375 results about "Routing decision" patented technology

The invention discloses an Internet of Things data transmission optimization system based on intelligent routing, and relates to the technical field of data transmission optimization. The heterogeneous network access module supports adaptive switching of multiple protocols such as 5G, NB-IoT and LoRa, and monitors link quality; intelligent routing decision selects an optimal path by means of deep reinforcement learning, and end-to-end delay is reduced; data distribution scheduling split data streams are transmitted in a multi-path parallel mode, and forward error correction is combined to guarantee reliability. The network state prediction uses an LSTM to predict bandwidth, and a route is adjusted in advance; the secure transmission integrates iris authentication and dynamic encryption. According to the invention, efficient, reliable and secure transmission is cooperatively realized, energy collection and dormancy strategies are combined, the equipment endurance is greatly prolonged, the network energy consumption is reduced, and the method is suitable for industrial control, telemedicine and other scenes with high requirements on real-time performance and reliability.

The invention belongs to the field of data processing, and discloses an archive arrangement multi-mode intelligent AI classification system and method. Comprising the following steps: introducing a cross-modal consistency loss function and structured constraint loss into a deployed dual-channel generator, and carrying out facies evaluation on original multi-modal data and evolutionary multi-modal data based on a modal quality evaluator; through conflict identification and digestion mechanism loop optimization, features are extracted based on the enhanced multi-modal data to obtain a quality scoring triple; a PPO algorithm is used for performing dynamic routing decision, quality evaluation closed-loop optimization is introduced, path rules are executed according to decision actions, a multi-modal feature matrix is constructed based on comprehensive features, a multi-modal classifier is designed based on Attention-Transformer, a modal weight adaptive adjustment module is introduced, a classification result verification mechanism is added, and closed-loop optimization is formed; and intelligent AI classification of the multi-modal archives is realized.

The invention, which relates to the technical field of customer relationship management, discloses a multi-channel interactive customer relationship management system comprising a dynamic routing decision module, a multi-modal data fusion module and an intelligent feedback optimization module. The dynamic routing decision module evaluates channel load through a deep neural network, dynamically allocates client requests to an optimal node by utilizing reinforcement learning, and realizes load balancing and service continuity; the multi-modal data fusion module integrates text, voice and image data, constructs a space-time correlation graph, identifies a cross-channel behavior mode, and ensures data consistency through multi-dimensional verification; the intelligent feedback optimization module combines customer satisfaction evaluation and multi-modal sentiment analysis, optimizes a service strategy by using a genetic algorithm, and synchronizes the service strategy to a cross-channel knowledge graph to realize adaptive iteration; according to the method, the problems of unreasonable multi-channel load distribution, insufficient data fusion and consistency verification and inaccurate service strategy optimization are effectively solved, and the customer service quality and experience are improved.

The invention relates to the artificial intelligence technology, can be applied to service system platforms such as medical health and financial science and technology, and discloses a data processing method, device, equipment and medium based on a hybrid expert model.The method comprises the steps that to-be-processed data is input into the hybrid expert model, routing probability distribution is obtained, the Tsallis entropy of the routing probability distribution is calculated, and the Tsallis entropy of the routing probability distribution is calculated; generating a prediction result according to the Tsallis entropy; evaluating loss based on an auxiliary entropy loss function, constructing a subspace, and adjusting pre-training parameters in the subspace to obtain an optimized hybrid expert model; and obtaining a target prediction result according to the optimized hybrid expert model. According to the dynamic routing mechanism, experts are flexibly selected, an entropy loss function is assisted to optimize a routing decision, uncertainty is reduced, and model convergence is accelerated; and interference of new tasks on old tasks is eliminated through re-parameterization and subspace design, so that the model achieves good balance between stability and plasticity, and generalization and adaptability of model data processing are improved.

The invention provides a large model lightweight reasoning deployment method under limited hardware resources, and the method comprises the steps: quantifying the weight importance of a large model through a composite index of gradient sensitivity and activation frequency, and carrying out pruning operation in combination with an improved index weighted moving average strategy, thereby obtaining a structured sparse model; the sparse model is divided into sub-networks by adopting double rules, a routing decision network is trained, and an adaptive feature shunting architecture model is constructed; a multi-precision weight set is generated through a nested quantization technology, quantization bit width is dynamically adjusted, and edge equipment hardware parameters are adapted to complete reasoning environment initialization; after a reasoning request is received, an optimal sub-network is selected based on the trained routing decision network, corresponding weights are loaded in parallel, and a reasoning result is fused and output; and converting a reasoning result format, and dynamically optimizing a scheduling strategy based on a system real-time monitoring index. The method is compatible with a mainstream large model and a hardware platform, and an efficient and universal deployment scheme is provided for end-side AI engineering landing.

