63 results about "Balancing network" patented technology

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

The invention discloses a self-adaptive dynamic routing method and system for an NoC (Network on Chip), and relates to the technical field of routing management. Comprising the steps that the local congestion degree of each routing node is collected through a communication link, the global congestion degree is generated through a weighted average mechanism, each routing node integrates delay counters in the east direction, the west direction, the south direction and the north direction according to a local router, the delay counters are used for monitoring the time interval when adjacent data packets leave a buffer area in real time, and the global congestion degree is generated; taking a time interval of leaving a buffer area of adjacent data packets as output delay, spreading the output delay to neighbor routing nodes, calculating local congestion degree and global congestion degree of the routing nodes, generating all possible shortest paths and a candidate direction set corresponding to output directions according to addresses of a source routing node and a destination routing node, and sending the candidate direction set to the destination routing node; screening an optimal path through the local congestion degree and the global congestion degree in sequence; and when a plurality of equivalent paths exist, the routing direction is randomly selected to balance the network load.

The invention discloses a dynamic topology ad hoc network method based on environmental perception and adaptive optimization, and relates to the field of wireless sensor networks. Parameters (such as temperature, obstacle distribution and node energy consumption) of nodes are collected in real time; a multi-dimensional quantization parameter matrix including an obstacle coverage matrix, an energy consumption distribution matrix and a signal attenuation coefficient matrix is constructed, and path loss, communication delay and node energy consumption are fused and calculated to generate a node weight value by adopting an adaptive weight calculation mechanism, so that the priority of a communication link is dynamically adjusted. Compared with the prior art, the method has the advantages that the self-healing capability and the real-time performance of the network in a complex environment are remarkably improved, the global calculation overhead is reduced, the real-time performance and the expansibility of the system are improved, the node layout and the dynamic optimization of a communication path are realized, and the network load and the stability are balanced.

The invention provides a network attack traffic category balancing method and system based on an improved ACGAN, and the method comprises the steps: carrying out the preprocessing of a network traffic data set, and obtaining a preprocessed network traffic data set; wherein the category of the network traffic data set comprises normal traffic data and attack traffic data; carrying out preliminary enhancement on the preprocessed attack traffic data by using an oversampling technology SMOTE to obtain a preliminary balanced data set; and enhancing the attack traffic data in the preliminary balanced data set by using a generative adversarial network ACGAN to obtain a final balanced data set, and completing balancing of network attack traffic categories. According to the technical scheme, the problem that extreme categories of data samples are unbalanced in the prior art is solved.

The embodiment of the invention relates to an analog simulation method based on electric power spot transaction. The analog simulation method comprises the steps of obtaining static structure data such as power grid topology and unit parameters and dynamic time sequence data such as load and new energy prediction; performing preprocessing and multi-time scale division on the dynamic data; constructing a multi-agent model with differentiated characteristics based on the processed data; each agent generates and reports energy block information such as a time-sharing block, a continuous block, a curve block or a variable block according to a perceived market environment; constructing a market clearing model considering system power balance, network power flow, unit operation and energy block characteristic constraint based on all declarations, and solving to obtain clearing electricity price and electric quantity; a result is fed back to the intelligent agent, and the strategy is updated through deep reinforcement learning; and performing rolling co-simulation between the day-ahead market and the real-time market, wherein a day-ahead clearing result is used as a boundary condition of the real-time market. The authenticity, accuracy and practicability of market simulation are improved.

The invention provides a hybrid routing method and device based on congestion awareness, and the method comprises the steps: determining a source node and a target node of a data packet, the source node representing a starting router node which initiates data transmission, the target node representing a termination router node which finally transmits data, and the destination node representing a termination router node which finally transmits the data; acquiring an input buffer occupation state of a target adjacent node corresponding to the source node, if the input buffer occupation state of the target adjacent node is an idle state, transmitting the data packet through a target shortest path corresponding to the target adjacent node, and if the input buffer occupation state of the target adjacent node is a congestion state, transmitting the data packet through a target shortest path corresponding to the target adjacent node. If yes, other nodes of the source node in the second target direction are determined, other shortest paths corresponding to the other nodes are obtained, and the data packet is transmitted to the target node through the other shortest paths. After the output port with the lightest load is selected for the data packet to bypass the congestion area near the source node, the algorithm is simplified while the network load is balanced, and overhigh routing overhead is avoided.

