34 results about "Global optimization problem" patented technology

The invention discloses an electric vehicle collaborative optimization scheduling method based on a'network-station-vehicle 'layered architecture. The method comprises the following steps: constructing a double-layer distributed scheduling model; in the upper-layer model, modeling is carried out on aggregation flexibility of the charging stations, and an aggregation charging and discharging plan of each charging station is formulated by taking maximization of the total benefit of the system as a target; in the lower-layer model, the charging and discharging behaviors of the individual electric vehicle are scheduled by following an upper-layer plan instruction and taking the comprehensive satisfaction degree of a vehicle owner as a target; a second-order cone programming relaxation technology is adopted to convert a non-convex power flow constraint equation in the upper-layer model into convex constraint; decomposing the global optimization problem into a plurality of independent sub-problems which can be solved by each charging station in parallel by adopting an improved Lagrange relaxation dual method; and iteratively updating the Lagrangian multiplier through a sectional self-adaptive step length strategy until the algorithm is converged and a global optimal scheduling plan is generated. According to the method, the scheduling effect close to centralized optimization can be obtained while the privacy of the user is protected; and the problem of curse of dimensionality is effectively solved.

The invention discloses a magnetic source target positioning method, which belongs to the technical field of geophysical exploration, is used for positioning a magnetic source target, and comprises the following steps: collecting and preprocessing plane grid magnetic anomaly data, constructing a simplified Euler deconvolution model under the constraint of derivative features, segmenting observation data by using a local sliding window, and resolving to generate an Euler solution point cloud. The method comprises the following steps: converting a selection problem of an equivalent scale factor into a one-dimensional global optimization problem, adaptively searching an optimal solution through a Bayesian optimization method, identifying a magnetic source high-probability solution point and a noise solution point by using a density-based spatial clustering method, introducing a density penalty term to strengthen a high-probability solution constraint, and constructing a final positioning target function. And outputting the optimal position solution of the magnetic source. According to the method, a traditional Euler deconvolution model is simplified, noise interference is effectively suppressed, the stability and robustness of the inversion process are improved, uncertainty caused by manual parameter selection is avoided, and the positioning precision is improved.

The invention provides a rapid trajectory planning and guidance method for a rocket landing section based on flight airspace grid division. The method mainly comprises the steps of establishment of a rocket landing column coordinate system, flight airspace grid division, a sequence convex optimization model for solving an absolute optimal trajectory, online interpolation reduction of the absolute optimal trajectory, online sequence convex optimization trajectory optimization and polynomial guidance. The invention aims to solve the problem of high-precision fixed-point landing of a reusable carrier rocket power landing section, three-dimensional motion of a rocket landing process is decomposed into two-dimensional motion and tangential motion of a longitudinal plane by using a landing column coordinate system, the complexity of an optimization model is reduced, an interpolation table is established by using a flight airspace grid division method, and the reliability of the model is improved. According to the method, the absolute optimal trajectory is rapidly obtained on line, the global optimization problem is converted into the local optimization problem, the solving scale of an optimization model is reduced, and the real-time performance of online optimization is improved while various constraints and fuel optimization are met.

The application relates to the field of intelligent control and information technology, and provides a natural gas generator set power device multi-working condition prediction control method. The application divides different intervals based on torque, designs optimal control strategies in each interval, converts a global optimization problem into a local optimization problem, and the algorithm idea is to guide a local minimum value optimization algorithm through the global optimization, and finally, better results can be achieved in control.

The application provides a kind of R&D resource visual scheduling method and system, method includes: based on original data, including several task entities and several resource entities are constructed, the matching degree between task entity and resource entity is obtained;The static business value score corresponding to the task entity, the time urgency factor, the task dependence coupling strength are obtained, and then the priority coefficient is obtained;Based on task entity and resource entity, decision variable is constructed, and final multi-objective optimization function is constructed, resource scheduling scheme is obtained based on final multi-objective optimization function, and resource scheduling scheme is displayed through visual board.The matching degree and priority coefficient are combined, the final multi-objective optimization function is constructed, the subjective choice of weighing between multiple optimization targets in traditional method is converted into global optimization problem, and the automatic and intelligent resource scheduling strategy weighing is realized.