The invention discloses a multi-dimensional computing power dynamic perception routing decision-making method and system based on SRv6 driving. The method comprises the following steps: firstly, screening an optimal target computing power node based on a service demand and a node real-time load; and then, based on the dynamic network topological graph fusing the service sensitivity, calculating an optimal network path reaching the computing power node by using an improved SPFA algorithm, and realizing a global collaborative decision of selecting the most suitable computing point and searching the most efficient connection path. Meanwhile, based on an SRv6 driving service chain dynamic generation and closed loop execution method, a target computing power node and necessary network functions are abstracted into a programmable SID, an SID sequence (service chain) is dynamically constructed according to a double-stage decision result and is packaged in an SRH head, second-level path issuing and state monitoring are achieved through a programmable controller, and a real-time monitoring result is obtained. According to the method, the problems of poor real-time performance, single dimension, weak cooperative capability and the like in the prior art are solved, and efficient, stable and intelligent development of a future-oriented distributed intelligent computing network system can be promoted.

The invention provides a multi-satellite cooperative distributed routing method for multi-agent reinforcement learning, and the method comprises the steps: building a satellite network static topology model through a time slicing technology, and building a target function for minimizing the end-to-end time delay of a data packet; constructing a satellite agent network based on the satellite network static topology model, and obtaining satellite interaction experience data; wherein the satellite agent network comprises a multi-satellite cooperative hybrid network and a satellite decision network corresponding to each agent; training a satellite agent network according to the satellite interaction experience data to obtain a trained satellite agent network; and respectively deploying the satellite decision networks in the trained satellite agent network to corresponding agents, so that the agents perform routing decision based on the deployed satellite decision networks. Through cooperative work of the satellite decision network and the multi-satellite cooperative hybrid network, efficient distributed routing decision of a low-orbit satellite constellation can be realized, time delay is reduced, and load is balanced.

The invention provides an adaptive path selection system based on GNN and multi-agent DRL, which is based on a layered multi-domain SDN architecture and comprises a root server layer, a domain controller layer and a data layer. The root server layer comprises a root server and a root agent, and the root agent is used for calculating a global routing path strategy and optimizing and adjusting a routing path according to feedback of a local network domain. And the root server distributes the routing path to a domain controller layer according to a routing path strategy to carry out routing path forwarding and adjustment of a local network domain. The domain controller layer comprises a plurality of domain controllers and domain intelligent agents, the domain controllers construct a topological structure composed of boundary switches, and a routing decision is made in a local domain according to an optimal global routing path strategy. And the domain intelligent agent performs modeling analysis on the topological structure by using a deep reinforcement learning algorithm in combination with a graph neural network to realize intra-domain path selection. And the data plane layer feeds back the network topology information to the domain controller, and receives and executes a routing decision sent by the domain controller at the same time.

The invention provides a knowledge base and business system collaboration method under an AI platform, and belongs to the technical field of AI platform digital data processing.The method includes the steps that triple semantic analysis is conducted on query by constructing a multi-level semantic vector representation module, a parallel data retrieval engine is started, vector retrieval, graph reasoning and real-time data pulling are executed at the same time, and the query efficiency is improved; establishing a dynamic confidence evaluation mechanism to evaluate the quality of the data source, executing a weight distribution algorithm based on reinforcement learning, dynamically calculating the weight of the data source according to a query type by adopting a graph convolutional network intelligent routing decision model, and implementing multi-source data fusion and consistency verification to solve data conflicts through a weighted voting mechanism. An intelligent result sorting and filtering system is established, a multi-dimensional evaluation strategy is adopted to output high-quality answers, a continuous learning and feedback optimization loop is constructed, system performance is continuously optimized through active learning and a graph shortest path algorithm, and the technical problem that knowledge base data and a service system cannot effectively and uniformly make decisions is solved.

The invention provides a multi-layer satellite network routing method and system based on multi-agent deep reinforcement learning. The method comprises the following steps that: a plurality of low-orbit satellites train respective executor networks by adopting a distributed multi-agent reinforcement learning algorithm to realize routing decision; and one medium-orbit satellite trains a global evaluation network according to the local parameters of the performer network uploaded by the plurality of low-orbit satellites, and provides the global parameters of the trained global evaluation network to each low-orbit satellite, so that the low-orbit satellites update respective local evaluation networks according to the global parameters. And updating respective executor networks according to the updated local evaluation network. According to the multi-layer satellite network routing method and system based on multi-agent deep reinforcement learning, dynamic optimization of a routing strategy can be realized in a medium and low orbit satellite network, and the overall performance of the satellite network is remarkably improved.

