Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

193results about How to "Avoid local optima" patented technology

Secure computing unloading method for cache-assisted ultra-dense heterogeneous MEC network

The invention discloses a cache-aided ultra-dense heterogeneous MEC (Mobile Edge Computing) network-oriented secure computing unloading method, which comprises the following steps of: acquiring network basic information of an ultra-dense multi-access mobile edge computing network, constructing a network system according to the network basic information, and constructing an optimization problem under the constraint of the network system; obtaining an initial solution according to the optimization problem, defining the initial solution as an initial population, searching the initial population by adopting an improved whale algorithm to obtain a target population, and outputting the position of the globally optimal whale in the target population; and executing safety calculation efficiency optimization configuration according to the global optimal whale position. According to the method, multi-base-station unloading is achieved, cache factors are considered, rate, mobile equipment energy consumption, time delay, security vulnerability cost and task and transmitting power proportion constraints are met, and the target of minimizing local total energy consumption of all mobile equipment can be well achieved.
Owner:EAST CHINA JIAOTONG UNIVERSITY

Battery safety management system and method

The invention discloses a battery safety management system and method, and relates to the technical field of battery management. After multiple times of loop iteration based on an optimization algorithm, an optimal charging strategy is output for a battery, the battery is divided into a plurality of monitoring areas in the charging process of the battery, and after multiple items of data of the monitoring areas are acquired in real time, the battery safety management is realized. And substituting multiple items of data into a TOPSIS algorithm to calculate a relative close index of each monitoring area, performing comprehensive weighting calculation on the relative close index of each monitoring area according to the weight coefficient, predicting whether the battery is abnormal in the charging process, and generating a corresponding management strategy based on a prediction result. The management system can dynamically adjust the charging strategy in combination with the historical charging data of the battery in the charging process of the battery, the charging efficiency of the battery is ensured while the charging safety of the battery is ensured, and the charging abnormity of the battery is predicted, so that the battery can be processed in time before the abnormity occurs, and the safety of the battery is further ensured.
Owner:SHANGHAI HAOZHE ELECTRONIC TECH CO LTD

Photovoltaic array fault diagnosis method based on Transform-BiLSTM model

PendingCN119416052ADiversified search pathsvalid updatePhotovoltaic energy generationGenetic algorithmsAlgorithmPhotovoltaic arrays
A photovoltaic array fault diagnosis method based on a Transform-BiLSTM model relates to a photovoltaic array fault diagnosis method, and comprises the following steps: S1, inputting photovoltaic array sample data under normal and different fault working conditions, carrying out standardized conversion on the sample data, and dividing a training set and a test set through VMD modal component decomposition; s2, initializing parameters of a Transform-BiLSTM model and parameters of an IIVYA algorithm, and setting the number of hedera helix populations, the upper and lower limits of independent variables, the maximum number of iterations and the dimension size; s3, initializing a hedera helix population; s4, calculating the fitness value of the new position of the hedera helix according to the position updating formula of each hedera helix population; s5, judging whether the number of iterations reaches the maximum or not, comparing the optimal fitness values stored in all iterations, and outputting optimal parameters; and S6, constructing an optimal Transform-BiLSTM model by using the optimal parameter value obtained by the IIVYA algorithm, and carrying out diagnosis test on the photovoltaic array fault.
Owner:SHENYANG INSTITUTE OF CHEMICAL TECHNOLOGY

Flexible continuum robot motion control method based on improved angle and exendinium optimization algorithm

The invention discloses a flexible continuum robot motion control method based on an improved Kerr-Ren optimization algorithm. The method comprises a novel flexible continuum robot 3D printing piece structure and a closed-loop model-free control method based on the improved Kerr-Ren optimization algorithm. According to the method, the improved angle exenx optimization algorithm is introduced, the Henon chaotic mapping theory, the linear adaptive weight and the mirror reflection learning model are introduced on the basis of the original algorithm, the diversity of population initialization is increased through improvement, the optimization precision of the algorithm is improved, and the optimization precision of the algorithm is improved. The global search capability and the local search capability of the algorithm in each stage are balanced, the convergence speed is effectively improved, and the algorithm is prevented from falling into local optimum; through combination of the improved angle exid optimization algorithm and the closed-loop model-free control method provided by the invention, path planning and attitude adjustment of the novel flexible continuum robot are realized, and a guarantee is provided for high-precision control of the novel flexible continuum robot.
Owner:CHINA UNIV OF MINING & TECH

