1055results about How to "Achieve optimization" patented technology

A digital camera which is an image processing apparatus is controlled so that a first image and a second image are acquired for the same subject in different timings. A plurality of correction tables for use in shading correction are prepared in advance in a shading ROM of the digital camera. When a first image and a second image are acquired, multiplication of one correction table is performed on both images by a shading correction circuit. As a result of this, a shading correction is effected on both images. The first image and the second image that have been subjected to the shading correction are taken into respective image memories, and after being subjected to a positioning process by a general controller, they are synthesized to generate a synthesized image.

The invention provides a secondary battery pack charging and discharging management system. The secondary battery pack charging and discharging management system comprises a battery pack, a step-up and step-down module and a main control module, wherein the battery pack is formed by a plurality of individual battery modules through serial connection; a self-control part of each individual battery module is used for controlling a serial-in switch and a bypass switch to be turned off or turned on; the main control module is used for generating a signal for controlling a battery pack step-up and step-down circuit short-circuiting switch and a signal for selecting a single battery to be serially connected into the battery pack according to the working state of the battery pack, and controlling a step-up and step-down circuit to adjust charging and discharging voltage and current. By adopting the secondary battery pack charging and discharging management system provided by the invention, under the situation that the capacity and the internal resistance of each single battery are different, the utilization ratio, the service life and the safety of the entire battery pack are improved; by calculating balance use battery quantity, under the situation that any step-up circuit is not used, different voltage can be obtained; the operating costs of electricity-consuming equipment such as electric vehicles which use the secondary battery pack charging and discharging management system can be reduced and the overall working efficiency of the equipment is improved.

The invention discloses a multi-objective optimization scheduling method for an electric vehicle charging station including a photovoltaic power generation system, and belongs to the technical field of smart power grids. Power purchase cost minimization and circulating electric quantity minimization of an energy storage system are used as objective functions, and a multi-objective optimization scheduling model of the electric vehicle charging station including the photovoltaic power generation system is built; the decision variable of the scheduling model and the constraint condition of the decision variable are determined; basic data are determined; solution is carried out through a multi-objective optimization algorithm to obtain a non-dominated solution leading surface, and then multiple Pareto optimal solutions are obtained; according to the low comprehensive cost of the main circulating electric quantity of the energy storage system and the power purchase cost, the charging station scheduling optimal scheme is selected at last. The multi-objective optimization scheduling method is suitable for the electric vehicle charging station including the photovoltaic power generation system of various cities with rich light sources; the scheduling scheme of the electric vehicle charging station including the photovoltaic power generation system is optimized; theoretical bases and technical supporting can be provided for the scheduling of the electric vehicle charging station including the photovoltaic power generation system; the operation economical efficiency of the charging station is improved.

The invention discloses a robot navigation positioning system and method, which are used for map construction, positioning and path planning of a robot. The method comprises the following steps: S100,positioning is carried out, in the positioning step, the robot detects surrounding environment information through multiple sensors, and later, based on an adaptive particle filtering SLAM algorithmand in match with different odometers, real-time map construction and positioning are completed; and S200, path planning is carried out, in the path planning step, a two-phase hybrid state A*-based path planning algorithm is adopted, after a path length and the number of extended nodes are obtained when path planning is carried out on a rasterized map, a higher rasterized map is obtained through parsing and extension, and the acquired path length and the acquired number of extended nodes are used as input of fuzzy reasoning, a heuristic weight is obtained through fuzzy reasoning and is used asinput of search of a second stage, and path planning is performed on a higher rasterized map. The system and the method disclosed in the invention can not only adapt to different environments but also can perform dynamic path planning.