The invention discloses an intelligent policy question-answering method and system based on retrieval enhancement generation and a medium. The method comprises the following steps: constructing a universal knowledge base and a plurality of mutually independent domain knowledge bases; in response to user input, the following steps are executed: performing routing analysis on the user input through a first large language model to generate a structured routing decision; retrieving the general knowledge base to obtain related general knowledge text segments and vectorization expressions thereof; according to the routing decision, a domain knowledge base corresponding to the at least one policy domain identifier is retrieved in parallel, and related policy text fragments corresponding to the at least one domain knowledge base and vectorization expressions of the related policy text fragments are obtained; obtaining an initial answer set based on retrieval results of the general knowledge base and the domain knowledge base; and performing intelligent fusion processing on the initial answer set through a second large language model, and generating and outputting response content. According to the method, the problem of knowledge updating lag is effectively solved, and the accuracy, timeliness and cross-domain specialty of policy questions and answers are improved.

The invention discloses a hydrological data synchronization method and device based on distributed storage and a medium, and the method comprises the steps: dividing hydrological data collected by different hydrological monitoring point sensor devices in real time into a plurality of different grades according to the data emergency degree, and adding a space-time label to the hydrological data; fragmenting the hydrological data based on the space-time dimension to obtain a standard data block, and determining a regional storage node corresponding to the standard data block; according to the real-time network state, determining an optimal transmission path corresponding to different levels of hydrological data; and based on the consistency transmission rules corresponding to the multiple different levels, synchronizing the standard data blocks to the corresponding regional storage nodes through the optimal transmission path. Through level-to-level management and intelligent routing decision making of hydrological data, an optimal transmission path is selected according to characteristics of different levels of data and network states, resource waste and transmission delay caused by a traditional unified synchronization strategy are avoided, and the data synchronization efficiency is remarkably improved.

The invention discloses an eBPF-based kernel optimization method, system, equipment and medium, belongs to the technical field of Linux kernel network optimization, and aims to solve the technical problem of how to improve the transmission performance for a specific data stream and specify the data stream to carry out a special routing and transmission process so as to provide services for a special network adapter and improve the transmission performance of the specific data stream. According to the technical scheme, the method comprises the steps that data packet interception is carried out on a plurality of kernel eBPF loading points and program interception points selected by a user through an eBPF kernel hook program, routing query is carried out with a routing kernel module, and selection of an operation function and routing parameters is completed; the routing kernel module completes a routing decision based on various complex routing requirements, configuration conditions, data packet conditions and environment conditions, and returns a query result to the eBPF kernel hook program, and the query result is executed by the eBPF kernel hook program. The system comprises an eBPF user mode management program module, a user mode control tool, an eBPF kernel hook program module and a routing kernel module.

The invention provides an outbound routing method and system based on multi-dimensional index coupling and large model decision, and the method comprises the steps: carrying out the preprocessing of the historical data of each line provider, predicting the future performance index through an LSTM neural network, and optimizing the prediction result through a Kalman filter, and eliminating the noise. In the route decision-making link, the system calculates the Q value of each line provider based on the deep Q learning network, comprehensively considers the call completing rate, average cost, negative feedback rate and other factors of each line provider, and intelligently selects the optimal line provider in combination with the user portrait, scene recognition and real-time load state. The system also constructs a complete feedback closed loop, continuously optimizes the model by collecting call result data, and automatically adjusts decision parameters. The call completing rate and the service quality of the outbound service are improved, and the communication cost is reduced.

A routing method for supporting differential service requirements in a satellite internet comprises the following steps: step 1, constructing multi-dimensional network state perception as a base layer which is responsible for collecting and quantifying physical states and logic states of a satellite network in real time, and obtaining objective network states and subjective QoS (Quality of Service) requirements of services; 2, generating a nonlinear adaptive weight driven by a service as a strategy layer, mapping an objective network state and a subjective QoS (Quality of Service) demand of the service, dynamically adjusting the value orientation of routing judgment, and generating a dynamic weight; and step 3, taking the dynamic cost routing decision as an execution layer, calculating the comprehensive cost based on the dynamic weight, and executing next hop selection and path maintenance, and through cooperative work of the dynamic cost routing decision, the comprehensive cost and the next hop selection and path maintenance, realizing intelligent path selection and high-reliability transmission for different service flows in a high-dynamic and resource-limited environment of the low-orbit satellite. According to the invention, while the routing computation overhead is effectively reduced, multiple requirements of differentiated services on low time delay and high reliability can be considered.