The application discloses a power distribution network voltage coordination control method based on an event triggering mechanism and belongs to the technical field of power distribution control. The method comprises the following steps: a global optimization model is constructed; the power distribution network is divided into multiple sub-regions, coupling constraints of each region are established, and the global optimization problem is decomposed into regional optimization problems; a pre-designed ETM-DAL-ADMM algorithm is adopted, a Lagrange function of the regional optimization problem is linearly processed and dynamically adjusted in a penalty coefficient, local decision variables of each region including boundary variables are updated, information exchange between regions is carried out according to a pre-set event triggering mechanism, and final local decision variables are iteratively solved; and reactive power outputs of distributed generators in each region are adjusted according to the final local decision variables, so that the voltage of the power distribution network meets the constraints and branch loss is minimized. The application can solve the problem that the existing control method cannot guarantee the stability of the power distribution system when the communication capability is insufficient.

The invention discloses a high-fidelity identification method and system for a pump and a pipe network system based on active resistance modulation, and the method comprises the steps: firstly, manually setting at least two pipeline resistance states with obvious differences through adjusting devices such as valves, carrying out the frequency conversion operation of a pump in each state, and collecting a plurality of groups of pressure and frequency data; then, an integrated system physical model is constructed, and the model forcibly uses uniquely shared pump performance parameters and pipe network static lift parameters to synchronously fit all data clusters in different resistance states; and finally, by solving the global optimization problem of the model, identifying a real performance curve of the pump, the static lift of the pipe network and the independent resistance coefficient in each state at one time with high precision. According to the method, a flow sensor is not needed, the ill-conditioned problem of a passive data analysis method is solved, on-site in-situ calibration comparable with laboratory precision is achieved, and a reliable basis is provided for construction of a digital twin model and a virtual flowmeter and system health assessment.

The invention provides a hierarchical collaborative optimization and settlement method for large-scale industrial demand response, and relates to the technical field of power systems, and the method comprises the steps: dividing large-scale industrial demand response resources into a plurality of hierarchies, and constructing an optimization model; disassembling a global optimization problem into local sub-problems, defining a consistency variable to constrain global consistency, and generating a scheduling instruction through iterative calculation; acquiring actual response data in response to the scheduling instruction, and performing hash operation and digital signature processing on the actual response data to generate a signature abstract; verifying the signature abstract by using an intelligent contract, executing real-time settlement after verification is passed to obtain real-time settlement data, and summarizing to obtain periodic settlement data; and calculating a performance index according to the scheduling instruction and the actual response data, and performing numerical correction on the real-time settlement data and the periodic settlement data by using the performance index to obtain adjusted settlement data. According to the invention, rapid scheduling and convergence of large-scale resources are realized, and the enthusiasm and fairness of users participating in scheduling are effectively improved.

The invention relates to the technical field of energy internet and electricity-hydrogen coupling, in particular to an electricity-hydrogen coupling intelligent management and control platform. The platform comprises an energy input module, an electricity-hydrogen conversion module, a hybrid energy storage module and an intelligent management and control center. The distributed optimization method is adopted, the global optimization problem is decomposed into a plurality of micro-grid sub-problems, interaction is carried out only through boundary information, the calculation complexity is remarkably reduced while privacy of all micro-grids is protected, an electrolytic cell serves as a flexible load to be connected into the micro-grids, the operation power is improved in the low-electricity-price period, the hydrogen energy yield is increased, and the energy efficiency is improved. The operation state index is monitored in real time, the power distribution of the electrolytic cell is dynamically adjusted, that is, the operation power is automatically reduced and the equipment overload risk is prevented when abnormal fluctuation of the power is detected, and the method is suitable for an electrolytic cell control strategy of a wind and light power supply with strong fluctuation; the service life of the electrolytic bath is prolonged, and the hydrogen production cost of the system is reduced.

This invention discloses a multi-agent collaborative optimization scheduling method for medium- and low-voltage distribution networks under extreme scenarios, relating to the field of energy dispatching technology. The method includes: S1, unified modeling based on probability distribution, which constructs random state vectors and uses probabilistic statistical methods to uniformly characterize the random characteristics of different types of resources; S2, extreme scenario generation based on conditional variational autoencoders, which generates a set of scenario samples satisfying conditional distributions by learning the mapping relationship between historical meteorological data and operating states; S3, collaborative optimization based on Lagrange price signals, which achieves multi-agent collaborative decision-making driven by constructing a revenue function coupled with the scenario; S4, multi-agent coupled modeling based on graph neural networks, which constructs a distribution network topology graph and uses graph convolution operations to aggregate features of node states and extract the interaction relationship between electrical distance and power; and S5, a distributed optimization solution method based on the alternating direction multiplier method, which decomposes the global optimization problem into sub-problems of each agent.

