418 results about "Mathematics model" patented technology

The invention relates to the technical field of cloud computing and data center energy efficiency optimization, and discloses a carbon neutralization-oriented cloud data center energy efficiency optimization method, which comprises the following steps of: 1, acquiring server computing load, power consumption, cooling state and carbon emission data of a data center based on a distributed sensor network; step 2, constructing an energy efficiency mathematical model of the data center based on a deep learning model of AI training; 3, training a task scheduling and cooling system optimization strategy based on a reinforcement learning agent; 4, analyzing the real-time carbon emission condition of the data center by using an AI prediction algorithm; 5, based on the carbon emission trend predicted by AI, triggering an intelligent carbon compensation strategy in combination with the real-time energy consumption condition of the data center; and step 6, dynamically adjusting task scheduling, a cooling control strategy and a carbon compensation scheme by adopting an AI adaptive learning method based on data feedback. And dynamic scheduling of calculation tasks of the data center is realized through a reinforcement learning agent trained by AI.

The invention discloses a code generation method, electronic equipment, a readable storage medium and a program product, and relates to the technical field of artificial intelligence. The method comprises the steps that code generation demand information and code generation cue words are input into a language model with a code generation function, and a code generation initial result is obtained; determining a semantic space model for defining a semantic space to which the code generation result belongs according to the code generation demand information, and converting the semantic space model into a mathematical model corresponding to the code generation scene to obtain a semantic verification model; inputting the code generation initial result into a semantic verification model, reasoning the code generation initial result, and determining a semantic verification result according to whether an end node can be reached or not; and determining a code generation result according to the semantic verification result and the code generation initial result. The problem that high-quality codes cannot be generated due to large model illusion in the prior art can be solved, and the code generation accuracy and reliability are effectively improved.

The invention discloses a ship formation autonomous obstacle avoidance control method based on offset intervention guidance. The method comprises the steps of obtaining a nonlinear mathematical model of a three-degree-of-freedom ASV; acquiring a smooth reference path for USV ship formation according to the virtual ship L1VS, and acquiring a ship obstacle avoidance virtual reference path according to the obstacle avoidance virtual ship VS in combination with a DWA algorithm; designing a kinematics virtual control law of the USV ship formation based on the constructed offset intervention guidance mechanism; on the basis of the constructed dynamic event triggering mechanism, a robust neural damping technology is combined, and a robust kinematics error is obtained according to the kinematics error of the USV ship formation; and designing an adaptive robust command filtering controller according to the robust kinematics error so as to realize autonomous obstacle avoidance control of the ship formation. The problems that according to a traditional ship formation control method, when a collision avoidance task in an unknown water area is executed in a complex marine environment, information of moving and fixed obstacles cannot be obtained in time, and communication load balancing and obstacle avoidance control precision cannot be effectively achieved are solved.

According to the new energy automobile flexible assembly line multi-target dynamic intelligent scheduling optimization method coping with the uncertain disturbance influence, the flexibility of the assembly manufacturing unit and the flexibility of the assembly manufacturing process sequence are comprehensively considered, and a dynamic intelligent scheduling optimization scheme is generated for multiple optimization targets. The method comprises the following steps: firstly, constructing a multi-target dynamic intelligent scheduling optimization problem mathematical model of an automobile flexible final assembly line; furthermore, the optimization problem is converted into a Markov decision process, and states, actions, multi-target reward components and reward component aggregation thereof are defined. And then, carrying out a multi-target deep reinforcement learning training process oriented to a new energy automobile flexible final assembly line, and constructing and forming a scheduling agent. And finally, realizing self-adaptive decision making of the automobile flexible assembly line under an uncertain disturbance condition based on the scheduling intelligent agent. According to the method, the robustness and the execution efficiency of the new energy automobile flexible assembly line in a dynamic environment can be remarkably improved, and the method has good practical value and popularization prospect.

