68 results about "Offline optimization" patented technology

The invention discloses an aircraft intelligent trajectory planning and cooperative control method based on offline optimization-intelligent learning-online guidance, and the method comprises the steps: firstly building an aircraft motion model and an aircraft-target relative motion model, and constructing a full-airspace and full-feature-point optimal trajectory data set; secondly, parameterized representation is carried out on the three-dimensional optimal trajectory, and a proportional guidance coefficient data set is constructed; then neural network training fitting is carried out on the proportional guidance coefficient data set, and a multi-constraint guidance method based on a neural network is designed; and finally, designing a three-dimensional collaborative guidance law by combining a multi-constraint guidance method and a residual time accurate estimation technology. According to the invention, the problems of insufficient real-time performance, incapability of meeting index optimality, incapability of coping with multiple constraints, difficulty in realizing multiple flight modes and the like of the existing trajectory planning and collaborative guidance technology are solved; the multi-constraint and optimality of the trajectory are realized through offline optimization, the global adaptability and small calculation amount of the trajectory are realized through intelligent learning, and a multi-cooperative flight mode is realized through online cooperative guidance.

The invention discloses a full-automatic intelligent simulation method and system for an electric power system driven by a large language model, and belongs to the field of electric power system simulation calculation. The simulation method comprises an offline optimization stage and an online application stage. An off-line optimization stage: acquiring a simulation parameter file of the power system and a modification range of each parameter, and designing a training command pool containing executable and non-executable commands; constructing a command generator and a cue word optimizer based on the large language model to form a generative adversarial framework; system cue word optimization is realized based on the framework, and an optimal cue word is obtained; the online application stage comprises the steps that after a user inputs a natural language command, a large language model recognizes the command and matches an optimal cue word, and the command is input into each parameter modifier to adjust simulation parameters; and verifying whether the parameters are within an allowable range, and if the verification is passed, calling simulation software to carry out power system simulation calculation. The simulation automation level and the intelligent interaction experience of the power system are remarkably improved, and the practical value is high.

The invention relates to the field of computer servers, and discloses a modularized computing power and storage integrated server which comprises a computing module, a storage module, a bus module, a control module and an optimization module. The calculation module and the storage module support modular hot plug through a standardized interface, and can dynamically adjust configuration according to task requirements; the bus module is used for realizing high-speed data transmission between the calculation module and the storage module and supporting dynamic allocation of bandwidth priority; the control module generates a resource extension suggestion and a bandwidth allocation strategy based on fuzzy reasoning; and the optimization module performs offline optimization on the fuzzy rule base and membership function parameters of the control module through a genetic algorithm so as to improve the overall performance of system regulation and control. According to the method, elastic expansion and dynamic management of computing and storage resources are achieved, the resource utilization rate, the bandwidth allocation efficiency and the system operation stability are improved, and the method is suitable for high-performance computing and large-scale storage tasks with complex loads and multi-scene requirements.

The invention relates to the field of vehicle control, in particular to a system based on vehicle speed planning and energy management strategy collaborative optimization and an application method. The system comprises an offline optimization subsystem and an online application subsystem, wherein the offline optimization subsystem comprises a scene pre-analysis module and a multi-dimensional dynamic planning algorithm module; the online application subsystem comprises a double-LSTM (Long Short Term Memory) model module and a vehicle-mounted control module; according to the invention, the scene pre-analysis module obtains and analyzes the scene of the urban continuous intersection; then solving a vehicle speed sequence and a fuel cell control sequence with the optimal comprehensive performance through a multi-dimensional dynamic programming algorithm module, and taking the vehicle speed sequence and the fuel cell control sequence as a training data set; the economic vehicle speed and the fuel cell output power are obtained through the double-LSTM model module; and the output power of the power battery is obtained through the vehicle-mounted control module. Therefore, the optimal output power of the fuel cell and the power cell at each moment is determined by considering external interference factors from driving scenes such as traffic lights and front vehicles and economic vehicle speeds.

