74results about How to "Improve optimization results" patented technology

The invention discloses a short-term photovoltaic power prediction method based on VMD-IPSO-GRU, and belongs to the technical field of photovoltaic power generation and grid connection. Firstly, a historical photovoltaic power time sequence is decomposed into sub-sequences with different frequencies through variational mode decomposition, geographic information and component parameters contained in photovoltaic sequence data are fully mined, and signals and noise of original data are separated; secondly, main meteorological factors influencing photovoltaic output are determined through Spearman and Pearson correlation coefficients; and finally, gating cycle unit network models are established for the sub-sequences decomposed by the VMD respectively, and the GRU nerve is optimized through an improved particle swarm algorithm and an adaptive moment estimation algorithm, thereby improving the network convergence rate and the data fitting effect, accurately and efficiently finishing short-term photovoltaic power prediction, and avoiding errors caused by manual parameter adjustment.

The invention provides a distribution method of container quay berths and shore bridges. By adopting a rolling type plan distribution method, berth and shore bridge distribution models based on multi-objective planning are constructed; the models are based on a continuous quay wall line and are more closer to the actual berth conditions of a quay; a hybrid algorithm on the basis of combining a heuristic algorithm and a parallel genetic algorithm is adopted, and the performance of the hybrid algorithm is evaluated by a distribution simulation system of the container quay berths and the shore bridges; when a berth and shore bridge distribution scheme is generated, the simulation system simulates the distribution scheme, acquires corresponding performance indexes, compares with other schemes, and determines whether the scheme is better; and a method combining simulation and a gene repair technology is adopted to repair infeasible schemes, thereby being favorable for reducing the time in port of a ship, and reducing the horizontal transport distance when the ship is loaded or unloaded, the energy consumption of the shore bridges, and the fine that the quay pays to a ship owner, and further reducing the loading and unloading cost on the quay, improving the service quality of the quay and realizing the purpose of the invention.

Embodiments of the invention disclose a deep neural network training method and system and electronic equipment. The method comprises the following steps of: in a forward propagation process, carryingout scene analysis detection on a sample image by utilizing a deep neural network model so as to obtain a first scene analysis prediction result output by a middle network layer and a second scene analysis prediction result output by a tail network layer; determining a first difference between the first scene analysis prediction result and scene analysis labeling information of the sample image and a second difference between the second scene analysis prediction result and the scene analysis labeling information of the sample image; and in a counter-propagation process, adjusting parameters of a first network layer according to the first difference and adjusting parameters of the a second network layer according to the first difference and the second difference, wherein the first networklayer comprises at least one network layer between the middle network layer and the tail network layer, and the second network layer comprises other network layers except the first network layer. According to the method and system and the electronic equipment, better network model optimization results can be obtained.

The invention discloses a method and a system for planning an optimal route and an optimal driving mode of an electric automobile. The method comprises the following steps: based on reachability analysis, performing sieving so as to obtain a practicable route set of the electric automobile under different driving modes and a corresponding charging mode set; after a user weight coefficient is obtained, calculating evaluation indexes of each practicable route under different driving modes, wherein the user weight coefficient is the coefficient preset by a user or is the coefficient which is automatically obtained according to the driving habits of the user; and based on the calculated evaluation indexes, planning the optimal route and the optimal driving mode. Through the adoption of the method disclosed by the invention, according to the demand indexes of the user, the evaluation indexes of corresponding weighting are calculated, so that the optimal practicable route, and the optimal driving mode, namely the combination of a driving mode and an air conditioner mode are obtained through sieving; and automobile characteristics of the electric automobile can be sufficiently utilized, so that the optimal route planning result and the optimal driving mode planning result are obtained, the optimizing result is good, and the method can be widely applied in the control field of the electric automobile.

The invention provides an energy management method considering a random forecast error of an independent microgrid. The independent microgrid comprises a wind power-photovoltaic-micro gas turbine-energy storage battery and an energy management system (EMS), wherein the EMS is used for monitoring a state-of-charge (SOC) state of the energy storage battery in real time, monitoring wind power and photovoltaic actual generation power and measuring actual powers of the micro gas turbine and the energy storage battery, a random forecast error of a net load is obtained according to powers which are forecasted and actually monitored, the contribution of the controllable micro gas turbine and the energy storage battery are arranged and adjusted by a multi-variable and multi-constraint mathematicaloptimization method and a set process according to the size of the random forecast error of the net load, so that the energy storage battery normally runs. According to the method provided by the technical scheme, the unfavorable influence brought by the forecast error of the load and a regenerative energy power is overcome, the contribution of the controllable micro source and energy storage arearranged and adjusted by the mathematical optimization method according to the size of the random forecast error of the net load, and stable, reliable and economic running of the microgrid is achieved.

