Multi-agent task coordination method and device, computer equipment, readable storage medium and program product

Abstract

The invention relates to a multi-agent task coordination method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the steps of obtaining task information of a to-be-processed task; analyzing the task information, and decomposing the to-be-processed task into a plurality of sub-tasks to obtain sub-task information corresponding to each sub-task; issuing each piece of subtask information to a block chain; obtaining a task response scheme corresponding to each subtask; the task response scheme is obtained by the candidate agent node corresponding to the subtask according to subtask information and node evaluation information obtained from the block chain and resource state information of the candidate agent node and is submitted to the block chain; according to the task response scheme of each candidate agent node corresponding to the sub-task, obtaining a task allocation result of each sub-task; and issuing the task allocation result to the block chain. By adopting the method, the reliability of multi-agent cooperation can be improved.

Description

Technical Field

[0001] This application relates to the field of intelligent agent collaboration, and in particular to a method, apparatus, computer device, computer-readable storage medium, and computer program product for multi-agent task coordination. Background Technology

[0002] With the development of information technology, intelligent agents with autonomous perception, decision-making and execution capabilities are gradually being applied to various scenarios such as the Industrial Internet of Things and distributed robots.

[0003] In related technologies, distributed agent collaboration is typically used to process complex tasks, with agents cooperating to divide and execute tasks. However, due to the lack of a reliable trust foundation among agents, the reliability of their collaboration is difficult to guarantee. Summary of the Invention

[0004] Therefore, it is necessary to provide a multi-agent task coordination method, apparatus, computer device, computer-readable storage medium, and computer program product to address the aforementioned technical problems.

[0005] Firstly, this application provides a multi-agent task coordination method, including:

[0006] Obtain task information for tasks to be processed; the task information is submitted to the blockchain by the intelligent agent node.

[0007] The task information is parsed, and the task to be processed is decomposed into multiple sub-tasks to obtain sub-task information corresponding to each sub-task.

[0008] Publish the information of each subtask to the blockchain;

[0009] Obtain the task response scheme corresponding to each of the sub-tasks; the task response scheme is submitted to the blockchain by the candidate intelligent agent node corresponding to the sub-task; the task response scheme is obtained by the candidate intelligent agent node based on the sub-task information and node evaluation information obtained from the blockchain and the resource status information of the candidate intelligent agent node;

[0010] Based on the task response scheme of each candidate agent node corresponding to the subtask, the task allocation result of each subtask is obtained; the task allocation result is used to indicate the execution agent node corresponding to each subtask.

[0011] The task allocation results are published to the blockchain.

[0012] In one embodiment, obtaining the task allocation result of each subtask according to the task response scheme corresponding to each subtask includes: obtaining the node evaluation information of each candidate agent node corresponding to the subtask, as well as the estimated timeliness information and estimated resource information of the candidate agent node for the subtask, according to the task response scheme corresponding to each subtask; obtaining the matching degree information between the candidate agent node and the subtask according to the node evaluation information, the estimated timeliness information and the estimated resource information; determining the execution agent node corresponding to the subtask according to the matching degree information of each candidate agent node; and obtaining the task allocation result of the subtask according to the execution agent node.

[0013] In one embodiment, after recording and publishing the task allocation result to the blockchain, the process includes: if a reassignment request for the subtask is detected, updating the task allocation result of the subtask according to the task response scheme of each subtask; wherein the reassignment request is submitted to the blockchain by the execution agent node corresponding to the subtask; and publishing the updated task allocation result to the blockchain.

[0014] In one embodiment, the method further includes: obtaining task execution data of each of the intelligent agent nodes from the blockchain; obtaining node evaluation information of the intelligent agent nodes based on the task execution data; and publishing the node evaluation information to the blockchain.

[0015] In one embodiment, the task execution data includes the agent node's task completion record, task data verification record, error log information, execution timeliness information, and task collaboration record. Obtaining the agent node's node evaluation information based on the task execution data includes: obtaining the agent node's execution quality index based on the task completion record, the task data verification record, and the error log information; obtaining the agent node's response timeliness index based on the execution timeliness information; obtaining the agent node's collaboration contribution index based on the task collaboration record; and weighting and fusing the execution quality index, response timeliness index, and collaboration contribution index according to preset node evaluation weights to obtain the agent node's node evaluation information.

[0016] In one embodiment, the method further includes: if the node evaluation information of the agent node meets a preset incentive condition, then executing a node incentive measure for the agent node; if the node evaluation information of the agent node meets a preset penalty condition, then executing a node penalty measure for the agent node.

[0017] Secondly, this application also provides a multi-agent task coordination device, comprising:

[0018] The first acquisition module is used to acquire task information of the task to be processed; the task information is submitted to the blockchain by the intelligent agent node.

[0019] The task parsing module is used to parse the task information, decompose the task to be processed into multiple sub-tasks, and obtain sub-task information corresponding to each sub-task.

[0020] The first publishing module is used to publish the information of each subtask to the blockchain;

[0021] The second acquisition module is used to acquire the task response scheme corresponding to each of the sub-tasks; the task response scheme is submitted to the blockchain by the candidate intelligent agent node corresponding to the sub-task; the task response scheme is obtained by the candidate intelligent agent node based on the sub-task information and node evaluation information acquired from the blockchain and the resource status information of the candidate intelligent agent node;

[0022] The task allocation module is used to obtain the task allocation result of each subtask according to the task response scheme of each candidate agent node corresponding to the subtask; the task allocation result is used to indicate the execution agent node corresponding to each subtask.

[0023] The second publishing module is used to publish the task allocation results to the blockchain.