The invention provides an online convex optimization method and system for unmanned aerial vehicle HTN task planning, relates to the field of artificial intelligence and intelligent planning, and adopts an online convex optimization (OCO) technology to adjust the time cost and resource consumption of task execution in real time according to the task planning requirements of an unmanned aerial vehicle in an uncertain and dynamic environment. According to the method, a task model is established through PDDL, and a task decomposition structure and target constraints are defined; decomposing the complex task into a plurality of executable actions (actions) by utilizing an HTN (Hypertext Transport Network) planning algorithm; a convex set is constructed based on loss in action execution, a cost index is defined, and action consumption is optimized online; and dynamically updating indexes according to actual feedback, adjusting subsequent action strategies, iteratively reducing consumption, and finally generating an optimized real-time plan (plan). Different from traditional static planning or off-line optimization, the method remarkably improves the action execution efficiency of the unmanned aerial vehicle, reduces resource and time waste, and can flexibly adapt to actual changes.

The invention discloses a PID parameter optimization system based on offline system identification, and the system comprises a data selection module which is used for screening modeling data in a historical database; the data preprocessing module is used for performing real data complementation and data filtering on the screened modeling data; the model system identification module is used for establishing a controlled object model by using the preprocessed data; and the PID parameter optimization module is used for performing off-line optimization of PID parameters on the controlled object model. The defects in the prior art can be overcome, any on-line closed-loop test is not needed, and the optimal PID parameter is output at a time.

The invention provides a man-machine collaborative dynamic scheduling method fused with a double-layer optimization mechanism, and relates to the field of man-machine collaborative dynamic scheduling. According to a double-layer optimization mechanism provided by the invention, an offline optimization layer and an online real-time scheduling layer are organically combined: firstly, the offline optimization layer is based on a workpiece set, stations and robot resources, and is combined with a simulated annealing algorithm for adaptive optimization to obtain a global optimal solution; and then, taking a global optimal solution output by the offline optimization layer as an initial scheduling scheme of the online real-time scheduling layer, and performing real-time adjustment by adopting a multi-agent adaptive near-end strategy optimization algorithm. According to the method, the low efficiency of online scheduling from zero exploration is avoided, and the high efficiency of decision making is ensured. Besides, the online real-time scheduling layer constructs a collaborative decision-making system composed of a task allocation agent, a resource scheduling agent and a disturbance response agent, decision-making dimension pressure faced by a single agent is remarkably reduced through specialized labor division, and a complex interaction relation in man-machine collaborative scheduling is delicately processed.

The application discloses a large language model driven full-automatic intelligent simulation method and system of a power system, and belongs to the field of power system simulation calculation. The simulation method comprises an offline optimization stage and an online application stage. In the offline optimization stage, a simulation parameter file of the power system and a modification range of each parameter are acquired, and a training command pool containing executable and non-executable commands is designed. A command generator and a prompt word optimizer are constructed based on a large language model, forming a generative adversarial framework. The optimization of system prompt words is realized based on the framework, and the optimal prompt words are obtained. In the online application stage, after a natural language command is input by a user, the large language model identifies the command and matches the optimal prompt words, and inputs the command to each parameter modifier to adjust the simulation parameters. Whether the parameters are within the allowed range is verified, and if the verification is passed, the simulation software is called to perform power system simulation calculation. The power system simulation automation level and intelligent interaction experience are significantly improved, and the method has strong practical value.

The invention belongs to the technical field of hypersonic flight vehicle control, and relates to a hypersonic flight vehicle intelligent control system based on a large language model. The purpose of the invention is to provide intelligent control of an aircraft through natural language instructions. The method comprises system design of an intelligent decision-making layer, a weight mapping layer, an offline optimization database, an interpolation calculation layer and a control execution layer, and stable tracking of a reference instruction is realized. According to the method, intelligent mapping from a natural language instruction to an accurate control action can be realized, the operation complexity is remarkably reduced, and the task adaptability and the intelligent level of the system are improved. And the method has a wide application prospect.

