91 results about "Deterministic system" patented technology

The invention discloses a feeder voltage adjustable potential multi-target evaluation method and device considering a flexible resource operation domain. The method comprises the following steps: constructing a flexible resource adjustment operation constraint model; establishing a deterministic model of node flexible resource operation domain evaluation considering the adjustment cost; in consideration of output of renewable energy sources in nodes and source load uncertainty of loads, a multi-scene stochastic optimization method is adopted to convert a deterministic model into a node flexible resource operation domain stochastic evaluation model, and upper and lower adjustable ranges of node power, baseline power and segmented unit adjustment cost after flexible resource aggregation are obtained; and finally, comprehensively considering feeder network constraint, adjustment cost after flexible resource aggregation and feeder source load prediction, and constructing a multi-target evaluation stochastic optimization model of the adjustable potential of the feeder voltage with minimum node voltage deviation and network power loss to obtain a quantized adjustable range of the feeder voltage.

The invention relates to the technical field of industrial process control, and discloses a digital twinning-based fermentation process temperature adaptive control system, which comprises a data acquisition and state representation module, a probability generation type digital twinning module, an adaptive risk management module, an opportunity constraint model prediction control module and a model iteration trigger module, according to the method, probability distribution of future states is predicted by constructing probability generation type digital twinning, and the uncertainty of the model is quantified in real time; the system can dynamically adjust the risk tolerance, drives the chance constraint model prediction controller, seeks an optimal control solution on the premise of ensuring the process safety probability, triggers the model adaptive update by means of an online evaluation mechanism, and realizes the optimal control solution on the basis of effectively managing the process uncertainty. The fermentation temperature is subjected to high-performance self-adaptive control considering safety and economy, and the problem that a traditional deterministic model is insufficient in control robustness is solved.

The invention discloses an automatic wharf import container position distribution method with embedded deep reinforcement learning. The method comprises the steps that firstly, for a container lifting area, an objective function is established based on the total processing time of a current batch of containers and the total turnover time of all containers in the future; then, robust peer-to-peer conversion is carried out on the established mathematical model, an uncertainty model is converted into a deterministic model, and constraints of position layout selection are correspondingly added; thirdly, carrying out statistics and simplification on the possible layout of the bay positions so as to greatly reduce the number of constraint conditions in the model; and solving a container turnover upper bound value in the objective function after robust peer-to-peer conversion through a deep reinforcement learning method, and finally obtaining a container position distribution scheme by utilizing a commercial solver. According to the method, on the basis of remarkably improving the solving efficiency of the container position distribution strategy, it is ensured that the wharf has higher stability and reliability when facing the import container operation with high uncertainty, and powerful decision support is provided for wharf operation management.

A computer-implemented method of controlling an output from a dialogue system, the method comprising:receiving, by way of an input, first input data relating to speech or text provided by a user;selecting a first state from a plurality of states of a deterministic model, at least some states of the plurality of states being associated with a corresponding portion of a language model prompt including an instruction to call a corresponding function;responsive to the first state being associated with a corresponding portion, generating a first language model prompt comprising at least part of the corresponding portion associated with the selected first state;providing the first language model prompt as input to a language model to generate a first language model output;determining whether to execute a function based on the first language model output;responsive to determining to execute a first function based on the first language model output, executing the determined first function to generate a first function output;selecting a second state from the plurality of states based on the first function output;responsive to the second state being associated with a corresponding portion of a language model prompt, generating a second language model prompt comprising at least part of the corresponding portion associated with the selected second state;providing the second language model prompt as input to the language model to generate a second language model output;determining whether to provide an output to the user based on the second language model output; andresponsive to determining to provide an output to the user based on the second language model output, outputting, by way of an output, speech or text to the user.

The presently disclosed subject matter provides a state-space model composed of one or more neural networks. The state-space model is configured to randomly model the technical system by modeling uncertainty in both hidden states of the technical system and weights of the one or more neural networks. Thus, the state-space model may be capable of capturing both accidental uncertainty (inherent unpredictability in observations) and cognitive uncertainty (uncertainty in parameters or weights of the model). During training and subsequent use for model predictive control, moment matching across neural network layers is used, which can ensure that the predictions of the model are consistent and close to real system behavior. The above measures can improve the control and security of many technical systems.

