46results about How to "Improve convergence speed" patented technology

The invention relates to the technical field of wireless sensor networks, is applied to a soil humidity monitoring scene in agriculture, and particularly relates to a soil humidity monitoring wireless sensor network coverage optimization method based on a chaos adaptive dung beetle optimization algorithm, which comprises the following steps: constructing a soil humidity monitoring wireless sensor network coverage model of a monitoring area; an improved dung beetle optimization algorithm is adopted to carry out coverage optimization on the soil humidity monitoring wireless sensor network coverage model, an optimized coverage result is obtained, and the improved dung beetle optimization algorithm is obtained through improvement of a Logistic chaos operator and an adaptive quasi-reverse learning strategy. According to the method, the coverage rate and robustness of the agricultural soil humidity monitoring wireless sensor network can be improved, local optimum can be avoided while the convergence speed is guaranteed and global search is improved, and the optimal coverage effect is achieved.

The embodiment of the invention relates to the field of foundation pile reverse design, in particular to a foundation pile reverse design method and equipment based on multi-objective optimization and machine learning. One specific implementation mode of the method comprises the steps that a foundation pile actual measurement data set is collected; updating processing is conducted, and a foundation pile actual measurement updating data set is obtained; dividing the foundation pile actual measurement update data set to obtain a foundation pile actual measurement sample data set and a foundation pile actual measurement test data set; based on the foundation pile actual measurement sample data set and the foundation pile actual measurement test data set, constructing a machine learning model for predicting the bearing capacity of the foundation pile; generating an initial foundation pile design data set; based on a machine learning model for predicting the bearing capacity of the foundation pile, optimizing the initial foundation pile design data set to obtain target foundation pile design data; and sending the target foundation pile design data to a foundation pile construction control terminal to construct a foundation pile. According to the implementation mode, the foundation pile design efficiency and accuracy are improved.

The invention provides a grey wolf optimization feature selection method, device and equipment based on an adaptive mechanism, and relates to the technical field of data processing. The method comprises the following steps: acquiring a data set to be subjected to feature selection, and initializing a grey wolf optimization algorithm; the fitness value of each grey wolf is calculated with the purpose of obtaining the current optimal solution, and three current optimal wolves are obtained; then taking the current optimal three wolves as search guidance, combining a nonlinear parameter control strategy, adopting an improved grey wolf step length updating calculation mode, adaptively adjusting variation intensity in the search process, updating and calculating related parameters, positions and fitness values of each grey wolf, and continuously converging the distance between the prey and the grey wolf in the search process; and calculating the score of each grey wolf, and selecting the solution candidate with the highest score as the optimal solution in the current population until the maximum number of iterations is reached, thereby obtaining a global optimal solution. The method is used for processing feature selection in high-dimensional and complex classification, and is high in convergence speed and high in accuracy.

The invention discloses a non-invasive blood pressure prediction method based on signal dimension extension and a lightweight network, which comprises the following steps: firstly collecting PPG blood pressure data and ABP data, then denoising the collected PPG blood pressure data and ABP data, and cleaning the denoised data. The processed data is used for data signal dimension expansion, a PPG blood pressure measurement model is constructed, and the PPG blood pressure measurement model comprises a basic convolution layer and a residual grouping convolution module Group ResidualBock; and then deploying the constructed PPG blood pressure measurement model into the wearable device to obtain the wearable blood pressure monitoring device. And finally, measuring the processed PPG signal by using wearable blood pressure monitoring equipment to obtain a blood pressure value. According to the method, the PPG signal dimension is expanded, the data processing link is optimized, and the model blood pressure prediction accuracy is improved; and meanwhile, a lightweight algorithm architecture is adopted, so that resource occupation is reduced, the system can be deployed in wearable embedded equipment, and the security of personal privacy data is guaranteed.

The invention discloses a YOLOv10-based power transmission line insulator detection method and system, and the method comprises the steps: constructing an insulator data set; then, an insulator detection model based on the lightweight YOLOv10n network is constructed, that is, a ghost convolutional layer GhostConv based on GhostBottleneck module stacking is used for replacing an original common convolutional network of a Backbone network; a feature enhancement module ConcatBiFPN is used for replacing an original connection layer network of the Neck network, and an attention module ContextAggregation is arranged behind a feature extraction module C2FCIB of the Neck network; and finally, the insulator data set is sent to an insulator detection model based on lightweight YOLOv10n for model training to obtain an insulator detection model for insulator detection, so that the algorithm operation efficiency and accuracy are greatly improved.

