33 results about "Gaussian mutation" patented technology

The invention discloses a method for splitting images based on clustering of immunity sparse spectrums, mainly solving the problem that a spectrum clustering method has poor stability and high complexity. The method comprises the following steps: (1) extracting features of images to be split; (2) normalizing feature data to eliminate the magnitude effect among data; (3) carrying out real coding onthe normalized feature data; (4) randomly generating initial species groups of the coded data and calculating the affinity; (5) cloning based on the affinity size of an antibody; (6) carrying out Gaussian mutation on the cloned antibody species groups and selecting the antibody of the highest affinity as an input for the next round; (7) iterating the set maximum iterations to obtain a final selected sample subset; (8) carrying out greed spectrum dimensionality reduction on the selected sample subset, clustering the dimensionally reduced data and outputting the final image splitting result. Compared with the prior art, the method has the advantages of no need of priori knowledge, high accuracy and low calculation complexity. The method can be used for object detection and object identification.

The invention discloses a plate-fin heat exchanger core structure optimization method based on dynamic pixel granularity. The method comprises the following steps: establishing a plate-fin heat exchanger core structure optimization design model according to a plate-fin heat exchanger core runner structure, providing dynamically-updated pixel granularity, enlarging the population search range and keeping the population diversity; and providing a pixel distance calculation model of non-head and tail particles and head and tail particles, calculating the cross and mutation operation probabilities in a self-adaption manner according to the pixel distances of the particles, and respectively adopting random cross and Gaussian mutation, so as to enhance the global population search capability, improve the local population search efficiency, prevent the algorithm from getting into local optimum and realize the purposes of wide coverage and uniform distribution of Pareto optimal solutions. According to the method, the heat exchanger core structure design efficiency can be improved, and relatively reasonable design parameters are provided. The plate-fin heat exchanger optimally designed by the method has the obvious characteristics of uniform passage load, small secondary heat transfer temperature difference, small flow resistance and high heat exchange efficiency.

The invention discloses a production scheduling method supporting peak-shift power demand response. The main steps are as follows: (1) establishing a production scheduling mathematical model supporting peak-shift power demand response; (2) initializing parameters and randomly generating N initial fireworks populations; (3) calculating a fitness value of each firework in the population; (4) calculating explosive spark of each firework; (5) selecting L fireworks from the N firework to generate Gaussian mutation sparks; (6) providing a local search strategy to replace worst explosive spark and the Gaussian mutation spark of each firework; (7) calculating a global optimum firework position Xb and an optimum fitness value fb; (8) determining whether an end condition is satisfied, if not, selecting N new fireworks according to a selection strategy, and repeating the step (4); if so, outputting an optimal scheduling scheme. The production scheduling method provided by the invention has good stability and effectiveness, and can effectively support implementation of power demand response projects.

A multi-target optimization method for a main bearing of an RV speed reducer comprises the steps of analyzing an external loading condition of the main bearing, building a quasi-static model of the main bearing of the RV speed reducer, and outputting target functions to be angular rigidity, a friction moment and an axial rated dynamic load of the main bearing of the RV speed reducer; randomly generating an initial population, calculating three optimization target function values of each individual by the model, performing non-dominated sorting, calculating a congestion distance, performing selection, heuristic crossover and Gaussian mutation to generate a new population based on a binary system tournament, and calculating a target function of each individual; and combining a parent population and the new population to form a big population, extracting the best individual to be used as a population entering next iteration, removing a repeated individual of the population after combination, detecting whether a current algebra reaches a set algebra or not, and outputting an optimization result. By the multi-target optimization method, a group of pareto-optimal design parameters which can be reference data for a designer can be acquired, whether the current design is a theoretical non-dominated solution can also be determined, and meanwhile, the algorithm efficiency is improved.