The invention discloses an optimal scheduling method for a comprehensive energy system, relates to the field of energy management, and is used for solving the problems that a traditional scheduling method is difficult to realize overall optimization, the traditional scheduling method is limited, the traditional scheduling method generally simplifies an energy system into a deterministic model, the uncertainty of energy supply and demand is ignored, and the energy consumption is low. And the scheduling scheme is poor in adaptability in actual operation. The method comprises the following steps of S101, data comprehensive acquisition and fine preprocessing, S102, precise system model construction, S103, uncertainty scene scientific generation, S104, optimal scheduling model elaborate construction, S105, model efficient solving and verification, and S106, scheduling scheme robust implementation and dynamic adjustment. The invention provides an optimal scheduling method for a comprehensive energy system, which can fully consider the energy supply and demand uncertainty and various energy coupling characteristic factors, realizes the economic and efficient operation of the comprehensive energy system, and reduces the energy cost and environmental pollution.

The invention discloses a self-adaptive micro-grid operation optimization method, system and device and a medium, and belongs to the technical field of micro-grid operation optimization, and the method comprises the steps: obtaining the operation state data of a micro-grid; performing depth feature coding and space-time attention processing on the micro-grid operation state data through a state sensing module to generate unified state representation; carrying out online strategy learning and optimization through a strategy evolution module based on unified state representation, and outputting a strategy optimization result; a new energy uncertainty model is adopted to generate multi-scene prediction data, the prediction data and a strategy optimization result are combined, and a prediction control problem is solved through a rolling optimization module; and according to a real-time risk assessment result, the decision weights of the strategy evolution module and the rolling optimization module are adjusted, and a micro-grid operation scheduling instruction is output. According to the method, a new state deep perception-dynamic strategy learning-robustness optimization normal form is constructed, and intelligentization and adaptive learning of the micro-grid are realized.

The invention relates to the technical field of scene modeling, and solves the problems of multi-source heterogeneous data fusion distortion, uncertainty expression deficiency and insufficient static model self-adaption. According to the core scheme, physical sensing data and subjective description data are processed in a unified mode through a differential fuzzy processing mechanism (continuous quantities are mapped through trapezoidal membership functions, and discrete quantities are mapped through semantic rules); a dynamic weight superposition strategy is adopted, data collaboration is achieved based on a spatial incidence matrix, and dynamic optimization of an additional aging attenuation factor is achieved; synchronously outputting a deterministic model framework and a fuzzy boundary thermodynamic diagram during scene reconstruction; and establishing a real-time feedback closed loop optimization membership function parameter and weight strategy. The innovation points are as follows: semantic distortion is eliminated by a confidence conversion mechanism, an adaptive superposition operator switches an operation mode according to a confidence threshold, and three-dimensional semitransparent rendering visualization uncertainty distribution is realized. The method has the advantages that the multi-source fusion semantic distortion rate is reduced by 62%, the disaster early warning accuracy rate is improved by 55%, and the continuous operation precision attenuation rate of the model is controlled below 3%.

The invention belongs to the technical field of power system evaluation, and particularly discloses an in-province power balance margin rapid evaluation method considering source load uncertainty and a market mechanism. Comprising the following steps: modeling marginal distribution of a source-load output prediction error by using nonparametric kernel density estimation; performing correlation modeling on random samples generated based on the probability model and generating the random samples; simultaneously obtaining two groups of corresponding random samples which are mutually independent and obey standard normal distribution; constructing an in-province power balance margin evaluation model, taking the two groups of sample sets as the input of the in-province power balance margin evaluation deterministic model, and establishing a BASPC-based power balance margin proxy calculation model; taking the random sample as the input of the BASPC agent model to calculate the final intra-province power balance margin; and quantifying the blocking risk of the line under the influence of the source load uncertainty through the load rate of the system line. According to the method, the power balance margin can be quickly and accurately evaluated, and the line blocking risk in the province can be found in time.

