199 results about "Quantitative model" patented technology

The invention provides a target recognition model reasoning optimization method and device, and the method comprises the steps: firstly carrying out the structural analysis and sensitivity evaluation of a pre-training model, extracting the structural features of each network layer, activating the distribution features, carrying out the quantitative sensitivity scoring, and constructing a data set reflecting the hierarchical features and fault-tolerant capability; and querying a quantitative configuration knowledge base based on the data set to generate a heterogeneous quantitative strategy. Layered low-bit quantization is executed according to the strategy, and a layered weighted loss function is introduced to carry out quantization perception training, so that precision loss caused by bit width compression is effectively compensated. According to the method, through hierarchical heterogeneous quantification, the model recognition precision is preserved to the maximum extent while high compression ratio and reasoning acceleration are achieved, and particularly, the performance of a high-sensitivity layer is protected. The generated heterogeneous quantitative model remarkably reduces memory occupation and power consumption, is suitable for an edge hardware platform with limited resources, forms a set of complete automatic process from analysis and configuration to training compensation, and has good universality and engineering practical value.

The invention discloses a damping thermal insulation material performance evaluation method and system for a battery pack, and relates to the technical field of battery pack thermal management, and the method comprises the steps: extracting a calibration sample set, applying transient thermal excitation, and collecting an infrared thermal image sequence to extract thermal response features; performing industrial CT scanning to obtain defect quantitative indexes, and forming a calibration data set with the thermal response characteristics; establishing a quantitative correlation model based on the calibration data set, and determining a batch quality threshold according to model output distribution; executing the same excitation and collection on an online sample to be detected to extract the characteristics of the sample, inputting the model to obtain a defect quantitative prediction value, comparing the defect quantitative prediction value with a batch quality threshold, and outputting an evaluation result; when a preset period or a trigger condition is met, a new sample is supplemented and extracted, and newly added calibration data is obtained to update the model and the threshold value. According to the method, accurate material performance evaluation is provided by establishing a quantitative model of thermal response characteristics and defect quantitative indexes, and evaluation consistency and stability are enhanced through a dynamic updating mechanism.

The invention discloses a meta-model decision-making-based intelligent prediction method for inclusion quality in an electroslag remelting process. The method comprises the following steps of: constructing a sample data set containing process and component characteristics and target variables; obtaining a first-layer basic model based on an SHAP value cumulative contribution rate screening method; a heterogeneous learner is adopted to construct a first layer structure of the stacked integrated learning model, and Bayesian is adopted to carry out adjustment and optimization; adopting logistic regression as a meta-learner to construct a second-layer structure of the stacked integrated learning model; training by adopting a five-fold cross validation strategy, and predicting the performance by using a multi-index quantitative model; and deploying the D-type inclusion prediction model in the electroslag remelting process to an actual process, collecting process parameters as input data in real time by using a multi-sensor group, performing D-type inclusion risk prediction of a corresponding heat, and outputting an inclusion risk prediction result. According to the method, the accuracy and reliability of D-type inclusion prediction can be remarkably improved.

The invention discloses a method for heterogeneous scheduling of computing power under a large model training reasoning framework, relates to the technical field of computing power scheduling, and solves the technical problem that task interruption or resource waste is easily caused by incomplete task migration, splitting and recovery mechanisms when computing power is insufficient or equipment is abnormal. According to the method, parameters such as the load rate, the residual computing power and the temperature of the CPU, the GPU and the FPGA are dynamically collected, the calculation amount, the duration and the delay requirement of the task are combined, a quantitative model is established, the resource utilization rate is increased, a temperature prediction model is established in stages, a dynamic threshold value is set in combination with the equipment type, the environment temperature and the health degree, temperature early warning and migration in the task execution process are achieved, and the task execution efficiency is improved. The frequency reduction risk caused by overheating of equipment is reduced, the task interruption loss is ensured to be reduced by more than 50% through task splitting, intelligent migration and pause recovery mechanisms aiming at insufficient computing power or equipment abnormity, and the stability of the system during peak flow or hardware failure is improved.

