641 results about "Differential evolution algorithm" patented technology

The invention discloses a quick feedback analyzing system in a tunnel constructing process. The system adopts a scheme: understanding currently adopted designing construction parameters; establishing a tunnel excavation three-dimensional finite element numerical grid calculation model; acquiring surrounding rock layering and convergent displacement monitoring information after a tunnel is excavated; establishing a non-linear support vector machine model; fixing an anchoring parameter according to the actual construction parameter, and optimally identifying rock mechanic parameters by adoptinga differential optimization algorithm; optimizing the construction parameter of an anchoring scheme by adopting a differential evolution algorithm; and optimizing the rock mechanic parameters by calling the differential evolution and optimization algorithms to further solve the construction parameter of the anchoring scheme, and outputting the construction parameter of the optimized anchoring scheme as a construction scheme through a computer display screen to guide the constructors to construct. The quick feedback analyzing system ensures that the monitoring information is used for optimizing the anchoring parameter while being used for identifying the surrounding rock parameters, so that the dynamic information construction is improved to a level of quantitative analysis.

The invention discloses a radial basis function (RBF) neural network parameter self-optimizing-based multi-step prediction method for water quality. The method comprises the following steps of: first storing the data of each monitoring station into a database of a local server by using the remote transmission of an online water quality monitoring instrument; then performing normalization processing on a water quality sample sequence, calculating an autocorrelation coefficient to determine an input variable of an RBF neural network, and converting sample data into a standard dynamic sequence data format trained and predicted by the RBF neutral network; next searching for and determining an optimal value of a spreading coefficient spread of the RBF neural network by utilizing a differential evolution algorithm and taking a relative standard error as a target function to obtain an optimal prediction model; and finally sampling water quality data in real time, performing multi-step prediction by using the obtained optimal prediction model and adopting a single-point iteration method, and evaluating a water quality prediction result to realize an early warning function. The water quality can be intelligently warned.

The invention provides a dynamics performance parameter optimizing method of a high-speed train, relates to the field of parameter design optimizing based on the dynamics simulation analysis of the high-speed train, and aims at effectively replacing a dynamics simulation model of the high-speed train by a comprehensive target neural network agent model and combining the design analysis and the multi-target optimization algorithm of the high-speed train in the multi-disciplinary field to analyze and optimize the dynamics simulation approximation model of the high-speed train. The method specifically comprises the steps of building a multi-rigidity dynamics simulation model for the high-speed train; determining related important input/ output design spaces; selecting sampling strategy to obtain a design space sample set suitable for the dynamics performance analysis of the high-speed train; improving the generalization accuracy of the comprehensive target neural network by the bayesian regularization method; adjusting the number of nodes in a hidden layer to build the comprehensive target neural network agent network model of which the error is controlled to be within a certain of range; performing multi-target optimization through the intelligent differential evolution algorithm by using the improved comprehensive target neural network agent network model to obtain the optimized high-speed train design parameters. The method is mainly applied to the dynamics analysis and design optimization of the high-speed train.

The invention discloses an ORP (optimal reactive power) method of a distribution network of an electric power system in the technical field of ORP of the distribution network of the electric power system. The method comprises the following main steps: introducing an accelerated evolution operation and an investigation operation in an ABC (artificial bee colony) algorithm to a basic differential evolution operation; and judging whether conditions of convergence of a hybrid algorithm are met, and ending the optimization and outputting the optimal result if the conditions of convergence are met.The hybrid algorithm for solving the ORP problem exerts the advantages that the operation is simple, robustness is good and the like, of the differential evolution algorithm, and can be used to shorten the running time of the algorithm and improve the probability of finding out the global optimal value.