The invention discloses an approaching rainfall forecast post-processing method based on a frequency transformation diffusion model, and belongs to the technical field of meteorology. The method comprises the following steps: firstly, acquiring a radar echo signal predicted and output by a deterministic model; secondly, performing post-processing on the radar echo signal output by the deterministic model based on a rainfall forecast post-processing model of frequency domain transformation, and outputting a corrected radar echo signal; and finally, the corrected radar echo signal is utilized to carry out approaching rainfall forecasting. According to the post-processing model, high and low frequency components of radar signals are decoupled through a frequency decomposition module, high and low frequency component optimization is carried out through a high and low frequency diffusion model, and complete radar echo signal output is reconstructed through a frequency reconstruction module. According to the method, a high-fidelity radar echo prediction signal with a physical structure and local details can be generated, the problem of fuzzification of existing close rainfall deterministic prediction is solved, and the precision and accuracy of close rainfall prediction are remarkably improved.

The invention discloses a robust optimization method for slab yard crown block scheduling of an iron and steel enterprise, relates to a method for optimizing a crown block operation process, and provides a robust optimization model for solving a slab yard crown block scheduling problem, so that the crown block operation process is effectively optimized, and the in-out warehouse efficiency of slabs is improved. According to the actual starting time of slab hoisting operation, a mathematical model is established by taking the minimum slab warehousing time as a target function. Establishing a deterministic crown block scheduling model under an ideal condition; due to the fact that many uncontrollable factors can occur in slab operation, crown block scheduling can be affected, an original operation plan is not feasible, and a robust optimization model of crown block scheduling under uncertain factors is established on the basis of a deterministic model. And solving the model by using LINGO, and verifying the correctness and feasibility of the model. And testing the influence of uncertain parameters on the crown block scheduling problem by using a sensitivity analysis method, and verifying the robustness of the model.

Described herein are techniques for stacking machine learning models to better capture deterministic relations in a dataset. In some instances, a first machine learning model may not be capable of capturing all of the deterministic relations in a dataset due to the limitations of the model. Supplemental models may be trained so that the corrections generated by the supplemental models, when combined with the first machine learning model, perform better at capturing the deterministic models in the dataset. Techniques are described for training supplemental models to capture deterministic relations associated with ordinal data and nominal data and continuous data.

The embodiment of the invention provides a rail transit intelligent ray tracing channel modeling method based on local phase error calibration, and aims to solve the problem of inaccurate channel modeling caused by environment modeling errors and dynamic phase distortion in a traditional ray tracing technology in a high-speed dense rail transit scene. The method comprises the following steps: firstly, constructing an intelligent railway communication system three-dimensional geometric model comprising a transmitting end and a receiving end, generating multipath parameters by using ray tracing simulation, and establishing a deterministic frequency response model based on the multipath parameters; secondly, an independent and identically distributed von Mises phase error # imgabs0 # is superposed in the deterministic model, and a random channel model is formed; and finally, optimizing a material electromagnetic parameter vector by using an actually measured data set, minimizing the difference between ray tracing prediction and actually measured data, and improving the generalization ability of the model at an unknown position.

The invention discloses a power transmission network extension planning method and device considering optimal consumption of new energy, and the method comprises the following steps: determining a net load disturbance acceptable domain, and calculating the reliability cost; establishing a power transmission network expansion planning model, and determining an objective function and constraint conditions of the model; processing the power transmission network expansion planning model, and solving the model; wherein the objective function of the model is line investment cost, standby cost, reliability cost and unit operation cost minimization, the quadratic term of the line cost part can be piecewise linearized, and a dual method is adopted to convert a constraint condition uncertainty model into a deterministic model for solving; according to the method, the whole model is converted into a deterministic model, so that the calculation efficiency of the model and the robustness of a planning result are greatly improved.

The embodiment of the invention provides a thin interlayer and phase control physical property cause sand body configuration model construction method and system, and belongs to the technical field of reservoir modeling. The method comprises the following steps: identifying fluvial facies reservoir distribution characteristics based on exploration data, and constructing a corresponding cause sand body reservoir configuration model; obtaining a corresponding small-layer sand body deterministic model based on the cause sand body reservoir configuration model; on the basis of the deterministic model of the small-layer sand body, performing representation of an internal configuration interface of the sand body based on a lateral lamination model and an adaptive triangular mesh generation algorithm; and generating a distribution model of each lithofacies unit based on representation of a configuration interface in the sand body, and constructing a physical property model for fine modeling and heterogeneity description of high and low permeability sections in the configuration based on a phase control stochastic simulation method. According to the scheme, the limitations of insufficient precision, unreasonable transition and low calculation efficiency in traditional modeling are effectively overcome, and high-precision geologic model support is provided for reservoir management and oil reservoir development.

