654 results about "Edge server" patented technology

The invention discloses a multi-edge device collaborative reasoning method and system for a hybrid expert large model, and the method comprises the steps: collecting and analyzing system data, and adjusting the expert layout according to the system data; in the first stage, the number of experts required by each layer of a server is dynamically determined by balancing activation diversity and memory resource limitation; in the second stage, according to the expert number and the activation mode of each layer obtained in the first stage, low-time-delay reasoning is achieved by minimizing the calling times of remote experts, and the experts are distributed to all the servers. The system includes a global scheduler and an edge server participating in the system. By using the method, edge multi-machine joint reasoning is achieved, the deployment method is optimized, online adjustment deployment can be performed according to data changes, and the reasoning speed is increased compared with other deployment technologies. The method can be widely applied to the technical field of distributed machine learning.

The invention discloses a machine tool thermal error compensation system based on digital twinning and cloud edge collaboration, which belongs to the technical field of precision manufacturing and comprises a physical error control layer, a digital twinning data layer, a virtual error control layer, an edge server and a cloud server. And performing real-time prediction on the acquired temperature and thermal error data through a long-sequence space-time fusion parallel network model deployed in an edge server, and issuing a compensation instruction to the CNC controller. And the cloud server is responsible for training and updating the model. According to the method, high-precision modeling and low-delay real-time compensation of the thermal error are realized through the cloud edge cooperation and digital twinning technology, the thermal error of the main shaft is reduced by 80% at most, the response delay of the system is reduced by about 40%, and the machining precision and stability of a machine tool are remarkably improved.

The invention discloses a cloud side-end collaborative architecture and task unloading method for an airport apron intelligent monitoring system, and relates to the technical field of cloud side-end collaborative computing. Comprising a terminal layer, an edge server layer and a cloud server layer, the task unloading method of the system under the cloud side-end collaborative architecture is designed and comprises the steps that a task model, a time delay model, an energy consumption model and an accuracy rate model of the system are constructed, and a multi-objective optimization problem of time delay-energy consumption-accuracy rate is formed; modeling an optimization problem into a Markov decision process, and designing a state space, an action space and a reward function required by deep reinforcement learning; and designing a task unloading method based on multi-agent deep reinforcement learning, and finding an optimal task unloading strategy of the system. According to the method, the computing tasks can be dynamically, scientifically and reasonably distributed and cooperatively scheduled among the terminal, the edge and the cloud, so that the time delay, the energy consumption and the accuracy are comprehensively optimized, and the overall performance of the system is improved.

The invention discloses a collaborative secret state task matching method based on edge calculation in mobile crowdsourcing. The method comprises the steps of system initialization and key generation, requester-cloud platform registration authentication, worker-edge server registration authentication, login authentication and attribute submission, requester task issuing, cloud platform task delegation, secret state matching, worker task content decryption, worker answer submission, answer forwarding and decryption. According to the invention, cooperative authentication of workers and requesters and precise task matching based on attributes are realized under an edge-cloud architecture. A fuzzy extractor and an authentication encryption method with associated data are introduced, so that low-cost user local authentication and registration and authentication of a joining system are realized, the security is improved, and the calculation cost is reduced. Through function hidden inner product encryption and a Paillier cryptographic algorithm, task-worker matching and task answer submission are realized in a ciphertext state, and the requirements of a mobile crowdsourcing application scene with high safety, high privacy and dispersed cost are met.

The invention discloses an industrial Internet of Things federated learning method based on a federated increment decision tree, and the method comprises the steps: a cloud server deploys and initializes a federated learning global model and the federated increment decision tree, and sets a statistical histogram bucket boundary set of each feature value; in each federated learning iteration, the cloud server broadcasts a federated learning global model parameter, each industrial device adopts a local data set to train and count to obtain a local gradient histogram parameter, the edge server performs local aggregation on the local gradient histogram parameter and the federated learning model parameter, and the edge server performs local aggregation on the local gradient histogram parameter and the federated learning global model parameter; and the cloud server globally aggregates the local aggregation parameters of the gradient histogram and then incrementally trains the federated increment decision tree, simultaneously aggregates the parameters of a federated learning global model, and adaptively calculates an aggregation weight based on a second-order gradient value during local aggregation and global aggregation of the parameters of the federated learning model. The federal learning overhead can be effectively reduced, and the performance of the federal learning model is improved.

