88 results about "Spatial search" patented technology

The invention provides a track board drilling parameter intelligent optimization method and system based on deep learning, and relates to the technical field of deep learning, and the method comprises the steps: recognizing a feature region in three-dimensional scanning data through a deep learning semantic segmentation network, building a local coordinate system, calculating hole site deviation, constructing a pose correction field, and achieving the intelligent correction of drilling process parameters. And an optimal drilling track is generated by using a state space search algorithm. The method can adapt to actual geometric errors of the track plate, the drilling precision and efficiency are improved, the drilling failure rate is reduced, and the manufacturing quality of the track plate is improved.

According to the large power grid reactive power optimization method and device, the storage medium and the computer equipment provided by the invention, the advantages of the two algorithms are fully exerted through the hybrid chaos quantum particle swarm optimization algorithm and the dimension-by-dimension convex space search algorithm. According to the chaotic quantum particle swarm algorithm, the global search capability and the capability of jumping out of local optimum of a particle swarm are enhanced by utilizing the characteristics of quantum behaviors and chaotic mapping, and the problem of premature convergence of a traditional heuristic intelligent algorithm is avoided. And according to the dimension-by-dimension convex space search algorithm, fine search is carried out on each excellent particle in different dimensions, a local optimal solution is determined, and the search precision and efficiency are further improved. According to the design of the hybrid algorithm, special optimization is carried out aiming at the characteristics of a reactive power optimization problem model, such as variable property difference, constraint complexity and the like, and the technical defects of poor optimization effect and optimization efficiency of an optimization solution algorithm in the prior art are effectively overcome.

The invention discloses a high-frequency transformer transmission extreme value frequency rapid calibration and tolerance analysis method, which comprises the following steps of collecting operation parameters of a high-frequency transformer in real time, performing standardized preprocessing, and integrating the preprocessed parameters to construct operation condition vectors in a unified format; defining a frequency response range of the high-frequency transformer, constructing a frequency search space based on the frequency response range, and establishing a target function for fitness evaluation based on the frequency search space; searching extreme value frequency points in the frequency search space by adopting a quantum evolutionary algorithm; based on the working condition vector and the extreme value frequency point, a working condition frequency deviation prediction model is constructed, and the working condition frequency deviation prediction model outputs an extreme value frequency shift offset; and calculating the dynamic tolerance bandwidth according to the predicted offset and the fluctuation range of the working condition vector. According to the invention, rapid and accurate calibration of the extreme value frequency of the high-frequency transformer and dynamic determination of the tolerance range can be realized.

The invention relates to the technical field of navigation assistance, in particular to a dynamic obstacle prediction and obstacle avoidance path planning method based on multi-sensor fusion, which realizes omnibearing perception of a dynamic obstacle through multiple sensors, provides a rich and reliable data basis for subsequent processing, and improves the accuracy of the dynamic obstacle prediction and obstacle avoidance path planning. Meanwhile, motion prediction is carried out on a dynamic obstacle through data processing fusion and model prediction, the motion inertia of the obstacle is fully considered, track prediction of the obstacle is closer to a real physical rule, a global path search problem is converted into a local but large enough space search problem through a mode of delimiting a final obstacle avoidance target area, and the search efficiency is improved. According to the method, the number of grids needing to be processed and the search calculation amount are remarkably reduced, the harsh requirement of the mobile device for the real-time performance is met, the area possibly occupied by the obstacle in the future is avoided in real time in the mode of constructing the cost function, then the obstacle avoidance planning path is generated, and the path safety is improved.

