107results about How to "Improve global convergence" patented technology

The invention discloses a photovoltaic power station power generation amount predication method based on an adaptive mutation particle swarm and a BP network. The method comprises the steps of acquiring and preprocessing a data sample, then constructing a BP neural network structure, optimizing the network by means of an adaptive mutation particle swarm algorithm, and introducing an optimal individual which is obtained after optimization into the BP neural network for predication. According to the method of the invention, the adaptive mutation particle swarm algorithm and the BP neural network are combined for performing real-time power generation amount predication on the photovoltaic power generation system; namely power generation state at next time period is predicated through power generation amount information of a former time period. According to the method of the invention, a relatively high global convergence capability of the adaptive mutation particle swarm algorithm is used for optimizing the initial weight and threshold of the network, thereby improving convergence speed and precision in power generation amount predication.

Embodiments of the invention disclose a multi-target-based improved gray wolf optimization algorithm which is used to solve the technical problems that a standard gray wolf algorithm falls into a local optimal value easily and has a low convergence speed and other defects while processing a multi-target optimization problem in the prior art. The method of the embodiments comprises the following steps: S1, setting a wolf pack initialization parameter and a direction correction probability, and randomly initializing wolves' individual positions; S2, calculating an adaptability value of each wolf individual according to a solving target, and selecting the three wolf individuals ranking top; S3, optimizing the wolves' individual positions of a wolf pack, generating moderate wolves, and updating a wolf pack position; S4, executing direction correction operation on the updated wolf pack, controlling the upgraded wolf pack to participate correction of the size of dimensions according to the direction correction probability, and obtaining a corrected wolf pack position; and S5, determining whether an iteration frequency reaches a preset maximum iteration frequency, outputting the corrected wolf pack position as a final optimization result if the iteration frequency reaches the preset maximum iteration frequency, and, if the iteration frequency does not reaches the preset maximum iteration frequency, turning to the S3 so as to continue performing iteration searching.

The invention provides an unmanned aerial vehicle static and dynamic path planning method based on a digital map, and relates to the technical field of unmanned aerial vehicle path planning. The method comprises the steps of: constructing an unmanned aerial vehicle flight three-dimensional digital map; setting initial and end position coordinates of flight of an unmanned aerial vehicle, and initializing an unmanned aerial vehicle path planning constraint condition; according to a hybrid particle swarm simulated annealing algorithm, updating a speed and positions of particles; drawing an unmanned aerial vehicle global static optimal flight path; judging whether the unmanned aerial vehicle encounters an emergent threat, and if yes, according to the hybrid particle swarm simulated annealing algorithm, re-planning a static flight path; and drawing an unmanned aerial vehicle global dynamic optimal flight path. By the method, in a static case, an accurate global flight track can be rapidly planned; and in a case of the emergent threat in the flight process, the threat can be rapidly avoided to effectively carry out path re-planning so as to guarantee flight safety of the unmanned aerialvehicle.

A micro-grid grid-connected optimal scheduling method based on improved subgradient particle swarms comprises the steps of establishing a micro-grid grid-connected model including an energy storage device; according to the actual situation, establishing an optimal scheduling objective function at the smallest total micro-gird power generation costs and environmental pollution control costs; establishing operating constraints in a micro-grid system, and separately establishing system power balance constraints, storage battery charge and discharge power constraints, micro-power output power limits and electricity purchasing and selling constraints for interaction between the micro-grid and large grids; improving the standard particle swarm optimization; separately improving the inertia weight and acceleration factors; and proposing to use the sub-gradient optimization method to update the velocity of the particles in the particle swarm optimization. According to the micro-grid grid-connected optimal scheduling method based on the improved subgradient particle swarms, while the micro-grid grid-connected optimization involving the energy-storage device is solved, advantages such as a good optimization searching effect and a fast convergence speed are realized.