Safety engineering optimization method and system based on artificial intelligence

The invention discloses a safety engineering optimization method and system based on artificial intelligence, and belongs to the technical field of safety engineering optimization, and the method comprises the steps of data preparation, data preprocessing, elevator running state evaluation, elevator fault prediction and elevator safety engineering optimization. According to the method, the elevator running state is evaluated by adopting the long-short-term memory model for enhancing parameter optimization, so that the sparrow population diversity is increased, the global search capability of the algorithm is improved, the parameter optimization is better performed, and the model prediction effect is improved; an auto-encoder long-short-term memory parallel model is adopted for elevator fault prediction, through parallel processing of an auto-encoder and a long-short-term memory network, rich feature representation is better captured, the model prediction precision is improved, integrated learning is carried out through a gradient lifting technology, and the model robustness is enhanced; the elevator running state evaluation information and the elevator fault prediction information are integrated, elevator safety engineering optimization is achieved, the influence of subjective factors is reduced, and the overall stability and reliability of an elevator system are effectively improved.
Owner:QUANLIAN FUJI ELEVATOR (FOSHAN) CO LTD

Verification method and device based on genetic algorithm

The embodiment of the invention provides a verification method and device based on a genetic algorithm, and the method comprises the steps: randomly selecting a plurality of groups of initial key parameters from a preset key parameter value range, and calculating the fitness of each group of initial key parameters in a processor verification process, processing the plurality of groups of initial key parameters according to the fitness, a preset selection probability, a preset crossover probability and a preset mutation probability to finally obtain a plurality of groups of target key parameters, taking the plurality of groups of target key parameters as the plurality of groups of initial key parameters, executing the steps again until a preset termination condition is met, and stopping iteration to obtain a plurality of target key parameters; and calculating and outputting a processor verification coverage rate according to the plurality of groups of target key parameters and the key parameter value range. The target key parameters are obtained based on the genetic algorithm, local optimum is avoided, the parameters are continuously optimized in the iteration process, the high processor verification coverage rate can be obtained, and the processor verification efficiency and quality are improved.
Owner:SHANDONG YUNHAI GUOCHUANG CLOUD COMPUTING EQUIP IND INNOVATION CENT CO LTD

Photovoltaic power short-term probability prediction system and method

The invention discloses a photovoltaic power short-term probability prediction system and method. The system comprises a convolutional neural network, a sequential network and a probability density estimation module. The time sequence network is composed of a self-organizing mapping neural network, a time sequence encoder and a time sequence decoder; each of the time sequence encoder and the time sequence decoder is formed by alternately connecting an autocorrelation mechanism unit, a time sequence decomposition unit and a feed-forward layer; the time sequence decoder is composed of an upper branch sub-layer and a lower branch sub-layer; the convolutional neural network screens meteorological factors to obtain a high-dimensional mapping relation between meteorological characteristics and optical power; the sequential network performs progressive decomposition and autocorrelation extraction on an optical power sequential sequence to obtain periodicity and tendency characteristics of the optical power sequence; the probability density estimation module optimizes and combines the high-dimensional mapping relation with periodicity and tendency characteristics of the optical power sequence to obtain optical power probability distribution parameter output; according to the method, the numerical weather forecast information and the weather type forecast published by the meteorological observatory are integrated to realize optical power probability prediction, so that the risk of misinformation of a single meteorological source is effectively reduced.
Owner:TIANJIN UNIV

