123 results about "Scale-free network" patented technology

A machine learning algorithm is trained to learn to cluster a plurality of original-destination routes in a network for transporting cargo into a plurality of clusters based on similarities of the original-destination routes, and to learn to cluster the plurality of clusters into a plurality of subgroups based on customer behavior. Influencing criteria associated with each of the subgroups may be determined and based on the influencing criteria, a price elasticity curve for each of the subgroups may be generated. Based on the price elasticity curve and current network traffic, cargo transportation price associated with each of the subgroups may be determined.

The invention discloses a medical image synthesis and classification method based on a conditional multi-judgment generative adversarial network. The method comprises the following steps: 1, segmenting a lesion area in a computed tomography (CT) image, and extracting a lesion interested area (Region of Interest, ROIs for short); 2, performing data preprocessing on the lesion ROIs extracted in thestep 1; 3, designing a Conditional Multi-Discriminant Generative Adversarial Network (Conditional Multi-) based on multiple conditions The method comprises the following steps: firstly, establishing aCMDGAN model architecture for short, and training the CMDGAN model architecture by using an image in the second step to obtain a generation model; 4, performing synthetic data enhancement on the extracted lesion ROIs by using the generation model obtained in the step 3; and 5, designing a multi-scale residual network (Multiscale ResNet Network for short), and training the multi-scale residual network. According to the method provided by the invention, the synthetic medical image data set with high quality can be generated, and the classification accuracy of the classification network on the test image is relatively high, so that auxiliary diagnosis can be better provided for medical workers.

A deep learning model based on a multi-scale network and application in brain state monitoring are provided. A model establishing method comprises steps of: preprocessing and multi-scale transforming a measured multichannel signal; obtaining a multi-scale weighted recursive network and a cross-recursive rate matrix corresponding to the multi-scale weighted recursive network of the multichannel signal in all scales; extracting the network indexes of the multi-scale weighted recursive network at different scales; at each scale, retaining relatively large elements in the cross recursive rate matrix and obtaining an unweighted adjacent matrix and a multi-scale unweighted recursive network corresponding thereto; for each value of a variable in the set range, obtaining the multi-scale unweighted recursive network and the adjacent matrix corresponding to the multi-scale unweighted recursive network, extracting the network indexes of the multi-scale unweighted recursive network at different scales, calculating the integral of the network indexes when the variable is changed in the set range, and the integral as the final network index of the multi-scale unweighted recursive network under each scale; and training the deep learning model and monitoring a brain state.

A method for identifying a bridge node in a network using a processor and memory unit in a specially programmed special purpose-purpose computer including the steps of, for each node in a plurality of nodes in the network: determining a global metric proportional to total traffic flow in the network and through the node; determining a local metric proportional to traffic flow between the node and each second node in the network connected to the node and traffic flow between each second node and each third node in the network connected to a second node; determining a second local metric proportional to the respective traffic flows between each node and each second node; and calculating a respective combination of the global metric and the first and second local metrics; and selecting, a bridge node from among the plurality of nodes based on the respective combinations.

The invention relates to a particle swarm optimization method based on a complex network. The particle swarm optimization method is used for solving the multiobjective optimization problem in the real world. The particle swarm optimization method based on the complex network comprises the steps that the population network topology is established according to a scale-free network generation mechanism, the optimization space, the population size, the positions of particles and the speeds of the particles are determined, the adaptive value is calculated according to a fitness function, the historical best position of each particle, the historical best position of the corresponding neighbor particle and the global historical best position of the particles are recorded, the positions and the speeds of the particles are updated in an iteration mode every time, the adaptive value is calculated again until iteration is completed, and the global best position is output. The particle swarm optimization method based on the complex network further provides four indexes for evaluating the optimal performance of center particles and non-center particles, the influence in neighborhood, the information transmission capacity, the advantages and disadvantages of the adaptive value and the capacity for maintaining population activeness. By means of the particle swarm optimization method based on the complex network, the local optimum can be effectively avoided, and the convergence rate and the optimization effect for resolving targets are balanced through the application of the particle swarm optimization algorithm.

The invention discloses a method for establishing a scale-free sensor having survivability, mainly solving the problems of bad survivability, low adjustability and low universality in the prior art. The method comprises the following steps: selecting cluster heads according to games among nodes, and evenly distributing the cluster heads; then performing topology evolution on the cluster heads according to evolution means of a scale-free network to obtain survivable mathematical models of the scale-free sensor network; and finally constructing the sensor network by fully considering randomly adding points, randomly removing points, randomly trimming and reconnection according to actual application characteristics of the sensor network. The sensor network established by the method in the invention has the scale-free characteristic, excellent adjustability and universality, and the network has excellent survivability to random and deliberate attacks; the network can satisfy the requirements of severe environments such as natural disasters, military application and the like on robustness of the sensor network.

