89results about How to "Improve search accuracy" patented technology

The invention relates to the field of particle size distribution measurement, in particular to a particle size distribution measurement method and system based on machine learning and a light scattering method.The method comprises the steps that a to-be-measured particle sample is irradiated through a light source, and meanwhile light scattering intensity data of different scattering angles and different wavelengths are collected through a multi-angle scattered light detector; constructing a multi-modal deep learning model, and forming an integrated learning architecture in combination with a support vector machine model and a random forest model; inputting the preprocessed multi-wavelength light scattering intensity data into the trained integrated learning architecture to obtain a preliminary particle size distribution result of the to-be-detected particle sample; based on the improved Mie scattering theory, a high-precision theoretical model considering the particle shape, the refractive index, the surface roughness and the internal structure is constructed; an improved Bayesian optimization algorithm is adopted, and the initial particle size distribution result is optimized and adjusted by minimizing the error between the theoretical light scattering intensity and the actually collected light scattering intensity, so that a corrected particle size distribution result is obtained.

The invention discloses a teaching video reconstruction method and system based on an intelligent agent. The method comprises the following steps: extracting a knowledge content text and time information from a teaching video; constructing a semantic vector library and a graph database; constructing a cue word template according to a mode layer of the knowledge graph; defining a core function module of the intelligent agent system; the core scheduler receives a user question and calls the knowledge base retriever to generate a related knowledge content text; the knowledge base retriever queries the graph database to obtain the corresponding time of the video frame required for answering the question; and the video reconstructor reconstructs and generates a teaching video required by the user by using the video frame time and the original video. Precise semantic understanding and matching of teaching video content and user questions are achieved, the teaching video is reconstructed according to the multi-dimensional requirements of the user for knowledge points, learning targets and learning ability, specific field fine adjustment does not need to be conducted on a large model, the method can adapt to teaching video reconstruction of different subjects and difficulty levels, and the user experience is improved. Wide application and popularization values are realized.

The invention relates to a data calibration method suitable for an atmospheric pollutant sensor deployed in an embedded terminal. The method comprises the following steps: acquiring original response data of different sensors with different concentrations in different atmospheric environments after temperature compensation, and corresponding temperature, humidity, air pressure data and sampling time; collecting real concentration values of detection objects corresponding to different sensors at corresponding moments given by a national control station or a provincial control station; preprocessing the obtained data to obtain a final data set; a random forest algorithm is used as a teacher model, a gradient elevator algorithm is used as a student model, and hyper-parameters of the gradient elevator algorithm and the random forest algorithm are optimized by using a hedera helix algorithm; substituting a hyper-parameter optimal solution found by optimization into a corresponding algorithm, and retraining the teacher model and the student model by using hyper-parameter combinations; and deploying, updating and displaying the model in the embedded equipment. And the model size is reduced to realize the deployment of the embedded end on the premise of ensuring the accuracy.

The invention relates to the technical field of crowd sensing, in particular to a cooperation mode excitation method and system based on a Stackelberg game. The method comprises the steps of obtaining a task set of crowdsourcing tasks and user historical data; calculating a quality score of the crowdsourcing task according to the urgency degree and the energy consumption of the crowdsourcing task; performing task division on the crowdsourcing task based on the quality score of the crowdsourcing task, and generating a cooperative task based on the task division; performing capability value evaluation on the user according to the historical data of the user; and task allocation is carried out based on a capability value evaluation result and the generated cooperative task, a Nash equilibrium solution is solved in the task completion process by using a Stackelberg game, and an optimal allocation strategy is obtained. By introducing concepts of capability values, shelving levels and quality scores, tasks which exist in a system for a long time and are high in difficulty are combined, the tasks are completed through cooperation of users, and the receiving completion rate of the tasks is guaranteed.

The invention discloses a PID (Proportion Integration Differentiation) parameter optimization method based on an improved graffiti search algorithm, and relates to the technical field of PID. The invention aims to solve the problems of insufficient local search capability, slow convergence speed and poor system stability in the prior art, thereby realizing a better PID parameter optimization effect. Specifically, the PID parameter optimization method based on the improved graffiti search algorithm comprises the following steps: constructing a PID controller model; a step of replacing a fixed perception probability and a step length in the model with an adaptive perception probability and a Levy flight step length; a step of adding a crow misleading prevention factor into the model; and acquiring preset parameters in the PID controller model based on a particle swarm algorithm. The preset parameters comprise proportional (Kp), integral (Ki) and differential (Kd) parameters. Parameters of the Levy flight step length are dynamically adjusted through a Bayesian optimization algorithm. The method is suitable for being applied to PID parameter optimization work.