The invention discloses an edge computing network task unloading and resource allocation method based on optimal service quality, which comprises the following steps of: constructing a system architecture comprising a cloud server, a plurality of edge servers and mobile equipment, and establishing a multi-dimensional system model covering task characteristics, service cache, communication transmission, computing resources, service cost and service quality; constructing an optimization problem aiming at maximizing the long-term service quality of all mobile equipment, and forming a mixed integer nonlinear programming model under the constraints of mobile equipment cost constraint, storage capacity limitation, bandwidth, computing resources and the like; a double-time-slot hierarchical decision-making mechanism is designed, and multi-dimensional joint optimization of service caching, task unloading and resource allocation is realized through a collaborative mechanism that short-term resource allocation is constrained through a long-term caching decision and short-term performance feedback optimizes long-term caching. According to the method, the overall quality of service (QoS) can be effectively improved and the task processing delay and energy consumption can be reduced under the condition that the cost of the mobile equipment and the system resource limitation are met.

The invention provides a mode selection scheme oriented to edge reasoning energy consumption optimization. Firstly, an equipment independent reasoning mode, an edge server-equipment collaborative reasoning mode and a multi-outlet edge collaborative reasoning mode are defined, and all information required by mode selection is acquired through an edge server equipped with an access point. On the basis, dividing a problem of mode selection for optimal energy consumption of edge reasoning into three sub-problems, and respectively calculating the minimum energy consumption of independent reasoning of equipment, the minimum energy consumption of edge server-equipment collaborative reasoning and the minimum energy consumption of edge collaborative reasoning based on multiple outlets; and energy consumption calculation and decision are completed on the edge server which obtains the information. According to the method, the optimal model segmentation strategy is determined through the enumeration-convex optimization joint algorithm, and dynamic mode selection based on the real-time environment state is achieved. According to the method, the optimal mode can be selected from the edge server under the condition of meeting the requirements of precision, time delay and computing power.

The embodiment of the invention provides a multi-level federal multi-mode large model privacy protection enhancement method and electronic equipment, and belongs to the technical field of intelligent traffic systems. According to the method, a'end-edge-cloud 'three-level federated learning architecture is constructed, firstly, local differential privacy disturbance is applied to local multi-mode traffic data at a traffic terminal node, and an end-side local model is trained; then aggregating a plurality of end side models on an edge server, and generating a personalized small model reflecting regional characteristics; then cooperatively training a plurality of personalized small models in a cloud center server to generate a global multi-modal large model; finally, knowledge of the global large model is fed back to a lower-level model through knowledge migration, and a continuously evolved iterative closed loop is formed. According to the method, privacy protection is carried out at a data source, so that original sensitive data is ensured not to go out of the local, the problems of data islands and privacy leakage in traffic large model collaborative training are effectively solved, and efficient and credible collaborative modeling is realized on the premise of ensuring data sovereignty.

The invention discloses a BDDR backdoor detection and data recovery method and system oriented to a large model, and belongs to the field of backdoor defense. Comprising the following steps: constructing a knowledge distillation architecture under federal learning, including an edge server and a plurality of clients, and obtaining distillation data; inputting the distillation data into a randomly initialized model for training, recording the loss change of each batch of data, and screening out abnormal batches to form a backdoor data set; the edge server initializes two independent models, respectively uses a distillation data set and a backdoor data set for training, and guides learning of backdoor features; using probability distribution to calculate and correct a backdoor label, generating a clean data set by adding noise, finely adjusting a large model, detecting residual backdoor feature intensity, and adjusting probability distribution calculation parameters to further weaken backdoor features according to the residual backdoor feature intensity so as to obtain a final repaired data set; while the generalization ability of the large model is improved, backdoor attacks can be effectively identified and defended, the data privacy of the client is protected, and the model security is ensured.

The invention provides a cloud edge collaborative diffusion model reasoning method and system based on a block chain and reinforcement learning, and the method comprises the steps: obtaining a current text prompt word submitted by a user in a block chain network composed of a cloud server and an edge server, and calling a semantic matching model through an intelligent contract, and judging whether a historical intermediate result can be reused or not; obtaining a server environment state, generating a collaborative reasoning strategy by using a pre-trained multi-agent attention actor-commentator model, and determining cloud edge denoising step number distribution; if the image cannot be reused, the cloud server executes partial denoising to generate an intermediate result and stores the intermediate result to the block chain, and the edge server continues to complete residual denoising to generate a final image; and if the edge server can be reused, the edge server directly completes denoising based on the historical intermediate result. According to the method, historical intermediate results can be effectively reused, dynamic intelligent task allocation is realized, the diffusion model reasoning efficiency and the image generation quality are improved, and the credibility and the traceability of a distributed reasoning process are guaranteed.

