67 results about "Optimal decision" patented technology

The application discloses a path planning method based on self-attention enhancement under unstructured terrain and relates to the technical field of automatic driving decision planning. The method comprises the following steps: constructing an offline global training field containing diversified slope combinations; establishing a physical mapping model of terrain slope and energy consumption to quantify path cost; constructing a normalized multi-channel terrain feature map, introducing a self-attention mechanism to capture long-distance spatial dependence to predict slope trends; building a double deep Q network and combining distance-energy coupling reward functions for training to learn an optimal decision model; and in the online stage, only local features are obtained by using vehicle-mounted sensors to output an optimal action sequence in real time. The application realizes autonomous navigation of a vehicle under no global terrain information through an offline global learning and online local perception mode, effectively overcomes decision failure caused by lack of prior map data in traditional methods, and significantly improves the working capacity of the vehicle in unknown unstructured terrain.

This invention provides a termite control method based on remote intelligent control using the Internet of Things (IoT), comprising: collecting multi-dimensional environmental data related to termite activity in real time through an IoT sensor node network deployed in the monitoring area; uploading the multi-dimensional environmental data to a remote intelligent control platform, analyzing and making decisions in real time based on a preset algorithm model, and generating optimal control instructions; and remotely controlling one or more control execution devices in the monitoring area according to the optimal control instructions to start or adjust their working modes and complete targeted termite control operations. This invention aims to solve the problems of inaccurate risk quantification, lack of scientific decision-making, low control precision, and lack of dynamic optimization mechanisms in existing termite control technologies. By using IoT multi-dimensional sensing, dynamic pressure field modeling, and multi-agent collaborative game decision-making, it achieves accurate quantification of termite activity risk, optimal decision-making for control schemes, and remote intelligent control, thereby achieving efficient, economical, and environmentally friendly termite control goals.

The application discloses a satellite data optimization method and system based on two-stage decision, and a processing system, belongs to the technical field of remote sensing satellite data processing, and solves the technical problems that the existing technology has low manual selection efficiency and cannot select optimal data. The application obtains a map selected range vector and a satellite standard scene to be selected, attaches a single scene contribution area of each satellite standard scene to be selected and a sorting value field of a pre-computed vector spliced based on priority to an attribute table of the corresponding satellite standard scene to be selected, performs twice sorting on the satellite standard scene to be selected, adopts a first-stage global coverage decision and a second-stage local optimal decision, and obtains optimal satellite data. The application is used for selecting optimal data from selected data on the basis of unchanged coverage area, minimizing delivery area, and greatly improving the selection efficiency of users and data processing personnel.

The application discloses an automatic driving behavior decision-making method based on double heuristic tree search, and belongs to the technical field of automatic driving.The method is as follows: actions are abstracted as a combination of lateral lane-changing decision and longitudinal acceleration-deceleration decision, a decision time domain and a decision tree depth are set, and an initial behavior decision tree is constructed; a heuristic search strategy is designed by fusing field knowledge priori and data-driven priori, a reference action sequence is generated to guide the search process; a decision tree pruning mechanism based on field knowledge is proposed, and the tree structure is dynamically reconstructed with the last period optimal decision as a root node; the comprehensive cost of the candidate decision sequence after pruning is evaluated, and the sequence with the minimum cost is selected as the optimal decision sequence.The application effectively solves the technical problems of high computational complexity, poor real-time performance and insufficient decision rationality of traditional methods in a long decision time domain by constructing a semantic action decision tree and fusing a double heuristic guide and a field knowledge pruning mechanism.

The application discloses a spacecraft autonomous avoidance of space debris based on a deep deterministic policy gradient algorithm, classifies a scene, constructs different training environments according to different types of orbit threats, constructs a constrained Markov decision process, makes avoidance decisions while satisfying multiple constraints of the spacecraft, in an offline training stage, makes the intelligent agent interact with different types of environments continuously, and learns and trains corresponding neural networks, in an online application stage, first, threat identification is performed, corresponding trained neural networks are extracted, avoidance actions are quickly generated, the spacecraft can learn to make corresponding optimal decisions in different scenes and environment states, and the avoidance of space debris is realized online. The technology can enable the spacecraft to face space debris threats, autonomously formulate avoidance strategies, autonomously complete avoidance actions, autonomously grasp the timing and strength of space debris avoidance, and effectively respond to the space debris threats at the minimum cost (minimize interference with task execution).

