20 results about "Design strategy" patented technology

The invention discloses an electric power field training set dynamic construction method and system based on BERT and reinforcement learning. The method comprises the following steps: S1, constructing a terminology library and a knowledge graph in the power field; s2, performing power field adaptation on the BERT model by using a terminology library and a knowledge graph to obtain a field adaptation model; s3, designing a strategy network based on reinforcement learning so as to dynamically select unlabeled samples and generate pseudo labels; and S4, based on the domain adaptation model and the strategy network, dynamically constructing and optimizing a training set through an iteration process. According to the method, the training set quality and the model performance are jointly improved, manual labeling is avoided, sample selection can be adaptively adjusted, the labeling cost is low, and the model accuracy is high. According to the method, the pseudo labels are automatically generated through the reinforcement learning strategy network, manual labeling requirements are reduced, labeling efficiency and model performance are improved, the quality of a training set can be optimized, and adaptability is enhanced.

The application discloses a centrifugal pump impeller design method, which comprises the following steps: initial parameter determination; parameterized geometric modeling, wherein the profile lines of the front cover plate and the rear cover plate of the impeller are constructed based on a cubic spline curve, and a blade surface model is generated by controlling point coordinates, a curvature radius and an inlet and outlet angle parameter; manufacturing process coupling design, wherein three-dimensional modeling software is used for parameterized modeling of the impeller, and a cutter kinematics equation of a five-axis machine tool is synchronously introduced; CFD numerical simulation analysis; construction of an artificial neural network; and compound shape reverse optimization. The application solves a three-dimensional RANS equation in a rotating column coordinate system to perform CFD verification, and combines the design strategy of the artificial neural network and the compound shape optimization algorithm to quickly and optimally select an optimal impeller model. The design of the impeller is based on the cubic spline curve to construct an impeller geometric model, and the generative machining technology of the five-axis numerical control machine tool is combined to realize deep cooperation of hydraulic design, three-dimensional modeling and manufacturing process.

The invention discloses a self-test case defect complementing method and device based on a large model, and electronic equipment, and relates to the field of financial science and technology, the method comprises the following steps: comparing a self-test case with a necessary input field set and a necessary design strategy set in a case rule base through a feature matching algorithm, obtaining a missing case field and a missing design strategy scene of the self-test case; the missing case field and the missing design strategy scene are assembled to a missing supplement request, the missing supplement request is sent to the financial case large model, and a missing scene case is generated through the financial case large model; and receiving a missing scene case returned by the financial case large model, and assembling the missing scene case into the self-test case by using the template framework to obtain a complemented self-test case and case element prompt information. According to the method and the device, the technical problem that the use satisfaction of a user is influenced due to obvious defects of the self-test case caused by lack of automatic identification and test steps of the self-test case developed in the related technology is solved.

The invention relates to the technical field of bridge design, and discloses a bridge type finding design method based on reinforcement learning and topological optimization, which comprises the following steps: defining the size, boundary condition, load position and support position of a bridge design domain, and initializing the design domain into a solid material; generating a bridge design strategy under the randomly initialized boundary layout; performing evolutionary optimization on the intelligent agent through a genetic algorithm, and updating the neural network weight of the intelligent agent so as to converge to an optimal boundary layout; in each iteration, a variable density method is adopted to optimize material distribution, so that materials are concentrated towards a high-bearing area; under the obtained optimal boundary layout, material density distribution is iteratively updated through a variable density method, and the structural rigidity and the material utilization rate are optimized; inputting a random variable into the optimized intelligent agent, and generating a diversified bridge form scheme meeting the structural performance requirement; the method can adapt to diversified design requirements and complex engineering conditions so as to realize more efficient and more reasonable bridge design.

The invention discloses a bill data generation method and device, equipment, a medium and a product. The method comprises the following steps: in response to a bill data generation request, obtaining original bill data generated by interaction between target systems according to business operation interface information of each target system, and determining a market category corresponding to each original bill in the original bill data; based on a domain-driven design strategy, performing sub-domain division on the original bills of different market categories, and constructing a domain model corresponding to the original bill data according to a sub-domain division result; and performing mapping implementation on the domain model through a preset test component, and performing configuration on an image user interface, so as to generate bill test data based on the test component and the image user interface which are realized through mapping. According to the technical scheme, the internal feature association of the bill data can be automatically extracted and analyzed, and effective and accurate test bill data is constructed in combination with the test component, so that the subsequent software test efficiency is improved.