The invention provides a control method and a control device of an underwater robot. The control method of the underwater robot comprises the following steps: establishing an underwater robot system model; performing obstacle avoidance path planning by using a sampling algorithm to obtain a discrete geometric path point set connecting the starting point and the ending point; a full-state reference sequence matched with the control method underwater robot system model is generated based on the control method discrete geometric path point set; constructing a cost function for trajectory optimization based on the control method underwater robot system model and the control method full-state reference sequence; generating an initial control sequence and an initial state sequence according to the deviation between the current state of the underwater robot in the control method and the full-state reference sequence in the control method; and performing iterative optimization by taking the control method cost function as an optimization target and taking the initial control sequence and the initial state sequence of the control method as hot start initial values to obtain a control instruction sequence of the underwater robot of the control method.

The invention relates to a gearbox fault diagnosis method and system. The method comprises the following steps: acquiring fault data of a gearbox; performing feature extraction by using the convolutional neural network fault data to obtain fault features; inputting the fault features into an extreme gradient lifting classifier to obtain a fault diagnosis result of the gearbox; wherein the extreme gradient lifting classifier adopts an improved big sugarcane-mouse algorithm to carry out parameter optimization, and the improved big sugarcane-mouse algorithm is combined with a nonlinear weight, an adaptive factor, sine and cosine and a sparrow warning mechanism. According to the invention, the features of the gearbox fault data are mined by using the convolutional neural network, accurate fault classification and identification are carried out on the basis of the features by using the extreme gradient lifting classifier, and the hyper-parameters of the extreme gradient lifting classifier are optimized by introducing the improved big sugarcane-mouse algorithm. And the optimization capacity, the stability and the convergence speed are improved, so that the overall performance of the model is effectively improved.

The invention discloses a phased array system transmitting beam forming method based on a Riemann-Newton method, which converts a beam pattern matching error minimization problem under a constant modulus constraint into an unconstrained optimization problem under a complex circle manifold, and adopts a second-order Riemann-Newton method to solve. Therefore, it is ensured that the direction of each iteration update is always the descending direction of the target function, and the super-linear convergence speed is achieved. The method has better performance advantages in the aspects of main lobe fitting precision and sidelobe level reduction, and the convergence speed is higher.

The invention discloses a silicon carbide vacuum sintering furnace optimization method based on computational heat transfer theory simulation, which is used in the field of artificial intelligence and comprises the following steps: establishing a heat transfer model; the heat effect in the silicon carbide vacuum sintering furnace is predicted by setting the boundary conditions of initial temperature, heat flow implementation and convective heat transfer and solving the heat transfer model by using an implicit numerical method; based on the optimized heat transfer model, simulating the physical process in the silicon carbide vacuum sintering furnace by using a multi-physical field simulation technology, and analyzing the influence of the operation parameter data of the silicon carbide vacuum sintering furnace on the heat transfer process; comparing the operation data with a simulation result, and based on a comparison result and an analysis result; and comparing and analyzing the actual performance data and the simulation result, and evaluating the adjustment effect of the operation parameter data. Through quality detection and energy efficiency evaluation, the energy utilization efficiency in the sintering process is measured and optimized, and the link of energy waste is identified.

The invention provides a magnetic element performance optimization model design method under multiple factors, and belongs to the technical field of electric energy conversion of power electronics. The technical problem that the accurate calculation of the magnetic core loss under different temperature conditions is low is solved. According to the technical scheme, the method comprises the following steps of S1, performing data measurement to form an original data set, and performing data processing to establish a new data set; s2, improving an iGSE magnetic core loss function; s3, fitting unknown parameters of the magnetic core loss function by a nonlinear least square method; s4, constructing a weighted single-target optimization model; and S5, solving the optimal solution of the target function by a fire bird optimization algorithm. The method has the beneficial effects that the improved iGSE magnetic core loss function can be suitable for various temperature changes, the constructed weighted single target model can more comprehensively optimize the performance of the magnetic element, and the fire bird optimization algorithm can more quickly converge the optimal solution of the target function.