The application discloses a kind of pumped storage power station weekly regulation external economic benefit intelligent evaluation method, belong to electric power system economic evaluation technical field.The technical problem to be solved is that existing method lacks the modeling ability of weekly regulation time scale, external economic benefit has not formed unified evaluation system, weight determination relies on artificial experience, qualitative index lacks standardization quantization method, and the stability of evaluation result is insufficient.The technical solution points are to construct a multi-level evaluation index system for weekly regulation, to establish a weekly scale operation scheduling model for pumped storage-power generation-backup coordination to quantify each index, to determine the comprehensive weight by using a three-source fusion game weighting method combining AI virtual expert, AHP subjective weighting and entropy weight method objective weighting, to obtain the comprehensive evaluation result by combining fuzzy comprehensive evaluation model, and finally to output comparative analysis and optimal decision suggestion under different operation modes.

This invention belongs to the field of multi-energy dispatching technology, specifically a method and system for multi-energy coordinated adaptive dispatching of photovoltaic, biogas, and thermal storage loads. The method includes establishing a multi-energy coupled system model encompassing photovoltaic, biogas, heat pump, thermal storage, and electrothermal loads; characterizing the nonlinear influence of fermentation temperature on biogas production rate and dynamic heat balance; constructing a two-stage rolling optimization framework: the first stage uses the fermentation cycle as a window to determine the temperature setpoint and biogas storage boundary; the second stage uses a 24-hour window to jointly optimize grid interaction, heat pump heating, thermal storage charging and discharging, and adjustable loads; constructing a Markov decision process in the day-ahead dispatching stage, and solving it using an improved approximate dynamic programming approach; and issuing dispatching commands based on the optimal decision sequence to complete closed-loop execution. This invention achieves coordinated optimization across fermentation cycles and day-ahead dispatching, effectively addressing uncertainties in photovoltaic output and load, and improving renewable energy absorption rate and system operating economy.

The application discloses a kind of based on the combat decision-making method of progressive evolution, including collecting expert combat data information under combat game simulation environment, obtains situation information;Expert combat data information is imitated and learned, and the same level of imitated learning intelligent agent as expert data is obtained.This based on the combat decision-making method of progressive evolution first imitates and learns from expert experience to obtain the same level of imitated learning intelligent agent as expert data by basic combat rule and decision-making method, then the imitated learning intelligent agent is reinforced learning training using reinforcement learning, improve intelligent agent level, finally again using alliance learning constantly improves decision-making ability, obtains optimal decision-making intelligent agent, solves the problem that intelligent agent training, exploration time is long in prior art, reward shaping is difficult and easily falls into local optimum and leads to effect cannot be improved.

The present application relates to a power distribution network-irrigation park electricity-water interaction coordination method considering photovoltaic access, belongs to the technical field of power distribution network optimization and energy internet, builds a power distribution network-irrigation park electricity-water physical structure; builds a double-layer interactive game framework of power distribution network-irrigation park and constructs a game model; according to the framework, a master-slave game model is constructed, the upper power distribution network takes the comprehensive operation cost, transformer load balancing and irrigation node load peak-valley difference as the optimization target, an optimal decision model of power distribution network operation mode and time-of-use electricity price is established; the lower irrigation park takes the time-of-use electricity price released by the power distribution network as the target, and makes optimal decision of irrigation under photovoltaic access based on its own cost and satisfaction; the interactive game coordination model proposed by the present application is solved by using particle swarm algorithm and CPLEX toolbox, and example verification is carried out; the agricultural irrigation water mode is changed from extensive and inefficient to saving and efficient, and the safe and reliable operation of rural low-voltage power distribution network is effectively supported, achieving a win-win situation.

