1052 results about "Algorithm optimization" patented technology

The invention discloses a compressed sensing-oriented block-sparse signal reconfiguring method, and particularly relates to a block-sparse signal reconfiguring algorithm, which aims to solve the problems that the optimization complexity of a mixed l2 / l1 optimization algorithm in the conventional block-sparse signal reconfiguring algorithm is relatively higher and that overmatching phenomenon is easily caused by a block-sparse matching pursuit algorithm or a block-sparse orthogonal matching pursuit algorithm. The method of the invention comprises the following steps of: correcting labels, in ameasurement matrix, of column vectors of a recovery matrix calculated in the iteration operation of the (l-1)th time by performing the iteration of the lth time, and for a block-sparse signal x with the block sparsity of K, reconfiguring the block-sparse signal x by performing the iteration for not more than K times. The method is applied to the reconfiguration of the block-sparse signal, particularly to the reconfiguration of a binary block-sparse signal.

The invention discloses a method for establishing virtual reality excavation dynamic smart load prediction models. The method includes the steps that the knowledge excavation technology is adopted so that a virtual reality analysis environment can be formed, the influence relation between fixed quantities is explored, and an input variable candidate set is determined; smart load prediction models of a support vector machine of a self-adaptive structure and an Elman neural network and the like are established, wherein input variables are determined by the support vector machine through the attribute screening technology and parameters are optimized by the support vector machine through a flora tendency differential evolutionary algorithm; a region load smart load prediction model based on data slice excavation is established; a load curve prediction model combined with dynamic electrovalence factors, user characteristics and the user response electric quantity is established, so that linked correcting prediction of loads, electrovalence and the response electric quantity is achieved. According to the method, the prediction models suitable for the actual condition of a smart power grid of China are established, the scale of construction of renewable energy sources is reasonably planned, more efficient power utilization of users is facilitated, and reasonable arrangement of power supply resources of power enterprises is facilitated.

The invention relates to a method for planning robot paths on the basis of path expansion ant colony algorithms. The method has the advantages that the ant colony algorithms are applied to the field of robot path planning, path expansion ant colony algorithm optimization strategies are proposed, the robot path optimizing efficiency can be optimized, information element distribution time-varying characteristics, information element updating strategies, path location inflection point optimization and local optimal path expansion are introduced, and location inflection point parameters and general evaluation are additionally used as evaluation standards for the paths; as verified by simulation analysis and practical experiments on the three algorithms, the method is high in robot path planning and searching capacity on the basis of the path expansion ant colony algorithm optimization strategies and is high in algorithm efficiency, and the found paths are short; phenomena that the algorithms run into local optimization can be effectively inhibited, the optimal paths of robots can be searched, and the robots can quickly avoid obstacles to safely arrive at target points.

A propylene polymerization production process optimal soft-measurement meter based on genetic algorithm optimized BP neural network comprises a propylene polymerization production process, a site intelligent meter, a control station, a DCS databank used for storing data, an optimal soft measurement model based on genetic algorithm optimized BP neural network, and a melting index soft-measurement value indicator. The site intelligent meter and the control station are connected with the propylene polymerization production process and the DCS databank; the optimal soft-measurement model is connected with the DCS databank and the soft-measurement value indicator. The optimal soft measurement model based on genetic algorithm optimized BP neural network comprises a data pre-processing module, an ICA dependent-component analysis module, a BP neural network modeling module and a genetic algorithm optimized BP neural network module. The invention also provides a soft measurement method adopting the soft measurement meter. The invention can realize on-line measurement and on-line automatic parameter optimization, with quick calculation, automatic model updating, strong anti-interference capability and high accuracy.

The invention discloses an image marking method based on multi-mode deep learning. The method comprises the following steps: firstly, a depth neural network is trained by utilization of images without labels; secondly, each single mode is optimized by utilization of counter propagation; finally, weights among different modes are optimized by utilization of on-line learning power gradient algorithm. The method employs a convolution neural network technology to optimize parameters of the depth neural network, and the marking precision is raised. Experiments of public data sets show that the method can raise the image marking performance effectively.