Mine car optimization scheduling method and system based on improved PSO-ACO

PendingCN119067281ASolve the technical problems of low scheduling efficiencyRealize evaluationData processing applicationsInternal combustion piston enginesLearning factorOptimal scheduling
The invention discloses a basically improved PSO-ACO mine car optimization scheduling method, which comprises the following steps of: firstly, based on the actual transportation condition of a mine, establishing a mine car total transportation cost minimization mathematical model comprising the vehicle fixed cost, the time window cost, the oil consumption cost, the carbon emission cost and various constraint conditions; and then, taking the model as a fitness evaluation function of a PSO algorithm, carrying out random disturbance adaptive adjustment on a group learning factor, an individual learning factor and an inertia weight in the PSO algorithm according to the number of iterations and the vehicle scale, and solving a mine car total transportation cost minimization mathematical model by using the improved PSO algorithm. And finally, the optimal solution obtained by the improved PSO algorithm is converted into the enhancement quantity of the initial information of the ACO algorithm, the state transition rule and the pheromone updating mode of the ACO algorithm are improved in combination with the actual transportation condition of the mine, and then the improved ACO algorithm is used for solving the mathematical model for minimizing the total transportation cost of the mine car. Therefore, optimal scheduling of the mine cars in the mining area is realized.
Owner:WUHAN INST OF TECH +2

Task scheduling method under heaven and earth edge computing network

The invention discloses a task scheduling method under a heaven and earth edge computing network, and relates to the technical field of network communication and virtualization, and the method comprises the following steps: taking a scheduling sequence vector and a scheduling position vector of all subtasks in a dependent task as optimization variables; constructing a dependency task scheduling problem model of the target system model by taking the minimum average execution time delay of the dependency task and the average energy consumption of the unmanned aerial vehicle as optimization targets; each unmanned aerial vehicle in the target system model collects a dependency task within a time interval, and the dependency tasks collected by the unmanned aerial vehicles are decomposed into a plurality of sub-tasks with dependency relationships; and the improved multi-objective evolutionary algorithm based on decomposition is utilized to solve the dependency task scheduling problem model to obtain an optimal scheduling scheme, and the task scheduling quality under the sky-ground edge computing network is improved.
Owner:GUIZHOU UNIVERSITY OF FINANCE AND ECONOMICS

Multi-target optimization method for pressure and flow characteristics of CDC electromagnetic pilot overflow valve

The invention relates to a multi-target optimization method for pressure and flow characteristics of a CDC electromagnetic pilot overflow valve, and belongs to the technical field of automobile suspensions. The method comprises the following steps: (1) establishing a three-dimensional equivalent model of the CDC electromagnetic pilot overflow valve; (2) building a dynamic model according to the built three-dimensional equivalent model and a working principle; (3) building a hydraulic simulation model according to the kinetic model; (4) forming a training sample to train the neural network model; (5) initializing ant colony algorithm parameters; (6) starting ant colony algorithm iteration; (7) updating pheromones; (8) judging whether a termination condition is met or not; (9) outputting an optimal parameter combination and an optimized pressure-flow characteristic curve; and (10) determining the importance degree of the influence factors through a Pareto method. The ant colony algorithm and the annealing algorithm are combined, the local optimum problem is effectively avoided, and global convergence is accelerated.
Owner:SHANDONG UNIV

Calibration and repair management system of optical detection equipment and management method of calibration and repair management system

The invention discloses a calibration and repair management system and method for optical detection equipment, and particularly relates to the technical field of optical detection equipment calibration, and the method comprises the following steps: firstly, determining whether a current measurement error exceeds a standard or not through a standard test, and if the current measurement error exceeds the standard, calculating an environment change coefficient; acquiring a current measurement error, an environment change coefficient and calibrated times, performing fuzzy reasoning, selecting a proper calibration strategy, and executing calibration; if the measurement error still exceeds the standard after calibration, the related optical elements or control parts are repaired; repeating the steps until the measurement error meets a preset standard; the calibration strategy can be dynamically adjusted, different environment change conditions can be adapted, the requirement for frequent recalibration is reduced, the measurement precision of the optical detection equipment is ensured under different environment conditions, the measurement error of the equipment under the stable and dynamic environment can be comprehensively corrected, and the measurement accuracy of the equipment under the stable and dynamic environment can be further improved through a cyclic calibration mechanism. And high precision and reliability of the equipment in long-term use are ensured.
Owner:SHENZHEN KING TEAM PRECISION TECH +1