The invention discloses a high-precision hand-eye calibration algorithm based on nonlinear optimization. The algorithm comprises the following steps: S1, data acquisition; s2, constructing a rotation matrix error model and performing precision analysis; s3, formulating an optimal calibration track rule according to the error model; s4, quaternion parameterization is used, and a global optimization problem is established; s5, solving a global optimization problem through a convex relaxation technology; and S6, outputting and verifying a calibration result. The invention relates to the technical field of robot vision and automatic control, and has the beneficial effects that a rotation matrix error analysis model is constructed (step S2), and the internal mathematical relationship between the hand-eye calibration precision and the motion track parameters (such as the rotation angle and the rotation axis included angle) of the mechanical arm is quantitatively revealed for the first time.

The invention relates to the technical field of virtual power plants, in particular to a virtual power plant energy scheduling method and system based on distributed optimization, and the method comprises the steps: constructing a distributed energy node state sensing network, collecting the operation state data and load prediction data of each distributed energy unit in real time, and transmitting the data to a distributed optimization scheduling center; based on an improved distributed model predictive control algorithm, establishing a multi-target energy scheduling optimization model, and decomposing a global optimization problem into local sub-problems of each node; parallel iterative solution of each sub-problem is realized through a distributed collaborative solution mechanism, and a scheduling scheme is dynamically corrected; and based on the real-time operation deviation and the external environment change, adaptively adjusting a scheduling strategy and issuing the scheduling strategy to each energy unit for execution. According to the invention, distributed energy resources are integrated through a distributed optimization architecture, efficient collaborative scheduling of virtual power plant energy is realized, the energy utilization rate is improved, the operation cost is reduced, and the system operation stability and flexibility are enhanced.

The present application belongs to one-dimensional signal processing technical field, and particularly relates to a Duffing oscillator weak signal enhancement method with automatic parameter optimization. The present application takes a double-stage parallel Duffing random resonance system as a mixed signal processing model, uses a quantum particle swarm algorithm with outstanding performance in optimization solution as a parameter optimization algorithm, converts the double-stage parallel random resonance parameter adjustment problem into a multi-parameter global optimization problem, and performs signal enhancement processing on the signal to be enhanced input into the system according to the best parameters, and outputs the enhanced signal. The present application uses parameter self-adaptive optimization to quickly obtain the best parameters of the double-stage parallel Duffing random resonance system, greatly improves the efficiency and precision of the system best parameter setting, and can quickly and high-quality output the enhanced useful signal.

This invention provides a rapid trajectory planning and guidance method for rocket landing phase based on flight airspace grid partitioning. It mainly includes establishing a rocket landing column coordinate system, flight airspace grid partitioning, a sequential convex optimization model for solving the absolute optimal trajectory, online interpolation to restore the absolute optimal trajectory, online sequential convex optimization trajectory optimization, and polynomial guidance. This invention aims to solve the problem of high-precision fixed-point landing of reusable launch vehicles' powered landing phase. By utilizing the landing column coordinate system, the three-dimensional motion of the rocket landing process is decomposed into two-dimensional motion in the longitudinal plane and tangential motion, reducing the complexity of the optimization model. An interpolation table is established using the flight airspace grid partitioning method, enabling rapid online acquisition of the absolute optimal trajectory. This transforms the global optimization problem into a local optimization problem, reducing the solution scale of the optimization model and improving the real-time performance of online optimization while satisfying various constraints and fuel optimization.

The invention discloses a rural electricity-gas-heat integrated energy system distributed collaborative optimization scheduling method under a weak communication condition, and relates to the technical field of integrated energy system scheduling, and the method comprises the steps: collecting energy parameters and data in a rural integrated energy system, carrying out the function division, and dividing the energy parameters and data into subsystems; modeling is carried out on the divided subsystems, constraint conditions are established at the same time, and a global optimization objective function is established; decomposing a global optimization problem into independent sub-problems of subsystems based on the constructed subsystem model and constraint conditions; establishing a distributed collaborative optimization scheduling model by taking the operation cost minimization of each subsystem as a target and considering the coupling constraint with other subsystems at the same time; and performing distributed iterative solution based on the distributed collaborative optimization scheduling model. According to the method, the overall optimization problem is divided into a plurality of sub-problems through the alternating direction multiplier method, each sub-system can independently complete calculation locally, only limited information interaction is needed, and therefore the adaptability of the system is remarkably improved.