The invention discloses a wireless ad hoc network routing and resource cross-layer joint optimization method based on a genetic algorithm. The method comprises the following steps: step 1, aiming at a plurality of end-to-end communication services in a wireless ad hoc network, carrying out mathematical modeling on a normalized difference sum of each service flow path hop count and a theoretical minimum value thereof and a constraint condition; performing joint optimization on link layer resource allocation and network layer routing decision; and step 2, solving the mathematical model based on a genetic algorithm to realize the targets of reasonable distribution of network resources and optimization of service quality. According to the method, link layer resource allocation and network layer routing decision are jointly optimized in a global mode under the condition that network communication resources are limited, it is ensured that a service flow is close to shortest path transmission as far as possible, end-to-end delay and network loads are reduced, and effective allocation of resources and flexible construction of routing are achieved.

The invention relates to the technical field of computer-aided engineering, in particular to a pure cotton dark-color fabric dye-resistant area dynamic prediction method based on a dye diffusion model. The method comprises the following steps: acquiring a parameter data set representing a fabric microstructure, dye properties and a printing process; on the basis, a multi-scale coupling mathematical model is constructed, the dye diffusion range is predicted, the diffusion process is decomposed into seepage diffusion between macroscopic yarns and diffusion in microcosmic fibers through the model, and the diffusion rate difference of the dye in the fabric direction is represented through anisotropic diffusion tensors; through finite element analysis of structure perception, running a model to calculate dye space-time concentration distribution, and generating geometric data for predicting a bleeding boundary according to a preset concentration threshold value; according to the data, printing process parameters and model parameters are adjusted in a self-adaptive mode, and dynamic prediction of the dye-resistant area is achieved; according to the method, computer simulation and finite element analysis technologies are utilized, and the accuracy of dynamic prediction of the dyeing-resistant area of the pure cotton dark-color fabric is improved.

The invention discloses a super-multi-target vehicle path planning method based on deep reinforcement learning assisted evolution, and aims to dynamically guide a search direction through deep reinforcement learning for multi-target optimization and path planning problems. The core of the method comprises the following steps: 1) constructing a super-multi-target vehicle path problem mathematical model; 2) designing a deep Q network as a decision model, and adaptively selecting a high-potential search direction in iteration; 3) designing a global search strategy, and balancing global search and local optimization; 4) generating a filial generation and evaluating the fitness, and executing environment selection to update the population; and 5) calculating a reward value according to a population evolution effect, and optimizing DQN network parameters. According to the method, through a collaborative mechanism of reinforcement learning and an evolutionary algorithm, a search strategy is dynamically adjusted, the convergence speed and stability of the algorithm are effectively improved, the adaptability to a complex path planning problem is enhanced, and the method has remarkable advantages in solving efficiency, solution set quality and dynamic environment response.

The invention relates to the technical field of cooperative control, in particular to an adaptive event triggering fixed time consistency control method for a multi-agent system, which comprises the following steps of: S1, constructing a nonlinear high-order multi-agent system mathematical model with uncertain dynamics; s2, constructing a Lyapunov function containing an uncertain parameter lower bound based on coordinate transformation; s3, designing an adaptive event trigger controller with a positive bounded function by using a backstepping control method, wherein the adaptive event trigger controller comprises a virtual controller, an adaptive law and an event trigger mechanism; and S4, through stability analysis, it is proved that the high-order nonlinear multi-agent with the uncertain dynamic state can achieve actual fixed time output consistency, Zeno behaviors are avoided, and the self-adaptive event triggering fixed time consistency control method of the multi-agent system has the advantages of being high in practicability and easy to popularize. The problem of actual fixed time output consistency of a high-order nonlinear multi-agent system is solved, and an existing actual fixed time output consistency result is popularized.