Disclosed are an active distribution network optimal configuration structure and a configuration method of the active distribution network optimal configuration structure. The active distribution network optimal configuration structure is formed by connecting an active distribution network optimizing system, a fuel oil generator set, a wind power generation set, a photovoltaic power generation set, a storage battery energy storage device, an EV charging pile, an AMI and a load into a distribution network through respective converters, and the active distribution network optimizing system, the fuel oil generator set, the wind power generation set, the photovoltaic power generation set, the storage battery energy storage device, the EV charging pile, the AMI and the load are connected into a large power grid through a public connecting point. Isolated island operation and grid-connected operation of the active distribution network can be achieved by controlling the active distribution network through an active distribution network optimizing system. A dual-layer optimization plane design model suitable for the active distribution network is provided, and comprehensive optimization of distributed power supplies of the active distribution network and scheduling optimization of the distributed power supplies can be achieved.

The invention discloses a method, a device and a system for optimizing a GPS track and relates to the information processing technology. The method comprises the following steps of: after a valid starting sampling point of the GPS is determined, beginning to cache CPS sampling points from the valid starting sampling point of the GPS, deleting error points during the caching; screening the cached GPS sampling points and then transmitting the Cached GPS sampling pints to a Gis server; generating a track vector map by the Gis server according to the screened GPS sampling points; and realizing optimization of the GPS track after the GPS sampling points are subjected to the operations of deletion and screening of the error points, and the track vector map is generated by the Gis server.

The embodiment of the invention provides a routing decision method based on deep reinforcement learning under an SDN framework and a device thereof. The method is applied to SDN controllers and comprises the following steps: obtaining real-time traffic information in a network; determining priority of each traffic; and inputting the real-time traffic information into a pre-trained deep Q network DQN and successively determining routing of each traffic according to priority level sequence of each traffic. The routing decision method based on deep reinforcement learning under an SDN framework and the device thereof are capable of achieving load balance of the networks in the networks of various topological structures, reducing the occurrence of network congestion and achieving optimization of routing policies in the network environment with highly dynamic change of the network traffic.

The invention discloses a relative attitude measurement real-time dynamic filter method based on a dual-inertial measurement unit / differential global positioning system (IMU / DGPS) combination. The method comprises the following steps of: resolving through strapdown inertial navigation in real time by adopting a dual optical fiber strapdown inertial navigation system to obtain navigation information of a master inertial navigation system and a slave inertial navigation system; judging whether information of a differential global positioning system (DGPS) is updated, and generating two situations: when the information of the DGPS is updated, the master inertial navigation system and the slave inertial navigation system perform filter correction to construct a measurement equation of a combined navigation filter, when the information of the DGPS is not updated, the master inertial navigation system is used for performing the filter correction on the slave inertial navigation system to construct a measurement equation of the combined navigation filter; discretizing the combined filter measurement equations obtained according to the two situations, constructing a recurrence equation of a discrete kalman filter, and resolving to obtain a pitching angle, a transversely rolling angle and a heading angle of each of the master inertial navigation system and the slave inertial navigation system; and then resolving a relative attitude matrix to obtain main values of relative attitude angles of the master inertial navigation system and the slave inertial navigation system. According to the method, the stability of a navigation system is improved; the speed information, the position information and the attitude information of the navigation system during measurement can be output in real time; and the measurement range is large.

The invention relates to a rare earth separation method with the material linkage cyclic utilization function. The rare earth separation method comprises the following steps: the organic phase of loaded rare earth prepared by extractant A and rare earth soap stock through mixing is used for the follow-up linkage extraction separation, and inorganic acid in the residual water phase is extracted and concentrated by extractant C and is then reused for material dissolving or recovering rare earth with oxalic acid precipitated therein; purified rare earth solution is subjected to extraction separation, the rare earth is precipitated through oxalic acid, sediment mother solution containing oxalic acid and inorganic acid is mixed with extractant B, the oxalic acid is extracted to be reused for precipitating rare earth, and the residual inorganic acid is directly used for washing and back extraction processes or is used for material dissolving after being concentrated by the extractant C. With the rare earth separation method provided by the invention, intermediate materials generated in the rare earth separation process can be recycled in process sections in a linkage manner, the alkali saponification process is avoided, and the material dissolving, washing and back extraction processes can be finished only by the recycled inorganic acid, so that the rare earth separation and purification process disclosed by the invention does not consume alkaline and inorganic acid, the cost is low, and the rare earth separation method is green and environmental-friendly.