The invention relates to the technical field of data processing, in particular to a parameter optimization method for a powder laser 3D printing forming process, which comprises the following steps: determining a solution space of a particle swarm algorithm, performing finite element analysis on each process parameter, determining an objective function of the particle swarm algorithm, and obtaining simulated particles and actually measured particles in the solution space, in the process of carrying out iterative training on a pre-constructed support vector machine by utilizing the actually measured particles, calculating the credibility of the simulated particles during any training, and when the product of the credibility and the normalized simulated fitness value is greater than a preset value and the actually measured fitness value of the simulated particles is greater than a conversion threshold value, judging that the simulated particles are not subjected to the training; and taking the simulated particles as actually measured particles in the subsequent training process, repeating the training process, and obtaining the target process parameters after the training is completed. According to the method, the accuracy of the determined target process parameters can be improved while the calculation amount and the time cost are reduced.

The invention discloses a tumor postoperative treatment staged planning method based on artificial intelligence. The method comprises the following steps: S1, constructing a multi-modal medical data set; s2, performing standardization processing on the acquired multi-modal medical data set to form a standardized data set which can be used for feature extraction; s3, generating a unified patient feature vector; s4, according to the characteristics of tumor postoperative treatment, dividing the whole treatment process into a postoperative immediate recovery stage, a postoperative recovery middle stage and a long-term follow-up visit stage, and defining a corresponding optimization target for each stage; s5, based on the patient feature vector and the optimization target of each stage, generating an initial treatment plan by using a deep ant colony algorithm; and S6, outputting a staged treatment scheme. According to the method, the professional and technical thresholds required by traditional modeling are reduced, and the universality and convenience of three-dimensional modeling are improved.

The invention discloses a two-dimensional MUSIC spectrum peak searching method based on an improved Archimedes algorithm, and the method comprises the steps: carrying out the one-dimensional MUSIC spatial spectrum function derivation of a signal model, and expanding the one-dimensional MUSIC spatial spectrum function to a two-dimensional MUSIC spatial spectrum function; secondly, taking a spatial spectrum function of the two-dimensional MUSIC as a target function of an Archimedes optimization algorithm; and then, in combination with an objective function, introducing a fusion chaos algorithm strategy to initialize a population, introducing a mutation algorithm, disturbing positions, and improving an Archimedes optimization algorithm. And finally, carrying out two-dimensional MUSIC spectral peak search through an improved Archimedes optimization algorithm, and taking the direction of arrival, namely, a pitch angle and an azimuth angle, obtained in a mode of solving maximum and minimum values through a target function as spectral peak values. The two-dimensional MUSIC spectrum peak searching method has the advantages that the optimization speed is higher, the optimization precision is higher, the robustness is better, and the two-dimensional MUSIC spectrum peak searching can be quickly and accurately carried out.

According to the multi-modal intelligent question answering method and device, the storage medium and the computer equipment, when the query question is received, the target query text of the query question is extracted to avoid input carrier limitation of the query question, then the target query text is vectorized into the target text vector, and the query question is queried according to the target text vector. The target text block is obtained based on the target query text and the target text vector, so that the retrieval accuracy is improved; and then optimizing the target text block by adopting the target text optimization model and outputting a high-accuracy target answer. If the target answer contains the picture address, a picture request is initiated to the server after a request address corresponding to the picture address is generated, and the server performs security verification on the request address and returns picture content corresponding to the picture address after the request address passes the security verification, so that the data security is improved; and finally, replacing the picture address in the target answer with the picture content, and displaying the final target answer to the user through a multi-modal mode of the text and the picture, so that the intuition and the accuracy are improved.

The invention provides an unmanned vehicle obstacle avoidance path planning method, and relates to an unmanned vehicle path planning method, which comprises the following steps of: adding an improved gravitational potential field to a target point, adding an improved repulsive potential field to an obstacle, meanwhile, a road boundary potential field and a speed repulsion potential field are introduced to ensure that the unmanned vehicle does not collide with a road boundary and can avoid static and dynamic obstacles; when the path nodes are beyond the maximum influence distance of the obstacle, a traditional gravitational potential field is added to the two adjacent path nodes in the bat position updating formula, and the path length is shortened; in the updating process of the local new solution of the bat algorithm, two strategies of Gaussian distribution and Cauchy distribution are used for searching to generate the local new solution, so that the searching speed and the searching precision are improved. Based on the improvement, static and dynamic obstacle avoidance can be realized while the path is ensured to be as short as possible in the driving process of the unmanned vehicle, and a smooth path is generated.

