437 results about "Multi-agent system" patented technology

The invention belongs to the field of natural language processing, and provides a task complexity driven graph semantic multi-agent collaborative decision-making method and system.The task complexity driven graph semantic multi-agent collaborative decision-making method comprises the steps that a task text is obtained and subjected to semantic coding to obtain a task semantic vector, evaluation is conducted based on the task semantic vector to obtain a complexity vector, and a task complexity score of the complexity vector is calculated; the task semantic vector and the complexity vector are fused to obtain a task representation vector, an agent capability relation graph is constructed, the participation probability of each agent node is obtained according to the task representation vector and the agent capability relation graph, and a dynamic agent combination scheme is formed; and performing task decomposition according to the agent combination scheme, constructing a sub-task dependency graph, scheduling the execution sequence of the sub-tasks through topological sorting, realizing cooperative execution of the agents, and generating a task result. According to the method, precise matching and efficient cooperation of the agent combination are realized, and the capability of processing complex tasks and the resource utilization efficiency of the multi-agent system are remarkably improved.

The invention discloses a hierarchical multi-agent robot cluster control method and system, and aims to solve the problems of partial observability and environment non-stability of a multi-agent system in a complex environment. According to the method, a three-layer layered reinforcement learning architecture is constructed, a high-layer strategy is responsible for global task decomposition and role allocation, a middle-layer strategy converts tactical intention into a cooperative behavior mode, and a low-layer strategy executes accurate motion control; a graph neural network is adopted for cluster modeling, global graph representation and local neighborhood features are extracted in parallel through graph convolution and an attention mechanism, and hierarchical decision making is supported; a centralized graph enhancement evaluation network is designed to be combined with an MAPPO algorithm for collaborative optimization, and dynamic adversarial training is introduced to improve strategy robustness. According to the method, effective decoupling of global planning and local control is realized, and the cluster cooperation efficiency, the strategy interpretability and the adaptive capacity in a dynamic environment are improved.

The invention discloses a data-based predefined time heterogeneous multi-agent formation collision avoidance method. The method comprises the following steps: establishing a nonlinear heterogeneous multi-agent system; designing a bimodal artificial potential field function to perform dynamic obstacle avoidance and prevent regional escape; establishing a self-adaptive robust controller used for generating an obstacle avoidance safety motion trail of the root leader; designing a self-adaptive formation zooming mechanism of the leader, and constructing a predefined time affine observer of the follower based on the self-adaptive formation zooming mechanism; designing a unified obstacle function; constructing a virtual control law for processing tracking errors based on the unified obstacle function; designing a neural network estimator; designing controllers of the leader and the follower according to the virtual control law; and forming a collision avoidance decision of the heterogeneous multi-agent formation based on a self-adaptive robust controller, a predefined time affine observer, a neural network estimator and controllers of the leader and the follower. According to the method, safe, efficient and robust cooperative control of the formation in a complex environment is realized, and the safety and task execution efficiency of the heterogeneous multi-agent formation in a limited and unknown environment are remarkably improved.

The invention discloses a visual algorithm self-training method based on multi-agent collaborative optimization, and the method comprises the following steps: constructing a multi-agent system architecture which comprises a user interaction layer, an intelligent scheduling layer, an A2A protocol communication layer and a professional agent cluster layer; the user interaction layer analyzes a user task intention and generates an execution plan; the scheduling agent calls the professional agent to complete data processing, model construction, training, testing and deployment; a task process is coordinated through a standardized communication mechanism, and task execution is supported by combining an MCP tool set, a knowledge base module and a memory system; and when the task fails, automatically executing rescheduling operation, and finally outputting a self-training result. According to the method, the development efficiency, the self-adaptability and the intelligent level are remarkably improved, and the method is suitable for computer vision tasks such as industrial detection, intelligent security and protection and automatic driving.

The invention provides a distributed formation control method driven by a recursive balance network under a communication attack, and relates to the technical field of cooperative control of a multi-agent system, and the method comprises the steps: constructing a distributed multi-agent system model, and constructing an RBN controller heterogeneous communication network architecture and a DoS attack model; the RBN controller design is optimized, the distributed RBN controller architecture is realized, the stability of the RBN controller is analyzed based on the shrinkage mapping theory, and the exponential convergence of the system state difference is proved by constructing a Lyapunov function; training the optimized RBN controller by adopting a difficulty priority confrontation training strategy, designing a multi-objective loss function, and dynamically adjusting the training weight of each difficulty level; and simulating the distributed multi-agent system optimized by the above steps, testing the robustness of the distributed multi-agent system under incremental DoS attack intensity, and performing comparative analysis of heterogeneous and isomorphic communication network configuration and comprehensive performance comparison of an RBN controller and a traditional MPC method.

