72results about How to "Avoid getting stuck in local optima" patented technology

The invention belongs to the technical field of unmanned driving, and discloses a target following and dynamic obstacle avoidance control method for a speed difference slip steering vehicle, and the method comprises the steps: building four neural networks through employing a depth determinacy strategy in reinforcement learning; constructing a cost range of the obstacle so as to determine a single-step reward function of the action; determining continuous action output through an actor-critic strategy, and updating network parameters continuously through gradient transmission; and training a network model for following and obstacle avoidance according to the current state. According to the method, the intelligence of vehicle following and obstacle avoidance is improved, and the method canbetter adapt to an unknown environment and well cope with other emergencies. the complexity of establishing a simulation environment in the reinforcement learning training process is reduced. By utilizing a neural network prediction model trained in advance, the position and posture of each step of the target vehicle and the obstacle can be obtained according to the initial position and posture ofthe target and the obstacle and the action value of each step, so that the simulation accuracy and efficiency are improved.

The invention discloses a topological map based unmanned boat route searching system and method. The method comprises following steps: step one, obtaining the external information of the searching area of an unmanned boat, wherein the external information comprises data of start position, final position, number of barriers, and edge point sets of barriers, based on the external information, carrying out route information analysis, based on the analysis result, establishing navigation topological relationship of the unmanned boat, and based on the topological relationship, establishing a spatial topological map; and step two, based on the established spatial topological map, a Dijskra algorithm, and navigation conditions of the unmanned boat, calculating the optimal path, and after the optical path is selected, controlling the track of the unmanned boat according to the optimal path. The method has the advantages that the optimal path that can meet the navigation requirements can be calculated out, the algorithm complexity is reduced, and the situation that route searching is trapped by local optimum is avoided.

The invention discloses a multi-hop positioning method for lightweight wireless sensor networks. The method comprises the following steps that: 1, all nodes to be positioned acquire positioning reference information per se; 2, the nodes to be positioned establish weight restraining models for multi-hop positioning of the nodes; 3, the nodes to be positioned determine feasible regions of coordinates per se; 4, the nodes to be positioned acquire samples of coordinates per se in a meshing mode; 5, the nodes to be positioned search approximate optimal solution of the coordinates per se from the samples; and 6, the nodes to be positioned refine estimation coordinates per se. In the method, the feasible regions of the coordinates of the nodes to be positioned can be determined by a method of intersections of restraining square loops, so the restraining range of node coordinate estimation is reduced; the global approximate optimal solution of the node coordinates can be acquired by using a lightweight mesh scanning method, so while the calculated amount is reduced, the positioning accuracy and network topology adaptive capability can be improved. The method has practical value and wide application prospect in the technical field of wireless sensor network positioning.

The invention discloses a pressure guide wire temperature compensation method of an improved Particle Swarm Optimization neural network. The method includes the following main steps: collecting pressure guide wire output voltage and parameters related to the environment where a pressure guide wire is, and performing normalization processing on data; building a three-layer front feedback neural network model having an error back propagation capability; utilizing improved Particle Swarm Optimization (PSO) to optimize the weight and threshold value of the built neural network; training the neural network after the weight and threshold value are optimized; and utilizing the neural network model obtained by training to perform temperature compensation on pressure guide wire measured data. The pressure guide wire temperature compensation method of the improved PSO neural network utilizes the improved PSO neural network algorithm to build a pressure guide wire measurement inverse model, the trained model is high in compensation precision, generalization ability and stability, and the defects that a Back Propagation (BP) neural network is easy to fall into local optimum and a standard PSO BP neural network is easy to skip global optimum are overcome.

The invention discloses a hierarchical planning method for a power distribution network containing a distributed power supply, and belongs to the field of power distribution network planning of a power system. The implementation method comprises the following steps of: obtaining a target object; establishing a multi-objective optimization model by taking the annual minimum comprehensive investmentoperation cost of the power distribution network as an objective, converting the multi-objective optimization model into a hierarchical planning model, establishing an objective function by taking the annual minimum comprehensive investment operation cost of a line as an objective in upper layer planning, solving a line decision variable, obtaining an optimal grid structure, and transmitting theoptimal grid structure to a lower layer; On the basis of the upper-layer net rack, the lower-layer planning establishes an objective function with the minimum sum of the annual average investment construction and operation maintenance cost of the distributed power supply DG, the line network loss cost, the power purchase cost from the upper-level network and the environmental pollution treatment cost avoided by accessing the DG; And solving the upper and lower layer models by using a PSCO optimization algorithm to obtain a final grid structure and DG access position and capacity configuration.The method has lower annual comprehensive economic cost and more stable system voltage level, and the power supply reliability can be improved.