The invention discloses a front subframe lightweight design method and device based on RMNN, a medium and a program product, and the method comprises the steps: (1) constructing three simulation models of weight, stress and modal of a front subframe, and deducing a lightweight design model containing stress and modal constraints; (2) generating a population based on a uniform experimental design and a correlation criterion, performing simulation evaluation on the population, and constructing a database; (3) constructing a network architecture and training an RMNN; (4) generating a candidate frame set according to a random preferential strategy and the RMNN; and (5) establishing a radial basis function model, screening an optimal frame, performing simulation evaluation, updating the database and the population, returning to the step (3) until the optimization target reaches the standard, and outputting an optimal solution of the optimization parameter. According to the method, the feasible region is quickly searched through the RMNN, the problems that the high-dimensional design space search efficiency is low and the feasible region is difficult to recognize under multiple constraints are further solved in combination with the radial basis function model, and quick optimization can be carried out for the lightweight design of the front subframe.

The invention discloses a robust transmission method for time-sensitive network fault uncertainty, and relates to the field of industrial Internet of Things. According to the method, the fault uncertainty model and the double-layer robust optimization framework are constructed, the method adapts to complex and changeable link fault scenes in the industrial environment, deterministic transmission of key flow can be guaranteed under the worst fault condition, unpredictable interference such as temperature fluctuation and mechanical vibration can be effectively handled, and the anti-risk capacity of the network is remarkably improved. On the basis of an optimization model of mixed integer linear programming, in combination with constraint linearization and hierarchical solution architecture, the exponential complexity of full-space search is reduced to a polynomial level, and routing and scheduling optimization of a large-scale network can be completed within millisecond time. Through a space-time dimension conflict avoidance and resource reservation mechanism, the bandwidth utilization rate is maximized and the occupation of redundant resources is reduced on the premise of ensuring the real-time performance of high-priority services.

The invention discloses a diffusion model reasoning acceleration method based on optimal time step sequence search and knowledge distillation, which constructs a unified search space covering a time step sequence and corresponding model architecture configuration on the premise of not performing overall fine adjustment on a pre-trained diffusion model, and searches an optimal time step sequence. And designing a knowledge distillation training method based on the optimal diffusion time step sequence obtained by searching, constructing a main distillation loss function in a discrete time step subspace limited by the optimal diffusion time step sequence, and introducing adjacent time step consistency loss to relieve discrete errors introduced by a non-uniform time step span. In the reasoning stage, a teacher model is frozen, a distilled student model is combined with an efficient sampler ACDMS based on UniPC and fused with flow matching dynamic correction, finite step sampling is carried out on an optimal time step sequence, and reasoning acceleration under generation quality maintenance is achieved.

Disclosed in the present invention are an HM-TRW and HAGENN structure search-based collision risk prediction method and system for an emergency rescue autonomous driving vehicle. The method comprises: abstracting the relationship between an emergency rescue vehicle and surrounding moving objects into a dynamic heterogeneous graph, using HM-TRW to capture the importance and dynamics of each surrounding moving object and various interaction relationships, and fusing same with original features of each surrounding moving object, and inputting same into a HAGENN to perform collision risk prediction. The method which embeds the dynamic heterogeneous graph is applied to a hierarchical attention-based neural network to further learn the heterogeneous characteristics and the dynamic change rules of the surrounding moving objects; and during attention calculation of different hierarchies, a positioning space is used to determine an application position of attention and a parameterized space is used for searching an attention function, and a multi-stage differential search is introduced to accelerate the search process. The present invention can more comprehensively and accurately predict the collision risks of emergency rescue autonomous driving vehicles during operation processes.

The application discloses a circuit input sensitivity analysis method based on Undef SAT, which comprises five steps of netlist analysis and input division, two groups of CNF clause construction, S set assignment and initial solving, first group of solution increment SAT processing and termination and result output. The method uses the incremental SAT processing mode to gradually determine the final assignment of the input variables of the L set, effectively avoids full space search, significantly reduces the search space, and improves the analysis and determination efficiency. The method constructs two groups of CNF clauses supporting irrelevant item propagation, directly processes the irrelevant input variables of the S set by using the Undef SAT solver, reduces the number of logic propagation, and thus reduces the operation time. The integrated process of the circuit input sensitivity analysis formed by the method is suitable for mainstream EDA tool chains, has good engineering applicability, and is suitable for different scenes such as logic synthesis stage, formal verification, functional equivalence checking and loop oscillation detection.

