Power dispatch model solving optimization system and method

Abstract

The application belongs to the field of power system automation, and discloses a power dispatching model solving optimization system and method, which comprises a bottleneck identification module, a mixed integer programming acceleration module, a random mixed integer programming acceleration module and an optimization engine integration module. The bottleneck identification module is used for identifying the solving bottleneck of the power dispatching model and generating a solving optimization strategy according to the solving bottleneck. The mixed integer programming acceleration module is used for processing the mixed integer programming model in the power dispatching model, ensuring fast convergence and high-precision solving. The random mixed integer programming acceleration module is used for processing the random mixed integer programming model in the power dispatching model, ensuring the efficiency and reliability of the dispatching scheme of the solving result in a complex uncertain environment. The optimization engine integration module is used for modularly packaging the power dispatching model solving algorithm and establishing a unified calling interface for calling and dynamically calling the calculation resources of the power dispatching model solving process. The application can meet the efficiency, real-time performance and reliability requirements of the power dispatching model solving of the new power system and ensure the stable operation of the power system.

Description

Technical Field

[0001] This invention belongs to the field of power system automation and relates to an optimization system and method for solving power dispatch models. Background Technology

[0002] With the global energy transition and the large-scale integration of renewable energy, the dispatch and operation of power systems are facing unprecedented complexity and challenges. Traditional power dispatch methods are mainly based on deterministic optimization theory, but under the new power system, the dispatch problem exhibits higher complexity and uncertainty due to the instability of new energy output, fluctuations in load demand, and the spatiotemporal coupling characteristics of the power system.

[0003] Power dispatch optimization mainly includes problems such as economic dispatch, unit allocation, and power flow optimization. These problems are typically solved using methods such as linear programming (LP), nonlinear programming (NLP), and mixed-integer programming (MIP). However, traditional optimization methods have the following limitations: Low computational efficiency: When faced with large-scale mixed-integer programming models, the solution speed of traditional methods is insufficient to meet the requirements of real-time dispatch in actual power systems. Simple structure: Most methods do not fully consider the complex spatiotemporal coupling structure of the power grid and cannot efficiently handle coupling constraints and variables in dispatch problems. Insufficient handling of uncertainties: Traditional optimization methods rely on deterministic models and lack effective means to handle the randomness and uncertainty brought about by the integration of new energy sources, affecting the reliability of dispatch schemes.

[0004] Under the new power system, the output of renewable energy sources such as wind and solar power exhibits significant randomness and intermittency, posing a significant challenge to the supply-demand balance of the power system. For example, the output of renewable energy is difficult to predict accurately, increasing the uncertainty of dispatching schemes, and the rapid fluctuations in renewable energy output place higher demands on the frequency stability of the power grid. Simultaneously, with the widespread application of electric vehicles, energy storage devices, and distributed energy resources, electricity load exhibits diversified and random characteristics. For instance, the uneven spatiotemporal distribution of load demand exacerbates the complexity of grid dispatching, and the random fluctuations in electricity demand increase the difficulty of supply-demand matching. Furthermore, the complexity of the power dispatching optimization problem is constantly increasing, involving a large number of continuous and discrete variables, highly coupled power flow constraints, network security constraints, and generation output constraints, and the dispatching problem involves different time scales and spatial ranges. It is evident that under the new power system, the large-scale integration of renewable energy, the diversification and uncertainty of load demand, and the complexity of the power dispatching optimization problem make power dispatching models increasingly complex. Therefore, how to optimize the solution of power dispatching models to accelerate the solution speed and improve the quality of optimal solutions, thereby ensuring the stable operation of the power system, has become an urgent problem to be solved. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a power dispatching model solving optimization system and method.