The application discloses an AI-based Beidou conformal microstrip antenna multi-target collaborative design optimization method, relates to the technical field of antenna optimization design, and comprises the following steps: acquiring surface feature vectors of geometric models of multiple typical carrier surfaces; performing clustering analysis to form multiple surface feature categories; obtaining an initial design model covering multiple performance trade-off points and storing the initial design model in association with corresponding surface feature categories; determining the surface feature category to which a target design surface belongs; calling the associated initial design model; and based on the initial design model, solving final antenna design parameters by using a fast optimization algorithm. The application has the advantages that: by intelligently clustering multiple typical carrier surfaces in advance, an initial design model library covering different surface feature categories is constructed, when a new target design surface is faced, a corresponding high-quality initial design can be quickly matched and called, and thus a complex global optimization problem is converted into efficient local fine adjustment.

The invention relates to the technical field of intelligent building and household energy management, and provides a household wind and light storage system two-stage optimization method for thermal comfort and energy efficiency collaboration, and the method comprises the following steps: S1, constructing a household wind and light storage system high-fidelity model, and precisely representing the coupling relation and operation constraint of electric and thermal multi-energy flow; s2, constructing a multi-objective optimization problem taking minimization of daily electricity consumption cost, maximization of renewable energy source self-use rate and precise maintenance of thermal comfort as the core; and S3, designing and solving a two-stage collaborative optimization framework: decomposing a global optimization problem into two successive stages of day-ahead scheduling and real-time control to realize efficient solving. According to the method, the source-load uncertainty is effectively stabilized, the power consumption cost is reduced, the renewable energy sources are efficiently consumed, the indoor thermal comfort degree is maintained at high precision, and the performance is remarkably superior to that of a traditional method.

An initial alignment method based on speed optimization, a combined navigation system, and a GNSS receiver. The method comprises: in a dynamic window involved in initialization, moving a combined navigation system on the basis of a preset rule (100); acquiring a relational expression between a GNSS velocity solution and acceleration and an angular velocity observation between adjacent epochs in the dynamic window (101); on the basis of the preset rule and the relational expression based on the GNSS velocity solution, acquiring a rough initial value of a parameter to be estimated of an initial attitude parameter (102); and using the rough initial value to construct a global optimization problem, and solving the global optimization problem to acquire the initial attitude parameter (103). The method can realize fast initial alignment of GNSS / INS in complex environments, and can provide reliable and accurate initial alignment results in complex and obstructed environments.

The invention discloses a sensor network sound field reconstruction method based on a graph signal and distributed optimization. The method comprises the following steps: constructing a sensor network of a target area into an undirected weighted graph; according to the method, a physical topological structure of a sensor network is combined with spatial smoothing prior of a sound field by introducing a graph Laplacian consistency regular term; the overall spatial correlation of the sensor network is utilized to suppress estimation abrupt change caused by local noise and measurement errors, so that the accuracy and continuity of a reconstruction result and the robustness in a complex noise environment are remarkably improved; according to the method, outer layer iteration is carried out through an alternating direction multiplier method to update a sound field sparse coefficient vector, an auxiliary variable and a dual variable at each node, and a complex global optimization problem is decomposed into local sub-problems which can be processed in parallel at each node. And each node only needs to communicate with a direct neighbor, so that the bottleneck effect of a central node is avoided, and the communication overhead and calculation delay of the system are remarkably reduced.

The application provides a hierarchical collaborative optimization and settlement method for large-scale industrial demand response, relates to the technical field of power systems, and comprises the following steps: dividing large-scale industrial demand response resources into multiple levels and constructing an optimization model; decomposing a global optimization problem into local sub-problems, defining consistent variable constraints for global consistency, and generating scheduling instructions through iterative calculation; obtaining actual response data in response to the scheduling instructions, performing hash operation and digital signature processing on the actual response data to generate a signed digest; verifying the signed digest by using a smart contract, executing real-time settlement after verification to obtain real-time settlement data, and summarizing to obtain periodic settlement data; calculating performance indicators according to the scheduling instructions and the actual response data, and performing numerical correction on the real-time settlement data and the periodic settlement data by using the performance indicators to obtain adjusted settlement data. The application realizes fast scheduling and convergence of large-scale resources, and effectively improves the enthusiasm and fairness of users participating in scheduling.