Provided are a multi-agent construction method and apparatus, a task processing method and apparatus, a computing device, a storage medium, a computer program product, and a chip. The multi-agent construction method comprises: splitting a task into a plurality of sub-tasks; determining, from an agent pool, agents respectively matching the plurality of sub-tasks, wherein the agent pool comprises a plurality of agents, and the agents matching the sub-tasks can process the sub-tasks by calling at least one of tools and other agents which are configured by the agents; and constructing a multi-agent system on the basis of the agents respectively matching the plurality of sub-tasks, wherein the multi-agent system is used for processing the task.

The invention relates to the technical field of medical information, in particular to a method and a system for processing medical record documents by utilizing artificial intelligence, and particularly relates to a method and a system which can receive and process multi-modal input information including images, videos and voices, can work cooperatively through a multi-agent system and can process the medical record documents. The invention discloses a system for automatically generating, controlling quality and safely inputting medical records in combination with a medical knowledge base and an implementation method thereof. The invention relates to an automatic medical record writing system based on multi-modal input and multi-agent driving. The automatic medical record writing system comprises a multi-modal input module, a multi-agent processing platform and a safety intranet input module. According to the method, establishment of the data channel between the medical intranet and the external AI system is proposed for the first time, the problem of medical intranet isolation is solved, industrial pain points are solved in a breakthrough mode, and a new medical AI landing path is developed.

The invention relates to the technical field of intelligent unmanned systems, in particular to a task allocation method of a sea area cross-medium unmanned system based on a graph attention network and deep reinforcement learning, and the method comprises the following steps: S1, constructing a dynamic graph and calculating task priority; s2, screening and selecting schedulable tasks; s3, intelligent agent distribution and state updating; s4, dynamic reordering and cycle control are carried out; according to the method, a space-time coupling dynamic graph structure is constructed, task attributes and agent states are coded into node features, a four-layer graph attention network is designed to extract task priority distribution, and a self-adaptive scheduling decision is realized in combination with a near-end strategy optimization algorithm; a task dependence verification and resource availability check mechanism is established, and approximate optimal agent allocation is realized; a multi-dimensional reward function is included, training stability is guaranteed in combination with an experience playback mechanism and a gradient clipping technology, and task allocation efficiency of a multi-agent system in a complex environment is remarkably improved.

The invention belongs to the technical field of smart power grids, and discloses a distributed photovoltaic cluster-based smart power grid optimal scheduling method and system, and the method comprises the steps: fusing meteorological and power grid data through a cross-modal sensing network, combining a meta-learning framework to quickly adapt to a new power station, and narrowing a fluctuation interval through a prediction-correction dual-channel mechanism; according to the method, implicit association between weather and a power grid is mined, prediction robustness of low-probability events is improved through extreme scene intensive training, output prediction precision is remarkably improved, and the problem that traditional prediction lags behind actual fluctuation is effectively relieved; a dynamic role multi-agent system is adopted, and a credit scoring mechanism of block chain evidence storage and an improved auction algorithm are combined, so that distributed efficient decision making is realized; the intelligent agent dynamically switches roles according to the load state, the credit condition is linked with the scheduling priority, the problem that the topology adaptability of centralized decision making is poor is solved, and a trusted collaborative environment is constructed through transaction records which cannot be tampered.

The invention relates to the technical field of multi-agent cooperative control, and discloses a multi-agent autonomous decision-making method based on deep reinforcement learning. The method comprises the steps of synchronously detecting an initial collaborative state of a cluster, performing joint situation assessment, and judging a collaborative operation mode according to a quantitative situation. And analyzing the capability of each agent and the real-time task load, and constructing a distributed task knowledge graph. And utilizing the atlas to drive a deep reinforcement learning network, coupling computing resource allocation and a task path, and generating a preliminary behavior strategy of each agent. And performing cluster-level conflict detection and iterative negotiation adjustment, and finally issuing an executable action instruction sequence. According to the method, integrated optimization of resource allocation and action paths is realized, the situation understanding and negotiation mechanism is enhanced through the knowledge graph to guarantee the collaborative consistency, and the collaborative decision-making efficiency and task execution robustness of a multi-agent system in a dynamic environment are improved.

