1003 results about "Greedy algorithm" patented technology

A differential compression method and computer program product combines hash value techniques and suffix array techniques. The invention finds the best matches for every offset of the version file, with respect to a certain granularity and above a certain length threshold. The invention has two variations depending on block size choice. If the block size is kept fixed, the compression performance of the invention is similar to that of the greedy algorithm, without the expensive space and time requirements. If the block size is varied linearly with the reference file size, the invention can run in linear-time and constant-space. It has been shown empirically that the invention performs better than certain known differential compression algorithms in terms of compression and speed.

The invention discloses a delay and energy consumption-based efficient offloading method for a computing task of a mobile edge computing system. The method comprises the steps of building a macro basestation with multiple users and one deployment mobile edge computing (MEC) server and a single-cell cell model of a small base station without an MEC server; then building a mathematical model basedon a wireless communication theory; and finally, designing a delay and energy consumption-based efficient offloading solution for a computing task of the MEC system through a parallel greedy algorithm, and analyzing and providing the performance of the solution. Compared with existing efficient energy consumption offloading solution, the method well weights the delay sensitive demand and energy consumption demand of a terminal user, and lowers delay and total energy consumption of the system.

Embodiments automatically generate an accurate network of watertight NURBS patches from polygonal models of objects while automatically detecting and preserving character lines thereon. These embodiments generate from an initial triangulation of the surface, a hierarchy of progressively coarser triangulations of the surface by performing a sequence of edge contractions using a greedy algorithm that selects edge contractions by their numerical properties. Operations are also performed to connect the triangulations in the hierarchy using homeomorphisms that preserve the topology of the initial triangulation in the coarsest triangulation. A desired quadrangulation of the surface can then be generated by homeomorphically mapping edges of a coarsest triangulation in the hierarchy back to the initial triangulation. This quadrangulation is topologically consistent with the initial triangulation and is defined by a plurality of quadrangular patches. These quadrangular patches are linked together by a (U, V) mesh that is guaranteed to be continuous at patch boundaries. A grid is then preferably fit to each of the quadrangles in the resulting quadrangulation by decomposing each of the quadrangles into k2 smaller quadrangles. A watertight NURBS model may be generated from the resulting quadrangulation.

The invention discloses a speaker recognition method based on depth learning. The method comprises the following steps: S1) carrying out pre-emphasis and overlapping-type framing windowing on collected voice signals; S2) carrying out endpoint detection on the collected voice signals by utilizing a dual-threshold endpoint detection method based on short-time energy and short-time zero-crossing rate, and judging and indentifying the staring moment, transition stage , noise section and ending moment of the voice; S3) carrying out feature extraction on the voice signals; S4) forming a depth belief network model based on restricted boltzmann machine hierarchy, training the established depth belief network model by utilizing layer-by-layer greedy algorithm and with speaker voice feature parameters being combined, and adding a Softmax classifier to the top layer of the depth belief network model; and S5) inputting the voice features of a speaker to the depth belief network model obtained after being subjected to training, calculating the probability that the model outputs voice features similar to the voice features of the other speakers, and selecting the speaker corresponding to the maximum probability as recognition result.

The invention provides an electric car intelligent charging system and method on the basis of a mobile device and belongs to the technical field of electric car charging positive intelligent control. The system comprises a power grid management center, a station level management server, the mobile device, a charging device and a power battery. The station level management server comprises a data input module, a data processing module, a data feedback module and a historical data memory module and is used for calculating, solving and achieving the real-time optimum order charging scheme, outputting specific charging orders to the charging device and the mobile device and feeding back the charging information. According to the method, the local optimum greedy algorithm is used for solving the real-time optimum order charging scheme, the user interaction character is achieved, and the order charging scheme of an electric car is obtained by carrying out data mining and predicting on the user data and the power grid data and carrying out performance analysis on the power battery. Under the circumstance of ensuring that battery loss is small, the electric car intelligent charging system and the method greatly meet the individual requirements of the users and achieve effective optimization of a power grid.

