146 results about "Agent architecture" patented technology

A computing system for adaptive electronic message classification employs a multi-agent architecture comprising a media feature analysis system, a user context refinement system, and a response synthesis system. The media feature analysis system generates pillar scores including message type, intent, and link risk scores with associated confidence values using trained classification models. When pillar scores and confidence values do not satisfy predetermined threshold conditions, the user context refinement system dynamically constructs contextual prompts using the pillar scores and confidence values as input parameters. User responses generate score modification data that refines the pillar scores and contextual response data for recommendation generation. The response synthesis system generates refined classifications and personalized recommendations using the refined pillar scores and contextual response data. An orchestration system coordinates agent interactions using learned uncertainty points and implements asymmetric influence algorithms with variable weighting based on content and URL analysis concordance.

The disclosed technology provides for an improved approach to AI code generation. In various embodiments, the disclosed technology provides for an AI agent architecture platform for generating, revising, testing, and debugging code using a customizable team of agents with specific tasks.

The invention provides an urban virtual power plant operation method and device based on an intelligent agent architecture, relates to the field of artificial intelligence, and solves the problem that in the prior art, operation strategies related to a virtual power plant are mostly based on static rules and manually set optimization models. And rapid and efficient response is difficult to realize in a multi-participant, multi-constraint condition and dynamic electricity market environment. The method comprises the following steps: collecting multiple pieces of first data in an urban energy system in an urban virtual power plant; performing arrangement and task scheduling on the first data on the basis of data types required by the plurality of agents to execute tasks, determining the first data required to be processed by each agent, and calling the plurality of agents to process the first data required to be processed; and generating a regulation and control decision and a market transaction strategy of the urban virtual power plant based on the processing results output by the plurality of agents and the operation constraint conditions. The method is used in the operation process of the urban virtual power plant.

The invention relates to a water-light storage area power grid optimization method based on federated and layered multi-agent reinforcement learning. The method comprises the steps of constructing a federated and layered multi-agent model, defining states, actions and reward spaces of layered agents, implementing federated reinforcement learning offline training with privacy enhancement, deploying the model and executing online distributed optimization. Through the federal learning framework, it is ensured that private data of all the energy main bodies are locally calculated, the privacy leakage risk caused by data collection in a traditional centralized method is avoided, and a technical basis is provided for building collaborative trust among multiple main bodies. Meanwhile, the designed hierarchical intelligent agent architecture decouples a complex power grid system into subproblems with clear hierarchies, and when the scale of the system is expanded, new energy units can be accessed in a modular mode, so that the overall maintenance and upgrading difficulty of the system is reduced, and the expandability of the system is improved.

The invention provides an automatic consumption tag analysis system and method based on multi-agent collaboration, and the system comprises a data processing agent which is used for collecting data from a social media platform and carrying out the data preprocessing; the label identification intelligent agent is used for extracting consumption labels of different dimensions from the preprocessed data; the sentiment analysis agent is used for carrying out context modeling and sentiment tendency recognition and binding the recognized sentiment tendency to the corresponding consumption label; the label normalization agent is used for performing clustering and normalization processing on all consumption labels bound with emotional tendencies to generate a structured multi-layer label atlas; the central scheduling agent is used for scheduling other agents, generating a label analysis result by using the multi-layer label atlas and sending the label analysis result to the user; according to the invention, based on a multi-agent architecture, structured analysis is carried out on user tags, behavior attributes and consumption intentions in social media contents, so that high-precision and high-efficiency intelligent consumption insight is realized.