The application discloses a kind of improved logistic chaotic sequence hybrid measurement matrix construction method.The element of measurement matrix in the application is by the sequence with independent and identically distributed statistical properties generated by a modified logistic chaotic system, after two-level processing, it meets Bernoulli statistical distribution, satisfies the RIP criterion required to be obeyed by measurement matrix of compressive sensing technology, finally, the measurement matrix obtained by sequentially filling according to the toprizi matrix construction rule, with the randomness characteristics of chaotic deterministic system, and the measurement matrix has the advantages of low data amount required to be generated, easy to realize hardware resources etc., and the reconstruction accuracy of signal is better than random Gaussian measurement matrix.

The application provides a quality-uncertain waste power battery recycling and disassembling integrated optimization method, which firstly takes minimization of total operation cost as an objective function, considers coupling constraints, recyclable packaging related constraints, vehicle related constraints, demand satisfaction constraints, inventory related constraints, disassembling line related constraints and decision variable value range constraints, and constructs a stochastic programming model of waste power battery recycling-disassembling integrated optimization; a sample average approximation method is used to convert the stochastic model into an equivalent deterministic model, and a linearization treatment is performed on nonlinear terms to enhance the solving performance; finally, a two-stage heuristic algorithm is used to solve the proposed mathematical model, and the required optimal scenario number can be selected according to the actual problem. The application realizes the integrated optimization of the recycling and disassembling two-stage problems of waste power batteries in the quality-uncertain environment, and provides a scientific solution for the complex optimization problems in the waste power battery recycling field.

The invention discloses a grid-connected photovoltaic system power prediction method adopting a hybrid meta learning method. The method comprises the following steps: extracting meteorological data and historical power generation data; carrying out interval interpolation on the extracted data by using a nearest neighbor interpolation method, and preprocessing the data; calculating coplane irradiance considering dust; generating combined characteristics of cloud amount and coplane irradiance considering dust; generating statistical characteristics according to the power generation modes in different time periods; inputting the combined features and the statistical features into a plurality of models for training, and performing dimensionality reduction on the features by using principal component analysis; stacking a plurality of model prediction results and a deterministic model prediction result together, and combining with dimension-reduced data to form an extended feature vector; and learning the extended feature vector by using a long short-term memory network to generate a final power prediction result. According to the method, high-precision prediction of the photovoltaic power generation power in all time periods within 24 hours in the future can be realized, and the method has a relatively high practical application value.

The invention relates to the technical field of new energy, and provides a cascade water-wind-light long-term complementary scheduling plan optimization method, electronic equipment and a computer program product. Comprising the following steps: performing multiple uncertainty quantization on new energy output and runoff to obtain a quantization result; a double-layer recursive optimization model including a planning stage and an adjusting stage is constructed based on the quantification result, the double-layer recursive optimization model takes the maximization of the cascade water-wind-light long-term power generation capacity as an optimization target, the planning stage is defined to generate an initial scheduling plan, and the adjusting stage dynamically corrects the planned output deviation based on the quantification result, so that the output of the cascade water-wind-light long-term power generation capacity is optimized. State information coupling between the two stages is realized through a transmission variable; and iteratively solving the double-layer recursive optimization model, and outputting an optimization scheduling plan meeting constraint conditions. The problem of long-term scheduling under multiple uncertainties of water, wind and light is solved, the limitation that the plan cannot be dynamically corrected by a traditional deterministic model is broken through, and the robustness and new energy consumption capability of the scheduling plan are remarkably improved.

