36results about How to "Efficient search" patented technology

The application belongs to the field of workshop scheduling, and particularly discloses a semiconductor wafer probe test scheduling method and system based on an improved genetic algorithm, which adopts double-layer process-machine coding, and combines auxiliary resource taboo coding as an individual, effectively converts a complex scheduling problem into a coding form through an innovative coding and decoding strategy, and avoids traversal search on all selectable auxiliary resources; global-local hybrid search is carried out through the improved genetic algorithm, on the basis of machine level exchange with the traditional critical path-key block, search on the auxiliary resource immediately preceding work is increased, and neighborhood exploration is carried out based on a load balancing strategy, so that the search efficiency and solution quality are improved. The application can quickly and efficiently solve the semiconductor wafer probe test scheduling scheme, effectively improve the equipment utilization, speed up the scheduling efficiency, and break through the bottleneck of traditional scheduling.

The utility model discloses a kind of optical fiber sensing formula light-curing lamp of deformability, it is related to medical instrument technical field, including handle, the outer surface of handle is fixedly installed with lamp pole, the outer surface of handle is fixedly installed with display screen, the side of handle away from lamp pole is fixedly installed with charging end;The inside of lamp pole is provided with the adjusting mechanism of angle convenient to adjust, the adjusting mechanism includes the optical fiber light guide stick of deformability, the optical fiber light guide stick is movably installed in the outer surface of lamp pole.The optical fiber sensing formula light-curing lamp of deformability, angle adjustment is carried out by control switch groove silk drive optical fiber light guide stick, and optical fiber light guide stick is installed with sensor, the distance between real-time sensing and oral cavity interior photosensitive resin is facilitated, to provide position reference for medical staff, and angle adjustment is carried out by the optical fiber light guide stick of deformability, more accurate irradiation to oral cavity interior is realized, reduce the operation time of doctor, guarantee the solidification effect to photosensitive resin.

The application discloses a multi-modal large model multi-agent collaborative scheduling and distribution method, and relates to the technical field of agent deployment.The steps of the method include: receiving a complex task input by a user, decomposing the complex task into atomic subtasks, and generating a task decomposition graph; establishing an agent capability evaluation model, obtaining the comprehensive capability scores of the agents according to the capability behaviors of the agents in the candidate agent set; using the subtask types and the comprehensive capability scores of the agents to dynamically match and assign roles through a decision function, and generating a candidate agent set; and constructing a task-agent ant colony distribution algorithm according to the task decomposition graph and the comprehensive capability scores of the agents, and solving a distribution scheme.The application solves the problem that the existing multi-agent system lacks a dynamic role adjustment mechanism, cannot flexibly cooperate according to task progress and agent state, and is difficult to meet the requirements of an efficient and flexible task processing requirement of a multi-modal large model application development platform.

This invention provides a method for predicting the dominant frequency of blasting vibration by integrating parameter fluctuation processing and the Dream Optimization Algorithm (DOA), belonging to the field of underground engineering safety control technology. The system includes a data acquisition and uncertainty modeling module, a data preprocessing and feature construction module, a hyperparameter optimization module, and a prediction model training and result output module. The method obtains relevant parameters through on-site investigation and monitoring, and generates an extended sample set using a combination of probabilistic perturbation modeling and fuzzy triangular modeling. The data is preprocessed and features are selected. The DOA algorithm is used to globally search the hyperparameters of the Support Vector Regression (SVR) model, and the optimal hyperparameter combination is selected by combining a robustness fitness function. The SVR model is trained, and prediction results and uncertainty prediction intervals are generated. This invention can explicitly characterize the uncertainty of geological parameters, achieve efficient global optimization of hyperparameters, improve prediction accuracy, robustness, and generalization ability, and is applicable to different geological conditions and blasting scenarios. It can be extended to various blasting dynamic response prediction tasks.