According to one embodiment, an edge device includes a communication interface connectable to an edge server, a storage unit for storing a dialogue LLM trained to interact with a user for setting a function of the edge device and a database storing reference information related to functions of the edge device. A controller unit receives request text input indicating a user's desired process to be provided using the edge device, then inputs the request text input to a generative AI that is based on the dialogue LLM and can access the database to identify a function of the edge device that can provide the user's desired process. The controller unit outputs confirmation question text to confirm the identified function corresponds to the user's desired process, and then sets the edge device to execute the identified function if confirmed.

The invention discloses an adaptive joint segmentation federated learning method based on hierarchical reinforcement learning in combination with a near-end strategy optimization algorithm. The method comprises the following steps: firstly, establishing a cloud-edge-end cooperative computing architecture consisting of a vehicle terminal, an edge server and a cloud server; secondly, providing an HSHPPO algorithm based on hierarchical reinforcement learning and near-end strategy optimization for model segmentation and resource optimization problems in an intelligent Internet of Vehicles scene; and finally, based on an HSHPPO algorithm, joint solution is performed on model cutting layer selection and computing resource allocation, so that a joint segmentation federated learning strategy capable of adaptively adjusting a segmentation position and minimizing time delay and energy consumption is obtained, and efficient training and privacy protection are realized.

A method and apparatus for dynamic orchestration of distributed artificial intelligence in a network including a user device, an edge server, and a cloud server. The method includes receiving, at the user device, input data comprising at least one of audio, video, image, or text; identifying a requested operation based on the input data; obtaining dynamic environmental information of the network relating to computing resources and network conditions of the user device and at least one of the edge server or the cloud server; determining, based on the requested operation and the dynamic environmental information of the network, a distributed allocation of the requested operation among the user device, the edge server, and the cloud server; and orchestrating the requested operation according to the distributed allocation.

The invention discloses a real-time cooperative task scheduling method and device for a vehicle-mounted operating system, and belongs to the field of intelligent vehicle-mounted systems and edge computing. The method comprises the following steps: collecting vehicle tracks and terminal and server state data in a vehicle-edge-cloud cooperative system in real time; using a PatchTST model to predict the future position of the vehicle based on the historical trajectory; constructing a collaborative scheduling model taking the vehicle and the edge server as intelligent agents; based on local observation, the vehicle agent generates an unloading decision, and the edge server agent generates a task migration decision; a multi-agent near-end strategy optimization algorithm for partial reward decoupling is adopted, and a strategy is optimized under a centralized training and distributed execution framework; task unloading, migration and dynamic result return are executed according to the decision; and calculating rewards based on the task completion condition and the system state, and continuously updating model parameters. According to the method, the task completion rate of vehicle-mounted task scheduling in a dynamic complex environment is effectively improved, and the average time delay is reduced.

The invention discloses an alliance chain-based unmanned aerial vehicle cross-domain service function link accessing method, which comprises the following steps of: 1, initializing a system and configuring an alliance chain, generating a global password parameter and a root key by a trusted mechanism, and finishing domain registration and certificate chain storage by each management domain edge server; 2, blockchain-driven identity management is carried out, and the unmanned aerial vehicle completes chain registration and anti-counterfeiting registration certificate acquisition through a domain edge server to which the unmanned aerial vehicle belongs; 3, deploying a flexible threshold signature algorithm to realize multi-domain joint signature and Byzantine fault tolerance; 4, executing a cross-domain SFC security authentication protocol, including SFC pre-verification and security authorization certificate issuing, first node verification starting, hop-by-hop key negotiation and handover certificate transmission, and on-chain auditing; and 5, based on the topology centrality and the path coverage, dynamically electing an orchestrator to realize load balancing. According to the invention, safe access and identity authentication of the cross-domain service function chain of the unmanned aerial vehicle are realized, safety, efficiency and expandability are balanced, and reliable guarantee is provided for cross-domain cooperation of the unmanned aerial vehicle.

