344 results about "Iterative method" patented technology

A park-factory-coordinative optimized scheduling method based on a Jacobi iteration method, which belongs to the technical field of energy scheduling optimization. A park-factory hierarchical game model is established, and by means of coordination and optimization between energy scheduling under a complex production constraint and production management and control that takes profit optimization as the goal, an optimal price regulating strategy and an optimal production scheduling strategy are obtained, such that a lower industrial park energy consumption and a higher factory production profit are achieved. A coordinative optimized scheduling model for a park control center and a factory management and control center is a multi-variable multi-constraint complex optimization issue; the solution of a high-order partial differential equation set involved in the model is calculated, and a cyclic iterative optimization algorithm is designed on the basis of a Jacobi iteration matrix, thereby greatly reducing the operation complexity; moreover, such a cyclic iterative algorithm requires a smaller storage capacity of a computer, and is thus suitable to be applied in practice by an engineer.

Lowest cost usage scheduling of an energy system during a time interval of interest is achieved by utilizing two components of learning and optimization. First, a number of learning approaches including Linear Least Square Regression (LLSR), Auto Regressive Integrated Moving Average (ARIMA), and Multi-Layer Perceptron Deep Learning (MLPDL) are used to forecast energy production and storage of the energy producing components of a given group of energy components as a function of historical production data and weather data collected from weather and geo reports. Then, an optimization problem is formulated to create the optimal schedule of energy usage during time intervals of interests for the given group of energy components subject to scheduling and equipment constraints. Two alternative solutions, namely Simplex Branch and Bound (SBB) and Interior Point Branch and Bound (IPBB) are presented to solve the optimization problem. Accordingly, integrated iterative methods, programs, and systems are described aiming at minimizing the cost of energy consumption for the given group of energy components within time intervals of interest.

Embodiments of the present application provide a bidirectional iterative method for a pre-trained model and downstream sequence tasks, a device, and a medium. In the embodiments of the present application, a novel soft-prompt-based fine-tuning training approach is provided, and each round of fine-tuning training comprises both a feedback link from historical downstream tasks to a pre-trained model and an adaption link from the pre-trained model to a current downstream task. For each current round, on the basis of the feedback link, the pre-trained model undergoes a first fine-tuning by using historical downstream tasks that have appeared prior to the current round, so as to enhance the capability of the pre-trained model; and on the basis of the adaption link, the already fine-tuned pre-trained model undergoes a second fine-tuning by using the current downstream task of the current round, so as to train a task model more adapted to the downstream task. Thus, the pre-trained model can be practically applied to downstream tasks, and improved model performance can be achieved especially in few-shot scenarios.

An iterative method for monitoring a computing device characterized by metric data to be monitored, including, for each iteration, of collecting metric data over a predetermined interval of time, detecting a seasonality pattern of said metric data over said predetermined interval of time, determining an interval-specific model representing the detected seasonality pattern, calculating modelled data using said determined model and the collected metric data, comparing the calculated modelled data with the collected metric data to calculate a score characterizing the difference between the calculated modelled data and the collected metric data, calculating an anomaly likelihood for each data of the collected metric data using the calculated score, detecting an anomaly on a data when probability that the value of said data is an anomaly is greater than a predetermined threshold.

The invention provides a numerical simulation method and system based on rock mass fracture extension and seepage coupling, and belongs to the field of rock mechanics. Rock is regarded as a set composed of particles and mechanical contact between the particles, and a triangulation network is constructed based on the circle center of each particle; according to the geometrical characteristics of the triangulation network, a seepage pipeline network containing the interior and the periphery is built, and network nodes and boundary nodes are determined; calculating the permeability coefficient according to the particle positions on the two sides of the pipeline, and forming a penetrating seepage channel and fracture network by taking the pipeline which is not 0 and penetrates through and is connected with boundary nodes as a passage pipeline; and constructing a flow conservation equation set based on network nodes, solving pore pressure by using a Jacobian iteration method, calculating fracture pressure, and applying the fracture pressure to the model to realize coupling of a stress field and a seepage field. According to the method, the advantages of particle discrete elements in calculation of rock mass fracture extension and the advantages of a pipeline network method in calculation of fracture seepage are fully exerted, and accurate simulation of rock mass fracture initiation extension, seepage channel recognition and seepage field calculation is achieved.