The invention is suitable for the energy management technology of an extended-range electric vehicle, and provides a range extender fuzzy adaptive multi-point operation mode control method, which comprises the following steps: optimizing an APU working point offline based on a dynamic programming algorithm, and defining a comprehensive performance evaluation index; dividing an APU working area based on a demand power probability density function, and calculating a power and torque interval; designing a mixed point-linear energy management strategy, and dynamically adjusting a threshold parameter in combination with a fuzzy logic controller; and performing multi-objective optimization on the parameters through a multi-objective backbone particle swarm optimization method, and realizing quantitative decision based on a Pareto solution set. The problems that in the prior art, the APU is low in high-load working condition efficiency, complex in control logic, insufficient in consideration of battery life factors and the like are solved, the energy consumption economy and emission performance of the whole vehicle are remarkably improved, and the service life of the battery is remarkably prolonged.

The application relates to a user fraud behavior detection method and device, equipment and medium. The method comprises the following steps: extracting a generated to-be-detected feature vector from a user behavior event stream in a business platform according to a plurality of detection dimensions of an implied fraud common feature; inputting a current effective convolutional neural network model for forward calculation and outputting a quantitative score representing a fraud risk; when the quantitative score exceeds a first risk threshold, sending a hard intervention instruction to a delayed delivery system to stop a transaction business process of a corresponding user; in response to a model performance index meeting a preset degradation trend condition or in response to receiving an incremental fraud sample, triggering an adaptive hyperparameter optimization process to perform offline optimization on a hyperparameter combination of the model to generate an updated model for model hot switching update. The application can realize long-term automatic evolution of a convolutional neural network model for fraud behavior identification and limit compression of illegal arbitrage cycles of black production users.

The invention provides the system, the device and the method for determining the optimal bidding strategy of which the performance is better than that of the bidding adjustment mode in a scene of repeatedly participating in the first-level price auction. The invention provides a method for predicting inventory. According to the method, an online real-time repeated auction scene can be converted into an offline optimization problem; specifically, auction is summarized through a small number of parameters, and joint distribution of the parameters is predicted based on historical data. Meanwhile, the invention further develops a method named as strategy search, and the method adopts a combined optimization means to search a bidding strategy capable of realizing the optimal performance of the advertisement activity on the premise of meeting specific constraint conditions.

The present invention discloses an industrial Internet adversarial training-based unsupervised intrusion detection system and method based on autoencoder model optimization. The present invention uses a data decompression module to collect the communication timing traffic of the industrial Internet system, obtains the input data set through data cleaning, data normalization and data sampling using spectral residual technology, binary encodes the architecture and parameters of the AE network model, designs an AE network model offline optimization platform based on binary genetic optimization technology, and uses adversarial training to evaluate individual fitness. After iterative optimization, the optimal AE network model of the industrial Internet intrusion detection system is automatically obtained. The present invention can not only realize the adversarial training and automatic optimization design of the AE model for the industrial Internet unsupervised intrusion detection system, improve the model training stability and intelligent design level of the industrial Internet intrusion detection system, but also improve the performance indicators such as the recall rate and F1 score of the industrial Internet system intrusion detection.

This invention relates to a dynamic optimization method for apparel supply chain inventory that integrates intelligent forecasting. The invention synchronously collects multi-source data, including structured fulfillment data, customer behavior maps, and strategic configuration data. It generates a channel value tensor through normalization and semantic fusion, and introduces gating activation and dynamic attention mechanisms to achieve adaptive weighting of multi-dimensional business indicators. The replenishment optimization objective function integrates time-varying weights and channel fairness constraints, and outputs differentiated replenishment strategies through convex programming. By monitoring channel replenishment revenue deviations in real time, it applies reverse fine-tuning and learning mechanisms to continuously optimize channel value representation and replenishment decisions. This invention combines a time-aware attention mechanism with a strategy-coupled optimization framework, achieving a sensitive response to real-time business conditions and a multi-objective collaborative balance in replenishment decisions, significantly outperforming the lag and rigidity problems of traditional fixed-rule or offline optimization models.