The invention relates to a multiple-unmanned-aerial-vehicle cooperated multi-target distribution method which comprises the steps of firstly numbering a plurality of unmanned aerial vehicles and a plurality of targets, wherein U represents the number of the unmanned aerial vehicles and T represents the number of the targets; then constructing a flight cost model according to magnitudes of U and T and an unmanned aerial vehicle flight cost parameter, wherein the unmanned aerial vehicle flight cost parameter comprises a flight range cost, an execution time cost and a damage cost; and finally performing optimized solving on the flight cost model by means of a heuristic genetic algorithm until a requirement for an optimized target is satisfied. According to the multiple-unmanned-aerial-vehicle cooperated multi-target distribution method, through adding the unmanned aerial vehicle damage cost in the flight cost, three most basic conditions are extracted for modeling for aiming at a relationship between the unmanned aerial vehicles and the number of targets so that a model approaches to reality; and in target distribution, the heuristic genetic algorithm is utilized. Through introducing heuristic information, algorithm execution efficiency is effectively increased and a precocity problem of the genetic algorithm is prevented. The multiple-unmanned-aerial-vehicle cooperated multi-target distribution method is better than a basic genetic algorithm and a differential evolution algorithm at aspects of convergence speed and convergence value.

Provided is a method for improving mobile sensor network coverage rate. A mobile sensor network that the method is based on comprises a plurality of heterogeneous mobile sensor nodes, wherein each mobile sensor node has a specific perception radius and a communication radius and can obtain position of itself. The method comprises: meshing the continuous mobile sensor network monitor area and distributing each mobile sensor node to the meshed monitor area randomly so that initial position of the each mobile sensor node is obtained; calculating coverage rate of the mobile sensor network; determining strategy of improving mobile sensor network coverage rate by employing coverage hole-directed distributed differential evolution algorithm (CHDDE), that is, determining the position where each sensor node is to move; and carrying out the movement operation for each mobile sensor node. The method is more comprehensive and practical with the rate of convergence and system energy consumption taken into account; besides, based on distributed computation of node, and through coverage hole-directed distributed differential evolution algorithm, the new position of the node is calculated without foreknowing position information of all nodes, so that calculating speed is quickened and communication cost is saved, node energy consumption is saved, and network lifetime is prolonged.

The invention discloses an addressing and capacity configuration method used for a UPFC. A UPFC addressing and capacity configuration model with transient stability constraints is established, a difference evolution algorithm serves as a framework, and the installation position and the capacity of the UPFC are optimized; in addition, a centrality correction interior point method is adopted for carrying out optimization and fitness assessment on continuous variables of each difference evolution unit. The difference evolution algorithm is simple in operation, high in searching capability, fewer in adjusting parameters and suitable for mixed optimization, and the interior point method is good in convergence and strong in robustness, so that the provided mixed algorithm can well solve the problems of UPFC addressing and capacity configuration, the method is highly efficient and convenient, and the application prospect is good.

The invention relates to a self-tuning PID energy-saving temperature control method. The method comprises the steps that a target temperature value, a control period, a sampling period and a thresholdtemperature difference are set first; load heating power is increased; the current regional temperature sampling value of a controlled object is acquired, and the sampling period is combined to calculate the temperature rise rate; a PSO-DE algorithm combining particle swarm optimization (PSO) and differential evolution (DE) is used to adjust and calculate PID control parameters; the duty cycle ofthe output pulse of a control period is adjusted according to the operation result; heating power output is adjusted, so that the sampled temperature sampling value is approximately equal to the actual temperature value of the controlled object; a temperature error caused by the heating inertia is reduced; heating power is accurately controlled; and the energy saving effect is obvious. A self-tuning PID energy-saving temperature control module comprises the self-tuning PID energy-saving temperature control method, can accurately control the heating power, and has an obvious energy-saving effect.

The invention discloses a recurrent neural network short-term power load prediction method for improving a whale algorithm, and relates to the technical field of short-term power load prediction. A recurrent neural network is used for short-term power load prediction, similar daily load data of a day to be predicted is used as input data of the recurrent neural network, and the number of input neurons, the number of output neurons, the number of hidden layers, the learning rate and the gradient descent algorithm of the recurrent neural network are determined. And a prediction model of the recurrent neural network is constructed. And the whale optimization algorithm is improved by using a differential evolution algorithm, so that the high-dimensional global optimization capability of a common whale algorithm is improved. An improved whale algorithm is adopted to pre-train the weight in the recurrent neural network, after pre-training is finished, the trained weight is put into a recurrent neural network model, then a gradient descent algorithm is adopted to train the recurrent neural network model, and after training is finished, a neural network model with the fixed weight is obtained, and then load prediction is carried out.