The present invention relates to a calculation method for the general scour pattern of a cross-sea bridge considering the turbulence effect. The research area including the cross-sea bridge project is determined and meshed, the elevations of each grid node are obtained, typical computational hydrological boundary conditions are selected, a non-hydrostatic three-dimensional mathematical model including the entire cross-sea bridge project is established, the hydrodynamic fields before and after the bridge construction are calculated, the tidal level and flow velocity are verified, the pulsating velocity at each grid node, the sediment settlement velocity in still water and the settlement velocity under the influence of turbulent kinetic energy are calculated, the sediment carrying capacity is calculated according to the estuarine fine sediment model, and finally the general scour amplitude of the seabed at each grid node position is calculated, thereby obtaining the characteristics of the general scour amplitude in the upstream and downstream areas of the bridge. Combining the plane three-dimensional numerical model and the sediment carrying capacity calculation method considering the turbulence effect, through the dynamic equilibrium theory, the accurate prediction of the general scour pattern of the cross-sea bridge including the depth and range is realized, and the calculation accuracy of the general scour pattern of the cross-sea bridge is improved.

The invention discloses an optimal scheduling method considering participation of a multi-element load polymer in demand response, and aims to improve the safety and economy of a power system. The method comprises the following steps: firstly, constructing a mathematical model of a load, representing an adjustable domain of the load, and clustering the load by using a K-means algorithm; and then, carrying out internal approximation on adjustable domains of various loads through scaling and translation factors, optimizing parameters by adopting a linear programming method, and finishing aggregation processing of the loads. And finally, in combination with power supply parameters, a load aggregation domain, market demand prediction and the like, an MILP optimization model with the minimum operation cost is established, power dispatching is performed, a dynamic excitation electricity price mechanism is introduced, the price is adjusted according to the adjustment capability of the load, and the load is excited to actively participate in demand response. According to the scheme, by accurately describing the load adjustable domain and flexibly guiding the load to participate in the demand response, the scheduling of the power system is effectively optimized, the consumption level of renewable energy sources is improved, the operation cost of the system is reduced, and the method has a relatively strong engineering application prospect.

The invention provides a collaborative scheduling and income distribution method for energy storage power stations in an alliance chain, and belongs to the technical field of energy storage power stations. The method comprises the steps of constructing an energy storage capacity sharing alliance block chain taking the energy storage power stations as nodes; an intelligent contract is constructed, the intelligent contract comprises a mathematical model for solving the shared scheduling plan, and the benefit of the shared capacity of each energy storage power station under the optimal shared scheduling plan is obtained based on the intelligent contract; each energy storage power station solves the shared scheduling plan through real-time information, the smart contract compares the earnings under the shared scheduling plans to obtain the shared scheduling plan with the maximum total earnings as the optimal shared scheduling plan, and each energy storage power station executes the optimal shared scheduling plan. And the maximization of the total capacity is realized.

The invention discloses a multi-objective reinforcement learning man-machine cooperative assembly task allocation method based on neighborhood parameter migration, and the method comprises the steps: building a mathematical model which aims at minimizing the physiological fatigue accumulated value of a human operator and minimizing the maximum completion time for a multi-objective optimization problem of task allocation in a man-machine cooperative assembly system; a multi-target problem is decomposed into N standard sub-problems by adopting a weighting and decomposition strategy, and training is accelerated through a neighborhood parameter migration strategy; each sub-problem is solved based on a near-end strategy optimization algorithm of an Actor-Critic framework, Gaussian noise is added to an Actor network to simulate environment uncertainty, an action mask mechanism is introduced to process priority constraints of assembly tasks, and it is ensured that a generated task allocation scheme is always feasible. According to the method, the convergence speed and diversity of the Pareto solution set can be remarkably improved, and the assembly efficiency and operator fatigue are effectively balanced.