The invention discloses an amplitude-comparison and phase-comparison hybrid direction finding system and method based on a particle swarm optimization algorithm. The method comprises the following steps: modeling the problems of searching an optimal receiving channel group and detecting the center frequency of a frequency band; initializing, iteratively optimizing and terminating the problem to obtain an optimal receiving channel group and a center frequency of a detection frequency band; a correlation interferometer direction finding method is adopted to obtain a target arrival azimuth angle estimation precise value and a pitch angle estimation precise value; and judging whether multiple values exist in the azimuth angle estimation precise value and the pitch angle estimation precise value when the target arrives, namely phase ambiguity occurs or not. According to the method, a fitness function fusing amplitude and phase information is constructed, a direction finding problem is converted into a global optimization problem, the global search capability of a particle swarm optimization algorithm is utilized, an optimal receiving channel group is iteratively searched, the center frequency of a detection frequency band is detected, coarse value estimation of amplitude comparison direction finding and fine value correction of phase comparison direction finding are combined, phase ambiguity is eliminated, and the optimal receiving channel group is obtained. And high-precision and high-robustness direction of arrival estimation is realized.

The invention discloses an air route network optimization method and device, and belongs to the technical field of air route network optimizing.The method comprises the steps that after airport operation data are obtained, a constructed multi-target air route network optimization model is solved through a fuel oil backlog and carbon emission control algorithm; the method comprises the following steps of: after acquiring a fuel pressure accumulation amount, a total fuel pressure accumulation amount and a balance parameter of each airline network, converting an objective function into a non-time-varying airline network optimization problem by adopting a Lyapunov optimization technology, and then solving by adopting a heuristic algorithm to obtain an airline network optimization scheme. The balance parameters are used for ensuring balance among passenger transportation cost, pollutant discharge and two-station stop cost; and optimizing the air route network according to the air route network optimization scheme. According to the method, a time-varying problem is converted into a non-time-varying problem, that is, a global optimization problem is converted into a local optimization problem, so that the model is effectively solved, and the reduction of operation cost and pollutant emission and the improvement of passenger satisfaction are realized.

The invention discloses a virtual power plant distributed cooperative regulation and control method and system based on a consensus ADMM and a graph convolutional network, and the method comprises the following steps: constructing a global optimization problem of aggregation regulation and control of a virtual power plant, the target being to minimize the total operation cost, and constraints comprising total power balance and individual power limitation; decomposing the global optimization problem into local optimization sub-problems of the distributed resource intelligent terminals by adopting a consensus ADMM framework; each intelligent terminal generates a local power decision function through a pre-trained graph convolutional network GCN based on the local state information, the dual variables and the neighbor information; the intelligent terminals exchange dual variables and consistency variables through a communication coordination layer, and iterative updating is carried out until convergence; and the output execution layer converts the final decision into an actual control instruction and issues and executes the actual control instruction. By fusing the consensus ADMM and the graph convolutional network, the method shows significant performance advantages in multiple aspects in virtual power plant distributed cooperative regulation and control.

The invention provides a cable intelligent wiring method and device based on a heterogeneous graph neural network and deep reinforcement learning, and a related product, and relates to the technical field of industrial automation design and artificial intelligence. According to the method, an intelligent and self-adaptive cable wiring scheme is provided, a heterogeneous graph is adopted to represent a collaborative architecture of learning and sequence decision optimization, and a cable wiring problem is converted into a global optimization problem on the graph; a heterogeneous graph structure containing five types of nodes and eight types of relation edges is designed, and various complex relations between signals can be explicitly expressed; compared with the prior art, the method has the advantages that signals and associated entities thereof are modeled into a graph structure in a unified mode, the multi-hop transfer relation can be expressed, and constraint conditions serve as learnable graph nodes; and through a sequence decision optimization algorithm, a prospective planning capability is provided, a local optimal trap is avoided, and a global optimal boxing decision is realized.

The invention provides a global gradient optimization method based on score matching, and aims to solve the problem of global optimization through a deterministic gradient method. Comprising the steps of providing a general method for constructing a micronized hierarchical target optimization function, providing a global gradient optimization algorithm based on the hierarchical target function, applying the algorithm to the field of generative modeling, and providing a generative modeling method based on global optimization. Compared with a traditional gradient descent algorithm, the global gradient optimization method based on score matching overcomes the defect of local optimum, and can be used for solving the global optimization problems of non-convex, non-differentiable, discrete, complex constraint and the like. Compared with a traditional generation normal form based on probability sampling, the generation normal form based on global optimization provided by the invention can obtain a better reasoning effect theoretically.