The invention discloses an integrated circuit wiring method, a computer device and a storage medium, an inertia weight in an improved DPSO algorithm is determined by an acceleration coefficient used by the improved DPSO algorithm, so that more targets and constraint conditions are introduced, the inertia weight can be iterated together with the acceleration coefficient for convergence, and the convergence efficiency is improved. The convergence speed of the inertia weight and the overall convergence speed of the improved DPSO algorithm can be enhanced, so that the efficiency of the algorithm can be improved, the wiring speed can be increased, the circuit design efficiency can be improved, the possibility that the improved DPSO algorithm falls into a local optimal solution can be reduced, and the improved DPSO algorithm can find a global optimal solution. The circuit is widely applied to the technical field of circuits.

The invention discloses a lightweight distracted driving detection method. Relates to the field of safety detection of an intelligent traffic system and an automatic driving assistance system, in particular to the technical field of distracted driving behavior detection based on deep learning. The problems that according to an existing detection method, the number of computing resources is large, lightweight application cannot be carried out, and inaccurate detection is likely to occur are solved. The method comprises the following steps: S1, data acquisition and division: collecting distraction driving behavior images, and dividing the distraction driving behavior images into a training set, a verification set and a test set; and S2, constructing a distracted driving detection model, and improving the existing YOLOv8s model framework by adopting a hierarchical improvement mechanism. S3, training the distraction driving detection model: training the distraction driving detection model through the training set and the verification set, and testing a training result through the test set to obtain the distraction driving detection model which can be applied; and S4, detecting the driving behavior by adopting the distraction driving detection model which can be applied.

The invention belongs to the technical field of digital signal processing, and particularly discloses a space-time interference cancellation algorithm parallelization FPGA implementation device and method, and the device comprises an input vector generation module which delays and arranges reference signals of multiple channels, and obtains space-time input vectors of the reference signals; the correlation matrix calculation module is used for acquiring an autocorrelation matrix and a cross-correlation vector; the matrix decomposition module is used for performing matrix decomposition on the autocorrelation matrix to obtain an upper triangular matrix; the weight determination module is used for carrying out weight solving by utilizing the upper triangular matrix and the cross-correlation vector to obtain an optimal weight; the delay matching module is used for delaying the reference signal and carrying out time sequence alignment on the reference signal and the optimal weight to obtain a delay matching result; and the weighting and combining module is used for weighting the reference signals subjected to the time sequence alignment according to the optimal weight to obtain weighted signals and combining the weighted signals with the received signals so as to realize interference elimination of the received signals. According to the invention, the real-time tracking capability can be improved and the processing delay can be reduced.

The invention discloses a motor adaptive backstepping controller design method for a high self-locking force load, and the method comprises the steps: building a high self-locking force load torque mathematical model containing a time-varying unknown parameter and a DC motor mathematical model, and defining a virtual control quantity through a backstepping method control theory; designing an adaptive law to realize real-time estimation of time-varying unknown parameters; a self-adaptive backstepping controller is designed through a piecewise projection function and a quadratic Lyapunov function, the system convergence speed and actuator buffeting are optimized, and motor mute control is achieved; the method has the advantages that the time-varying unknown parameters can be estimated in real time, the mute control of the direct current motor is realized, and the hardware cost and the computing power cost are low.

The invention discloses a sensor network-oriented rapid positioning method and device, a terminal and a storage medium, which can improve the positioning speed of a linear positioning method. The method comprises the following steps: on a two-dimensional plane, establishing linear constraint equations corresponding to three measurement quantities of distance, azimuth angle and angle; constructing a first linear positioning algorithm corresponding to the linear constraint equation by adopting a least square method; in order to improve the speed of obtaining the real position of a sensor node by a first linear positioning algorithm, a linear constraint condition is set, and when the sensor node meets the linear constraint condition, a decision variable is introduced into a first positioning matrix of a sensor network to generate a second positioning matrix; constructing a positioning speed model by adopting the minimum characteristic value of the second positioning matrix; solving the optimal solution of the positioning speed model by adopting the designed centralized optimization algorithm or distributed optimization algorithm; and obtaining a target position of each free node in the sensor network based on the optimal solution, the first positioning matrix and a first linear positioning algorithm.