The invention relates to a binocular depth estimation method based on a depth neural network, which comprises the following steps: 1) preprocessing the input left and right viewpoint images to enhancedata; 2) constructing a multi-scale network model for binocular depth estimation, wherein the model comprises a plurality of convolution layers, an activation layer, a residual connection, a multi-scale pooling connection and a linear upsampling layer; 3) Designing the loss function to minimize the results in the continuous training process, so as to obtain the optimal network weights; 4) inputting the image to be processed into the network model to obtain the corresponding depth map, and repeatedly repeating the above steps until the network converges or reaches the training times. The invention adopts the idea of unsupervised learning, and only the left and right viewpoint images obtained by the binocular camera are used as network input. The adaptive design of the network sets the internal and external parameters of the camera as a single model parameter, so it can be applied to multiple camera systems without modifying the network.

The present application provides a method for implementing a device to device (D2D) communication between UEs, wherein a UE and a network obtain each other's capability of Proximity Service (ProSe); the UE may activate the capability of ProSe that both the UE and the network support and initiates a request for a ProSe service with another UE to the network; the network authenticates the request for the ProSe service according to subscription information for the UE's device to device communication; and the network establishes a device to device PDN bearer between the two UEs for the request for the ProSe service passing the authentication. The present application further provides a user equipment. The present application is the basis for implementing the ProSe, the application thereof can satisfy service demands of public safety communication and realize network resource optimization configuration.

The invention relates to a frequency planning method in GSM network. The frequency planning method in GSM network comprises the following steps: A, infrastructure data collection and collation; B, coverage optimization; C, neighboring optimization; D, optimization of same frequency and same BSIC (Base Station Identity Code); E, measured data collection and collation; F, interference matrix building; G, frequency model judgment; H, planning object selection; I, frequency planning; and J, cutting implementation according to frequency planning results; Step B is based on data of Step A, in Step B, a comprehensive general survey on the planning network is made, an antenna of the network, a repeater, an indoor distribution system and an extension system are understood and mastered comprehensively; and then coverage is optimized; Step G is based on data of Step A, through the analysis of network mean carrier frequency configure number and density of mean carrier frequency configure per unit area, the method for optimal frequency multiplexing model is acquired; Step G is based on data of Step A, according to requirements of the planned residential areas, a region, scale and capacity, the method for planning the region is acquired. The invention has high response speed.

The invention discloses a deformation skin for an aircraft. The deformation skin for the aircraft is formed by connecting N continuous cellular units (21), wherein the N is a natural number. The deformation skin for the aircraft is characterized in that: elastic substrates (3) are filled in the cellular units (21). According to the deformation skin for the aircraft, the capability of the deformation skin of transferring load is improved as high as possible on the premise of ensuring the deformation capability by a method for compounding different performance materials, the problem of contradiction between higher elastic deformation and transfer of airload under the action of lower driving force is solved, the local flow fields are actively controlled further by mean of an electrostrictive effect of polyurethane elastomers, and the deformation skin has the capability of deforming in different scales. According to the deformation skin, the properties of intelligent morphing aircrafts are comprehensively optimized.

The invention discloses a multi-target optimal hybrid power flow algorithm of a regional comprehensive energy system. The multi-target optimal hybrid power flow algorithm includes: building the mathematical models of a power distribution system, a gas pipe network and an energy center, building a multi-target optimal scheduling model of regional comprehensive energy system economic cost and pollution gas discharge quantity, selecting constraint conditions, calling OpenDSS in MATLAB software to perform power flow calculation, performing optimization on the basis of an improved non-dominated sorting genetic algorithm to obtain the Pareto frontier of the regional comprehensive energy system economic cost and pollution gas discharge quantity, and comparing the Pareto frontier with a single-target optimization result using the economic cost and pollution gas discharge quantity and performing analysis to prove the correctness of the algorithm and further obtain the current optimal scheduling scheme of the regional comprehensive energy system. The multi-target optimal hybrid power flow algorithm has the advantages that many scheduling schemes can be provided for operation staff while system operation constraint conditions are satisfied, and the final operation scheme can be decided according to actual needs.

