178 results about "Network complexity" patented technology

A distributed processing, interactive computer network and method of operation is described. The network is designed to provide very large numbers of simultaneous users access to large numbers of applications which feature interactive text / graphic sessions. The network includes one or more host computers having application data stores; a plurality of concentrator computers, also including application data stores, the concentrator computers being connected in groups of one of more to each of the host computers; and a plurality of reception system computers connected in groups of one or more to each of the concentrator computers, the reception system computers being arranged so that respective users can request interactive applications at the reception system computers. In accordance with the design, the reception system computers also include application data stores. The method for operating the network includes steps for generating the interactive text / graphic sessions from objects that include data and / or program instructions. Additionally, the method features steps for distributing objects among the data stores of the network computers, and, thereafter, permitting the reception system computer at which an application is requested to selectively collect objects required for the application from the network and the respective reception system so that the requested application may be presented at the reception system based on the objects collected. This operation decreases processing demand on the higher-level network elements, permitting them to function primarily as data supply and maintenance resources, thereby reducing network complexity, cost and response time.

A method is given for diverting injected slickwater in a hydraulic fracturing treatment. The diversion fluid is preferably a substantially proppant free viscous fluid that causes a net pressure increase and plugging of some of the microfractures in the initial fracture system created, which induces formation of supplementary microfractures connected to the initial fracture network and in-creases the contact area with the formation rock. The method generates a greater fracture network complexity and thus a higher contact area with the reservoir during a single treatment cycle.

The present invention relates to a method, mobile station and base station in a soft handover cellular communications system. A problem addressed is the complexity imposed by a soft handover combination node in the network for combining packets received from one mobile station over parallel diversity links. The combination node is a problem with respect to the network complexity, transmission capacity and retransmission delays. The present invention solves the problem by selecting on a packet-by-packet basis one of an active set base stations for forwarding the packet further uplink. Thereby no combination node in the network is needed. The mobile stations select the base station after transmitting a packet or prior to transmitting it. The selection is based on measures of the radio link qualities made during transmission of the relevant packet or made prior to its transmission.

The invention belongs to the technical field of wireless communication, and relates to an anti-interference method for communication based on deep deterministic gradient reinforced learning. The method provided by the invention comprises the steps of firstly building an interference environment model according to the quantity of interference sources and a wireless channel model; building a utilityfunction according to a legal user communication quality index, using the utility function as a return in learning; forming spectrum information sampled at different time slots into a spectrum time slot matrix, and describing an interference environment state by using the matrix; and then building a convolutional neural network according to a deep deterministic gradient reinforced learning mechanism, and when an anti-interference decision is made, an environment state matrix achieves anti-interference strategy selection of the corresponding state in a continuous space via a target actor convolutional neural network. According to the method provided by the invention, the continuous anti-interference strategy selection in communication is completed based on the deep deterministic gradient strategy reinforced learning mechanism. The quantization error caused by quantized discrete processing on the policy space is overcome, the quantity of cells output by the neural network and the network complexity are reduced, and the anti-interference algorithm performance is improved.

The invention discloses a method for locating faults in a power communication network. The method comprises the steps that a bipartite graph model is built according to the many-to-many uncertainty of faults and symptoms; fault effect weight factors are led in on the basis of the bipartite graph model; the fault effect degree is calculated and corrected through credible parameters to obtain suspected fault sets. According to the method, the many-to-many uncertainty of faults and symptoms is modeled through a weighting bipartite graph, so that the causal relationship between the faults and the symptoms is expressed, the modeling capacity is good, noise immunity is high, and the method can adapt to real environment with fault burstiness and network complexity. The fault effect weight is led in, the total probability and the Bayes thought are utilized under the bipartite graph model, prior fault probability is transformed into condition probability, and the fault effect degree is calculated; finally, credible parameters are added to control effect of suspected faults, the cover degree and the contribution degree are combined, and the suspected fault sets of which the effect degree is within the controllable parameter range are picked out.