The invention discloses a collaborative optimization method for a power system containing multiple virtual power plants based on a feasible region equivalent projection theory. The method comprises the following steps: establishing a main network operation model; modeling wind and light uncertainty by adopting an opportunity constraint optimization method and establishing a virtual power plant operation model; the chance constraint is converted into a deterministic constraint through quantiles; giving an objective function of virtual power plant-main network collaborative optimization, and establishing an integrated dispatching model of a power system containing multiple virtual power plants; eliminating internal variables of the virtual power plant in the target function by using an upper mirror graph theory; describing a multi-period power-cost feasible region projection of the virtual power plant at the common coupling point based on a feasible region equivalent projection theory and a vertex projection convex hull search method; and establishing a VPP-major network collaborative optimization model based on feasible region equivalent projection, and calling a GUROBI solver for solving in a Matlab2021a environment. According to the method, the virtual power plant-main network collaborative optimization model in multiple periods is solved in a non-iterative mode, the solving time is greatly shortened compared with a traditional iterative method, and the method has the same optimality as the traditional method.

The invention discloses a hydrogeological parameter space-time heterogeneity geostatistics inversion method. The method comprises the following steps: constructing a to-be-solved hydrogeological parameter column vector based on a geologic model and a'space-time equivalence 'principle; a priori covariance matrix of partition parameters in a time domain is assembled in a blocking mode; carrying out dimension reduction on the covariance matrix based on a principal component analysis method; carrying out parameter sensitivity matrix approximate calculation, solving a collaborative Kriging equation, and calculating an estimated value of a to-be-solved parameter by utilizing an iteration mode; and determining a target function, and performing iterative updating according to the target function to obtain a to-be-solved parameter column vector. The invention designs a partition geological statistics dimension reduction inversion method for describing the spatial-temporal heterogeneity of the hydrogeological parameters, and provides an efficient and accurate solution for describing the spatial-temporal heterogeneity of the parameters in a complex spatial-temporal process.

The invention discloses a three-level ANPC converter model prediction control method and device based on Newton iteration method vector optimization, and belongs to the technical field of converter control, and the method comprises the steps: building a mathematical model of a three-level ANPC converter under a three-phase static coordinate system based on the Kirchhoff theorem; performing forward Euler discretization on the mathematical model of the three-level ANPC converter to obtain a current prediction equation model; establishing a cost function according to the control target; performing online search on the optimal voltage vector according to a cost function; voltage vector synthesis is carried out, voltage vectors selected when the optimal voltage vector falls in 36 different small sectors, a voltage vector action sequence and action time distribution are obtained through calculation, and then the action time of a pair of redundant positive and negative small vectors in the voltage vectors is adjusted to carry out neutral point potential balance control. According to the invention, rapid optimal voltage vector search and three-level ANPC converter output current and neutral point potential control are realized.

The invention discloses a circular foundation pit diaphragm wall deformation calculation method and system based on a vertical elastic foundation beam. The method comprises the steps that S100, a design scheme and design parameters of a circular foundation pit are collected; s200, establishing a vertical elastic foundation beam calculation model of the circular foundation pit supporting structure, and inputting the design scheme of the circular foundation pit into the model; s300, dispersing the beam into a plurality of calculation points, dividing the excavation step into a plurality of load increment steps, and calculating each load increment step of the beam; and S400, based on the load increment step of the beam, calculating an accumulated deformation amount, and outputting the accumulated deformation amount. A bending equation of a vertical elastic foundation beam is adopted to establish a calculation model of a circular foundation pit supporting structure, an inner supporting structure is introduced into the bending equation of the beam, the excavation process is decomposed into a plurality of excavation steps and increment steps, an iteration method is adopted for calculation, and the soil pressure distribution and deformation process of each stage in the excavation process is accurately simulated. Therefore, the accuracy of deformation calculation of the circular foundation pit diaphragm wall is improved.

The invention discloses a fast scanning seismic travel time calculation method based on a process function equation, and relates to the field of seismic travel time calculation, and the method comprises the steps: firstly, reading a speed model and related parameter information; then carrying out seismic source initialization, setting a slowness fixed item, calculating a travel time fixed item, and setting a travel time disturbance item according to the position of a seismic source point; next, Gauss-Seidel iterative calculation is carried out, in four alternating directions, a Gauss-Seidel type iterative method is used for solving an eikonal equation of factorization, a proper travel time disturbance term is selected through a causal condition, and finally, a travel time calculation result meeting a termination condition is output. According to the high-precision seismic travel time calculation method based on the factorization eikonal equation, the problem of seismic source singularity is well solved, high precision is guaranteed, meanwhile, the algorithm has high calculation efficiency, and the accuracy of subsequent seismic exploration data processing is improved.