This invention discloses a method for optimizing the dynamic transmission error distribution characteristics of a harmonic reducer, comprising: establishing a static transmission error probability model to obtain the probability distribution of the overall static transmission error; constructing a dynamic transmission error mathematical model considering static transmission error and dynamic parameters; constructing a high-precision surrogate model of dynamic transmission error including the probability distribution of static transmission error and the range of dynamic parameters; obtaining the probability distribution of dynamic transmission error; and dynamically adjusting the range of dynamic parameters based on a particle swarm optimization strategy to find the parameter range that optimizes the dynamic transmission error distribution. By introducing static transmission error into the dynamic model of the harmonic reducer, considering the influence of the processing and assembly of the harmonic reducer on the transmission error, a probability-range hybrid uncertainty model is used to mathematically describe the probability distribution characteristics of static transmission error and the range of dynamic parameters, and a Chebyshev polynomial method is used to construct an approximate model of dynamic transmission error.

The invention discloses a scheduling method and device for a new energy power system containing small hydropower stations, equipment and a medium, and the method comprises the steps: constructing a hydropower space-time matrix model of the system based on a topological structure and water flow motion parameters between nodes of a watershed; substituting the load prediction value and the operation parameter prediction value of each unit into a deterministic model to obtain a reference value of a system optimization target; and according to the uncertainty model, the reference value and the certainty model of the system, constructing an IGDT scheduling model of the system and solving the IGDT scheduling model to obtain an output plan and an optimization target of each unit in the system. By constructing a hydropower space-time matrix model of the system, a topological structure and a water flow motion state between each small hydropower station and a conventional hydropower station are considered when a decision is made, and the accuracy and the reliability of a scheduling result are improved; the influence of water load uncertainty on a scheduling result is quantified through an information gap decision theory, a decision maker provides scheduling strategies under different risks and different effects, and the reliability and economical efficiency of a final scheduling scheme are guaranteed.

User speech or text data is received. A state from a plurality of states of a deterministic model is selected. At least some states are associated with a corresponding portion of a language model (LM)

The application discloses a kind of space-time graph prediction method and system based on diffusion model and space-time feature decoupling.It is determined that the present space-time prediction is in conflict with the problem of optimization of certainty and uncertainty, and a two-part network containing double-flow time-frequency encoder and conditional diffusion model is designed;The double-flow time-frequency encoder extracts the time-domain dynamics and frequency-domain periodic features of the space-time data in parallel through bidirectional Mamba-2 and adaptive spectral block, pre-trains the deterministic model, and obtains the fusion features using the pre-trained network;Its fusion features are injected into the diffusion model as a condition, and a token statistical self-attention mechanism is introduced to capture spatial dependencies and model probability distributions.The application belongs to the field of artificial intelligence and intelligent transportation technology, and can be used for high-precision prediction of urban traffic flow and air quality.

A real-time artificial intelligence and / or machine learning (AI / ML) system for an aircraft can include an AI / ML module configured to receive one or more inputs and to calculate an AI / ML control output. The AI / ML module can include a non-deterministic model for processing the inputs and outputting the AI / ML control output. The system can include a deterministic module configured to receive one or more inputs and the AI / ML control output from the AI / ML module. The deterministic module can include a deterministic model for processing the inputs and / or AI / ML control output to calculate a deterministic condition. The deterministic module can be configured to check the AI / ML control output against the deterministic condition to determine whether to output the AI / ML control output.

The invention discloses a high-resolution rainfall ensemble forecasting method, system and equipment and a medium. The method comprises the following steps: acquiring a historical atmosphere reanalysis data set and a multi-source fusion rainfall data set; the constructed deterministic neural network learning model is trained to output target duration accumulated average rainfall; training a probability precipitation model constructed by a variational auto-encoder model and a denoising diffusion probability model; acquiring meteorological data at a to-be-predicted moment, and inputting the meteorological data into the deterministic rainfall model to obtain an accumulated average rainfall prediction result within a target duration; sampling from Gaussian distribution to obtain noise, and performing denoising processing in a hidden space by using the trained denoising diffusion model; the denoising result is restored by the trained variational auto-encoder model and added with the output of the deterministic model to obtain the accumulated rainfall of the target time length; and constructing ensemble forecast based on the multiple groups of target duration accumulated rainfall forecast. According to the method, high-resolution rainfall distribution can be generated, the extreme rainfall prediction accuracy is improved, and the uncertainty of rainfall is evaluated.