The embodiments of the invention disclose an optimization method and device for a neural network. The method comprises the steps of: acquiring a first neural network satisfying a set precision condition, and processing a training sample set based on the first neural network to obtain a first feature vector of each training sample in the training sample set; constructing second neural networks to be trained based on a set network construction condition; training the second neural networks according to the first feature vectors and the training sample set, and determining a second neural network satisfying the set precision condition; and determining the second neural network as a target neural network of the first neural network. By using the method, another newly-constructed small-scale neural network can be directly trained and learnt according to the optimization condition and then determined as a target optimization network of the neural network to be optimized, so that when feature recognition is performed based on the optimized neural network, the purposes of improving the recognition speed, shortening the recognition time and reducing the spatial occupation of a memory, a running memory, a display memory and the like can be fulfilled.

The invention discloses an optimization method of a convolution neural network. In a convolution neural network training process, a filter group with poorer effects and a filter group with better effects are selected out, and parts of coefficients in the filter group with poorer effects are replaced by parts of coefficients in the filter group with better effects, parameter update between convolution layer filters is achieved. After a modified network has been trained for a certain time, error magnitude generated by the network after and before the modification is compared, whether the previously achieved parameter update of the filters is valid or not, selection is performed in the two networks, and the network with good performance is used for following trainings. By continuously repeating the above processes, a network with better performance can be obtained through the training and the feature extraction ability of the network is improved compared with the traditional network.

The invention relates to a bus line optimizing method based on taxi GPS data and mobile phone signaling data, and aims to solve the problem of low operation efficiency of a bus line because the existing bus line is unreasonable in setting, cannot meet needs very well and does not conform to the actual situation. The bus line optimizing method comprises the following specific processes: 1, acquiring the taxi GPS data, acquiring the mobile phone signaling data, acquiring a land use map and acquiring bus stop data; 2, processing the data acquired in the step 1; 3, calculating bus line optimizingindexes according to the step 2; 31, calculating the distances from an origin point and a destination point to the nearest bus stop; 32, determining the number of bus lines included in the bus stop and the traffic flow; 33, calculating the average running speed of a road vehicle; 4, setting optimizing measures according to the optimizing indexes. The bus line optimizing method is applied to the field of public transportation planning.

The invention discloses a boiler control and combustion optimization method and solves the problem that the boiler system is difficult to control for a long time. The method includes the steps of 1, establishing a target function; 2, recording an optimized target function value and an operating parameter to be optimized when the boiler system stably operates; 3, adding a preset step value to the operating parameter to be optimized; 4, after the boiler system stabilizes and the target function increases, continuing to add a step value until the target function value decreases; if the target function value decreases, performing reverse step optimization, decreasing by one step value until the target function value increases, and confirming that the boiler system is in the optimal combustion area. The boiler control and combustion optimization method has the advantages that system stability is ensured, system performance is improved, labor intensity is reduced for workers, the system is integrally optimized, and economic benefit is high.

The invention provides a method for energy optimization of a distributed power generation and energy supply system, belonging to the technical field of electric power energy conservation. The method is mainly applied to evaluation of energy-saving and emission-reduction benefits of the distributed power generation and energy supply system when the distributed power generation and energy supply system is connected to an electric network. The method comprises the following steps of: firstly, establishing an energy optimization model of the distributed power generation and energy supply system with four sub-goals of minimizing the total annual planning cost, maximizing the generating capacity of renewable energy sources, minimizing the annual power failure quantity and minimizing the annual capacity shortage; secondly, defining goal optimization functions of the four sub-goals, unifying dimensions of the pollution discharge capacity, the power failure quantity and the capacity shortage of the system through gross penalty, and establishing a single-goal optimization function of the system by adopting a linear weighted sum method, wherein weighing coefficients are determined by adopting a dualistic contrast constant weight method; and finally, performing simulation and solving on the distributed power generation and energy supply system by adopting a heuristic global optimization algorithm to obtain values of the four optimization sub-goals. By the method, the multi-goal energy optimization of the distributed power generation and energy supply system is realized, and the aims of energy saving and emission reduction are achieved.