The invention relates to the technical field of active noise control, and provides a binaural cooperative active noise control method based on a combined cost function. The objective of the invention is to solve the problems of asynchronous convergence rates of left and right ears and inconsistent steady-state errors caused by factors such as physiological structure difference of left and right ear canals, inconsistent wearing fitness of earphones and asymmetrical directivity of loudspeakers. According to the method, a combined cost function including an overall error term, an instantaneous error term and an accumulative error term is constructed, the contribution proportion of each error term is automatically adjusted through a dynamic weight adjustment mechanism, and the weight of a control filter is updated based on a gradient descent method, so that the convergence speed and the steady-state error of a left sound channel and a right sound channel tend to be consistent. The consistency and balance of the binaural active noise control performance can be improved, and the method can be suitable for application scenes such as active noise reduction earphones and intelligent cabins.

The application discloses a biological environment text named entity recognition method, medium, equipment and product, relates to the biological environment entity recognition technical field, and the method comprises the following steps: acquiring a biological environment text, converting the biological environment text into training samples and test samples; inputting the training samples into a fine-tuning language model, optimizing the fine-tuning language model by using a proximal policy optimization algorithm, calculating a reward function by using the cosine similarity of the embedding vector of the input sample, calculating an advantage function by using a dynamic discount factor and a dynamic balance factor, and obtaining a trained language model; inputting the test sample into the trained language model, testing the trained language model, and using the tested language model to perform named entity recognition on the biological environment text. The application can optimize the update stability of the model and the adjustment of the model parameters.

This invention provides a method for predicting the dominant frequency of blasting vibration by integrating parameter fluctuation processing and the Dream Optimization Algorithm (DOA), belonging to the field of underground engineering safety control technology. The system includes a data acquisition and uncertainty modeling module, a data preprocessing and feature construction module, a hyperparameter optimization module, and a prediction model training and result output module. The method obtains relevant parameters through on-site investigation and monitoring, and generates an extended sample set using a combination of probabilistic perturbation modeling and fuzzy triangular modeling. The data is preprocessed and features are selected. The DOA algorithm is used to globally search the hyperparameters of the Support Vector Regression (SVR) model, and the optimal hyperparameter combination is selected by combining a robustness fitness function. The SVR model is trained, and prediction results and uncertainty prediction intervals are generated. This invention can explicitly characterize the uncertainty of geological parameters, achieve efficient global optimization of hyperparameters, improve prediction accuracy, robustness, and generalization ability, and is applicable to different geological conditions and blasting scenarios. It can be extended to various blasting dynamic response prediction tasks.

The application discloses a guided wave tomography method based on full waveform inversion of a recurrent neural network, comprising the following steps: step 1, time series forward modeling based on the acoustic formula of the finite difference method is realized by using sequence calculation of the recurrent neural network, and an initial model is obtained; step 2, the wave field value of each time step at each excitation position is calculated in turn, and a simulation value is generated; step 3, after actual observation values are processed, a least square target equation of the simulation value and the observation value is generated; step 4, the least square target equation is iteratively minimized by an inversion part, including an automatic differentiation method and an Adam algorithm, which are respectively responsible for gradient calculation of the target equation and iterative updating of a velocity model and a UNet weight; step 5, the updated velocity model is subjected to depth image prior UNet regularization; and step 6, steps 1 to 5 are repeated, and the next iteration calculation is performed until the target equation is lower than a specified threshold value, so that a final velocity model is obtained, and the final velocity model is converted into a thickness model by using a dispersion relation.

The invention provides an integrated multi-target dynamic economic dispatching method for electric vehicles. Firstly, a multi-target dynamic economic dispatching mathematical model integrated with the electric vehicle is established; secondly, boundary constraint conditions of a multi-target dynamic economic dispatching problem integrated by the electric vehicle are determined; secondly, on the basis of the built mathematical model, a moth flame algorithm is introduced and improved; and finally, scheduling the electric vehicle integration based on the improved chaos moth flame algorithm, performing comparative analysis on other conventional algorithms through software simulation, and verifying the effectiveness and superiority of the chaos moth flame algorithm provided by the invention. Results show that compared with a conventional electric vehicle integration algorithm, the method has better convergence characteristics, can be optimized to obtain smaller optimal fuel cost and worst fuel cost, can effectively reduce the running time of a CPU, and can improve the efficiency of multi-target dynamic economic dispatching of electric vehicle integration. The method has a good level in the aspects of convergence speed and reliability.