This invention belongs to the field of hypersonic aircraft control technology and relates to a hypersonic aircraft intelligent control system based on a large language model. The purpose of this invention is to provide intelligent control of an aircraft using natural language instructions. The method includes a system design consisting of an intelligent decision-making layer, a weight mapping layer, an offline optimization database, an interpolation calculation layer, and a control execution layer, achieving stable tracking of reference instructions. This invention can achieve intelligent mapping from natural language instructions to precise control actions, significantly reducing operational complexity and improving the system's mission adaptability and intelligence level. It also has broad application prospects.

This invention discloses an adaptive polishing and grinding method for aero-engine blades, comprising: constructing an aero-engine blade polishing and grinding process database; performing collaborative preprocessing and deep feature extraction on multi-source data in the database; constructing an adaptive decision model based on a dual-path strategy of offline generation and online adjustment to generate and adjust process parameters in real time; and optimizing the process parameters offline based on measured quality feedback of the ground blades, and updating the process database and the adaptive decision model. This invention achieves adaptive decision-making and optimization control of process parameters.

The invention belongs to the technical field of new energy vehicle control, and discloses a new energy vehicle intelligent energy management strategy offline optimization and online implementation method and system, and the method comprises the steps: carrying out the offline processing of vehicle real-time driving data, and building a driving condition basic database and a driving style basic information database; based on the vehicle daily driving data, the driving condition basic database and the driving style basic information base, performing rolling optimization on the control parameters related to the driving condition characteristics; and estimating the road gradient in real time, and correcting the control parameters after rolling optimization to realize intelligent energy management of the new energy automobile. According to the method, the rolling optimization of the control parameters along with the long-period operation data and the dynamic adjustment of the multi-mode driving data are realized through the process of ''offline processing + rolling optimization + online implementation'', so that the intelligent energy management of the new energy automobile is ensured.

The invention provides an SM9-based client-multi-server rapid self-adaptive key negotiation method. The SM9-based client-multi-server rapid self-adaptive key negotiation method mainly comprises the following steps: an initialization stage, a private key extraction stage, a broadcasting stage, an offline preparation stage, an online response stage and a session key generation stage. The method has the following characteristics: 1, the method is suitable for a client-multi-server scene, a client initiates a communication request, a plurality of servers are automatically adapted to communicate with the client according to states such as self load and communication delay, and a key negotiation process only needs three times of communication; 2, aiming at a client-server communication mode, computing load optimization is carried out; different from the traditional P2P key negotiation with the same calculation amount of both parties, the method migrates the calculation overhead to the server side as much as possible; and 3, the client does not need to carry out complex bilinear pairing operation, and particularly, the calculated amount in an online response stage can be almost ignored. And meanwhile, an online / offline optimization technology is adopted, so that the client can carry out pre-calculation before communication, and the online response efficiency is improved.

The present invention discloses a SAR image anti-backdoor classification method and system based on backdoor feature extraction optimization. The method includes: constructing a backdoor feature extraction subnetwork and a classification subnetwork based on network blocks; performing particle position and velocity encoding on the neural network architecture parameters, backdoor feature selection parameters, and pruning position parameters of the SAR image anti-backdoor classification model composed of the two subnetworks; taking the parameter quantity, backdoor attack success rate, and classification accuracy of normal samples of the SAR image anti-backdoor classification model as optimization targets; designing an offline optimization platform based on a three-objective particle swarm optimization method; and obtaining an online deployment model for SAR image classification that balances lightweight, high precision, and strong backdoor robustness. The present invention realizes the automatic optimization design of the SAR image anti-backdoor classification model. The constructed backdoor feature extraction subnetwork expands the processing method of backdoor features, improving the model's lightweight, classification accuracy, and backdoor robustness.

This invention discloses an input-output scaling calibration method and system. In the ECO layout stage, the method constructs an independent validation set offline and uses an optimization algorithm to solve for the input scaling factor 'a' and the output compensation factor 'b'. In online application, for long interconnects whose line length exceeds the training distribution, a two-step calibration is performed: the line length is divided by 'a', and the input is the machine learning model; the predicted value is multiplied by 'b', and the output is the model. This invention achieves a calibration mechanism that enables one-time offline optimization and efficient online invocation through the collaborative work of a sample generation module, an offline optimization module, an online calibration module, and an ECO decision module. This significantly improves the prediction accuracy for long interconnects outside the distribution, avoids buffer insertion omissions due to latency underestimation, thereby reducing the number of ECO iterations and improving timing convergence efficiency.