The invention relates to the technical field of electromagnetic release structure optimization, and discloses an electromagnetic release optimization method based on a whale algorithm, and the method comprises the steps: solving an optimal structure parameter through the whale algorithm, taking a position with the maximum counterforce as a design point, and taking the maximum electromagnetic suction value of the point as an optimization target, according to the national standard requirement, the condition that tripping is not carried out when the test current is equal to 80% of the short circuit setting current serves as the constraint condition, and mathematical modeling is carried out on the relation between the structure parameters and the electromagnetic attraction through SPSS analysis software. And then the whale algorithm and the optimization result of the orthogonal test are compared and analyzed, so that the rapid breaking of the electromagnetic release is better realized, and the short-circuit protection reliability of the electromagnetic release is improved.

The invention discloses an intelligent traffic signal lamp control method based on an improved particle swarm optimization algorithm. The intelligent traffic signal lamp control method comprises the following steps: step 1, data acquisition: 1.1, real-time data acquisition; 1.2, preprocessing the data; 2, initializing a particle swarm: 2.1, setting the number of particles; 2.2, initializing particle positions and speeds; 2.3, setting parameters; step 3, evaluating particle positions; 4, updating the speed and the position of the particle; 5, updating individual optimum and global optimum; step 6, introducing a self-adaptive inertia weight; 7, a local search enhancement mechanism is added, and the process is as follows: after each iteration, local search is performed on particles with relatively high fitness values, and the positions of the particles are further optimized; 8, when the maximum number of iterations is reached, iteration is stopped; and step 9, outputting a global optimal position, namely an optimal green light time allocation scheme. According to the invention, the green light time of the signal lamp can be dynamically adjusted to optimize traffic flow and reduce traffic jam.

The invention is suitable for the technical field of data search, and provides a data search method and device for an electromagnetic signal image-text, equipment, a medium and a product, and the method comprises the steps: obtaining a target electromagnetic signal image and target text content corresponding to to-be-searched data through employing an image-text generation model; extracting image semantic features of the target electromagnetic signal image and text semantic features of the target text content; fusing the image semantic feature and the text semantic feature to obtain a target image-text semantic feature; searching target reference image-text semantic features corresponding to the target image-text semantic features from a retrieval map corresponding to a knowledge base; and in the knowledge base, determining the reference electromagnetic signal image or the reference text content corresponding to the target reference image-text semantic feature as a search result of the to-be-searched data. The accuracy of searching the electromagnetic signal image is improved through searching the image-text semantic features and the retrieval atlas.

The invention belongs to the technical field of resource scheduling, and provides an online resource dynamic scheduling method and system for a power transformation intelligent patrol cloud platform, and the technical scheme comprises the steps: obtaining historical resource scheduling data of the power transformation intelligent patrol cloud platform; training a resource quantity prediction model based on the obtained historical resource scheduling data, and predicting the resource quantity required by the to-be-scheduled task based on the trained resource quantity prediction model; establishing a fitness function according to the total time and the total cost of the to-be-scheduled task, taking the fitness function as an optimization target, and scheduling idle resources of the service terminal by using a multi-target longhorn cattle herd combined artificial fish swarm hybrid algorithm to obtain an optimal resource scheduling scheme; the advantages of good global convergence of the artificial fish swarm algorithm and local rapid convergence, easy realization and the like of the Tiancattle swarm algorithm are comprehensively utilized, the defects that the convergence speed of the artificial fish swarm algorithm is low and the later global convergence of the Tiancattle swarm algorithm is poor are overcome, the advantages of the artificial fish swarm algorithm and the Tiancattle swarm algorithm are fully combined, and the optimal allocation problem of cloud resource scheduling can be well solved.

