18results about How to "Improve search capabilities" patented technology

This invention relates to the field of battery state prediction technology and discloses a method for predicting the health status of UAV power batteries. The key technical points of this method are: data acquisition and quantum encoding; quantum feature extraction; construction of a quantum-heuristic deep learning model; model training and optimization; and health status prediction and assessment. Through quantum encoding, a quantum-heuristic model architecture, and a dynamic optimization mechanism, this method overcomes the limitations of traditional methods in modeling complex nonlinear relationships and local optimization, providing a high-precision and robust solution for the health management of UAV power batteries, and has significant engineering application value.

The present application belongs to the technical field of tracked vehicles, and particularly relates to a method for determining the pre-tensioning force of a high-speed tracked vehicle, comprising the following steps: establishing a topological graph of a high-speed tracked vehicle system; establishing a discretized mass model of a lower track section; establishing a continuum verification model of the lower track section; checking the discretized mass model; establishing a dynamic model of the road wheels and the vehicle body; establishing a multi-level coupled vibration model of the vehicle; establishing an excitation model of road unevenness and different vehicle speeds; constructing an objective function of simultaneous passability and task load stability; setting the pre-tensioning force as a design variable and defining its variation range; using an optimization algorithm to optimize the pre-tensioning force under different structural masses and excitation conditions; and substituting the optimal solution of the pre-tensioning force into the multi-level coupled vibration model for solving and verification. The present application can provide theoretical support for setting the pre-tensioning force of a high-speed tracked vehicle.

The application provides a file retrieval method and system based on HBase, and relates to the field of computer systems.In the application, the field library defined by each business system in the incremental data is uploaded by each project group when uploading the file to the bank storage system through busid, and the field library extracted from the unstructured text by using an algorithm is trained by using a bert model or a chatGLM model in a natural language processing technology.The application is based on the technical scheme of hbase+es, the search feature of es is used to quickly and accurately search the unstructured data stored in the hadoop cluster, the unstructured data files meeting the conditions are searched in the storage system through the self-definable conditions, and the search capability of the unstructured data files of the storage system is enhanced.

The application discloses a method for cooperative jamming based on intelligent optimization algorithm, comprising: constructing a jamming decision model, the jamming decision model comprising a cooperative jamming decision matrix, a gain matrix, a jamming matrix, a jamming gain matrix, a jamming bandwidth ratio factor, a jam-to-signal ratio, and a jamming benefit; establishing an objective function and a constraint condition of the jamming decision model according to the jamming benefit; and using an artificial bee colony algorithm to take the jamming benefit as a fitness function and optimize the cooperative jamming decision matrix A. In different complex electromagnetic spectrum environments such as limited spectrum resources and the same frequency band shared by jamming devices and illegal users, the limited jamming resources are reasonably distributed under the condition that the jamming device of the own side can normally communicate, so that greater jamming benefit is achieved; the algorithm convergence speed and search ability are improved, and the method is helpful for making a decision with higher jamming benefit in a shorter time.

The application discloses a mask image recognition method based on an evolutionary support vector machine. The application utilizes an improved differential evolution algorithm to evolve training parameters of a support vector machine based on a polynomial kernel function, and then utilizes the evolved support vector machine based on the polynomial kernel function to recognize whether a face image is a mask image. In the improved differential evolution algorithm, selection probabilities of mutation strategies are calculated first, then a population is divided into a guide subpopulation and a repulsion subpopulation, and meanwhile, directional information of guide individuals and repulsion individuals is utilized to improve search performance of the algorithm, so as to improve recognition precision of the mask image.

The application discloses an internet intrusion detection method and device based on an optimized related vector machine, and the method comprises the following steps: acquiring an NSL-KDD data set, including a training set and a test set; performing non-numerical one-hot encoding and numerical normalization processing on the NSL-KDD data set to obtain the preprocessed training set and test set; inputting the preprocessed training set into a related vector machine model which is pre-constructed and optimized based on an improved crow search algorithm to perform training, and obtaining the optimized and trained related vector machine model; inputting the preprocessed test set into the optimized and trained related vector machine model to perform testing, and obtaining a classification detection result. The application has better convergence, and can improve the classification accuracy of the intrusion detection data set and reduce the false positive rate.

The application relates to a heterogeneous multi-sensor joint deployment method for low-altitude target detection, comprising the following steps: constructing three-dimensional monitoring grid units covering the ground and low-altitude airspace of a region to be monitored, and determining a multi-sensor set to be deployed and continuous state variables and discrete state variables thereof; judging the line-of-sight unobstruction of each three-dimensional monitoring grid unit by using a digital elevation model, and calculating the comprehensive detection probability of each three-dimensional monitoring grid unit for cooperative targets and non-cooperative targets under multi-sensor cooperative detection according to the result of the line-of-sight unobstruction judgment; constructing a multi-objective optimization function with the minimum total cost and the maximum monitoring coverage rate as the targets, wherein the monitoring coverage rate is determined based on the comprehensive detection probability; and solving the multi-objective optimization function by using a hybrid evolutionary algorithm to output a multi-sensor deployment scheme. The scheme can improve the global search capability and convergence efficiency of the algorithm, and obtain a sensor station deployment scheme considering the network construction cost and the three-dimensional coverage rate.