The invention relates to the technical field of routing context connection, in particular to a zero-code visualization and automatic routing method for large-model servitization orchestration, which comprises the following steps of: analyzing a user session text, extracting a core entity and an intention dependency relationship, constructing an entity intention graph, compressing to generate a state snapshot, and distributing a global session ID (Identity); the method comprises the steps that a snapshot is bound to a distributed routing decision tree node and stored, cross-node state continuity is established, when a new request is received, the snapshot and new session semantics are fused, a routing weight offset correction decision table is calculated, and a routing decision logic flow is generated by dragging an interface configuration rule and strategy. The method improves the service matching degree, reduces the development threshold, and is suitable for agile construction of large-model collaborative application.

The invention discloses a dynamic bypass fault-tolerant router in a network-on-chip. The dynamic bypass fault-tolerant router comprises input and output ports, an input buffer area, a crossbar switch, a routing calculation unit, a self-checking circuit and a bypass channel. Key features comprise a fault information unit, a bypass selector, a single flit buffer area and an arbitration unit. The router detects the self-fault through the self-checking circuit, and the fault information unit stores the information. And if the router fails, the bypass selector transmits the data packet through the bypass channel and bypasses the failed router. If the router is healthy, the router exchanges fault information with an adjacent router, the fault information is used for routing decision of a data packet after integration, and meanwhile, the router can select a routing path with smaller congestion in an effective direction according to the flow distribution condition in a network. According to the method, the area overhead is relatively low, the routing hop count of the data packet is remarkably reduced, the transmission delay of the data packet is reduced while the high reliability is maintained, and the network throughput is improved.

The invention relates to the technical field of intelligent decision making, can be applied to business scenes such as financial science and technology and medical health, and discloses a routing decision making method, device and equipment for data processing and a medium, and the method comprises the steps: constructing a label strategy SQL training set, and carrying out the supervision training of the SQL training set, and obtaining a trained language model; receiving a to-be-processed SQL task, generating a structured feature vector, generating a routing strategy through a reinforcement learning module, and determining a processing engine of the task; after the execution path identifier is analyzed, the task is distributed to a corresponding engine through message middleware; and generating a feedback data set based on the task execution index and the execution path identifier, and updating the strategy parameter based on the data set. According to the method, the label strategy SQL training set is constructed, decision optimization is carried out in combination with the reinforcement learning module, the optimal processing engine of the calculation task can be judged, calculation overhead and resource waste are reduced, and particularly the calculation performance is improved in a large-scale data processing scene.

The invention discloses a computing power network redundant path construction method based on multi-modal multi-objective optimization and a scheduling system, and belongs to the technical field of routing optimization in a computing power network. According to the method, firstly, for a constructed multi-objective optimization function including the minimum average time delay of all service traffic, the minimum maximum link bandwidth utilization rate and the maximum successful routing rate of all traffic, a transformer-based deep reinforcement learning and truncation selection strategy is adopted to realize multi-modal multi-objective optimization of routing decision; compared with the prior art that an equivalent path cannot be obtained only through an optimal path, the solving method provided by the invention can make an optimal choice for problems such as routing link faults on the premise of ensuring Qos.

The invention relates to a satellite network distributed routing method, a satellite network distributed routing system, satellite network distributed routing equipment and a storage medium, belongs to the technical field of wireless communication, and solves the problems of dynamic topological structure change and overlarge load of part of nodes caused by non-uniform service volume distribution in the existing satellite network routing method. The satellite network distributed routing method comprises the following steps: firstly, modeling a routing problem, constructing a problem model, then defining a state space, an observation space, an action space and a reward function of a Markov decision process, constructing a DDQN-based satellite network routing algorithm training framework based on the Markov decision process, and constructing a satellite network routing algorithm training framework based on the DDQN. And then a satellite network routing algorithm training framework based on DDQN is utilized to realize satellite network routing decision, and the network load is reduced with the purpose of minimizing end-to-end delay. The method is not influenced by the dynamic change of the network topology, and can adapt to the frequent network topology change of the low-orbit satellite, thereby ensuring the real-time performance and robustness of data transmission.