The invention discloses a quick feedback analyzing system in a tunnel constructing process. The system adopts a scheme: understanding currently adopted designing construction parameters; establishing a tunnel excavation three-dimensional finite element numerical grid calculation model; acquiring surrounding rock layering and convergent displacement monitoring information after a tunnel is excavated; establishing a non-linear support vector machine model; fixing an anchoring parameter according to the actual construction parameter, and optimally identifying rock mechanic parameters by adoptinga differential optimization algorithm; optimizing the construction parameter of an anchoring scheme by adopting a differential evolution algorithm; and optimizing the rock mechanic parameters by calling the differential evolution and optimization algorithms to further solve the construction parameter of the anchoring scheme, and outputting the construction parameter of the optimized anchoring scheme as a construction scheme through a computer display screen to guide the constructors to construct. The quick feedback analyzing system ensures that the monitoring information is used for optimizing the anchoring parameter while being used for identifying the surrounding rock parameters, so that the dynamic information construction is improved to a level of quantitative analysis.

The invention discloses a pH (potential of hydrogen) value predicting method of a BP (back propagation) neutral network based on a simulated annealing (SA) algorithm optimization. The pH value predicting method comprises the following steps: step one, selecting a sample according to a sample selecting strategy and inputting; step two, according to the BP theorem, determining the structure of the BP neutral network; step three, according to a network training strategy, applying the simulated annealing algorithm to optimize the BP network weight parameter; training the BP network by using the input sample, and determining the optimal weight and optimal hidden node number of the BP network; step four, according to the well trained BP neutral network, structuring a predicting model of the pH value. The pH value predicting method overcomes the randomness of the BP network in terms of weight selection, improves the rate of convergence and study ability of the BP neutral network. Besides, the method optimizes the selection of the training sample and the network hidden neutral element number, and improves the generalization ability of the BP neutral network. Moreover, the pH value predicting method is high in predicting accuracy of pH value and good in nonlinear fitting ability.

The invention relates to a vehicle type identification method based on a support vector machine and used for an earth inductor. The vehicle type identification method includes the following steps: vehicle type waveform data which require to be identified are collected by the earth inductor; a plurality of numeralization features are extracted from waveforms, effective data are screened out, and the features are normalized; multilayer feature selection is performed according to the extracted features, and an optimal feature combination is picked out; a vehicle type classification algorithm based on the clustering support vector machine is built, and parameters in a classification function are optimized by adopting a particle swarm optimization algorithm; a binary tree classification mode is built, classifiers on all classification nodes are trained, and a complete classification decision tree is built; and earth induction waveforms of a vehicle type to be identified are input to obtain identification results of the vehicle type. The vehicle type identification method builds a waveform feature extraction and selection mode, simultaneously adopts the classification algorithm based on the support vector machine and the particle swarm optimization algorithm, greatly improves machine learning efficiency, and enables a machine to identify vehicle types rapidly and accurately.

The invention discloses an electrocardiogram classification method based on a deep learning model, and the method is characterized in that the method comprises the steps: data obtaining, data processing, model construction, algorithm optimization, and model training. A technical problem to be solved in the invention is to carry out the discrimination of arrhythmia through electrocardiogram data, to provide assistance and reference for a doctor, and to solve problems that doctors are not sufficient in some places and the wrong diagnosis and diagnosis leakage rate are higher. According to the embodiment of the invention, the invention has the following beneficial effects that the method employs the deep learning method, and achieves the discrimination of arrhythmia in the electrocardiogram information through the building of a large-scale convolution neural network. Compared with the conventional model, the method saves the cost, and is higher in accuracy.

A novel and useful mechanism for the skinning of 3D meshes with reference to a skeleton utilizing statistical weight optimization techniques. The mechanism of the present invention comprises (1) an efficient high quality linear blend skinning (LBS) technique based on a set of skeleton deformations sampled from the manipulation space; (2) a joint placement algorithm to optimize the input skeleton; and (3) a set of tools for a user to interactively control the skinning process. Statistical skinning weight maps are computed using an as-rigid-as-possible (ARAP) optimization. The method operates with a coarsely placed initial skeleton and optimizes joint placements to improve the skeleton's alignment. Bones may also be parameterized incorporating twists, bends, stretches and spines. Several easy to use tools add additional constraints to resolve ambiguous situations when needed and interactive feedback is provided to aid users. Quality weight maps are generated for challenging deformations and various data types (e.g., triangle, tetrahedral meshes), including noisy, complex and topologically challenging examples (e.g., missing triangles, open boundaries, self-intersections, or wire edges).