[0024] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0025] Obtain task information for tasks to be processed; the task information is submitted to the blockchain by the intelligent agent node.

[0026] The task information is parsed, and the task to be processed is decomposed into multiple sub-tasks to obtain sub-task information corresponding to each sub-task.

[0027] Publish the information of each subtask to the blockchain;

[0028] Obtain the task response scheme corresponding to each of the sub-tasks; the task response scheme is submitted to the blockchain by the candidate intelligent agent node corresponding to the sub-task; the task response scheme is obtained by the candidate intelligent agent node based on the sub-task information and node evaluation information obtained from the blockchain and the resource status information of the candidate intelligent agent node;

[0029] Based on the task response scheme of each candidate agent node corresponding to the subtask, the task allocation result of each subtask is obtained; the task allocation result is used to indicate the execution agent node corresponding to each subtask.

[0030] The task allocation results are published to the blockchain.

[0031] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0032] Obtain task information for tasks to be processed; the task information is submitted to the blockchain by the intelligent agent node.

[0033] The task information is parsed, and the task to be processed is decomposed into multiple sub-tasks to obtain sub-task information corresponding to each sub-task.

[0034] Publish the information of each subtask to the blockchain;

[0035] Obtain the task response scheme corresponding to each of the sub-tasks; the task response scheme is submitted to the blockchain by the candidate intelligent agent node corresponding to the sub-task; the task response scheme is obtained by the candidate intelligent agent node based on the sub-task information and node evaluation information obtained from the blockchain and the resource status information of the candidate intelligent agent node;

[0036] Based on the task response scheme of each candidate agent node corresponding to the subtask, the task allocation result of each subtask is obtained; the task allocation result is used to indicate the execution agent node corresponding to each subtask.

[0037] The task allocation results are published to the blockchain.

[0038] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0039] Obtain task information for tasks to be processed; the task information is submitted to the blockchain by the intelligent agent node.

[0040] The task information is parsed, and the task to be processed is decomposed into multiple sub-tasks to obtain sub-task information corresponding to each sub-task.

[0041] Publish the information of each subtask to the blockchain;

[0042] Obtain the task response scheme corresponding to each of the sub-tasks; the task response scheme is submitted to the blockchain by the candidate intelligent agent node corresponding to the sub-task; the task response scheme is obtained by the candidate intelligent agent node based on the sub-task information and node evaluation information obtained from the blockchain and the resource status information of the candidate intelligent agent node;

[0043] Based on the task response scheme of each candidate agent node corresponding to the subtask, the task allocation result of each subtask is obtained; the task allocation result is used to indicate the execution agent node corresponding to each subtask.

[0044] The task allocation results are published to the blockchain.

[0045] The aforementioned multi-agent task coordination method, apparatus, computer device, computer-readable storage medium, and computer program product first acquire task information of the pending tasks submitted to the blockchain by agent nodes. Then, they parse the task information to decompose the pending tasks into multiple sub-tasks and obtain sub-task information corresponding to each sub-task. Next, they publish the sub-task information to the blockchain. Subsequently, they acquire task response schemes submitted to the blockchain by candidate agent nodes corresponding to each sub-task. These task response schemes are obtained by the candidate agent nodes based on the sub-task information, node evaluation information, and resource status information obtained from the blockchain. Then, based on the task response schemes of each candidate agent node corresponding to the sub-task, they obtain task allocation results to instruct the executing agent nodes corresponding to each sub-task. Finally, they publish the task allocation results to the blockchain.

[0046] This scheme, by storing task information, sub-task information, task response plans, and sub-task allocation results on the blockchain, effectively ensures data trustworthiness and tamper-proofness during the collaboration process. It also enables information sharing and synchronization among intelligent agent nodes via the blockchain, effectively guaranteeing the consistency and non-repudiation of task coordination decisions. Furthermore, by having candidate intelligent agent nodes derive task response plans for sub-tasks based on sub-task information, node evaluation information, and resource status information, and then coordinating and allocating sub-tasks based on these plans, the scheme incorporates trusted node evaluation information and self-provided resource status information from the intelligent agents on the blockchain into the allocation decision, improving the fairness and efficiency of task allocation. Therefore, by introducing blockchain technology into distributed intelligent agent collaboration, this scheme leverages the decentralized, tamper-proof, and consensus-based data consistency characteristics of blockchain to provide a reliable trust foundation for multi-agent collaboration, thereby enhancing its reliability. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a diagram illustrating the application environment of a multi-agent task coordination method in one embodiment.

[0049] Figure 2 This is a flowchart illustrating a multi-agent task coordination method in one embodiment;

[0050] Figure 3 This is a flowchart illustrating the process of obtaining the task allocation results for each subtask in one embodiment;

[0051] Figure 4 This is a schematic diagram of the process for obtaining node evaluation information of an agent node in one embodiment;

[0052] Figure 5 This is a flowchart illustrating a multi-agent task coordination method in another embodiment;

[0053] Figure 6 This is a schematic diagram of the structure of a distributed intelligent agent cooperative system in one embodiment;

[0054] Figure 7 This is a structural block diagram of a multi-agent task coordination device in one embodiment;

[0055] Figure 8 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0057] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0058] The multi-agent task coordination method provided in this application can be applied to, for example... Figure 1 The application environment shown may include multiple nodes of a blockchain, which may include multiple agent nodes. In this application, a node of the blockchain can act as a task coordination participant. After a task coordination participant joins the blockchain, its system corresponds to a node of the blockchain. This node can execute the operation of the multi-agent task coordination method according to the smart contracts deployed on the blockchain. The blockchain may include multiple agent nodes, and each agent node may correspond to one or more agents. Figure 1 In the application environment shown, intelligent agents can be, but are not limited to, various IoT devices, robotic devices, etc. Blockchain nodes can be servers or intelligent agents.