This invention belongs to the field of marine engineering antifouling material technology and artificial intelligence optimization technology. It discloses a collaborative optimization method for the antifouling stability and safety of antifouling materials used in debris-blocking nets, comprising the following steps: Step 1, collecting multi-dimensional data and preparing samples of the antifouling materials for the debris-blocking nets; Step 2, constructing and training an AI prediction model: selecting key feature parameters from the multi-dimensional data, using these key feature parameters as core feature vectors to construct a dual-objective prediction model, which outputs predicted values ​​for the material's antifouling stability index and environmental safety index; Step 3, collaborative optimization under multiple constraints: setting multiple constraints for the material; using the hybrid meta-heuristic algorithm ALO-KHO to solve the multi-objective optimization problem; selecting the optimal solution from the Pareto optimal solution to obtain the parameter combination of the antifouling materials for the debris-blocking nets with the highest score. The prediction model of this invention has high accuracy, strong generalization ability, and significant multi-objective collaborative optimization effect.

The invention discloses a photovoltaic MPPT (Maximum Power Point Tracking) control algorithm, which relates to the technical field of photovoltaic control algorithms, utilizes the characteristics of simple structure, high convergence speed, high global optimization capability, wide applicability and the like of a gradient optimization algorithm, and improves the defects of premature convergence and sensitivity to control parameters of the gradient optimization algorithm. Meanwhile, the photovoltaic output energy efficiency and the stability of the photovoltaic output energy efficiency serve as the evaluation basis, an MPPT algorithm performance evaluation model is constructed through an analytic hierarchy process, photovoltaic MPPT performance is evaluated, and GTO performance is verified and improved.

This invention discloses a method, system, and device for detecting weak seismic signals based on instantaneous phase analysis, relating to the field of reservoir prediction. The method includes: 1. Inputting a three-dimensional seismic data volume; 2. Extracting the amplitude data of the current seismic trace and recording it as a one-dimensional array; 3. Performing a -90° phase shift on the one-dimensional array to obtain a phase shift array; 4. Performing a time-frequency continuous wavelet transform on the phase shift array to obtain multiple common-frequency amplitude arrays; 5. Performing a Hilbert transform on each common-frequency amplitude array to obtain the corresponding instantaneous phase array; 6. Calculating the second derivative of each instantaneous phase array with respect to the time variable n to obtain a three-dimensional data volume indicating weak signals at the current seismic trace's common frequency; 7. Changing the seismic trace and repeating steps 2-6 to obtain multiple three-dimensional data volumes indicating weak signals at the common frequency; 8. Selecting the optimal weak signal indicator data volume. This invention can effectively find the three-dimensional data volume indicating weak signals, improving the accuracy of detection.

The invention discloses a full-closed-loop radar signal level simulation system and method based on intelligent optimization. The system comprises a simulation scenario configuration module, a radar signal generation module, a target and environment model module, a radar signal processing module, a data processing and tracking module, a performance evaluation module and a closed-loop control engine. The method comprises the steps that firstly, a simulation scenario configuration module initializes a simulation scenario; then a radar signal generation module generates a radar emission signal, and a target and environment model module performs modeling to generate a synthetic echo; then the radar signal processing module receives the echo signal and extracts a target trace point, and the data processing and tracking module processes the target trace point to form a target track; and finally, the performance evaluation module executes radar signal level simulation and outputs performance indexes, and a closed-loop control engine judges whether acceptance conditions are met or not and outputs optimal parameters and performance reports. According to the invention, intelligent full-closed-loop automatic simulation is realized, and the precision and flexibility of radar simulation are improved.

The application discloses a monitoring point layout optimization method and system based on regional scale nitrogen leaching, and relates to the field of agricultural environmental protection.The method comprises the following steps: acquiring a regional area and regional data; based on the regional data, drawing a slope grade map, a groundwater level grade map and a nitrogen accumulation map of the region respectively; performing pixel-to-pixel alignment, superposition and combination calculation on the slope grade map, the groundwater level grade map and the nitrogen accumulation map of the region to generate a plurality of basic units; under a constraint condition, performing initial layout of monitoring points in each basic unit according to the regional area, an obstacle layer and a preset monitoring point spacing; based on an average distance minimization criterion, constructing a target function of the monitoring points; and based on the target function, optimizing the monitoring points in the initial layout by using a space simulated annealing algorithm to obtain an optimized layout of the monitoring points.The application can improve the monitoring accuracy of regional scale nitrogen leaching and the layout efficiency of monitoring well points.