The invention relates to an improved artificial fish school optimization method based on vehicle path planning. The existing algorithm has the defects of large calculation amount and long time consumption. The invention effectively combines the bionics principle of the artificial fish school and the subjective bias of a decision maker, improves the artificial fish school algorithm, and introduces the concept of a fish school sensing range for reconstructing an optimization formula of the artificial fish school algorithm by a repulsion region, a neutral region and an attracting region. The moving step length of the artificial fish, the view field range and the adjacent field value are dynamically regulated through judging the demands of nodes on the vehicle path, the reset transportation capability of the current vehicles, and the subjective bias of the decision maker, so the overall searching capability and the searching speed of the artificial fish school algorithm can be improved. Finally, the improved artificial fish school algorithm is used for improving the cargo taking and cargo sending behaviors of the vehicles and completing the path dispatching problem of return trip cargo taking vehicles. The invention is superior to the traditional optimization algorithms such as the genetic algorithm, the simulation annealing algorithm and the like in aspects of calculation precision and stability, and has good optimization capability.

The invention, which belongs to the field of intelligent algorithm optimization, provides a reinforced learning robot path planning algorithm based on a potential field in a complex environment, thereby realizing robot path planning in a complex dynamic environment under the environmental condition that a large number of movable obstacles exist in a scene. The method comprises the following steps:modeling environment space by utilizing the traditional artificial potential field method; defining a state function, a reward function and an action function in a Markov decision-making process according to a potential field model, and training the state function, the reward function and the action function in a simulation environment by utilizing a reinforcement learning algorithm of a depth deterministic strategy gradient; and thus enabling a robot to have the decision-making capability of performing collision-free path planning in a complex obstacle environment. Experimental results showthat the method has advantages of short decision-making time, low system resource occupation, and certain robustness; and robot path planning under complex environmental conditions can be realized.

A novel analog circuit early fault diagnosis method includes the steps of (1) acquiring time domain response signals of an analog circuit and taking the time domain response signals as output voltage signals of the analog circuit; (2) performing wavelet transform to the acquired voltage signals; (3) performing fractal analysis to original signal patterns and wavelet sub patterns to generate wavelet fractal dimensions of different patterns; (4) performing kernel entropy component analysis to candidate feature vector data composed of the wavelet fractal dimensions to acquire low-dimension feature vector data; (5) creating a multi-class classifier of a least squares support vector machine, and optimally selecting penalty factor and width factor of the least squares support vector machine, which are used for distinguishing overlapped early fault categories, by a quantum-behaved particle swarm optimization algorithm; and (6) sending the low-dimension feature vector data into the multi-class classifier of the least squares support vector machine and then outputting early fault diagnosis results. The novel analog circuit early fault diagnosis method can effectively detect early faults of analog circuits.

The invention provides a method for predicting an immune neural network of a cigarette sensory smoke panel test and smoke indexes. The method comprises the following steps: detecting analysis indexes and storing the obtained data into a database; normalizing sample data; dividing single material cigarettes and cigarette products into a plurality of groups according to styles; finishing the final ranking of all the single material cigarettes and cigarette products in the database; establishing corresponding immune neural networks for N physical and chemical index samples of different types of the single material cigarettes of the cigarette products respectively; sending the normalized sample data into the corresponding immune neural networks; using an immune algorithm to optimize the network weight and structure until ranking targets are reached; storing the network weight and the structure in a knowledge base; and judging whether the sample data to be analyzed is new and unclassified. The method has the functions of higher ranking accuracy, more accurate mapping relations between the physical and chemical indexes of all kinds of cigarettes and the sensory smoke panel test and the smoke indexes, fewer manual smoke panel tests and detecting times, auxiliary formulation design, and improvement of work efficiency.

The invention discloses a hydraulic aerial cage operation platform trajectory control device. The control device comprises an operation mechanism (A), a detection device (B), a display and alarm device (C), a hydraulic operation loop (D) with pressure compensation, a coordinate location module (E), a forward resolving module (F), a deflection compensation module (G), a speed setting module (H), an algorithm optimization module (I), and a programmable controller (J). The control device provided by the invention can significantly improve the work efficiency of the hydraulic aerial cage, reduce the labor intensity of operators, and lower the energy consumption as well as the use cost.