The invention relates to a method for controlling physical quantity of underwater explosion near a naval vessel based on a multiphase flow model, and belongs to an anti-impact technology of underwater explosion near the naval vessel. According to the method, an additional correction equation related to total mixed energy is created, and a model is expanded into seven equation models. A mixed energy correction equation is introduced, and a more accurate gas state equation is combined, so that the difference between the predicted thermodynamic states under the shock wave condition is solved; an algorithm program of an improved six-equation model is provided on an unstructured grid system, a homogeneous hyperbolic equation is solved by using a second-order MUSCL-Hancock format based on least square reconstruction and a Barth-Jespersen limiter and a two-phase flow HLLC Riemann solver, and then an instantaneous pressure relaxation equation is solved by using a Newton-Raphson iteration method. By improving the six-equation model in the diffusion interface method, the interface capture precision and the numerical algorithm stability are improved.

The invention discloses a water level prediction method based on an improved intelligent algorithm, a readable storage medium and a device, and belongs to the technical field of water level prediction.The method comprises the steps that normalization preprocessing is conducted on original hydrological data of multi-source information, and an original water level sequence is obtained; performing VMD decomposition on the original water level sequence to decompose the original water level sequence into modal components; inputting the modal component into an LSTM model to construct a prediction model; and inputting an original water level sequence of recent multiple days into the prediction model, generating a predicted value of a future date based on an iteration mode, and restoring to obtain real water level data through reverse normalization processing. A computer program is stored in the readable storage medium, and the method is carried out when the computer program is executed. The apparatus includes a memory, a processor, and a terminal program stored in the memory and operable in the processor. Water level influence factors can be comprehensively considered, multiple variables are fused, the water level condition closer to the reality is obtained through analysis, and the water level prediction precision is improved.

The invention belongs to the related technical field of cross-scale particle transport simulation, and discloses a particle transport discrete unified gas kinematics simulation method and system especially suitable for complex non-conservative collision operators. According to the method, high-precision compact reconstruction of a distribution function at a characteristic line is realized in a compact space of a single grid through a universal compact reconstruction method, and particle transport complex non-conservation collision implicit terms under time trapezoidal integration are effectively processed through a universal estimation-correction method. The defects of application limitation of an existing discrete unified gas kinematics method on a complex non-conservation collision term operator transport model or low calculation efficiency, error accumulation, numerical instability and the like existing in an existing repeated iteration method, a time splitting-discrete unified gas kinematics method and the like are overcome. And the problem that the calculation precision is relatively low when an existing interface reconstruction method based on a plurality of adjacent grids is used for solving a particle transportation problem with complex substance arrangement or relatively high heterogeneity is solved.

The invention belongs to the related technical field of power system optimization operation, and particularly relates to a non-iterative interconnected multi-microgrid system optimal power flow determination method based on a dimensionality reduction projection method, and the method comprises the steps: reconstructing a multi-microgrid centralized operation model, and obtaining a sub-problem of each sub-microgrid; taking a multi-dimensional variable space formed by all public variable sets in each sub-problem and a preset sub-microgrid operation cost variable as a projection space of the sub-problem; determining an initial vertex set of the projection of the convex hull corresponding to the original operation space constraint of each sub-problem in the projection space, and translating the coordinate origin to the interior of the convex hull enclosed by the vertex set; based on the convex hull and the outer normal vector of each boundary surface, circularly searching vertexes satisfying original operation space constraints outside each boundary surface in the projection space, and taking a space enclosed by a final vertex set as a projection space of the operation space constraints of the sub-problem; and reconstructing the sub-problem in the space and solving the sub-problem. According to the method, the optimal power flow can be solved under the condition that only public variables are disclosed.

The invention discloses an iterative satellite internet confrontation method based on momentum, and mainly solves the problem of low success rate of black box migration attack in the existing method. According to the method, a momentum item is introduced in an iterative optimization process of adversarial sample generation and is used for recording a historical direction of gradient updating. According to the momentum iteration method, the gradient updating direction in the adversarial sample generation process can be more stable, meanwhile, the adversarial sample is prevented from falling into a local extreme point of an optimization problem, the mutual restriction of the white box attack effect and the migration capability of the adversarial sample generated by the existing method is effectively relieved, and the black box migration attack success rate is improved.