Polymer flooding is an important tertiary oil recover technique, but the price of polymer is expensive, and the responding time is relatively lagged. In order to scientifically establish a development scheme and to obtain better economic benefit, the invention provides a polymer flooding development scheme optimization design method on the basis of the dynamic plan technique. The optimization of the polymer flooding scheme can be specifically described as a distributed parameter system dynamic plan problem with state and control inequality constraints. For a fixed slug length injection optimization problem, a numerical solving algorithm based on iteration dynamic plan is provided, a polymer flooding injection-production process is dispersed on the aspect of time and on the aspect of space, and error caused by the dispersion is compensated by utilizing an iteration method. For the injection optimized problem that the slug length is variable, an iteration dynamic plan numerical solving algorithm with variable monitor is provided, so the injection concentration and the slug length can be simultaneously optimized. The method is implemented in the well groups of Gudong eighth area of Shengli Oil Field, and the economic benefit of the polymer flooding is remarkably improved.

The polymer is an expensive chemical product, and the injection of the polymer is a long and complex process. In order to scientifically decide a developing scheme for obtaining better economic benefit, the invention provides an optimal design method of a polymer flooding scheme which is based on an optimal control theory. The performance index of the optimal control problem can be the maximum net present value, the governing equation is the transfusion fluid mechanics equation of the polymer flooding process, the inequality constraints of the polymer amount and the injection density are considered, and an optimal control model of a polymer flooding distributed parameter system with the inequality constraints is built. By utilizing the Pontrjagin maximal principle, the necessary conditions of the polymer flooding optimal control problem are deduced, an improved subspace cutting off Newton method is provided and is used for solving the segment optimization problem with a decision variable boundary constraint during an SUMT (sequential unconstrained minimization technique) iterative process. The optimal design method provided by the invention is put into use in the second Gudong area of Shengli oil field, and the economic benefit of the polymer flooding is obviously improved.

The invention relates to a remote sensing image segmentation algorithm based on a convolutional neural network, and the algorithm comprises the steps: employing ResNet34 pre-training weight parametersas the depth guarantee of a network layer, so as to solve a problem of gradient disappearance when the network layer is deepened; a pyramid pooling module is adopted to aggregate context informationof different areas in the image so as to improve the capability of obtaining global information; designing a loss function by adopting a method of combining a cross entropy loss function and regularization constraints, wherein the cross entropy is used for judging the closeness degree of actual output and expected output in the multi-classification problem; the regularization term can reduce the complexity of the model so as to prevent overfitting and improve the generalization ability of the model. According to the method, the designed composite convolutional neural network algorithm is applied to semantic segmentation of the remote sensing image, so that the speed of information recognition in the remote sensing image and the capture rate of effective information are greatly improved, and the method is of great significance to information analysis of the remote sensing image.

The invention discloses a high-fitness interactive microgrid configuration method which is adaptive to various requirements including operation condition requirements, voltage loss requirements and distributed electrical connection requirements. According to the method, various distributed generators are fully considered, restriction is made before configuration according to the geographical conditions, wind and light conditions and operation site conditions of various regions, so that the configuration result has optimization meaning and application value. The optimization target is that the overall cost of a microgrid is minimum, and the life-cycle cost of the distributed generators, the cost of power distribution networks and benefits of the microgrid are also taken into account. The configuration method is based on the improved genetic algorithm, global searching ability is improved, and complicated power supply optimizing configuration can be achieved.

The invention discloses an electric wheel automobile chassis integration system and a multi-disciplinary optimization method thereof. The electric wheel automobile chassis integration system comprisesa differential power-assisted steering system, a wheel hub motor driving system and a semi-active suspension system, and four independent hub motors are used for power-assisted steering and power driving. Movement of a complete automobile in all directions affects each other, subsystems interact, steering feel, steering sensitivity, steering energy consumption, driving energy consumption and suspension smoothness serve as subsystem-level objective functions, the comprehensive performance of the complete automobile serves as a system-level objective function, parts of structural parameters ofthe chassis integration system serve as optimization variables, and multi-disciplinary optimization is performed on the electric wheel automobile chassis integration system by the aid of the topological decoupling multi-disciplinary optimization method and a glowworm optimization algorithm.