The invention provides a multi-agent-based task cooperative execution method and device, and the method comprises the steps: carrying out the calculation based on the task type of a target task and the current system state of a multi-agent system, and obtaining an index weight; determining a task allocation scheme of each subtask based on the index weight and the agent bidding information of each subtask; and determining an execution strategy of each sub-task based on the task attribute of each sub-task, the task allocation scheme and the environment perception data of the multi-agent system. According to the method and device provided by the invention, the index weight is calculated according to the task type of the target task and / or the current system state of the multi-agent system; the task allocation scheme of each subtask is determined based on the index weight and the bidding information of each subtask corresponding to the plurality of agents, so that dynamic task allocation adapting to environment change is realized, the task allocation rationality is improved, and the task execution efficiency and task completion quality of the system are improved.

The invention discloses a multi-agent-based context knowledge internalization, sharing and multiplexing system and method. The system comprises a multi-source knowledge capture module, a situation information extraction and structuring module, a knowledge internalization and incremental learning module, a situation awareness knowledge reuse and recommendation module and a knowledge quality evaluation and maintenance module, wherein each module is composed of a special intelligent agent or a functional unit and works cooperatively; according to the method, interaction, cooperation and process data and expert input of a task execution full cycle of a multi-agent system are captured through a multi-source module, preprocessing is carried out through a situation extraction module, a situation is associated to generate structured knowledge, and then the structured knowledge is fused into a dynamic knowledge system through an internalization module to achieve incremental learning. In a new task, matching knowledge is retrieved and recommended through the context awareness module, and finally the knowledge quality is periodically maintained by the evaluation module to ensure accuracy and timeliness; according to the method, the autonomous learning ability, the knowledge reuse efficiency and the intelligent level of solving complex problems of the multi-agent system are remarkably improved.

The invention discloses a heterogeneous agent distributed multi-alliance game control method, device and equipment and a medium, and relates to the field of multi-agent system cooperative control and game theory cross application, and the method comprises the steps: dividing agents in a heterogeneous unmanned cluster system into a plurality of alliances; establishing a high-order heterogeneous linear state dynamical model for the intelligent agent in each alliance; constructing an internal and external double-layer alliance communication topological graph according to an actual communication condition; and on the basis of the model and the topological graph, through a distributed Nash equilibrium search algorithm containing a search layer and an output adjustment layer, performing iterative optimization on an agent decision state to obtain a multi-alliance game strategy, and controlling the agent according to the strategy. According to the method, the cooperative control problem of the heterogeneous agents in a multi-alliance game scene can be effectively solved, the Nash equilibrium point is quickly searched, the decision state is adjusted to be optimal, and the cooperative control performance and task execution efficiency of a multi-agent system are improved.

The embodiment of the invention provides a context information management method and device based on hierarchical memory, equipment and a medium, and the method comprises the steps: analyzing a task request to generate a sub-task node sequence containing a target description and acceptance function, and arranging a plurality of agents to cooperatively execute a task, and after the audit is passed, a complete execution context is stored in a long-term memory, and a structured abstract is generated and stored in a short-term memory, so that the problem that the context management mechanism of the existing multi-agent system is single is effectively solved, stable maintenance of cross-session cognitive continuity is realized, and the efficiency of the multi-agent system is improved. The method improves the execution reliability of a complex long-process task, optimizes the information storage and retrieval efficiency through hierarchical memory, reduces the risk of semantic deviation accumulation and amplification, clarifies the task target and acceptance standard of each stage, enhances the multi-agent cooperation consistency, and improves the efficiency of task execution. And the execution efficiency and the task completion quality of the intelligent agent system in a complex application scene are comprehensively improved.

The invention provides a multi-agent proposition method and system based on collaborative reinforcement learning. The method comprises the following steps: constructing a multi-agent proposition framework comprising a test question generation agent and a quality inspection agent based on a large language model; constructing training data on the basis of a target subject test question set, respectively supervising and adjusting large language models on which a test question generation agent and a quality inspection agent are based, and training a reward model for evaluating the output quality of each agent for the agent; and performing multi-agent collaborative reinforcement learning training on the test question generation agent and the quality inspection agent by using the reward model to obtain a trained agent model weight. And loading the trained agent model weight into a multi-agent collaborative proposition framework, and receiving proposition requirements to generate final test questions and reference answers. A multi-agent cooperative reinforcement learning mechanism is introduced into a multi-agent system, so that the agents can form a stable and efficient cooperation strategy in a real proposition task interaction environment.