The invention discloses a special vehicle driving task performance and sitting posture comfort coupling-oriented test evaluation system and method, and the system comprises a data collection module which is used for obtaining multi-modal driving data; the index calculation module is used for extracting and quantifying characteristic indexes from the multi-modal driving data to obtain a driving task performance evaluation index set and a driving sitting posture comfort evaluation index set; the comprehensive evaluation and time domain modeling module constructs a driving task performance-sitting posture comfort coupling-oriented comprehensive quantitative model based on the driving task performance quantitative model and the driving sitting posture comfort quantitative model; and finally, comprehensively scoring the state of the driver by using the comprehensive quantitative model, and establishing a nonlinear mapping relationship between the driver state comprehensive score and the task time to obtain a dynamic change trend of the driver state comprehensive score.

The invention provides a reasoning acceleration method, electronic equipment, a readable storage medium and a program product. The reasoning acceleration method comprises the steps that an input text is received, a quantitative model is called, in the quantitative model, all tensors comprise m-bit tensors expressed by m bits and n-bit tensors expressed by n bits, and m is larger than n; decomposing the input text into input lexical elements; calling an m-bit tensor or an n-bit tensor in the quantitative model to calculate the input lexical elements to obtain key vectors and value vectors of the input lexical elements, and storing the key vectors and the value vectors in a key value cache; calling an n-bit tensor in the quantization model, and generating an output lexical element according to a key vector and a value vector in the key value cache; and obtaining a reasoning result based on the output lexical elements.

The invention relates to a fault prediction method and system for parameter fine tuning and adaptive structure pruning based on business influence driving, and belongs to the technical field of artificial intelligence, time series data analysis and intelligent operation and maintenance. Comprising the following steps: S1, business influence data modeling: integrating multi-source operation and maintenance data and historical business fault event data, and constructing a business influence quantitative model; s2, efficient fine tuning of service influence guide parameters: loading the pre-training fault prediction model, and performing efficient fine tuning of the parameters on the basis of service influence signals; s3, business influence driven adaptive structure pruning; S4, dynamic reasoning and alarm generation: realizing a dynamic reasoning mechanism of a pruned fault prediction model, and generating a structured and business value oriented fault prediction alarm according to a reasoning result; and S5, performing closed-loop optimization and continuous evolution. According to the invention, the configuration accuracy of the operation and maintenance resources is improved.

The invention discloses a smart city planning method and system based on digital twinning, relates to the technical field of city planning, and solves the problems that the scheme is not feasible due to dependence on manual interpretation of regulation texts, lack of automatic generation and matching of index constraint lists and possible neglect of key constraint conditions in planning suggestions, and meanwhile, the scheme is high in subjectivity and low in cost. And a unified quantitative model is lacked. Through the NLP technology, planning suggestions and regulation terms are compared in real time, conflict indexes are automatically marked, planning conditions and preliminary index calculation results are combined, the generated project planning suggestions strictly follow a policy framework, and manual interpretation errors or violation risks are avoided. A collaborative quantitative evaluation model of the single building and the regional landscape is established, and the landscape effect of the planning suggestion is evaluated from the dimensions of spatial layout, visual vestibule, skyline, color coordination and the like. The fusion degree of the project and the surrounding environment is visually reflected through quantitative scoring, and the deviation of subjective evaluation is avoided.