Film thickness determination method and training method of film thickness estimation model

The invention provides a thin film thickness determination method and a thin film thickness estimation model training method. The film thickness determination method comprises the following steps: acquiring initial reflection spectrum data of a to-be-measured film, the initial reflection spectrum data being associated with the film thickness of the to-be-measured film; performing data conversion on the initial reflection spectrum data to obtain target reflection spectrum data; inputting the target reflection spectrum data into a film thickness estimation model to obtain a film thickness interval of the to-be-measured film, the film thickness estimation model being obtained by training a candidate film thickness estimation model according to a simulation spectrum training set associated with the film thickness; and based on a preset optimization algorithm, the film thickness of the to-be-measured film is determined according to the film thickness interval, and the preset optimization algorithm comprises a global optimization algorithm based on a progressive non-convexity solver.
Owner:TIANJIN UNIV

Wave height prediction method for multi-strategy snow ablation optimization

The invention provides a wave height prediction method for multi-strategy snow ablation optimization, and belongs to the field of ocean engineering. The wave height prediction method comprises the following five steps: preprocessing sea condition data, performing feature extraction on the sea condition data by using a CNN model, training the model, optimizing parameters and evaluating the model. According to the method, the multi-strategy snow ablation optimization algorithm, the convolutional neural network, the bidirectional gating circulation unit and the multi-head attention mechanism are comprehensively utilized, and efficient processing and accurate prediction of the sea condition data are achieved; the introduction of the multi-strategy snow ablation optimization algorithm increases the diversity of the population, prevents the model from falling into local optimum, and improves the robustness of the algorithm. The CNN and the bidirectional GRU are combined, so that the model has excellent performance in the aspect of processing high-dimensional, nonlinear and time sequence data; the application of the attention mechanism can highlight important information, perform weighting processing on features, and further improve the prediction performance of the model.
Owner:CCCC FOURTH HARBOR ENG INST CO LTD

Wastewater treatment control strategy based on fuzzy neural network

PendingCN119087802AOvercoming slow convergenceincrease training timeAdaptive controlChemical oxygen demandWater quality
The invention discloses a wastewater treatment control strategy based on a fuzzy neural network, and the strategy comprises the steps: S1, collecting sample data of a wastewater treatment system, and dividing the sample data into a training set and a test set according to a fixed proportion; determining an input parameter and a pre-measurement parameter; s2, performing data normalization processing on the sample data; s3, constructing a PSO-BP-based wastewater treatment prediction model; S4, training the wastewater treatment prediction model through the training set sample data after data preprocessing; s5, performing performance evaluation on the wastewater treatment prediction model through test set sample data; s6, performing water quality prediction on the wastewater treatment system through the trained wastewater treatment prediction model; the problems that the cost is high and the real-time performance is poor when parameters of effluent chemical oxygen demand and inflow suspended solids of a sewage treatment system are monitored are solved, and the method has the advantages that the key parameters of water quality can be predicted through an improved neural network model, so that high-precision prediction data are obtained, and the real-time performance is high.
Owner:INNER MONGOLIA FUQING ENVIRONMENTAL TECH CO LTD +1

Communication interference strategy generation method based on multistage pruning lightweight network

The invention discloses a communication interference strategy generation method based on a multistage pruning lightweight network, and belongs to the technical field of artificial intelligence and reinforcement learning. The method comprises the steps that a PDQN-based multi-layer reinforcement learning network model is used for communication interference strategy generation modeling, multi-stage pruning is used for model lightweight so as to conform to deployment on a low-computing-power platform, the lightweight model comprises Q-Network and Actor, interference types and corresponding parameters are output respectively, the training speed is increased by adopting a dual-order threshold strategy, and the model weight is reduced. And finally, a reward function based on an adversarial strategy is constructed for guiding frequency alignment, and the model realizes the target bit error rate with less interference resources. Through the communication interference strategy generation method based on the multistage pruning lightweight network, intelligent cognitive interference of a communication countermeasure party to a communication party is realized, the deployment problem of the communication countermeasure party on a platform with relatively low computing power is solved through a lightweight method, and meanwhile, a plurality of updating acceleration methods are designed for the model, so that the learning efficiency is improved, and the communication countermeasure requirement is met.
Owner:THE 54TH RESEARCH INSTITUTE OF CHINA ELECTRONICS TECHNOLOGY GROUP CORPORATION