[0059] In one exemplary embodiment, such as Figure 2 As shown, a multi-agent task coordination method is provided, which can be applied to... Figure 1 Taking a node in the example, the explanation includes the following steps:

[0060] Step S201: Obtain task information for the task to be processed. This task information is submitted to the blockchain by the intelligent agent node.

[0061] The task to be processed can be initiated by a smart agent node on the blockchain. This smart agent node can encapsulate the task information into a transaction and submit it to the blockchain. The task information may include, but is not limited to, task objectives, task requirements, and task deadlines. Nodes on the blockchain can obtain the task information from the blockchain network.

[0062] Step S202: Parse the task information, decompose the task to be processed into multiple subtasks, and obtain the subtask information corresponding to each subtask.

[0063] After obtaining task information, the node can parse the information to break down the task to be processed into multiple subtasks.

[0064] For example, a node can decompose the task to be processed based on the task objective, task requirements, and agent function information corresponding to each agent node in the blockchain network contained in the task information, to obtain multiple subtasks that can be executed by the agent corresponding to a single agent node. After decomposing into multiple subtasks, a unique identifier and task parameters can be generated for each subtask, and the subtask information can be obtained based on the unique identifier and task parameters. For example, the task parameters of a subtask may include, but are not limited to, priority parameters, resource requirement parameters, and task deadlines.

[0065] Step S203: Publish the information of each subtask to the blockchain.

[0066] After obtaining the subtask information for each subtask, this subtask information can be published to the blockchain so that all nodes in the blockchain network can obtain the subtask information.

[0067] Step S204: Obtain the task response plan corresponding to each subtask. This task response plan is submitted to the blockchain by the candidate intelligent agent node corresponding to the subtask. The task response plan is obtained by the candidate intelligent agent node based on the subtask information, node evaluation information, and resource status information obtained from the blockchain.

[0068] In this context, the candidate agent nodes corresponding to a subtask can be agent nodes in the blockchain network that have submitted task response schemes for that subtask. These task response schemes may include, but are not limited to, the agent nodes' evaluation information, estimated timeliness information, and estimated resource information for that subtask. Nodes on the blockchain can obtain the task response schemes submitted by the candidate agent nodes for each subtask from the blockchain network.

[0069] Specifically, intelligent agent nodes in a blockchain network can obtain sub-task information by listening to the blockchain and determine whether they meet the task requirements of the sub-task based on their own intelligent agent function information, resource status information, etc. When an intelligent agent node determines that it meets the task requirements of the sub-task, it can derive a task response plan for the sub-task based on the sub-task information and node evaluation information obtained from the blockchain, as well as its own resource status information, and submit the task response plan to the blockchain in the form of a transaction.

[0070] For example, an intelligent agent node can determine the estimated resource information for a subtask based on the resource requirement parameters included in the subtask information and its own resource status information, including computing power, storage resources, and energy reserves. It can also determine the estimated timeliness information for the subtask based on the task deadline included in the subtask information and its own resource status information, including current load. Furthermore, it can obtain its own node evaluation information from the blockchain. Finally, based on the estimated resource information, estimated timeliness information, and node evaluation information, a task response plan for the subtask is derived. For example, the estimated resource information can be used to indicate the resources required for the intelligent agent corresponding to the intelligent agent node to execute the subtask; the estimated timeliness information can be used to indicate the estimated completion time for the intelligent agent node to execute the subtask; and the node evaluation information can be evaluation information obtained based on the intelligent agent node's historical task execution data, which can be used to indicate the intelligent agent node's overall performance in executing various subtasks historically.

[0071] Step S205: Based on the task response schemes of each candidate agent node corresponding to the subtask, obtain the task allocation result for each subtask. This task allocation result is used to indicate the executing agent node corresponding to each subtask.

[0072] For each subtask, task allocation can be performed based on the task response schemes of its corresponding candidate agent nodes. The executing agent node for the subtask is then selected from multiple candidate agent nodes, resulting in the task allocation result for that subtask. For example, in task allocation, candidate agent nodes can be selected according to preset rules based on information such as node evaluation information, estimated timeliness information, and estimated resource information included in the task response scheme. For instance, different sorting priorities can be set for node evaluation information, estimated timeliness information, and estimated resource information, and candidate agent nodes can be sorted and selected according to these priorities to identify executing agent nodes with good historical performance, high execution efficiency, and low resource costs.

[0073] Step S206: Publish the task allocation results to the blockchain.

[0074] After obtaining the task allocation results for each subtask, these results can be published to the blockchain so that all nodes in the blockchain network can access them. The executing agent nodes in the blockchain network can then execute the corresponding subtask through their respective agents after obtaining the task allocation results.

[0075] In the aforementioned multi-agent task coordination method, by storing task information, sub-task information, task response plans, and sub-task allocation results on the blockchain, the reliability and immutability of data during the collaboration process can be effectively guaranteed. Furthermore, it allows each agent node to share and synchronize information through the blockchain, effectively ensuring the consistency and non-repudiation of task coordination decisions. Simultaneously, by having candidate agent nodes derive task response plans for sub-tasks based on sub-task information, node evaluation information, and resource status information, and then coordinating and allocating sub-tasks based on these plans, the reliable node evaluation information and self-provided resource status information of the agents on the blockchain can be incorporated into the allocation decision, improving the fairness and efficiency of task allocation. Therefore, by introducing blockchain technology into distributed agent collaboration, this scheme leverages the decentralized, tamper-proof, and consensus-based data consistency characteristics of blockchain to provide a reliable trust foundation for multi-agent collaboration, thereby improving its reliability.