The invention relates to the technical field of industrial internet of things and mobile edge computing crossing, in particular to a computing task unloading and service caching collaborative optimization method, which comprises the following steps of: acquiring an IIOT equipment state, channel quality and edge server resources in real time by constructing a dynamic network awareness architecture, and analyzing a task DAG topological structure; designing a service cache delay compensation mechanism, parallelizing a cache process and task execution, predicting future required services based on a task dependency relationship, and calculating a delay compensation factor; a deep reinforcement learning algorithm based on Transform is adopted, and an unloading decision is generated in combination with a hierarchical attention mechanism and an improved DDQN network; tasks are dynamically allocated to the local, the edge or the cloud through a three-level unloading decision tree, task completion time is minimized, a resource constraint modeling and task preemption mechanism is further included, and the system response capability and the resource utilization rate are improved.

The invention relates to the technical field of edge computing resource management, in particular to an edge server load evaluation method based on analytic hierarchy process and performance loss. According to the method, firstly, edge servers are divided into a common type, a computing type, a memory type and an I / O type according to heterogeneous types; aiming at each type of edge server, constructing a judgment matrix by adopting an analytic hierarchy process and solving a feature vector so as to calculate load weight coefficients of CPU (Central Processing Unit), memory and disk I / O (Input / Output) resources; calculating a real-time load based on the weight coefficient, wherein the real-time load is the weighted sum of the resource utilization rates; the performance loss is further calculated according to the nonlinear relation between the real-time load and the total load ratio; and finally, restraining the real-time load through the load threshold value, and evaluating the performance state of the server in combination with the performance loss. According to the method, the accuracy and reliability of load evaluation of the heterogeneous edge server are effectively improved, the resource utilization rate is optimized, overload of the server is avoided, and the system stability is enhanced.

The invention provides an Internet of Vehicles cross-domain identity authentication method based on PUF and certificateless, and the method comprises the steps: firstly generating a system master key and a master public key through a KGC, and disclosing system parameters; secondly, the vehicle generates a physical response and extracts a stable key, interacts with the KGC to complete generation of a certificateless key pair, and stores registration information to a block chain; then the edge server registers in the KGC, obtains a certificateless key pair, initiates an authentication request to the KGC, and obtains a block chain data reading authority; finally, bidirectional authentication is carried out between the vehicle and the edge server based on the certificateless signature and the PUF response, and a session key is generated; and the cross-domain vehicles are assisted by the edge server to complete mutual identity authentication and session key negotiation. According to the method, the key escrow and certificate management burden is eliminated by adopting a certificateless cryptosystem, the high computing load is released to the edge server in combination with the PUF hardware security characteristic and the block chain distributed trust, the computing and communication overhead is remarkably reduced, and the method is suitable for a large-scale Internet of Vehicles cross-domain authentication scene.

The invention discloses a mobile social network resource allocation method based on cloud side-end cooperation, and the method comprises the steps: screening out a malicious ES through building a trust evaluation mechanism, guaranteeing the safety of a task unloading process, improving the data transmission efficiency between an MU and the ES through semantic transmission, constructing a two-layer task unloading frame and an incentive mechanism, and achieving the resource allocation of the MU and the ES. And optimizing an unloading strategy of the MU and the ES and a pricing strategy between the ES and the cloud. In the aspect of system composition, a mobile user, an edge server and a cloud server form a cloud edge-end cooperative mobile social network, and a trust evaluation machine comprehensively considers direct trust degree, indirect trust degree, intimacy degree, contribution history and the like, so that trust evaluation based on historical interaction records of the mobile user and the edge server is realized; for task unloading and excitation, an excitation-oriented two-layer task unloading scheme is designed, a Stackelberg game model is constructed to optimize strategies of all parties, and the problems of limited computing resources, high data transmission delay, low communication efficiency, security threat and unreasonable resource allocation of mobile equipment are solved.

The invention relates to the technical field of supply chain monitoring, and discloses an intelligent monitoring and early warning method, device and equipment for a supply chain and a storage medium, and the method comprises the steps: deploying a plurality of types of Internet of Things sensors at each target node in the supply chain, collecting the state data of the supply chain, and executing time sequence analysis and anomaly detection in an edge server, obtaining local early warning information; performing graph neural network analysis based on the local early warning information and a preset supply chain network structure to obtain a node risk representation vector and a risk propagation probability matrix; according to the node risk representation vector and the risk propagation probability matrix, early warning strategy calculation is carried out through a grouped multi-agent deep reinforcement learning algorithm, and an early warning strategy set is obtained; and performing risk propagation simulation calculation on the early warning strategy set and the supply chain network structure to obtain comprehensive early warning decision information. According to the invention, the risk propagation path and rate can be accurately identified, and the accuracy of supply chain risk propagation prediction is improved.

