33 results about "Space optimization" patented technology

The invention relates to the technical field of space planning, in particular to an intelligent optimization method for territorial space planning fusing multi-objective evolution. The method comprises the following steps: carrying out multi-modal territorial space characterization analysis on territorial space heterogeneous data to generate multi-modal territorial grid characterization space data; performing territorial multi-modal feature analysis processing on the multi-modal territorial grid representation spatial data to generate territorial planning unit multi-modal feature data; carrying out evolution trend analysis on the territorial planning unit multi-modal characteristic data to generate territorial planning unit evolution trend data; performing territorial space planning multi-target optimization strategy analysis based on the territorial planning unit evolution trend data to generate territorial space planning multi-target optimization strategy data; and carrying out multi-objective evolution territorial space planning intelligent optimization processing through the territorial space planning multi-objective optimization strategy data to generate territorial space planning intelligent optimization data. According to the invention, intelligent land and land space optimization planning is realized.

This invention discloses a method and system for spatial optimization of vegetation carbon sinks based on interpretable machine learning. The method first acquires multi-source environmental data, including administrative divisions, environmental factors, and target variables for the target area, forming a spatial tabular dataset. Based on this dataset, an XGBoost regression model is trained to predict vegetation carbon sink capacity and evaluate model performance. A SHAP interpreter is then used to quantify the contribution of each environmental factor to carbon sink, generating feature importance and interpretability maps. The Lowes algorithm is used to fit the relationship curve between environmental factors and carbon sink response, extracting the optimal threshold interval. Further, based on the optimization threshold, carbon sink gain is predicted, optimization zones are delineated, and a spatial optimization scheme is formed. Finally, a vegetation carbon sink optimization recommendation report is output. This method enables the scientific quantification of carbon sink spatial optimization, improving the interpretability and decision support value of carbon sink enhancement strategies.

The method for constructing urban ventilation corridors based on local climate zone classification and remote sensing inversion belongs to the technical field of urban climate regulation and space optimization. The present application constructs a multi-scale urban thermal environment analysis framework, from macro, meso to micro level, to identify the coupling mechanism of urban form and thermal effect, and to realize the fine modeling and regulation path identification of thermal environment. At the same time, combined with machine learning and explainable analysis method, the scientificity and transparency of thermal environment driving factor identification are enhanced. Through remote sensing inversion, form factor construction and three-dimensional simulation simulation, a complete technical path of "cold source identification-resistance modeling-path optimization-microclimate response evaluation" is put forward, which effectively improves the ventilation and cooling potential identification ability of blue-green infrastructure, and is suitable for ventilation corridor planning and climate adaptive updating design of various urban spaces, and has wide application prospect.

This invention discloses a method for testing jailbreak attacks on large language models based on latent space optimization. First, standard rejection vectors pointing to rejection concepts at multiple network layers of the target large language model are obtained. Then, the set of key layers performing security defenses in the target large language model is located, and adaptive weights are assigned to each layer in the set of key layers. A cognitive hijacking prefix targeting malicious query commands is generated and compressed. Using the positive guiding part of the cognitive hijacking prefix as the optimization target, an initial suffix is ​​generated through a first-stage adversarial suffix optimization. Finally, using the complete cognitive hijacking prefix as the optimization target, combined with the standard rejection vectors and the adaptive weights, a second-stage adversarial suffix optimization is performed on the initial suffix to generate the final adversarial suffix. This method transforms the attack target from shallow output probability manipulation to deep cognitive state alignment, effectively eliminating the "pseudo-jailbreak" phenomenon.

The invention discloses a robust classification model training method based on adversarial data enhancement. The method comprises the following steps: S1, actively collecting samples; s2, generating an adversarial sample model; s3, performing quality filtering and hardness scoring on the adversarial sample; s4, performing sample layering and adaptive training; s5, cross-task meta-learning enhancement is carried out; according to the robustness classification model training method based on antagonism data enhancement, through StyleGAN2 four-level potential space optimization projection and in combination with perception maintenance loss and potential code regularization, a generated confrontation sample not only meets the concealment but also has high fatality, and the robustness classification model training method based on antagonism data enhancement has the advantages that the robustness classification model training method based on antagonism data enhancement has high robustness and high robustness. Samples are divided into an easy layer, a medium layer and a difficult layer based on hardness scores, and high-hardness samples are dynamically introduced through a function increasing progressively along with training rounds, so that the model is prevented from wasting computing power on the easy samples, and training resources are ensured to focus on key samples for improving robustness.