The invention discloses an image cutting method based on a multi-target intelligent body evolution clustering algorithm. The problems that the image cutting technology is prone to local optimum and an algorithm is not high in robustness are mainly solved. The image cutting problem is converted into a global optimization clustering problem. The process includes the steps of extracting gray information of pixel points of an image to be cut, initiating parameters and establishing an image intelligent body network, calculating the energy of an image intelligent body, conducting non-domination sequencing, conducting neighborhood competition operation, conducting Gaussian mutation operation, calculating the energy of the image intelligent body, conducting non-domination sequencing, conducting self-learning operation, selecting the optimal clustering result according to the crowding distance, outputting a clustering label, and achieving image cutting. Multiple targeting is achieved for the image processing process, the convergence effect is good, the robustness of the method is enhanced, the image cutting quality can be improved, the cutting effect stability can be enhanced, and the extraction, recognition and other subsequent processing of the image targets are facilitated.

The invention provides a prediction model method based on an improved moth optimization algorithm, which comprises the steps of loading a data set and performing standardized processing on sample data; Gaussian mutation strategy and chaotic agitation are used to improve the moth flame optimization algorithm, and support vector machine model and/or limit learning machine model are constructed by using the improved moth optimization algorithm. The implementation of the invention can not only increase the diversity of the population and enhance the searching ability of the algorithm, but also prevent the algorithm from falling into the local optimum and quickly finding the global optimum solution.

The invention discloses an ore rock strength soft sensing method based on adaptive artificial bee colony optimization. The method of the invention adopts a three-layer perception neural network as a soft sensing model for ore rock strength, and utilizes the adaptive artificial bee colony algorithm to optimize a connection weight and an offset value of the designed neural network. In the adaptive artificial bee colony algorithm, search scaling factors are generated adaptively according to the feedback information of adaptive values, and a Gaussian mutation strategy based on neighbor optimal individuals and global optimal individuals is designed to adaptively generate new individuals. The method of the invention can raise the precision of soft sensing of the strength of ore rocks, and raises the efficiency of measuring the strength of the ore rocks.

The invention discloses an image enhancement method using Gaussian inverse harmony search. In the invention, a Gaussian inverse harmony search algorithm is used to optimize [alpha] and [belta] parameters of an incomplete Beta function, and the optimized incomplete Beta function is used to non-linearly transform an image to enhance the image quality. In the Gaussian inverse harmony search, mean information of a harmony library is fused into a Gaussian mutation operator and the reverse learning operation is performed with a certain probability to speed up the convergence of the algorithm to improve the effect of image enhancement. The image enhancement method of the invention can speed up the image enhancement to a certain extent and improve the effect of image enhancement.

The invention discloses a power supply line fault identification method based on quantum variation multivariate universe optimization, and relates to the field of power supply line fault identification. A quantum particle swarm optimization algorithm is introduced into a traditional multivariate cosmic optimization algorithm, and the average optimal position of particles is introduced into the algorithm, so that the algorithm has better convergence precision and speed; a Cauchy-Gaussian mutation strategy is adopted to solve the problem of local optimal stagnation caused by rapid assimilation of individuals in the later stage of iteration of a traditional multivariate universe optimization algorithm; the short-circuit fault is effectively identified, so that good fault data information is provided for maintenance personnel, the short-circuit fault of a line is removed in time, accident expansion is avoided, and stable operation of a power system is ensured.

The invention discloses a feature extraction optimizing method based on image figure face expression recognition. The method includes the following steps: inputting a face image and representing the face image in the form of a matrix, and obtaining projection characteristic vectors after axes of projection are arranged; setting a projection characteristic covariance matrix of a given image, and using the track of the matrix to represent an optimal projection direction function; updating the optimal projection direction function on the basis of an overall distribution matrix of a training sample image, wherein a vector set formed by the axes of projection satisfying the maximum value of the function updates the projection characteristic vectors and forms a matrix representing expression characteristics; endowing each element in the characteristic matrix with a weight, and optimizing the optimal projection direction (that is to say, optimizing a global optimal solution) through a particle swarm algorithm modified by Gaussian mutation; and dividing the global optimal solution in the particle swarm algorithm into secondary groups including leaders and followers, and carrying out secondary optimization and recognition on multiple main parts representing image figure face expressions. According to the invention, expression characteristic results of a figure face can be effectively distinguished and optimized.