The invention discloses a privacy protection and robustness test method and system for large model fine tuning, and belongs to the technical field of machine learning security. The method comprises the steps that a three-layer distributed architecture comprising an edge server, a cloud server and a plurality of edge clients is constructed, the edge clients distill local privacy data and cooperate with the edge server to train a global model, and a candidate detection sample set is formed; screening a sample set based on the potential feature deviation evaluation index, and sending the sample set to a cloud server for vulnerability detection to obtain an optimal backdoor detection candidate sample set; and multi-trigger parallel and progressive trigger sequence backdoor implantation is respectively used for scenes of single fine tuning and multiple fine tuning of the large model, an optimal backdoor detection candidate sample set is combined with a preset trigger to generate a backdoor test sample set, the backdoor test sample set is mixed with a clean data set, and then the robustness of the backdoor test sample set is tested through fine tuning of the large model. Large model fine tuning and robustness testing of privacy protection can be realized in a heterogeneous model cooperative training environment.

The invention discloses a highway network traffic flow detection and congestion situation dynamic modeling system, and relates to the technical field of highway traffic management, the highway network traffic flow detection and congestion situation dynamic modeling system comprises a sensing layer, a transmission layer, a first processing layer, a second processing layer and an application layer, the sensing layer deploys multi-modal detection equipment to realize traffic flow data acquisition of highway key nodes; the transmission layer is provided with an edge server cluster for realizing real-time transmission of data acquired by the sensing layer; the first processing layer is used for realizing an RGBT multi-mode traffic flow detection subsystem, the second processing layer is used for realizing GNN congestion situation dynamic modeling, and the application layer provides visual display and decision support for a traffic management department. The RGBT detection technology and GNN modeling need to be deeply fused, a perception-modeling-analysis-early warning closed-loop system is constructed, the requirements of a traffic management department for accurate management and control and efficient scheduling are met, and the practical requirements of current highway management pain points are met.

The invention relates to the field of intelligent scheduling allocation, in particular to an edge cloud computing resource allocation optimization method based on deep learning, which adopts a space-time prediction algorithm based on multi-head attention and gating mechanism optimization to design time coding and space coding. The spatial relationship and interaction between time sequence characteristics of the computing power load and edge server nodes are captured, and meanwhile, a multi-head attention mechanism and expansion causal convolution are combined, so that instantaneous computing power load fluctuation can be captured, and the long-term trend of the computing power load can be mined; therefore, a reliable basis is provided for subsequent computing power scheduling by predicting an accurate computing power load. The invention designs an alternating direction multiplier method based on genetic algorithm optimization, which is not only suitable for a nonlinear and multi-constraint optimization problem, but also can be expanded to a larger-scale distributed edge node cloud computing system, and meanwhile, a global optimal solution is quickly approached through the genetic algorithm, so that the quality of an initial solution is improved, and model convergence is accelerated; and the distributed collaborative allocation scheduling efficiency is improved.

The invention relates to a federated learning-based privacy protection personalized recommendation system, which comprises a user side, an edge side and a cloud side, and is characterized in that the user side is a basic unit for data generation and local training, each user maintains a personalized recommendation model on own local equipment, the edge side is used as an intermediate aggregation layer, and the cloud side is used as a cloud side; the cloud end is used for preliminarily aggregating local models of users in a certain geographic area or logic area, the edge end is generally deployed in an edge server or an area data center, and the cloud end is a central node for updating and distributing a global model and is used for secondarily aggregating area aggregation models uploaded by the edge ends to form a global recommendation model; according to the scheme, user privacy is strictly protected, and collaborative recommendation is realized under the condition of not transmitting original user data through a differential privacy and security aggregation dual protection mechanism.