This invention discloses a multi-UAV cooperative search trajectory planning method based on 3D action decoupling, aiming to solve the core problems of low efficiency in 3D action coupling learning, insufficient target uncertainty modeling, imbalance between exploration and utilization, and poor multi-UAV cooperation in multi-UAV cooperative search under 3D unknown environments. This invention models the multi-UAV cooperative search process as a partially observable Markov decision process, designs a 3D action decoupling Actor network to achieve decoupled control of horizontal plane motion and vertical altitude motion, constructs a non-sparse multi-dimensional reward function to provide dense feedback for policy learning, and introduces a dynamic noise attenuation mechanism based on target confidence to achieve spatial optimization allocation of exploration resources and complete iterative optimization of the cooperative search strategy. This method can significantly improve the target detection rate, reduce flight energy consumption and conflict rate, and can efficiently adapt to the needs of multi-UAV cooperative search trajectory planning in complex 3D unknown environments.

The invention relates to an ecological protection priority area identification method and system based on multi-source data and space optimization, and the method comprises the steps: collecting and preprocessing multi-source space data in a target area, and generating a theme space data set; quantifying the ecosystem service in the target area, and obtaining a standardized magnitude spatial distribution map; combining the species distribution point location in the target area with the theme space data set, and simulating to generate a suitability space distribution diagram; fusing the standardized magnitude spatial distribution map and the suitability spatial distribution map to form a regional comprehensive protection value curved surface; and carrying out iterative calculation on the regional comprehensive protection value curved surface by adopting a space optimization algorithm to obtain an ecological protection space layout scheme, and delimiting different protection level priority ecological protection regions of the target region. According to the method, the accuracy, scientificity and operability of an ecological protection priority area identification delimiting process are greatly improved, and an efficient technical approach is provided for cooperative protection of biodiversity and ecological system service.

The invention belongs to the crossing field of oil and gas field development and artificial intelligence optimization methods, and particularly discloses an old oil field well pattern reconstruction optimization method based on GSO-Niche multi-objective collaboration, which comprises the following steps: according to dynamic and static parameters of an oil reservoir and an oil-water well, determining well pattern density and well pattern type, initializing well position parameters, calculating evaluation indexes, and normalizing the evaluation indexes; on the basis of the evaluation indexes, a dominant set well pattern deployment scheme is obtained, the dominant set is continuously updated, well position parameters obtained after a follow-up process are updated through a firefly optimization algorithm, and finally the optimal well pattern deployment scheme is determined by means of the niche technology according to the dominant set. According to the method, the firefly group optimization algorithm and the niche technology are fused, five-dimensional index collaborative optimization is combined, the high-dimensional parameter space optimization problem and the blindness of multi-target solution set screening are solved, the optimization efficiency is greatly improved, the order of magnitude is improved, and the method is particularly suitable for encryption adjustment and secondary development of land high-water-content old oil fields.

The invention discloses an electricity-carbon cooperative scheduling method and system based on a target random path signature algorithm, and belongs to the field of power systems, the method comprises the steps of obtaining prediction data and a scheduling optimization problem of a power system, the scheduling optimization problem being an optimization problem of calculating a multi-stage scheduling model under a constraint condition; and inputting the prediction data into a multi-stage scheduling model, calculating the multi-stage scheduling model through a target random path signature algorithm to obtain a scheduling scheme, and scheduling the power system of the next day according to the scheduling scheme, according to the target random path signature algorithm, a function space optimization problem corresponding to the scheduling optimization problem is projected to a signature space through signature transformation, and dimension reduction processing is performed on the signature space through a vortex embedding technology, so that the function space optimization problem is converted into a nonlinear programming problem for calculation, and therefore, through implementation of the method and the system, the scheduling optimization problem can be effectively solved. And efficient and accurate power dispatching can be realized.