The application discloses a multi-target intelligent decision system and method, and relates to the technical field of multi-target optimization and artificial intelligence. The application collects factory equipment operation parameters, pollutant emission data and historical case data; constructs a green performance intelligent agent; defines a comprehensive emission sequence, builds a gray breakpoint model through inverse accumulation reconstruction and background value calculation; establishes a difference equation, calls a parameter optimization sub-intelligent agent to independently optimize model parameters, and outputs a load and emission correlation prediction model; deduces single-class emission, constructs a multi-target optimization equation, calls an optimization guide sub-intelligent agent to execute a simulated annealing improved non-dominated sorting genetic algorithm, and solves a Pareto optimal solution set; generates an SBM-DDF model to calculate a performance index, and calls an optimal decision sub-intelligent agent to screen an optimal load point and regulate equipment operation. The application realizes qualitative change from static solution to dynamic evolution.

This invention discloses a method and system for interaction between intelligent agents in power grid dispatching that supports dynamic negotiation and result fusion, belonging to the field of power grid dispatching technology. The method includes establishing a contract network protocol and a verifiable dynamic negotiation framework among agents. Agents obtain dynamic negotiation process data through multiple rounds of bidding for resources and allocating tasks. Confidence assessment results for local decision schemes are obtained based on multi-level, multi-dimensional evidence. A multi-level evidence aggregation algorithm is used to dynamically weight all local decision schemes, perform preference attribution alignment, and conflict resolution reasoning, fusing them into a globally optimal decision scheme. A full-process decision tracing graph is constructed based on the globally optimal decision scheme and the dynamic negotiation process data. This invention solves the problems of existing power grid dispatching technologies lacking a dynamic negotiation and result fusion mechanism, leading to the inability to collaboratively optimize under multi-objective conflicts, the difficulty in scientifically fusing multi-source decision results to form a globally optimal solution, and the lack of traceability in the decision-making process.

The application discloses a decision fusion method and system in a multi-listening scenario in covert communication and belongs to the technical field of wireless communication. The method comprises the following steps: configuring a limited total block length for each listening node, and dividing the total block length into a detection block length and a transmission block length; each listening node performs local signal sensing, generates a local decision result by using the detection block length, and sends the local decision result to a fusion center by using the transmission block length; the fusion center allocates an optimal fusion weight for each decision result based on the local detection performance and the transmission performance of each listening node according to the criterion of minimizing the global detection error probability; the fusion center calculates a weighted sum based on the fusion weight and the corresponding decision result, compares the weighted sum with an optimal decision threshold of the fusion center, and makes a global decision on whether the target signal exists. The application realizes the optimal performance of a whole link from sensing to fusion, and improves the reliability and accuracy of signal detection.

This invention discloses an intelligent shading control method and system for building atriums based on multi-source data fusion and multi-objective optimization, relating to intelligent control of green buildings. The method includes the following steps: defining thermal control zones and shading zones; acquiring multi-source prediction data sequences within the prediction time domain; dynamically predicting the regional direct solar heat gain and regional spatial heat generation from people in each thermal control zone at a given time node based on the multi-source prediction data sequences; performing optimization to obtain the optimal decision variable matrix that minimizes the objective optimization function; and updating the target opening angle corresponding to the first time node in the optimal decision variable matrix to the shading adjustment angle for the next control cycle. This invention's method, combining the real-time mapping of the spatial location of the light spot heat zone with the pedestrian flow heat map, enables precise shading actions to target local hotspot areas with pedestrian activity, significantly reducing ineffective shading and wasted natural lighting caused by traditional large-area uniform actions such as full opening and closing.

The application discloses an intelligent agent multi-target navigation method based on deep reinforcement learning in a dynamic environment, and aims at the problems of separation of pedestrian behavior prediction and decision, insufficient modeling of behavior difference and insufficient multi-target trade-off in the prior art, and constructs a decision framework integrating pedestrian impatience prediction and Nash game. First, a simulation environment containing intelligent agents and pedestrians is established, and state and action spaces are designed; second, a multi-target reward function of safety, sociality and efficiency is constructed, and a meeting angle is introduced to depict the interaction relationship; further, a pedestrian impatience evolution model is established to realize dynamic prediction of the behavior trend of pedestrians; finally, the prediction result is embedded into a Nash equilibrium game model to guide the optimal decision; a safety penalty and a cost Critic network are introduced on the basis of a Soft Actor-Critic algorithm to realize parameter collaborative optimization.