The invention relates to a digital twinning-based multi-device dynamic cooperative control method for a cut tobacco production line, and the method comprises the steps: 1-5, constructing a digital twinning body of the cut tobacco production line through employing a three-dimensional unified modeling frame of a geometric model, a physical model and a behavior model, constructing a distributed cooperative control network architecture based on edge calculation, and constructing a multi-device dynamic cooperative control system of the cut tobacco production line; data interaction is realized among all layers by defining an interface function, modeling is performed on interaction among devices by adopting a GNN neural network, a global coupling relation quantitative model of a device cluster is established, time sequence alignment is performed on multi-sampling frequency data by adopting a time distortion function, abnormal data is detected and eliminated by adopting an anomaly detection method based on a mahalanobis distance, and the device cluster is obtained. A three-dimensional evaluation system of key indexes is constructed, a comprehensive evaluation function is calculated, a two-stage architecture of a steady-state optimization layer and a dynamic adjustment layer is adopted, and global collaborative optimization of the cut tobacco production line under a dynamic working condition is realized by combining digital twinning rehearsal verification, virtual and real millisecond-stage synchronization, three-stage early warning and fault collaborative self-healing.

The invention relates to the technical field of financial risk assessment, in particular to an intelligent financial risk assessment method and system based on multi-mode dynamic graph combined efficiency improvement, and the method comprises the steps: mining the implicit correlation characteristics of an enterprise / individual in a complex financial network through the combination of multi-mode data fusion and a dynamic knowledge graph; and precise evaluation and real-time early warning are realized in combination with time sequence analysis and an external event quantitative model. According to the method, implicit associated feature information of enterprises / individuals in a complex financial network is mined by adopting a mode of combining multi-modal data fusion and a dynamic knowledge graph, and accurate assessment and real-time early warning of financial risks are realized by combining time sequence analysis and an external event influence quantitative model, so that the risk assessment efficiency is improved. The financial risk assessment efficiency and accuracy can be improved, and supervision compliance verification is supported.

The invention provides a smart city data linkage method based on digital twinning, and the method comprises the steps: constructing a fusion modeling layer, constructing a lightweight digital twinning model through CIM, BIM and GIS technologies, packaging a scenarized simulation model through a micro-service architecture, and forming a data linkage space carrier; establishing a space-time-event double-main-line data linkage mechanism by taking the scene as a reference, accessing multi-source data through a stream processing engine and associating the multi-source data to a space coordinate system to form a dynamic data link; a semantic mapping layer based on Ontology unifies data modes and associates and maps the data modes, a cross-domain data linkage channel is constructed, and heterogeneous data semantic-level linkage is achieved; recognizing the state change of the facility through a time sequence model, and generating a cross-department automatic instruction chain in combination with spatial association to form scenarized closed-loop linkage; and constructing a decision support linkage system by using feedback data, integrating a quantitative model and a deduction engine, and converting the data into visual decision information to complete a full-chain scenarized closed loop. The method solves the problems of traditional data dispersion and the like, improves the linkage accuracy and processing efficiency, and provides support for urban governance.

The invention provides a power distribution system source network load storage multi-time scale layering and grouping scheduling control method, and relates to the technical field of power distribution system scheduling control. The power distribution system source network load storage multi-time scale layering and grouping scheduling control method comprises the following steps of 1, source network load storage adjustable resource quantitative evaluation, 2, multi-time scale power prediction, 3, layering and grouping architecture construction, 4, multi-time scale scheduling strategy making, and 5, scheduling strategy execution and rolling updating. According to the invention, by constructing the quantitative model containing the algorithm formula and the hierarchical grouping architecture, the problems of power balance, scheduling economy and power grid safety under high-proportion new energy access are solved, the new energy consumption rate and the system flexibility are improved, the source-grid load-storage collaborative interaction is realized, and the new energy consumption capability and the power grid operation stability are improved.

The invention discloses a battering ram multi-station online detection system and detection method, and belongs to the technical field of battering ram production detection. The system comprises a single-station basic detection unit, a multi-station traceability linkage core module, an abnormal conduction path analysis module and a self-adaptive threshold optimization module which interact through an industrial Ethernet, and is additionally provided with a key process multi-dimensional reinspection unit, a dynamic interference compensation module and the like. The method is based on the system, and detection is realized through station data acquisition and preprocessing, dynamic interference compensation, comprehensive quality evaluation, quantitative traceability, grading processing, threshold optimization and data storage. The core of the method is that a quality problem root station is accurately positioned through a quantitative model, an abnormal conduction path is traced, a detection threshold value is dynamically optimized in combination with traceability data, a'detection-traceability-optimization 'closed loop is formed, the problems that traditional manual traceability is low in precision, poor in efficiency and disjoint in traceability and detection are solved, the method is adaptive to a flexible production scene, resource waste is reduced, and production efficiency is improved. And the production quality and efficiency of the battering ram are improved.