[0076] In one exemplary embodiment, such as Figure 3As shown, based on the task response schemes corresponding to each subtask, the task allocation results for each subtask can be obtained, including:

[0077] Step S301: Based on the task response schemes corresponding to each subtask, obtain the node evaluation information of each candidate intelligent agent node corresponding to the subtask, as well as the estimated timeliness information and estimated resource information of the candidate intelligent agent nodes for the subtask.

[0078] Specifically, based on the task response scheme corresponding to the sub-task, the node evaluation information of the candidate intelligent agent node corresponding to the task response scheme, as well as the estimated timeliness information and estimated resource information of the node for the sub-task can be obtained.

[0079] The estimated resource information can be used to indicate the resources required for the agent corresponding to the candidate agent node to execute the subtask. For example, the estimated resource information may include internal resource requirement information and external resource requirement information. The internal resource requirement information can be used to indicate the internal resources required for the agent corresponding to the agent node to execute the subtask, and the external resource requirement information can be used to indicate the external resources required for the agent corresponding to the agent node to execute the subtask. These external resources may be resources shared by all agent nodes in the blockchain network.

[0080] Among them, the estimated timeliness information can be used to indicate the estimated completion time of the agent corresponding to the candidate agent node in executing the sub-task; the node evaluation information can be information stored on the blockchain to indicate the comprehensive performance of the agent node in executing various sub-tasks in history, which can be obtained from the candidate agent node's historical task execution data.

[0081] Step S302: Based on the node evaluation information, estimated timeliness information, and estimated resource information, obtain the matching degree information between the candidate intelligent agent node and the subtask.

[0082] Among them, the matching degree information between each candidate agent node and the subtask can be obtained based on the node evaluation information, estimated timeliness information and estimated resource information of each candidate agent node corresponding to the subtask.

[0083] For example, in this step, the node evaluation index, response time index and resource cost index of the candidate intelligent agent node can be obtained first based on the node evaluation information, estimated timeliness information and estimated resource information of the candidate intelligent agent node, and then the matching degree information between the candidate intelligent agent node and the subtask can be obtained by weighted fusion.

[0084] For example, based on the node evaluation information of the candidate intelligent agent node, a node score can be obtained for that candidate intelligent agent node, and this score can be used as a node evaluation index. A higher index value indicates better historical performance of the node. For example, based on the estimated timeliness information of the candidate intelligent agent node and the task deadline of the subtask, the estimated task completion time of the candidate intelligent agent node can be calculated, and a corresponding response timeliness index can be obtained. A higher response timeliness index value indicates an earlier estimated task completion time for the subtask. For example, based on the estimated resource information of the candidate intelligent agent node, the internal and external resource requirements of the node for the subtask can be obtained, and a corresponding resource cost index can be derived. A higher resource cost index value indicates lower internal and external resource requirements for the subtask.

[0085] Then, based on preset weights, the node evaluation index, response time index, and resource cost index are weighted and fused to obtain the matching degree value between the candidate agent node and the subtask, and the matching degree information between the candidate agent node and the subtask is obtained based on this matching degree value. A higher matching degree value between the candidate agent node and the subtask indicates that the agent corresponding to the candidate agent node is more suitable to undertake the subtask. For example, the weights corresponding to the node evaluation index, response time index, and resource cost index can be 40%, 30%, and 30%, respectively.

[0086] Step S303: Determine the execution agent node corresponding to the subtask based on the matching degree information of each candidate agent node.

[0087] Based on the matching degree information of each candidate agent node, the candidate agent node with the highest matching degree to the subtask can be selected as the execution agent node corresponding to the subtask.

[0088] Step S304: Obtain the task allocation result of the subtask based on the executing agent node.

[0089] Specifically, based on the executing agent node corresponding to each subtask, the task allocation result for each subtask can be obtained. For example, the task allocation result for a subtask may include the unique identifier of the subtask and the node identifier of the executing agent node corresponding to that subtask.

[0090] In this embodiment, the matching degree information between each candidate intelligent agent node and the sub-task is obtained by considering the node evaluation information, estimated timeliness information and estimated resource information contained in the task response scheme submitted by each candidate intelligent agent node. Then, the corresponding executing intelligent agent node of the sub-task is determined based on the matching degree information. This can break through the traditional single-factor task allocation mode and allocate sub-tasks by considering multi-dimensional factors such as reputation, timeliness and cost, thereby achieving a balance between efficiency and fairness.

[0091] In an exemplary embodiment, after recording and publishing the task allocation result to the blockchain, the process may include: if a subtask reassignment request is detected, updating the subtask's task allocation result according to the subtask's response scheme; wherein the reassignment request is submitted to the blockchain by the execution agent node corresponding to the subtask; and publishing the updated task allocation result to the blockchain.

[0092] Within this system, the executing agent node corresponding to a subtask can obtain data generated by its corresponding agent during the execution of the subtask. When the executing agent node obtains data regarding abnormal execution of a subtask, it can submit a reallocation request for that subtask to the blockchain. Abnormal execution conditions of a subtask can include, but are not limited to, execution delays or insufficient resources.

[0093] In this blockchain, nodes can listen to the blockchain and, upon detecting a subtask reallocation request, trigger the reallocation of that subtask. Specifically, during subtask reallocation, a new executing agent node is selected from the candidate agent nodes corresponding to the subtask based on the corresponding task response schemes, thus updating the subtask's allocation result. The updated allocation result is then published to the blockchain. The new executing agent node, upon receiving the updated allocation result, can then execute the subtask through its corresponding agent.