The invention relates to a scale-aware geographic flow clustering method and device, electronic equipment and a storage medium, and the method comprises the steps: introducing a pre-constructed scale factor to construct a relation mapping between a spatial search neighborhood of a flow and a flow length, and generating a to-be-processed flow connection cluster set based on the relation mapping and a plurality of flows; any one to-be-processed stream connection cluster is selected from the to-be-processed stream connection cluster set, and whether any one to-be-processed stream connection cluster is a strong connection stream cluster is judged; if any to-be-processed stream connection cluster is a strong connection stream cluster, adding any to-be-processed stream connection cluster into a stream clustering cluster set, otherwise, taking any to-be-processed stream connection cluster as a weak connection stream cluster, and processing the weak connection stream cluster to obtain a potential strong connection stream cluster meeting a preset condition; and obtaining all strong connection flow clusters and all potential strong connection flow clusters in the to-be-processed flow connection cluster set to generate an actual flow clustering result of the flow. Therefore, the technical problems that local spatial distribution characteristics of stream data are neglected when spatial heterogeneous stream data are clustered, accurate perception of stream analysis scales is lacked, stream clusters with different densities cannot be correctly divided and the like in related technologies are solved.

The invention relates to the technical field of data processing and engineering project process management, in particular to a full-process intelligent management method and system for bridge and tunnel maintenance engineering. The method comprises the following steps: extracting a disease geometric contour based on bridge and tunnel inner wall three-dimensional point cloud data, and calculating a centroid coordinate, a volume and a surface contour area; constructing a space retrieval sphere by using a target disease centroid, screening an adjacent disease set, and calculating a local aggregation density by combining the volume and the centroid distance; calculating the average depth of the target disease, and deducing the overlapping sharing rate of the working plane; constructing a group topology consumption equivalent through linear analysis; and inputting the extreme gradient lifting model to output an optimal repair material dosage table and a material distribution operation instruction book. According to the method, the disease space distribution and aggregation degree can be accurately evaluated, the construction loss surface overlapping probability and the material sharing potential are scientifically evaluated, the material preparation accuracy is improved, waste and cost are reduced, and the whole-process intelligent management of bridge and tunnel maintenance engineering is realized.

The invention discloses an electric vehicle charging load prediction method and system based on model-data cooperative driving, and belongs to the field of electric vehicle charging load prediction. According to the method, an improved Monte Carlo inversion mechanism and a double-flow GCN-Transform combined prediction framework are used, so that the problem that prediction mechanism consistency, data adaptability and space-time heterogeneity are difficult to consider at the same time in an existing single driving mode is solved. The method comprises the following steps: firstly, constructing a charging behavior-path cost double-layer coupling physical model, inverting user behavior features through parameter space search, and generating a multi-modal feature set; secondly, capturing a spatial topological relation by adopting a graph convolutional network, and deeply fusing spatial and temporal features by utilizing a shared embedded layer and a Transform network; the method is remarkably superior to the prior art in the aspects of prediction precision, robustness and cross-scene generalization ability, and particularly, stable and reliable prediction performance can still be kept in a data incomplete or load sudden change scene.

The present disclosure discloses an emergency rescue autonomous driving vehicle collision risk prediction method and a system based on HM-TRW and HAGENN structure search, abstracts the relationship between the emergency rescue vehicle and the surrounding moving objects into a dynamic heterogeneous diagram, captures the importance and dynamic of each surrounding moving object and various interactive relations by using HM-TRW, and merges with the original characteristics of each surrounding moving object. Then input HAGENN to predict the collision risk, and apply the dynamic heterogeneous graph embedding method to the hierarchical attention neural network to further learn the heterogeneous characteristics and dynamic changes of the surrounding moving objects. When calculating the attention force of different levels, the positioning space is used to determine the application position of attention, the parameterized space is used to search the attention function, and multi-stage differential search is introduced to accelerate the above search process. The present disclosure can more comprehensively and accurately predict the collision risk in the operation process of the emergency rescue autonomous driving vehicle.