The invention specifically discloses a data skeleton matching method and system based on an intelligent Internet of Vehicles environment, and the method comprises the steps: carrying out the type division of a vehicle according to the automatic driving level of the vehicle, calculating the specific data demand intensity, scene emergency degree score and communication constraint degree of the vehicle, constructing a quantitative model based on a spatial view level demand, calculating a view demand factor, and determining the spatial view level demand; global fusion is carried out on the vehicle automatic driving level, the specific data demand intensity of the vehicle, the scene emergency degree score, the communication constraint degree and the view demand, and the precision level and scale of a final data skeleton are output; and according to the precision grade and the scale of the final data skeleton, the road side unit encapsulates the sensing data and sends the sensing data to the requesting vehicle. By adopting the technical scheme, the personalized and accurate matching of the data response in the vehicle-road cooperation system is realized through a multi-level and self-adaptive decision process, and the communication resource utilization rate, the data transmission real-time performance and the vehicle driving safety are improved.

The invention discloses a coal-based solid waste road base full life cycle dynamic carbon accounting method and system, and relates to the field of carbon accounting, and the method comprises the steps: dividing solid waste types according to component characteristics, building an identification system, and collecting the dynamic parameters of each stage through a monitoring device to form a basic database; building a staged quantitative model based on the data, and correcting an accounting result in combination with a scene-carbon emission correction matrix; predicting a carbon emission curve of 5-10 years of an operation period by using an LSTM neural network, and establishing a carbon balance equation of the waste stage recycled aggregate; and finally, analyzing parameter uncertainty through Monte Carlo simulation, dynamically updating a database and outputting a carbon emission confidence interval. The coal-based solid waste road base carbon accounting method has the advantages that the coal-based solid waste road base carbon accounting is accurately, dynamically and comprehensively completed by refining the full life cycle stage and combining the staged quantitative model, scene correction, LSTM prediction, regenerated carbon balance calculation and Monte Carlo uncertainty analysis, and a basis is provided for long-term planning.

The invention discloses a novel information system reliability determination method for a multi-input dependent starting scene, and belongs to the technical field of system reliability. Comprising the following steps: drawing a system structure chart containing a multi-input dependency starting module, and dividing a three-stage system structure of a core module-a support signal module-an auxiliary guarantee module; a GO graph is drawn, the multi-input dependent starting module is uniquely identified as a type 29 operator, and the number of input ports of the operator is in one-to-one correspondence with the number of support signals of the corresponding core module; and determining the signal effective probability, the correlation coefficient and the element reliability according to a standardized process, calculating the module reliability by combining a type 29 operator quantitative model, and finally determining the system reliability. According to the method, the problem that an existing GO method operator cannot adapt to a multi-input dependent starting scene is solved, the blank of a special tool for GO method modeling in the scene is filled up, and the application range of the GO method in a novel information system is expanded.

The invention provides a power grid equipment investment comprehensive benefit evaluation and optimization method and system and a medium, and belongs to the technical field of power grid benefit evaluation and optimization. According to the method, the comprehensiveness of evaluation indexes is ensured by constructing a multi-dimensional index system, covering functions and economy and environment multiple dimensions, and meanwhile, the evaluation efficiency is improved by constructing a function contribution degree quantitative model, an investment and income allocation model, a function and cost coupling benefit model and a region and equipment collaborative optimization model. A coupling model between equipment operation characteristics and investment cost is established, an area and equipment collaborative optimization mechanism is introduced, quantitative evaluation, dynamic comparison and optimal configuration of power grid equipment investment benefits are realized, and optimization of decision demands is realized while comprehensiveness and stability of power grid equipment investment comprehensive benefit evaluation are improved.