This invention provides an intelligent prediction and control method for the jacking attitude of large-size rectangular pipe jacking. The method includes: defining the attitude control principle of the rectangular pipe jacking and determining the decision variables and control range of the actual jacking process; collecting real-time actual pipe jacking attitude parameters and multi-source heterogeneous data for prediction; based on the attitude control principle and control range of the decision variables, using a trained BO-GANDALF model to solve for the optimal decision variables; and realizing intelligent control of the pipe jacking attitude based on the optimal decision variables. The BO-GANDALF model processes the model input through a GFLU feature extraction module composed of stacked gated learning units, extracts optimized feature representations, and outputs a feature parameter contribution matrix. A feedforward neural network then outputs the predicted values ​​of the rectangular pipe jacking attitude parameters based on the optimized feature representations and the feature parameter contribution matrix.

The application discloses a tobacco field soil fertilizer machine flow regulation control method, and belongs to the technical field of tobacco field soil fertilizer control. The method comprises the following steps: obtaining tobacco field soil parameter data, tobacco field soil meteorological data and tobacco field soil remote sensing data; extracting the temporal and spatial characteristics of the tobacco field soil parameter data to obtain soil parameter temporal and spatial characteristics; performing multi-modal data fusion on the soil parameter temporal and spatial characteristics and the tobacco field soil remote sensing data to obtain multi-modal fusion data; obtaining a state vector based on the multi-modal fusion data and the tobacco field soil meteorological data; inputting the state vector into a pre-trained deep reinforcement learning strategy network to generate an optimal decision action vector; and converting the optimal decision action vector into a physical signal for driving the action of an execution mechanism to realize accurate regulation of the fertilizer machine flow of the execution mechanism. The application can comprehensively analyze the factors influencing the fertilizer application amount, realize accurate matching of the fertilizer application scheme and actual production, and is beneficial to accurate tobacco production.

The application relates to the technical field of artificial intelligence and decision optimization, in particular to a dynamic system cooperative fast-consumption e-commerce pricing prediction method based on deep reinforcement learning. The method proposes an end-to-end cooperative decision model. The model fuses system state multi-source data through a multi-modal feature extraction network, outputs initial prediction values of each decision variable by using a time series prediction network, and introduces an Actor-Critic framework-based reinforcement learning optimizer. The optimizer maximizes a hybrid objective function which combines a profit target and a supervision constraint, jointly optimizes and adjusts the variables, and finally outputs the cooperatively optimal decision values. The application realizes joint dynamic optimization of coupled variables, effectively improves the accuracy, training stability and overall profitability of the decision.

The application discloses a kind of construction methods of matrix solver based on DNA computing, it is related to molecular computing technology field, comprising the following steps: obtaining matrix to be calculated, the element of the matrix to be calculated is converted into DNA concentration difference signal and DNA circuit signal, generates double-track encoding matrix signal;Based on the double-track encoding matrix signal, construct simple operation module and complex operation module;With chemical reaction network, the output concentration difference of the simple operation module and the output concentration difference of the complex operation module are uniformly operated, and a steady-state output signal is generated;The steady-state output signal is subjected to weight calibration and calculation result convergence, and a matrix solver based on DNA computing is generated.Benefiting from this, the low computational efficiency of the existing optimization and control algorithm is solved, the optimal decision cannot be realized quickly, stably and with high precision, resulting in the path smoothness of robot in actual operation being insufficient, the energy consumption being relatively high, and the environmental adaptability being weak.