The invention discloses a short message monitoring method that determines identity of senders based on social network mechanism. Short message is a shortcut method that applies wireless communication technology, which is not different from face to face communication, letter communication, voice conversation and Email communication in essence. The social network belongs to scale-free networks in complex network category. Clustering is a basic characteristic of the social network. The invention starting from clustering of the social network to implement short message preprocess and re-real time filtering based on communication historical records between short message senders and receivers and between receivers determines identity of the senders to block spam short messages. A short message monitoring centre is used in a hardware platform using symmetrical multi-process SMP technology. The monitoring centre is characterized in free maintenance, high blocking rate, low mis-judgment, which is not applicable for text short message filtering, but also effective to voice short message, multi-media short message in 3G generation.

The invention discloses a hierarchical network construction method for the data compression storage of a massive road network. The method comprises the steps that hierarchy division is conducted on massive network data, the level of division can be set by parameters; based on the network layer division, a network overlay map is constructed, topology features of an upper layer network based on the shortest path are reconstructed, so that the upper layer network still has connectivity; the network is partitioned on the basis of hierarchical network overlay map; on the basis of the hierarchical network overlay map, partitioning is conducted on the network; on the basis of the hierarchical partition construction, the nodes in a region are compressed, and by calculating reachable nearest neighbor partition boundary nodes, a node is attached to the boundary nodes and relevant information is saved, so that the compression of the massive network data is achieved. The hierarchical network construction method is mainly used for the hierarchical construction and compression storage of a large-scale road network, the whole structure and topological features of the network can be kept well after the large-scale compression of the network, and the efficiency of the analysis algorithm for a sub network can be improved.

The invention provides a construction method for a low-density parity-check code. A check matrix of the low-density parity-check code comprises variable nodes and check nodes. The method is characterized by comprising the following steps of: giving a degree distribution sequence of the variable nodes according to the power-law distribution of a scale-free network, and simultaneously limiting the degree of each check node to be two constant values; controlling the arrangement of the variable nodes from the left to right of the matrix according to an ascending sequence of degrees; constructing the check matrix under the constraint of the step 2 by using a progressive edge growth algorithm; and checking whether the check matrix obtained by the step 3 comprises four rings or not, and if the check matrix obtained by the step 3 comprises the four rings, finding the four rings, deleting a certain number l of rings by using a searing algorithm for the four rings to obtain a matrix without the four rings to obtain a final check matrix. The performance of the low-density parity-check code obtained by the method is not remarkably different from that of the conventional high-quality code, the complexity of the low-density parity-check code is remarkably lowered, and iterative decoding time is shortened.

The invention discloses a scale-free network based router-level topology modeling method. According to characteristics of increasing and preferential connection of a scale-free network, new added nodes are connected to old nodes, and many of the old nodes are connected so that nodes in a router-level topology are distributed in a power law mode; according to functions, organization relations and types of the nodes, the nodes in the router-level topology are divided into three layers of a core layer, a leaf layer and a main layer, and the layers are expandable; when the router-level topology is flattened to a plane, a radar chart is used for dividing connection quadrants, and the inside of the quadrants can be continuously refined and split; and as for the router-level topology, new nodes can be added to the layers/the quadrants according to ratios, and namely, probability calculation is performed according to weight to determine areas which the new added nodes belong to. By means of the mixing modeling method, degrees of coincidence with actual conditions are higher.

The invention provides a flight scheduling method and a flight scheduling device. According to the flight scheduling method, a generation mechanism of a scale-free network is adopted when the landing time of each flight is determined through the flight scheduling arrival time and the waiting cost in the air of each plane so as to generate a particle swarm network topology of the landing time of each flight, connection between P particles whose flight waiting cost is the minimum and other particles in the network topology are increased, and connection between Q particles whose flight waiting cost is the maximum and other particles in the network topology is reduced at the same time. According to the invention, a connection weight of the particles which are the closest to an optimal flight scheduling time result is increased, and a connection weight of the particles which are the furthest to the optimal flight scheduling time result is reduced, thereby enabling the convergence time of a finally acquired flight landing time result to be shortened, and improving the efficiency in determining the landing time of each flight. Due to a heterogeneity characteristic of the scale-free network, the diversity of the particle population network topology is ensured, and a result acquired by adopting a regular network is prevented from running into local optimum.