A survey station optimization deployment problem is one of important problems in use of a spatial measurement positioning system. The present invention provides a survey station layout intelligent optimization method of a spatial measurement positioning system, so that an optimized survey station layout can fully cover a tested area under certain costs and a requirement of measurement accuracy can be met. According to the present invention, a proper positioning error model is established from three aspects of constraint analysis, optimization objective, and optimization means, a multi-goal optimizing function is defined, and survey station layout optimization of the spatial measurement positioning system is implemented by using a practical intelligent optimization algorithm. According to the survey station layout intelligent optimization method of a spatial measurement positioning system, a survey station network optimization deployment problem of the spatial measurement positioning system in an engineering application is effectively resolved. As a quantity of survey stations increases, the method has good expansibility, provides a new method for a multi-station networking measurement layout optimization problem based on an angle intersection principle, and has important theoretical value and practical significance.

The invention discloses a fault diagnosis method for a PWM inverter of a motor drive system. The fault diagnosis method is characterized in that an inverter fault diagnosis model based on both wavelet packet decomposition and an RBF neural network is designed, wavelet packet transformation is utilized to extract a feature vector of a fault signal of the inverter, and the feature vector is taken as the input quantity of the RBF neutral network; a wolf pack-simulated annealing algorithm is adopted to optimize structural parameters of the RBF neutral network; and 22 groups of learning samples and 6 groups of test samples are utilized to train and examine the RBF neutral network. Simulation experiment analysis shows that when the fault diagnosis method is used for an open-circuit fault of the PWM inverter of a three-phase motor drive system, the fault position of an inverter TGBT power tube can be accurately positioned, fault diagnosis is quick, accurate and efficient, and the fault diagnosis method contributes to improving the running reliability of the motor drive system.

The invention provides a simulated annealing particle swarm based air-conditioning energy consumption model parameter identification method. The method includes: taking an energy consumption model function, created via a mechanism analysis method, of an entire air-conditioning system as an objective function, optimizing the objective function with a simulated annealing and particle swarm combined algorithm, and determining an optimal parameter combination under certain load. The concept of simulated annealing is introduced into the particle swarm algorithm, and parameters can be optimized and controlled in the process of particle evolution to be changed adaptively according to the algorithm, so that on the basis of existing advantages of the algorithm, global convergence of the algorithm is enhanced, convergence precision of the algorithm is improved, and the shortcoming that the convergence process stops easily and is trapped in local optimization easily is overcome.

The invention relates to a seismic data processing method, particularly relates to a full waveform inversion velocity field modeling method, and provides a reflected wave waveform inversion method based on a phase correction objective function. A new objective function is obtained through changing the traditional phase calculation method, and a corresponding adjoint seismic source is solved by using an adjoint state method; the objective function is optimized by an optimization method to obtain an effective background velocity field; and finally, in combination with the traditional waveform inversion, a high-precision velocity field modeling result is obtained. In view of the problem of instability of the traditional phase inversion method, the traditional phase wrapping problem is avoided, and through separation between phase and amplitude information, the corresponding objective function has higher global convergence; and middle and low wave number components in the deep layer of themodel can be effectively updated, an accurate initial model is provided for full waveform inversion, and dependence on the initial model by full waveform inversion can be effectively reduced.

The invention discloses a novel multi-objective optimization method for an electric-hydraulic composite power steering system. A booster part of the electric-hydraulic composite power steering system consists of an electric booster steering module and a hydraulic booster module, and controlling the of the two booster modules is realized through a coupler, so that the problems that the booster characteristic adjustability of a hydraulic booster system and an electrical-controlled hydraulic booster system, which are used by a conventional large-scaled bus is poor, and when the large-scale bus travels at a middle speed or at a high speed, the steering road feel of the conventional large-scaled bus is poor are solved. Besides, through multi-objective optimizing of the electric-hydraulic composite power steering system, the steering road feel and the steering energy consumption are used as targets, the steering sensitivity is used as a constraint condition, and based on an improved multi-island genetic algorithm which is amended by simulated annealing, so that the selected parameters of the electric-hydraulic composite power steering system are optimally designed, and compared with the optimization result of the multi-island genetic algorithm, the optimized electric-hydraulic composite power steering system can obtain better global convergence, higher convergence rate, better steering road feel and better steering economy.