The invention discloses a logistics node facility layout optimization method based on computer vision. The method comprises the following steps: introducing Multi-Agent technique and CA (cellular automata) simulation technique to construct a geographical simulation system based on computer vision technique; representing plot terrain and constraint conditions of logistics facility by computer graphics; constructing a three-layered Multi-Agent model with optimized logistics node facility layout; generating an initial solution based on the advantage of a construction algorithm on creating an initial scheme; and combining a genetic algorithm and simulated annealing algorithm to provide a logistics node facility layout optimization algorithm by adopting the idea of an improved algorithm. The method provided by the invention can achieve optimization of multi-constraint large-scale logistics node facility layout under complex topographic conditions, and can be widely used in the fields of urban planning, transportation, large-scale integrated circuit design, architectural design, machinery manufacturing, aerospace, and so on.

The invention discloses a control method of a memory device and a relevant device. The memory device is arranged in a data processing system host machine, a controller of the data processing system host machine performs unified management to FLASH of the memory device, superincumbent unified management and control the FLASH of the memory device can be achieved, and achieving of resource optimization is facilitated.

A structure finite element model correcting method based on multi-element uncertainty comprises the following steps: (1) building an initialized parameterization equivalent finite element model in finite element software; (2) screening out significance parameters; (3) obtaining sample points, and constructing an incomplete variable high-order response surface model; (4) judging validity of the response surface model, if the validity of the response surface model meets the requirement, executing the next step, and if the validity of the response surface model does not meet the requirement, executing the step (3) again; (5) building a rapid random sampling analysis model with the combination of a high-order response surface and the Monte Carlo method, and conducting statistics on a mean value and a covariance matrix of simulation output responses; (6) conducting statistics on a mean value and a covariance matrix of test output results; (7) constructing a weighting objective function of the mean values and covariances of tests and simulation; (8) reversely estimating a mean value and a covariance matrix of input parameters; (9) judging whether the mean value and the covariance matrix of the input parameters meet correction accuracy or not, if yes, stopping iteration, and if not, executing the step (8) again. According to the structure finite element model correcting method, the calculated amount of the iteration is reduced, the application range is wide, and optimization of a large-scale parameter range is achieved.

The invention discloses a coordination and optimization control method for a multi-energy complement comprehensive energy system, optimization control for comprehensive energy system cooling heating and power energy flow is realized through a hierarchical regulatory mechanism, wherein a scheduling optimization layer serves operation cost minimization as a target, according to system operation constraint conditions, cooling heating and power load requirements are combined to perform current plan optimization, according to the current load plan obtained by the scheduling optimization layer, a coordination control layer combines current operation condition of the system to obtain a real-time cooling heating and power load instruction, and the coordination control layer transmits the instruction to an automatic control system of equipment relative to the comprehensive energy system through a real-time control layer. According to the invention, coordination dispatching and control of the comprehensive energy system in current and real time can be realized, negative influences for system optimization caused by energy demand uncertainty and load prediction error are eliminated, economic operation of the comprehensive energy system is realized, so that energy utilization rate of the system is improved.

The invention discloses an overall energy saving control method a central air conditioner. The method is characterized by including: generating the behavior feature models of each equipment and the operating environment of the central air conditioner, updating the feature models, performing optimization calculation while the central air conditioner water system balance and heat and mass balance conditions are satisfied and on the basis of behavior feature predicting models, and executing optimal operating work conditions. The prediction-based overall energy saving control method of the central air conditioner has the advantages that the operation data of each equipment and the operating environment, the structure-uniformed and universal behavior feature predicting models of each equipment and the operating environment are generated and updated on the basis of the system identification technology, the behavior feature predicting models can constantly approach the actual operating features of each equipment and the operating environment in an online manner, and the equipment which is used for a long time and has performance degradation and deviation can be accurately predicted; the central air conditioner water system balance and heat and mass balance conditions are introduced into the prediction, and overall work condition optimization is achieved.