The invention provides a five-axis side milling machining cutter path smoothing method. The method comprises calculating the initial cutter location, and calculating the machine tool turning axle angle corresponding to the direction of a cutter axis through a dynamitic model of a specific machine tool; performing interpolation for the machine tool turning axle angle of the initial cutter location and the cutter nose point location, and obtaining a cutter nose point location trajectory and a machine tool rotation angle change curve; according to the cutter nose point location trajectory and the cutter axis vector, showing a side milling machining cutter path; calculating the rigidity matrix, corresponding to the cutter path, of the machine tool rotation angle curve; according to a differential evolution algorithm, calculating the point that the distance between the point to the cutter enveloping surface is shortest among the points on the design surface; and according to a weighted least square method, establishing a five-axis side milling machining cutter path smoothing model, and obtaining an optimized the initial cutter location and the cutter nose point location curve and a smooth machine tool rotation angle curve after solution. The five-axis side milling machining cutter path smoothing method solves the problem about smoothing and geometric deviation control of the five-axis side milling machining cutter path, and is suitable for five-axis side milling machining on a free form surface, a ruled surface or a curved surface similar to the ruled surface.

The invention discloses a feedback analytical method and a feedback analytical device during tunnel construction and based on an extreme learning machine. The feedback analytical method comprises the following steps: the optimal input layer weight and the optimal hidden layer offset are utilized to train and learn a training sample set I and a training sample set II through the extreme learning machine so as to obtain a rock classification evolution extreme learning machine model and a rock parameter identification evolution extreme learning machine model; rock classification influence factors disclosed during the tunnel construction process are obtained; the rock classification influence factors are inputted, and a rock classification result is outputted through the rock classification evolution extreme learning machine model; the rock displacement of a tunnel is monitored and obtained; according to the rock classification result, and within different rock classification ranges, the obtained rock displacement is combined, and the rock mechanical parameter is obtained by utilizing the differential evolution algorithm and the rock parameter identification evolution extreme learning machine model. According to the invention, the rock classification result and the rock mechanical parameter can be rapidly obtained, the predication is accurate, and the accuracy is high.

The invention discloses a method for automatically designing and optimizing a railway vertical profile. The method comprises the steps of carrying out ground line smoothening processing to an original ground line, fitting an initial slope to the smoothened ground line, carrying out vertical profile design constraint condition processing to the initial slope, forming a vertical profile automatic slope design scheme, optimizing the vertical profile based on a differential evolution algorithm, setting optimization control parameters, initializing the populations according to the vertical profile automatic slope design scheme, using a target function to assess the advantages and disadvantages of the individual vertical profile scheme, calculating an individual target function value, carrying out evolution among the populations by using a mutation operation, an interlace operation, a restoring operation, setting of railway tunnel and a selecting operation until the final evolution condition is reached, and outputting a vertical profile diagram and a standard check table. The method for automatically designing and optimizing railway vertical profile has the advantages of high automation degree, strong practicality, fast computing speed, and high promotion and application value in design and optimization of the railway vertical profile.

The invention discloses a calibrating method for an electronic compass. The calibrating method is used for promoting measuring accuracy of the electronic compass based on a self-adaption differential evolution algorithm and a Fourier neural network theory and is especially suitable for an orienting system with low cost and higher precision. The calibrating method comprises the following steps: utilizing a Fourier neural network to perform error modeling on the electronic compass and utilizing an improved self-adaption differential evolution algorithm to optimize a weight value of the Fourier neural network, so as to acquire an accurate error model to compensate a measuring value of the electronic compass. The error model which is established according to the calibrating method is capable of realizing accurate mapping of a sample space and has a higher nonlinear approaching capability. According to the calibrating method, a minimum local part is avoided, the defects of over-slow convergence rate and oscillation of the neural network are overcome and the influence of an outside magnetic field on the electronic compass is efficiently compensated, thereby greatly promoting the measuring accuracy of the electronic compass.