This invention belongs to the field of robot safety control technology and discloses a method for safe motion planning of a dual-redundant robotic arm. A linear mapping relationship is established between the joint spatial velocities of the redundant robotic arm and the velocity of the end effector. End effector velocity constraints, safety distance constraints, and joint amplitude constraints are established, and a multi-objective joint constraint model is established by integrating a differential kinematics model with the three types of constraints. The optimal decision variables satisfy the Cartesian-Kuhn-Tucker conditions. The original expression of the Cartesian-Kuhn-Tucker conditions is clarified, and a nonlinear complementary function is introduced to reconstruct the Cartesian-Kuhn-Tucker conditions. Based on the improved Cartesian-Kuhn-Tucker conditions, a vector error function is defined, and an adaptive zeroing neural network is designed. The vector error function is iteratively solved using the adaptive zeroing neural network to obtain the optimal decision variables of the multi-objective joint constraint model. The optimal decision variables are then used as the joint motion control commands for the robotic arm and output, completing the robotic arm motion planning.

The present application relates to the technical field of network communication, and more particularly to a video dynamic code rate intelligent adjustment method and system, which comprises the following steps: collecting encoding parameters, transmission state data and playing state data comprehensively to provide basis for intelligent adjustment and ensure that the decision is in line with the actual scene; deconstructing the encoding parameters, extracting features and fusing to generate a content-aware feature vector, so that the code rate adjustment is in line with the video characteristics and the visual quality is effectively improved; predicting the network quality change trend and evaluating the playing risk index to provide a forward-looking guide for code rate adjustment, timely adjust the code rate and other parameters, and reduce the possibility of playing interruption; generating an optimal decision vector by comprehensively considering multiple targets, balancing the visual quality, the probability of freezing and the switching smoothness, and realizing the overall optimization of video transmission and playing performance; and adjusting the parameters of each link of the video through cross-layer self-adaptation to fully exert the advantages of different levels and ensure smooth and high-quality video playing.

The application relates to the technical field of unmanned mine vehicle control, and discloses a mine open area vehicle avoidance control method. The method realizes real-time sensing of dynamic obstacles and terrain through the fusion of high-precision radar and visual sensors, significantly improves the safety and efficiency of the mine automatic driving system, uses a graph neural network to predict the future motion trajectory of the obstacles, combines the unique operation rules of the mine to analyze the behavior, can effectively reduce the risk caused by the dynamic obstacles, in addition, the introduction of the reinforcement learning model enables the avoidance strategy to be dynamically adjusted, multiple factors such as safety distance, task delay, fuel efficiency and ride comfort are comprehensively considered, and optimal decisions can be made by the vehicle in a complex environment in time, so that the intelligent level of the avoidance control is improved, the limitations of traditional fixed route planning are reduced, the overall efficiency of mine transportation is optimized, and the risk of accidents is reduced.

This invention discloses a real-time decision-making method and system for power balance in distribution networks based on topology awareness. The method includes: collecting distribution network topology data, real-time power flow data, and source-load prediction data, and preprocessing the collected data; performing chromosome encoding and initial population construction based on a topology awareness strategy; constructing a multi-objective function and dynamic constraint penalty weights, and constructing an individual fitness function; adopting a hierarchical tournament selection mechanism, designing a topology-preserving crossover operator and a hierarchical mutation operator, updating the population, and maintaining an external archive set; performing real-time decision optimization based on a rolling time-domain window and prediction compensation, determining whether the convergence criterion is met, and outputting the optimal decision instruction if the convergence criterion is met. This invention solves the problems of many invalid solutions, poor constraint adaptability, high topology destruction rate, and insufficient real-time performance in traditional methods, and realizes multi-objective collaborative optimization and rapid decision-making for power balance in distribution networks.

This invention relates to a digital monitoring method, system, and medium for biosecurity in pig houses, belonging to the field of pig house monitoring technology. The invention utilizes data statistics and key indicator calculations based on control data to generate a visual dashboard. Through spatiotemporal aggregation analysis of multi-source data on pigs, it identifies the spatial and temporal clustering of abnormal pigs and encodes multi-dimensional features to form a high-dimensional feature vector. Based on this high-dimensional feature vector, real-time risk monitoring and early warning are performed, and contact network tracing is conducted to generate tracing results. Based on the tracing results, propagation path simulation is performed to generate optimal decision-making suggestions, which are then displayed according to a preset method. This invention, through early detection and early intervention, can reduce the risk of disease spread in breeding pig farms and reduce losses such as pig deaths and production stoppages caused by epidemics. Calculations show that annual losses per farm can be reduced by 15%-30%.