The invention discloses a social network-based vehicle-mounted self-organization network routing method and belongs to the technical field of a vehicle-mounted wireless network. The method comprises the steps of (1) utilizing neighbor node information to calculate the direction angles and the effective values of nodes; (2) adopting a greedy algorithm added with a cache mechanism for the nodes on a road section, wherein intersection nodes adopt the neighbor nodes with the maximum effective values larger than those of the current nodes in an angle threshold value range as the next-hop transmission relay; (3) enabling vehicle nodes to study from the self history transmission actions by a Q learning algorithm assisted by a routing algorithm, wherein the nodes select the neighbor nods enabling a reward function to achieve the maximum convergence value as the next-hop transponder. The complexity of the routing algorithm is reduced, the system cost is reduced, and the Q learning algorithm is used for assisting the routing selecting, so the data packets are enabled to be transmitted along the path with the minimum hop number, and the time delay is reduced; the delivery rate of the data packets is improved and the end-to-end time delay and the consumption of system resources are reduced.

The invention relates to a travel demand analysis-based bus station point deployment method. According to the travel demand analysis-based bus station point deployment method, travel demand characteristics of Beijing are analyzed based on massive taxi data; passenger hot areas are obtained through trajectory data; data such as road traffic congestion conditions, road types, road width and the number of roads are obtained through GIS map data extraction, and the data are adopted as consideration factors for station site deployment; and station site deployment optimization targets are processed, and a greedy algorithm is adopted to solve problems in station site deployment. Bus station sites obtained through adopting the method is the basis of intelligent bus line design; and routes can be arranged between the bus station sites, so that travel cost of passengers and road resource consumption can be effectively reduced.

The invention discloses a multi-mode intelligent configurable method for implementing optimization of a wireless network. Firstly, the network optimization demand and object are raised according to the user's own network and then the demand and object are analyzed and used as a basis for determining the network optimization mode and establishing a simple model of wireless network, and the network optimization plan and configuration network optimization parameters are prepared. Subsequently, the network is optimized by the cost functions (such as capacity, coverage and network quality) at five different angles (antenna, power, address, frequency and load balance) in combination with different optimization algorithms. The optimization algorithms include three heuristic algorithms (simulated annealing, particle bee colony and ant colony) and the conventional greedy algorithm, and increase the network performance to the ideal level. Finally, the optimization results of the wireless network are sorted to provide a network optimization plan to the user for reference and reality basis. The method is intelligent and configurable, can meet reasonable requirements of users, can also achieve the purpose of optimizing 2G / 3G dual-network coexistence, has strong flexibility, and provides a good reference for the present network performance.

The invention provides a three-dimensional reconstruction method for binocular stereo vision. The method comprises the steps of performing image pair collection: obtaining images of a calibration plate and an object; performing camera calibration: performing calibration on a camera based on a Zhengyou Zhang calibration method to obtain initial internal and external parameters of the camera; performing stereo rectification: performing rectification on a distortion parameter through epipolar constraint to obtain a rectified image pair; performing image processing: obtaining a region of interest,and balancing brightness difference between left and right images, thereby searching for matching points; performing stereo matching: implementing a multi-grid intensive stereo matching algorithm toobtain a continuous smooth disparity map; and performing point cloud rebuilding and model rebuilding: obtaining a point cloud chart of the object, performing point cloud optimization, and obtaining acomplete object model through triangulation of a greedy algorithm. The smooth model with no obvious difference with the real object can be obtained; and the method can be used for three-dimensional reconstruction of multiple scenes.

The invention belongs to the field of the environment perception and autonomous obstacle avoidance of quadrotor unmanned aerial vehicles and relates to a depth Q learning-based UAV (unmanned aerial vehicle) environment perception and autonomous obstacle avoidance method. The invention aims to reduce resource loss and cost and satisfy the real-time performance, robustness and safety requirements ofthe autonomous obstacle avoidance of an unmanned aerial vehicle. According to the depth Q learning-based UAV (unmanned aerial vehicle) environment perception and autonomous obstacle avoidance methodprovided by the technical schemes of the invention, a radar is utilized to detect a path within a certain distance in front of an unmanned aerial vehicle, so that a distance between the radar and an obstacle and a distance between the radar and a target point are obtained and are adopted as the current states of the unmanned aerial vehicle; during a training process, a neural network is used to simulate a depth learning Q value corresponding to each state-action of the unmanned aerial vehicle; and when a training result gradually converges, a greedy algorithm is used to select an optimal action for the unmanned aerial vehicle under each specific state, and therefore, the autonomous obstacle avoidance of the unmanned aerial vehicle can be realized. The method of the invention is mainly applied to unmanned aerial vehicle environment perception and autonomous obstacle avoidance control conditions.