The invention discloses a pipeline leakage signal denoising method based on improved WOA-VMD combined with SVD, and relates to the technical field of signal processing, and the method comprises the steps: obtaining a pipeline leakage signal at a pipeline defect position; taking a penalty factor alpha and a modal number K in the VMD algorithm as optimization targets of a whale algorithm, namely a WOA algorithm; setting a VMD algorithm according to the optimal penalty factor alpha and the optimal modal number K obtained by the WOA algorithm, and performing VDM decomposition on the pipeline leakage signal to obtain K IMF components; identifying the K IMF components by using a correlation coefficient method to identify effective IMF components; carrying out SVD decomposition and denoising processing on each effective IMF component by adopting an SVD algorithm, and removing a noise signal in each effective IMF component; and synthesizing each effective IMF component after the de-noising processing to obtain a de-noised pipeline leakage signal. According to the method, the WOA algorithm, the VMD algorithm and the SVD algorithm are combined to form an integrated signal processing framework, and compared with a single method, the performance is remarkably improved.

The present invention discloses a GIS electromagnetic field coarse simulation data enhancement method, system and storage medium based on differential fusion neural network, comprising obtaining a GIS electromagnetic field coarse grid simulation data set; inputting the GIS electromagnetic field coarse grid simulation data set into a differential deep learning network model, the differential deep learning network model comprising an enhancement network and a structural similarity network connected in sequence, the enhancement network comprising a self-adjusting module and a differential convolution module; using the self-adjusting module and the differential convolution module to calculate the coarse grid structure data and the coarse grid field strength data respectively, to obtain fine grid enhancement grid structure characteristics and fine grid enhancement field strength characteristics; using the structural similarity network to match the fine grid enhancement grid structure characteristics and the fine grid enhancement field strength characteristics, and calculating the similarity characteristics; based on the similarity characteristics, calculating the GIS fine grid enhancement data; the present invention can take into account both the accuracy and efficiency of GIS electromagnetic simulation.

The invention discloses a path planning method and system based on an improved self-adaptive ant colony algorithm, and the method comprises the following steps: initializing, loading traveling salesman problem instance data of a plurality of places based on basic parameters and dynamic parameters of the improved self-adaptive ant colony algorithm, constructing a distance matrix between the places based on Euclidean distance, and performing path planning on the distance matrix. Initializing a symmetric pheromone matrix; iterative optimization is carried out, each ant starts from a randomly selected initial site, a complete path is constructed step by step through an improved state transition probability rule, the path length of each ant is calculated, and the current iterative optimal path and the length thereof are recorded; iteration is terminated, and when the number of iterations reaches the preset value, the algorithm stops running; according to the method, statistics is carried out on results of multiple rounds of operation, the average optimal distance of the multiple rounds of operation and the minimum distance in the multiple rounds of operation are obtained, namely, the optimal planning path is obtained, a self-adaptive volatilization mechanism is introduced, the pheromone volatilization rate is dynamically adjusted according to the number of iterations, and the local optimum problem is solved.

The application discloses a pipeline leakage signal denoising method based on improved WOA-VMD combined with SVD, relates to the technical field of signal processing, and acquires a pipeline leakage signal of a pipeline defect position; a penalty factor alpha and a mode number K in a VMD algorithm are used as optimization targets of a whale optimization algorithm (WOA) algorithm; the VMD algorithm is set according to the optimal penalty factor alpha and the optimal mode number K obtained by the WOA algorithm, and VDM decomposition is carried out on the pipeline leakage signal to obtain K IMF components; the K IMF components are identified by using a correlation coefficient method, and effective IMF components are identified; SVD decomposition and denoising processing are respectively carried out on each effective IMF component by using an SVD algorithm, and noise signals in each effective IMF component are removed; and each effective IMF component after the denoising processing is synthesized to obtain a pipeline leakage signal after denoising. The application combines the WOA, VMD and SVD algorithms, forms an integrated signal processing framework, and has a significant improvement in performance compared with a single method.

This invention belongs to the field of sliding mode control technology and relates to a sliding mode control method based on a variable exponent double power-law reaching law. Addressing the slow convergence speed of traditional double power-law reaching laws, this invention establishes a mathematical model of a permanent magnet synchronous motor (PMSM): the voltage equation of the PMSM in the three-phase coordinate system is converted into a voltage equation with the rotor coordinate axis as the reference frame; an adaptive variable gain double power-law reaching law is designed; a sliding mode disturbance observer is designed based on the adaptive variable gain double power-law reaching law to observe sliding mode disturbances; a speed loop controller is designed based on the novel variable gain reaching law and sliding mode disturbances, and the PMSM system is controlled through the speed loop controller. Compared with the traditional double power-law reaching law, the variable gain double power-law reaching law of this invention has a superior convergence speed across the entire domain, and the proposed variable gain control strategy does not cause chattering, thereby improving the stability of control performance.