The invention relates to the technical field of data processing, and discloses an intelligent agent architecture-based education large model autonomous teaching planning method and system. The method comprises the following steps: inputting a cognitive ability matrix, a knowledge mastery degree vector and a learning preference vector into an improved Bayesian knowledge tracking algorithm, and adding an agent parameter consistency constraint term and a teaching planning coordination factor to obtain a learning track state vector; based on the learning track state vector, multiple agents cooperatively generate a recommendation content identification sequence, a difficulty level sequence and a time distribution sequence; and monitoring the knowledge mastery degree vector variation, and reordering the recommended content sequence when the knowledge mastery degree vector variation exceeds a preset threshold value to generate a teaching sequence. According to the invention, the Bayesian knowledge tracking algorithm is improved, the agent parameter consistency constraint term and the teaching planning coordination factor are fused, and coordination and unification of multi-agent teaching decisions in a communication-free environment are realized.

The invention provides a traffic corridor signal cooperative control method based on single agent reinforcement learning, and relates to the technical field of traffic management and control, and the control method comprises the steps: obtaining the real-time traffic state data of a traffic corridor comprising a plurality of signal intersections; acquiring a current signal control scheme of the traffic corridor; constructing a state vector according to the real-time traffic state data and the current signal control scheme; inputting the state vector into a pre-trained single-agent reinforcement learning model to obtain a corresponding action vector; based on the action vectors, phase division of all the signal intersections is synchronously adjusted, and a new signal control scheme is generated; according to the invention, signal timing of all signal intersections in a traffic corridor is cooperatively controlled by adopting a centralized single-agent architecture, so that the problems of system complexity and training instability caused by local observation, distributed decision and communication coordination among agents in a multi-agent scheme are fundamentally avoided.

The invention provides a multi-type micro-grid cross-layer collaborative scheduling method based on multi-agent reinforcement learning, and relates to the technical field of power system and micro-grid scheduling, and the method comprises the steps: firstly constructing a multi-type micro-grid layered collaborative architecture, a multi-agent reinforcement learning scheduling model, and a centralized coordination layer where global agents are deployed in a platform layer; and the local agents are deployed at each distributed micro-grid node of the platform layer, a cross-layer cooperative training mode is adopted to train the model, and finally, cooperative decision results of the global agents and the local agents are integrated to generate a global-local cooperative scheduling scheme. According to the method, global optimization and local flexibility can be considered, and the hierarchical multi-agent architecture adopts a multi-type micro-grid hierarchical collaborative architecture, so that the method can be suitable for multi-type micro-grid operation scheduling of a novel power system, and multi-type micro-grid collaborative scheduling considering global optimization and local flexibility is realized.

The invention discloses a computing power network resource scheduling method and system, and the method comprises the steps: responding to a service request, carrying out the coding of a network state through an edge attention graph neural network, and enabling the network to integrate an edge feature vector into attention weight calculation in message transmission, so as to sense a link state; based on a coding result, a scheduling decision of joint optimization calculation and network resources is generated through a self-adaptive multi-target reinforcement learning strategy network, and the reward function weight can be dynamically adjusted according to the network state; decisions are executed based on a distributed multi-agent architecture, and all agents realize global coordination through a message coordination mechanism. According to the invention, efficient, adaptive and extensible scheduling of multiple targets in the dynamic computing power network is realized.

The invention discloses a multi-agent resource scheduling method based on a large language model in an industrial scene, and the method comprises the following steps: 1, setting initial scheduling parameters and user input of a multi-agent scheduling system; 2, the planning agent performs intention recognition, feature value extraction and coding; and 3, the scheduling agent receives the scheduling coding information, selects a scheduling algorithm tool and generates a scheduling scheme. And 4, the fault detection agent receives the initial scheduling scheme and receives key parameters of the working agent. And 5, the fault detection agent judges that a fault exists and optimization exists, if yes, the scheduling agent is fed back, and the step 3 is executed again, and if not, an optimal scheduling result is output. According to the method, a multi-agent architecture driven by a large language model is adopted, and a dynamic algorithm tool selection mechanism and a prompt project are introduced. Minimization of task total time consumption and maximization of fault detection accuracy are taken as optimization objectives, and adaptivity of industrial scene resource scheduling and system reliability are remarkably improved.