[0006] To achieve the above objectives, the present invention employs the following technical solution:

[0007] In a first aspect, this invention provides a power dispatching model solution optimization system, comprising: a bottleneck identification module, used to analyze the variables, constraints, and branch-and-bound nodes of the power dispatching model, identify the solution bottlenecks of the power dispatching model, and generate a solution optimization strategy based on the solution bottlenecks; a mixed-integer programming acceleration module, used to perform variable and constraint preprocessing on the mixed-integer programming model in the power dispatching model, introduce a heuristic search strategy to determine the priority solution variables, adjust the branch order in the branch-and-bound solution process based on an adaptive branch variable selection strategy and a reinforcement learning algorithm, process the sub-problems in the branch-and-bound solution process in parallel based on multi-threading technology, and introduce a cutting plane in the feasible region; a stochastic mixed-integer programming acceleration module, used to perform multiple random variable modeling, uncertainty constraint transformation, stochastic scenario simulation, stochastic scenario parallel solution, and adjust the branch order in the branch-and-bound solution process based on a reinforcement learning algorithm on the stochastic mixed-integer programming model in the power dispatching model; and an optimization engine integration module, used to modularly encapsulate each power dispatching model solution algorithm to obtain each algorithm module and establish a unified calling interface for calling, and dynamically call the computing resources of the power dispatching model solution process.

[0008] Optionally, the analysis of variables, constraints, and branch-and-bound nodes in the power dispatching model, identification of solution bottlenecks in the power dispatching model, and generation of solution optimization strategies based on the solution bottlenecks include: analyzing the impact of variables on solution efficiency and the distribution and density of variables to obtain high-impact integer variables, high-impact semi-continuous variables, and high-complexity integer variables in the power dispatching model, which are used as variable bottlenecks in the solution bottlenecks of the power dispatching model; analyzing the constraints of the power dispatching model to obtain highly coupled constraints, nonlinear constraints, and redundant constraints in the power dispatching model, which are used as constraint bottlenecks in the solution bottlenecks of the power dispatching model; and analyzing the branch-and-bound nodes of the power dispatching model. The system identifies invalid branch-bound nodes and slowly converging branch-bound nodes in the power dispatch model, which serve as efficiency bottlenecks in solving the model. For variable bottlenecks, high-impact integer and semi-continuous variables are relaxed into continuous variables, and high-complexity integer variables are decomposed into simpler variables. For constraint bottlenecks, highly coupled constraints are merged or dimensionality-reduced, redundant constraints are eliminated, and nonlinear constraints are transformed into linear forms. For branch-bound node bottlenecks, invalid branch-bound nodes are pruned through constraint relaxation, and heuristic searches are performed using simulated annealing, tabu search, or genetic algorithms, with the search process conducted in parallel.

[0009] Optionally, the bottleneck identification module is also used to acquire power grid operation data and scheduling requirements, and to establish a power scheduling model based on the power grid operation data and scheduling requirements.

[0010] Optionally, the variable and constraint preprocessing of the mixed integer programming model in the power dispatch model includes: for unconstrained variables in the mixed integer programming model, transforming them into constrained forms through variable substitution; for redundant constraints in the mixed integer programming model, merging constraints based on the equivalent constraint merging technique; the introduction of a heuristic search strategy to determine the priority solution variables includes: introducing a heuristic search strategy to prioritize solving the variables in the mixed integer programming model that contribute significantly to the objective function.

[0011] Optionally, the parallel processing of subproblems in the branch and bound method solution process based on multi-threading technology includes: assigning different subproblems to different computing threads for parallel solution; wherein, during the parallel solution process, each subproblem executes several linear programming sub-solvers in parallel, and the computing threads with large computational loads are allocated more thread resources than the computing threads with small computational loads.

[0012] Optionally, introducing a cut plane in the feasible region includes: introducing a Gomory cut plane, a cover cut plane, a power flow constraint cut plane, and a spatiotemporal characteristic cut plane in the feasible region; dynamically checking whether the current solution violates a cut plane during the solution process, and dynamically adding cut plane constraints if it does; and generating new cut plane constraints in real time as the search tree expands.

[0013] Optionally, the multi-random variable modeling of the stochastic mixed integer programming model in the power dispatch model includes: adopting a two-stage stochastic programming approach, wherein the first stage solves for deterministic variables, the second stage sets the probability distribution of uncertain variables, and solves for uncertain variables through stochastic scenario compensation.