The invention relates to a method and device for obtaining grouting pressure distribution, a storage medium and electronic equipment, and the method comprises the steps: determining a monitoring hole for monitoring a grouting hole in a target grouting region according to a monitoring layout strategy; a grouting pressure sensor, a soil pressure sensor and a pore water pressure sensor are respectively arranged at different depths of the monitoring hole; the monitoring holes are backfilled, and grout is injected into the grouting holes in a segmented mode according to the grouting technology; in the grout injection process, a grouting pressure sensor, a soil pressure sensor and a pore water pressure sensor are synchronously triggered to collect grouting pressure data, soil pressure data and pore water pressure data respectively, and segmented grouting parameters are adjusted based on the soil pressure data and a soil pressure threshold value as well as the pore water pressure data and a stabilization time threshold value; a pressure diffusion quantitative model is constructed on the basis of synchronously collected grouting pressure data, soil pressure data and pore water pressure data, and grouting pressure distribution is obtained on the basis of the constructed pressure diffusion quantitative model. And the grouting efficiency can be improved.

The application discloses a TBM support grade decision-making method and system based on images and deep learning, and belongs to the technical field of TBM intelligent monitoring. Taking the type and structure parameters of a surrounding rock support structure as indexes, a surrounding rock support grade decision-making scheme is generated based on a screened historical picture data set; a rock slag discharge picture segmentation model is established, a trained rock slag discharge picture segmentation model is used to perform segmentation processing on newly collected TBM rock slag discharge pictures to output a rock slag segmentation prediction graph; a multi-dimensional rock slag parameter calculation system is constructed to determine a large block area in the rock slag segmentation prediction graph; the degree of surrounding rock crushing is judged based on the rock slag particle parameters of the large block area; and a matched support decision-making scheme is matched in combination with the surrounding rock support grade decision-making scheme. The application converts experience judgment into data-driven quantitative decision-making by collecting rock slag discharge pictures to construct a data set, correlating rock slag image features with geological parameters, and generating a support grade decision-making scheme in combination with clustering analysis and a quantitative model.

The application provides a malodorous smell walk monitoring method based on pattern recognition and concentration simulation coupling, characterized in that the malodorous smell walk monitoring is carried out in a pattern recognition and concentration simulation coupling mode, the source of the malodorous smell in the environment is preliminarily screened through pattern recognition, the corresponding odor concentration simulation quantitative model is automatically selected according to the pattern recognition result, and when the component of the malodorous gas in the environment changes, the corresponding model is reselected through pattern recognition judgment. The application solves the problem of low accuracy of the instrument in monitoring the odor concentration during the malodorous smell walk monitoring, improves the actual effect of the malodorous smell walk monitoring, improves the accuracy of the industry judged by the pattern recognition method, and meets the requirement of accurately judging the industry during the malodorous smell walk monitoring.

The invention discloses an anti-transfer identification method based on straddle structure and surface random texture binding. The anti-transfer identification method comprises the steps of preparing an identification base material with a straddle structure, preparing a random texture, constructing three types of quantitative models, collecting parameters, substituting the parameters into the models for calculation, completing traceability evaluation and the like. According to the method, collaborative binding of the perforation structure and the surface texture and quantitative evaluation of anti-transfer and traceability are achieved by constructing the quantitative model, the problem that no quantitative basis exists in the prior art is solved, the reliability and traceability accuracy of the anti-transfer identifier are improved, and the method is suitable for the technical field of identifier anti-counterfeiting.