The invention relates to the field of data assessment, and discloses an enterprise risk assessment method based on a multi-agent risk assessment algorithm MA-ERC, and the method comprises the following steps: S1, constructing an enterprise multi-dimensional risk knowledge graph; s2, designing a risk assessment algorithm MA-ERC to assess initial risks of the enterprises and quantify propagation risk values among the enterprises; s3, constructing an MNF-GNN model to carry out enterprise risk assessment; s4, constructing an enterprise risk assessment interpretable graph EG; according to the method, a multi-agent architecture based on a large language model and a risk propagation algorithm based on a graph are combined, the risk propagation process between enterprises is more accurately and reasonably simulated, and the assessment precision of a risk assessment model can be effectively improved through the calculated propagation risk value; and a new thought is provided for the combination of the multi-agent based on the large language model, the knowledge graph, the graph neural network and other technologies.

The invention provides a natural language interaction AI agent system for an intelligent tower crane and a method thereof. Wherein the agent system comprises a natural language understanding module, a task planning module, a tool scheduling engine, a large language model interface, a data adaptation layer, a result generation module and a learning optimization module; the proxy method comprises the following steps: S1, natural language input processing; s2, intention understanding and task planning; s3, data acquisition and preprocessing; s4, intelligent analysis and reasoning; s5, verifying and optimizing a result; s6, generating a natural language answer; and S7, feedback learning and optimization are carried out. According to the method, natural language interaction between the user and the intelligent tower crane system is achieved by constructing a special AI agent architecture, the large language model is automatically called to intelligently analyze the running state of the tower crane, and analysis results and decision suggestions which are easy to understand are provided for the user.

The invention provides a truss structure autonomous design system and method based on multiple agents. A layered multi-agent architecture is adopted, a general control agent is taken as a core, and a high-level design intention is analyzed and a plurality of downstream special agents are arranged for cooperative work; the whole design process realizes automatic closed loop through a graph-based workflow engine; the core innovation of the method is a hybrid intelligent optimization algorithm of LLM heuristic exploration and genetic algorithm accurate fine tuning, and the algorithm is driven by an original and quantifiable design quality scoring system to efficiently carry out analysis-diagnosis-optimization iteration so as to cooperatively improve the safety and economy of the design. The design time can be shortened from several hours to several minutes, about 7.0% of material consumption is saved on average, it is ensured that the scheme completely conforms to industry specifications, the design efficiency, economy and reliability are greatly improved, and meanwhile the use threshold of professional design software is remarkably reduced through natural language interaction.

The invention provides a code auditing method and device based on a multi-agent architecture, a medium and a product, and relates to the field of data processing, and the code auditing method comprises the steps of obtaining multi-source input data; performing distributed cooperative processing on the multi-source input data through the multi-agent architecture to obtain a plurality of processing results output by the multi-agent architecture; integrating the plurality of processing results through a multi-agent architecture to generate comprehensive judgment information; wherein the comprehensive judgment information is used for indicating a code auditing result; according to the comprehensive judgment information, a corresponding code auditing report is generated, and the method provided by the invention is used for achieving the technical effect of guaranteeing the security and compliance of code auditing.

The invention relates to the technical field of network communication, and provides a hierarchical reinforcement learning scheduling and routing method for a multi-domain TSN, and the technical scheme comprises the steps: collecting the global state information of the multi-domain TSN, and carrying out the coding processing of the global state information to generate comprehensive state representation; performing a cross-domain routing decision, and outputting an inter-domain path and a time delay budget of a cross-domain flow; executing intra-domain scheduling, determining a sending sequence and a specific path of a domain flow, and generating gating list configuration; through hierarchical coordination and strategy optimization, a cross-domain routing decision and intra-domain scheduling are updated based on reward feedback of an intra-domain scheduling result; and deploying the finally updated intra-domain scheduling to a switch of the multi-domain TSN network for execution, thereby realizing deterministic transmission of the time-triggered flow. According to the invention, through a layered agent architecture, a hybrid neural network coding mechanism and a cross-domain collaborative optimization strategy, challenges of complexity, expandability, dynamic adaptability and the like of a joint routing and scheduling problem in a multi-domain TSN environment are effectively solved.