The invention discloses a design method of a barrier function rehabilitation robot safety tracking control system, and relates to the technical field of medical robots. The method comprises the steps that on the basis of an existing double-joint single-arm robot mathematical model, a mathematical model with uncertain model parameters, uncertain external disturbance and uncertain external force borne by the tail end is established, and the established model is closer to a real model of a rehabilitation robot; the uncertainty in the model is classified into one class, and a disturbance observer is used for estimating and compensating the uncertainty; a logarithmic obstacle Lyapunov function is adopted to constrain the joint angle; and designing an adaptive trajectory tracking controller based on the disturbance observer and the obstacle Lyapunov function method, and tracking a set trajectory. The high-precision tracking of the target trajectory is ensured, the actual motion trajectory always moves within a certain range on the two sides of the target trajectory, and the secondary injury of the patient in the rehabilitation process is avoided.

The invention provides a self-adaptive fixed time tracking control method based on event triggering, which is used for solving the technical problem that a virtual control law is easy to generate singularity in the traditional backstepping method design. The method comprises the following steps: establishing a mathematical model of a non-strict feedback nonlinear system with input time delay, and introducing an auxiliary system to counteract delay influence; an error dynamic system is obtained through coordinate transformation; according to system fixed time convergence, designing a piecewise function, introducing a backstepping method to design a controller and an adaptive law, using an RBFNN to approach an unknown function therein, designing a virtual control signal corresponding to each error signal, and obtaining an actual controller and an adaptive law; according to the actual self-adaptive controller, event triggering conditions are designed, when the event triggering moment is reached, the event triggering conditions containing input are triggered, and information transmission and response are carried out. An event triggering mechanism is designed, the number of times of control instruction transmission is reduced, communication resources are saved, and rapid convergence and safe operation of the system state are achieved.

The invention relates to a hydrogen production and refueling station hydrogen leakage range prediction method based on meshless particle method simulation, and belongs to the technical field of hydrogen energy safety and numerical simulation. Aiming at the problems that a traditional CFD grid method is poor in complex geometric boundary adaptability, low in calculation efficiency and insufficient in dynamic early warning capacity, a three-dimensional scene of a hydrogen production and refueling station is constructed by combining a geographic information system (GIS) and a building information model (BIM), a mathematical model is discretized by adopting a smoothed particle hydrodynamics (SPH) method, and the three-dimensional scene of the hydrogen production and refueling station is obtained. Real-time sensor data are integrated, and simulation is accelerated in parallel through a GPU. And carrying out numerical simulation of hydrogen leakage of the hydrogen production / hydrogen refueling station under multiple factors in combination with large eddy simulation, and generating high-precision simulation of evolution and distribution of a hydrogen leakage concentration field of a hydrogen production / hydrogen refueling station scene. The method has the advantages of being high in modeling speed, high in simulation precision, wide in applicability and the like, the problem of precise simulation of hydrogen leakage and diffusion under the complex geometric structure of the hydrogen production and refueling station is effectively solved, and reliable technical support is provided for safe operation of the hydrogen production and refueling station.

The invention provides a collaborative collision and obstacle avoidance strategy of an underactuated unmanned ship considering input and state quantification. The collaborative collision and obstacle avoidance strategy comprises the following steps: establishing a mathematical model of the underactuated unmanned ship; a distributed guidance rule is constructed in the USV kinematics subsystem by using an auxiliary variable method, and a repulsive force function of an improved artificial potential field method is introduced to reconstruct the guidance rule, so that collaborative collision avoidance and obstacle avoidance at the kinematics level and expected trajectory tracking of the underactuated USV are realized; designing an extended state observer to solve the influence of unknown items in the system and quantitative state variables in the communication process on the formation control system; a linear time-varying model is adopted to describe a quantizer, so that a bottom-layer quantization controller does not need to predict specific information of quantization parameters; and proving the stability of the constructed USV formation tracking control system according to an input-state stability theory. According to the invention, starting from two aspects of input quantization and state quantization, the problem of USV formation tracking control in the marine environment with limited communication bandwidth and complex quantization environment is solved.