Pilot frequency power control method and system based on multi-strategy differential evolution

The invention relates to the technical field of wireless communication, and discloses a pilot power control method and system based on multi-strategy differential evolution, and the method comprises the following specific steps: S1, constructing an uplink training stage signal model; s2, initializing a multi-strategy differential evolution algorithm; s3, finding an initial optimal individual; s4, an unbiased variation evolutionary strategy is adopted, an optimal individual is selected, and the number of iterations is updated; s5, dividing into three sub-populations when the maximum number of iterations is reached, otherwise, repeating the step S4; s6, on the basis of a second-stage multi-strategy differential evolution algorithm, corresponding variation strategies are adopted for the three sub-populations, each individual is updated, and the number of iterations is updated; s7, if the maximum number of iterations is reached, selecting a current global optimal individual, otherwise, repeating the step S6; and S8, outputting the pilot frequency and the pilot frequency power coefficient corresponding to the pilot frequency. According to the method, the problems of premature convergence and local optimum during solving in the prior art are solved, and the method has the characteristic that the system throughput can be effectively increased.
Owner:CHINA ELECTRONICS TECH GRP NO 7 RES INST

AGV obstacle avoidance path planning method in unmanned warehousing system

The invention discloses an AGV obstacle avoidance path planning method in an unmanned warehousing system, and the method comprises the steps: employing nested Logistic mapping and Skew Tent mapping to generate a chaotic initialization population to replace a population generation mode of a conventional differential evolution algorithm, carrying out the Levy flight operation after the selection operation, and carrying out the global path search on an environment map, thereby achieving the optimal path planning. And obtaining an AGV global feasible path avoiding obstacles in the environment map, and when the AGV has an unknown obstacle on the feasible path, re-planning the local path by adopting a dynamic window method so as to enable the AGV to realize an obstacle avoidance process in an interval where the unknown obstacle is located. The improved differential evolution algorithm enhances the global search capability through chaos initialization and a Levy flight strategy, and the AGV can quickly respond to an unknown obstacle through a dynamic window rule. The adaptability and the accuracy of path planning are improved, and the obstacle avoidance capability and the safety of the AGV in a complex environment are enhanced at the same time.
Owner:MASCH TECH DEV CO LTD

Combined fleet cargo transportation optimization method and system based on data processing

The invention relates to the technical field of cargo transportation optimization, and discloses a combined fleet cargo transportation optimization method and system based on data processing, and the method comprises the following steps: S101, collecting cargo feature data and ship feature data, and respectively representing the data through first feature data and second feature data; s102, preprocessing the first feature data and the second feature data to obtain third feature data, fourth feature data and fifth feature data; s103, constructing a matching degree prediction model according to the improved graph neural network model, and outputting the matching degree of the cargo to the ship; the cargo-ship bipartite graph structure comprises cargo nodes, ship nodes, edges, initial features of the cargo nodes, initial features of the ship nodes and initial features of the edges; and S104, optimizing the allocation scheme according to the improved genetic algorithm to obtain an optimal allocation scheme. The efficiency, accuracy and flexibility of cargo and ship distribution are improved.
Owner:TIMES TIANHAI (XIAMEN) INTELLIGENT TECH CO LTD

Loess landslide detection method and system based on grade differentiation learning grey wolf optimization

The invention belongs to the technical field of image processing, and discloses a loess landslide detection method and system optimized by a multi-population grey wolf optimization algorithm based on grade differentiation learning. Performing data enhancement on the landslide data set by using a data enhancement generator; designing a grey wolf optimization algorithm model based on multi-level differential learning through the segmentation network model, and optimizing hyper-parameters in the segmentation network model by using an improved algorithm to realize adaptive adjustment and cooperative calculation of three hyper-parameters, namely a learning rate, an iteration number and a batch processing number, in the segmentation network; and finally carrying out loess landslide detection. According to the method, algorithm improvement is carried out, so that the balance capability of global exploration and local development is enhanced, and later diversity loss and local optimum are avoided. By combining the excellent semantic segmentation capability of VIG and the improved grey wolf optimization algorithm, the method can effectively overcome the feature extraction difficulty caused by factors such as geological structure and vegetation coverage, and significantly improves the accuracy of loess landslide recognition.
Owner:NORTHWEST UNIV