Methods for interactively preventing and detecting conflicts and errors (CEs) through prognostics and diagnostics. Centralized and Decentralized Conflict and Error Prevention and Detection (CEPD) Logic is developed for prognostics and diagnostics over three types of real-world constraint networks: random networks (RN), scale-free networks (SFN), and Bose-Einstein condensation networks (BECN). A method is provided for selecting an appropriate CEPD algorithm from a plurality of algorithms having either centralized or decentralized CEPD logic, based on analysis of the characteristics of the CEPD algorithms and the characteristics of the constraint network.

The invention discloses a building method of a complex network. The building method comprises main steps of generated network determination, adjacent matrix calculation, degree distribution polynomial determination, complex network adjacent matrix calculation and the like. The complex network obtained by adopting the building method simultaneously has the self-similarity and worldlet characteristics, the self-similarity characteristic derives from the fractal matrix form adjacent matrix generated through Kronecker product iteration of the generated network adjacent matrix, and the worldlet characteristic derives from the network with the diameter not exceeding two times of the diameter of the generated network. In addition, the degree distribution of the complex network can be theoretically worked out through the number multiplication and coefficient multiplication operation of the degree distribution polynomial expression form. Studies show that the degree distribution of the complex network does not meet the power-law distribution, i.e., the complex network is not a scale-free network.

The invention discloses a calculation method for optimizing a municipal drainage network plane layout design. The calculation method is characterized by mainly comprising three steps that firstly, a construction cost fitness function which has high matching degree with drainage network assessment is constructed; secondly, an initialized population is obtained from a tree structure genetic algorithm initializing scheme based on the divide-and-conquer method; thirdly, an expanded variable population is obtained by utilizing a line-adding and circle-breaking method. The calculation method for optimizing the municipal drainage network plane layout design has the advantages that the matching degree between the constructed construction cost fitness function and the drainage network assessment is high, so a drainage network can be well planned; the tree structure genetic algorithm initializing scheme based on the divide-and-conquer method enables an initialized scheme of a tree structure to be completed rapidly, and large-scale initialization of the large-scale pipe network can be achieved; a variation algorithm of the drainage network is simpler by utilizing the line-adding and circle-breaking method, a novel pipe network planning map can be generated after each line is cut out in a circle with added lines, and variation diversity is guaranteed.

The invention provides a dynamic link predication method of a network structure. The method comprises the following steps: step one, inputting a network structure corresponding to a service object; step two, performing Jaccard distance conversion on the input network structure to obtain a processed network structure; step three, calculating the distance between every two nodes in the network structure; step four, obtaining a network structure with a marked priority at current time; step five, repeatedly executing the step one to the step four at next time to obtain a network structure with a marked priority at the next time, wherein the priority of each link at the next time is postponed to the priority at the current time, and the priorities of the links are successively marked from high to low according to time; and step fix, taking a network structure with a marked priority at each time as a predication result of one network structure for a user to perform analysis processing on the service object. According to the invention, based on a dynamic network topology structure, a dynamic evolution mechanism of a complex network is taken into consideration, the calculation complexity is quite low, and the method provided by the invention is applied to link prediction of a large-scale network.

The invention relates to a distribution network topology reconstruction method based on an adaptive sparse regression method, belonging to the technical field of distribution network topology analysis. The distribution network topological structure reconstruction method of the invention designs an algorithm which does not need prior knowledge of the distribution network and branch measurement dataaccording to the characteristic that branch measurement equipment is difficult to install in the distribution network, and can complete the topological reconstruction of the distribution network onlythrough the sequential voltage data of the distribution network bus bar, and the method is simple and easy to operate. The invention solves the problem of biased estimation by using the adaptive Lasso algorithm on the original Lasso algorithm. At the same time, a supplementary criterion is added to correct the wrong estimation when the algorithm does not meet the feasible conditions, which improves the accuracy of the algorithm. The method can be used in loop-free network or loop-network, and the topology reconstruction of distribution network can be carried out in a short time.

The invention provides a compressive sensing method based on a scale-free complex network LDPC (Low Density Parity Check) code, which comprises two processes of the sensing matrix construction and the signal reconstruction arithmetic. The method is characterized by comprising the following steps: sparsely representing a signal x by adopting appropriate basis function, constructing a check matrix H of an irregular low complexity and scale-free network LDPC code, taking the check matrix H of the scale-free network LDPC code as the sensing matrix phi of the compressive sensing arithmetic, calculating the measured value y which equals to phi x, and reconstructing an original signal from the measured value by utilizing a belief propagation (BP) decoding algorithm. The method has the advantages that the constructed scale-free network LDPC code with good performance is applied to the compressive sensing; compressive sensing signal reconstruction is realized by utilizing the decoding method; the method can be applied to the fields such as signal processing, image processing, error correction coding and radar imaging; and the application prospect is wide.