This invention relates to the field of navigation assistance technology, and in particular to a dynamic obstacle prediction and obstacle avoidance path planning method based on multi-sensor fusion. This invention achieves comprehensive perception of dynamic obstacles through multiple sensors, providing a rich and reliable data foundation for subsequent processing. At the same time, it predicts the motion of dynamic obstacles through data processing fusion and model prediction, fully considering the motion inertia of obstacles to make their trajectory prediction closer to real physical laws. Furthermore, by "delineating the final obstacle avoidance target area", the global path search problem is transformed into a local but sufficiently large spatial search problem, significantly reducing the number of grids to be processed and the amount of search computation, meeting the stringent real-time requirements of mobile devices. Moreover, by constructing a cost function, it avoids areas that obstacles may occupy in the future in real time, thereby generating an obstacle avoidance planning path, which helps to improve path safety.

A physics-informed data-driven oil and gas pipeline network fault diagnosis method and system for type-imbalanced data scenarios, relating to the field of oil and gas pipeline network fault diagnosis, and aiming at solving the overfitting problem of current intelligent fault diagnosis models during processing of oil field type-imbalanced data sets, and reducing the risks of false alarms and missed alarms. Main steps are as follows: deeply understanding propagation and attenuation mechanisms of negative pressure waves when oil and gas move in a pipeline network, and establishing a negative pressure wave attenuation physical model reflecting the operating state of the pipeline network; on the basis of a long short-term memory network, constructing a deep generative adversarial model suitable for processing time series data; designing a reasonable series-parallel mechanism to fuse the physical model and a data-driven model, so as to construct a hybrid generative adversarial model; using the trained hybrid generative adversarial model to generate pipeline network fault data, and balancing an original training set; and training an intelligent fault diagnosis model to achieve pipeline fault type identification. The present invention considers both the prior knowledge from the physical model and the learning capability of the data-driven model and integrates same, and compared with simple data-driven models, uses the prior knowledge of the pipeline network contained in the physical model to reduce a parameter space search domain, thus reducing the number of estimated parameters, improving the interpretability and generalization performance of a deep generation model, and improving the physical rationality and feature distinguishability of generated fault data. Therefore, the present invention effectively overcomes the negative impact of type-imbalanced data sets on the performance of diagnosis models, further improving the accuracy of intelligent fault diagnosis of pipelines.

The present invention discloses a method and system for diagnosing mechanical faults of high-voltage circuit breakers based on a MIDBO-SVM model. The method first obtains a vibration signal of the circuit breaker, and then uses the obtained vibration signal in combination with a pre-built MIDBO-SVM model to diagnose the mechanical faults of the circuit breaker. During the construction of the MIDBO-SVM model, a good point set strategy and a circle chaos map are used to jointly generate an initial population method to improve the uniform ergodicity of the population. A sine-cosine search strategy and an adaptive weight factor are introduced during the individual position update stage of a rolling ball dung beetle to cause the solution to attenuate and oscillate until it approaches the global optimal solution. A Cauchy-Gauss mutation strategy is used to mutate the individual with the best current fitness in a manner away from the optimal solution, thereby expanding the spatial search range. The method also solves the problems of slow search speed and susceptibility to falling into local optimality in the later stages of traditional algorithms, thereby improving the accuracy of circuit breaker fault classification.

The application discloses a kind of optimization method and device for identifying Cu-Zn alloy cluster structure potential energy surface, comprising: randomly generating several Cu-Zn alloy clusters, constitute population structure library n pop ; develop shell script that can efficiently call Gaussian16 quantum chemistry software, carry out first local structure optimization to n pop ; innovative surface atom self-adapting reorganization strategy is proposed, iteratively move cluster surface atoms to self-adapting generated empty space sites, execute second stage structure optimization;Through crossover, variation operation, in combination with similarity detection and potential energy comparison, extract the optimal structure of binary Cu x Zn y Cluster.Using the technical scheme of the application, the local and global spatial search capability of genetic algorithm is optimized and improved, and the stability and electronic structure of Cu-Zn alloy cluster can be efficiently analyzed by calling quantum chemistry software.