The invention provides a hand-eye calibration parameter identification method, a hand-eye calibration parameter identification system based on a differential evolution algorithm and a medium. The hand-eye calibration parameter identification method comprises the following steps of moving a robot end of a robot vision system to different poses to collect robot joint data and camera image data; separately calculating a pose matrix of the robot end relative to a robot base coordinate system and a pose matrix of a calibration board relative to a camera coordinate system; defining a rotary component calibrated error function and a translational component calibrated error function; determining and solving a multi-target optimization function of a hand-eye calibration problem; and separately calculating calibrated errors of a rotating part and a translational part of the robot vision system and verifying and identifying the optimum hand-eye calibration parameter. The global optimality of a calibration result can be acquired, the acquired calibration result falls onto a special Euclidean group SE (3), and additionally introduced calculation of calibrating orthogonalization of the acquiredrotating matrix.

The invention discloses a micro-grid operation optimizing method in consideration of the real-time electricity price and the controllable load. The real-time electricity price and the controllable load are introduced, difference minimization of the user electricity cost, an HVAC and an EWH with the corresponding target temperatures serves as the optimization target, the proportion of power supplied to the load by a distributed power source to charging power for stored energy with the distributed power source, the delay time of a delay load and the working state of a schedulable load serve as the decision variables, optimization and simulation are performed through a differential evolution algorithm, and a simulation and optimization operation result is analyzed. The delaying performance of the delay load and the selectivity of the operation state of the schedulable load are excavated and reflected, and the effectiveness and applicability of the optimizing method are verified. The method fully reflects the guide function of the real-time electricity price on the user electricity habit; the controllable load is divided into two parts, and regulation of the load is reflected; resources on the demand side are fully utilized, and a theoretical support is provided for economic operation of a micro-grid.

The invention provides a method for predicting the residual life of a lithium ion battery based on a WDE optimization LSTM network, and relates to the technical field of lithium ion batteries. The method comprises the steps of firstly, constructing two groups of lithium ion battery monitoring indexes; acquiring lithium ion battery monitoring data, and extracting lithium ion battery monitoring index data and lithium ion battery capacity data from the lithium ion battery monitoring data; then determining an LSTM network structure, and building a model for indirectly predicting the residual lifeof the lithium ion battery based on LSTM; utilizing a weighted differential evolution algorithm to optimize key parameters in the model for indirectly predicting the residual life of the lithium ion battery; utilizing optimization data to determine an optimal model for indirectly predicting the residual life of the lithium ion battery; and finally, predicting later lithium ion battery capacity data by utilizing the optimal model for indirectly predicting the residual life of the lithium ion battery. According to the method for predicting the residual life of the lithium ion battery based on the WDE optimization LSTM network, the change rule of the lithium ion battery capacity data can be accurately predicted, and the residual life of the lithium ion battery can be effectively evaluated.

The invention provides a layered dynamic regulation method for multiple energy media. The method comprises the steps that on the basis of an energy gradient utilization scheme, an energy medium layered dynamic regulation frame is built; according to the energy medium layered dynamic regulation frame, a regulation scheme of an energy subsystem is built; according to the regulation scheme of the energy subsystem, an optimization model and a corresponding solution mode which are suitable for layered dynamic regulation are built; according to the optimization model and the solution mode of layered dynamic regulation, a collaborative optimization algorithm of bionic intelligence is designed to solve the optimization model; dynamic regulation is carried out on multiple energy media according to the result obtained through solution. According to the layered dynamic regulation method for multiple energy media, a regulation strategy suitable for layered step collaborative optimization of multiple energy media is built, multi-cycle multi-target dynamic collaborative optimization of energy regulation is achieved, the optimization model is built, a differential evolution algorithm with a self-learning mechanism is designed to solve the optimization model, collaborative optimized dispatching of multiple energy media in a complex environment for steel production is achieved, and the method is good in comprehensive economical efficiency and effectiveness.