The application relates to the technical field of high-frequency inverter output current sharing, and discloses a high-frequency inverter parallel current sharing method and system. According to the application, for multiple parallel high-frequency inverters, the output currents of the high-frequency inverters flow into a parallel confluence node and subsequent circuits after being balanced by a preset impedance balancing network. The impedance balancing network is modular, and comprises multiple modular current sharing branches. One end of each current sharing branch is connected to the output end of each high-frequency inverter in one-to-one correspondence, and the other end of each current sharing branch is connected to be suspended. The parameter design of the impedance balancing network is determined based on the connection between the output end of each high-frequency inverter and the parallel confluence node and the subsequent circuits. The balancing method has the characteristics of few devices, wide frequency range and strong extensibility of multi-inverter current balancing.

The invention provides a radio frequency switch circuit, which is characterized in that a voltage-sharing network is connected to a body region of a switch transistor, the voltage-sharing network comprises a first branch and a second branch, and the first end of the first branch and the first end of the second branch are both connected to the body region of the switch transistor; when the radio frequency switch circuit is in a turn-off state, the first branch circuit and the second branch circuit are conducted, the second end of the first branch circuit is connected to a source-drain connection node or a source electrode of another switch transistor, and the second end of the second branch circuit is connected to a source-drain connection node or a drain electrode of another switch transistor. The first branch leads the leakage current on the body region of the switch transistor to the source-drain connection node or the source electrode of other switch transistors, and the second branch leads the leakage current on the body region of the switch transistor to the source-drain connection node or the drain electrode of other switch transistors. The phenomenon of uneven voltage division of the switch transistor caused by the fact that the voltage of the body region of the switch transistor is raised is improved.

The application discloses a kind of anonymous routing methods of double-layer satellite network for delay-sensitive service, it is applied to VLEO / LEO double-layer satellite network, this network is based on software-defined network SDN architecture implementation, including data plane, control plane and application plane;Data plane includes VLEO layer, VLEO / LEO cross layer and LEO layer;Application plane is deployed with the routing method of the application upper;The method comprises: constructing layered decoupling routing calculation framework, to divide routing decision into including LEO layer inner routing, VLEO / LEO cross layer routing and VLEO layer inner routing multilayer path selection;Using the collaborative computing capability of SDN controller and satellite node, multilayer path selection is reconfigured and weight disturbance in real time, generates optimal routing path.The application can effectively reduce end-to-end delay, balance network load, improve the real-time performance, robustness and security anonymity of double-layer satellite network.

The application relates to a kind of digital power energy storage system dynamic network reconfiguration methods in the technical field of power energy storage, S1. Build energy storage system;It is composed of intelligent layer, control layer and data layer;S2. Energy storage network modeling;Adopt graph capsule neural network to learn the optimal control strategy of energy storage network;S3. Dynamic network reconfiguration;Based on Markov decision process, generate fault emergency response model;By systematic optimization of the topological structure and optimal control strategy of energy storage network, the quick isolation of fault area is realized, and the dynamic adjustment and reconfiguration of the working mode of energy storage equipment and load equipment are realized, to ensure that the energy storage network can still operate stably under fault situation, improve the flexibility and elasticity (fault tolerance) of energy storage system in dynamic environment;By reducing the energy loss in the transmission process, balancing network load and preventing local failure, the technical problems existing in the prior art are solved, and the sustainable development of new renewable energy system is promoted.

The application discloses a resource allocation method of a hybrid space division multiplexing elastic optical network and belongs to the technical field of optical communication. The application constructs a hybrid network model containing a multi-fiber bundle link and a multi-core optical fiber link, and designs a load and crosstalk aware heuristic algorithm. By means of modulation format step degradation, spectrum window plane construction to implicitly satisfy spectrum continuity constraints, load aware routing to balance network load, crosstalk aware core selection to suppress inter-core crosstalk, and optical signal-to-noise ratio verification to ensure transmission quality, joint optimization of routing, spectrum and spatial channels is realized. The application solves the problem that the prior art is only directed to a single type of space division multiplexing network and is difficult to adapt to a hybrid scene, effectively reduces network blocking rate under the premise of ensuring transmission quality, takes into account crosstalk suppression and spectrum efficiency, and has good engineering practicability.