PID parameter estimation method based on MIWOA

The invention relates to a PID (Proportion Integration Differentiation) parameter estimation method based on MIWOA. The method comprises the following steps: initializing the current iteration times of the MIWOA; when the current number of iterations does not reach the maximum number of iterations, obtaining a new shrinkage factor and an adaptive weight according to the current number of iterations and the maximum number of iterations; obtaining an updated position updating model according to the new shrinkage factor and the adaptive weight; according to the updated position updating model, obtaining the updated position of each individual and the corresponding fitness; and processing each individual through a sparrow search algorithm inserter mechanism or a Cauchy Gaussian mutation strategy to obtain an optimal individual, and obtaining a globally optimal solution. Through the new shrinkage factor, the adaptive weight, the sparrow search algorithm inserter mechanism and the Cauchy Gaussian mutation strategy, the global search capability and the local optimization capability of the whale optimization algorithm are enhanced, local optimum is avoided, the convergence speed and the optimization precision are significantly improved, and the overall performance of the PID control system is improved.
Owner:XIDIAN UNIV

Wind power output analysis method based on WOA-K-means

The invention provides a wind power output analysis method based on WOA-K-means. A K-means algorithm is used as a basis, volatility and output efficiency are used as characteristic indexes to carry out dimension reduction processing on a daily output curve, an initial clustering center is optimized through global search capability of a WOA algorithm, a comprehensive difference distance suitable for processing the curve is obtained through shape and dimension fluctuation difference and generating capacity difference, and the comprehensive difference distance is used for the K-means algorithm to measure similarity of different samples. And finally, the clustering result is evaluated through the improved contour coefficient. Therefore, the problems that the traditional K-means algorithm is sensitive to the initial clustering center and the intrinsic characteristics of the daily output curve are difficult to reflect through the Euclidean distance are solved. The method has the advantages that the clustering result is high in accuracy and stability, the efficiency of the clustering algorithm is improved, the calculation speed of the clustering algorithm is increased, a systematic clustering analysis process is formed, and scientific and efficient clustering analysis of the wind power daily output data is achieved.
Owner:HUBEI QINGJIANG HYDROPOWER DEV +1

Insulated cable process data processing method

The invention provides an insulated cable process data processing method, and relates to the technical field of process parameter optimization. The method comprises the following steps: acquiring multiple groups of process operation parameters of a target cable and product performance parameters corresponding to each group of process operation parameters, extracting a process data set of each product performance parameter, and determining multiple target process parameters corresponding to each product performance parameter; performing data enhancement on each process data set to obtain expanded data sets, screening out a plurality of first data enhancement sets from the expanded data sets, and performing cross validation to determine a plurality of second data enhancement sets; and performing data expansion on the multiple groups of process operation parameters and the product performance parameters corresponding to each group of process operation parameters to obtain a process performance data set, and processing the process performance data set through the process parameter optimization model to obtain optimized process parameters of the target cable. According to the invention, the optimization efficiency of the process parameters of the insulated cable is improved.
Owner:SHIN YA WIRE & CABLE (SHENZHEN) CO LTD

Active-disturbance-rejection controller parameter optimization method based on two-stage training reinforcement learning

The invention discloses an active disturbance rejection controller parameter optimization method based on two-stage training reinforcement learning, and belongs to the technical field of computer system control. The problems that when an active disturbance rejection controller is adopted in an existing method, the parameter adjusting efficiency is low, parameter optimization cannot be guaranteed, and initial parameter selection is relied on are solved. According to the two-stage training method based on pre-training and advanced reinforcement learning, a speed error-based reward function is designed in the pre-training stage, an error threshold-based sectional reward mechanism is designed in the advanced training process, and the speed tracking process of the motor is divided into different intervals, so that the speed tracking accuracy of the motor is improved. The intelligent agent training is guided by different rewards in different intervals, performance indexes are decoupled, local optimization of parameters is effectively avoided, and parameter optimization of the active disturbance rejection controller is realized. The method can be applied to parameter optimization of the active disturbance rejection controller.
Owner:江淮前沿技术协同创新中心 +1