The invention discloses an online scheduling method for an electricity-heat comprehensive energy system based on near-end strategy optimization, and the method comprises the steps: constructing a real-time operation cost model of the electricity-heat comprehensive energy system for the intermittency of wind energy, the randomness of a real-time power market and the uncertainty of a user load; andthen, adopting a deep reinforcement learning method to convert a dynamic energy conversion problem into a discrete finite Markov decision process, and adopting an approximate strategy optimization algorithm to solve the decision problem, so that a system operator can adaptively determine the wind power conversion rate through online learning, the uncertainty of user load requirements, the flexibility of real-time electricity price and the uncertainty of wind power generation are also solved, and the income maximization of the electricity-heat comprehensive energy system is realized.

The invention discloses an interactive optimization method for improving an internal environment of a building. The method comprises the steps of building an indoor environment simulation module based on software for calculating fluid mechanics, and configuring relevant data interfaces; building a data interaction module to realize data interaction between a simulation procedure and an optimization algorithm; building an optimization module based on an evolutionary algorithm, and searching for the optimal values of the air supply temperature and the speed of an HVAC (Heating Ventilation Air Conditioning) by applying the evolutionary algorithm so as to enable the internal environment of the building to be optimal and the energy consumption of the building to be minimal. According to the method provided by the invention, different modules are formed by integrating a plurality of software by using an environment simulation interface and scientific calculating software, therefore, the influence on environment parameters caused by space distribution can be fully considered by the optimization method; furthermore, compared with an existing optimization method for the environment of the building, the interactive optimization method provided by the invention has the characteristics of good universality, high precision, etc.

The invention belongs to the technical field of ITS intelligent traffic systems, and particularly relates to an adaptive full-chain urban area network signal control optimization method. Traffic flow parameters are collected by utilizing a machine vision technology, traffic flow prediction is carried out based on a pre-trained traffic flow prediction algorithm by using obtained data, a microscopic traffic simulation model is constructed according to predicted traffic flow data, an original signal timing scheme and traffic network basic data, and a network-level signal optimization model is constructed. Active optimization is carried out on the network-level signal optimization model by adopting a Bayesian optimization algorithm so as to obtain an optimal signal timing scheme of the target area network. The method has good integration, an inner and outer circulation feedback closed loop is formed, an inner and outer circulation feedback mechanism can achieve interaction of a network signal optimization model and a microscopic traffic simulation model, it can be guaranteed that an optimization result scheme adapts to dynamic changes of the external environment, and then instantaneous dynamic optimization and long-term steady-state optimization are achieved.

The invention discloses a multi-supervision image super-resolution reconstruction method based on a generative adversarial network. The method comprises the following steps: registering slice images with different resolutions of the same slice; making a training data set by using the registered slice images; on the training data set, training a generative adversarial model by using a multi-supervision multi-stage generation model; and reconstructing the low-resolution image into a high-resolution image by using the trained generative adversarial model. According to the method, the image shot under the low-power lens is reconstructed into the high-resolution image, the imaging time can be saved, the hardware space for storing the image can also be saved, and blurring and artifacts of commonmethods on pathological data are overcome.

The invention provides an unmanned aerial vehicle path (UAV) planning method and device. According to the method, a topological network is constructed by taking ground base stations as nodes, a three-dimensional coordinate system is established by interaction between an UAV and the topological network, and ground coordinates of all nodes in the coordinate system and a coordinate projection of a flight path of the UAV on the ground are known. The method comprises determining a ground node access sequence according to the coordinates of the ground nodes or the number of neighbor nodes of the ground nodes; and searching path points in an effective transmission area of the ground nodes through a convex optimization method, and connecting given starting and terminal points and all the path points found in the path point searching step according to the access sequence to obtain an optimal path. The method can solve the problem that time minimization and path change influence each other in the prior art, the path of the UAV is optimized, the total task time is reduced to the greatest extent, and data collection and distribution between the UAV and the ground nodes are more efficient.