The invention relates to a registration method and system for mouse brain tissue slice images. On the basis of a two-dimensional mouse brain tissue slice image to be registered and a section parameter prediction model, spatial positioning parameters are determined, and tedious three-dimensional space search and iterative matching in the prior art are effectively replaced by one-time prediction based on the section parameter prediction model, so that the matching efficiency is improved, and the robust spatial positioning parameters are given; determining deformation mapping information based on the spatial positioning parameters, the two-dimensional mouse brain tissue slice image and the three-dimensional average mouse brain template map, determining a two-dimensional labeled slice based on the spatial positioning parameters and the three-dimensional standard map, and determining a registered mouse brain tissue slice image based on the two-dimensional labeled slice and the deformation mapping information, according to the method and the system, subsequent fine alignment processing is carried out on a two-dimensional level, so that the calculation complexity and the algorithm difficulty are greatly reduced, an automatic workflow of intelligent three-dimensional positioning and efficient two-dimensional registration can be formed, and powerful support is provided for high-throughput analysis and clinical application.

The present invention discloses a method, system, computer device and storage medium for scenario behavior deduction and reliability prediction analysis of a multi-source data fusion dynamic system. The method, based on the Markov / CCMT dynamic reliability prediction analysis method, combines multi-source data fusion and assimilation methods, and uses Monte Carlo probability model random sampling to simulate and statistically analyze the complex dynamic behavior characteristics of digital process control with strong interactive coupling, nonlinearity and high uncertainty. Then, through dynamic search and analysis of the system state transition probability matrix model, forward deduction analysis and reliability prediction of the system operation state are achieved. The present invention can realize the adaptive update construction of the state transition probability matrix of large-scale complex digital process control systems and the dynamic deduction and analysis of scenario behavior, avoid the problem of high-dimensional system state space search explosion, and can accurately simulate and map the system dynamic behavior characteristics to realize system dynamic reliability prediction analysis.

The prompt word determination method, device, electronic equipment and storage medium provided by the present application relate to the technical field of prompt word processing. The method obtains baseline prompt words and a test case set based on a preset service requirement; performs multi-round response testing on a target model by using the baseline prompt words and the test case set, obtains model behavior logs corresponding to each round of response testing, and obtains a prompt word position effect matrix and a vocabulary sensitivity knowledge base based on the model behavior logs corresponding to each round of response testing; and performs mixed variable space search processing by using the prompt word position effect matrix, the vocabulary sensitivity knowledge base and the model behavior logs corresponding to each round of response testing, to obtain effective prompt words, thereby improving the optimization efficiency of prompt words and enhancing the verifiability of results.

The application provides a multi-modal data security analysis technology based on a deep fuzzy wavelet learning model. The deep fuzzy wavelet learning model is widely used in the field of security analysis due to its self-adaptive ability in complex data environment. The method improves the deep fuzzy wavelet learning framework and applies it to multi-modal data security analysis. The model first uses the method combining wavelet transform and fuzzy clustering to perform feature decomposition and noise reduction processing on multi-modal data (text, speech, video, image), extracts key semantic, timing and spatial features, and provides high-dimensional information representation for subsequent security analysis. Secondly, a deep fuzzy wavelet neural network is constructed, the extracted features are input into the adaptive fuzzy decision layer, and the wavelet learning model is combined to capture the spatio-temporal correlation features, improve the perception ability of the model to complex attack patterns. Finally, the security analysis result is strengthened by using the space search optimization algorithm, so as to effectively detect malicious samples, identify backdoor attacks, and improve the defense ability against attacks, and finally realize accurate multi-modal data security evaluation.