A building body flexible regulation and control potential prediction method comprises the steps that a building group energy consumption model is established based on a back propagation neural network, each energy consumption system sub-model takes characteristic data of influence parameters of energy consumption of each energy consumption system as input, and the maximum power saving rate of the corresponding energy consumption system is taken as a first-level optimization target; outputting energy consumption predicted values of the energy consumption systems in different time periods; determining the energy consumption prediction minimum value of each energy consumption system in different time periods by adopting a distribution entropy improved adaptive grid search particle swarm method and taking the maximum power saving rate of the building group as a second-level optimization target; updating the minimum energy consumption prediction value by using the deviation between the extreme value of each energy consumption system and the power saving rate of the building group; the energy consumption prediction minimum values of all the energy consumption systems in different time periods are spliced according to the time sequence to serve as the energy consumption prediction value of the building group; and taking the proportion of the difference value between the actual energy consumption value of the building group and the predicted energy consumption value of the building group in the actual energy consumption value of the building group as a prediction result of the flexible regulation and control potential of the building body, and realizing accurate prediction.

The invention relates to a wind-solar power fluctuation stabilizing method based on a hydraulic turbine set and energy storage, and belongs to the technical field of power grids. The method comprises the following steps: firstly, constructing a wind-light-water storage integrated power generation system; performing adaptive moving average filtering on the wind and light output power to obtain power meeting grid connection requirements and stabilizing reference power; then, parameters K and # imgabs0 # of variational mode decomposition are optimized through a whale optimization algorithm, reference power is stabilized through parameter-optimized VMD decomposition, and a mode component is obtained; performing Hilbert marginal spectrum analysis on each modal component obtained by VMD decomposition, and determining high and low frequency components according to a modal aliasing phenomenon; low-frequency signals are distributed to a hydraulic turbine set, high-frequency signals are distributed to a lithium battery, and wind and light power is stabilized. The method can effectively stabilize the fluctuation of the wind-light output power and is easy to popularize and apply.

The invention relates to the field of data processing, in particular to a pedestrian retrieval method and device based on trajectory data, a medium and equipment, and the method comprises the steps: obtaining a second actual trajectory point and a second actual elapsed time period corresponding to a target pedestrian through combining the position information and time information of the target pedestrian provided by wifi equipment, a public mobile communication base station and GPS equipment; according to the method, actual track information of a target pedestrian is clarified, second initial acquisition videos are acquired according to second actual track points and a second actual elapsed time period, and second target acquisition videos are screened from the second initial acquisition videos according to second similarity between each second initial acquisition video and a preset image of the target pedestrian. The screening accuracy of the second target acquisition video is improved, and the retrieval accuracy and the tracking accuracy of the target pedestrian are further improved.

The invention relates to an asynchronous motor fault identification method and system, and the method comprises the steps: collecting a vibration signal generated in the operation process of an asynchronous motor through a sensor device, and carrying out the time-frequency analysis of the collected vibration signal, and obtaining a time-frequency diagram; a plurality of movable points are initialized on the time-frequency graph, the standard deviation of the time-frequency graph in a window with the movable points as the center is calculated, and the horizontal movement direction and the step length of the movable points are determined based on the standard deviation; after horizontal movement, moving upwards or downwards by the step length in the vertical direction at the same probability, and if the value of the movable point in the time-frequency diagram after movement is smaller than that before vertical movement, determining the movement probability and the non-movement probability according to the movable point before vertical movement and two points above and below the movable point; updating the measurement matrix after the movable point moves to a new position; and extracting a ridge line of the time-frequency graph based on the final measurement matrix, and identifying an asynchronous motor fault corresponding to the ridge line.

The invention relates to the technical field of knowledge management and intelligent retrieval, and discloses a knowledge enhancement method and device based on context adaptive retrieval, equipment and a storage medium, which are used for improving the accuracy of a retrieval result. The knowledge enhancement method based on context adaptive retrieval comprises the following steps: acquiring post portraits and task context information, and generating context representation vectors; calculating a knowledge granularity weight vector according to the context representation vector; determining a target retrieval hierarchy based on the knowledge granularity weight vector, and performing retrieval operation on the target retrieval hierarchy to obtain a candidate knowledge set; performing fusion processing on the candidate knowledge set according to the knowledge granularity weight vector and the retrieval similarity to obtain a target knowledge result; and sending the target knowledge result to a service system.

This invention relates to the field of multi-target search using swarm robots, proposing a multi-source localization method based on virtual targets and a virtual population using particle swarm optimization (PSO) algorithm. This method utilizes a swarm of robots with certain communication and sensing capabilities to achieve multi-source target search and localization. The method equally divides the search area into multiple units, each with a virtual target at its center. Only one group of robots traverses all virtual targets and uses the PSO algorithm to search the corresponding regions of each virtual target, ultimately covering the entire search space. Common grouping methods in this field suffer from the difficulty in determining the number of groups, and their performance decreases as the number of source targets increases. The method proposed in this invention requires only one group to complete multi-source target localization, thus greatly reducing the number of robots needed and reducing physical costs. Furthermore, this method can be adapted to large-area passive scenarios.