The application discloses an artificial intelligence psychological evaluation method and system based on multi-modal input, and belongs to the technical field of artificial intelligence, and solves the problem that the existing system does not consider individual differences of the tested person to generate a customized evaluation strategy, resulting in insufficient evaluation targeting of the tested person, the method comprises the following steps: generating a personalized evaluation psychological evaluation strategy of the tested person based on basic information of the tested person, collecting multi-modal evaluation information of the tested person in real time, and identifying and analyzing a comprehensive state characteristic vector based on a pre-constructed psychological evaluation model; the application utilizes basic information to construct an initial feature portrait, uses test appeal as a constraint, adopts a collaborative attention network to generate strategy demand information, and individually adjusts item priorities based on an adjusted feature portrait, so that an evaluation strategy that adapts to individual cognitive preferences and emotional characteristics is generated, different individual differences can be adapted to, and implicit and defensive psychological problems of the tested person can be accurately captured, and the accuracy and efficiency of psychological evaluation are improved.

The application relates to the technical field of network security, and discloses an encrypted traffic identification system and method based on multi-modal fusion, which realizes intelligent type identification of network attacks on encrypted traffic in network communication by combining different network attack type standard encrypted traffic abnormal connection interaction feature information established by combining artificial intelligence algorithms and standards with network attack type standard encrypted traffic abnormal connection interaction feature information; scientific detection of network attack types on encrypted traffic in network communication is realized by combining different network attack type standard encrypted traffic abnormal connection protocol feature information established by combining data analysis and science, multi-level intelligent analysis of network attack types based on communication time sequence features, communication packet size features, communication traffic statistical features, communication connection interaction features and communication connection protocol features of encrypted traffic is realized, and intelligent and efficient supervision of communication network anomalies based on multi-modal behavior features of encrypted traffic is realized.

The utility model discloses a file archive frame for collecting and arranging market research data. The file archive frame comprises a frame body, radial partition plates and a rotating base. The rack body is formed by connecting at least three layers of coaxial circular shelves through vertical supporting columns, each layer of shelf is divided into a plurality of fan-shaped file containing areas through radial partition plates, and an annular rotating groove is formed in the bottom of the rotating base and can drive the rack body to rotate stably. By means of the multi-layer coaxial rotation design, fast file searching is achieved in cooperation with fan-shaped partition, and the problems that a traditional frame body needs to be surrounded for searching, and rotation is not stable are solved; materials such as the 304 stainless steel supporting column and the ABS partition plate are adopted, existing auxiliary devices such as folders and labels are matched, calibration and fixation are conducted through a screwdriver and a gradienter during installation, files can be taken and used by rotating the shelf during use, maintenance is convenient and fast, and the device is suitable for efficient arrangement and retrieval of market research data.

The invention discloses a simulated annealing standardization correction method and system for density logging data, and belongs to the technical field of logging data processing. Comprising the steps of collecting density logging data of a standard well and original density logging data of a to-be-corrected well, preprocessing the original density logging data of the to-be-corrected well, establishing density data histogram distribution based on a linear mapping model, establishing a histogram overlapping comparison evaluation system, and calculating a comprehensive matching score. Optimizing calculation is conducted on the linear mapping model based on a simulated annealing algorithm, linear transformation parameters are automatically optimized, the optimal parameters are applied to full-well-section density data correction, and comparison between original density data and an anti-normalization processing result is achieved through anti-normalization processing. The method effectively overcomes the defects that a traditional method depends on artificial experience and is low in correction precision, data quality and processing efficiency are remarkably improved, and a reliable data basis is provided for reservoir fine evaluation.

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.

This paper proposes a baseline-length-constrained high-precision relative positioning method for BeiDou, comprising the following steps: First, establishing a double-difference relative positioning model using the pseudorange and carrier phase raw observations of the reference station and the rover station; Second, obtaining a strictly baseline-length-constrained objective function based on the double-difference relative positioning model according to the least squares criterion; Third, employing an integer ambiguity search space amplification strategy to achieve effective ambiguity search, and accurately estimating the rover station's position relative to the reference station. This invention utilizes the often underutilized baseline length information of dual antennas in real-world scenarios to improve the relative positioning solution performance, exhibiting higher reliability compared to the traditional unconstrained case. The method employs an integer ambiguity search space amplification strategy, thereby extending the standard LAMBDA algorithm to baseline-length-constrained relative positioning, achieving effective integer ambiguity search.