The invention relates to an ODRL strategy generation method based on a large language model multi-agent architecture, which is applied to a system of a multi-agent architecture in an orchestrator-worker mode, and comprises an orchestrator agent and a group of multi-category worker agents, the method comprises the following steps: receiving natural language strategy input of a user through an orchestrator agent, identifying a structure type of language logic of the natural language strategy, and selecting at least one worker agent to form a processing chain according to the structure type; and calling a worker agent in the processing chain, processing the natural language strategy, generating an ODRL strategy, and outputting the ODRL strategy.

The invention discloses an industrial image anomaly detection method based on a multi-agent arrangement heterogeneous algorithm. The method comprises the following steps: S1, constructing an algorithm component library for shielding data representation differences among different architecture algorithms; s2, constructing a multi-modal large model system based on a multi-agent architecture; the multi-agent architecture comprises the following steps: receiving an image by using a visual expert agent, and outputting structured physical metadata and an unstructured visual suggestion text; the planner agent receives the physical metadata and the visual suggestion text, and generates an algorithm configuration tree based on an algorithm component library; the optimizer agent optimizes the algorithm configuration tree in the closed-loop feedback stage; s3, using the algorithm configuration tree to construct a reference model representing good product distribution according to the normal training sample images; s4, performing anomaly detection on the to-be-detected sample image by using the algorithm configuration tree to generate an anomaly judgment result; and S5, optimizing the algorithm configuration tree.

The invention relates to the field of sign language emotion interaction, and provides a multi-mode sign language emotion interaction system and method based on an intelligent agent architecture, and the system comprises a recognition module which is used for recognizing a current task type based on multi-mode data of a user, and generating a task scheduling instruction; the large language model module is used for generating feedback content by utilizing a large language model according to the task scheduling instruction; the emotion module is used for extracting keywords from the dialogue history and environment context information of the user, judging the emotion of the user through an emotion recognition algorithm, generating prompt words with humor / comforting elements by utilizing a large language model in combination with the emotion and the keywords of the user, and sending the prompt words to the user; according to the cue word, utilizing a large language model to select proper language content to generate feedback content with emotion; and the generation module is used for displaying the feedback content with the emotion to the user in a sign language form and / or a voice form through a digital person. According to the method, the defects that an existing system is fragmented and cannot be expanded, and situations cannot be understood are overcome.

Systems and methods are disclosed for automated generation, validation, and deployment of infrastructure-as-code (IaC) using a multi-agentic artificial intelligence / machine learning (AI / ML) architecture. An extraction agent parses multimodal artifacts (e.g., diagrams and / or configuration data) to derive or generate infrastructure set-up parameters. A coding agent generates IaC units based on the extracted parameters. A validation agent determines the syntactic and semantic compliance of the generated IaC, classifying the IaC as executable, non-executable, or identifying corrective actions. A deployment agent transmits executable IaC to target computing environments and manages automated provisioning of the IaC in the target environments. In cases of validation failure, error indicators are provided to a user (e.g., via an agentic chat bot) for clarification or correction, enabling iterative refinement.

The invention relates to the technical field of power distribution network protection, and discloses a power distribution network fault scene constant value verification method based on reinforcement learning, and the method comprises the steps: constructing a hierarchical multi-agent verification system; a hierarchical multi-agent verification system is constructed, a high-level agent performs fault situation awareness and decomposes a global constant value verification target, sub-tasks are allocated to a low-level agent for execution, and task allocation and execution are realized through collaborative optimization; after a large number of distributed photovoltaics are connected into the power distribution network, the complexity and uncertainty of a power distribution network fault scene are enhanced, fault types comprise two-phase short circuit, two-phase grounding short circuit and three-phase short circuit, the protection behavior difference under different fault positions is obvious, the fault characteristics are changed, the increasing assisting and drawing-out effects are generated, and the reliability of the power distribution network fault scene is improved. Therefore, over-current protection range changes, protection mismatch, override trip and maloperation are caused, and power failure range is expanded. For a complex fault scene, a layered multi-agent architecture is adopted, and efficient task allocation and execution are realized through high-level coordination and bottom-level negotiation.