The invention relates to a Markov chain and neural network based traffic congestion state combined prediction method. The Markov chain and neural network based traffic congestion state combined prediction method comprises the following steps of 1 adopting a similar-PageRank Markov chain method to perform traffic congestion state prediction so as to obtain a first prediction result, 2 adopting a quantum multi-agent algorithm optimized back-propagating (BP) neural network method to perform traffic congestion state prediction so as to obtain a second prediction result, 3 obtaining the weight of the first prediction result and the weight of the second prediction result based on information entropy, 4 obtaining a final prediction result according to the first prediction result, the second prediction result and the corresponding weights. Compared with the prior art, the Markov chain and neural network based traffic congestion state combined prediction method has the advantages of being good in prediction real-timeliness, high in accuracy, good in extension and the like.

The invention discloses a matching navigation method based on local gravity field approximation. The method comprises the following steps that: at first, a flight path indicated by an inertial navigation system in a period of time is acquired as an initial flight path for matching, a gravity abnormal value corresponding to the point of the flight path indicated by the inertial navigation system is acquired, grid data of a local square region are intercepted in a gravity anomaly reference graph through the flight path indicated by the inertial navigation system and a confidence interval thereof, and a function of a gravity anomaly graph of the square region is acquired; and then terrain contour matching algorithm is used for rough matching, difference pretreatment is carried out on a gravity measuring sequence, then rough matching is carried out by using the terrain contour matching algorithm, and matching results are used as an initial flight path for matching algorithm of a related extreme value; and a matching algorithm optimization interval of the related extreme value is acquired based on the rough matching results obtained through terrain contour matching and the confidence interval, an iterative random initial value is generated by an average random number and enters into BFGS for optimization, an optimal matching flight path is calculated based on optimal solution of BFGS optimization, and the original flight path is updated. The method provided by the invention enables available matching precision to be higher through construction of a local gravity field reference graph.

The invention provides an intelligent traffic flow congestion dispersal method through vehicle group autonomous cooperative scheduling. An intelligent vehicle-mounted terminal and a wireless communication module are integrated, a cooperative scheduling algorithm optimization index and vehicle driving parameter acquisition and processing process is built according to a vehicle following model in the case of traffic flow congestion, the optimization index reflects the vehicle group passing efficiency, the intelligent vehicle-mounted terminal and an adjacent vehicle build a wireless communication network, a guiding instruction is generated according to the cooperative scheduling algorithm and the protocol, a vehicle queuing position and estimated passing time are provided for a driver through a man-computer interface, and the driver is guided to select a proper speed and a proper direction. The method of the invention has the advantages of low cost, high terminal integration, simple mounting, high sensitivity, quick response speed and the like, and is applicable to various motor vehicles, such as freight cars, buses and cars, which run in traffic dense areas such as urban areas and freeways.

The invention discloses a method and a device for optimizing the goods loading of a container. According to the size and bearing power of the container, and related information about the lengths, widths, heights and weights of various case goods, pre-loading is performed, and search is optimized by an algorithm to find an optimal loading scheme, thereby improving the utilization rate of resources and the loading efficiency.

The invention discloses an efficient energy-saving optimizing method for numerical control milling processing process parameters based on a Taguchi method. The efficient energy-saving optimizing method comprises the following steps: analyzing characteristics of a numerical control milling processing process energy consumption time interval and establishing a numerical control milling processing energy efficiency function; designing orthogonal experiment by applying an orthogonal table in the Taguchi method, and adopting signal to noise ratio evaluation to obtain an interference relationship between process parameters and processing time as well as specific energy consumption; obtaining a regression equation of each target by adopting a response surface process, and establishing an efficient energy-saving multi-target optimizing model for numerical control milling processing process parameters; and searching out Paretro optimal solution through a particle swarm optimization. Correlation relationship of specific energy, processing time and process parameters in the processing process is analyzed through experimental data and algorithm optimizing results, so that a complex coupling mechanism between energy consumption efficiency and process parameters is disclosed in the numerical control milling processing process.