This invention discloses a parameter estimation method, system, device, and storage medium for a nonlinear diffusion model. The parameter estimation method includes: constructing a spatial basis function based on system data snapshots for a nonlinear Fisher-type diffusion system applied to the heating and cooling process of an iron rod, and separating spatiotemporal variables using an orthogonal decomposition method; sparsely sampling the nonlinear terms using a discrete empirical interpolation method to obtain the optimal low-order approximation of the high-order system; initializing a particle swarm to obtain n particles corresponding to the m-dimensional solution vector of the low-order time series model, and calculating the fitness of each particle according to the calculation formula of the parameters to be identified and the objective function; iterating and optimizing each particle, and outputting the position of the particle with the global optimum value; adding a non-Gaussian Levy process to the traditional particle swarm algorithm, and avoiding premature concentration of the particle swarm in the same direction through the random jump of the Levy process, thereby increasing the mutual learning ability between particles.

This invention discloses a path optimization method based on cooperative aggregation and branch deviation, comprising the following steps: (1) Initialize the population and variables, and divide the population into two subpopulations s1 and s2; (2) Calculate the fitness of individuals in subpopulations s1 and s2 according to the fitness function; sort the fitness of individuals in the two subpopulations from smallest to largest, and select the optimal individual; (3) Keep the i1 and i2 individuals with the highest fitness in subpopulations s1 and s2 respectively, and take the remaining individuals as clustered individuals; (4) Apply the cooperative aggregation strategy to the clustered individuals; (5) Take the top m1% and m2% of s1 and s2 respectively according to fitness to synthesize a new population s3; (6) Update the individuals in s3 randomly using the replacement strategy, and update the global optimal individual; (7) Determine whether the termination condition has been met. If not, increase the population diversity using the crossover operator and mutation operator; (8) Iterate according to steps (2)-(7) until stopping, and output the optimal path.

This invention provides a reinforcement learning-based method for finding the nearest point and navigating a patrol robot, comprising: Step 1: Setting two routes A and B that the patrol robot needs to inspect; Step 2: Obtaining the initial position coordinates of the current patrol robot and activating the LiDAR; Step 3: Obtaining the current yaw angle of the patrol robot, allowing the robot to move straight for m seconds, and obtaining the current position coordinates of the patrol robot, unifying the coordinate system; Step 4: Selecting the nearest path point as the initial point for tracking; Step 5: Connecting the initial point with the next path point to obtain the angle relative to the IMU coordinate system, controlling the point-to-point movement of the patrol robot; Step 6: When patrolling route A is completed, starting patrolling route B, using a reinforcement learning algorithm to select the nearest path point to route B as the initial point for patrolling. This invention's method has high efficiency and accuracy, significantly reducing the computational complexity of existing technologies.

The application discloses a neural network distributed automatic parallel training method based on AC reinforcement learning. Firstly, the application performs performance data sampling analysis on a neural network model based on a performance analysis method, extracts performance data of model operators, and realizes operator grouping fusion. Secondly, global feature vector representation of a calculation graph is realized, and a state search space of reinforcement learning is constructed. Then, an AC algorithm based on time series difference sampling is adopted to complete iterative optimization of the reinforcement learning model, so as to search for an optimal distributed parallel strategy. Finally, a multi-thread simulation executor based on event driving is constructed to simulate the calculation process of the neural network model. The application reduces the state search space scale of the AC reinforcement learning, improves the search performance and universality of the parallel strategy, reduces the strategy iteration execution time, and reduces the hardware execution cost in the iterative search process.

The application discloses a reinforcement learning driven heuristic solving method for multi-ship co-dock maintenance scheduling, and three global search operators are designed for the multi-ship co-dock maintenance scheduling problem. One is a ship crossover operator, which is used to adjust and exchange the maintenance order of the ship. The second is a maintenance task destruction operator, which breaks the original task structure by extracting part of the task from the current maintenance task sequence, providing a larger optimization space for subsequent search. The third is a maintenance task recombination operator, which inserts the extracted maintenance task into the existing maintenance task list to form a new maintenance task order, thereby improving the possibility of obtaining a better scheduling scheme. The execution process of the global search operator is adaptively guided, and a more suitable search operation is dynamically selected according to the state information in the solving process, so as to reduce the invalid search overhead in the meta-heuristic algorithm, improve the utilization efficiency of the computing resources, and further improve the solving efficiency and solving effect of the multi-ship co-dock maintenance scheduling problem.