The invention discloses a fuzzy control system optimization method based on a differential evolution-local unimodal sampling algorithm. The fuzzy control system optimization method comprises the following steps that (1) a fitness function is determined; (2) a differential evolution algorithm is used, the appropriate number of times of iterations is set to repeatedly perform variation, intersection and selection operation until meeting the convergence precision or reaching the maximum number of times of iterations, and corresponding parameter combinations are obtained; and (3) a local unimodal sampling algorithm is used, the parameter combinations finally obtained through the differential evolution algorithm act as initial values, search range reducing and target vector transferring operation is repeatedly performed according to the fitness value until the end of iteration, and the parameter combination of the best fitness value acts as the optimal result. The beneficial effects of the fuzzy control system optimization method are that the two algorithms are combined, the number of times of iteration of the differential evolution algorithm is reduced, and switching to the local unimodal sampling algorithm is performed after obtaining the preliminary optimal solution so that the advantages of the two algorithms can be fully exerted, the defects of the two algorithms can be compensated and the optimization computation efficiency and the global convergence can be greatly enhanced.

The invention provides a building energy consumption prediction method based on an artificial bee colony algorithm and a neural network. The method comprises the steps that firstly, the artificial bee colony algorithm is utilized for conducting weight value optimization on the neural network; secondly, the optimized neural network is utilized for predicting building energy consumption. The artificial bee colony algorithm is an optimizing algorithm simulating a bee colony and has the advantages that control parameters are fewer, implementation is easy, and calculation is convenient; compared with a particle swarm algorithm, a genetic algorithm and other intelligent computing methods, the artificial bee colony algorithm has the prominent advantages that in each iterative process, global search and local search are both performed, the probability of finding an optimal solution is greatly increased, local optimum is avoided to a great extent, and global convergence is enhanced. Thus, when the artificial bee colony algorithm is adopted to optimize the initial weight value of the neutral network, the accuracy of the neutral network predicting the building energy consumption is improved, and meanwhile the defects existing in weight value optimization of the neutral network at present can be overcome obviously.

The invention discloses an NSGA-II-based automatic storage goods allocation optimization method, and the method comprises the steps: carrying out the goods classification through an ABC classificationanalysis method in combination with the goods information; determining an optimization target and a constraint condition of the automatic storage goods allocation method; establishing a constrained multi-objective optimization problem mathematical model; and solving the model by adopting an NSGA-II-based optimization algorithm to obtain an optimal Pareto solution set, and distributing optimization weights according to the actual specific situation of the automatic storage space to obtain a unique non-dominated solution as an optimal solution. The method can be suitable for small and medium-sized automatic storage such as intelligent express cabinets, intelligent vending machines and self-service storing and taking cabinets; the automatic storage space utilization rate and the goods storing and taking execution efficiency can be effectively improved, the working intensity of workers is relieved, the labor cost and the device maintenance cost are greatly reduced, the safety and reliability of automatic storage are improved, and the good practical value and the wide application value are achieved.

The invention discloses a satellite constellation system multidisciplinary design optimization method based on an agent model, and belongs to the field of spacecraft constellation application. The method comprises the steps of on the basis of the Walker-delta constellation configuration and a small-sized earth observation satellite structure, establishing a key subject analysis model of the satellite constellation system by comprehensively considering the design requirements of the constellation configuration and a satellite subsystem; adopting a sequence radial basis function based on a support vector machine and a discrete-continuous variable sampling method and taking the satellite constellation system mass as an objective function to optimize a preselected design variable; mapping thediscrete variables in a continuous space through a discrete-continuous variable sampling method; replacing an original analysis model with an RBF agent model; identifying an interest interval by adopting an SVM (support vector machine), and performing sequence sampling in the interval to update and manage an RBF agent model, so that a design scheme which meets the requirements of constellation configuration and a satellite subsystem and has the minimum total mass of the satellite constellation system is efficiently obtained, the calculation cost of the satellite constellation system is furtherreduced, and the optimization efficiency is improved.