[0094] In this embodiment, by reallocating the subtask when a subtask reallocation request is detected, it is possible to react and handle abnormal execution of the subtask in a timely manner, which is conducive to ensuring the efficient and reliable execution of the task as a whole.

[0095] In an exemplary embodiment, the method may further include: obtaining task execution data of each agent node from the blockchain; obtaining node evaluation information of the agent node based on the task execution data; and publishing the node evaluation information to the blockchain.

[0096] In this embodiment, the node evaluation information corresponding to each intelligent agent node can be obtained based on the task execution data of each intelligent agent node.

[0097] Specifically, the blockchain can store task execution data for each intelligent agent node. This data can include data submitted by the node during historical sub-task execution, as well as data obtained through verification or statistical analysis of the submitted data. For example, during sub-task execution, the executing intelligent agent node submits its task execution progress, intermediate results, and error log information to the blockchain. Nodes on the blockchain can verify the data submitted by the executing intelligent agent node to obtain corresponding task data verification records, and can analyze the submitted data to obtain task completion records, execution timeliness information, etc., for that executing intelligent agent node, and then submit this information to the blockchain.

[0098] This can be achieved by retrieving task execution data from each intelligent agent node on the blockchain and analyzing and statistically processing it according to preset evaluation rules to obtain node evaluation information for each intelligent agent node. Subsequently, the node evaluation information for each intelligent agent node can be published to the blockchain. For example, task execution data for each intelligent agent node can be retrieved according to a preset period, and its respective node evaluation information can be updated accordingly. This preset period can be a preset time interval or the task execution cycle.

[0099] In this embodiment, task execution data obtained from the blockchain is used to obtain node evaluation information for each intelligent agent node. Since all data originates from immutable records on the chain and is evaluated through unified rules, an objective and reliable evaluation of the intelligent agent node can be achieved. The obtained node evaluation information can accurately reflect the task execution performance of the intelligent agent node.

[0100] In one exemplary embodiment, task execution data may include task completion records of the agent node, task data verification records, error log information, execution timeliness information, and task collaboration records. For example... Figure 4 As shown, based on task execution data, the node evaluation information of the agent nodes can be obtained, which may include:

[0101] Step S401: Based on the task completion record, task data verification record, and error log information, obtain the execution quality index of the agent node.

[0102] Specifically, in this step, the task completion rate, data accuracy, and error rate of the agent node can be calculated based on the task completion record, task data verification record, and error log information of the agent node, and then the execution quality index of the agent node can be calculated by weighted summation.

[0103] For example, the task completion record of an agent node may include the number of completed subtasks and the total number of subtasks assigned to the agent node. The task completion rate of the agent node can be expressed as: Task completion rate = (number of completed subtasks / total number of subtasks) × 100%.

[0104] For example, the task data verification record of an agent node may include the total number of data submission entries submitted by the agent node to the blockchain, and the number of correct data entries that have passed verification. The data accuracy of the agent node can be expressed as: Data accuracy = (number of correct data entries / total number of submission entries) × 100%.

[0105] For example, the error log information of the agent node may include the total number of operations performed by the agent corresponding to the agent node during the execution of the subtask and the number of errors in the task execution. The error rate of the agent node can be expressed as: Error rate = (Number of errors in task execution / Total number of operations) × 100%.

[0106] For example, the execution quality index of an agent node can be calculated by weighted summation based on its task completion rate, data accuracy, and error rate. For instance, the execution quality index can be expressed as: Execution Quality Index = (Task Completion Rate × 30%) + (Data Accuracy × 40%) + (Error Rate × 30%).

[0107] Step S402: Based on the execution timeliness information, obtain the response timeliness index of the intelligent agent node.

[0108] The execution timeliness information of an agent node can include the estimated and actual completion times of each subtask previously executed by the agent node. The estimated completion time is the time the agent node estimated would be needed to complete the subtask when submitting its response plan, while the actual completion time is the time the agent node actually spent completing the subtask. Based on the agent node's execution timeliness information, its corresponding response timeliness metric can be calculated.

[0109] For example, the time deviation rate can be calculated for each subtask executed by the agent node, and then the response timeliness index of the agent node can be obtained based on the time deviation rate of each subtask. For example, the time deviation rate of the agent node executing a subtask can be expressed as: Time Deviation Rate = (Estimated Completion Time - Actual Completion Time) / Estimated Completion Time × 100%. Then, the response timeliness index of the agent node can be obtained based on the time deviation rate of each subtask, for example, by averaging or selecting the median.

[0110] Step S403: Obtain the collaboration contribution index of the agent node based on the task collaboration record.

[0111] The task collaboration record of the intelligent agent node can include its on-chain data sharing record and assistance request log. The on-chain data sharing record can include the number of times the intelligent agent node submits key shared data to the blockchain during the execution of sub-tasks. This key shared data can be non-mandatory shared data actively submitted by the intelligent agent node and applicable to the sub-task execution process by other intelligent agent nodes. The assistance request log can include the number of times the intelligent agent node provides assistance to the corresponding intelligent agent node after receiving assistance requests from other intelligent agent nodes.

[0112] For example, the data sharing frequency of an agent node can be calculated based on its on-chain data sharing records, and can be expressed as: Data sharing frequency = (number of times data is shared / total number of subtasks) × 100%.

[0113] The collaboration contribution index of an agent node can be calculated by weighted summation based on the data sharing frequency and the number of times it assists another agent node. For example, the collaboration contribution index can be expressed as: Collaboration contribution index = (data sharing frequency × 60%) + (number of times it assists another agent node × 40%).