The invention relates to a polymer flooding production optimization method and system based on a Monte Carlo algorithm. The method comprises the steps that an optimum control function for oil reservoir optimization is generated by a simulator according to an oil reservoir value; a covariance matrix is generated by a solving model, an initial disturbance vector is generated nearby an optimum control variable by utilizing the Monte Carlo approach algorithm, the initial disturbance vector is processed by adopting the covariance matrix to generate a target disturbance vector; the Monte Carlo algorithm including the target disturbance vector is adopted to solve the optimum control function, and an optimum injection-production parameter corresponding to optimum exploitation benefit is generated. The calculation process of the method is simple and easy to achieve, the optimization efficiency is high, the convergence rate is high, the system is easily combined with any oil reservoir value simulator to calculate polymer flooding production, injection-production well production parameters and injection polymer concentration can be simultaneously optimized, an oil reservoir development effect can be remarkably improved, and a basis can be provided for scientific, reasonable and efficient oil field development. The method can be widely applied to the field of oil-gas field development.

The present invention discloses a virtual power plant scheduling optimization method based on Riemann integrals, in order to solve the optimal scheduling problem when the renewable power output and the load inside the virtual power plant are continuously changed. Considering that an aggregated unit of the virtual power plant comprises gas turbines, wind turbines, pumped storage power stations andloads, for the continuous change of the renewable power output and the load, by using the idea of the Riemann integrals and taking the continuity of variables over time into consideration, a virtual power plant scheduling optimization model based on Riemann integrals is established; and according to the definition of Riemann integrals, by dividing, making sum, and seeking the limit of the virtualpower plant scheduling time intervals, the integral problem is transformed into a limit summation problem, so that the problem can be solved. The method disclosed by the present invention can providean optimal scheduling scheme for a virtual power plant that adapts to the continuous change of the renewable energy output and load for a continuous period of time, provides effective support for thedecision maker to select the optimal strategy, and has certain engineering practical value.

The invention discloses a compact range collimator based on a plane wave synthesis technology and an optimization method thereof. The collimator is a plane wave comprehensive array antenna, based on the array near-field synthesis technology, the feed amplitude and the phase of each antenna unit in the plane wave comprehensive array antenna are controlled by adopting a particle swarm optimization algorithm to generate quasi-plane waves required by testing, and a dead zone is realized in a darkroom. The plane wave comprehensive array antenna adopts a rotational symmetry structure and is providedwith a parasitic unit and a decoupling structure; the plane wave comprehensive array antenna comprises an antenna array surface, the antenna array surface comprises a plurality of antenna units. According to the array near-field synthesis technology, plane wave synthesis is carried out by using a near-field directional diagram of each antenna unit in the plane wave synthesis array antenna, and array near-field synthesis is carried out by using the in-array near-field directional diagram considering the coupling effect between array elements. The collimator is advantaged in that the collimatoris high in convergence efficiency, good in optimization effect and high in design precision, and an ideal quiet zone is achieved in a large frequency band.

The invention discloses a multi-rotor aircraft noise suppression method based on phase angle control, and the method comprises the following steps that a system obtains the three-dimensional coordinates of a target noise reduction position, and transmits the position information to a noise prediction module; the flight control module takes the attitude angle and the speed as characteristic parameters and sends the characteristic parameters to the noise prediction module; the noise prediction module calculates noise information generated by a rotor at a target position through a rapid noise prediction method and sends the noise information to the phase angle optimization module; the phase angle optimization module obtains an optimal phase angle combination based on an optimization algorithmand sends the optimal phase angle combination to the phase synchronization control module; and the phase synchronization control module adjusts the phase angle positions of the plurality of rotors according to the received optimal phase angle combination in combination with the phase angle position information to realize noise reduction of the target position. According to the invention, the target function can be adjusted and optimized according to the noise reduction requirement, and compared with a passive noise reduction method, the method of the invention can control noise components ina targeted mode and has a better inhibition effect on multi-rotor-wing low-frequency-band noise.

The invention provides a light source and mask alternate optimization method based on an Abbe vector imaging model. According to the method, a graphics pixel value of a light source and transmissivity of an opening part and a light resistance part in a mask are set; variable matrixes omegaS and omegaM are established; a target function D is constructed to be a square of an Euler distance between a target graphics and an image in photoresist corresponding to the conventional light source and the conventional mask; and the alternate optimization process of the light source graphics and the mask graphics is guided according to the variable matrixes omegaS and omegaM and the target function D. Compared with the conventional method for independently optimizing the light source or the mask and synchronously optimizing the light source and the mask, the method has the advantage that the resolution of a photoetching system is effectively improved. Furthermore, the light source and the mask which are optimized by the method are applicable to small numerical apertures (NA) and NA which is more than 0.6. Moreover, according to gradient information of the optimized target function and a steepest speed reduction method, the efficiency for optimizing the light source graphics and the mask graphics is high.