The application discloses a photovoltaic data management method based on a self-organizing network and related devices, and relates to the technical field of data processing. The method comprises the following steps: establishing a connection between each wireless communication node based on the performance interaction of a communication link and the connectivity between adjacent wireless communication nodes to form a self-organizing network; determining a target transmission path based on the time delay overhead and the signal-to-noise ratio of each communication link by using a hybrid algorithm to transmit photovoltaic data; classifying based on the feature matrix of the photovoltaic data to obtain photovoltaic data of each category; performing security requirement analysis on the photovoltaic data of each category to obtain a target requirement security level; performing service requirement degree analysis on the photovoltaic data of each category to evaluate the cold and hot degree thereof; and determining a storage strategy based on the cold and hot degree and the requirement security level of the photovoltaic data of each category in combination with storage node performance analysis to store the data into corresponding storage nodes. The application can achieve a more ideal effect for the storage management of photovoltaic data.

The application discloses a kind of based on simulated annealing algorithm's repair tree construction method, including obtaining node number and corresponding adjacency matrix;Set initial parameter and control parameter;Randomly generate initial Prufer sequence and execute simulated annealing algorithm, record the maximum bottleneck bandwidth corresponding Prufer sequence as current solution;Current solution produces disturbance and obtains legal new solution;Calculate the bottleneck bandwidth of new solution and current solution and decide whether to accept new solution;Repeat the above two steps until the set condition is met, obtain the Prufer sequence corresponding to the maximum bottleneck bandwidth and decode to get rootless tree, the auxiliary joint in rootless tree is used as root node and the final repair tree is obtained, and the application also discloses a kind of data recovery method comprising the repair tree construction method based on simulated annealing algorithm of the application.The encoding and decoding speed of the application is fast, algorithm is flexible, search effect is good and reliability is high.

The invention belongs to the technical field of quadrotor unmanned aerial vehicles, and particularly relates to a quadrotor unmanned aerial vehicle mapping and planning method based on deep reinforcement learning, and the method comprises the following steps: S1, feature extraction and screening, S2, feature map creation, S3, feature map position calculation based on value iteration, and iteration through geometric verification and using a TD algorithm, and S4, state estimation. A planning method of quadrotor unmanned aerial vehicle mapping based on deep reinforcement learning is used for planning a generated flight trajectory of quadrotor unmanned aerial vehicle mapping, and the planning method comprises the following steps: S5, generating a strategy network model of the trajectory; according to the method, the common feature points and the feature point set between the frames can be rapidly extracted, the correlation graph of the feature points and the feature set is determined, and finally the corresponding relation and the position of the graph are calculated through a reinforcement learning method to obtain the accurate feature point relation.

The embodiment of the invention discloses a target object searching method and device, electronic equipment and a storage medium. The target object searching method comprises the steps that feature data of a to-be-searched target object is obtained, and a reference monitoring camera corresponding to the to-be-searched target object is captured for the first time; determining a candidate monitoring camera group based on the monitoring range of the reference monitoring camera, the trajectory features of the target to be searched and the spatial-temporal distribution relationship of the monitoring cameras; performing hierarchical feature matching on the image data acquired by the candidate monitoring camera group; and according to the matching result and the coverage range of the candidate monitoring camera group, generating a search result corresponding to the to-be-searched target object, so that the target object search efficiency and precision can be improved, and the search effect is improved.

This application provides a key gene network search method and system based on intelligent agents, belonging to the field of gene search technology. The method includes: acquiring multiple search data sets for searching key genes of target genes; for each search data set, scoring each gene in the whole genome based on the search data; determining a candidate gene set corresponding to the search data set from the whole genome based on the scoring results; performing set operations on the candidate gene sets corresponding to each search data set to obtain a target candidate gene set; for each target candidate gene in the target candidate gene set, treating the target candidate gene as a network node, and determining the node attributes of the network node based on the score of the target candidate gene in each candidate gene set; generating edges between each network node based on preset gene association evidence and each target candidate gene to obtain a key gene network. This application can improve the accuracy of key gene search.

The application belongs to the technical field of power systems and discloses a power distribution network distributed resource optimization aggregation method based on an ant colony simulated annealing algorithm; the method first acquires parameters of distributed energy in a virtual power plant, constructs an objective function with the aim of minimizing the total cost of the virtual power plant, sets a constraint condition, and finally solves the objective function by combining an ant colony algorithm and a simulated annealing algorithm to output an optimal solution, namely, an optimal aggregation method of the distributed resources of the power distribution network; the application can jump out of the trap of a local optimal solution, enhance the global search capability, improve the convergence speed, and significantly improve the probability of finding a global optimal solution; the cooling strategy of the simulated annealing algorithm accelerates the late convergence, the pheromone mechanism of the ant colony algorithm strengthens high-quality solutions, and the cooperation of the two can improve the economy and stability of the scheme; by setting the constraint, the aggregation scheme can meet various operation requirements, reduce the scheduling cost, improve the resource utilization rate, realize the efficient utilization and optimal allocation of resources, and the like.