The invention provides a dialogue reply method and device, equipment, a medium and a program product, and the method comprises the steps: obtaining dialogue data, inputting the dialogue data into a trained large language model based on a deep learning network, and obtaining a dialogue reply feature, the deep learning network comprises at least one first module, each first module comprises a second module with at least one first basic unit and a second basic unit which are connected in sequence, or the second module with at least one second basic unit and the first basic unit which are connected in sequence, and the first basic unit and the second basic unit comprise a normalization layer and a sub-layer; the output of the normalization layer is used as the input of the sub-layer, the input of the normalization layer in the first basic unit and the output of the sub-layer are subjected to residual connection, the output of the normalization layer in the second basic unit and the output of the sub-layer are subjected to residual connection, and reply is carried out according to dialogue reply features. The stability and model performance of the network are improved by balancing the internal structure of the network.

Disclosed is a high-power resonant inverter system (100) for wireless power transfer. The system (100) includes a control block (102) that generates control input signals. A synchronous gate driving arrangement (104), isolated from the main power circuit, operates a series-parallel configuration of plurality of switching elements (106). The series connection of switching elements increases voltage handling capability, while the parallel connection increases current handling capability. Furthermore, a voltage balancing network (108) evenly distributes voltage stress across the plurality of switching elements (106). Additionally, a resonant tank circuit (110) is coupled to the voltage balancing network (108) for efficient wireless power transfer.

The invention discloses a collaborative environment-oriented service deployment and request scheduling optimization method and device, and relates to the technical field of edge-cloud collaborative mobile edge computing in the power industry, and the method comprises the steps: constructing a multi-instance queuing network model and a multi-dimensional hierarchical soft actor-commentator algorithm; based on the multi-instance queuing network model, modeling a multi-instance execution relationship in a power station task processing process, and providing a performance evaluation basis for a request scheduling strategy by quantifying task transmission delay, queuing delay and calculation delay; and based on the multi-dimensional hierarchical soft actor-commentator algorithm and in combination with the current state of the power information system, determining the deployment number and distribution position of each micro-service instance, and formulating a request scheduling strategy. According to the method, the dependency relationship and the multiplexing characteristic between the micro-services are fully considered, the request response delay can be effectively reduced, the network load and the system energy consumption can be balanced, and the service quality and the resource utilization efficiency of the electric power station are improved.

The invention discloses an energy storage grading charging and discharging regulation and control method for stabilizing photovoltaic power fluctuation, and relates to the technical field of photovoltaic power generation and energy storage cooperative control, and the method is applied to a photovoltaic power station system comprising a group cascade self-balancing networking unit, a station-level central networking energy storage system and a station-level cooperative networking controller. Power deviation caused by local shadow is monitored and compensated in real time through a group cascade unit, power cooperative scheduling and reactive power dynamic distribution are carried out on a station level layer, cooperative optimization of photovoltaic array internal power balance and grid-connected point power grid support is achieved, and meanwhile, through virtual synchronous machine cluster control and a protective off-grid / re-grid-connected mechanism, the photovoltaic array internal power balance and the grid-connected point power grid support are optimized. According to the method, power fluctuation can be effectively stabilized in a complex shielding environment, the power generation efficiency of the photovoltaic power station and the interaction capability of a power grid are improved, and the stability and reliability of a power system are enhanced.

The present invention provides a routing selection method suitable for large-scale LEO constellations, including route preselection: based on the predictability of satellite operating positions, each satellite preselects a sending port when transmitting a data packet; and path convergence: using a queuing factor (SVF) and a power factor (PVF) to converge the routing path. This method reduces routing storage and overhead, improves constellation throughput and reduces end-to-end latency as the probability of interruption increases. The queuing factor reduces queuing latency and balances network load, while the power factor reduces further consumption of satellites with relatively low power consumption, extending the constellation's operating time. The proposed path convergence strategy based on SVF and PVF, in addition to using the queuing factor and power factor, can also select factors such as end-to-end latency, number of routing hops, and link failure rate based on service QoS requirements.

This invention relates to a method and apparatus for collaborative trajectory planning and energy efficiency optimization of UAV networks for multi-target surveillance. The method achieves deep integration of routing pairing, trajectory planning and power control by constructing a multi-target collaborative optimization framework, designing a hierarchical linkage algorithm, and enhancing scenario adaptability and dynamic response capabilities. This solves the problem of balancing network performance, lifetime and energy efficiency in multi-target surveillance scenarios and promotes the large-scale application of UAV swarms in complex surveillance tasks.