[0114] Step S404: According to the preset node evaluation weights, the execution quality index, response time index, and collaboration contribution index are weighted and fused to obtain the node evaluation information of the agent node.

[0115] In this step, the execution quality indicators, response time indicators, and collaboration contribution indicators of each agent node can be weighted and fused according to the preset node evaluation weights to obtain the node score of each agent node, and then the node evaluation information of the agent node can be obtained based on the node score.

[0116] For example, the node evaluation weight corresponding to the execution quality indicator can be 50%, the node evaluation weight corresponding to the response time indicator can be 30%, and the node evaluation weight corresponding to the collaboration contribution indicator can be 20%. The node score of the agent node can be expressed as: Node score = (execution quality indicator × 50%) + (response time indicator × 30%) + (collaboration contribution indicator × 20%).

[0117] In this embodiment, by calculating the execution quality index, response time index, and collaboration contribution index of the agent node respectively, and then obtaining the corresponding node evaluation information through weighted fusion, a multi-dimensional objective evaluation of the agent node can be achieved. The obtained node evaluation information can provide more reliable decision support for subsequent task allocation, which is conducive to improving the fairness of sub-task allocation.

[0118] In an exemplary embodiment, the method may further include: if the node evaluation information of the agent node meets the preset incentive conditions, then executing node incentive measures for the agent node; if the node evaluation information of the agent node meets the preset penalty conditions, then executing node penalty measures for the agent node.

[0119] After obtaining the node evaluation information of each agent node, the node incentive measures or node penalty measures can be executed accordingly based on whether the node meets the preset incentive conditions or penalty conditions.

[0120] For example, the agent nodes can first be sorted according to their node evaluation information, from highest to lowest score. The top-ranked agent nodes (either a first number or a first proportion) can be considered as meeting the incentive conditions, allowing for the implementation of node incentive measures. Conversely, the bottom-ranked agent nodes (either a second number or a second proportion) can be considered as meeting the penalty conditions, allowing for the implementation of node penalty measures.

[0121] For example, node incentive measures performed on an agent node may include, but are not limited to, allocating more shared resources to the agent node, or setting a higher priority for subtasks assigned to the agent node. For example, node penalty measures performed on an agent node may include, but are not limited to, reducing the shared resources allocated to the agent node, or setting a lower priority for subtasks assigned to the agent node.

[0122] In this embodiment, by implementing node incentive measures or node penalty measures on intelligent agent nodes whose node evaluation information meets preset incentive or penalty conditions, it is possible to build a self-driven collaborative ecosystem on the chain. This is beneficial to enhance the integrity awareness, task execution quality, and collaborative enthusiasm of intelligent agent nodes, thereby improving the overall execution effect of tasks and effectively improving trust and security issues among intelligent agent nodes.

[0123] In one exemplary embodiment, such as Figure 5 As shown, a multi-agent task coordination method is provided.

[0124] Specifically, the method in this embodiment can be applied to, for example... Figure 6The illustrated distributed intelligent agent collaboration system includes a consortium blockchain network built on blockchain technology. This network can include multiple authorized nodes, such as verification nodes, consensus nodes, and multiple intelligent agent nodes. The consensus nodes are responsible for executing the consensus algorithm to ensure transaction consistency and block generation; the verification nodes verify transactions submitted by the intelligent agent nodes to guarantee data legitimacy; and the intelligent agent nodes can integrate blockchain node modules, perception modules, decision-making modules, and execution modules. Multiple intelligent agent nodes can collaborate within the distributed intelligent agent environment to complete tasks.

[0125] The consortium blockchain network can deploy smart contracts, which may include code logic for executing the multi-agent task coordination method in this embodiment. Nodes in the consortium blockchain network can, upon receiving a transaction interacting with the smart contract, invoke the corresponding content of the smart contract and execute corresponding operations according to predetermined rules and logic. During system initialization, various smart contracts, including task management contracts, allocation contracts, trust assessment contracts, and incentive contracts, can be deployed and initialized in the consortium blockchain network. Each agent node completes registration and identity authentication, uploading its basic information, resource status, and capability parameters to the consortium blockchain as the basis for subsequent task allocation and collaboration. The smart contracts deployed on the consortium blockchain can be continuously updated based on data from the consortium blockchain to adjust parameters such as task allocation rules, node scoring rules, incentive strategies, and penalty strategies to optimize system performance.

[0126] Among them, such as Figure 5 As shown, the multi-agent task coordination method in this embodiment may include the following steps:

[0127] Step S501, Task Publication. When a smart agent node has a task requirement, it can encapsulate the task information into a transaction and submit it to the consortium blockchain. Nodes in the consortium blockchain can obtain the task information, call a smart contract to parse the information, break it down into multiple subtasks, and obtain the subtask information for each subtask. Then, the subtask information can be written to the consortium blockchain through a consensus mechanism.

[0128] Step S502: Task response plan submission. Each agent node in the consortium blockchain can listen to the consortium blockchain to obtain sub-task information and its own node evaluation information. Then, based on the sub-task information, node evaluation information, and the agent node's resource status information, when it determines that it meets the task requirements of the sub-task, it submits a task response plan for the sub-task to the consortium blockchain. For example, the task response plan may include the agent node's node evaluation information, as well as the agent node's estimated timeliness and estimated resource information for the sub-task.

[0129] Step S503, Intelligent Agent Node Screening. Nodes in the consortium blockchain, after obtaining the task response plan for each subtask, can invoke a smart contract to screen candidate intelligent agent nodes for each subtask. Specifically, based on the node evaluation information, estimated timeliness information, and estimated resource information of each candidate intelligent agent node for the subtask, the matching degree information between each candidate intelligent agent node and the subtask can be obtained. Then, based on the matching degree information of each candidate intelligent agent node, the executing intelligent agent node corresponding to the subtask is determined.