The invention relates to an optimization method in the field of electric systems, in particular to an optimization method of extra-high voltage transient stability control. The optimization method of the extra-high voltage transient stability control comprises the following steps: (1) setting up a power grid coordination control optimization model, (2) converting a constrained optimization problem of the power grid coordination control optimization model to an unconstrained optimization problem, (3) solving the unconstrained optimization problem by adopting an optimization method of a particle swarm with wavelet mutation, and (4) confirming a globally optimal solution in the particle swarm. The optimization method of the extra-high voltage transient stability control facilitates that electric power system planners and analysts grasp transient stability emergency control, and provides the basis for confirmation of emergency safety control measures of ultra-high voltage grids. The optimization method of the extra-high voltage transient stability control has the advantages of increasing wavelet mutation operation, improving global convergence characteristics, being capable of processing optimization problems of thousands of nodes, conveniently combining parallel computing techniques, and further improving computation speed by using large-scale computers.

The invention discloses a method for optimizing a PI parameter of a high-voltage direct-current transmission constant-current controller. The method includes the following steps of firstly, using a high-voltage direct-current transmission quasi-steady-state model as a calculation model for parameter optimization; secondly, calculating a target function; thirdly, initiating an algorithm parameter; fourthly, calculating the feasible region of the PI parameter; fifthly, optimizing a genetic algorithm; sixthly, embedding a simplex method to judge the condition; seventhly, optimizing a mixed algorithm. The method relies mainly on the genetic algorithm while the simplex method subsidiary, the simplex method is embedded after optimization of the genetic algorithm tends to stop, and the advantages of local optimization of the simplex method are utilized while the individual variation and the global optimization capacity of population of a traditional genetic algorithm are kept. The simplified mode, namely, the quasi-steady-state model which is widely adopted during direct-current power transmission engineering calculation, is adopted, and therefore the calculation speed is high, the calculation accuracy meets the engineering requirements, the method is suitable for optimizing the parameter of the direct-current transmission system controller, and the degree of combination between the method and practical engineering is high.

The invention discloses a KP model density function identification method based on the self-adaption bat search algorithm and belongs to the technical field of identification. The KP model density function identification method based on the self-adaption bat search algorithm aims to introduce the self-adaption concept into the bat search algorithm to enable a density function selected by the bat search algorithm during flight searching to change in a self-adaption mode according to the algorithm, so that global convergence is improved. According to the method, an objective function is set first, the position of a bat individual corresponding to a parameter to be identified is generated randomly by the self-adaption bat search algorithm, updating is conducted by means of a position and speed updating formula to generate a new bat position, whether a random number is met is judged, the fitness function value of each bat individual is calculated, a group is ranked according to fitness values, and whether the maximum iteration number is met is judged. By the adoption of the method, defects of the prior art are overcome, and the optimal KP model density function combination is identified by means of the self-adaption bat search algorithm.

The invention provides a bee colony optimization based network traffic scheduling method under multiple QoS (quality of service) constraints. According to the method, a multi-objective optimization problem is solved with a multi-objective artificial bee colony optimization algorithm, a fitness function is improved in combination of the algorithm with a Pareto sorting mechanism and a crowding distance, solution selection is performed with a Boltzmann strategy, a found Pareto solution is recorded with an external file, and neighborhood search of the colony is guided according to global information, so that the found Pareto optimal solutions are uniformly distributed at the real Pareto optimal front end. The degree of importance of each objective is analyzed according to actual conditions, an optimal traffic scheduling scheme is determined, so that after traffic scheduling, the network traffic is scheduled as required, the service level for users is improved, the utilization rate of network resources is increased, the load balancing purpose is achieved, and the traffic scheduling effect is optimal. With the application of the method, the high-utilization and low-consumption traffic scheduling of the network traffic under the multiple QoS constraints can be realized.