Food-borne germ classification method based on neural network structure search

The invention discloses a food-borne pathogen classification method based on neural network structure search, which comprises the following steps: acquiring surface enhanced Raman spectrum data of multiple seed pathogens corresponding to multiple food-borne pathogens, and preprocessing the surface enhanced Raman spectrum data; automatically designing a U-Net neural network model architecture according to the hybrid network; training the U-Net neural network model, and adjusting the weight of each module of the U-Net neural network model by using a block differentiable neural architecture search strategy; carrying out the sampling of the super network through the pre-trained and optimized weight distribution, and determining the optimal module combination of each layer of the U-Net neural network model; and constructing a new U-Net structure by using architecture information obtained by sampling, training and generating a classifier, and realizing classification of unknown food-borne germ samples. According to the method, the U-Net model architecture is automatically designed through the hybrid network, and the structure and parameters of the network can be adaptively optimized, so that the model has better performance when different subclasses of germs are classified, and the classification accuracy is remarkably improved.
Owner:CHINA JILIANG UNIV

Multi-target operation optimization method for combined cooling heating and power comprehensive energy system

The invention belongs to the technical field of energy system optimization, and provides a combined cooling heating and power comprehensive energy system multi-target operation optimization method, which comprises the following steps: by combining chaotic mapping and a cuckoo algorithm, introducing a dynamic adaptive step factor and a dynamic discovery probability, and introducing disturbance reverse learning, constructing a multi-target optimization scheduling model; by improving the cuckoo optimization algorithm, performing population initialization by utilizing Logistic chaotic mapping, enhancing the randomness and global search ability of the algorithm, introducing a dynamic adaptive step factor and a dynamic discovery probability, increasing the diversity and flexibility of bird nest initialization, improving the global search ability, and introducing disturbance reverse learning, the algorithm has the advantages of improving the robustness of the algorithm, and improving the robustness of the algorithm. The algorithm performance is fully improved, and falling into a local optimal solution is avoided; compared with a common micro-grid, the improved comprehensive energy power grid has the advantages that the economic benefit can be effectively improved, pollutant emission is reduced, and the total cost of the whole system is reduced to a certain extent.
Owner:TIANJIN POLYTECHNIC UNIV

Thermal power plant circulating water pump operation optimization method using multi-objective culture differential evolution algorithm, storage medium, equipment and computer program product

The invention discloses a thermal power plant circulating water pump operation optimization method using a multi-objective culture differential evolution algorithm, a storage medium, equipment and a computer program product. Comprising the steps of collecting historical frequency converter instructions in the thermal power plant circulating water system, circulating water pump parameters under the corresponding frequency converter instructions and condenser vacuum degrees in the thermal power plant circulating water system, determining the normal use range of the frequency converter instructions, and establishing a comprehensive objective function; and in the application range of the frequency converter instruction, with the purpose of minimizing the comprehensive objective function, optimizing the frequency converter instruction by using a multi-objective culture differential evolution algorithm, determining the optimal frequency converter instruction, issuing the optimal frequency converter instruction to obtain the optimal circulating water pump parameter and the optimal condenser vacuum degree, and operating the thermal power plant circulating water pump. The power generation capacity of the thermal power plant is improved on the basis of reducing the energy consumption of the thermal power plant circulating water system, so that the energy consumption is reduced, and the power generation cost is reduced.
Owner:JIANGSU GUOHUACHENJIAGANG POWER GENERATION CO LTD +2