The invention provides a distributed large-scale anti-traceability elastic network intelligent routing method and system, belongs to the field of network communication and network security, and is suitable for intelligent routing decision optimization in a dynamic network environment. The method is based on a multi-agent reinforcement learning framework, network nodes are mapped into independent agents, and path traceability risks are blocked through local information constraints; dynamic feature aggregation of a neighbor link state is realized by adopting a lightweight graph attention network, and the local sensing efficiency of a large-scale network is improved; a QMIX algorithm is introduced, and network parameters are optimized by nonlinear fusion of a local Q value and global graph state representation through a hybrid network; and in combination with a self-adaptive exploration mechanism driven by action entropy, the sudden change scene strategy response capability is enhanced. According to the system, in military anonymous communication, dark network data transmission and cross-border sensitive services, the anti-traceability and transmission efficiency balance can be remarkably improved, the characteristics of high concealment, high elasticity and low resource consumption are achieved, and a systematic routing solution is provided for a dynamic network environment.

The invention provides an unmanned aerial vehicle cluster intelligent opportunistic routing method based on multi-agent deep reinforcement learning, and aims to optimize a data transmission path in an unmanned aerial vehicle cluster and improve communication efficiency. The method comprises the following steps: firstly, establishing a communication channel model and a motion model of an unmanned aerial vehicle cluster, and constructing a distributed partial observation Markov decision process model; based on the model, a multi-agent near-end strategy optimization reinforcement learning algorithm is adopted to carry out routing decision training, and after training is completed, a collaborative strategy of the agents is loaded into an airborne computer of the unmanned aerial vehicle, so that intelligent opportunistic routing is realized. In the training process, the unmanned aerial vehicle selects a proper next-hop node to optimize data forwarding by continuously collecting neighbor information and making a local decision. Through the method, the unmanned aerial vehicle cluster can dynamically adapt to a complex environment, and the end-to-end transmission efficiency of a network and the successful transmission rate of data packets are remarkably improved.

The invention discloses a wireless ad hoc network routing and resource cross-layer joint optimization method based on a genetic algorithm. The method comprises the following steps: step 1, aiming at a plurality of end-to-end communication services in a wireless ad hoc network, carrying out mathematical modeling on a normalized difference sum of each service flow path hop count and a theoretical minimum value thereof and a constraint condition; performing joint optimization on link layer resource allocation and network layer routing decision; and step 2, solving the mathematical model based on a genetic algorithm to realize the targets of reasonable distribution of network resources and optimization of service quality. According to the method, link layer resource allocation and network layer routing decision are jointly optimized in a global mode under the condition that network communication resources are limited, it is ensured that a service flow is close to shortest path transmission as far as possible, end-to-end delay and network loads are reduced, and effective allocation of resources and flexible construction of routing are achieved.

The invention discloses a large-scale low-orbit satellite network domain-division intelligent routing method, which belongs to the technical field of satellite communication, and comprises the following steps: constructing a network model based on multi-dimensional resource characteristics, and providing resource constraint conditions and space-time correlation characteristics for routing decision by combining with a regional flow prediction algorithm of an improved Transform architecture; virtual nodes and an autonomous domain are divided based on a geographic area, and a low-orbit satellite network routing process is decoupled into an intra-domain part and an inter-domain part; a multi-agent deep Q network is adopted in a domain, a sum tree mechanism-based deep Q network is adopted between domains, a routing algorithm and a resource scheduling algorithm are respectively designed, and a domain intelligent routing system with flexibility and expandability is constructed through hierarchical routing and resource decoupling scheduling. According to the method, the service capability of the low earth orbit satellite network in extreme scenes such as topology high-frequency change, strict resource limitation and traffic space-time mutation is remarkably enhanced, and a high-reliability and low-delay routing solution is provided for space-ground integrated communication, emergency disaster early warning and global real-time data transmission.

The invention discloses a Mesh routing selection method based on a graph neural network and reinforcement learning. The Mesh routing selection method comprises the following steps: step S1, constructing a Mesh network graph structure model; s2, extracting an embedded representation of each router node by using a graph neural network; s3, designing a routing decision mechanism based on Q-learning, and generating a Q value in combination with the current router node state and neighbor information; s4, introducing a Masking mechanism to process the dynamic action space, and shielding illegal next-hop action; and step S5, defining a multi-target reward function which is used for guiding a Q-learning training process. The invention provides a Mesh routing selection method based on a graph neural network and reinforcement learning, and aims to solve the problems of poor adaptive capacity to a dynamic network environment, low routing efficiency and lack of intelligent decision support in the prior art and improve the data transmission performance of a Mesh network.