The invention provides a dialogue chain multi-dimensional semantic enhancement method based on an MCP agent negotiation and voting mechanism, and relates to the field of artificial intelligence, and the method comprises the steps: outputting a plurality of candidate dialogue chains based on a large language model; performing multi-dimensional semantic enhancement on the plurality of candidate dialogue chains to obtain a plurality of enhanced dialogue chains; outputting a corresponding dialogue chain score for each enhanced dialogue chain through each preset agent; screening the plurality of dialogue chain scores, and obtaining a target consistency score of which the score value is greater than a preset ranking in the plurality of dialogue chain scores; determining a target conversation chain corresponding to each target consistency score in the plurality of enhanced conversation chains; and calculating a ticket selection number corresponding to each target dialogue chain, and outputting the target dialogue chain corresponding to the highest ticket selection number as an optimal dialogue chain. The problems that an existing multi-agent system mostly drives agents to execute tasks through a single control center, an independent expression and consensus decision-making mechanism of individual agents is lacked, and multi-round dialogue efficiency is low are solved.

The invention discloses an agent contribution degree evaluation method and device of a multi-agent system, a storage medium, electronic equipment and a computer program product. The evaluation method comprises the steps that according to received environment state information, a graph neural network model corresponding to the multi-agent system is determined, and the environment state information at least comprises agent state information, task requirements and environment information; training an action network and a graph neural network model according to the environment state information; determining action information of nodes of the graph neural network model through the trained action network according to the environment state information; and determining the contribution degree of the intelligent agent through the trained graph neural network model according to the environment state information and the action information.

The invention relates to the technical field of computers, and discloses a fragmentation block chain federal learning method based on a large language model multi-agent, and the method comprises the steps: S1, initializing a client; s2, dynamic fragmentation scheduling and distribution; s3, generating and uploading local knowledge; s4, intelligent agent collaborative routing and knowledge acquisition; s5, knowledge fusion and model updating; and S6, repeatedly executing the steps S3 to S5 until the model converges or reaches a preset number of iterations. According to the invention, under a decentralized and fragmented federated learning architecture, the complex reasoning ability of a large language model and the autonomous cooperation mechanism of three multi-agent systems, namely a fragmented scheduling agent, a fragmented knowledge state agent and a global knowledge routing agent, are deeply fused; according to the mechanism, dynamic optimization of a bottom layer fragment structure and intelligent routing of high-value knowledge are achieved in an intelligent mode, and therefore the overall efficiency and model performance of a system under the condition of heterogeneous data and heterogeneous equipment are remarkably improved.

The invention discloses an electric vehicle charging station energy storage scheduling method based on a multi-agent system. The method comprises the following steps: obtaining and standardizing operation basic data of a plurality of new energy vehicle charging stations; setting five types of agents, defining observation variables and action space, and constructing a multi-agent system model; constructing a global collaborative scheduling network, and executing strategy evaluation and strategy generation; setting a constraint boundary, and constructing a linear programming scheduling model; constructing a training sample, and performing offline training and updating of a global collaborative scheduling network; solving a local optimal scheduling amount based on the real-time operation basic data, and analyzing to generate a control instruction; and issuing an energy storage control instruction and a computing power unit control instruction, acquiring a cooperative scheduling result to generate a final scheduling result set, and submitting the final scheduling result set to an upper-layer scheduling system. According to the method, a multi-agent collaborative scheduling system is constructed, strategy optimization and linear programming are fused, and efficient, stable and executable global collaborative scheduling of energy storage and computing power tasks of the charging station is achieved.

The invention relates to a workshop scheduling system and method based on combination of a graph neural network and multiple agents. The system mainly comprises three modules: a graph neural network module, a multi-agent system module and a scheduling decision and optimization module. The graph neural network module extracts features of a complex relationship between a task and a machine through a double attention network (DAN). And the multi-agent system module performs distributed decision making and scheduling optimization based on a reinforcement learning strategy by using the feature information. And the scheduling decision and optimization module dynamically adjusts and optimizes the scheduling decision according to the real-time production condition, so that the efficient operation of the production process and the optimal utilization of resources are ensured. The method aims at improving the scheduling efficiency and optimizing the scheduling effect, remarkably improving the production efficiency of a workshop and reducing the production cost.