The application relates to an industry space adaptation optimization method and system based on a large model, which comprises the following steps: determining a target area, and obtaining a first data set and a second data set; constructing a digital twin base, and performing multidimensional analysis through a large language model; constructing an enterprise capability evaluation model and a migration risk early warning model based on the multidimensional analysis result, and obtaining enterprise capability information and migration risk grade information; obtaining interest point data and traffic network data, and determining geographic circle layer information; inputting the multidimensional analysis result, the enterprise capability information, the migration risk grade information and the geographic circle layer information into the large language model to generate an industry space optimization strategy; in summary, the application integrates multiple sources of data to construct a digital twin base, and uses a large language model to perform multidimensional analysis to generate an optimization strategy, solves the fragmentation problem in the traditional method, provides a decision basis, and has the effects of improving the scientific nature and operability of decision-making.

This application relates to the field of computer technology and discloses a hot-time-space task scheduling method and system for HBM in-memory computing architecture. The method includes: acquiring the thermal state information and spatiotemporal state information of a high-bandwidth in-memory computing system in real time, and performing hot spot prediction and performance prediction to obtain predicted state information; based on the thermal state information, the spatiotemporal state information, and the predicted state information, and according to preset optimization objectives and constraints, performing joint hot-time-space optimization decision-making to generate scheduling instructions including a space mapping strategy and a spatiotemporal execution strategy; and performing dynamic scheduling operations on computing tasks in the high-bandwidth in-memory computing system according to the generated scheduling instructions. This application effectively solves the performance degradation, fluctuation, and reliability risks caused by the coupling of thermal management, timing scheduling, and space resource allocation in HBM in-memory computing architecture, and improves the overall energy efficiency and stability of the system.

This invention relates to the field of space optimization technology and discloses a method for optimizing container loading space based on robot collaboration. This method constructs a multi-dimensional physical model of the package by acquiring point cloud data and weight distribution. A loading scheme including target pose, pressure tolerance threshold, and loading sequence is generated based on a reinforcement learning algorithm. The robotic arm generates motor current-position composite control commands based on a pre-trained grasping dynamics model to achieve precise grasping. During loading, the package pose is monitored in real time using a Kalman filter algorithm. When the deviation exceeds a threshold, a local spatial KD tree is constructed and the collision probability gradient field is calculated to generate a compensation scheme that satisfies stability constraints. Finally, the quadtree index is updated and fed back to the reinforcement learning algorithm for online optimization. This method effectively solves the efficiency decline problem caused by inaccurate grasping force control and accumulated execution deviations during robot loading, achieving efficient and precise automated loading.

The application discloses a resource space optimization configuration method and system for land planning, and relates to the technical field of land planning and resource space configuration; the application collects and pre-processes space data of a planning area, constructs complete topological adjacency relations, establishes a neighborhood correlation constraint matrix and a constraint processing priority sequence, filters candidate land use types through priority, completes initial configuration based on comprehensive configuration scores, detects and resolves neighborhood conflicts, adjusts land use structure deviation, and outputs a configuration scheme after constraint compliance verification, effectively solving the problem that neighborhood correlation constraints are difficult to be effectively coupled and processed in the optimization solving process, significantly improving the scientificity and rationality of land resource space configuration, and improving the constraint compliance rate of the configuration scheme and resource utilization efficiency.

The invention provides a pseudo-tag-guided single cell feature space optimization method, which comprises the following steps: acquiring a data set, and preprocessing the data set; constructing a feature extraction network model, and mapping the preprocessed data set to a unified embedding space to obtain an initialized embedding representation; inputting the samples into a classifier, generating a pseudo tag, taking the maximum value as a confidence score, comparing the confidence score with a preset confidence threshold, and dividing the confidence score into a high-confidence sample set and a low-confidence sample set; constructing a feature space structure optimization model, and designing a structured loss function; s5, minimizing the structured loss function, updating parameters of the feature extraction network model, and repeating the steps S3-S5 until the model converges; the target modal cells are predicted by using the convergent feature extraction network model and the classifier, so that category confusion caused by excessive alignment is effectively prevented through double structure constraints, and the accuracy of label migration, the F1 score and the performance of open set recognition are remarkably improved.

The invention belongs to the technical field of artificial intelligence, and discloses an unsupervised dimension reduction method based on flexible projection and adaptive graph optimization, and the method specifically comprises the following steps: 1), constructing a first-order similar graph in an original space, and obtaining a high-order graph structure through power propagation; 2) performing weighted fusion on the multi-order graph by learning a non-negative weight of which the sum is 1, and generating a multi-order fusion graph for guiding optimization of subspace composition; and 3) introducing a flexible projection mechanism to break through strict linear mapping limitation and capture potential information of the data. According to the method, a collaborative optimization objective function is constructed, and an iterative optimization strategy is adopted to update a projection matrix, a subspace optimization graph and a multi-order graph weight. According to the method, the problems of insufficient structure exploration, too strong linear constraint and the like when complex high-dimensional data is processed by a traditional method are effectively solved, the robustness and separability of dimension reduction characterization are remarkably improved, and the method has remarkable advantages in classification on a standard data set.