The present invention discloses a method for activating a Multimedia Broadcast / Multicast Service (MBMS) comprising the steps: a. sending a message which carries MBMS bearer capabilities of a user equipment (UE) from the UE to a SGSN which the UE belongs to after passing authorization; b. verifying whether the MBMS bearer capabilities of the UE are less than Required MBMS Bearer Capabilities, if the SGSN has the Required MBMS Bearer Capabilities; and c. rejecting a request for activating an MBMS Context if the MBMS bearer capabilities of the UE are less than the Required MBMS Bearer Capabilities, or creating an MBMS UE Context if the MBMS bearer capabilities of the UE are not less than the Required MBMS Bearer Capabilities. The method can simplify the signaling interaction between the network entities, reduce the network complexity and save the network resources.

The invention relates to the field of neural network computing technologies, and is a self-adaptive learning neural network implementation method based on an evolutionary algorithm. One or more types of known neural networks are used as an initial parent of the evolutionary algorithm; and characteristics of the neural networks which are the initial parent are integrated by using the evolutionary algorithm, so as to obtain an optimum output value. According to the invention, binary coding is performed on a circuit implemented by the neural networks, a result obtained by coding is used as a chromosome of an individual; and chromosomes constitute a primitive population of an organism, that is, the initial parent. According to the invention, a case in a conventional method that only the evolutionary algorithm is used to optimize a neural network weight is broken through; optimization for modes such as a neural network organization form, a connection weight among networks, and a network calculation method is simultaneously implemented by using the evolutionary algorithm; a network freedom degree is increased; an optimization scope is enlarged; and a relatively simple network is initially obtained; and in acquired learning, network complexity is increased by using an algorithm.

The embodiment of the invention discloses a push service negotiation method, which is applied to an IMS Based IPTV network. The method comprises the following steps: receiving push service informationcarrying a push service identifier; analyzing the push service information; triggering push service logic according to the push service information, and sending a push service request carrying a pushcontent identifier; and negotiating to complete the push service logic. Simultaneously, the embodiment of the invention also discloses a push service negotiation device and a push service system. Thepush service negotiation method, the push service negotiation device and the push service system trigger the push service logic through the push service identifier and the push service information tonegotiate to complete the push service logic, can meet the requirement of pushing content to users by initiating session through a network side, can realize the push function universal for services to continue various push services, and can effectively improve application deployment and simplify the network complexity.

The invention provides a dynamic host configuration method and system in a software defined network. The method comprises the following steps that: a controller computes virtual network configuration in an entire physical network, and transmits the virtual network configuration to an OpenFlow switch by using an OpenFlow protocol to form a plurality of virtual networks in the entire physical network; the OpenFlow switch receives a flow table control command issued by the controller, and performs matching rule configuration on an own flow table according to the flow table control command; the OpenFlow switch receives data packets from other OpenFlow switches or hosts, processes the data packets according to a matching rule recorded by the own flow table and the attributes of the data packets, and outputs a DHCP (Dynamic Host Configuration Protocol) data packet needing to be forwarded; and the OpenFlow switch forwards the DHCP data packet to the controller, a destination OpenFlow switch or a destination host. Through adoption of the dynamic host configuration method and system, centralized dynamic configuration of an access host in a multiple virtual overlay network is realized; the network efficiency is increased; and the problem of increase of the network complexity due to the multiple virtual overlay network in the software defined network is solved.

The invention relates to the related technical field of lightweight expression recognition methods, in particular to a deep learning-based lightweight expression recognition method, which comprises the following steps that firstly, quick and efficient face detection is realized through an MTCNN lightweight face detection network; then, the basic structure of the MobileNetV3 network is improved, and a linear bottle is improved on the basis of depth separable convolution, so that the network complexity is reduced, and over-fitting is prevented; and finally, an efficient attention module is designed to combine the depth and spatial information of the feature map, important feature extraction is more emphasized, an intra-class feature difference of the same expression is reduced by adopting aCenter loss function, and Softmaxloss expands a feature distance between different expression classes, so that the network has a better feature discrimination effect, and finally, real-time facial expression recognition is completed.