A computer-implemented system for routing artificial intelligence (AI) queries. The system utilizes a zero-copy data pipeline, which processes prompts in memory-mapped buffers to eliminate at least one memory copy operation, thereby reducing latency relative to conventional serialization pipelines. The system continuously verifies AI provider compliance by injecting synthetic prompts containing invisible, Ed25519-signed Unicode watermarks. Algorithmic bias is detected by generating counterfactual “digital twin” prompts and applying Fisher exact statistical testing.Routing decisions for multi-tier autonomous systems are governed by safety-level requirements (ASIL-D, ASIL-B, QM) and may be constrained by external routing directives received via a meta-identifier. A hash-chained manifest, cryptographically signed using Ed25519 and consumed by downstream gateways, is generated for each routing decision, with its Merkle root asynchronously anchored to a blockchain to create a tamper-evident audit trail for regulatory compliance.

Systems and methods for processing artificial intelligence user requests including receiving a user input from a user device, performing a routing analysis on the user input on multiple characteristics, making a routing decision based on the routing analysis to send the user input to one or both of an experiential reasoning agent model and an analytical reasoning model, routing the user input to at least one of the experiential reasoning agent model and the analytical reasoning model responsive to the routing decision, receiving one or more outputs from at least one of the experiential reasoning agent model and the analytical reasoning model, generating a final result by performing a result validation procedure on the one or more outputs, and transmitting the final result to the user device.

The invention provides a multi-platform informatization construction system and method based on a computer, and solves the problems that existing multi-platform data synchronization is not timely, formats are not compatible and the like. The system comprises a real-time data capture module, and a data change event is generated in real time through database transaction log monitoring; the intelligent routing decision module optimizes a transmission path in combination with rules and machine learning; the data conversion adaptation module realizes automatic adaptation of a format and an interface by utilizing dynamic template matching; the conflict detection and resolution module guarantees data consistency based on version control and priority judgment; the data synchronization execution module transmits data according to an optimized path, and the monitoring management module monitors the whole process in real time. The method comprises the steps of data capture, routing decision, format conversion, synchronous execution and monitoring feedback. According to the scheme, the data synchronization efficiency and accuracy are improved, the multi-platform integration operation and maintenance cost is reduced, and the method is suitable for multi-scene informatization construction of enterprises, government affairs and the like.

The embodiment of the invention provides an elastic fragment routing and cold and hot data optimization method based on multi-dimensional feature prediction and related equipment, and belongs to the technical field of distributed database storage optimization. The method comprises the following steps: receiving a data stream from Internet of Things equipment; performing analysis and feature extraction on the data stream to obtain multi-dimensional feature data; according to the multi-dimensional feature data and the time sequence prediction model, performing prediction analysis on a fragmentation strategy to obtain an optimal fragmentation strategy; analyzing the routing decision according to the data priority of the data stream and a dual-mode routing engine to obtain routing decision result data; generating a target index name and creating a target index according to the optimal fragmentation strategy and the routing decision result data; according to the target index name, the data stream is routed to the target index according to the fragmentation key for batch writing; the fragmentation keys are from multi-dimensional feature data. According to the embodiment of the invention, intelligent prediction of the fragment number and automatic hierarchical storage of data can be realized, and the problems of resource elastic scaling and cost efficiency optimization are fundamentally solved.

The invention provides a dynamic routing forwarding method and device for an AI large model. Constructing a collaborative architecture of a centralized scheduler and distributed execution nodes; collecting multi-dimensional performance indexes of each model instance in real time, wherein the multi-dimensional performance indexes comprise GPU utilization rate, video memory occupation and reasoning delay; extracting key features of the input request based on a natural language processing technology, wherein the key features comprise an input text length, a Token number and a task type; a load evaluation model based on an entropy weight method is constructed, and an intelligent routing decision is implemented in combination with the real-time load state and request characteristics of the instance; and instance fault detection and adaptive fault-tolerant processing are carried out based on a multi-level threshold mechanism. The invention provides a more intelligent and efficient dynamic routing method for specific requirements of multi-instance deployment of the AI large model. The load state and the performance index of the model instance are monitored in real time, and intelligent scheduling is performed in combination with the request characteristics, so that the resource utilization rate and the service quality are improved, and the operation and maintenance complexity and the management cost are reduced.