The invention belongs to the technical field of underground cavern axis generation, and provides an underground cavern axis generation determination method based on a multi-target space optimization model, and the method mainly comprises the steps: firstly, obtaining a main structural plane trend, a rock stratum trend and a maximum horizontal principal stress direction of an engineering region, and obtaining the engineering geological condition characteristics of the engineering region; secondly, establishing a multi-target space optimization model taking structural plane stability, stress state optimization and hydraulic performance guarantee as space optimization targets by taking the axis direction as an optimization variable; then, weights are distributed to each space optimization target based on engineering geological condition features; and finally, constructing a comprehensive scoring function, solving the axis direction when the comprehensive score is the highest by using the comprehensive scoring function, and taking the axis direction as the optimal axis direction. According to the method, the optimal axis direction suitable for the corresponding geological engineering condition characteristics can be quickly found under the complex geological conditions.

The application relates to the technical field of fault detection of industrial systems, in particular to an industrial system fault detection method and system based on distributed equivalent space optimization. The method comprises the following steps: a Hankel matrix is constructed to identify an equivalent matrix with zero projection under normal working conditions, and a residual signal is generated; based on local data of each sensor node, an average consistency algorithm is used to calculate a covariance matrix in an offline stage; parameter optimization design is performed on the equivalent matrix; the residual signal is calculated when the system is running online, and a globally consistent detection statistic is obtained through the average consistency algorithm; a detection threshold is determined in combination with normal operation data, and fault detection of a complex industrial process is realized. The application combines an equivalent space optimization method, a distributed sensor network cooperative computing mechanism and an average consistency algorithm, reduces communication and computing burden, improves system robustness and scalability, and enhances the sensitivity and reliability of fault detection.

The invention relates to the technical field of ship intelligent space optimization and operational research decision crossing, and particularly discloses an intelligent space optimization method for a ship to carry a complete set of goods, which comprises the following steps of: establishing a multi-target multi-constraint mathematical model taking cabin capacity utilization rate maximization and goods complete arrangement set number maximization as target functions; designing a ship intelligent arrangement decision tree framework based on an MCTS algorithm according to a cargo complete set carrying demand and the established mathematical model; based on a ship intelligent arrangement decision-making tree framework, carrying out iterative calculation on the complete-set cargo carrying mathematical model, and constructing a ship intelligent arrangement decision-making tree; and searching the constructed intelligent ship arrangement decision tree by adopting a UCB formula to obtain a globally optimal intelligent whole-set cargo ship arrangement scheme. By the adoption of the method, the breakthrough from local optimization to global optimization is achieved, the utilization efficiency of ship cabin capacity resources is remarkably improved, and the method is economical and practical.

The invention discloses a normalized stream missing data restoration method based on causal guidance, and the method comprises the following steps: data preprocessing and initial restoration: carrying out the standardization and coding of original missing data, and obtaining initial complete data for structure learning through a pre-restoration function; causal structure learning: utilizing a micro-optimizable causal discovery algorithm NOTEARS to automatically identify a directed dependency structure between variables from the pre-restored data, and providing causal constraints for subsequent generation modeling so as to ensure that the data restoration process accords with a real data generation mechanism; and performing normalized flow repair modeling and submerged space optimization based on a causal structure: embedding causal dependence in a flow model and introducing causal consistency constraint in a submerged space to realize high-precision data repair which has reversibility and conforms to a causal law.