The invention relates to a route protection converting method and device. The route protection converting method comprises the following steps of: configuring a three-layer virtual interface bound with a two-layer VPN (Virtual Private Network) on device nodes of two ends of the two-layer VPN; establishing a protection group relation between a protection route and a protected route according to the binding relation of the two-layer VPN and the three-layer virtual interface; and converting a link flow rate from the protected route to the protection route through the three-layer virtual interface according to the protection group relation when a protected link has failure. According to the route protection converting method disclosed by the invention, through the three-layer virtual interface ARP (Address Resolution Protocol) protection policy, without changing an original network topology and two / three-layer domain partition, the protection of an over two-layer network domain on a three-layer network domain route is realized, a problem of flow interruption of the three-layer route at the two / three-layer network connection caused by DOWN (Direct Memory Access Controller) is solved, and additional devices and a hardware lookup operation do not need to be additionally arranged; and in addition, the expandability of a network structure is improved, the network complexity is reduced and the maintenance cost is saved when conversion performances are ensured.

The invention relates to a cluster network super-resolution image reconstruction method of an Laplace pyramid structure. By dividing the images to be reconstructed into training data sets and test data sets and enhancing and preprocessing them respectively, a Laplace pyramid cluster network is constructed and the training data sets are input for training, and the test data sets are input into thetrained Laplace pyramid cluster network to reconstruct super-resolution images step by step. A Laplace pyramid structure is adopted, By progressively reconstructing high-resolution images, Step by step optimize reconstruction results, better solve super-resolution image reconstruction, the residual learning is applied to the network, Reducing network parameters and avoiding gradient explosion caused by network complexity increase. In each construction module, there are both forward and feedback connections between any two convolution layers, and the information between layers is updated alternately, which maximizes the information flow and feedback mechanism between layers, makes the connection between layers more dense and extracts more detailed features.

The invention discloses a pedestrian re-identification method based on video appearance and motion information synchronous enhancement. During training, pedestrian appearance and motion information ina backbone network are respectively enhanced through an appearance enhancement module AEM and a motion enhancement module MEM. The appearance enhancement module AEM uses an attribute recognition model obtained by training an existing large-scale pedestrian attribute data set to provide an attribute pseudo label for the large-scale pedestrian video data set, and enhances appearance and semantic information through attribute learning; the motion enhancement module MEM predicts pedestrian gait information by using a video prediction model, enhances gait information features with identity discrimination ability in a pedestrian feature extraction backbone network, and improves pedestrian re-identification performance. In practical application, only the pedestrian feature extraction backbone network needs to be reserved, and higher pedestrian re-identification performance can be obtained without increasing the network complexity and the model size. And the enhanced backbone network featuresobtain higher accuracy in a video-based pedestrian re-identification task.

The invention discloses a refrigerator-based smart home control system, and relates to the technical field of smart home control. The technical shortcomings of networking complexity and high cost of the conventional smart home control system are overcome. The system is arranged in a refrigerator body, and comprises a system master control module, and a network communication module, a liquid crystal display (LCD) touch module, a wireless fidelity (WIFI) module and a refrigerator master control module, which are connected with the system master control module respectively. The smart home control system is combined with an Internet of things refrigerator, so that the optimal selection of a smart home can be realized with the lowest cost, and the cost of a smart home system can be decreased. A WIFI technology is applied to the refrigerator, so that the smart home system has the advantages of convenience for networking and no need of wiring, the networking complexity of the smart home system is lowered, and the cost of the system is greatly decreased.