This invention discloses a comprehensive collaborative energy-saving optimization method and system for industrial pumps, encompassing the equipment, system, and operating conditions. This method abandons traditional single-point or offline optimization, constructing a real-time dynamic optimization system with a closed-loop "perception-analysis-decision-execution" mechanism. The core of this method lies in: deploying intelligent edge nodes at the equipment level to achieve microsecond-level adaptive adjustment based on real-time operating conditions; building a digital twin model of the pump group at the system level to perform global topology and scheduling optimization; and analyzing process requirements at the operating condition level to generate a dynamic "optimal operating condition curve." Using a unique "system energy efficiency entropy" model as a unified optimization target, and leveraging a cloud-based collaborative optimization engine, it coordinates decisions across the three levels in real time, achieving optimal energy efficiency across the entire chain, from single pump efficiency and pipeline losses to process matching. This invention achieves real-time, adaptive, and globally optimal energy-saving effects.

This invention discloses a dual friction parameter identification and adaptive synchronous control method for a multi-axis electro-hydraulic system, comprising: constructing a dynamic model of a hydraulic leveling system including a LuGre friction model; using a genetic algorithm to perform offline optimization identification of the friction parameters of the LuGre friction model to obtain initial estimated parameters; based on the initial estimated parameters, designing an adaptive parameter adjustment law based on control error to fine-tune the friction parameters in the LuGre friction model online in real time, forming a dual friction parameter identification framework consisting of offline identification and online adaptation; using the dual identification framework, designing an integrated adaptive feedforward integral sliding mode controller, introducing the online fine-tuned friction parameters into the sliding mode control law for feedforward compensation, and ensuring that synchronous control and online estimation of friction parameters are performed synchronously; based on the dynamic model, distributing the virtual torque in the overdrive coupling system, and designing an inner-loop voltage controller to achieve pressure control under constant back pressure.

The invention relates to a phishing website detection method based on multi-model cascading and genetic algorithm optimization, and the method comprises the following steps: S100, constructing an offline optimization module: based on a historical data set, an optimization target and a constraint condition, solving an optimal threshold combination through a genetic algorithm optimizer, the historical data set comprises prediction confidence scores of different models, real labels and prediction average time consumption of each model; and S200, an online detection module is constructed, specifically, an optimal threshold combination is loaded, real-time detection is conducted on the input website url to be detected in a multi-model cascade reasoning mode, multi-model cascade reasoning follows the quick judgment and on-demand deep principle, and balance of detection efficiency and precision is achieved through hierarchical screening. The objective of the invention is to automatically solve a set of optimal model decision threshold values by using the global search capability of the genetic algorithm, so that the online cascade system can realize the maximization of the detection efficiency on the premise of meeting strict performance constraints.

The application provides a federated learning model aggregation method based on aerial computing and energy collection, proposes a model aggregation design optimization algorithm using aerial computing technology and energy collection technology, considers the correlation of device energy in different training rounds, realizes reasonable scheduling of limited energy in the device by alternately optimizing the participating device, the transmission device transmission power and the beamforming vector of the receiving end, improves the model aggregation precision, and further improves the model test precision. The scheme provided by the application is an online optimization method, which is different from the existing offline optimization method. The design of the application in each communication round only depends on the current channel state information and the arrival energy, and is more suitable for actual application scenarios.