[0014] Optionally, the uncertainty constraint transformation includes: transforming the uncertainty constraint into an equivalent deterministic constraint by introducing a confidence level to the uncertainty constraint.

[0015] Optionally, the random scenario simulation and parallel solution of random scenarios include: generating several initial random scenarios using Monte Carlo simulation, and clustering the several initial random scenarios using K-means clustering to obtain several typical random scenarios; solving the several typical random scenarios in parallel to obtain the optimal solution of each typical random scenario and performing scenario weighted merging to obtain the final global weighted optimal solution.

[0016] In a second aspect, the present invention provides a method for optimizing a power dispatch model based on the aforementioned power dispatch model optimization system, comprising: during the power dispatch model solving process, analyzing the variables, constraints, and branch-bound nodes of the power dispatch model through a bottleneck identification module to identify the solution bottlenecks of the power dispatch model, generating a solution optimization strategy based on the solution bottlenecks, and optimizing the solution process based on the solution optimization strategy; during the power dispatch model solving process, performing variable and constraint preprocessing on the mixed integer programming model in the power dispatch model through a mixed integer programming acceleration module, introducing a heuristic search strategy to determine the priority solution variables, and adjusting the branch variable selection based on an adaptive branch variable selection strategy and a reinforcement learning algorithm. This paper discusses the branching order in the branch-and-bound method solution process, the parallel processing of subproblems in the branch-and-bound method solution process based on multithreading technology, and the introduction of cutting planes in the feasible region. In the power dispatching model solution process, a stochastic mixed-integer programming acceleration module is used to perform multiple random variable modeling, uncertainty constraint transformation, stochastic scenario simulation, and parallel solution of stochastic scenarios. Furthermore, a reinforcement learning algorithm is used to adjust the branching order in the branch-and-bound method solution process. Finally, in the power dispatching model solution process, a unified calling interface is used to call various algorithm modules in the optimization engine integration module, and the optimization engine integration module dynamically calls the computational resources for the power dispatching model solution process.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] This invention presents a power dispatching model solution optimization system. Based on a bottleneck identification module, it identifies the bottlenecks causing low efficiency or unsatisfactory results in power dispatching model solving, thereby implementing targeted optimization strategies to improve efficiency and result quality. A mixed-integer programming acceleration module preprocesses variables and constraints in the mixed-integer programming model of the power dispatching model, introduces a heuristic search strategy to determine priority variables, and introduces a cutting plane in the feasible region, significantly improving the solution speed of the mixed-integer programming model. Simultaneously, by using multi-threading technology to process sub-problems in parallel during the branch-and-bound method solution process, it fully utilizes computing resources to further accelerate the overall solution process. Furthermore, by using an adaptive branch variable selection strategy and reinforcement learning algorithms to adjust the branch order in the branch-and-bound method solution process, it can effectively reduce the size of the search tree and improve search efficiency. The mixed-integer programming acceleration module provides efficient mixed-integer programming model solving capabilities, ensuring fast convergence and high-precision solutions under large-scale mixed-integer programming models. This invention utilizes a stochastic mixed-integer programming acceleration module to perform multi-random variable modeling, uncertainty constraint transformation, stochastic scenario simulation, parallel solution of stochastic scenarios, and reinforcement learning-based adjustment of the branch-and-bound method in the power dispatching model. This improves the solution efficiency of large-scale stochastic mixed-integer programming models and provides core capabilities for handling uncertainties, ensuring the efficiency and reliability of the solution results under complex and uncertain environments. An optimization engine integration module is used to encapsulate various power dispatching model solution algorithms into modular units, establishing a unified calling interface for flexible combination and application to adapt to diverse dispatching needs. By dynamically allocating computational resources during the power dispatching model solution process, it fully utilizes computational resources and avoids idle resources. This invention meets the efficiency, real-time performance, and reliability requirements of new power systems for power dispatching model solutions, thereby effectively ensuring the stable operation of the power system. Attached Figure Description

[0019] Figure 1 This is a block diagram of the power dispatch model optimization system structure according to an embodiment of the present invention.