The application discloses a RSA factorization method and system based on three-dimensional space search, and the method comprises the following steps: obtaining an RSA modulus; setting a maximum iteration number to enter a loop step; randomly selecting a calculation S + =F + (k, p, q) and S ‑ =F ‑ (k, p, q); taking S + and S ‑ as a base point, dividing each hyperboloid into four sub-hyperboloids; if N is on a certain sub-hyperboloid, outputting (k, p, q); otherwise, returning to the sub-hyperboloid division step; until N is found or the search limit is terminated; if N is not found, entering a loop of the next k. The system comprises a memory and a processor for executing the RSA factorization method based on three-dimensional space search. By using the application, a large integer factorization task can be completed. The application can be widely applied in the technical field of cryptography.

The application relates to a kind of unmanned aerial vehicle auxiliary-based vehicle task unloading method, its steps are: first, the task queue model, calculation model, communication model of vehicle, roadside unit and unmanned aerial vehicle and the energy consumption and collection model of unmanned aerial vehicle are constructed;Second, under the constraint conditions such as meeting task queue stability, long-term energy sustainability of unmanned aerial vehicle, the optimization problem with the minimum long-term average total task processing delay of system as target is constructed, and the long-term optimization problem is converted into deterministic single time slot optimization problem using Lyapunov optimization;Finally, in each time slot, according to the current system state, the deterministic optimization problem is solved using genetic algorithm, so that the task unloading and unmanned aerial vehicle path planning scheme of current time slot are obtained.The method can effectively improve the task processing efficiency and guarantee the system queue stability, reduce the task processing delay, and thus improve the overall performance of the vehicle edge computing network in the dynamic uncertain environment combined with the Lyapunov optimization framework.

The invention discloses a solid wood furniture board intelligent nesting method considering texture direction constraint, which comprises the following steps of: synchronously generating a pixel-level defect segmentation image and a texture direction field from a board digital image by adopting a deep neural network; for each to-be-trepanned part, constructing a comprehensive value map by fusing a defect repulsive force field generated by the defect segmentation map and a texture coordination map generated by the texture direction field; iteratively searching the optimal placement pose of each part on the value map by adopting a hybrid optimization strategy combining an artificial potential field method and simulated annealing; and after the pose of one part is determined, the area occupied by the part is updated to be an unavailable area, and the process is repeated for the next part until nesting of all the parts is completed. The problems that in the prior art, defect perception and trepanning optimization are disjointed, texture constraint processing is insufficient, and an optimization algorithm is prone to falling into local optimum are solved, and the plate utilization rate and the product aesthetic quality can be improved.

The invention discloses a cyclone feeding concentration prediction method, relates to the field of iron ore grinding and separation feeding concentration prediction, and aims to solve the problem that an ore grinding process cannot be effectively optimized in real time due to the fact that ore grinding parameters are difficult to measure in real time in the current ore separation production process. According to the method, a lion group optimization algorithm is improved, the global search capability is improved by changing a lion king updating mode and introducing an attenuation factor, the problem that the algorithm is trapped in a local optimal solution is avoided, a dynamic learning strategy is introduced for a parent lion and learning lion updating mode, and the local search capability is optimized; the performance of the lion group algorithm in a complex prediction task is effectively improved, and the limitation of a traditional algorithm in solving a high-dimensional complex problem is avoided.

The application discloses a path planning method based on elephant trunk fish optimization strategy, which comprises the following steps: firstly, generating an initial environment map, initializing the iteration times and N elephant trunk fishes, and calculating the fitness of each elephant trunk fish; secondly, saving the elephant trunk fish with the minimum fitness as the optimal elephant trunk fish, detecting and generating sampling points for each elephant trunk fish, calculating the fitness of the sampling points, and if the fitness is better than the current position, the elephant trunk fish will move to the sampling point, if no sampling point with better fitness is found, the elephant trunk fish will move to the optimal position of the population or a random position in the solution space according to a random probability; then, calculating the fitness of the elephant trunk fish after updating, iterating, taking the optimal elephant trunk fish as the global optimal path, and finally converting the generated path into navigation instructions. The application can quickly find the optimal or nearly optimal path in a complex environment, effectively explore the search space, and avoid the local optimal solution trap.