The present invention provides a high-precision SAR image target detection method and system based on particle filtering, the method comprising: using a context-aware algorithm to perform context-aware operations; feature extraction and target initialization; performing context-aware operations to obtain environmental adaptability information; continuous state estimation and tracking; comprehensive update of state estimation; using a feedback mechanism to perform feedback operations; and adaptive adjustment. The present invention solves the technical problems of poor processing speed, real-time performance, accuracy, robustness, adaptability, and flexibility, as well as high sensitivity to shielding and interference.

The application provides a strategy generation method for combined heat and power based on a generalized Nash equilibrium algorithm, comprising the following steps: constructing a combined heat and power system under related constraints; wherein the upper limit of the thermal comfort of each consumer and the relationship between the consumer utility and energy consumption are considered; the problem to be solved by each company is determined by considering the decision information of all energy companies; wherein the objective function of the combined heat and power system model is determined according to the upper limit and the lower limit of the production efficiency of each company, the production plan of the company which needs to follow the power balance and the thermal balance, and the problem to be solved by each company; when the global decision information of other companies is obtained, the decision of each company is obtained by solving based on the edge-based distributed generalized Nash equilibrium algorithm under the global decision information, including the demand and the price of thermal energy and electric energy in each round of auction.

The invention discloses a vehicle adaptive cruise method based on vehicle speed planning, which comprises the following steps that: a vehicle receives a satellite signal to determine a current position, and obtains the altitude and speed limit data of a front road; an n-order Bezier function is constructed, and a vehicle speed curve of the vehicle speed relative to the road position is generated through the Bezier function; an evaluation function of the vehicle speed curve is constructed, and constraints are established; a wild horse optimization algorithm is improved, and Gaussian distribution is introduced to generate an initial population and an individual position updating strategy based on Levy flight so as to improve the global search capability and the convergence speed; inputting the Bezier function and the evaluation function into an improved wild horse optimization algorithm, and finally outputting an optimal cruise speed curve; based on the optimal cruise speed curve, the conversion relation between fuel energy and electric energy is determined by using an ECMS global optimization strategy and introducing an equivalent factor, a global optimal control strategy can be obtained by minimizing a Hamilton value, then an optimal control method of an engine and a motor at each moment is obtained, and the engine is controlled.

The present application relates to a flowered stereo calibration box and a camera calibration method based on an OCAM algorithm, wherein the flowered stereo calibration box comprises five checkerboards, each checkerboard is a regular row and column distributed black and white checkerboard, the bottom checkerboard is a square, the four side checkerboards are rectangles, the four side checkerboards are connected with one bottom side of the bottom checkerboard respectively, the four side checkerboards are equal to the included angle of the bottom checkerboard, and the four side checkerboards are divergent and inclined upward with the bottom checkerboard as the center; the checkerboard small grid of the bottom checkerboard is provided with a closed black frame, the black frame does not intersect with the checkerboard small grid; the three sides of the four side checkerboards which do not intersect with the bottom side of the bottom checkerboard are provided with black frames, the black frames are connected with the checkerboard small grid of the side checkerboard; and the bottom checkerboard is provided with four circular calibration points. The present application reduces the requirement for image quality, and improves the operation precision and operation efficiency of panoramic camera calibration.

The invention relates to a DDCPN-based time delay precision balance optimization method in vehicle infrastructure cooperative infrastructure perception, which belongs to the field of Internet of Vehicles and comprises the following steps: constructing an Internet of Vehicles system consisting of a plurality of vehicles and a single RSU; establishing an optimization problem to minimize the total delay of all task vehicles completing the perception task within the duration; modeling the optimization problem as a Markov decision process MDP; and task unloading and resource allocation optimization are carried out based on a discrete drive continuous policy network DDCPN algorithm. The algorithm has good convergence efficiency. Simulation results show that the algorithm has a faster convergence rate under the constraint conditions of meeting fast reasoning precision, energy consumption, computing resource utilization rate and the like, effectively reduces task execution delay, and better adapts to a dynamic and complex vehicle-mounted network environment.