The invention discloses an array sparse method for a broadband non-frequency-variable multi-beam imaging sonar. With the Bessel function, fitting of influences on array guiding vectors by different frequency points in the broadband signal bandwidth is performed and a broadband signal multi-beam forming model under the far-field situation is established; on the premise that the formed multiple beams approximate a reference beam, a minimum number of effective array elements are searched and multiple sets of weighting coefficients are calculated; a highly nonlinear sparse array optimization problem is transformed into a sparse signal reconstruction problem in the compressed sensing theory, a reconstruction weighting coefficient is calculated iteratively by an underdetermined system localizedsolution algorithm, and a sparse array structure is determined; a convex optimization theory is introduced so as to form a plurality of low-side-lobe beams and a multi-beam array sparse side-lobe suppression model for array element excitation is established. According to the invention, the main lobes of a plurality of formed beams are not extended with changes of signal operating frequencies; andpeak side-lobe levels of multiple beams formed by the sparse array are reduced effectively.

The invention belongs to the technical field of network security and particularly relates to an APDE-RBF (Affinity Propagation Differential Evolution-Radial Basic Function) neural network based network security situation prediction method. The APDE-RBF neural network based network security situation prediction method comprises the steps of dividing and clustering sample data by utilizing an AP clustering algorithm to obtain the number of nodes of hidden layers of the center and network of the RBF; obtaining population diversity by using AP clustering, changing a zoom factor and a crossover probability of a DE algorithm adaptively and optimizing the width and connection weight of the RBF; meanwhile, performing chaotic search on elite individuals and population diversity center of each generation of population in order to avoid falling into local optimization and jumping out of a local extreme point; and determining a final RBF network model, inputting a test dataset and outputting a situation prediction value. The APDE-RBF neural network based network security situation prediction method aims at improving the prediction precision for the network security situation while enhancing the generalization ability.

The invention discloses a multi-objective quantum Shuffled Frog Leaping Algorithm (SFLA) based water resource optimization and diversion method, comprising the following steps: firstly obtaining basic information data on water resources; establishing multi-objective optimization and diversion models for water resources; executing the shuffled Frog Leaping Algorithm (SFLA) to find the best solutions to Pareto of the multi-objective optimization and diversion of water resources; and finally according to a multi-objective decision theory, choosing the best theory to divert water resources by combining objective weights and subjective weights. According to the invention, optimization is achieved through an overall choosing process. Calculation efficiency is increased so as to meet the requirements for best multi-objective diversion programs in a water resource system.

The invention discloses a service request distribution method oriented to an edge computing environment. The method comprises the steps of (1) monitoring an available resource and a service request execution situation of each edge apparatus in a system in real time; (2) collecting service requests collected by all edge apparatuses in the system; (3) selecting an optimal edge server or a cloud server for all service requests by use of a GASD (Globally Aligned Spatial Distribution) algorithm; and (4) performing service request scheduling on all edge servers. Compared with the prior art, from twoaspects of distribution and scheduling of the service requests, the service response time is optimized to the utmost extent; meanwhile, a heuristic method is adopted, a temperature control mechanismin a simulated annealing method is introduced into a genetic algorithm, the convergence rate of the algorithm can be slowed down at the initial phase of the algorithm, thereby effectively avoiding local optimization; and the convergence rate is accelerated at the termination phase of the algorithm, thereby improving the efficiency of the algorithm.

The invention discloses a cell-oriented amorphous coverage small base station deployment method in a cellular network. The method aims at maximizing multi-user distribution system average throughout capacity. The multi-user distribution system average throughout capacity under different small base station position vectors is calculated by means of a given collaborative cell building and resource scheduling method; a small base station deployment position enabling the throughout capacity to be maximum is found based on a given position updating algorithm, and multi-user distribution is considered to the greatest degree. Compared with a traditional method, the cell-oriented amorphous coverage small base station deployment method is more applicable to an actual site; in consideration of the user distribution tidal phenomenon, when user distribution is changed, the determined small base station position enables adjacent small base stations to change the collaboration way more effectively in real time, so that a traditional fixed cell shape is changed, and the system performance requirement for distribution of different users is met. By the adoption of the cell-oriented amorphous coverage small base station deployment method, the system average throughout capacity, margin user performance and user fairness can be effectively improved.

The invention discloses an electroencephalogram signal classification method of an artificial bee colony optimized BP neural network. The method is specifically implemented according to the followingsteps of firstly, collecting electroencephalogram signals, and preprocessing the obtained electroencephalogram signals; mEMD decomposition being carried out on the preprocessed electroencephalogram signals to obtain more concentrated frequency band signals; screening effective frequency band signals from the obtained frequency band signals according to the maximum mutual information coefficient; reconstructing a component, and performing feature extraction on the reconstructed signal by using fuzzy entropy to form a feature matrix; dividing the data set of the electroencephalogram signals intoa training set and a test set, wherein the training set is used for training a model of the BP neural network; and inputting the obtained feature matrix of the fuzzy entropy into a trained BP neuralnetwork model, and outputting a classification result. The method solves problems that in the prior art, an artificial neural network is low in convergence speed, sensitive in initial weight, prone tofalling into local optimum and poor in global search capacity.