The invention discloses a distribution network expansion planning method based on strategy adaption differential evolution. A capital time value is considered, calculation is carried out according to equal amount installment payment capital recovery, a model is established by taking least annual investment and operation maintenance cost as a target, a differential evolution algorithm based on strategy adaption on the basis of the model is proposed, an optimal solution obtained by the algorithm is decoded, a planning line is drawn on an electric power GIS (geographic information system) platform according to a line number obtained by decoding to carry out structure repair to an individual, and the individual is enabled to satisfy a radiative network structure. An adaption evolutionary strategy is continuously updated in an iteration process to prevent the algorithm from being trapped in local optimum, and a planning process of a distribution network line is enabled to exhibit better interactivity, more intuitive planning result and flexibility in planning scheme regulation by a GIS.

The invention discloses a real-time yield prediction method for a hydrocracking device. Field real-time data are processed with a data reconciliation technology, and hydrocracking reaction kinetics parameters are corrected in real time in combination with an improved differential evolution algorithm, so that a mechanism model can accurately describe the actual running condition of the device. On the basis of the corrected model, effects caused by key operation/process conditions such as the raw material density, the sulfur content, the nitrogen content, the reaction temperature, the pressure, the hydrogen-to-oil volume ratio and the like on hydrocracked products are analyzed. Piecewise linearization is performed according to the effect trend, a linear equation is solved, corresponding Delta-Base yield data are acquired, the operation condition is associated with the Delta-Base data with a neutral network modeling technology, a yield surrogate model is established, the yield data calculation speed is increased, real-time prediction of the yield of products of the hydrocracking device is realized, and theoretical support is provided for establishing an accurate plan optimization PIMS (process industry modeling system) model.

The invention discloses a furnace temperature controlling method in a heating process of plate blank of a heating furnace. An optimal control scheme is determined by adopting an adaptive differential evolution algorithm according to the heat transfer characteristic in the heating furnace based on the constraint conditions such as heating mass requirements, production equipment safety and the likeof the plate blank, so that the setting temperature of the heating furnace is obtained and temperature control in the heating process of the plant blank is realized. On the premise that the production requirement is met and the temperature of the plant blank meets the hot rolling requirement, the adaptive differential evolution algorithm is improved, and mutagenic factors and crossover probability are adaptively regulated, so that the global searching ability of the algorithm can be improved, the algorithm convergence is quickened and the searching precision of the algorithm is improved. By implementing the method, the temperature of the heating furnace is reduced, over-burning is avoided, the oxidative burning loss of the plant blank, the energy consumption, the production cost and the discharge of waste gas are reduced, and the economic benefits and the social benefits of an enterprise are improved.

The invention relates to a method of optimallying designing a closed-winding permanent magnet brushless DC motor on the basis of differential evolution algorithm and finite element time-stepping method. The method includes: according to the particularity in the closed-winding permanent magnet brushless DC motor, by considering a body and a control part of the motor, using the cost in torque density and unit torque of a system as an optimization target; using constraint conditions such as parallel branch circulation, power factor, efficiency, torque output, fluctuation, heating, structural strength and the like. The method has the advantages that design schemes of each generation are varied, intersected and selected for differential evolution, and a Pareto optimal solution set is updated continuously; in order to ensure design accuracy, the finite element time-stepping method is applied to accurate dynamic simulation computation of the closed-winding permanent magnet brushless DC motor; the reliable and accurate method is provided for the design optimization of the closed-winding permanent magnet brushless DC motor, and the development of the novel motor is quickened.

The invention discloses a low sidelobe level series-fed microstrip antenna with non-uniform array element distance. Beam width, an array element size and the array element distance are used as constraint condition inputs, and the distance and the relative amplitude of an antenna array element, which are acquired, are optimized by a differential evolution algorithm. Compared with a series-fed microstrip antenna, the series-fed microstrip antenna has the advantages of lower sidelobe level and narrow beam width, and moreover, the conversion between the low sidelobe level and the beam width can be achieved by the differential evolution algorithm.