The application discloses a method for power heterogenous scheduling under a large model training and reasoning framework, and relates to the technical field of power scheduling. The application solves the technical problem that the task migration, splitting and recovery mechanism is imperfect, and task interruption or resource waste is easily caused when facing insufficient computing power or device abnormity. The application collects the load rate, residual computing power, temperature and other parameters of CPU, GPU and FPGA dynamically, establishes a quantitative model in combination with the calculation amount, time length and delay demand of the task, improves the resource utilization rate, establishes a temperature prediction model in stages, sets a dynamic threshold in combination with the device type, environmental temperature and health degree, realizes temperature warning and migration in the task execution process, reduces the risk of frequency reduction caused by device overheating, and ensures that the task interruption loss is reduced by more than 50% through the task splitting, intelligent migration and suspension recovery mechanism, thereby improving the stability of the system in peak traffic or hardware failure.

The application provides a prefabricated cabin switch cabinet state prediction method combining electric heating coupling and improved D-S, belongs to the switch cabinet operation and maintenance technical field, pre-processes the data collected by a multi-parameter sensing terminal, divides the data into two independent evidence sets of contact state evidence set and insulation state evidence set; based on the collected load current, environmental temperature and cumulative opening and closing times data, a contact electric-thermal-mechanical coupling quantitative model is constructed to make up for the defects of single temperature measurement; with the aid of time series fluctuation entropy and game theory dynamic weight, the D-S fusion rule is optimized, the multi-channel monitoring signal conflict problem is effectively alleviated, and the false alarm is greatly reduced; the historical comprehensive health degree sequence is taken as the input, and the future equipment health change trend is deduced through a light-weight LSTM time series prediction network. Thus, an integrated closed loop of collection, evaluation and regulation is formed, and the fault recognition accuracy is improved.

This invention discloses a method for detecting aflatoxin B1 based on a BIC terahertz metamaterial sensor. Based on electromagnetic theory and the time-domain finite-difference method, a continuous-domain bound-state (BIC) terahertz metamaterial sensor was designed. Using this BIC terahertz metamaterial sensor, terahertz enhanced spectra of aflatoxin B1 solutions of different concentrations were acquired. Analysis revealed that the amplitude of the transmission peak of aflatoxin B1 around 1.5 THz gradually decreased with increasing aflatoxin B1 solution concentration, while the frequency of the transmission peak gradually shifted to lower frequencies with increasing aflatoxin B1 solution concentration. PLS and LS-SVM quantitative models for aflatoxin were established. Analysis showed that the LS-SVM model performed better, and the detection limit for aflatoxin B1 was found to be 2.04 × 10⁻⁶. ‑7 μg / ml. This invention enables rapid, non-destructive, and highly sensitive detection of aflatoxin B1 in grains.

This invention belongs to the field of granulation, specifically disclosing an automated water suspension granulation control method based on online particle size monitoring, comprising: Step 1, during the granulation process, using an online particle size analyzer to monitor the particle size in real time and plotting a real-time particle size growth curve over time; Step 2, based on the morphological characteristics such as the slope change points of the real-time particle size growth curve, automatically dividing the granulation process into three characteristic stages; Step 3, for each stage divided in Step 2, establishing a quantitative model between particle size growth and key process parameters; Step 4, based on the model's predictive and feedback control, executing closed-loop control in a new granulation batch. This invention also discloses an automated water suspension granulation control system. This method and system, through real-time feedback and dynamic adjustment, ensure that the granulation process of each batch follows the optimal growth trajectory, significantly reducing product quality differences between batches.

The invention provides a sintering main exhaust fan energy-saving potential diagnosis method, device and equipment and a storage medium, and the method comprises the steps: collecting the operation parameter data of a fan system in a typical normal production period, including working condition parameters, electric energy parameters and process parameters; and performing fusion analysis according to the first potential diagnosis criterion, the second potential diagnosis criterion and the third potential diagnosis criterion to obtain an energy-saving potential diagnosis result. According to the method, through multi-criterion fusion diagnosis, three quantitative models of working condition deviation degree, motor load rate and system efficiency are constructed, fuzzy judgment depending on experience is converted into accurate quantitative analysis driven by data and models, comprehensive evaluation is carried out from three dimensions of equipment operation working condition, load matching and system energy efficiency, the one-sidedness of a single index is overcome, and the system efficiency is improved. The diagnosis conclusion is more reliable, required data are all parameters easy to obtain on site, the model calculation is rapid, the method is suitable for rapid screening and priority ranking, and the evaluation cost is remarkably saved.