The invention relates to the field of intelligent temperature control, and provides a refrigeration house PID control parameter optimization method and device, equipment and a storage medium. The method comprises the steps of collecting temperature data and energy consumption data of the refrigeration house at a constant-temperature stage; based on the temperature data and the energy consumption data, a mathematical model representing the mapping relation between the temperature and the energy consumption of the refrigeration house is established; constructing a deep reinforcement learning model based on a preset reward function and a mathematical model giving consideration to the temperature stability and energy consumption optimization of the refrigeration house; and through interactive learning of the intelligent agent of the deep reinforcement learning model and the refrigeration house environment, determining optimization parameters used for adjusting refrigeration house PID control. According to the refrigeration house PID control parameter optimization method, device and equipment and the storage medium provided by the invention, the mathematical model of the temperature and energy consumption mapping relation of the refrigeration house can be effectively established, the deep reinforcement learning model is constructed through the reward function and the mathematical model, and the optimization parameters of refrigeration house PID control are adjusted through deep reinforcement learning.

The invention discloses a robot optimal path planning method based on an improved Hemma optimization algorithm. A plurality of improvement measures are provided for problems possibly existing in path planning of a Hemma optimization algorithm. Firstly, a Circle chaotic mapping strategy is adopted to replace random initialization, so that the quality and distribution uniformity of an initial population are improved, and the global exploration capability is enhanced. Secondly, mathematical models of three core behavior stages of a standard HO algorithm are deeply optimized, and adaptive dynamic weight, a dynamic weighting strategy based on a current optimal solution and individual fitness and guided Levy flight direction adjustment are introduced; and finally, introducing a dynamic stage conversion and adaptive exploration intensity mechanism DPTAEI, and dynamically adjusting the participation proportion of the individual in different behavior stages and the intensity of the exploration behavior according to the real-time search progress of the algorithm. Through the improvements, the optimization precision, the convergence speed and the global search capability of the IHO algorithm in the aspect of robot path planning are remarkably improved.

The invention discloses a multi-agent system safety convergence control method based on dual-channel adaptive event triggering, and the method comprises the steps: firstly building a multi-agent system mathematical model, and designing a controller of each agent in a multi-agent system based on the multi-agent system mathematical model; secondly, establishing a sequence scaling attack model, and combining the sequence scaling attack model to a controller of an intelligent agent; and then designing an event trigger controller based on the controller of the intelligent agent containing the sequence scaling attack. And then, based on the event triggering controller, safety convergence control is carried out, and multi-agent control in the multi-agent system is completed. According to the method, malicious network attacks can be effectively resisted, the problem of limited bandwidth resources can be solved, and the applicability and the application range are improved and expanded.

The invention relates to the technical field of welding path planning, in particular to a welding robot path planning method and system based on an improved whale optimization algorithm, and the method comprises the following steps: carrying out the path planning modeling of a welding robot on a front cover of an automobile engine, and obtaining a welding path planning mathematical model; the whale optimization algorithm WOA is improved by integrating multiple strategies to obtain an optimal improved whale optimization algorithm GSWOA, and the performance of the optimal improved whale optimization algorithm GSWOA is verified through four test functions; on the welding path planning model, path planning is conducted through the optimal improved whale optimization algorithm GSWOA, and the optimal welding path on the automobile engine front cover is obtained. The whale optimization algorithm WOA is improved through multiple strategies, the improvement of the multiple strategies is integrated through a Stacking framework, the integration of the advantages of all the strategies is achieved, the influence of failure of a single strategy can be avoided, and the robustness of the algorithm is improved.

The invention relates to the technical field of petroleum engineering, and provides an oil and gas well cementation cement sheath integrity evaluation method which comprises the following steps: S1, simulation experiment: simulating underground environment and cement sheath stress changes under various operation conditions through an indoor physical simulation device to obtain reliable experiment data and experiment phenomena; s2, analysis and verification experiment: analyzing and verifying an experiment phenomenon through the mathematical model of the cement sheath; and S3, judgment: forming a cement sheath sealing integrity judgment method with practical significance through organic combination of a mechanical model and physical simulation. Through a method of combining a simulation experiment and mechanical model analysis, the integrity and the packing performance of the cement sheath can be evaluated more accurately, a more comprehensive and objective evaluation basis can be provided by experimental data and phenomenon acquisition and a stress analysis result, the defect of a single basis is avoided, the integrity of the cement sheath is evaluated accurately, and the sealing performance of the cement sheath is improved. And potential well cementation problems can be found in time.