The embodiment of the invention discloses a search interaction method and device, equipment, a storage medium and a program product, and the method comprises the steps: responding to a first access request of a target object to a target application, and displaying a content recommendation interface; if the target object is associated with the target media content, target search information is displayed on a content recommendation interface, and the target search information is obtained based on the target media content; in response to a trigger operation for the target search information, displaying a search result page; wherein the search result page comprises published content matched with the target search information. By adopting the embodiment of the invention, the search accuracy and search efficiency of the object can be improved.

The utility model discloses a functional composite reinforced guide rod for work dog smell recognition operation, and relates to the technical field of dog equipment support. A control handle is installed at the upper end of the rod body, a dog food storage box is installed on the rod body, the rod body is composed of a cavity part and a guide rod inner side wall, a metal elastic piece is installed in the cavity part, a lamp bead electric wire and a trigger mechanism used for triggering the metal elastic piece are arranged in the guide rod inner side wall, and the trigger mechanism is connected with the control handle; a dog food outlet is formed in the lower portion of the rod body, a telescopic thin rod and a lamp bead are installed at the tail end of the rod body, and the lamp bead is connected with a lamp bead wire. The control handle is provided with a clicker button, an elastic piece switch and a lamp bead switch, the clicker button and the elastic piece switch are in linkage with a trigger mechanism in the inner side wall of the guide rod, and the lamp bead switch is connected with the lamp bead through a wire. According to the utility model, rapid, accurate, orderly and efficient search training and using effects can be achieved, the purpose of strengthening trained dogs is achieved, and large-scale popularization and application are easy.

The present application relates to the technical field of deformation monitoring of water-wading slopes, and specifically to a deformation monitoring method, system and medium applied to rock water-wading slopes, the method comprising: obtaining point cloud data of each area in the water-wading slope in each sampling period; determining the rock deformation index of each cluster corresponding to each area in the water-wading slope in each sampling period; determining the rock complexity, slope instability value, and minimum segmentation threshold of each area in the water-wading slope in each sampling period, combining the spatial partitioning algorithm, and performing nearest neighbor search on the point cloud data of two adjacent sampling periods to obtain the nearest neighbor point cloud data corresponding to each point cloud data; determining the slope deformation degree of each area in the water-wading slope in each sampling period; and monitoring and warning the deformation of the rock water-wading slope. The present application is conducive to improving the search accuracy of the nearest neighbor point cloud data, reducing subsequent calculation errors, and improving the monitoring accuracy of areas with higher deformation risks in rock water-wading slopes.

The invention belongs to the technical field of energetic materials, and particularly discloses a high-throughput design method for a crystal structure of a ternary nitrogen-rich energetic material under high pressure, which is characterized in that specific dominant element components of the ternary energetic material are constructed based on an electronic counting rule, and a dominant structure search method is established in a structure search framework. Comprising the following steps: determining search components; searching a structure; calculating a convex hull graph; evaluating energy density and mass density; structure screening; evaluating the stability; analyzing an electronic structure; and evaluating detonation performance. According to the high-throughput design method for the crystal structure of the ternary nitrogen-rich energetic material under the high pressure, the search range is greatly narrowed in an extremely large search space phase of a ternary compound, the search precision and the search efficiency under specific proportioning components are improved, and the high-throughput design method for the crystal structure of the ternary nitrogen-rich energetic material under the high pressure is realized. And the crystal structure design time cost and the computing resource cost of the ternary nitrogen-rich energetic material are greatly reduced.

The application discloses a mobile device path planning method and an electronic device. The mobile device path planning method comprises the following steps: constructing a fitness function of mobile device path planning; initializing a particle swarm population, and a particle position is used for encoding a candidate path composed of multiple intermediate path point coordinates; performing multiple iterations of optimization, in each iteration, after each velocity and position update is completed, performing a dimension-by-dimension optimization operation on the entire population, the dimension-by-dimension optimization operation searches the path point coordinates of the candidate path encoded by all particle positions in the population dimension by dimension, and optimizes the corresponding path point coordinates in the global optimal position; after the iterations are completed, determining an optimal path based on the path encoded by the global optimal position. The application overcomes the defect that the effective local feature is lost due to simple elimination of low fitness particles in the existing PSO algorithm; meanwhile, repeated calculation of full path evaluation is avoided, and the real-time performance and high efficiency of algorithm operation are ensured while the optimization precision is effectively improved.