The invention relates to the technical field of robot interaction, in particular to a guide robot interaction method and system based on a large language model and an intelligent agent architecture, and the interaction method comprises the steps: S1, carrying out the input processing of a user; s2, identifying and classifying user intentions; s3, service intelligent routing is carried out; s4, executing control; and S5, performing interrupt processing. The interaction system comprises a user interaction layer, a core processing layer, a back-end service layer and an execution layer.

The invention relates to the field of artificial intelligence, discloses a full-duty investigation method and system based on a large model hierarchical agent architecture, and aims to solve the problems of efficiency bottleneck, knowledge blind areas, data stream solidification and the like in existing full-duty investigation. The method comprises the steps that a central task planning agent receives an instruction and decomposes and dynamically reconstructs a task; the agent modules in each field refine tasks and cooperatively analyze multi-source data; the central evidence management agent constructs a global evidence chain; the central report generation agent generates a report. The system comprises a central task planning module, a plurality of fields, a central evidence management module and a central report generation agent module. According to the technical scheme, intelligence, efficiency and depth of full-duty investigation can be remarkably improved, existing problems are solved, and through dynamic task reconstruction, knowledge fusion, evidence verification and full-link traceability, conclusion objectivity and reliability are improved, and systematic risks are reduced.

The invention discloses a cross-system collaborative to-do task query and processing method and system based on an AI Agent architecture, and particularly relates to the technical field of artificial intelligence and information system integration. According to the method, to-be-handled task data in each heterogeneous service system is automatically collected through a distributed deployment AI Agent module, and standardized processing is carried out through a unified task modeling module; identifying and combining repeated tasks by using a semantic understanding and deduplication module based on natural language processing, and generating a global unified task view; distributing tasks according to a dynamic strategy through an intelligent task routing and distribution module; realizing cross-system operation and real-time state synchronization by means of a task execution and state synchronization module; and finally, providing an aggregated task list and an operation interface for a user through a unified interaction interface. According to the system, unified integration, intelligent processing and collaborative management of multi-source heterogeneous tasks are effectively realized.

This invention discloses a secure voice-based intelligent customer service system and interaction method based on a multi-agent architecture. The system includes an ASR (Automatic Speech Recognition) module, multi-layered security barriers, a main retrieval agent, a specialized retrieval agent cluster, a parallel retrieval and result aggregation module, an LLM (Limited Language Modeling) synthesis engine, and a TTS (Text-to-Speech) module. Through multi-agent system architecture and methodological design, this invention achieves modular processing, parallel retrieval tasks, and closed-loop security control. By enabling independent calls between input, output, agents, and inference stages via multi-layered security barriers, the system effectively mitigates risks such as prompt word injection, indirect injection attacks, harmful content generation, and information leakage. Sensitive information identification, access permission verification, and content compliance review are embedded in each stage of input, retrieval, information synthesis, and output to prevent sensitive data leakage and meet the application requirements of high-security business scenarios (such as finance, telecommunications, and government services).

The application discloses a multi-position defect repairing method based on repairing sequence scheduling and multi-agent cooperation. First, the defect analysis agent is used to identify the code error root cause and generate a defect analysis report. Then, the report, error code, local context, failed test cases and test error information are combined to build a repair context. Then, the multi-agent architecture is used to iteratively generate patches: the coordinator agent analyzes the dependency relationship between code blocks and schedules the repair sequence, the proposer agent generates a local candidate patch, and the coordinator evaluates and selects the optimal patch. Repeat the process until all repairs are completed and a complete patch is assembled. Finally, according to the compilation error and test failure report, the patch is executed in two stages of syntax and semantics refinement. The application overcomes the defects of the existing large model repairing method, such as lack of sequence scheduling and difficulty in handling code block dependencies, and realizes efficient repair of complex multi-position defects.