The invention relates to the technical field of intelligent control, and discloses a multi-agent consistency detection system, which effectively improves the cooperative control capability of a multi-agent system in a dynamic environment. A real-time topology updating mechanism guarantees rapid convergence when a network structure changes, a prediction compensation module significantly reduces control delay caused by environmental interference, and a multi-dimensional health monitoring system improves the accuracy and efficiency of fault recovery. Functional decoupling is achieved through the modular design, system expansion and maintenance are facilitated, autonomous adjustment and optimization of a recovery strategy are achieved through a strategy optimization mechanism driven by reinforcement learning, the fault recovery time is shortened, and the success rate is increased. A secondary health verification mechanism ensures effectiveness of recovery operation, and secondary faults caused by continuous operation of the system in an incomplete recovery state are avoided.

The invention provides a multi-agent system and a multi-agent dialogue management method. The system comprises a main agent, a dialogue management module and a plurality of sub-agents, and the sub-agents correspond to processing of tasks in different business fields respectively; the main agent is used for calling the dialogue management module under the condition that a user request is received; the dialogue management module is used for acquiring historical information corresponding to the user request, analyzing the historical information and returning an analysis result to the main agent; the main agent is further used for determining corresponding request task data and a target sub-agent in the multiple sub-agents according to the analysis result and the user request, and sending the request task data to the target sub-agent; the sub-agent is used for processing based on the request task data, obtaining a task processing result and returning the task processing result to the main agent; and the main agent is also used for returning a corresponding natural language reply to the user according to the task processing result of the sub-agent.

The invention relates to the technical field of control and information, and particularly discloses a distributed optimization control method for a security constraint uncertain multi-agent system, which comprises the following steps of: constructing a topological graph according to a network structure of the multi-agent system, and determining an adjacent matrix of the topological graph; determining a state equation; determining a target function needing to be optimized and inequality constraints needing to be met; according to the state equation, the target function needing to be optimized and the inequality constraint needing to be met, based on a Lyapunov function control method, an optimal solution search condition is established, and based on a barrier function control method, a system safety maintenance condition is established; determining an input constraint condition according to the multi-agent system; calculating by using a quadratic programming method to obtain optimal control input; and controlling the multi-agent system according to the optimal control input. According to the method, the long-term reliability of the system can be remarkably improved, the communication and calculation overhead is remarkably reduced, and the calculation efficiency is remarkably improved.

The invention provides a multi-agent trajectory prediction method based on space-time causal interaction modeling, and belongs to the field of behavior prediction of a multi-agent system. The problem of low trajectory prediction precision caused by data deviation and false correlation in the existing method is solved. The method comprises the following steps: extracting two-dimensional position coordinates of a lane center line corresponding to driving of each vehicle; calculating the deviation degree between the vehicle position and the center line of the corresponding lane, modeling the spatial dependency relationship between the vehicles, outputting vehicle spatial interaction characteristics, carrying out time sequence modeling, and extracting vehicle time sequence characteristics; feature coding: outputting lane center line features; fusing the time sequence features and the lane center line features to generate vehicle fusion features; constructing a causal discovery network, and outputting a causal probability matrix; based on the vehicle fusion features and the causal probability matrix, space-time causal interaction features between the vehicles are extracted and input into a trajectory decoder, and multiple possible future trajectory sequences of each vehicle are generated. The method is used in the fields of artificial intelligence, robots and intelligent driving.

The invention relates to a motorcade consistency formation control method and device based on a dynamic event triggering mechanism, electronic equipment and a storage medium. The method comprises the steps of generating a communication topological graph of a multi-agent system composed of a preset number of unmanned vehicles according to the theoretical basis of a graph theory, and building a mathematical model of each unmanned vehicle; based on the mathematical model, designing a combined measurement function, a measurement error and a consistency controller of the multi-agent system; distributed event triggering conditions and self-triggering conditions are designed based on Lyapunov parameters; according to a dynamic event triggering condition and a self-triggering condition, a triggering time interval expression of the intelligent agent is obtained through calculation; by constructing a multi-agent system consistency controller and combining distributed event triggering conditions and self-triggering conditions, multi-vehicle formation consistency control is completed. According to the fleet consistency formation control method disclosed by the invention, accurate and efficient control of a fleet formed by multiple unmanned vehicles can be realized.