The invention discloses a particle swarm classifying method based on automatic clustering, which mainly solves the problems in the prior that the reference of domain information is limited, and the target function accessing standard is single. The method comprises the following processes: (1) carrying out an automatic clustering method on training data so as to obtain a class mark of the automatic clustering method; (2) carrying out the particle swarm optimal classifying method on the training data so as to obtain the class mark of the classifying method; (3) calculating the fitness value of the particle, and calculating the optimal relationship matrix; (4) replacing the positions of the particles; (5) updating the maximum historical fitness value and the maximum comprehensive historical fitness value of the particle; (6) determining whether the algorithm meets the terminating conditions, if so, stopping iterating, if not, carrying out step (3); (7) determining the class mark of the data based on the particle cluster; and (8) calculating the accuracy of classifying. The particle swarm classifying method based on automatic clustering has the advantages of obvious UCI (Uplink Control Information) data classifying effect, and can be applied to classifying the texture image.

The invention belongs to the technical field of multi moving station passive positioning, and relates to a multi moving station passive time difference positioning method of performing high-speed and high-accuracy positioning on a target. The multi moving station passive time difference positioning method includes the steps: obtaining a real distance value between a target and four base stations by means of a multi moving station time difference positioning module; by means of the four base stations, obtaining three groups of time difference, and according to the conversion relation between the distance difference and the time difference, solving the three groups of time difference; forming a time difference observation matrix by means of three groups of time difference equation set; and performing maximum likelihood estimation on the observation matrix to obtain a likelihood function, and converting the likelihood function into solution of optimization problem so as to derive a fitness function. The multi moving station passive time difference positioning method distribute particles according to grids, so that the particles can be searched globally, and can effectively avoid the particles running into local optimum.

The invention discloses a method for predicting a frequency spectrum of a CRN (Cognitive Radio Network) on the basis of a GCV-RBF neural network. The method comprises the following steps of: S1: acquiring channel historical data information; S2: using the channel historical data information as a preset input sample of an RBF neural network, training the RBF neural network by an OLS algorithm, and acquiring an optimal RBF neural network structure by a GCV evaluation method; and S3: according to the channel historical data information, by the optimal RBF neural network structure, predicting a current frequency spectrum state. Compared to the prior art, according to the invention, the optimal RBF neural network structure is obtained by the GCV evaluation method, so that the problem of overfitting in the training process is solved, and prediction accuracy is improved. Further, the RBF neural network structure, as a local approaching network, has the advantages of simple structure, high convergence rate, high real-time performance and the like, and can be sufficiently adaptive to changes of the network and improve self-adaptation of the network.

The embodiment of the invention discloses an intrusion detection method for optimizing a regularization extreme learning machine through an improved longicorn swarm algorithm, and the method introduces an LU decomposition method to solve an output weight through iteration, reduces the calculation complexity, and improves the intrusion detection accuracy. The improved longicorn swarm algorithm is introduced for RELM neural network parameter optimization so as to improve the detection precision and the training speed of the RELM neural network; according to the improved longicorn swarm algorithm, Tent mapping reverse learning is used to initialize a population, a Levy flight population strategy and a dynamic mutation strategy, so that individuals dynamically learn the experience of the population in the moving process, the convergence rate of the algorithm is improved, the later exploration capability is enhanced, and the algorithm is prevented from falling into local optimum.

An improved multi-target particle swarm optimization-based complicated well track optimization method comprises the steps of (1) setting a parameter of a multi-target particle swarm optimization MOPSO; (2) initializing population; (3) calculating a target function value; (4) updating a position and a speed of each generation of particle; (5) performing mutation operation on the particle; (6) calculating a target function value of each particle in the population; (7) updating the process of individual optimal algorithm from iteration beginning to a current optimal position; (8) sorting a non-domination set nd; (9) sequencing non-inferior solutions in external document of the MOPOS according to a target function value in a descending order; (11) deleting subsequent remaining individuals by an intercept method; (11) performing global optimization; and (12) obtaining an optimal solution set with algorithm optimization, wherein the actual measurement length of the well track and actual control torque reach optimization relatively. By the improved multi-target particle swarm optimization-based complicated well track optimization method, multi-target well track parameter optimization under an actual well drilling condition is achieved, the drilling success rate is improved, and a theoretical decision foundation is laid for reduction of drilling cost.