The present invention provides an effective wind speed soft measurement method, comprising a wind turbine SCADA data acquisition module, an offline optimization calculation module based on an improved gray wolf optimized IGWO algorithm, an effective wind speed soft measurement modeling module based on a kernel extreme learning machine (KELM) algorithm, and the like. The signal output end of the wind turbine SCADA data acquisition module sequentially passes through a SCADA data preprocessing and normalization module, an offline optimization calculation module based on an improved gray wolf optimized IGWO algorithm, an effective wind speed soft measurement modeling module based on a kernel extreme learning machine (KELM) algorithm, an effective wind speed soft measurement model parameter optimization module based on the IGWO algorithm, and an effective wind speed soft measurement model performance evaluation module, and finally outputs the effective value of the wind speed. The present invention establishes an effective wind speed soft measurement model, which does not require wind tunnel experiments or virtual simulations to obtain the data required for effective wind speed soft measurement modeling, but only requires SCADA data of the wind turbines actually operating in the wind farm. The model has the characteristics of low cost, high prediction accuracy, and ease of implementation.

The invention relates to the technical field of converter control, discloses a converter control parameter optimization method and system based on an improved particle swarm, and adopts a two-stage architecture of offline optimization and online parameter configuration. An offline optimization stage: in the offline stage, firstly setting key parameters of an improved particle swarm, and initializing the number of iterations and the particle swarm; calculating a particle fitness value through a dual-objective function reflecting dynamic response performance and stability robustness, and solving an individual / global extreme value, a related vector mean value and a distance to determine an individual extreme value for updating; updating the particle speed and position; then simulating a disturbance condition to perform transient stability analysis, calculating a converter state variable and judging whether a system constraint is out of limit or not; the particle inertia factor weight and the particle historical optimal fitness value are updated; if not, iteration is repeated, and if yes, the optimal control parameter combination is output. The optimal parameters are configured to the converter control system in the online stage, the parameter setting problem is efficiently and reliably solved, and stable operation of the system under the complex working condition is guaranteed.

Systems, apparatuses, and methods for determining an optimal bidding strategy that outperforms bid shading in situations that involve repeated bids into first-price auctions. An approach termed Inventory Forecasting is developed, that functions to convert the online real-time repeating auction scenario into a problem that can be optimized offline. Inventory forecasting summarizes auctions with a small number of parameters and then forecasts the joint distribution of these parameters based on historical data. An approach termed Strategy Search is developed that uses combinatorial optimization methods to search for a bidding strategy that produces the best campaign performance while still meeting specific constraints.

The invention discloses an autonomous regulation and control method and device for a nominal frame angle of a variable topology CMG group, and belongs to the field of spacecraft attitude control. The method comprises the following steps: according to the axial symmetry of the mechanical installation configuration of the CMG group, summarizing all configurations of the CMG group in a full configuration state and an under-configuration state into a plurality of limited equivalent configuration classes; performing off-line optimization on each equivalent configuration class to sequentially obtain a nominal frame angle migration path corresponding to each equivalent configuration class; when the working state of the CMG group operating in orbit is changed, determining an optimal switching scheme from the migration path before the change to the migration path after the change according to the equivalent configuration classes before and after the change of the CMG group; and performing zero angular momentum switching on the CMG group according to the optimal switching scheme. According to the method, automatic switching and migration of nominal frame angles among CMG groups with different configurations can be realized, the on-satellite calculation and storage burden can be simplified to the greatest extent, and the system reliability is improved.

The invention discloses a waterproof coiled material processing roller surface temperature control optimization method based on swarm intelligence, and belongs to the technical field of control optimization, and the method mainly comprises the steps: S1, constructing a waterproof coiled material processing roller surface temperature PID control system; s2, introducing an improved cloud drift optimization algorithm to construct an improved cloud drift optimization algorithm module; s3, performing iterative optimization by using an improved cloud drift optimization algorithm to obtain an optimal control parameter; and S4, the optimal control parameters obtained through off-line optimization serve as static operation parameters to be solidified into a temperature PID controller module, and steady-state precision control over the roller surface temperature is achieved. According to the method, the improved cloud drift optimization algorithm serving as a swarm intelligence algorithm is introduced, off-line global setting is carried out on the roll surface temperature control PID parameters in the waterproof roll machining process, and the parameter matching precision and the system evolution stability of a control system facing a large-lag and strong-nonlinearity heat conduction model are effectively improved.