The invention relates to the technical field of unmanned aerial vehicle management and control, and discloses a low slow small unmanned aerial vehicle countering method and system, which eliminates space-time deviation and redundant information of different sensor data by acquiring multi-source data and performing fusion processing. Furthermore, the target tracking trajectory of the low-slow small unmanned aerial vehicle is processed and generated through an adaptive Kalman filtering algorithm, and real-time smooth tracking of the trajectory is realized. Furthermore, probabilistic region prediction is performed by combining a Monte Carlo method and a kernel density estimation method, and a threat region prediction result of the low-slow small unmanned aerial vehicle is generated, so that the possible motion range of the target is comprehensively covered, and the motion intention of the target is accurately captured. Furthermore, in combination with a preset constraint condition set, a target tracking trajectory and a threat area prediction result, the interception success rate mathematical model is solved by using an improved particle swarm algorithm, and a collaborative countering strategy is obtained, so that the limitation of an existing countering method is avoided, and the accuracy and efficiency of countering are improved.

A storage device stores a medical knowledge graph including nodes corresponding to medical care events, and edges indicative of a relationship between the nodes. A graph feature of the medical knowledge graph is expressed by a mathematical model characterized by patient background information. A processing circuitry obtains patient background information relating to a background factor of one or more target patients, computes variation of the graph feature relating to the target patient, based on the patient background information of the target patient and the mathematical model, and displays the variation of the graph feature on a display device.

The invention discloses a speed control method for a high-speed water pump engine, which belongs to the technical field of PID (Proportion Integration Differentiation) control optimization and specifically comprises the following steps of: 1, constructing a speed control system for the high-speed water pump engine; step 2, improving a whale migration optimization algorithm: D1, generating an initial population of the algorithm by using a mixed sine and cosine mapping strategy; d2, improving a whale migration optimization algorithm by using an adaptive convergence step length strategy fused with crossover variation to update a mathematical model of an individual position; thirdly, parameter optimization is conducted on a PID controller module in the high-speed water pump engine speed control system through the improved whale migration optimization algorithm; 4, applying the optimal control parameters obtained in the step 3 to a PID controller module in a high-speed water pump engine speed control system; a PID controller module in the high-speed water pump engine speed control system is optimized through the improved whale migration optimization algorithm, and the adaptability and robustness of the whole control system are improved.

The invention provides a steady-state temperature distortion inlet temperature test layout method, and belongs to the field of aero-engine tests, and the method comprises the steps: S10, constructing a temperature field mathematical model of a measurement cross section in a steady-state temperature distortion test and a test layout mathematical model employing a fixed total temperature probe for temperature measurement; s20, temperature field parameters used for describing temperature field distortion characteristics are determined, the relative error sum of the temperature field parameters in the prior flow field and the temperature field parameters at the corresponding coordinates of the test layout is calculated, and therefore the optimization target of the test layout is determined; and S30, taking the circumferential angle of the fixed total temperature probe in the test layout as an optimization parameter, taking the relative error sum of the temperature field parameters as an optimization target, and adopting an optimization algorithm to optimize the circumferential angle of the fixed total temperature probe in the test layout. Therefore, the circumferential optimal angles of the fixed total temperature probes with different circumferential numbers in the steady-state temperature distortion test are obtained. The method is good in convergence and high in calculation precision.