The invention relates to the technical field of porous lattice material forming, and discloses a buffering and energy-absorbing three-dimensional truss lattice structure design method which comprises the following steps that based on at least one predefined design strategy, a set of initial vertexes are generated in a single trigram limit, and the design strategy is used for optimizing the energy-absorbing performance of lattice unit cells; according to a symmetry rule, mapping the initial vertexes in a single trigram to other trigrams to obtain all vertexes of the dot matrix unit cells; defining a connection relationship between all the vertexes to form a truss structure, the connection relationship being configured such that the truss structure is in a bending dominant deformation mode under a compressive load; and generating a geometric model of the three-dimensional lattice unit cell based on all the vertexes and the connection relationship. Through strategy driving and rule constraint, automatic mapping from performance requirements to geometric configurations is achieved, and the defect that a traditional design method is mostly focused on elastic response or linear behaviors and cannot be suitable for reverse design with energy absorption as the target is overcome.

The invention provides a low-rolling-resistance and impact-resistant tire design method, and relates to the technical field of tire design, and the method comprises the following steps: acquiring sidewall impact stress data, rolling resistance performance data and historical impact working condition data of a target optimization tire; carrying out area division on the sidewall of the target optimization tire, determining an impact bump risk area, a tire shoulder area and a padding steel belt lower sidewall area, and establishing a sidewall reference thickness matrix of each area; based on the sidewall impact stress data, the rolling resistance performance data and the impact working condition characteristics, determining a thickness adjustment coefficient of each area, and generating a sidewall optimized thickness matrix; and generating a sidewall thickening scheme based on the sidewall optimized thickness matrix. The thickness of the tire side at the bump position is increased, the thickness of the tire shoulder position and the thickness of the lower tire side area cushioned with the steel belt are kept unchanged, the differential design strategy avoids the problem of rolling resistance increase caused by traditional thickening of the whole tire side, and the low rolling resistance performance of the tire is effectively maintained.

The application discloses an intelligent design method based on a thinking chain and brand DNA constraint, and comprises the following steps: design strategy proposal; constructing a "demand-logic-visual" thinking chain; brand DNA constraint; AIGC design scheme output; gestalt adaptive layout; scheme output and landing production. The application utilizes a thinking chain model combined with a multi-modal artificial intelligence algorithm to realize accurate conversion from fuzzy design requirements to structured design strategies, significantly shortens the creative deduction cycle, and improves the accuracy of design logic and the demand fit degree. The system of the application has high production flexibility and can quickly generate visual schemes suitable for different media scales, marketing nodes and theme contents based on parameterized constraints, effectively meeting the design requirements in diversified scenarios.

The invention discloses a dynamic design strategy generation method and device, equipment and a readable storage medium, and belongs to the field of architectural design, and the method comprises the steps: carrying out the matching of a design portrait from a portrait database according to the space preference data of a user; acquiring a corresponding design element correction coefficient from the design element priority correction table according to the basic feature data of the user; and dynamically correcting the priorities of the design elements in the design portrait by utilizing the design element correction coefficient, generating a corrected design element priority list, and generating and outputting a design strategy according to the corrected design element priority list. According to the method, the space preference of the user is accurately captured and analyzed in a data driving mode, and a more detailed and personalized design portrait is formed; and comprehensive analysis is carried out through further data subdivision demands and user feedback, and a data dynamic design feedback and adjustment mechanism is enhanced, so that design-oriented dynamic adjustment is realized, and a design scheme can flexibly respond to changes of user demands.