The invention provides a self-healing graph scheduling system and method for a multi-agent system. According to the method, in the multi-agent task flow graph, when any node fails or needs to be upgraded, bypass or hot replacement can be automatically completed within the time lower than a preset failure threshold value, and it is ensured that task topology is continuously acyclic, data are not lost, and services are not interrupted. According to the method, in a directed acyclic graph of a multi-agent task process, a main / standby node is configured for each edge, and the health state of the nodes is monitored in real time through dual-channel Gossip heartbeat; when the main node meets the failure condition, the flow is automatically redirected to the backup node at the millisecond level, the node state is recovered by using the XOR-delta snapshot, and the topology acyclic property is verified at the same time, so that the task is ensured to be continuous and traceable. Experimental results show that the average recovery time is reduced to 18 ms, the annual downtime is reduced by more than ten times, and the method is suitable for intelligent finance, industrial internet, automatic driving and other real-time scenes needing parallel agent collaboration and high availability.

The invention relates to the technical field of digitization systems, and particularly discloses an intelligent AI-driven archive digitization full-process processing system, which comprises a physical digitization module, an AI agent collaborative network and a digital resource management module which work cooperatively in sequence, and realizes automatic processing from an entity archive to a structured digital knowledge base; the physical digitization module is configured to perform high-fidelity scanning or shooting on the entity archive to generate an original digital image set; the AI agent collaborative network is a multi-agent system dispatched and managed by the central coordinator; according to the system, an end-to-end closed-loop processing flow from an entity file to a structured digital knowledge base is constructed through sequential collaboration of the physical digitization module, the AI agent collaboration network and the digital resource management module, connection of all links does not need to be manually intervened, manual review is only needed in a very few complex abnormal scenes, the labor cost is remarkably reduced, and the efficiency is improved. And the processing efficiency and the flow consistency are improved.

The invention discloses a Unicycle multi-agent dynamic coverage control method based on safety reinforcement learning, and the method comprises the steps: firstly, carrying out the modeling of a Unicycle agent, an anisotropic sensor, and a multi-agent system dynamic coverage control task; then, designing corresponding constraining force according to safety requirements such as input limitation and obstacle and collision avoidance of Unicycle so as to design a safety protection mechanism, and correcting a control instruction of the intelligent agent when the intelligent agent is about to do unsafe action so as to ensure the safety of the intelligent agent in a dynamic coverage process; and finally, training the dynamic coverage control security decision model in cooperation with a multi-agent deep reinforcement learning algorithm, improving the adaptive capacity of the dynamic coverage control security decision model to a variable environment based on an observation coding mode of state compression, and executing the dynamic coverage control security decision model to complete Unicycle multi-agent dynamic coverage control.

The invention provides a self-adjusting dialogue type multi-agent system and interaction method based on dynamic topology, and the system comprises a scheduling module which is used for receiving user input and routing a task to a corresponding field task agent in a conventional state; the domain task intelligent agents are used for processing the received tasks, evaluating the cognitive states in real time and sending out assistance signals when the cognitive states meet preset conditions; the shared knowledge management module is used for maintaining a dynamically updated distributed semantic knowledge graph; wherein the distributed semantic knowledge graph is used for sharing, associating and persisting structured knowledge entities and context information generated in the session process among all modules in real time; the topology control module is used for dynamically switching the communication control topology of the system from a first topology structure to a second topology structure in response to the received assistance signal; and when the assistance signal is eliminated, the communication control topology is recovered from the second topology structure to the first topology structure.

The invention discloses a multi-agent system double-layer Q learning control method and system based on Warisstein distance. The method comprises the following steps: establishing a multi-agent system state space model under fault and unknown distribution interference; constructing an inner-layer fault-tolerant control system and an outer-layer robust control system; designing an internal fault-tolerant control gain based on a Q learning algorithm; designing an outer layer robust control gain and a bias item based on the Warisstein distance and a zero-sum game framework; and designing a distributed consistency protocol in combination with neighborhood state information. According to the method, under the conditions that a system model is unknown, an executor has additive time-varying faults, external disturbance probability distribution is uncertain and only depends on limited samples, the influence of compensation faults and disturbance can be effectively reduced, the H infinity performance constraint is met, meanwhile, asymptotic state synchronization of all agents is achieved, and system stability is ensured.