A calculation method for an annual average loss of an external substructure reinforcement system based on interfacial shear stresses is provided. The calculation method includes: step 1: establishing a deterministic model for an external substructure reinforcement system, and generating seismic ground motion samples and structural samples by random sampling; establishing a probabilistic hazard model based on the seismic ground motion samples; generating a probabilistic vulnerability model based on interfacial shear stresses; establishing a total probabilistic loss expectation model by different seismic ground motion intensity levels; and using the total probabilistic loss expectation model to obtain an annual average probabilistic seismic loss of the external substructure reinforcement system based on the interfacial shear stresses. The calculation method can realize classified calculations of the seismic loss of the external substructure reinforcement system. Compared with the traditional seismic loss calculation method, the calculation method is more efficient in calculation.

The invention relates to a high-energy-load park affine optimization scheduling method considering wind and light correlation, and belongs to the field of energy optimization scheduling. The method comprises the following steps: acquiring prediction intervals of output and load demands of wind and light units in a park, constructing an initial parallelogram correlation model, and converting the initial parallelogram correlation model into a multi-constraint affine form model in combination with an improved adaptive polygon algorithm; a particle swarm optimization algorithm is used for dynamically adjusting constraint conditions, redundant intervals are eliminated, and irregular polygon affine constraints fitting the reality are generated; the method comprises the following steps: converting a deterministic model of electrolytic aluminum, energy storage and thermal power generating units into an affine form, and constructing an optimal scheduling model containing power balance, equipment operation and cross-time constraint; and solving by adopting a mixed integer linear programming solver in order to minimize the operation cost. According to the method, by dynamically correcting the wind-light correlation area, the conservative property of the uncertainty interval is reduced, the new energy consumption rate is improved, the starting and stopping cost of the thermal power generating unit is reduced, and collaborative optimization of economical efficiency and reliability is achieved.

A system and method for stochastic optimization of manufacturing and supply chain processes are disclosed. The system comprises a processor and memory configured to receive a plurality of tasks associated with a process, along with data items for each task. The data is processed through various modeling modules, including yield, cycle time, throughput, on-time delivery, cost, and financial results modeling modules. The system analyzes the data to determine metrics and provides recommendations for optimization. The method includes steps for determining statistical distributions, identifying bottlenecks, calculating cycle times, and refining estimates to improve process efficiency. The system outputs metrics and recommendations via a user interface, enabling proactive management and optimization of manufacturing and supply chain metrics. This approach addresses the limitations of traditional heuristic and deterministic models by providing a comprehensive, integrated solution for process optimization.

The invention discloses a method for calculating the line loss of a high-proportion new energy access power distribution network, and belongs to the technical field of calculation of the line loss of the high-proportion new energy access power distribution network. The calculation method for the line loss of the high-proportion new energy access power distribution network comprises the following steps: S1, carrying out preliminary data acquisition and preprocessing by the calculation method for the line loss of the high-proportion new energy access power distribution network, and providing basic line parameters for all other modules by a module; according to the method for calculating the line loss of the high-proportion new energy access power distribution network, the change characteristics of the line loss of the power electronic power distribution network accessed by the high-proportion distributed power supply and the optimization mechanism are calculated; comprising the aspects of network loss change characteristics and a line loss optimization mechanism under a deterministic model, an uncertainty processing scheme, a line loss and electric energy quality regulation and control method and the like, and a line loss regulation and control scheme based on an evolutionary algorithm is established aiming at the electric energy quality problem caused by high-proportion distributed power supply and nonlinear load grid connection.

The invention relates to a wireless signal simulation method and device, equipment, a storage medium and a program product, and relates to the technical field of wireless communication. The method comprises the following steps: determining geographic data and base station distribution data of a target area; and based on the geographic data, the base station distribution data and a pre-trained signal prediction model, determining the wireless signal intensity of at least one height at the target location in the target area. One-step end-to-end prediction can be realized, and the calculation overhead is reduced and fixed. Moreover, compared with a method based on a deterministic model, high-precision environment information does not need to be collected, and the cost of environment information collection is reduced.