The application discloses a serious game design process optimization method, which comprises the following steps: constructing a serious game design project task node management table, and constructing a total design progress index model and a total guiding vector index model; based on the actual progress data of the serious game design project when the project is executed according to the task node management table, the total design progress index model and the total guiding vector index model are used to evaluate the state of the serious game design project when the project is executed to the current progress, so as to obtain evaluation indexes of the total design progress, the progress state, the total design progress prediction accuracy, the concretization guiding vector and the perspective guiding vector of the serious game design project; and the task node management table is optimized based on the evaluation indexes. The application can quantitatively feed back the serious game design process, thereby monitoring the progress and the guidance of the project in real time, enabling the project manager to accurately master the actual state of the project, and timely adjusting the design strategy according to the actual progress, thereby improving the efficiency and the effect of the project management, and the transparency and the predictability of the design process.

The invention discloses a centrifugal pump impeller design method. The method comprises the steps that initial parameters are determined; parametric geometric modeling is carried out, molded lines of a front cover plate and a rear cover plate of the impeller are constructed based on a cubic spline curve, and a blade curved surface model is generated through control point coordinates, curvature radiuses and inlet and outlet angle parameters; manufacturing process coupling design is carried out, three-dimensional modeling software is used for impeller parameterization modeling, and a tool kinematics equation of a five-axis machine tool is synchronously introduced; performing CFD numerical simulation analysis; constructing an artificial neural network; and performing compound reverse optimization. According to the method, CFD verification is carried out by solving a three-dimensional RANS equation under a rotating cylindrical coordinate system, an optimal impeller model is quickly optimized in cooperation with a design strategy combining an artificial neural network and a complex optimization algorithm, an impeller geometric model is constructed on the basis of a cubic spline curve in the design of an impeller, and a generation type machining technology of a five-axis numerical control machine tool is combined. And deep collaboration of hydraulic design, three-dimensional modeling and a manufacturing process is realized.

This invention discloses a product design strategy generation method based on environmental feature extraction, belonging to the field of intelligent product design technology. The steps are as follows: constructing a multimodal dataset containing ambient sound, scene images, and odor data; extracting root mean square energy, zero-crossing rate, spectral centroid, and spectral flux features from the ambient sound; extracting explicit visual features from the scene images and implicit style features through an improved Swin-Transformer model; extracting intensity, pleasantness, and familiarity features from the odor data; performing semantic mapping through Mamdani fuzzy inference to generate a design guidance parameter vector, including sensory stimulation, morphological simplicity, emotional affinity, technological feel, interaction rhythm, stability, sense of security, and environmental integration; and generating a product design strategy based on this vector. This invention automatically transforms environmental perception into a product design strategy, improving the objectivity and efficiency of design decisions.

The application discloses a dynamic design strategy generation method, device and equipment and a readable storage medium, and belongs to the field of architectural design. The method comprises the following steps: matching a design image from an image database according to user space preference data; obtaining corresponding design element correction coefficients from a design element priority correction table according to user basic characteristic data; dynamically correcting the priority of design elements in the design image by using the design element correction coefficients, generating a corrected design element priority list, and generating a design strategy and outputting the same according to the corrected design element priority list. The application accurately captures and analyzes the space preference of a user in a data-driven manner, forms a more detailed and personalized design image, and further analyzes the data subdivision requirement and user feedback, enhances the data dynamic design feedback and adjustment mechanism, and then realizes dynamic adjustment of the design orientation, so that the design scheme can flexibly respond to the change of the user requirement.

The present application proposes a shell-based metamaterial design strategy based on level set function weighted combination. The framework is inspired by three-periodic minimal surfaces, and adopts a level set function weighted combination strategy to construct a shell-based metamaterial database. The effective elastic properties of the metamaterial are evaluated by the homogenization theory of voxel modeling. A backpropagation neural network model is developed, which can realize the reverse mapping between the geometric parameters and material properties of the shell-based unit. In addition, the present application verifies the reliability of the homogenization theory by using quasi-static compression tests, and evaluates the effective elastic properties and compression response of the proposed shell-based structure. The research results show that the shell-based metamaterial not only shows isotropic elasticity and high stiffness beyond the upper limit of Hashin-Shtrikman theory, but also has significant improvement in yield strength and energy absorption capacity under compression. The design finally forms a series of lightweight, high-stiffness structures, and is 3D printed by PolyJet technology.