The invention discloses a health prediction method and system for a new energy vehicle battery. The method comprises the following steps that: carrying out data analysis processing on vehicle data obtained in real time to obtain vehicle working condition data; independently executing data cleaning, data conversion and data reduction processing on the vehicle working condition data; on the basis of the vehicle working condition data subjected to data preprocessing, adopting a factor analysis method to extract data which influences a battery health degree, adopting a supervised learning method to mine a potential relationship between the data which influences the battery health degree and the vehicle working condition data, and constructing an initial battery health prediction model; carrying out model evaluation and algorithm optimization on the initial battery health prediction model to obtain an optimal battery monitoring prediction model, and finishing battery health prediction under a practical working condition. By use of the method, the dynamic prediction of the health state of the new energy vehicle battery is realized, the dynamic property and the economy of the vehicle can be improved, and the method has the advantages of being simple in operation and easy in implementation.

The present invention discloses a three-dimensional visualization engine and a WEB application calling method based on the BIM model. The method comprises: carrying out geomorphic modeling according to the digital elevation model and the terrain topography; according to the different modeling software, carrying out model uploading and registration based on the plug-in; optimizing the model resource according to the algorithm, calculating the multi-level LOD in real time, and storing the model resource; carrying out online transmission and offline download on three-dimensional scene data from a server to a client in a cloud+end mode; and supporting the model interaction on the client, and ensuring the scene to be smoothly displayed through scene optimization. According to the technical scheme of the present invention, a smooth three-dimensional visualization real-time rendering engine can be provided for the model with increasing volume at present, contradiction between the model volume and the client computing performance can be solved, and the technical scheme has significant value, and economic and social benefits to realize building information visualization and information collaborative sharing in the whole life cycle and to realize model efficiency maximization.

According to the invention, a probability integral method parameter prediction model for optimizing the BP neural network based on a combined algorithm (GP) of a genetic algorithm and a particle swarmoptimization algorithm, and an input layer of the BP neural network is optimized by adopting an average influence value algorithm (MIV), so that the complexity of the network is reduced, and the purpose of improving the prediction precision is achieved. An MIV-GP-BP model is established by taking actual measurement data of 50 working surfaces as a training set and a test set of the BP neural network, and the precision and reliability of a model prediction result are analyzed; the results show that: in five parameters, the root mean square error ranges from 0.0058 to 1.1575; the maximum relative error of q, tan beta, b and theta is not more than 5.42%, the average relative median error is less than 2.81%, the s / H relative error is not more than 9.66%, the average relative median error is less than 4.31% (the parameter itself is small), and the optimized neural network model has higher prediction precision and stability.

The invention discloses a binocular stereo matching method and a binocular stereo matching system based on dense network depth learning. Firstly, stereoscopic image pairs with real values are preprocessed to construct training samples for network learning. Then a dense convolution neural network model is trained to calculate the matching cost. Finally, cross-aggregation algorithm is used to optimize the matching cost, and the optimal disparity is calculated by WTA strategy, and the corresponding disparity map is obtained. A method for extracting features of convolution layer features that a dense convolution neural network model is used to extract features of convolution lay, Attempt to find a better matching cost, effectively solve the problem that pixel matching point can not be found accurately in weak texture region and the performance of detail features is poor, with higher computational accuracy and better robustness.

The invention discloses an optimization method for a track playback function in an indoor positioning system, wherein the optimization method comprises the following steps of: S1, importing an indoor CAD (Computer-Aided Design) map into a monitoring center of the indoor positioning system; S2, acquiring the boundary, the scale and the coordinates of obstacles of the imported current map by the monitoring center automatically; S3, identifying the obstacles by the monitoring center, formatting the current map area and optimizing the specification scenery color of the map; S4, recording and storing the track coordinate data of the electronic tag movement, carried by a positioned object, by the monitoring center via a communication network; S5, optimizing the recorded track data by the monitoring center by adopting an A* algorithm and recording optimal obstacle avoidance data; and S6, drawing the track point and the trajectory by the monitoring center according to the optimized track data. Compared with the prior art, the invention has the advantages of greatly improving the performance of track playback, and has the obvious optimal choice of a track path and an intelligent obstacle avoidance function.