Embodiments of the present application relate to a kind of simulation method based on power spot trading, comprising: obtaining grid topology, unit parameter and other static structure data and dynamic time series data such as load, new energy prediction;Dynamic data is preprocessed and divided into multiple time scales;Based on the data processed, construct a multi-agent model with different characteristics;Each agent generates and reports energy block information such as time block, continuous block, curve block or variable block according to the perceived market environment;Based on all reports, construct a market clearing model considering system power balance, network power flow, unit operation and energy block characteristic constraints and solve, obtain clearing price and power;The results are fed back to the agent, and the strategy is updated by deep reinforcement learning;Rolling collaborative simulation is carried out between day-ahead and real-time market, wherein the day-ahead clearing result is used as the boundary condition of real-time market.The present application improves the authenticity, accuracy and practicality of market simulation.

The application discloses a link bandwidth allocation method and allocation system for multi-mode mobile network multi-link aggregation, comprising: obtaining mobile network base station, device position, real-time environment data; using intelligent power control algorithm, dynamically adjusting transmission power according to environment data and network load; obtaining real-time signal strength data and historical data of network device to base station; predicting minimum transmission point of signal transmission based on real-time and historical data; adjusting antenna direction and position according to base station position, and designing connection using beamforming antenna; setting threshold value according to real-time weather and minimum transmission point data, and deploying signal enhancer; the application discloses a link bandwidth allocation method and allocation system for multi-mode mobile network multi-link aggregation, through intelligent algorithm, real-time monitoring of link quality, dynamic adjustment of network, balancing of network load, and combination of real-time weather data for evaluating signal transmission environment, deploying signal enhancer, and improving signal coverage range.

The present disclosure generally relates to optimizing load-balancing of network endpoints using tree collectives representing a logical network communication topology for the network endpoints. Systems and methods described herein eliminate the previously restrictive conditions imposed on tree-based communication collectives by generating collective trees with any arity and representing any number of physical network endpoints. The resulting collective trees ensure that each represented network endpoint has a number of outgoing flows and a number of incoming flows that are no more than the arity of the collective tree. In this way, the described systems and methods inject significant efficiencies into communication collectives within networked compute nodes by eliminating communication bandwidth latencies and bottlenecks.

The embodiment of the invention provides a system for balancing network transmission bandwidth, and the system comprises a state detection module which is configured to monitor a network transmission index and clock synchronization information; generating a network quality score based on the network transmission index and the clock synchronization information; the index detection module is composed of a sending end and a receiving end, and the sending end is configured to send audio data; the receiving end is configured to generate feedback data based on the received audio data and transmit the feedback data back to the transmitting end, and the feedback data comprises a buffer area change trend, the number of packet loss times caused by buffer area overflow and clock offset; the performance adjusting module is configured to determine a performance adjusting strategy based on the network quality score and the feedback data; and adjusting the audio transmission parameters based on the performance adjustment strategy.

The application belongs to the technical field of distributed online optimization and learning, and discloses a quantized distributed online proximal gradient descent optimization method based on Bandit feedback. The method aims to solve the distributed online composite optimization problem under the condition that the communication resource is limited, the network structure is unbalanced, and the gradient information of the loss function is difficult to obtain or even cannot be obtained. The technical scheme comprises the following steps: constructing an unbalanced network graph and determining the node neighborhood relationship, modeling a distributed online composite optimization problem based on Bandit feedback, designing a quantized distributed online proximal gradient descent algorithm combined with an adaptive uniform quantization strategy and a proximal gradient descent technology, and performing static regret analysis on the algorithm.

The application discloses an output driving circuit, which comprises an input unit, a multi-stage output unit, a load unit, a feedback network, a bias network and a balance network; the input unit, the multi-stage output unit and the load unit are connected in sequence, the feedback network is connected with the output of the multi-stage output unit, the bias network is connected with the input unit and the multi-stage output unit, and the balance network is connected with the multi-stage output unit. The output driving circuit has the advantages of large output swing, high output bandwidth and good reliability.