The invention relates to the technical field of vehicles, in particular to a vehicle roof design method and device, equipment, a medium and a product, and the method comprises the steps that on the basis of at least one structural performance working condition, optimization design is conducted on structural parameters of a roof sub-assembly, and a target roof structural design strategy meeting preset performance conditions is obtained; based on the target top cover structure design strategy, optimization design is conducted on at least one boundary condition parameter of the production and manufacturing process with the aim of minimizing the relative deformation amount between the top cover outer plate and the reinforcing component, and a target top cover process design strategy meeting preset process conditions is obtained; and based on the target roof structure design strategy and the target roof process design strategy, obtaining a roof design scheme of the current vehicle. Therefore, by synchronously introducing joint optimization of the performance working condition and the manufacturing boundary condition in the structural design stage, the problem that the top cover cross beam and the outer plate are degummed and fail due to the fact that the adhesive layer is not cured is fundamentally solved, and the later design change cost is reduced.

The invention relates to the technical field of reinforcement learning, and discloses a sports store data-based costume design dynamic adjustment strategy generation method and system, and the method comprises the steps: carrying out the format standardization processing of the real-time sales data, historical sales data and commodity information data of a market sports store, and obtaining a multi-source integrated data set of the market sports store; performing label mapping on the multi-source integrated data set based on a consumer portrait label system to obtain a label mapping data set of the multi-source integrated data set; carrying out attribute aggregation on the label mapping data set to obtain a commodity dynamic portrait of the market competitive store; performing popularity ranking on the real-time sales data in the multi-source integrated data set to obtain a sales trend index set and a commodity popularity sequence of the market competitive store; according to the method, the effectiveness and reliability of data processing can be improved, a high-quality data foundation is laid for subsequent design strategy analysis, and the long-term market competitiveness and self-adaptive ability of a design strategy are improved.

An automatic design optimization method for an assembly line workstation decoding algorithm based on LLM belongs to the technical field of intelligent manufacturing system optimization and computer intelligence cross, and automatically generates a high-performance decoding algorithm for solving a parallel mixed flow assembly line balance problem by constructing a self-evolution algorithm design framework. The method is integrated into an upper-layer genetic algorithm of a double-layer optimization algorithm based on reinforcement learning assistance, a traditional normal form depending on manual design is thoroughly changed, and the solving efficiency and quality of the parallel assembly line balance problem are remarkably improved; meanwhile, the self-evolution mechanism enables the decoding algorithm to have the continuous improvement capability, the design strategy can be automatically adjusted according to different problem characteristics, the adaptability and robustness of the algorithm are improved, and therefore the efficiency and quality of production line planning are remarkably improved.

The invention discloses a question generation method, a medium, computer equipment and a program product. The method comprises the following steps: acquiring knowledge points for generating a question; generating a question design strategy based on the obtained knowledge points and preset expected difficulty information, and inputting the question design strategy into a question generation model, so that the question generation model generates questions based on the question design strategy; the expected difficulty information is used for indicating the expected difficulty of the generated question; based on the difference between the difficulty of the currently generated question and the expected difficulty indicated by the expected difficulty information, a difficulty adjustment strategy is determined, and the difficulty adjustment strategy is used for indicating the difficulty relationship between the next generated question and the currently generated question; and performing first adjustment on the question design strategy based on the difficulty adjustment strategy, so that the question generation model regenerates a question based on the question design strategy subjected to the first adjustment.

The application provides an optimization design method and system for foundation pit support based on BIM, and belongs to the field of foundation pit support design. The method comprises the following steps: constructing a BIM foundation pit support model, obtaining a plurality of BIM free structure units and surface attribute parameters of each BIM free structure unit; collecting pollution monitoring data samples, analyzing support resistance indexes, and obtaining support resistance index calculation results; obtaining an identification optimization design strategy for the maximum support resistance index of each BIM free structure unit, generating a plurality of identification optimization design strategies, and optimizing the BIM foundation pit support model. The application solves the technical problem that the existing BIM-based foundation pit support design lacks targeted optimization adjustment after model construction, resulting in insufficient adaptability of the support structure, and achieves the technical effect of optimizing and adjusting the BIM foundation pit support model through support resistance index analysis, and improving the adaptability and reliability of the foundation pit support structure.