The invention relates to the field of vehicle route planning, in particular to a dynamic multi-target vehicle route planning method and system based on prediction and recombination, and the method comprises the steps: building a mathematical model of dynamic multi-target vehicle route planning, initializing a population through real number coding, adjusting the population to meet a constraint, carrying out the iteration through a genetic algorithm, and obtaining a candidate solution set; a Pareto front set is obtained and clustered, and a clustering center is recorded as a special individual; and when environment change is detected, predicting individuals in a new environment by using a gray model or a least square support vector machine, and constructing a new population until an optimal solution set is output. According to the method, the convergence speed of the algorithm is high, the response to the environment change is more sensitive and timely, the dynamic multi-target vehicle routing problem is effectively solved, and the application effect and the economic benefit of the vehicle routing in the actual life are effectively promoted.

The invention relates to an HPLC (High Performance Liquid Chromatography) + HRF (High Radio Frequency) dual-mode communication system and method, belonging to the technical field of communication, the system uses a single micro-control unit to replace a multi-chip architecture, dynamically manages hardware resources through a mathematical model, and solves the problem of high hardware cost caused by integration of HPLC and HRF chips in the traditional dual-mode communication system. The method has the advantages of low hardware cost and high resource utilization rate. According to the method, the optimal protocol is dynamically selected through mixed integer nonlinear programming and a variational Bayesian strategy based on an environment state main vector, the problems that protocol switching of a traditional system is rigid and the success rate is low in a complex power line environment are solved, the reliability and the success rate of a communication link can be improved, and the communication efficiency is improved. And meanwhile, the method has noise identification and modulation mode self-adaptive capabilities.

The invention provides a sepsis multi-task prediction method and system, and belongs to the technical field of disease prediction based on a mathematical model, and the method comprises the steps: obtaining a symptom data record of a to-be-detected patient; and processing the obtained symptom data records by using a pre-trained classification model to obtain a final task result. According to the method, a multivariable time sequence in medical health record data of a patient is utilized, time representation containing sepsis pathological characteristics is obtained through a time mode of modeling data based on a self-attention encoder, and the trend of the condition of the patient changing along with time is obtained; correlating different medical features according to features including patient condition change trends, so that each feature can perceive other features mutually, and mining global feature representation of the patient; and inputting the global feature representation into a task-aware hybrid expert system, and routing the unified global feature representation into different expert systems, thereby obtaining task-level representation and realizing multi-task prediction.

The invention provides an air-breathing engine incremental control method based on reinforcement learning, and relates to the technical field of air-breathing engine control. The deep reinforcement learning control module is used for outputting control increments of control variables such as the main fuel flow, the inner culvert and outer culvert thrust augmentation fuel flow and the nozzle throat area, the safety limiting module is combined for monitoring and cutting multiple monitoring parameters in real time, and finally a smooth and feasible execution signal is generated through the action space processing module. And multi-dimensional fine adjustment of the engine is realized. According to the scheme, multi-agent parallel training is adopted, switching between a low thrust mode and a high thrust mode is supported, under the condition that an accurate mathematical model is not needed, nonlinearity and coupling characteristics of an engine are flexibly coped, one-time large jump is prevented, safety and fuel efficiency are improved, and the requirements for high thrust and low fuel consumption in the flight process are met.

The invention discloses an ultrasonic guidance navigation system based on multi-modal detection and a CNB ejection mechanism, and belongs to the technical field of ultrasonic auxiliary diagnosis, multi-modal detection loads a directional bounding box model and a segmentation model when the system is initialized, and the directional bounding box model performs directional detection of a puncture needle on an input frame; in the result extraction stage, puncture needle information is extracted from a directional bounding box model detection result, a focus mask is extracted from a segmentation result, and the focus mask is adjusted to be the same as an input frame in size; when the focus mask is extracted, the validity of a segmentation result is verified, then first effective mask data are extracted from multiple types of segmentation results, the mask data are transmitted to a CPU from a GPU memory, a floating point probability value is converted into an integer binary mask, and puncture needle direction detection and segmentation of multiple types of anatomical structures are achieved. According to the CNB ejection mechanism, a mathematical model of the ejection process is established through an ejection trajectory calculation function, and a reliable mathematical basis is provided for CNB ejection operation diagnosis.