The invention provides a path planning method and system for personnel evacuation and obstacle avoidance, and belongs to the technical field of personnel evacuation dynamic obstacle avoidance path planning. The invention provides a Bi-RRT *-based personnel evacuation path planning algorithm with self-adaptive bias and a new improved self-adaptive extended step length. On the basis of the standard Bi-RRT * algorithm, an adaptive bias strategy which is jointly influenced by search experience and obstacles is added, so that the convergence speed of the algorithm is effectively improved, the path search efficiency is improved, and the ability of passing through narrow terrains is improved; based on a self-adaptive step length search strategy combining the repulsion size and the direction included angle, the obstacle avoidance capability is improved; in combination with improved adaptive bias, adaptive step length and bidirectional search capability, the adaptive capability to a complex evacuation space search path and the search efficiency of a multi-obstacle region algorithm are effectively improved. The method has remarkable advantages in the aspects of path planning efficiency, path quality, convergence speed, complex environment adaptability and the like.

The invention relates to a multi-target cooperative power flow real-time distribution method for a hybrid tracked vehicle considering dynamic stability constraints, and belongs to the technical field of hybrid tracked vehicle control. The method comprises the following steps: estimating a road adhesion coefficient in real time based on UKF and generating a real-time dynamic stability boundary; identifying a driving intention based on multi-dimensional features and vehicle speed self-adaption; constructing an all-condition off-line mapping table based on reverse modeling; and performing multi-target collaborative MPC predictive control and real-time distribution based on MAP initial value hot start. According to the method, through the hybrid architecture of providing the initial value through offline all-condition traversal and processing stability constraint through online MPC, huge computing power consumption caused by online full-space search is effectively avoided, and meanwhile, the safety and economical efficiency of the vehicle under the complex road surface working condition are guaranteed.

The invention discloses an intelligent production scheduling method and device for a production workshop. The method comprises the following steps: establishing a mixed integer linear programming model for performing workpiece scheduling on the production workshop; performing initial solution space construction, and generating an initial scheduling solution set corresponding to the target function; performing neighborhood solution space search, solution space destruction and solution space reconstruction, and constructing a reconstruction solution corresponding to the initial scheduling solution set; replacing the initial scheduling solution set with the reconstruction solution, and repeating the steps of neighborhood solution space search, solution space destruction and solution space reconstruction until the total processing duration of all workpieces corresponding to the target function of the obtained reconstruction solution is lower than a preset value; and carrying out model decoding operation on the reconstruction solution of which the total processing duration of all workpieces corresponding to the target function is lower than a preset numerical value in combination with the coding structure to obtain a production scheduling scheme of the production workshop. The method is used for effectively optimizing the production schedule of the production workshop, reducing the maximum completion time of production and improving the production efficiency.

This invention discloses a pole angle consistency improvement method based on residue enhancement, relating to the fields of radar signal processing, target detection and recognition technology. This invention maps the complex values ​​of poles to two variables: real and imaginary parts. It uses an optimization algorithm to dynamically optimize the unnormalized real and imaginary parts of the poles. During the optimization process, the optimized poles and the frequency domain response echoes reconstructed by the original residues are brought closer to the frequency domain response echoes of the original poles theoretically obtained by increasing the residues in the vicinity of the frequency range corresponding to the optimized poles. This combines the spatial search of pole positions (adjusting frequency position) with the gain effect of the residue amplitude (increasing response intensity), effectively solving the problem of pole angle flickering caused by low echo energy contribution during pole extraction. It corrects pole positions at their root, identifies hidden poles, improves pole angle consistency, and generates a complete pole feature database.