The application discloses a wireless power transmission system multi-parameter collaborative optimization method based on a VPSR-PSO algorithm, and belongs to the technical field of wireless power transmission, and comprises the following steps: step 1: according to the primary side input voltage, the primary side input current, the primary side output current, the primary side compensation inductance, the primary side compensation inductance internal resistance, the primary side coil self-induction internal resistance and the primary side coil self-induction of the system, the secondary side output current, the secondary side coil self-induction, the primary side coil series capacitor, the secondary side coil self-induction internal resistance and the secondary side coil series capacitor, and the mutual inductance between the primary side coil and the secondary side coil; step 2: constructing a target model according to the primary side compensation inductance, the mutual inductance between the primary side coil and the secondary side coil and the equivalent load resistance; and establishing the constraint condition of the target model; the application constructs a multi-objective model for simultaneously optimizing transmission efficiency and output power, and realizes efficient search and balanced regulation and control.

The invention relates to a method and equipment for generating a project combination scheme and a medium, and the method comprises the steps: firstly, recognizing four types of association relationships of collaboration, mutual exclusion, time sequence and conflict among projects, and constructing a four-layer multiplexing network for modeling; on the basis of a network identification cooperation and conflict project combination, value emergence, cost allocation effect and value suppression and resource competition effect brought by the combination are quantified respectively. On the basis, a multi-objective optimization model which takes project selection as a variable and minimizes the total cost and maximizes the total value is constructed, and model constraints include time sequence, mutual exclusion and strategic objective support requirements. And solving the model by adopting an improved genetic algorithm based on time sequence chain coding, generating a Pareto optimal solution set, and extracting a final scheme from the Pareto optimal solution set. The method can systematically quantify the combination effect generated by interaction between projects, so that the project combination scheme with the comprehensive optimal cost and value is generated under multiple constraints, and scientific decision making is assisted.

The invention discloses a software defect feature selection method, device and system and a storage medium. The method comprises the following steps: S1, carrying out feature coding on software measurement features of a software defect historical data set; s2, performing sectional probability initialization on the initial population by adopting a minimum redundancy and maximum correlation algorithm according to the coded software measurement characteristics to obtain a plurality of sub-populations; s3, performing parallel chromosome evolution operation and crossover mutation operation on each sub-population; and S4, after parallel evolution and overall crossover variation of each sub-population, putting a parent chromosome and a generated sub-chromosome sum into a chromosome pool, selecting an optimal individual chromosome according to a fitness descending order, then carrying out remodeling of the sub-population, and continuing to repeat genetic manipulation until the maximum number of iterations is reached. By adopting the technical scheme of the invention, a high-quality feature combination which is most effective to predict defects and has the lowest redundancy is quickly and accurately screened out from high-dimensional software features.

The invention discloses a robot autonomous target searching method for an unknown environment, and belongs to the field of artificial intelligence and robot control. The method comprises the following steps: converting environment data collected by a robot at a current position into point cloud, generating an observed area, detecting a target in the observed area, and outputting the position and confidence of the target; meanwhile, based on the observed area, an unobserved area is iteratively predicted through a pre-training generation model, and the probability of occurrence of a target is counted. And fusing the information of the two to construct a comprehensive belief state, simulating the movement of the robot in a discrete action space by using an Image-POUCT algorithm based on the state, and scoring and selecting an optimal movement position through a dense reward function. The process is continuously iterated until a sufficient number of targets with the confidence reaching the standard are found, and all target positions are output; according to the invention, efficient intelligent search of the robot in a large-range unknown environment can be realized.

The embodiment of the invention discloses a method for constructing a topology description function, which comprises a first function and a second function, wherein the first function takes the parameter coordinate as a variable and is used for determining a point which corresponds to the parameter coordinate and is used for determining a spline assembly center line of the variable thickness spline assembly in a preset design domain and a first coordinate in a world coordinate system, and the second function takes the parameter coordinate as a variable and is used for determining the thickness of the variable thickness spline assembly at the parameter coordinate; the first function and the second function are continuous differentiable mathematical operation functions; constructing a third function according to the first function and the second function; and constructing a topology description function according to the third function. A global continuous differentiable topology description function is constructed through the combination of a first function, a second function and a third function which are continuously differentiable, and the topology description function can promote the application of the topology description function in gradient-based topology optimization and can improve the effectiveness of optimization.