The invention provides a financial research report generation method based on a large model. Comprising a data collection module, a data analysis module and a report writing module. The data collection module is used for receiving a predefined data collection task and an autonomously generated data collection task, then inputting the tasks into a deep search agent, and constructing a structured multi-modal memory for a search result for storage; the data analysis module constructs a data analysis agent, and realizes the generation of a structured multi-modal small report through cooperation of a financial calculation tool, a chart generation tool and a reference addition tool in an actor-reviewer iteration mode in which a large language model is matched with a visual language model; and the report writing module constructs a report writing agent to obtain a chapter first draft and then inputs the chapter first draft into the post-processing agent to obtain a final report. The problems that an existing deep research system and a large language model agent in the financial field are lack of financial field knowledge, limited in multi-modal ability, limited in research and report generation depth and comprehensiveness and the like are solved.

The application discloses a kind of gas-liquid two-phase flow numerical simulation method considering coalbed gas desorption lag, comprising the following steps: using well logging, indoor experiment means to obtain productivity calculation parameter;Desorption equation of matrix system is defined in combination with desorption lag effect, diffusion equation;Gas-water two-phase percolation equation of fracture system is defined, and desorption equation of matrix system, diffusion equation is combined, and double medium gas-water two-phase coupling flow equation set is formed;The porosity and permeability change of matrix system and fracture system in production process is calculated respectively, and the physical property change equation set of double medium is deduced;Using the numerical solver of COMSOL, the coupling equation set is solved using iterative method, and the gas production and water production are calculated.The application is simple to operate, can more scientific and reasonable assess the influence of coalbed gas desorption lag effect on gas well productivity, while greatly reducing the cost and time of gas well numerical simulation calculation, and also has important inspiration for other unconventional gas reservoir development affected by desorption lag.

The embodiment of the invention discloses a nonlinear system dynamic characteristic calculation method based on a structural dynamic modification method, and the method achieves the high-precision reduced-order modeling of a nonlinear structure through enabling a complex nonlinear effect to be equivalent to an additional stiffness field and combining with a structural dynamic modification algorithm. And based on the decoupled linear substructure modal space, a nonlinear correction operator is introduced, and the global dynamic response of the nonlinear structure is solved through a resolving method after modification. Structural vibration frequency, vibration shape and energy dissipation characteristics under a nonlinear effect can be efficiently obtained, and dynamic characteristics under different nonlinear amplitudes can be obtained. Compared with a traditional iteration method, the method has the advantage that the calculation complexity is remarkably reduced by equivalently linearizing a nonlinear part.

The invention discloses a gas-liquid two-phase flow numerical simulation method considering coal bed gas desorption lag. The method comprises the following steps: acquiring productivity calculation parameters by using logging and indoor experiment means; defining a desorption equation and a diffusion equation of the matrix system in combination with a desorption hysteresis effect; defining a gas-water two-phase seepage equation of the fracture system, and forming a dual-medium gas-water two-phase coupling flow equation set in combination with a desorption equation and a diffusion equation of the matrix system; the porosity and permeability changes of a matrix system and a fracture system in the production process are calculated respectively, and a physical property change equation set of the double media is deduced; and solving the coupling equation set by using a numerical solver of COMSOL and adopting an iteration method, and calculating to obtain gas production and water production. The method is easy and convenient to operate, the influence of the coal bed gas desorption lag effect on the gas well productivity can be evaluated more scientifically and reasonably, meanwhile, the gas well numerical simulation calculation cost and time are greatly reduced, and important inspiration is also provided for development of other unconventional gas reservoirs influenced by desorption lag.

The invention discloses a series robot calibration algorithm based on an index product, and the algorithm comprises the following steps: A1, building an index product motion model according to a robot joint axis and coordinate points on the axis through an index product modeling method, a2, differentiating the positions of the robot hand-eye, the kinematics and the calibration object relative to the robot base coordinate system to obtain a hand-eye, kinematics and calibration object position error model based on an exponential product; b1, obtaining an initial value of a robot system hand-eye matrix # imgabs0 # and an initial value of a calibration object relative to the position Bp of a robot base through a closed solution method based on a Kronecker product; b2, taking the difference between the predicted value and the measured value of the base position of the calibration object under the robot base as a target function; and B3, establishing an optimization algorithm by taking the hand-eye, kinematics and calibration object position parameters of the robot as to-be-optimized parameters, and carrying out iterative optimization calculation on the hand-eye, kinematics and calibration object position parameters of the robot by using a Newton iteration method.