The invention discloses an RCS reduction method of an electromagnetic digital coding metasurface based on a GAN-PSO joint algorithm, and belongs to the field of electromagnetic digital coding metasurfaces and electronic countermeasures. The method comprises the following steps: firstly, constructing a 1-bit metasurface unit with a phase difference of 180 degrees + / -10 degrees; thirdly, training a generative adversarial network (GAN) to enable a generator of the GAN to map a continuous vector into a discrete metasurface coding matrix; then, a fitness function of a particle swarm optimization (PSO) algorithm is set to be far field scattering function maximum value minimization; and finally, GAN-PSO joint optimization is executed: PSO searches in a continuous space, a generator maps particle positions into candidate coding matrixes and calculates fitness, and iterative updating is carried out until an optimal coding matrix is output as a final arrangement scheme. Through GAN-PSO joint algorithm optimization, the bottleneck that a traditional coding strategy is low in degree of freedom and uncontrollable in scattering is broken through, and high-performance broadband RCS reduction is achieved.

The application discloses a kind of based on layered joint graph index's house rental multi metric space search method, comprising: obtaining new house object, find the k nearest neighbor house object of the new house object in each metric space, under each metric space, add undirected edge in new house object and its k nearest neighbor house object, adopt splitting and hierarchical update strategy based on tree, dynamically adjust graph structure, improve node distribution, to build layered joint graph index, wherein the metric space includes price metric space, house type metric space, geographical position metric space and comment metric space;Based on the pruning technique of best priority search strategy, multi metric space approximate nearest neighbor search is carried out using the layered joint graph index, and the approximate-k nearest neighbor house object of given query house object is obtained.

The invention discloses a rural area truck and unmanned aerial vehicle cooperative distribution path planning and resource scheduling method, and belongs to the technical field of path planning. According to the method, a multi-target truck-unmanned aerial vehicle path problem with a flexible time window is modeled as a reward maximization problem in deep reinforcement learning; a time window is allowed to be deviated within tolerance time, and punishment is applied to serious deviation; a multi-agent model is adopted, a distribution unit composed of a truck and a carried unmanned aerial vehicle is regarded as an agent, space-time topological features of a complex logistics network are embedded and captured through a multi-order space-time diagram convolutional network, the attention mechanism of the multi-order space-time diagram convolutional network is utilized to perform weighted summation on adjacent node features to generate specific weights of client nodes, and the state sensing ability of the agent is enhanced; and based on a multi-agent near-end strategy optimization algorithm, the collaborative distribution path of the multiple truck-unmanned aerial vehicle units is optimized, and the parallel processing method is combined with ST-MGCN embedding, so that the action space search dimension is effectively reduced, and the network convergence speed and the solving efficiency are improved.

This invention belongs to the field of power system technology, specifically relating to a method and system for mining rules for electricity price and fee consistency verification. The method includes: acquiring a verification dataset and calculating the equivalence partitions of quantity, price, and fee data; using a depth-first spatial search strategy to mine data consistency relationships, obtaining a set of consistency relationships in the data; using the obtained set of consistency relationships to perform rule parsing on the quantity, price, and fee consistency relationships; and identifying rare data in the quantity, price, and fee data through rule parsing. This invention applies a novel consistency rule mining method for categorized data to the verification of electricity price and fee data. It can proactively mine potential rules in the data, effectively overcoming the shortcomings of previous manual rule definition methods, such as high workload, long time consumption, insufficient rule coverage, and the ability to only discover abnormal business data covered by defined rules. This assists business personnel in completing the verification of electricity marketing quantity, price, and fee data more intelligently and comprehensively.

The present invention provides an audio and video retrieval method, system and storage medium based on a large language model, relating to the field of computer information technology. The method comprises the following steps: obtaining natural language request information related to audio and video retrieval content; vectorizing the request information according to a large language model tool to extract retrieval information of the request information; searching in a vector database using a vector space search algorithm to screen out vectors matching the retrieval information; searching in a relational database by matching the index information of the vector to obtain associated information corresponding to the vector, the associated information including the video name, text content and start and end time; sending the target video or target audio corresponding to the vector and its associated information to a client, and being able to select a specified time point from the start and end time to play the target video or target audio.