The invention discloses an order insertion scheduling method based on a genetic algorithm and a fireworks algorithm. The order insertion scheduling method is characterized by comprising the steps of setting initial parameters; selecting an order insertion scheduling mode; obtaining an original order and an order insertion order, generating an initial population, and starting iteration by taking the initial population as a current population; performing population crossover and mutation operation on the current population; calculating the firework scale to obtain a firework group; performing firework explosion and Gaussian mutation operation on the firework group; selecting fireworks; carrying out genetic selection; judging whether a set total number of iterations is reached or not after one-time iteration, and if yes, outputting a hybrid firework selection and genetic selection population; otherwise, continuing iteration until the set total number of iterations is reached; the method has the advantages that the production orders with the inserted orders are scheduled through the method, the production scheduling efficiency is improved, the time loss is reduced, and the precision of the solving result is improved.

The invention provides a firework algorithm-based clustering method and device, and the method comprises the steps: carrying out the initialization processing of a to-be-clustered sample point set, obtaining a first-generation firework group comprising a preset number of fireworks, and obtaining a next-generation firework group according to the Gaussian mutation operator and cost function of eachsample point and the fitness of each firework, judging whether the next-generation firework group meets a preset termination condition or not, if yes, determining K clustering centers included in fireworks with the highest fitness in the next-generation firework group as target clustering centers, and obtaining a clustering result according to the distance between each sample point in the to-be-clustered sample point set and each target clustering center. According to the embodiment of the invention, the next-generation firework group is obtained based on the fitness and the Gaussian mutationoperator, so that the clustering center can be obtained globally, the problem that clustering falls into local optimum is avoided, and the clustering accuracy is improved.

The invention relates to a method for accurately extracting an unknown phase shift amount based on an optimized quantum particle swarm algorithm. The method comprises the following steps: firstly, approximately calculating a Fresnel diffraction field statistical average value of each phase shift amount in four-step phase shift based on statistical properties of a diffraction field; secondly, solving an intermediate value according to a phase random condition, and representing an approximate phase shift quantity by using the intermediate value; and finally, setting an average error function asa fitness function, and improving a quantum particle swarm algorithm to further approach the true value of the phase shift amount, Gaussian mutation operation is added to the quantum particle swarm algorithm to enable the algorithm to accord with the initial distribution law of particles. The method is reasonable in conception and realizes theoretical improvement, according to the method, approximate pre-estimation of the unknown phase shift amount in four-step phase shift can be achieved, the accuracy of the calculation direction is improved, the accuracy and speed of calculating the true value of the unknown phase shift amount can be improved, the sawtooth solution phenomenon near the true value in a traditional iteration method is effectively overcome, and the method has the advantagesof being extremely high in speed, excellent in solving quality and the like.

The invention discloses a lithium ion battery state-of-health prediction method based on ISSA coupled DELM. The method employs a DELM network prediction battery state-of-health module and an ISSA optimization DELM network parameter module to realize battery state-of-health prediction, wherein a DELM network comprises two ELM-AE structures. According to the lithium ion battery state-of-health prediction method, 30% of excellent sparrows are used as elite sparrows, and a search space of an SSA algorithm is further expanded by solving reverse solutions of the sparrows; a Cauchy-Gaussian mutation operator is adopted to relocate the position of the optimal sparrow, so that the whole population moves to the vicinity of an optimal solution as far as possible, and the algorithm is prevented from falling into local optimum; the optimal hidden layer weight and bias of the DELM network are solved based on the improved SSA algorithm, and the prediction precision of the DELM network is further improved; and an ISSA-DELM lithium ion battery SOH estimation model is high in prediction precision and can be used for accurately predicting the state of health of a lithium ion battery under the condition of random discharge.

The invention discloses a data fusion method for a gas sensor array. A traditional atom search algorithm is improved by adopting a dynamic self-adaptive position updating coefficient and Gaussian mutation operation; seven standard test functions are used to carry out optimization performance test on the improved atom search algorithm; finally, the minimum value of the test function can be found. In addition, in all the seven test functions, the sum of the precision and the convergence rate of the algorithm is superior to those of FOA, PSO and GA algorithms, the algorithm is applied to a generalized average algorithm of gas sensor array data fusion, and the result shows that a satisfactory gas sensor array data fusion result can be successfully obtained by applying the method.