A data packet greedy-algorithm forwarding method for vehicle-mounted Ad hoc network fitting urban environment comprises following steps: (1) modeling urban road environment as undirected graph possessing weight, (2) dynamically selecting and updating fork in data packet forwarding path, (3) each node maintains a neighbor list and predicating and maintaining neighbor list, (4) updating position information of target node,(5) forwarding data packet based on greedy-algorithm forwarding policy. The inventive method fully utilizes characteristic of vehicle-mounted Ad Hoc network node, excels present algorithm in data packet delivery rate, time delay, average hop number and total data quantity of physical layer; the operation step is simple and easy, computation is not complex, demand for real time transmission is satisfied.

The invention discloses a self-adaptive reconstruction and an uncompressing method for power quality data based on a compressive sensing theory. A power quality data compression process with concurrent sampling and compression is achieved through a random measurement matrix, compressive sensing thoughts are used to perform sparse decomposition on the power quality data, sparse signals are subjected to Gaussian measurement encoding, and a self-adaptive matching pursuit algorithm is applied to reconstruct signals. According to the self-adaptive reconstruction and the uncompressing method, the random measurement matrix is simple in structure and quick in operation, in no need of intermediate variable storage space and independent of power disturbance signal characteristics, and has universality; compared with greedy algorithms of an orthogonal matching pursuit and the like, known sparseness is not needed, self adaption and regularization processes are provided, the operation time is short, and accurate reconstruction can be achieved; and constraints of compression after sampling of traditional data compression methods are broken through, little sampling can recover original power quality signals well, and accordingly, requirements for hardware can be reduced, and the compression efficiency is improved.

The invention discloses an enhanced map learning path planning method for an indoor mobile robot. The method includes the steps that (1) ambient environment information is acquired, and an obstacle probability density model is established; (2) path planning is carried out through a greedy algorithm and an enhanced map learning method; (3) path selection and self-adaptation speed strategy adjustment are conducted on the indoor mobile robot. By the adoption of the enhanced map learning path planning method, a current optimal path can be planned in real time according to the current condition of the indoor mobile robot and the inherent non-holonomic constraint of the robot, meanwhile, the obstacle crossing ability, the target point convergence ability and the planning efficiency of the indoor mobile robot can be considered through the self-adaptation speed strategy adjustment, and therefore the indoor mobile robot can arrive at a specific location safely and effectively.

The present invention relates to an intelligent express distribution method on a mobile platform. According to the intelligent express distribution method, a database, which comprises a starting point, each distribution address and a final destination, of a mobile terminal is established, an SDK of a map application is combined, a greedy algorithm with a roulette and a global k-opt method are used for planning an optimal distribution path before departure, modification is carried out according to changes and details of an optimal route are displayed in a character and map form. The intelligent express distribution method overcomes the defects that express distribution efficiency is seriously low, labor is wasted and express distribution delay is caused. The intelligent express distribution method adopts the global k-opt method, so that the express distribution route before distribution can be optimized; and due to adoption of the k-opt method, the subsequent path can be optimized by constructing a planning address set on the basis of not influencing the passed route, not only the distribution efficiency of the express company can be basically guaranteed, but also couriers also can be liberated from the load of planning the distribution path and real wisdom logistics is realized.

A method and system for ordering treatment or transport of victims of a mass casualty incident, prior to the incident, by establishing simulated casualty severity scores for each of the victims and considering the casualty severity scores and simulated resource availability. The order of treatment is provided on a card, chart, table or graph. Ordering treatment occurs through an analysis of triage simulations using analytical or mathematical programming techniques considering the severity scores, time periods of treatment delay resulting from constrained resources, estimated survival probabilities, and / or deterioration rates applied to the severity scores due to treatment delay. Alternatively, the order of treatment can be determined through data mining, pattern recognition, or greedy algorithm optimization techniques analyzing survival probability estimates associated with the simulated casualty severity scores, and changes to the survival probability estimates due to time periods of treatment delay resulting from constrained resources.

The invention provides a caching optimizing method of internal storage calculation. The method includes the steps that monitoring codes are inserted into a Spark source program, and dynamic semantic analysis is performed on an application program to construct a DAG; out-degrees of all vertexes in the DAG are calculated, RDDs of the vertexes of which the out-degrees are larger than one are screened, and the screened RDDs are RDDs needing to be cached to an internal storage; according to a greedy algorithm, the execution sequence of Action is adjusted so that the access sequence of RDD data calculation can be optimized; the weights of the RDDs are calculated, and the replaced RDDs in the internal storage are determined according to an internal storage replacement algorithm; it is determined how to process the replaced RDDs according to a multi-level caching algorithm. By the utilization of the caching optimizing method of internal storage calculation, a programmer does not need to examine and weigh internal storage using and display the RDDs of the appointed loading internal storage in the process of programming, programming loads of the programmer are reduced, meanwhile, the utilization rate of the internal storage is improved, and then the speed of processing big data is increased.