The present invention discloses a software defect prediction optimization method based on a differential evolution algorithm, and belongs to the field of quality assurance in the software engineering.The method comprises the following steps: arranging modules in the software project, cleaning annotations and the like in the code, and establishing a software defect data code set; arranging the given defect set, including the defect metric design, the defect data marks, and the like, to generate a software defect data set; and with a differential evolution algorithm, creating a ratio of a majority class to a minority class as 2:1 for a defect prediction data set by using a minority class oversampling method, determining an optimal value of the neural network hyper-parameter, using a trainedneural network classification model to test in a test set, and if the performance indicators are satisfied, representing that a software defect prediction model is successfully established. Accordingto the method disclosed by the present invention, corresponding parameter factors in the classification model construction can be automatically classified according to the difference of the data sets, a parameter combination most suitable for the current data set and the classification model can be found, the performance of the software defect prediction model can be improved, and the workload ofparameter searching in the model construction can be reduced.

The invention provides a parking guidance method. The parking guidance method comprises the steps: acquiring the real-time road traffic data and the real-time parking lot data in an urban road network area, wherein the real-time road traffic data includes the road network data; according to the road network data, constructing a traffic network topological diagram; according to the destination and the real-time parking lot data, generating candidate parking lot sets; taking the user demand data as an objective function, according to a high dimension multi-target differential evolution algorithm, and selecting the optimal parking lot set from the candidate parking lot sets; and utilizing a single-target differential evolution algorithm to generate a parking guidance result according to the positional data of the user and the target parking lot which is selected from the optimal parking lot set by the user. The parking guidance method can realize static high dimension multi-target optimal parking lot selection and path guidance before a trip, can also provide dynamic real-time high dimension multi-target optimal parking lot selection and path guidance during the driving process, can fully consider dynamic changes of the parking lot and the road traffic information, and can effectively improve the accuracy and the intelligent degree for parking guidance.

The invention discloses a multi-target task scheduling method and a multi-target task scheduling system. Fuzzy clustering processing is performed on a task and a virtual machine by using a fuzzy clustering algorithm; the obtained task clustering set is matched with the virtual machine clustering set; obtaining a matching set, differential evolution algorithm, scheduling the tasks in the matching set to the virtual machines in the matching set; obtaining initial scheduling results, according to the load value of the virtual machine in the initial scheduling result; dividing the virtual machinesinto a heavy load set, a balanced load set and a light load set; selecting a task set needing to be rescheduled from the heavy load set, obtaining a rescheduling task set and adopting a Q-value algorithm; tasks in a rescheduling task set are scheduled to virtual machines in a light load set, the technical problem that a scheduling strategy obtained through a traditional task scheduling algorithmis poor is solved, local resource tasks are reallocated through a Q-value method. The load balancing of global resources is achieved, and the resource utilization rate is increased.

The invention provides a wing profile optimal design method of a parallel difference evolutionary algorithm based on open computing language (Open CL). The wing profile optimal design method is used for wing profile design. A standard wing profile function and a profile function are selected, the profile function serves as a design variable, an optimization objective function is determined, steps of the difference evolutionary algorithm are divided into different stages according to processed data, and all stages are packaged in different cores for operating based on Open CL. In population updating, mutation operators are used for generating test vectors and crossover operators are used for generating descendants; individuals in a population are restored to the wing profile shape and tested whether to meet geometric constraint; computational fluid dynamics (CFD) analysis is used for obtaining fitness of the individuals and seeking the optimum individual, and whether the optimum individual meets performance constraint is tested; and finally iteration is finished and the optimum result is copied back into a host internal storage. The wing profile optimal design method achieves parallel processing of a wing profile design process, performs sufficient search in effective space, shortens design period, achieves cross-platform wing profile optimal design, and improves design efficiency.