The invention discloses a self-learning wheel chair control method based on change of a gravity center of a human body, and belongs to the field of pattern recognition and intelligent systems. According to the self-learning wheel chair control method, a pressure sensor is installed between a wheel chair seat and a framework so as to collect force distribution under a sitting position of the human body, two-dimensional areal coordinates are calculated, and real-time data of the center of the gravity are stored in an embedded type computer; and algorithm optimization is conducted to the number of neurons in an output layer, network initial weight value, a network neighborhood radius adjusting rule and the like according to a basic learning process of a normal self-organizing feature map (SOFM) algorithm, and therefore operating complexity is reduced, calculating instantaneity of the algorithm in application is improved, and the purpose that algorithms are controlled to be different according to difference of people is achieved. By utilizing the improved SOFM algorithm, and in the process of driving habit learning, rate of convergence of an SOFM clustering algorithm and learning efficiency are greatly improved, instantaneity of the algorithm and accuracy of cluster are improved, the requirement of wheel chair real-time learning and controlling is met, and the problem that manual parameter adjustment is fussy due to difference of driving habits of users is solved.

The invention discloses a method for optimizing the fault prediction and diagnosis of a wind power plant unit mainshaft bearing of an SVM (Support Vector Machine) on the basis of a Cuckoo algorithm. Historical moment sampling data is subjected to principal component characteristic extraction to establish an SVM model, and the Cuckoo algorithm is used for optimizing the performance parameter of the SVM. After the sampling data which contains fault information is subjected to real-time prediction, an expert system makes an effective fault diagnosis, and a diagnosis result is presented on a human-computer interaction interface. PCA (Principal Component Analysis) is used for carrying out dimensionality reduction on the data, classification accuracy is improved, and the training time of the classifier is greatly shortened. Meanwhile, compared with other traditional optimizing methods, the Cuckoo algorithm has the advantages that a global optimal value is obtained by quick convergence, and has an obvious advantage on an aspect of prediction accuracy, and a guarantee is provided for the expert system to accurately obtain a diagnosis result.

The invention belongs to the technical field of computer vision, and particularly relates to an image stitching real-time performance optimization method. The method comprises the following steps of image NCC region matching, SURF (speeded up robust feature) threshold value estimation and feature point matching. Under the condition of precisely solving the transformation matrix, through algorithm optimization, the detection feature point number is greatly reduced; meanwhile, the image overlapping region size is pre-estimated through a local region matching algorithm NCC; the feature point search range in the image stitching process is reduced through locking the overlapping region. The NCC algorithm obtains a cross-correlation maximum value window for estimating the approximate matching condition of the image; the feature point finding among local images is avoided; the real-time performance of the image stitching is improved through the overlaying use of two methods.

The invention provides a fractal image generation and rendering method based on a game engine and CPU parallel processing, relates to the technical field of CPU parallel processing in an image game engine, and aims to solve the technical problems in the prior art that image data information division is unreasonable, the rendering performance and the efficiency are low due to limited arranging structures of commands such as parallel processing and synchronous operation, and the performance requirements of real-time rendering cannot be met. In addition, the invention provides a new technical problem which specifically is how to realize highly-efficient and high-quality image rendering by simultaneously using more than two fractal algorithms in parallel. A software rendering generation algorithm is constructed based on fractal characteristics; and the CPU parallel processing technology is utilized to process fractal image geometrical characteristics in real time, a control flow instruction parallel algorithm is utilized to optimize rendering generated pipe lines, rapid generation and rendering of a fractal image model are realized, and a rapid and accurate display effect in a user PC machine is achieved.

A computer-based airplane flight path planning method based on the pigeon-inspired optimization (PIO) algorithm includes steps of establishing an uncertainty prediction model, determining the path to be optimized, and obtaining an optimal path using the PIO algorithm for a flight controller onboard to execute. The PIO algorithm treats a pigeon flock as a scale-free network, applies map and compass operators to the scale-free network, and performs landmark operations to obtain the optimal path. The device that performs the path planning includes an access module for obtaining the regional environment information and a flight controller onboard the airplane. The flight controller includes a building module for setting up the trajectory prediction model including uncertainties; a determining module to determine the trajectories which need optimization; an optimization module, which uses the PIO algorithm to optimize the flight path; and a computer memory module.