Methods, systems and articles of manufacture for identifying semantic nearest neighbors in a feature space are described herein. A method embodiment includes generating an affinity matrix for objects in a given feature space, wherein the affinity matrix identifies the semantic similarity between each pair of objects in the feature space, training a multi-bit hash function using a greedy algorithm that increases the Hamming distance between dissimilar objects in the feature space while minimizing the Hamming distance between similar objects, and identifying semantic nearest neighbors for an object in a second feature space using the multi-bit hash function. A system embodiment includes a hash generator configured to generate the affinity matrix and train the multi-bit hash function, and a similarity determiner configured to identify semantic nearest neighbors for an object in a second feature space using the multi-bit hash function.

Systems and methods for synchronizing two or more vehicles operating on an electric transportation line to optimize energy consumption. A controller is provided having a computer memory component storing a set of computer-executable instructions, a list of braking intervals, and a list of acceleration intervals for the vehicles. The controller also has a processing component configured to execute the set of computer-executable instructions to operate on the list of braking intervals and the list of acceleration intervals to minimize an energy consumption of the electric transportation line over a determined period of time by shifting acceleration intervals to synchronize with braking intervals. A dedicated heuristic greedy algorithm and an energy model are implemented in the controller as part of the computer-executable instructions to achieve the improved energy consumption.

The invention discloses a social network influence maximization method based on a community structure. The method comprises the following specific processes that: (1) dividing a community, forming a candidate node set, dividing a network to identify a kernel node and a boundary node in the network, and forming the candidate node set; (2) heuristically selecting nodes, and for each node in the candidate node set, verifying potential influence through a node degree, a community scale, a connected community number and influence weight so as to heuristically select and add the node with the highest influence into a seed set; and (3) executing a greedy algorithm, and utilizing the greedy algorithm to select and add the node with the highest marginal income into the seed set. By use of the method, the function of the community structure in influence spreading is analyzed to further improve accuracy and operation efficiency for mining initial seed nodes, and the problem of social network influence maximization is effectively solved.

An area is divided into regions or cells each having an associated multicast group address. The cells are organized using a quad-tree data structure that is dynamically updated. Mobile entities within a given cell elect one of their members as coordinator for assessing cost data associated with subdividing or merging the cell with other sibling cells. The coordinators communicate this cost data to a partitioning entity, which may be a central server, that computes a new partition scheme using a greedy algorithm. The new partition scheme is then broadcast to the mobile entities for use in subsequent communication. Mobile entities join multicast groups that intersect with their individual vision domains. In this way, mobile entities can acquire information from other mobile entities without the need to monitor all multicast group transmissions.

The subject invention provides systems and methods that facilitate AIDS vaccine cocktail assembly via machine learning algorithms such as a cost function, a greedy algorithm, an expectation-maximization (EM) algorithm, etc. Such assembly can be utilized to generate vaccine cocktails for species of pathogens that evolve quickly under immune pressure of the host. For example, the systems and methods of the subject invention can be utilized to facilitate design of T cell vaccines for pathogens such HIV. In addition, the systems and methods of the subject invention can be utilized in connection with other applications, such as, for example, sequence alignment, motif discovery, classification, and recombination hot spot detection. The novel techniques described herein can provide for improvements over traditional approaches to designing vaccines by constructing vaccine cocktails with higher epitope coverage, for example, in comparison with cocktails of consensi, tree nodes and random strains from data.

The present invention relates to a wireless communication system. More specifically, the present invention relates to a method for scheduling a terminal in a wireless communication system permitting overlapping between base station clusters, and an apparatus therefor, the method comprising the steps of: grouping a plurality of terminals into a plurality of groups satisfying predetermined conditions; and selecting one more terminals for scheduling by means of a group-based greedy algorithm for each of the plurality of groups, wherein, in the group-based greedy algorithm, a check is skipped for whether the total number of base station antennas corresponding to each group is equal to or greater than the number of terminals selected from each group.