The invention discloses a network management method for managing network users and network equipment in a same management system, which includes the steps of: taking the users and the network equipment communicating with the users as two management objects and establishing incidence relation between the two management objects; determining another management object correlating with the current management object according to the established incidence relation; at least providing one basis or means for carrying out management operation to another management object. The invention also provides a network management system which includes a management module and a memory module. The network management proposal provided by the invention can realize the management on the network equipment and the network users in the same network management system, thus reducing the complexity of the management operation.

The invention relates to a fully distributed intelligent power grid economic dispatching method based on deep reinforcement learning. The fully distributed intelligent power grid economic dispatchingmethod comprises the following steps: 1), obtaining a network topological structure, and building an economic dispatching model based on load distribution and unit combination; 2) obtaining a local optimal solution of the economic dispatch model through the deep reinforcement learning model to serve as a first Q function table; 3) loading the first Q function table into a pre-trained deep convolutional neural network to obtain a second Q function table; and 4) initializing the power of each unit according to the second Q function table, loading a unit power solving model, and updating the second Q function table according to the network topology structure to obtain a globally optimal solution. Compared with the prior art, the fully distributed intelligent power grid economic dispatching method has the advantages that economic dispatch optimization can be realized in an intelligent power grid environment with large data volume and complex network structure, and the fully distributed intelligent power grid economic dispatching method does not depend on a clear target function, can adapt to the plug-and-play characteristic of distributed energy, and has a good application prospect.

The invention provides a regional control phase timing optimization method based on lane saturation. The method includes identifying a congested road in a region, analyzing relevance roads, i.e., an adjacent signal control intersection and an uncongested traffic direction, i.e., a control direction of the signal control intersection, and adjusting the phase timing to achieve the optimization of aregional signal scheme. According to the regional control phase timing optimization method based on lane saturation, a control direction in which coordination and optimization can be made, i.e., the control direction in which the green light duration can be reduced can be identified according to the road saturation, then the green light duration of an overlapping phase can be adjusted through adding the overlapping phase / making phase late on early off, so that an intersection signal scheme can be effectively optimized through a method of phase sequence adjustment and green light duration optimization. As a result, it is possible to realize the regional optimization of the traffic signal, greatly improve the efficiency of the regional optimization, and solve the problem of area congestion caused by a situation that in a traditional optimization mode, only the green light duration of each phase can be adjusted but demands for green light of the traffic direction under the same stage areinconsistent.

The invention discloses a fuzzy sliding mode and neural network based composite control system of a permanent synchronous motor. The control system comprises a rotation speed control outer ring and a current control inner ring, wherein a control unit in the rotation speed control outer ring adopts a composite control module; the composite control module comprises a fuzzy sliding mode control module and a neural network control module; and the fuzzy sliding mode control module and the neural network control module are combined together through two weighing factors lambda n and lambda f. The control system has the advantages of both the fuzzy sliding mode control module and the neural network control module, so that the strong robustness and high stability of the control system of the permanent synchronous motor are realized, the steady state precision is improved, and moreover, the rapidness in control and the strong load disturbance resistance are achieved.

The invention discloses a boiler combustion multi-constraint multi-objective optimization expert system and its optimization method. The system includes a DCS control system module, an industrial coal type analysis module, an environmental pollutant monitoring module, an approximate case query unit, a historical database, and a boiler combustion model Library, boiler optimization expert system server program and man-machine interface. The boiler combustion multi-constraint multi-objective optimization expert system and its optimization method proposed by the present invention can realize the prediction of boiler combustion thermal efficiency and pollutant discharge; at the same time, reasoning and query are carried out according to the combustion model library, and the intelligent algorithm of the boiler optimization expert system is adopted , can find the parameter setting scheme to achieve the best working condition of the boiler, and then adjust the boiler operating state through DCS to realize the overall optimization of boiler combustion.