Fan main shaft bearing fault diagnosis method based on IPSO-VMD

The invention relates to a fan main shaft bearing fault diagnosis method based on IPSO-VMD, and the method employs an improved particle swarm optimization algorithm to carry out the global optimization of the parameters of a VMD model, prevents the later parameters of the algorithm from falling into local optimum, increases a particle disturbance factor, avoids the random selection of the number K of modal components and a penalty factor alpha according to experience in the prior art, and improves the fault diagnosis precision. According to the method, the problems that the optimal parameter combination of the VMD model is difficult to obtain and the model diagnosis fails are solved, the improved particle swarm algorithm is adopted to optimize the VMD model through global optimization given parameter combination, and finally the optimal parameter combination is found out, so that a local optimal solution caused by a random parameter optimization algorithm in the prior art is avoided, and the generalization ability of the model is improved.
Owner:CENT CHINA BRANCH OF CHINA DATANG CORP SCI & TECH RES INST CO LTD

Passenger flow prediction method based on optimized ICEEMDAN and bidirectional time convolution network

A subway passenger flow prediction method based on an optimized ICEEMDAN-bidirectional time convolution network model comprises the following steps: collecting passenger number data of subway passenger flow, and carrying out missing value processing; the data subjected to missing value processing are input into the optimized ICEEMDAN decomposition, and the data are decomposed; carrying out batch normalization on the data after data decomposition, inputting the obtained data into a bidirectional time convolution network model to carry out model parameter training, and carrying out feature extraction on the data through a causal expansion convolution layer, a linear activation layer, a discarding layer and residual connection; the full connection layers respectively output predicted values; and accumulating the predicted values of the decomposed data after reverse normalization to obtain a final prediction result. And inputting the test set into the trained prediction model to generate a final subway passenger flow prediction result. According to the subway passenger flow prediction method based on the optimized ICEEMDAN-bidirectional time convolution network model, the future change trend of the subway passenger flow can be well predicted, the passenger flow prediction precision is improved, and good adaptability is achieved. And decision support is provided for making a passenger flow management strategy.
Owner:XUCHANG MUNICIPAL RAIL TRANSIT CO LTD

Internet of vehicles computing power scheduling algorithm and system based on graph neural network and deep reinforcement learning

The invention relates to an Internet of Vehicles computing power scheduling method based on a graph neural network and deep reinforcement learning, and the method comprises the steps: S1, collecting the state information of network entities in the Internet of Vehicles and the association information between the entities in real time, and constructing a dynamic time-space attribute graph; s2, inputting the dynamic space-time attribute graph into a pre-trained graph neural network encoder, and outputting a node embedding feature set containing a high-order topological relation and global graph embedding features through multi-layer message passing and feature aggregation; s3, the node embedding feature set and / or the global graph embedding feature are / is used as the state of a deep reinforcement learning agent and input to a strategy network, the computing power scheduling action at the current moment is output, and the computing power scheduling action comprises the steps of assigning an unloading target node for a to-be-processed computing task and distributing corresponding computing and communication resources; s4, dispatching actions are distributed to the corresponding network entities to be executed, environment feedback is collected and used for model updating and next round of dispatching, and the method has the advantages of improving dispatching efficiency, adaptability and system energy efficiency and the like.
Owner:NANTONG SHIPPING COLLEGE

Intelligent electric meter service life prediction method based on improved Bi-lstm network and attention mechanism

The invention discloses an intelligent electric meter service life prediction method based on an improved Bi-lstm network and an attention mechanism, and the method comprises the steps: integrating the historical operation data and corresponding meteorological data of an intelligent electric meter, building a database, carrying out the preprocessing of the data in the database, and obtaining a sample set, calculating the service life of the intelligent electric meter of each sample in the sample set and adding into the sample set; a life prediction model is established, the model outputs a multi-head attention mechanism as the input of a Bi-lstm network, then the Bi-lstm network is optimized, an adversarial attention mechanism is introduced into the output layer of the Bi-lstm network, and finally a life prediction result is output; and training a life prediction model by using the sample set, and inputting the historical operation data and the meteorological data corresponding to the to-be-predicted smart electric meter into the trained life prediction model to obtain the predicted life of the to-be-predicted smart electric meter. The service life prediction capability of the intelligent electric meter is improved.
Owner:STATE GRID HUNAN ELECTRIC POWER COMPANY LIMITED +2