This invention discloses a crowd evacuation method that considers crowd disorder and congestion. It describes the state of crowd disorder and congestion using velocity entropy and spatiotemporal congestion, and incorporates these factors into the heuristic function of the A* algorithm to obtain a disorder impact factor and a congestion cost, which describe the influence of crowd disorder and congestion on the planned path. Finally, a social force model is used to drive pedestrians to evacuate along the planned path. This invention utilizes velocity entropy and spatiotemporal congestion as influence factors for the A* algorithm, ensuring that the path planned by the A* algorithm simultaneously considers the effects of crowd disorder and congestion. Then, a replanning mechanism is proposed based on velocity entropy and spatiotemporal congestion to meet real-time requirements. Finally, the social force model drives pedestrians to move along the evacuation path, thus retaining the advantages of the social force model in pedestrian-environment interaction while improving upon the lack of a pathfinding mechanism through the A* algorithm.

The present application relates to a kind of quantum approximation optimization-based multi-antenna wireless communication system beam forming method, system, belong to multi-antenna wireless communication field, comprising the following steps: S1, according to the channel state information and the wireless resource weight, construct the weighted channel correlation matrix reflecting the characteristics of multi-user integrated channel;S2, the continuous phase of each transmitting antenna is discretized, and a gradually doubled binary variable is sequentially assigned to the discrete phase, the phase combination of all antennas is jointly constructed to determine the state space corresponding to quantum approximation optimization algorithm corresponding to quantum state representation of quantum approximation optimization algorithm;S3, by quantum approximation optimization algorithm, quantum circuit evolution and observation are executed, and the power allocation quantum optimization initial parameter is updated and optimized by the gradient descent of classical algorithm, to obtain power allocation quantum optimization optimal parameter;S4, power allocation quantum optimization optimal parameter is converted into the phase combination of all antennas, and the optimal power allocation beam is obtained.

The present disclosure provides a layout structure determination method and device, equipment and a storage medium, relates to the technical field of computers, in particular to the technical field of quantum computers, quantum chips and quantum simulation. The specific implementation scheme is: obtaining a target search range of a geometric parameter of a chip layout structure to be designed based on a target parameter required by the chip layout structure to be designed; the target search range represents a value range of the geometric parameter of the chip layout structure to be designed; the value of the geometric parameter can affect a capacitance parameter of the chip layout structure; obtaining a target value of the geometric parameter of the chip layout structure to be designed in the target search range based on the target parameter; obtaining a target chip layout structure based on the target value of the geometric parameter; the value of the geometric parameter of the target chip layout structure is the target value, and difference information between a simulation parameter corresponding to the target chip layout structure and the target parameter satisfies a preset difference requirement.

The application discloses a single-page application search method, device and equipment and a computer storage medium. The single-page application search method comprises the following steps: receiving a hypertext markup language (HTML) file and data resources of a single-page application of a target user; parsing the HTML file to obtain a document object model (DOM) tree, and loading a JavaScript file relied on by the HTML file; executing the JavaScript file, replacing corresponding nodes in the DOM tree with the HTML file and the data resources of the target user, and generating a first static HTML resource; calling an anchor point parser, replacing anchor point links in the first static HTML resource with non-anchor point links, and obtaining a second static HTML resource; and sending the second static HTML resource to a gateway. In this way, the network crawler can comprehensively capture the content of the single-page application, and the search engine can effectively search the related content of the single-page application.

The application provides a MIMO antenna design optimization method, device, equipment and medium, and belongs to the antenna design field. The method comprises the following steps: obtaining historical design parameters and corresponding performance indicators of a MIMO antenna; preprocessing the historical design parameters and the performance indicators, constructing a mapping relationship between input data and output data, and training a forward prediction model and a generative model based on the mapping relationship; receiving a target performance indicator input, generating a candidate design parameter set by using the trained generative model, taking the candidate design parameter set as an initial population, combining the forward prediction model as a surrogate model, adopting a multi-objective evolutionary algorithm for iterative optimization, and outputting a Pareto optimal solution set as a target design parameter set. In this way, the application greatly speeds up the performance evaluation link with the help of the surrogate model, provides diversified non-inferior solution selection through Pareto frontier analysis, and can flexibly trade off different design targets according to actual requirements.