The embodiment of the invention provides a calculation method and device for boundary parameters of a supercritical fluid heat exchanger, computer equipment and a medium, and relates to the technical field of fluid mechanics and heat transfer science calculation.The method comprises the following steps that design parameters and target working conditions of the heat exchanger are obtained; calculating to obtain the single section length of each infinitesimal section; based on the initial inlet pressure guess value, according to the length of the single section and the pressure and the temperature of the infinitesimal section i, the pressure and the temperature of the infinitesimal section i + 1 are obtained through calculation, the difference value between the outlet calculation pressure and the target outlet pressure is calculated, and the difference value serves as a pressure error; if the pressure error is greater than the convergence tolerance, updating the inlet pressure guess value by adopting a numerical iteration method, and performing iteration again based on the inlet pressure guess value after iteration; otherwise, iteration is terminated, and the final inlet pressure value and the final outlet temperature value are output. According to the scheme, the calculation efficiency and accuracy of solving the inlet pressure and the outlet temperature of the supercritical fluid heat exchanger under the given target outlet pressure are improved.

The invention discloses a power module temperature hybrid model prediction method, and relates to the technical field of reliability of power electronic devices, and the method comprises the following steps: building a power module finite element thermal model based on the geometric dimension and internal characteristics of a power module; acquiring a temperature field sampling snapshot based on the power module finite element thermal model; extracting a stable evolution trend and an unstable evolution trend of a temperature field of the power module in the sampling snapshot based on a dynamic mode decomposition method; and predicting a stable trend by using an iteration method, and predicting an unstable trend by using a deep learning model. And superposing the predicted stable trend and the predicted unstable trend to obtain a predicted snapshot of the temperature field. According to the method, hybrid model prediction is adopted, the calculation efficiency can be improved as much as possible while the simulation precision is guaranteed, the chip temperature and internal temperature data of the power module are obtained, the number of needed sampling snapshots is reduced, and the modeling period is shortened.

The invention relates to the technical field of power systems, and discloses a wind and light output scene-oriented power system power flow data reconstruction method, system and device and a medium, and the method comprises the steps of constructing structured boundary input data; generating a voltage amplitude and a phase angle estimation vector of each node through a deep neural network model, and taking the voltage amplitude and the phase angle estimation vector as a power flow initial solution; performing physical constraint check on the initial solution, identifying out-of-limit nodes and determining an adjustable point set with reactive power regulation capability; and in the range of the adjustable point set, constructing an optimization function, solving to obtain a local adjustment quantity, correcting the initial solution, and generating a final power flow solution meeting the operation constraint of the power system. According to the method, data-driven modeling and physical constraint checking are fused, overall mapping is performed on a high-dimensional scene formed by wind and light output and load by using the deep neural network, the problem that the convergence performance is sharply reduced when the dimension is increased and the number of scenes is increased in a traditional iteration method is avoided, and rapid, stable and engineering-available power flow data generation is realized.

The invention discloses a low-altitude flight path planning method based on grid coding, which relates to the technical field of path planning, and is characterized in that on the basis of space and time grid coding, an A * path planning algorithm is improved, a neighbor node range in a classical algorithm is expanded, pruning optimization for removing invalid neighbors is applied to the A * path planning algorithm, and the path planning efficiency is improved. Relatively fast searching and establishment of a smooth grid path meeting the motion constraint become possible; an algorithm for interpolating a track on a grid path at equal time intervals is introduced; comprising the following steps of: thinning and smoothing a grid path; carrying out mathematical modeling on a flight state of the unmanned aerial vehicle; calculating flight segment initial and final speeds and flight segment middle cruising speeds which meet the motion performance constraint of the unmanned aerial vehicle and the consistency between flight segments by a backtracking iteration method; and calculating specific flight path points by speed integration. And a relatively accurate unmanned aerial vehicle flight plan track prediction generation method is provided.

Lowest cost usage scheduling of an energy system during a time interval of interest is achieved by utilizing two components of learning and optimization. First, several learning methods are used to forecast energy production and if needed energy consumption of a given group of energy components as a function of historical production data, weather, and solar irradiance data collected from weather and geo reports. A classification approach is used to select the learning model. A classification approach is used to select the learning model and approach. Then, an optimization problem is formulated and solved to create the optimal schedule of energy usage during time intervals of interests for the given group of energy components subject to scheduling and equipment constraints. Accordingly, integrated iterative methods, programs, and systems are described aiming at minimizing the cost of energy consumption for the given group of energy components within time intervals of interest.