The invention belongs to the field of teaching quality evaluation, and particularly relates to a teaching quality evaluation method and system, and the method comprises the following steps: S1, uploading the content of teaching courseware to a database, and carrying out the keyword labeling; s2, collecting video information during teaching, and constructing a target function P; s3, optimizing the Pm, the Pe and the Pf by using an improved DBO algorithm; and S4, feeding back the optimized parameters to the deep learning algorithm, continuously adjusting the weight coefficient of each layer, continuously reducing the error between the true value and the predicted value, substituting the predicted values Pm, Pe and Pf output by the deep learning algorithm into the step S2, and calculating to obtain a score P according to a target function. The method has the advantages that an improved DBO algorithm is utilized, and a random walk strategy and a crisscross strategy are added in a rolling behavior for disturbance; and a direction following strategy is added in the stealing process, so that the optimizing precision is improved, and the performance of the method in the aspects of global exploration and optimizing precision is improved.

This invention discloses a method and system for injecting knowledge into large-scale models for vertical industries, belonging to the field of large-scale model knowledge injection technology. The method involves: collecting professional knowledge data from vertical industries, performing denoising, structured transformation, and terminology standardization; dividing the data into knowledge modules and associating them with vectorized indexes and metadata tags to form a vertical industry knowledge base; parsing the semantic feature vector of the input question using a dynamic routing classifier and routing it to the target knowledge module in the vertical industry knowledge base; employing a hybrid retrieval strategy to recall relevant document sets from the target knowledge module and deduplicate them to generate enhanced retrieval results; based on this, concatenating the enhanced retrieval results with the input question using industry-specific prompt templates; and inputting the LoRA-tuned large-scale model to generate accurate answers conforming to industry standards. This invention solves the problem of insufficient knowledge adaptability of large-scale models in vertical industries, improving the accuracy and applicability of large-scale models in vertical industries.

The application discloses a kind of low-carbon scheduling method, system, equipment and medium of power system, belong to power grid low-carbon scheduling technical field, method includes initialization parameter and constructs collaborative space;By particle swarm module optimization reinforcement learning hyperparameter;Reinforcement learning module interactive experience data generation;Extract priority sample and elite particle realize two-way experience sharing;Converge after output scheduling strategy.System includes initialization module, collaborative space construction module, hyperparameter optimization module, interactive storage module, two-way experience sharing module and strategy output module.The application is coupled by constructing dynamic collaborative optimization space and two-way experience sharing mechanism, particle swarm and deep reinforcement learning.Feedback guides particle swarm to realize adaptive disturbance, avoid local optimum trap.At the same time, spontaneous optimization network parameter, significantly enhance the adaptability and generalization ability of algorithm.The scheme breaks through single algorithm decision bottleneck, improves low-carbon scheduling efficiency and power system reliability.

Stable conformation search systems, methods, and media storing stable conformation search programs are disclosed. The system executes the following steps: obtaining N (N: gt; an integer of 1) main chain particles and N side chain particles correspondingly linked to the respective N main chain particles are arranged in a lattice space; calculating the coordinates of the (i + 1)-th main chain particle by using the coordinates of the i-th main chain particle and a first state variable represented by the relative coordinates between the i-th main chain particle and the (i + 1)-th main chain particle; calculating the coordinates of an (i + 1)-th side chain particle corresponding to the (i + 1)-th main chain particle by using the coordinates of the (i + 1)-th main chain particle and a second state variable represented by the relative coordinates between the (i + 1)-th main chain particle and the (i + 1)-th side chain particle; calculating a value of energy of the model when any one of the first state variable and the second state variable changes; a first state variable and a second state variable having a local minimum are identified.

The invention provides a search method and device, equipment and a medium, and relates to the technical field of search, in particular to the technical field of artificial intelligence, intelligent agents and generative search. According to the implementation scheme, in response to a received target search request, a historical search request corresponding to the target search request is determined, and the relevancy between the historical search request and the target search request is not smaller than a preset threshold value; acquiring historical search data corresponding to the historical search request, the historical search data including interactive behavior information of a historical user for a historical search result of the historical search request; determining a target search intention based on the target search request and the historical search data; and determining a target search result based on the target search intention.