According to the credential and credential resource scheduling management system provided by the invention, a resource DNA map is constructed through a dynamic sensing module, unified representation of irrelevant features of heterogeneous computing node hardware is realized, and the problem of resource portrait fragmentation caused by instruction set difference of a traditional system is solved; the entropy decision module adopts manifold learning to construct a multi-layer decision hypersurface, converts a scheduling decision into multi-dimensional strategy space optimization, and effectively improves the decision precision in a complex load scene; the topology evolution module dynamically executes resource isolation and reconstruction based on a real-time chaos degree index to form a sensing-decision-execution closed-loop control system, and cascade performance degradation caused by cross-node communication is remarkably reduced. Through the synergistic effect of the three modules, dynamic balance between the calculation density and the memory access cost is achieved while the advantages of the NUMA architecture are kept, stable resource supply guarantee is provided for a low-delay scene, and the problem of scheduling jitter caused by state characterization and decision logic splitting in a traditional method is solved on the whole.

The application relates to the technical field of space optimization management, and discloses a land space precise management method based on key driving factor thresholds, which adopts technical coupling and comparison of the leading ESs supply-demand types identified by SOM and the key driving factor thresholds extracted by constraint lines. The method solves the technical problem that the data collection, model analysis and policy making links are disconnected in the existing land space evaluation and monitoring, resulting in the difficulty in converting the evaluation results into specific instructions, realizes the full-chain technical closed loop from data input to evaluation suggestion output, ensures the systematicness and integrity of the evaluation and monitoring method, adopts the XGBoost model to process the complex interaction between variables, and innovatively adopts the constraint line model to analyze the constraint boundary of data distribution, so that the technical problem that the traditional evaluation method relies on a linear model and is difficult to capture nonlinear relationships and the evaluation suggestions are mostly qualitative descriptions is solved.

This invention discloses a method for optimizing agricultural spatial layout to synergistically enhance ecological and production benefits. The method includes: data collection and processing, assessment of ecosystem service and economic value, construction of a multi-objective quantitative optimization model, construction of spatial optimization rules, spatial pattern optimization based on cellular automata (CA), scenario simulation, and scheme evaluation. This invention combines quantitative and spatial optimization, overcoming the blind spots of traditional experience-based layouts and providing a systematic and scientific decision support method for industrial spatial layout. This invention fully considers the topographic heterogeneity and industrial compatibility of ecologically vulnerable areas such as karst regions, and the optimization rules are rationally designed with strong local adaptability. The spatial simulation results based on the CA model are intuitive and visual, facilitating planners' understanding and application. The optimized layout scheme can be directly used to guide actual production.

The invention discloses an agricultural management measure space continuous optimization configuration method for a combinatorial explosion problem, and relates to the technical field of agricultural resource management, and the method comprises the steps: obtaining the feature data of a plurality of to-be-configured space units in a target region, and dividing the plurality of to-be-configured space units into a plurality of homogeneous clusters; for each homogeneous cluster, a continuous proportion decision variable is defined, the value range of the continuous proportion decision variable is [0, 1], and a target function model comprising at least two optimization targets which conflict with each other is constructed; performing optimization solution on the objective function model by adopting a multi-objective evolutionary algorithm to obtain a Pareto optimal solution set; and selecting an optimal solution from the Pareto optimal solution set, and generating a spatial decision scheme based on the continuous proportion decision variable value of each homogeneous cluster in the optimal solution. According to the method, the problem of large-scale space optimization is feasible in calculation, the requirements of decision makers of different levels are met, and the practicability and operability of the result are enhanced.

This invention discloses a low-cost, high-strength, and high-wear-resistant ductile iron design method based on few-sample machine learning, belonging to the fields of metal material preparation technology and computer science. This invention constructs an enhanced dataset by collecting the composition and mechanical properties of annealed ductile iron and extracting and screening alloy features to obtain key alloy features. Then, a multi-objective optimization function is constructed, and feature space optimization is performed using an improved NSGA-II algorithm to obtain the Pareto solution set, which is then mapped to features and experimentally verified. This invention overcomes the risk of overfitting with few samples through a hybrid enhancement strategy; it constructs an adaptive multi-scale recurrent convolutional network and a dual attention mechanism to accurately screen key alloy features; it uses an improved NSGA-II algorithm for multi-objective optimization, introducing a dynamic sharing mechanism and local search to make the Pareto front distribution more uniform; and it establishes a random forest mapping model and a closed-loop automatic retraining system to achieve accurate feature-component mapping and continuous optimization.