The disclosed systems, methods, and computer-readable media for attributing value to a plurality of members of a company can be configured to retrieve, from each user device associated with at least a respective member, first user input data indicating a respective set of initial scores; assign a respective preliminary first member valuation to at least one member of a group; update a respective preliminary first member valuation assigned to each member of the group by performing an iterative method; generate at least one feature vector based at least on the each first user input data and the respective preliminary first member valuation assigned to each member of the group; provide the at least one feature vector to a machine learning model that is configured to generate final member valuations for the first group for generation of a recommendation that at least one member of the first group be replaced.

The invention belongs to the technical field of new energy power generation, and particularly relates to a multistage collaborative random optimization scheduling method, device and equipment and a medium. The method comprises the steps of establishing a robust constraint for global new energy prediction uncertainty, and establishing a chance constraint for local new energy prediction uncertainty; for prediction error distribution of different areas, a kernel density estimation method is adopted to establish a global and local new energy prediction error probability distribution model, and the maximum value of a certain quantile of global and local new energy prediction errors is selected; and a global robust and local random multi-region power system random optimization scheduling model is constructed, a bipartite iteration method is adopted to optimize opportunity parameters of establishment of reserve capacity constraints of each region, and the target of economical optimization operation of the multi-region interconnection system is achieved.

The invention relates to the technical field of computer aided design, in particular to a neighborhood gradient-oriented large-curvature curved surface projection point calculation method, which comprises the following steps of: S1, calculating a space bounding box of a curved surface to be projected, and performing triangulation to generate a triangular mesh; s2, traversing curved surface control vertexes of the triangular mesh, recording an adjacent vertex set, and judging whether the vertex set is a boundary point or not according to the UV parameter coordinates; s3, storing the vertex information and constructing a BVH tree, traversing the BVH tree for a to-be-projected point, querying and obtaining the nearest vertex with the minimum distance from the to-be-projected point, judging whether the nearest vertex is a boundary point or not, and processing according to a judgment result to obtain an iteration initial value; and S4, constructing a target function by taking the iteration initial value as an initial value, and calculating by using a Newton iteration method to obtain an accurate projection point of the to-be-projected point on the curved surface. The curved surface projection point approximation result calculation method is high in calculation efficiency and robustness, and the influence of curvature, grid division and other factors on Newton iteration input is reduced.

The invention discloses a power distribution network load flow calculation method based on a full-pure embedding theory. The power distribution network load flow calculation method comprises the following steps: (1) establishing a full-pure embedding load flow calculation model of a two-node system consisting of a balance node and a PQ node by applying a full-pure embedding method; (2) deriving and expanding based on the full-pure embedded load flow model of the two-node system established in the step (1) to obtain a full-pure embedded load flow calculation model of a multi-node system comprising a balance node and N PQ nodes; and (3) completing power flow calculation of the power distribution network by using the full-pure embedded power flow calculation model of the multi-node system. Compared with the Newton-Raphson method, the method has the advantages that the difference is small, the calculation time is greatly shortened, and the accuracy and the high efficiency of the method are remarkably improved compared with those of the Newton-Raphson method which is an iterative method.

The invention discloses a multi-park integrated energy system collaborative optimization method based on an equivalent projection theory. The method comprises the following steps: establishing a park integrated energy system scheduling model considering fuel gas hydrogen doping; establishing a park integrated energy system robust collaborative scheduling model participating in the multi-park integrated energy system alliance by considering the uncertainty of hydrogen energy interaction and energy load; the established double-layer region integrated energy system robust collaborative scheduling model is equivalent to a single-layer mixed integer programming collaborative scheduling model based on an equivalent projection theory; deriving a decentralized solution of the established mixed integer programming collaborative scheduling model by using a dual decomposition algorithm; and establishing a multi-park integrated energy system collaborative optimization model based on the feasible region equivalent projection and solving the multi-park integrated energy system collaborative optimization model. According to the method, a two-stage robust problem is converted into an MILP type robust optimization problem through feasible region projection, the converted model does not need to be iteratively solved, and the solving time is greatly shortened compared with that of a traditional iterative method; and a decentralized solution is derived by using a dual decomposition algorithm, so that the convergence and optimality of the result are ensured.