candidate extended fusion terminal in the next predicted time period, is a hyperparameter and satisfies ; All candidate extended fusion terminals are ranked according to the calculated comprehensive resource utilization rates of the candidate extended fusion terminals, and the candidate extended fusion terminals with the highest rankings are selected as the task extended fusion terminals.

2. According to claim 1, a fusion terminal interaction method based on flexible edge services is characterized in that: The task table includes all currently published tasks and the status of the tasks, and the terminal information table includes the resource status, communication status of each fusion terminal in the current task network and the number of hops of the communication path between each fusion terminal and the group leader terminal.

3. The method for integrated terminal interaction based on flexible edge services according to claim 1, characterized in that: The determining of the adjacent fusion terminal based on the task initiating fusion terminal includes: When the number of hops of the communication path between the candidate adjacent fusion terminal and the task initiating fusion terminal is less than or equal to the set maximum number of hops When the candidate adjacent fusion terminal is the adjacent fusion terminal.

4. The method for integrated terminal interaction based on flexible edge services according to claim 1, characterized in that: The step of determining a group leader terminal and a group member terminal from among the neighboring fusion terminals and the task initiating fusion terminal, and determining a task network based on the group leader terminal and the group member terminals includes: If there are multiple group leader terminals in the adjacent fusion terminal, the task initiating fusion terminal joins the task network of the group leader terminal whose communication path with the task initiating fusion terminal has the smallest number of hops; If there is a group leader terminal among the adjacent fusion terminals, the task initiating fusion terminal joins the task network of the group leader terminal; If there is no group leader terminal in the adjacent fusion terminals, and there are multiple task initiating fusion terminals, in the adjacent fusion terminals, a group leader terminal is elected by comparing resources to obtain the group leader terminal, and the remaining fusion terminals after removing the group leader terminal from the adjacent fusion terminals are group member terminals, and the task initiating fusion terminal joins the task network formed by the adjacent fusion terminals; If there is no group leader terminal among the neighboring fusion terminals, and there is only one task initiating fusion terminal, the task initiating fusion terminal is selected as the group leader terminal.

5. The method for integrated terminal interaction based on flexible edge services according to claim 4, characterized in that: The election rule of the group leader terminal obtained by the resource comparison is: ， in, Respectively represent the task initiation fusion terminal CPU, memory, and bandwidth utilization; Initiate a fusion terminal for the task and adjacent fusion terminals The number of hops in the communication path between them; To initiate the fusion terminal in the task Set the maximum number of hops All fusion terminals within the range; and are weight parameters, which control the influence weights of resources and paths respectively; choose The fusion terminal with the largest value is the group leader terminal.

6. The method for integrated terminal interaction based on flexible edge services according to claim 1, characterized in that: The pre-allocating the tasks based on the task allocation network and the task table includes: The unassigned tasks are counted and the total amount of resources required is calculated. Then, the tasks are pre-assigned. The pre-assignment rules are as follows: The unassigned tasks in the task table are grouped and sorted by priority, and all fusion terminals are sorted. The sorting rules are: ， in, is the number of hops of the communication path from the fusion terminal to the group leader terminal, They represent the CPU, memory, and bandwidth utilization of the converged terminal respectively, and n is the number of tasks. For the task Required CPU, memory, bandwidth, and task size, is a hyperparameter; The tasks with higher priority groups are assigned to the fusion terminals with higher rankings. Each time a task is assigned, the ranking of the fusion terminals is updated. Based on the updated ranking of the fusion terminals, the remaining unassigned tasks are assigned according to the priority group ranking until the tasks are assigned or the comprehensive resource utilization of all fusion terminals is greater than the set value.

7. The method for integrated terminal interaction based on flexible edge services according to claim 1, characterized in that: The step of adjusting the task allocation network based on the result of the pre-allocation to form a task execution network includes: Will The fusion terminal without assigned tasks is removed from the task assignment network; If high priority tasks have not been assigned, and more than 60% of the fusion terminals in the task assignment network meet , then the task allocation network is expanded to increase the number of fusion terminals of the task allocation network.

8. The method for integrated terminal interaction based on flexible edge services according to claim 1, characterized in that: When executing the task in the task execution network, the priority groups of unassigned tasks in the terminal information table and the task table are updated, and unassigned tasks of corresponding priorities are assigned to the fusion terminal according to the updated terminal information table and the priority groups.

9. A fusion terminal interaction system based on flexible edge services, characterized in that: The system comprises: Terminal determination module: used to determine the adjacent fusion terminal based on the task initiating fusion terminal, determine the team leader terminal and the team member terminal from the adjacent fusion terminal and the task initiating fusion terminal, the team leader terminal determines the candidate extended fusion terminal based on the task network, predicts the resources of the candidate extended fusion terminal through the LSTM network based on the terminal information table, and selects the task extended fusion terminal from the candidate extended fusion terminals according to the prediction result; Network construction module: used to determine the task network based on the team leader terminal and the team member terminal, the team leader terminal constructs and updates the task table and terminal information table in real time; adds the task extension fusion terminal to the task network to form a task allocation network, and adjusts the task allocation network based on the pre-allocation result to form a task execution network; Task allocation module: used for pre-allocating the tasks based on the task allocation network and the task table; Task execution module: used for executing the task in the task execution network; The predicting the resources of the candidate extended fusion terminals through the LSTM network based on the terminal information table, and selecting the task extended fusion terminal from the candidate extended fusion terminals according to the predicted result, comprises: The input of the LSTM network is: ， The size is ,in ,in , indicating time CPU, memory, and bandwidth utilization of the alternative extended converged terminals; The output of the LSTM network is: Indicates the predicted CPU, memory and bandwidth utilization of the candidate extended converged terminals in the next time period; According to the prediction results, the comprehensive resource utilization of the candidate extended fusion terminal is calculated. The calculation formula is as follows: ， in is the number of hops of the communication path from the candidate extended fusion terminal to the group leader terminal, To set the maximum number of hops, respectively represent the CPU, memory and bandwidth utilization of the candidate extended fusion terminal in the next predicted time period, is a hyperparameter and satisfies ; All candidate extended fusion terminals are ranked according to the calculated comprehensive resource utilization rates of the candidate extended fusion terminals, and the candidate extended fusion terminals with the highest rankings are selected as the task extended fusion terminals.

Images