[0130] Step S504, Task Allocation. After determining the executing agent node corresponding to each subtask, the task allocation result for each subtask can be obtained. This task allocation result can then be published to the consortium blockchain.

[0131] Step S505, execution and data synchronization. During this step, the execution agent nodes corresponding to each subtask can begin executing the corresponding subtask after obtaining the task allocation results. During the execution of the subtask, the execution agent nodes can record information such as task execution progress and intermediate results on the consortium blockchain. In case of abnormal situations, a reassignment request can be submitted to the consortium blockchain to promptly initiate a new task allocation process.

[0132] Step S506: Node evaluation information update. Nodes in the consortium blockchain can obtain task execution data from each agent node and periodically call smart contracts to obtain node evaluation information for each agent node based on this data, which is then recorded on the consortium blockchain. This node evaluation information can be derived from factors such as the agent node's execution quality, response timeliness, and collaborative contribution.

[0133] Step S507, Incentive or Penalty. Nodes in the consortium blockchain can invoke smart contracts to implement node incentive measures for smart agent nodes whose node evaluation information meets preset incentive conditions, and to implement node penalty measures for smart agent nodes whose node evaluation information meets preset penalty conditions. For example, node incentive measures may involve allocating a certain number of on-chain resource points to smart agent nodes, and node penalty measures may involve deducting a certain number of on-chain resource points from smart agent nodes. These points can be used to access shared resources or to set higher priority for subtask allocation.

[0134] Step S508, Result Feedback. The node evaluation information of each agent node can be fed back to the next task cycle.

[0135] In this embodiment, a distributed intelligent agent collaboration system is constructed based on a blockchain network. This system enables task allocation and collaboration mechanisms, including task on-chaining and decomposition, response plan submission and allocation, and collaboration and execution. A trust and incentive mechanism is also established, using node evaluation information and incentive / penalty mechanisms to ensure trust and collaborative enthusiasm among intelligent agent nodes. Furthermore, the smart contracts are deployed in a modular manner and possess adaptive adjustment capabilities, allowing for optimization of system parameters based on historical data. Therefore, by leveraging the immutability and consensus mechanism of the consortium blockchain, this embodiment ensures the security and reliability of data interaction between intelligent agent nodes, effectively preventing malicious data tampering and forgery, reducing collaboration risks, and improving data security compared to traditional systems. Simultaneously, the task allocation and collaboration mechanism based on the consortium blockchain enables fair and transparent task allocation, improves task execution efficiency, reduces collaboration friction caused by trust issues, and effectively shortens task completion time. Moreover, by establishing a node evaluation and incentive mechanism, the collaborative enthusiasm of intelligent agent nodes is stimulated, the level of trust among intelligent agent nodes is improved, the healthy operation of the system is promoted, and the participation of intelligent agent nodes and the task success rate are increased. The adaptive adjustment capability of smart contracts enables the system to quickly adapt to different task scenarios and environmental changes, enhancing the system's flexibility and scalability, and reducing the cost of manual intervention.

[0136] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0137] Based on the same inventive concept, this application also provides a multi-agent task coordination device for implementing the multi-agent task coordination method described above. The solution provided by this device is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more embodiments of the multi-agent task coordination device provided below can be found in the limitations of the multi-agent task coordination method described above, and will not be repeated here.

[0138] In one exemplary embodiment, such as Figure 7 As shown, a multi-agent task coordination device is provided, comprising:

[0139] The first acquisition module 701 is used to acquire task information of the task to be processed; the task information is submitted to the blockchain by the smart agent node.

[0140] The task parsing module 702 is used to parse the task information, decompose the task to be processed into multiple sub-tasks, and obtain sub-task information corresponding to each sub-task.

[0141] The first publishing module 703 is used to publish the information of each subtask to the blockchain;

[0142] The second acquisition module 704 is used to acquire the task response scheme corresponding to each of the sub-tasks; the task response scheme is submitted to the blockchain by the candidate intelligent agent node corresponding to the sub-task; the task response scheme is obtained by the candidate intelligent agent node based on the sub-task information and node evaluation information acquired from the blockchain and the resource status information of the candidate intelligent agent node;

[0143] The task allocation module 705 is used to obtain the task allocation result of each subtask according to the task response scheme of each candidate agent node corresponding to the subtask; the task allocation result is used to indicate the execution agent node corresponding to each subtask.

[0144] The second publishing module 706 is used to publish the task allocation results to the blockchain.

[0145] In an exemplary embodiment, the task allocation module 705 is configured to: obtain, according to the task response schemes corresponding to the sub-tasks, the node evaluation information of each candidate agent node corresponding to the sub-task, as well as the estimated timeliness information and estimated resource information of the candidate agent nodes for the sub-tasks; obtain the matching degree information between the candidate agent nodes and the sub-tasks based on the node evaluation information, the estimated timeliness information, and the estimated resource information; determine the execution agent node corresponding to the sub-task based on the matching degree information of each candidate agent node; and obtain the task allocation result of the sub-task based on the execution agent node.

[0146] In an exemplary embodiment, the apparatus further includes: an allocation update module, configured to update the task allocation result of the subtask according to each of the task response schemes of the subtask if a reallocation request information of the subtask is detected; wherein the reallocation request information is submitted to the blockchain by the execution agent node corresponding to the subtask; and a third publishing module, configured to publish the updated task allocation result to the blockchain.