The invention relates to a proportion variation particle swarm algorithm-based reactive power optimization method for a power distribution network. A method of proportion variation is added on the basis of a traditional particle swarm algorithm; when a certain condition is met during search for an optimal solution, variation can be carried out on a particle position according to the condition of each particle, so that the particle is prevented from falling into local optimum in the search process; in addition, the inertia weight of traditional linear change is also improved, the inertia weight is dynamically changed according to a feedback value, a w value becomes larger when the feedback value is greater than an average value, the particle search range is expanded, otherwise, the particle is searched in a relatively small range; and therefore, the whole particle space can be more accurately searched, and the convergence speed is also increased. Through verification, the proportion variation particle swarm-based reactive power optimization method for the power distribution network can be used for efficient management of the power distribution network when combined with an existing controllable device of the power distribution network, and higher convergence and optimization degree can be ensured during running.

The invention relates to a task parameter optimization method for a distributed iterative computing system, and belongs to the technical field of distributed data processing. The method comprises the following steps: firstly, acquiring operating data of a historical task in the distributed iterative computing system, and constructing a historical database; secondly, when performing task parameter optimization, performing primary filtering on significantly unrelated operating data in the historical database according to a constraint condition; thirdly, performing similarity calculation of a directed acyclic graph on operating data in the historical database corresponding to tasks to be optimized and operating data after the primary filtering, and performing secondary filtering on operating data with a similarity lower than a certain threshold; and finally, calculating and sorting results after twice filtering, and taking a task parameter corresponding to the operating data after sorting as a task parameter optimization result. The task parameter optimization method for the distributed iterative computing system provided by the invention can automatically perform task parameter optimization of the distributed iterative computing system, is a plug and play adaptive optimization method, and can obviously reduce the threshold of using the distributed iterative computing system by users.

The invention discloses a multi-wind-curtailment-area wind storage capacity configuration method based on a multi-objective optimization technology. According to the method, the problem of energy storage capacity optimal configuration during wind storage system planning can be solved so as to deal with the problems of prediction assessment and wind curtailment of the wind power plant in the multi-wind-curtailment area. According to the method, on the basis of the principle that wind curtailment of a power station is reduced to the maximum extent in a multi-wind-curtailment-area wind power plant, and by adopting an energy storage system to compensate wind power short-term prediction errors and meet the requirement for reducing fines of the power station, a charging and discharging mathematical description model of a multi-wind-curtailment-area wind power plant energy storage system is established; on the basis, an energy storage system scheduling strategy is coupled, the operation of awind storage system is simulated, the wind storage capacity is optimally configured by utilizing a multi-objective optimization model, and the method is analyzed and evaluated by applying wind storagesimulation of an example wind power plant. Experiments show that the method can realize effective energy storage capacity configuration and reach optimization expectation.

The invention discloses an interval uncertainty multi-objective optimization method for a chassis system of a hub-motor-driven car. The method disclosed by the invention takes into account many uncertain factors existing in the optimization design process of the chassis system of the hub-motor-driven car, wherein the uncertain factors are prone to cause an optimization result to be difficult to satisfy original design requirements. The interval uncertainty multi-objective optimization method adopts an interval model to describe an uncertainty variable, and separately transforms an uncertaintyobjective function and an uncertainty inequality constraint into a certainty objective function and a certainty inequality constraint. A conjugate gradient algorithm is combined to perform the interval uncertainty multi-objective optimization design of the chassis integrated system of the hub-motor-driven car.

The invention discloses an optimum design method of the dynamic performance of an aero-engine fastening junction surface. The optimum design method comprises the steps of firstly, determining the type, the specification, the assembly technology and the machining precision of adopted bolts; then selecting the aero-engine part fastening junction surface as a design domain, and using a rectangular assembly for description; then adopting several small four-node quadrilateral meshes to divide the design domain, then constructing a rectangular level set function, equalizing the design domain into amaterial model to obtain a unit equivalent rigidity matrix, assembling the equivalent rigidity matrix, and extracting a mass matrix of parts to obtain a kinetic equation and then a target function; then constraining center coordinates of the assembly in the design domain, then conducting iterative optimization, and finally, conducting rounding processing to thus obtain the junction surface joint form with the optimal dynamic performance. By means of the optimum design method of the dynamic performance of the aero-engine fastening junction surface, through optimal design of the junction surface, the optimal solution of the relevant dynamic performance is solved, and a clear design boundary is obtained.