The invention provides an electric power system state estimation method based on a convolutional neural network. According to the method, offline power system power flow section data is taken as a sample for training; the quantity measurement is used as input data; The convolutional neural network state estimation method is characterized in that a state quantity is taken as an expected output, training parameters are reduced by means of local connection, weight sharing, downsampling and the like of a convolutional neural network, the network complexity is reduced, overfitting is prevented by means of back propagation of errors between input and output, and a convolutional neural network state estimation model based on quantity measurement is trained. The method estimates the new quantity measurement data through the trained convolutional neural network to obtain the state quantity of the system at the moment. According to the method, the convolutional neural network is adopted for state estimation, the defects that a traditional iterative least squares state estimation method is too long in calculation time and gradient transmission of a full-connection neural network state estimation method is difficult are overcome, and the training difficulty of the network is reduced while the calculation time is reduced.

The invention discloses a node injection power-based running topology identification method of a power distribution network. The node injection power-based running topology identification method comprises the steps of building a variance mode of node injection power measurement-based branch voltage deviation according to node injection powers sampled for many times by a miniature synchronous phase measurement unit Mu PMU; obtaining minimum spanning tree taking variance of the branch voltage deviation as a circuit weight by employing a Kruskal algorithm to achieve identification on a topological running structure of the power distribution network; and comparing and analyzing the sampling frequency of node injection power measurement and influence of network complexity on an topological identification error. A simulation result shows that the algorithm has relatively good reliability and practicability on running topological identification of the power distribution network.

This invention discloses a physical layer chip, used between the MAC chip and the mouth modules to transmit signal, the physical layer chips including conversion modules for low-speed side signals multiplexing M from the MAC duplicate N road for high-speed optical signal transmitted to the mouth module or for the modules from the light side of the mouth high-speed interface N road signals for M Road after low-speed signals transmitted to the MAC chip. The present invention also discloses two Ethernet physical layer signal transmission method for the MAC chip and optical modules mouth between the signal multiplexing and discloses an application of this technology to the switch. This invention makes a greater rate-mouth module can replace two or more smaller rate optical module mouth, decreasing the physical layer interface chip number and reducing the physical layer chip itself costs, and further reducing the cost of the equipment, and the realization of the complexity, which can be accessed and resolved within the cost of network complexity and Qos issues.

Disclosed are an access session controller, an IP multimedia subsystem and registration and session establishment methods thereof, where the access session controller includes a message reception module, a service trigger configuration module, a session control module and a message transmission module. The IP multimedia subsystem includes the access session controller and a network-wide home subscriber server system. The registration method includes: initiating, by the access session controller, an authentication information request to the network-wide home subscriber server system. The session establishment method includes: directly process, by the access session controller, the request message initiated from a user to establish a session. The foregoing solutions can simplify the signaling processing flow and reduce the network complexity.

Certain embodiments of the invention may include systems and methods for accelerating volt / VAR load flow optimization. According to an example embodiment of the invention, a method is provided for accelerating load flow for integrated volt / var control (IVVC) optimization. The method can include evaluating load flow on lines of an electrical network, identifying combinable network lines, combining the identified combinable network nodes to reduce network complexity, and determining load flow optimization for IVVC based at least in part on the reduced complexity network.

The invention discloses a deep learning characteristic generalization method based on a latent variable model. The method is carried out according to the following steps that: S1: on the basis of thelatent variable model, establishing one pair of bidirectional parametric mapping between data space and latent variable space, and combining with a weighted relationship to construct a characteristicgeneralization layer; S2: embedding the characteristic generalization layer into a deep neural network, and dividing the network into three parts, including a DNN (Deep Neural Network)-1, the characteristic generalization layer and a DNN-2; S3: determining the optimization objective of the model, and defining a target function; and S4: lowering characteristic pattern data complexity, and establishing a multi-branch parallel forward propagation structure. The method is favorable for improving the generalization ability of a deep network model, and an overfitting phenomenon during small-scale data training can be lightened. Compared with other methods which use a generative model to enhance data, the method disclosed by the invention simplifies network complexity and improves training efficiency.