[0147] In an exemplary embodiment, the device further includes: a third acquisition module, configured to acquire task execution data of each of the intelligent agent nodes from the blockchain; an evaluation acquisition module, configured to obtain node evaluation information of the intelligent agent nodes based on the task execution data; and a fourth publishing module, configured to publish the node evaluation information to the blockchain.

[0148] In an exemplary embodiment, the task execution data includes the agent node's task completion record, task data verification record, error log information, execution timeliness information, and task collaboration record. The evaluation acquisition module is used to: obtain the agent node's execution quality index based on the task completion record, the task data verification record, and the error log information; obtain the agent node's response timeliness index based on the execution timeliness information; obtain the agent node's collaboration contribution index based on the task collaboration record; and weight and fuse the execution quality index, response timeliness index, and collaboration contribution index according to a preset node evaluation weight to obtain the agent node's node evaluation information.

[0149] In an exemplary embodiment, the device further includes: a node incentive module, configured to execute node incentive measures for the agent node if the node evaluation information of the agent node meets preset incentive conditions; and a node penalty module, configured to execute node penalty measures for the agent node if the node evaluation information of the agent node meets preset penalty conditions.

[0150] Each module in the aforementioned multi-agent task coordination device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can invoke and execute the operations corresponding to each module.

[0151] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 8 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores task information for tasks to be processed, sub-task information for subtasks, corresponding task response schemes for subtasks, and task allocation results for subtasks. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When executed by the processor, the computer program implements a multi-agent task coordination method.

[0152] Those skilled in the art will understand that Figure 8 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0153] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0154] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0155] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0156] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0157] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0158] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0159] The above embodiments are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A multi-agent task coordination method, characterized in that, The method includes: Obtain task information for tasks to be processed; the task information is submitted to the blockchain by the intelligent agent node. The task information is parsed, and the task to be processed is decomposed into multiple sub-tasks to obtain sub-task information corresponding to each sub-task. Publish the information of each subtask to the blockchain; Obtain the task response scheme corresponding to each of the sub-tasks; the task response scheme is submitted to the blockchain by the candidate intelligent agent node corresponding to the sub-task; the task response scheme is obtained by the candidate intelligent agent node based on the sub-task information and node evaluation information obtained from the blockchain and the resource status information of the candidate intelligent agent node; Based on the task response scheme of each candidate agent node corresponding to the subtask, the task allocation result of each subtask is obtained; the task allocation result is used to indicate the execution agent node corresponding to each subtask. The task allocation results are published to the blockchain.

2. The method according to claim 1, characterized in that, The step of obtaining the task allocation result for each subtask based on the corresponding task response scheme includes: Based on the task response schemes corresponding to the sub-tasks, obtain the node evaluation information of each candidate agent node corresponding to the sub-task, as well as the estimated timeliness information and estimated resource information of the candidate agent nodes for the sub-tasks. Based on the node evaluation information, the estimated timeliness information, and the estimated resource information, the matching degree information between the candidate intelligent agent node and the subtask is obtained; Based on the matching degree information of each candidate agent node, the executing agent node corresponding to the subtask is determined; Based on the executing agent node, the task allocation result of the subtask is obtained.

3. The method according to claim 1, characterized in that, After the task allocation results are recorded and published to the blockchain, the following steps are included: If a reallocation request for the subtask is detected, the task allocation result of the subtask is updated according to the task response scheme of each subtask; wherein the reallocation request is submitted to the blockchain by the execution agent node corresponding to the subtask. The updated task assignment results are published to the blockchain.

4. The method according to any one of claims 1 to 3, characterized in that, The method further includes: Obtain task execution data from each of the intelligent agent nodes from the blockchain; Based on the task execution data, the node evaluation information of the agent node is obtained; The node evaluation information is published to the blockchain.

5. The method according to claim 4, characterized in that, The task execution data includes the task completion record, task data verification record, error log information, execution time information, and task collaboration record of the agent node; The step of obtaining the node evaluation information of the agent node based on the task execution data includes: The execution quality index of the intelligent agent node is obtained based on the task completion record, the task data verification record, and the error log information. Based on the execution timeliness information, the response timeliness index of the intelligent agent node is obtained; Based on the task collaboration records, the collaboration contribution index of the agent node is obtained; According to the preset node evaluation weights, the execution quality index, response time index, and collaboration contribution index are weighted and fused to obtain the node evaluation information of the agent node.

6. The method according to claim 4, characterized in that, The method further includes: If the node evaluation information of the agent node meets the preset incentive conditions, then the node incentive measures for the agent node are executed. If the node evaluation information of the agent node meets the preset penalty conditions, then the node penalty measures for the agent node are executed.

7. A multi-agent task coordination device, characterized in that, The device includes: The first acquisition module is used to acquire task information of the task to be processed; the task information is submitted to the blockchain by the intelligent agent node. The task parsing module is used to parse the task information, decompose the task to be processed into multiple sub-tasks, and obtain sub-task information corresponding to each sub-task. The first publishing module is used to publish the information of each subtask to the blockchain; The second acquisition module is used to acquire the task response scheme corresponding to each of the sub-tasks; the task response scheme is submitted to the blockchain by the candidate intelligent agent node corresponding to the sub-task; the task response scheme is obtained by the candidate intelligent agent node based on the sub-task information and node evaluation information acquired from the blockchain and the resource status information of the candidate intelligent agent node; The task allocation module is used to obtain the task allocation result of each subtask according to the task response scheme of each candidate agent node corresponding to the subtask; the task allocation result is used to indicate the execution agent node corresponding to each subtask. The second publishing module is used to publish the task allocation results to the blockchain.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images