The invention discloses a heterogeneous cellular network task unloading method oriented to location privacy protection, and belongs to the technical field of communication. The invention provides a heterogeneous cellular network task unloading method oriented to location privacy protection, aiming at solving the problem that location privacy leakage is possibly caused when a mobile user unloads a task to a micro base station edge server to which the mobile user belongs in heterogeneous cellular network edge calculation. According to the method, the privacy leakage risk of task unloading is evaluated according to user position distribution, the number of unit tasks to be unloaded and the resource state of each edge server, the position correlation between users and target servers is weakened by using a relay forwarding mechanism of a macro base station, and a deep reinforcement learning method is adopted, so that the task unloading efficiency is improved. And the position privacy risk and the task unloading strategy of the user are dynamically learned and optimized, so that the position privacy protection of the user is effectively realized and the average unit task energy consumption is minimized under the constraint of average unit task tolerance time delay.

The application discloses a kind of server fault monitoring systems related to server technical field, and the data acquisition terminal of deployment in computer room gathers the environmental monitoring data and software monitoring data of server, and as server monitoring data is sent to corresponding edge server.Edge server detects that software / hardware monitoring data exists exception, then from corresponding server monitoring data extraction relevant exception data, and generate exception report log.Central server determines that server exists fault according to exception report log and relevant exception data, then generate fault data information.Monitoring terminal locates the graphic object of fault server corresponding according to equipment identification in three-dimensional computer room virtual model, and the display state of server graphic object is updated to fault state, simultaneously generate fault data interaction label according to fault data information.The application can solve the problem that related technical fault response delay is larger, and can effectively improve server fault monitoring efficiency.

The present application relates to the technical field of intelligent Internet of Vehicles, and in particular to a deep reinforcement learning-based task unloading method in a vehicular edge computing environment, comprising: acquiring tasks to be executed generated by a task vehicle, uploading the tasks to an RSU, generating a scheduling decision on the basis of the current states of the tasks and available resources, and allocating the tasks to service vehicles or executing the tasks locally; prioritizing the tasks by means of the analytic hierarchy process, and acquiring the current state of each of the tasks to be executed; on the basis of the states and the priorities of the tasks, using a model for scheduling to acquire service vehicle numbers and a task vehicle number for the tasks, and a computing unloading and scheduling policy; acquiring available computing resources of the current service vehicle; and, by means of the Actor-Critic algorithm, training a sequence-to-sequence model, so as to obtain an optimal task partial unloading and scheduling policy. The present application can fully use computing resources of service vehicles and edge servers, so as to allow for shorter execution delays of all tasks throughout the entire time period and higher task execution success rates.

Disclosed in the present invention is a communication and computing resource joint allocation optimization method for multiple agents, capable of constructing a communication channel model, a computing model and a motion energy consumption model on the basis of a current industrial workshop environment, an edge server and agent parameters. Each agent interacts with the environment to obtain a current observation value, inputs the obtained observation value into an action network, obtains an action parameter corresponding to a next moment, calculates a current reward, updates neural network parameters, and trains current data, so as to obtain a communication and computing resource joint allocation strategy. The present invention can optimize the allocation method when the communication and computing resources are limited, so as to satisfy the requirements of low delay and high reliability in communication, and timely computing power, thereby reducing resource waste.

System and method for an over-the-air computation (AirComp) scheme for federated edge learning (FEEL) doesn't require channel state information (CSI) at the edge devices (EDs) or edge server(ES). The disclosure adopts the majority vote (MV) principle and defines multiple subcarriers and orthogonal frequency division multiplexing (OFDM) symbols for voting options, which reduces to frequency-shift keying (FSK) over OFDM subcarriers as a special case. Thus, FSK-based over-the-air computation is provided for federated edge learning without channel state information. Since the votes from EDs are separated on orthogonal resources, the scheme eliminates the need for truncated-channel inversion (TCI) at the EDs and allows the ES to detect MV with a non-coherent detector.

The invention relates to the technical field of industrial Internet of Things and edge computing, and discloses an edge-end cooperative computing task unloading method and system for an intelligent factory, and the method comprises the steps: building a task unloading decision mechanism of a multi-device multi-edge server, distributing a computing node for each task, determining task delay, and constructing a node speciality matrix; establishing an optimization model; formalizing an optimization problem of the optimization model into a Markov decision process; solving a Markov decision process by using a deep reinforcement learning algorithm, and realizing intelligent workshop environment interaction so as to store generated empirical data into an empirical cache region; synchronously updating network parameters of the deep reinforcement learning algorithm and a node speciality matrix corresponding to the edge server by using the data pairs of the experience cache region; and outputting a strategy network and a node speciality matrix until the deep reinforcement learning algorithm converges. And an optimal task allocation decision is dynamically generated, so that the task completion deviation is effectively reduced.