This invention discloses a self-healing decision optimization method for power information systems based on reinforcement learning feedback. The method includes: acquiring operational data to construct a dynamic graph structured state representation; formalizing the self-healing objective into a temporal logic reduction based on signal temporal logic; predicting the quantitative robustness of the state trajectory to the reduction, and generating a dense reward signal in the form of a potential function difference to reshape the feedback reward; finally, training the agent based on the reshaped reward, the graph network foundation, and the course learning mechanism, outputting an optimal decision sequence including topology reconstruction and resource scheduling. This invention eliminates the reward sparsity problem of traditional models, guides the agent to find the optimal solution under strict physical temporal constraints, effectively avoids the risk of cascading failures caused by trial and error, and significantly improves the convergence speed and online decision reliability of self-healing in high-dimensional complex power grids.

The application discloses a kind of multi-objective mathematical calculation model construction method and application system, it is related to mathematical modeling technical field, including the following steps: S1, real-time data acquisition;S2, data preprocessing;S3, dynamic weight adjustment;S4, model construction and solution.The multi-objective mathematical calculation model construction method and application system, through the dynamic frequency mechanism of basic frequency and trigger type frequency increase, according to the parameter fluctuation characteristics flexible adjustment acquisition strategy, avoid the key fluctuation information missing caused by fixed low-frequency acquisition, reduce the resource waste caused by fixed high-frequency acquisition, different change characteristics parameters can be realized efficient data capture, and rely on the linkage mechanism of real-time data acquisition and dynamic weight adjustment, can quickly respond to parameter change, combined with efficient model solution architecture, substantially shorten the period from parameter change to output optimal decision scheme, effectively avoid the decision lag problem of traditional fixed weight model.

The application provides a kind of CBF-based automatic driving decision method and system, comprising: obtaining the traffic state data of current frame, and extracting the vehicle state vector and obstacle state set from traffic state data;According to the vehicle state vector and obstacle state set, the intention risk factor between each obstacle and the vehicle is calculated;According to state matrix and intention risk factor, the minimum safety distance between each obstacle and the vehicle is calculated, and then based on the minimum safety distance, the linear constraint corresponding to each obstacle is generated using the second-order CBF method;Construct optimization objective function, extend linear constraint, obtain optimization linear constraint, then based on optimization linear constraint, use QP solver to solve optimization objective function, obtain optimal decision vector.The application solves the problems of excessive defense and mechanical division of safety boundary in existing automatic driving decision-making, which leads to low traffic efficiency and rigid interaction.

The application discloses a wind-solar coupling hydrogen production project site selection method based on multi-criteria decision in a fuzzy environment, and is specifically implemented according to the following steps: step 1, obtaining influence factors in a wind-solar coupling hydrogen production project site selection process, and constructing a wind-solar coupling hydrogen production project site selection evaluation index system; step 2, constructing an evaluation model of site selection decision; step 3, calculating comprehensive weights of evaluation indexes; and step 4, adopting a multi-attribute boundary approximate region comparison method to sort the alternative sites, and obtaining an optimal decision. The application solves the problem of poor accuracy of the existing site selection method.

A method and system for intelligent fault monitoring in distribution networks containing distributed energy resources are disclosed. The method first acquires multi-source information data of the distribution network containing distributed energy resources and extracts features to generate a node feature matrix. Then, based on the node feature matrix, a Markov decision model for fault detection in the distribution network containing distributed energy resources is constructed. The system's main body is used as the agent, and the agent is trained using an improved D4PG algorithm based on the Markov decision model. After training, the optimal fault detection strategy for the distribution network containing distributed energy resources is derived and deployed in the fault detection task for intelligent fault monitoring. This invention, through a deep reinforcement learning framework, transforms the complex fault detection task of the distribution network containing distributed energy resources into an optimal decision problem, successfully establishing a direct, fast, and reliable logical mapping relationship from the grid fault state to the optimal diagnostic action, providing a more robust and adaptive fault detection method.