The invention relates to a green scheduling optimization method of an intelligent manufacturing workshop for complex man-machine coupling, and belongs to the field of workshop operation scheduling optimization. The method is characterized by comprising the following steps: 1, determining the number of employees working at work, the electricity prices in different time periods, the characteristic parameters of available equipment and the like; 2, constructing a multi-objective function of green scheduling optimization; 3, constructing constraint conditions of green scheduling optimization; 4, making a greedy strategy according to the objective function of the problem; 5, solving the model through an improved non-dominated sorting genetic algorithm.

The invention relates to a reactive power optimization of a power system and specifically relates to a power system reactive power optimization method of a wind power field. The method includes random initialization of population, linear annealing weight introduction, gene fusion of genes of individuals in a new population and individual in a original population under a CR weight, target population generation, cross operation implementation, target individual fitness value calculation, one-to-one comparison of target individual fitness values and original individual fitness values, preferential saving, new population generation, and iteration search in the maximal evolution algebra range until the large evolution algebra is reached. According to the invention, dynamic adjustment is performed on parameters of a differential algorithm and a variation strategy of linear annealing is adopted for overlapped individuals in the population, so that a condition that the algorithm falls into local optimum is avoided, optimization and overall search capability are improved, the calculation time is shortened, influence on power grid reactive power distribution and voltage problems by the wind power field are eliminated, system grid loss is reduced and voltage level is improved.

The invention discloses a position auxiliary beam alignment method and system based on multi-arm steal. The method comprises the following steps: acquiring position information of a receiver; selecting a plurality of LOS beams of the LOS path according to the position information to form an LOS subset; selecting a plurality of NLOS beams of the NLOS path with the maximum UCB value from the remaining beams to form an NLOS subset; during beam alignment, selecting a plurality of beams with the maximum path gain according to the channel states of all paths in the LOS subset and the NLOS subset; inthe data transmission period, transmitting data streams in parallel through the selected wave beams, and obtaining UCB reward values of the wave beams; and updating the average return of the beam based on the UCB reward value. According to the invention, the beam alignment probability is increased, the calculation complexity is reduced, and the spectral efficiency is improved.

The invention relates to a multi-target moth algorithm-based small hydropower station optimal scheduling method. The method comprises the following steps of: firstly collecting target small hydropower station, and combining a storage capacity, a water yield, power generation scheduling, water supply and a boundary condition constraint to establish a mathematic model with targets of maximum power generation capacity and minimum ecological water deficit; and secondly taking the established model as a target function and substituting the target function into a multi-target moth algorithm to carry out optimal computation, and after carrying out the optimal computation through the algorithm, finally returning a set with optimal scheduling schemes so that decision makers can finally make a scheduling scheme through referencing the given optimal scheduling scheme set. The method provided by the invention emphasizes on improving the correctness and high efficiency of optimal scheduling of small hydropower stations and solving the problems existing on models and methods in the prior art, and has significance for pushing the development of the optimal scheduling of the small hydropower stations and improving the economic benefit.

The invention provides a method and a device for determining a path, relates to the field of robot path planning, and aims to increase the convergence rate of an ant colony algorithm and prevent the algorithm from falling into local optimum, thereby improving the speed and efficiency of routing optimization of an inspection robot. The method comprises the following steps: determining at least m paths between a starting position and an end position through the ant colony algorithm; updating pheromones on the effective path according to a first pheromone updating strategy; comparing the existingglobal optimal path with the local optimal path, and determining a current global optimal path; updating pheromones on the current global optimal path according to a second pheromone updating strategy; adding one to the cycle index c, and if the cycle index c does not reach n, repeating all the steps; and if the cycle index c reaches n, determining a current global optimal path as a target path of the inspection robot, wherein the target path is used for indicating the inspection robot to arrive at an end point position along the target path from the starting position.

The invention provides a proxy optimization calibration method for time-varying parameters of a large-scale hydrological model, and the method comprises the steps: selecting a distributed hydrological model, building a model with a drainage basin as a research object, and screening out sensitive parameters; in combination with the determined sensitive parameter values, verifying a model according to actual measurement data of the key indexes related to the sensitive parameter values, and performing numerical simulation of a long-time sequence after verification; counting key index analog quantities under different time scales and corresponding annual actual measurements based on the simulation results; dividing the long-time sequence into a plurality of segments; setting correction factors of empirical values of the sensitive parameters and establishing a proxy optimization calibration model; selecting a proper evaluation index to calibrate the correction factors to obtain posterior distribution of the correction factors in different seasons; and correcting the empirical values of the sensitive parameters by using correction factors to obtain variable parameter values under the seasonal scale, and comparing the variable parameter values with simulation precision. According to the method, rapid calibration of the large-scale model parameters is realized, and the problems of time consumption and low efficiency of large-scale model parameter calibration are solved.