This invention discloses an optimization method for restoring power supply to critical loads in active distribution networks based on user outage loss assessment. The method includes the following steps: collecting user outage time, duration, and economic and social loss information; establishing a quantitative model of outage loss; using the results of the outage loss model, combined with the grid topology and restoration difficulty, comprehensively evaluating and prioritizing loads to generate a critical load sequence; constructing a power restoration optimization model with the goal of minimizing outage loss; applying the optimization strategy to actual distribution network operation; collecting restoration effect data; dynamically adjusting; and forming an iterative optimization. This invention not only achieves critical load identification and prioritization based on user outage loss, enabling priority restoration of loads with the greatest economic and social impact under limited resources, but also significantly improves the restoration efficiency and scientific nature of power supply decisions in active distribution networks after sudden power outages, balancing power supply security, economy, and social benefits.

The application discloses a general optimization method for 3D printing sugar-reduced food, comprising the following steps: S1. Defining and measuring core parameters: including porosity P of a 3D printing food matrix, odor concentration C of an odor active component accounting for a total mass proportion of the food matrix, sucrose reduction amount Suc compared with a traditional full-sugar food, and a sensory evaluation index S of a sensory evaluator on a sweetness similarity degree between the sugar-reduced food and the traditional full-sugar food; S2. Constructing a quantitative model: under the condition of setting the sucrose reduction amount Suc value in advance, a plurality of different P, C and S corresponding data are obtained through experimental design, and based on the data, a quantitative model of the sweetness similarity S about the porosity P and the odor concentration C is obtained by using a multiple linear regression method; and the problem that the prior art cannot clearly guide how to set the printing parameters and the formula to obtain the best comprehensive sensory experience under the premise of ensuring a certain sugar reduction ratio is solved.

The invention relates to the technical field of space imaging, and provides a space target space-based microwave imaging parameter optimization method and device. According to the method, through predicting a target intersection section of a detection satellite and a space target, an average relative speed, a minimum intersection distance and an orbit prediction error component of the detection satellite and the space target are extracted from the target intersection section; determining the optimized detection time length and the optimized detection frequency of the detection satellite at least based on the minimum intersection distance and the orbit forecast error component, and determining the optimized imaging time length of the detection satellite at least based on the average relative speed and the minimum intersection distance, and multi-constraint cooperative solution is carried out on the pulse repetition frequency PRF of the detection satellite at least based on the average relative speed and the minimum intersection distance to obtain the optimized PRF, microwave imaging parameter optimization design of quantitative model and actual scene constraint is realized, the efficiency is high, the application scene is wide, and the method is suitable for large-scale popularization and application. And powerful support is provided for precision improvement and engineering landing of a space-based microwave imaging technology.

The application relates to the technical field of agent testing, and particularly provides an agent dynamic behavior safety testing method and system based on a risk transmission quantitative model, which comprises the following steps: collecting agent behavior data, containing a plurality of risk operation information and risk correlation relations; constructing a behavior risk correlation graph taking risk operations as nodes and correlation relations as edges, marking risk attributes and time limit factors on the nodes, and marking risk transmission strength on the edges; constructing a risk rule library based on the risk marking results and a preset risk judgment logic; inputting the behavior data, the behavior risk correlation graph and the rule library into a trained quantitative model to calculate risk transmission probability, influence range and critical nodes; real-time monitoring of behavior trajectories, task suspension, backtracking reasoning chain and intervention record generation when a multi-dimensional risk threshold is triggered; and test report generation by fusing multi-dimensional data. The application solves the problems of missed judgment and response lag in traditional testing, and improves the pertinence and reliability of safety testing.