The invention provides a network monitoring processing method and system, belonging to communication field, comprising: according to monitoring instruction, determining signaling flow matching with the monitoring instruction and determining communication party transmitting the signaling flow as monitored object; sending data information to the monitored object to obtain message and implementing monitoring of data information transmitted by the monitored object. And it replaces session boundary controller with monitoring processor set in kernel control party to monitor the communication party, where the setting method is simple and easy to operate, not increasing network complexity; it keeps the original basic structure of VOIP network, avoiding increasing network safety risk and unable to largely increase the cost.

The invention relates to a method for achieving an Arnold image scrambling effect on a quantum computer. The method is characterized in that the Arnold image scrambling effect can be achieved on the quantum computer. The method includes the steps that an original quantum image is provided, Arnold scrambling conversion is carried out on an x-axis coordinate of the original image, and Arnold scrambling conversion is carried out on a y-axis coordinate of the original image; scrambling conversion carried out on the x-axis coordinate further includes the procedures of calculation of x+y and execution of a modular arithmetic on 2n; scrambling conversion carried out on the y-axis coordinate further includes the procedures of calculation of 2y, calculation of x+2y and execution of a modular arithmetic on 2n. The algorithm is provided in a quantum circuit form, and it is guaranteed that the algorithm can directly operate on the quantum computer; compared with an existing method for achieving the Arnold image scrambling effect on the quantum computer, a modular arithmetic operation can be achieved by ignorance of carry directly, and network complexity is greatly reduced.

The invention relates to a knee joint moment estimation method based on surface myoelectricity and motion signals. The method comprises the following steps: firstly, establishing an offline knee jointmoment estimation model: establishing the knee joint moment estimation model by using sEMG of a pair of active muscles and antagonistic muscles under knee joint flexion motion, knee joint angles andknee joint angular velocities and by utilizing an LSTM network and optimizing the established model; and then carrying out online estimation on the knee joint torque signal. According to the invention, the characteristic quantity of the input model can be reduced, and the network complexity is reduced.

The invention discloses an EPON (Ethernet passive optical network) optical network unit supporting a CATV (cable television) optical access and fusing EoC (Ethernet over coaxial cable) functions, which comprises an EPON remote machine, an EoC head end, a CATV optical receiving module and an inter-plate connector, wherein the EPON remote machine and the EoC head end are connected by the inter-plate connector; the EPON remote machine completes the configuration management of devices and the PON (passive optical network) access of IP (internet protocol) data; the inter-plate connector transmits 100Mbps Ethernet data, reset signals and port connection status indication signals; the EoC head end completes the mutual conversion of the Ethernet data and low-frequency data of an EoC and the coupling of high-frequency / low-frequency data of the EoC; and the CATV optical receiving module is used for completing the conversion between optical signals of a CATV and radio-frequency signals of the CATV. The optical network unit disclosed by the invention can provide reliable IP data and CATV signal access without adding new CATV optical receivers and corollary devices, therefore, the networking cost, network complexity and construction process of an HFC (hybrid fiber-coaxial) are reduced.

The invention belongs to the technical field of data analysis and discloses a kernel incremental transfinite learning machine and a differential multi-population gray wolf mixed optimization method. Aiming at the problem that the kernel incremental transfinite learning machine (KI-ELM) has the redundant nodes with low learning efficiency and poor accuracy; At first, that invention utilize the differential evolution algorithm and the multi-population grey wolf optimization algorithm to propose a hybrid intelligent optimization algorithm--the differential multi-population grey wolf optimizationalgorithm, optimizes the node parameter of the hidden layer, and determine the effective node quantity, so as to reduce the network complexity and improve the learning efficiency of the network; Secondly, the depth structure is introduced into the kernel incremental transfinite learning machine, and the input data is extracted layer by layer to realize the high-dimensional mapping classification of the data and improve the classification accuracy and generalization performance of the algorithm. The simulation experiment results show that the hybrid intelligent depth kernel incremental transfinite learning machine provided by the invention has good prediction accuracy and generalization ability, and the network structure is more compact.