This invention discloses a nationwide integrated power balance method. First, a nationwide integrated balance mathematical model is constructed. Based on this model, the integrated balance network equations characterizing network structural constraints are derived. Second, based on the network equations, a new type of key balance objective for the power system is introduced, deriving an integrated balance dynamic equation consisting of four process forms: natural mutual support for supply guarantee, mutual support for supply guarantee and start-up, natural mutual support for absorption, and mutual support for absorption and start-up. Then, engineering boundary conditions are substituted into the integrated balance dynamic equations, deriving nine integrated balance configurations for engineering practice to achieve nationwide integrated power balance. The new power system integrated balance model and mechanism proposed in this invention effectively solve the fundamental principle problem of nationwide integrated balance. By proposing a basic model for integrated balance engineering practice and eight sub-configurations, the problem of diverse engineering practice scenarios and multiple objectives is effectively addressed.

The application discloses a cell camping method and device, which is used for balancing network load among different cells and ensuring continuity of terminal device service. The method comprises the following steps: after a terminal device camps on a first cell, a network device can send a first message to the terminal device, the first message containing information of a second cell, the first cell corresponding to a first CC and the second cell corresponding to a second CC; after the terminal device receives a second message, the terminal device sends the second message to the network device, the second message can indicate the first cell and / or the second cell, or the second message can contain measurement information of the first cell and / or the second cell, and then the network device can determine, based on the second message, that the terminal device camps on the first cell or the second cell, so as to balance network load among different cells, and meanwhile, the whole process does not need the terminal device to return from an RRC connected state to an RRC idle state, and the continuity of terminal device service can be ensured.

The present disclosure generally relates to optimizing load-balancing of network endpoints using tree collectives representing a logical network communication topology for the network endpoints. Systems and methods described herein eliminate the previously restrictive conditions imposed on tree-based communication collectives by generating collective trees with any arity and representing any number of physical network endpoints. The resulting collective trees ensure that each represented network endpoint has a number of outgoing flows and a number of incoming flows that are no more than the arity of the collective tree. In this way, the described systems and methods inject significant efficiencies into communication collectives within networked compute nodes by eliminating communication bandwidth latencies and bottlenecks.

Embodiments of the present application provide a data processing method, a network training method, an electronic device and a storage medium. The method comprises: obtaining input data of a network layer of a deep neural network, original weights of the network layer being decomposed into N sub-weights for quantization processing; performing convolution processing on the input data and the N sub-weights respectively to obtain N intermediate data corresponding to the N sub-weights; and performing summation processing on the N intermediate data to obtain output data of the network layer. In this way, the original weights of the network layer are decomposed into the N sub-weights for quantization processing, and the efficiency and performance of the network can be flexibly balanced by adjusting the value of N.

Disclosed are systems, apparatuses, methods, and computer-readable media for balancing replicated frames in a replication network. A method includes: receiving, at a redundancy receiver, first frames from a first path and second frames from a second path, determining a network performance imbalance between a redundancy edge device to the redundancy receiver based on receiving the first frames and the second frames, and in response to determining a threshold of memory will be exceeded based on the delay between the first frames and the second frames, providing an indication to slow frames transmitted along the second path. In some aspects, a path mismatch device is configured to track replication frames and determine network performance imbalance. For example, the path mismatch device can identify one network is slower than another, and may provide remediation to balance network performance that maintains the integrity of the replication network.

The invention belongs to the technical field of remote sensing image processing, and particularly relates to a multi-modal remote sensing data classification method based on a semi-supervised contrast regular network. The method comprises the following steps: generating high-quality pseudo labels for all label-free multi-modal remote sensing data by adopting a multi-modal label distribution method, screening samples with balanced categories according to the pseudo labels, constructing a label-free sample set and a label sample set, and performing enhancement processing on the sample sets; constructing a semi-supervised contrast regular network; performing supervised learning of the network by using the labeled sample set; performing consistent regular learning of the network by using the reliable unmarked sample; contrastive learning is executed, an anchor sample set, a positive sample set and a negative sample set are established according to reliability judgment, contrastive loss is calculated, and unreliable samples provide negative sample features for the loss; and jointly using three loss functions to jointly optimize the classification performance of the network. The multi-modal remote sensing data classification method is beneficial to realizing category-balanced network optimization in training, enhances the network robustness, improves the network classification performance, and can be widely applied to multi-modal remote sensing data classification.

Load balancing is provided between network devices to perform DPI operations. Network devices in a virtualized network device array can balance network traffic by using a network connection between the devices to exchange data packets for DPI operations. Furthermore, the network devices can use dedicated CPUs to offload DPI operations. In this way, the techniques presented allow existing network routing or deterministic routing to be maintained, enabling DPI to be performed on the fly without restricting network traffic to a specific route.