The invention provides a multi-factor collaborative ecological system carbon-water utilization efficiency space optimization regulation and control regionalization method and system, and relates to the crossing field of ecology and computer technology.The method comprises the steps that firstly, carbon-water utilization efficiency collaborative element data of an ecological system is collected, and a multi-factor collaborative conduction network is constructed by analyzing an element interaction path and injecting a conduction rule; then, a spatial conduction link matrix is generated based on the multi-factor collaborative conduction network in combination with regional geographic spatial attributes, then dynamic adaptation processing is carried out on an initial spatial unit of a region where the ecological system is located on the basis of the spatial conduction link matrix, a basic zoning unit with element conduction adaptation capacity is generated, and on the basis, the spatial conduction link matrix is dynamically matched with the basic zoning unit; and after the element cyclic feedback flow is constructed according to the operation state of the multi-factor collaborative conduction network, the dynamically adjusted basic zoning unit and the element cyclic feedback flow data are integrated to form a space optimization regulation and control zoning scheme, multi-factor collaboration and ecological system dynamic change are fully considered, and the carbon-water utilization efficiency of the ecological system is effectively improved.

The invention provides a resident sharing oriented community public space intelligent optimization design method and system, and the method comprises the steps: obtaining real track sets and environment data of different groups, filling a data sparsity blind area through employing a conditional generative adversarial network, and generating space-time thermodynamic diagrams of different group behaviors. Then, based on a space-time diagram convolutional network, fusing a grid topology, a thermodynamic diagram and a social association matrix, and analyzing a demand priority map; and coupling cross-modal features by adopting a dual-channel deep learning architecture, and generating physical design parameters conforming to constraints through a bidirectional attention mechanism. According to the method, the confrontation generation network is used for enhancing data, the physical fidelity of the space-time thermodynamic diagram is improved, and different group behaviors and space-time data are fused to overcome the priority deviation of public space design caused by subjective weight distribution. And finally, fusing a spatial topology gradient and a group behavior evolution rule by cross-modal feature fusion, thereby improving demand response precision and dynamic adaptability of public space optimization.

The application provides a water network space optimization method and device, wherein, based on each space optimization measure, an initial water network space is optimized to obtain several optimized water network spaces; each optimized water network space is subjected to implementation feasibility evaluation to obtain implementation feasibility evaluation results of each optimized water network space; a space optimization scheme of the initial water network space is determined based on the implementation feasibility evaluation results of each optimized water network space; and based on the space optimization scheme, the initial water network space is optimized to obtain a target water network space. The above method can improve the implementation feasibility and optimization effect of water network space optimization.

Method for determining parameters for operating a technical device by combining two algorithms, starting with input data using at least two optimization algorithms, wherein in a first step the input data are processed using discrete parameter space optimization and in a second step the input data are processed using continuous parameter space optimization.

The application discloses a kind of based on partial order prediction multi-fidelity microarchitecture design space optimization method, comprising the following steps: from microarchitecture parameter, the parameter that has influence to target optimization PPA index is selected, and multi-fidelity data set of SOC design space initialization is constructed;Nonlinear multi-fidelity Gaussian proxy model is constructed, and maximum expectation hyper volume is used on last high-fidelity proxy model improvement;Logical regression function is constructed and trained;After obtaining new sampling design point, first using EDA simulation tool carries out architecture level simulation, to obtain low-fidelity index data, whether need to enter next fidelity based on partial order prediction is judged;Iterative optimization, obtain Pareto optimal solution.The application will partial order prediction, fidelity modeling and multi-objective bayesian optimization are organically integrated, form an efficient exploration system suitable for complex microarchitecture DSE task, with strong practicality, universality and scalability.

The present application belongs to the field of intelligent additive manufacturing quality control and fluid mechanics, and provides a flow field uncertainty correlation mapping and robust optimization method and system for high-consistency aerosol jet printing: based on the original geometric structure of the aerosol jet printing head, an aerosol printing flow field simulation model is established, and then combined with the uncertainty parameters, the flow field uncertainty space mapping data set is obtained; a flow field reasoning agent model is constructed by combining a Gaussian process regression model; a multi-objective optimization function based on velocity-pressure field stability is constructed, and the flow field reasoning agent model is used for iterative solution, and finally the optimal process parameters of aerosol jet printing are obtained through the Pareto optimal boundary identification. The present application is based on the organic combination of computational fluid dynamics simulation and uncertainty modeling based on polynomial chaos expansion, and multi-objective global robust optimization algorithm based on Pareto equilibrium, and overcomes the multi-source airflow interference from the bottom aerodynamic mechanism and global space optimization angle.