The invention discloses a YOLO v3-based detection method for key objects of a transportation junction. An algorithm is designed on the basis of the idea of direct regression to achieve multi-scale detection and multi-label classification. A Darknet-53 network based on ResNet is designed as a feature extractor based on the defects of current target detection technology, so that the detection precision and speed of YOLO technology architecture are improved. Meanwhile, the defect that the YOLO technology architecture is not good at detecting small objects is improved. The Darknet-53 network considers both the network complexity and the detection accuracy, and compared with a common target detection feature extraction network VGG-16, the model operand is reduced. The latest progress in the field of artificial intelligence is introduced into main target detection in the transportation hub, the detection precision and the detection speed are good, and meanwhile the method has the potential of being expanded and applied to other fields.

The invention provides a bridging method based on a VF promiscuous mode. The bridging method is achieved through the steps of setting to support the VF promiscuous mode, setting to support a sniffingmode, enabling the VF promiscuous mode and configuring built-in bridging mode support port mirror. The invention further provides a built-in bridging system supporting the VF promiscuous mode. By adopting the built-in bridging method in an SR-IOV promiscuous mode, PFs can be directly called, and therefore the performances of virtual machines approach to those of physical machines; the method can support the mode of deploying multiple types of the virtual machines, support to control the multiple mode (unicast, multicast, broadcast, VLAN) promiscuous mode of a single VF, achieves a hardware network card-based built-in switch function and has the advantages of being high in performance, not occupying system resources such as a CPU and a memory, reducing the cost and the network complexity and the like.

Disclosed is an invention related to a system and device for actively defending a network infrastructure by implementing certain features that are attributed with lower performance cost and network complexity. The features implemented for protecting the network infrastructure comprises of: protecting the network from hostile scanning, providing a faster authenticated and limited access response to a network traffic request for sage guarding dedicated connections, intervening a TCP connection that is established between one or more clients and servers for terminating unwanted connections, and cleaning up SYN flood attacks to terminate one or more outstanding TCP connection.

The invention discloses a buffer adjustment method for a key chain based on project property, which comprises the following steps: step 1, generating an initial feasible scheduling scheme; step 2, calculating free time difference of task and identifying the key chain; step 3, calculating buffer size by two classical quantification methods, selecting the better quantification method to adjust the buffer size through an adjustment strategy; and step 4, generating a buffer scheduling scheme and outputting the result. According to the buffer adjustment method for the key chain based on project property, two important factors of network complexity and resource tightness which affect the robustness of the scheduling scheme are considered, and the network complexity and the resource tightness are introduced in the buffer quantification method, so that the scheduling scheme can respond to the uncertainty of a project in a better way. The method can obtain a reasonable buffer inserting scheme on the basis that the feasibility of the scheduling scheme is ensured. The buffer adjustment method for the key chain based on project property has practical value and vast application prospect in the technical field of optimizing and scheduling.

The invention provides a new method for identifying the type of a fuel on line, which comprises the following steps: extracting primary characteristic values of a time domain and a frequency domain of a flame from a radiation signal of the flame; subjecting the primary characteristic values first to data preprocessing to change the primary characteristic values into irrelevant orthogonalized characteristic value data, approximating the nonlinear function relationship between the characteristic value data and the fuel types with two layers of neural networks with nonlinear hidden layers to make the output of the neuron of an output layer between 0 and 1, and showing the degree of the similarity of a combusting fuel to a fuel of some type. The primary characteristic values extracted by the method are subjected data preprocessing before being input into the neural networks, so the method can use more useful characteristic information while obviously reducing the complexity of the networks, can increase the training speed of the neural networks and reduce the training time of the neural networks, and also can greatly increase the success rate of fuel type identification.