The application discloses a kind of data center cluster distributed collaborative regulation method and regulation system.Method includes: the collaborative optimization model of data center cluster with the minimum system total cost as goal is established;Based on collaborative optimization model, the power-computing coordination optimization submodel of each data center local is constructed;Based on submodel, the collaborative optimization model of inter-data center computing load balance and transmission balance is constructed;Based on collaborative optimization model, global optimization problem is decomposed into local optimization subproblem and global coordination subproblem which can be solved in parallel;Based on local optimization subproblem and global coordination subproblem, the distributed iterative solution algorithm of alternating direction multiplier method is designed;Based on the result of distributed iterative solution algorithm, judge whether to continue iteration.The method can reduce the total operating cost of data center cluster, effectively protect the commercial privacy of each party, while providing flexible load regulation capability for power system, supporting high proportion of renewable energy consumption.

The invention discloses an antenna housing design method, and belongs to the field of antenna design. Aiming at the problems of high design complexity and high cost of the existing radome, the invention provides the radome design method, which comprises the following steps of: determining the working broadband range of the radome, and dividing the working broadband range into a plurality of working broadband sub-frequency bands; solving a design parameter corresponding to each working broadband sub-band; performing intersection solution on the plurality of design parameters to obtain an optimal solution; and verifying whether the optimal solution accords with global performance or not. According to the invention, the working broadband range is decomposed into a plurality of sub-bands, optimization design is independently carried out in each sub-band, and finally, an intersection optimal solution capable of meeting the performance requirements of all the sub-bands is searched, so that the high-performance and high-robustness design of the antenna housing in the whole broadband is realized; and a complex high-dimensional global optimization problem is decomposed into a plurality of simple low-dimensional sub-problems, so that the optimization efficiency and the success rate are greatly improved, and the computing power cost is reduced.

The application discloses a magnetic source target positioning method, belongs to the technical field of geophysical exploration, and is used for magnetic source target positioning, comprising the following steps: collecting plane grid magnetic anomaly data and preprocessing, constructing a simplified Euler deconvolution model under derivative characteristic constraint, segmenting observation data by using a local sliding window, solving and generating an Euler solution point cloud, converting an equivalent scale factor selection problem into a one-dimensional global optimization problem, adaptively searching for an optimal solution by using a Bayesian optimization method, identifying magnetic source high-probability solution points and noise solution points by using a density-based spatial clustering method, introducing a density penalty term to strengthen high-probability solution constraint, constructing a final positioning target function, and outputting a magnetic source optimal position solution. The application simplifies a traditional Euler deconvolution model, effectively suppresses noise interference, improves stability and robustness of an inversion process, avoids uncertainty caused by artificial selection of parameters, and improves positioning precision.

This invention provides a method and apparatus for extracting key information from document images. The method includes: determining a preset number of candidate key values ​​for each key based on OCR recognition results; determining a total score for candidate paths for all keys based on the scores of the candidate key values ​​selected for each key; and determining the target key value selected for each key when the total candidate path score is the highest, to obtain the corresponding information extraction result. The score for each key selecting a corresponding candidate key value includes at least a score determined by the positional relationship between the key and the candidate key value. This method effectively avoids the error-prone problems caused by fixed keywords in rule-based information extraction methods. Furthermore, this method comprehensively considers all candidate key value selections, effectively utilizing the interrelationships and mutual exclusions between key information, thus achieving high accuracy and solving the global optimization problem of similar entity information extraction.

The invention belongs to the field of power grid dispatching, and particularly discloses a projection-enhanced distributed aggregation gradient tracking virtual power plant deaggregation method. According to the method, a global optimization problem which aims at minimizing the sum of local operation costs of all distributed energy sources and tracking an aggregation power instruction is constructed, and global optimization of a scheduling scheme is ensured. In distributed iteration, distributed energy sources update output values in parallel on the basis of neighbor variables, local feasible region constraints are strictly met through projection operation, and engineering feasibility is guaranteed; and simultaneously updating an aggregation variable and a gradient tracking variable to realize distributed estimation and propagation of global gradient information, so that local output adjusts and tracks a global optimal direction. And during iterative convergence, outputting the output value of each distributed energy source, so that a depolymerization result meets all local operation limits and approaches to global cost optimization. Compared with the prior art, the applicability and reliability of virtual power plant depolymerization are improved, optimization software or a solver does not need to be introduced in the whole process, and practical engineering application is guaranteed.