The invention provides a multi-cell large-scale MIMO system user dual-antenna pilot frequency interference elimination method. Pilot frequency pollution of an uplink and interference in the data transmission process of a downlink are alleviated by considering spatial correlation and controlling the beam forming weight vector of a user side. Meanwhile, pilot frequency distribution is introduced to the method, and a pilot frequency distribution strategy based on a greedy algorithm is put forward so that influence caused by pilot frequency pollution is further alleviated through combination of a user dual-antenna device, and thus the objectives of reducing channel estimation error and enhancing system speed can be achieved.

The disclosed resource allocation method for ensuring QoS in MIMO-OFDM system comprises: BS of MIMO- OFDM system takes SVD decomposition to user-feedback channel matrix to obtain parallel feature sub-channel; from every sub-band; on condition BS total power allocated evenly, the BS scheduler allocates some sub-band to user according to user support speed and PRI, then re-allocates power on all feature sub-channels, loads allocated power and bits according to greedy algorithm till loading complete, and taking out the residual sub-band and power to reallocate to others.

The present invention provides an enablement and optimization system and method for web advertisement placement and pricing. The invention provides an open advertising network that incorporates both the advertisers and content publishers into the advertising decision process to create an open marketplace. Similar to a greedy algorithm in Computer Science, this model pushes the final advertising decision to the web servers that serve the actual content. Rather than an advertising network providing a single ad to show, the content publisher receives a selection of ads. Each publisher in the network then makes the local optimum choice by collecting data on the effectiveness of ads previously shown, and using this data to select which ad will perform the best. Similarly, the advertisers collect data on the visitors they receive from advertising, and use this information to optimize their bidding. With each endpoint selecting a local optimum, the network of advertisers and content publishers as a whole becomes more effective.

The invention discloses an underwater acoustic sparse channel estimation variable step sparsity adaptive matching tracking method, which fully utilizes the underwater acoustic channel sparse multipathcharacteristic and avoids the waste of frequency spectrum resources caused by the excessive number of pilots in the traditional channel estimation technology. The method does not need sparseness as apriori information, and the size of the support set is the estimated sparseness at the end of iteration by expanding the support set through step size. In addition, the signal reconstruction processis divided into several stages by combining stage idea and variable step size, the number of atoms in the support set in a certain phase remains constant, and the adjacent phases gradually expand thesupport set by different step sizes. The invention improves the recovery accuracy on the premise of not significantly increasing the calculation amount, that is, obtains a better trade-off between thereconstruction accuracy and the calculation complexity. Compared with the prior classical greedy algorithm, the invention does not need sparseness as a prior information, and the step size adaptive change can give consideration to the algorithm accuracy and the operation efficiency.

The invention discloses a method and a device for route optimization of logistics delivery vehicle, and belongs to the technical field of logistics. The method comprises the following steps of: initializing a congestion matrix alpha and a distance matrix D, generating a delivery route weight matrix omega=alpha D, and initializing a population module N<ZQ>; selecting a population size N<X>, a maximum number of generations N<G>, a crossing-over rate beta, a mutation rate gamma and a number of generations n=0, generating an initial route r1 through a greedy algorithm, and performing mutation operation on the initial route r1 to generate N<ZQ>-1 new routes; calculating fitness A<n> of each route of a first generation population formed by the initial route and the new routes, selecting N<X> routes with the highest fitness from the current population by adopting selection operators, and performing crossover and mutation operations on the N<X> routes to generate a population of next generation; updating n=n+1, when n=N<G>, calculating the fitness A<n> of all the routes in the latest population, and selecting the delivery route with the highest fitness in the current population as the optimal route. According to the invention, when the logistics delivery vehicle delivers goods, the delivery time can be as less as possible, and the delivery route can be as short as possible.

A recognition ambiguity resolution method of a Chinese named entity belongs to the field of named entity extraction; the method comprises the following steps: (1) the invention adopts a feature induction to reduce a feature number, i.e., a classifier provided by the invention automatically selects a meaningful feature through training and learning; (2) after selecting the feature, an N-BEST best marking sequence is selected from the CRF (Conditional Random Field) model learning through a Viterbi algorithm, i.e., N marking sequences with maximum probability of an observation sequence are selected; (3) in consideration of an occurrence frequency and a word length of the Chinese named entity, the method disclosed by the invention adopts a modified greedy algorithm to implement the ambiguity resolution, thus the entity marking sequence is obtained.