Systems and methods are disclosed that create a deployment package for commissioning of a building control system. The deployment package may include a series of automated validation sequences that ensure each equipment and / or device is properly configured and installed. A validation sequence is performed, and the behavior is compared to expected behavior for the sequence. A brokering architecture facilitates the commissioning by providing a fixed schema to all applications communicating over the message bus, thus standardizing the deployment package and the validation sequences. Correlations between data may be found and used when data is missing in the same or similar equipment.

The invention discloses a semiconductor valve fluid simulation method based on deep reinforcement learning fusion, and relates to the technical field of semiconductor equipment manufacturing, and the method comprises the steps: collecting real-time working condition data, and dynamically calibrating the helium detection inner leakage rate, the flow coefficient CV and the adaptive threshold of the inner surface roughness in combination with the service degradation characteristics of a valve; a perception-decision double deep reinforcement learning agent architecture is established, a perception agent extracts multi-physics field high-dimensional features through a convolutional neural network and performs noise reduction, and a decision agent drives a simulation model to adaptively adjust structural parameters by adopting an improved double-strategy reinforcement learning algorithm; starting multi-physics field simulation and double-agent closed-loop iteration until a simulation result meets a self-adaptive threshold value requirement; and introducing a multi-target degradation compensation mechanism based on an iteration result, reversely correcting valve service degradation characteristic related parameters, and outputting a valve simulation optimization scheme. According to the invention, the problem of low simulation precision in the prior art is solved, and intelligent and accurate optimization of the semiconductor valve structure is realized.

The invention relates to the technical field of energy storage, and discloses a global protection method, device and equipment for an energy storage system and a medium, and the method comprises the steps: obtaining temperature data and environment data monitored by a global sensing network; a three-level multi-agent framework is adopted, a particle swarm optimization algorithm is utilized to output a cluster global decision and an equipment control instruction, protection modes are switched, a corresponding protection strategy is determined, and the protection modes comprise a normal operation protection mode, a fault early warning protection mode and an emergency isolation protection mode; according to the invention, a global sensing network is used for monitoring parameters, a monitoring blind area is eliminated, full-dimensional parameter monitoring is realized, a three-level multi-agent architecture is used for outputting a global decision and an equipment control instruction, compliance is guaranteed from the decision level, a multi-level protection mode is switched, and the protection mode is dynamically adapted. Safety and efficiency are balanced, and the hierarchical protection device executes a protection strategy, so that global protection is realized, and global safety protection capability is improved.

The application relates to the technical field of satellite intelligent access, and discloses a multi-satellite intelligent access method, system and equipment based on AI prediction, which comprises the following steps: collecting satellite orbit data and user terminal position data of a multi-satellite system, and performing satellite position prediction through an orbit prediction neural network to obtain satellite predicted positions and communication window information; each intelligent agent perceives communication environment parameters and constructs an environment perception vector; a multi-agent strategy network is driven based on the satellite predicted positions and the environment perception vector to generate a first access strategy; each intelligent agent performs resource conflict processing according to the first access strategy to obtain a second access strategy; the application adopts a multi-agent architecture to replace traditional centralized control, each intelligent agent has independent decision-making capability and realizes cooperation through a distributed negotiation protocol, single-point failure problems are avoided, robustness and scalability are improved, and resource competition and access conflicts are effectively reduced.

This invention discloses a method for agent task allocation based on deep reinforcement learning, comprising the following steps: design of a strong and weak agent architecture; design of a Markov decision process for cooperative behavior; design of a phased reward mechanism; design of the agent network structure; and design of a multi-head attention mechanism and target selection. This invention proposes a "one strong agent leading N weak agents" multi-agent structure, which can reduce system complexity and eliminate the weakness of multi-agent systems prone to interaction conflicts when dealing with complex problems. Through the multi-head attention mechanism and the phased reward mechanism, the efficiency and stability of training are effectively improved.