[0020] Figure 2 This is a flowchart of the power dispatching model optimization method according to an embodiment of the present invention. Detailed Implementation

[0021] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0022] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0023] The present invention will now be described in further detail with reference to the accompanying drawings:

[0024] See Figure 1 In one embodiment of the present invention, a power dispatching model solving optimization system is provided, which effectively solves the problems of low efficiency and poor solution quality in power dispatching model solving.

[0025] Specifically, the power dispatching model solving optimization system of the present invention includes a bottleneck identification module, a mixed integer programming acceleration module, a stochastic mixed integer programming acceleration module, and an optimization engine integration module.

[0026] The module comprises several components: a bottleneck identification module for analyzing variables, constraints, and branch-and-bound nodes in the power dispatching model; a mixed-integer programming acceleration module for preprocessing variables and constraints in the mixed-integer programming model of the power dispatching model; a heuristic search strategy for determining priority variables; an adaptive branch variable selection strategy and reinforcement learning algorithm for adjusting the branch order in the branch-and-bound process; parallel processing of subproblems in the branch-and-bound process using multi-threading technology; and the introduction of a cutting plane in the feasible region. A stochastic mixed-integer programming acceleration module for modeling multiple random variables, transforming uncertainties and constraints, simulating stochastic scenarios, parallel solving of stochastic scenarios, and adjusting the branch order in the branch-and-bound process using reinforcement learning algorithms. An optimization engine integration module modularly encapsulates the solution algorithms for each power dispatching model, establishes a unified calling interface, and dynamically calls the computational resources for the power dispatching model solution process.

[0027] This invention presents a power dispatching model solution optimization system. Based on a bottleneck identification module, it identifies the bottlenecks causing low efficiency or unsatisfactory results in power dispatching model solving, thereby implementing targeted optimization strategies to improve efficiency and result quality. A mixed-integer programming acceleration module preprocesses variables and constraints in the mixed-integer programming model of the power dispatching model, introduces a heuristic search strategy to determine priority variables, and introduces a cutting plane in the feasible region, significantly improving the solution speed of the mixed-integer programming model. Simultaneously, by using multi-threading technology to process sub-problems in parallel during the branch-and-bound method solution process, it fully utilizes computing resources to further accelerate the overall solution process. Furthermore, by using an adaptive branch variable selection strategy and reinforcement learning algorithms to adjust the branch order in the branch-and-bound method solution process, it can effectively reduce the size of the search tree and improve search efficiency. The mixed-integer programming acceleration module provides efficient mixed-integer programming model solving capabilities, ensuring fast convergence and high-precision solutions under large-scale mixed-integer programming models. This invention utilizes a stochastic mixed-integer programming acceleration module to perform multi-random variable modeling, uncertainty constraint transformation, stochastic scenario simulation, parallel solution of stochastic scenarios, and reinforcement learning-based adjustment of the branch-and-bound method in the power dispatching model. This improves the solution efficiency of large-scale stochastic mixed-integer programming models and provides core capabilities for handling uncertainties, ensuring the efficiency and reliability of the solution results under complex and uncertain environments. An optimization engine integration module is used to encapsulate various power dispatching model solution algorithms into modular units, establishing a unified calling interface for flexible combination and application to adapt to diverse dispatching needs. By dynamically allocating computational resources during the power dispatching model solution process, it fully utilizes computational resources and avoids idle resources. This invention meets the efficiency, real-time performance, and reliability requirements of new power systems for power dispatching model solutions, thereby effectively ensuring the stable operation of the power system.

[0028] In one possible implementation, the step of analyzing the variables, constraints, and branch-and-bound nodes of the power dispatching model, identifying the solution bottlenecks of the power dispatching model, and generating a solution optimization strategy based on the solution bottlenecks includes: analyzing the impact of variables on solution efficiency and the distribution and density of variables to obtain high-impact integer variables, high-impact semi-continuous variables, and high-complexity integer variables of the power dispatching model, which are considered variable bottlenecks in the solution bottlenecks of the power dispatching model; analyzing the constraints of the power dispatching model to obtain highly coupled constraints, nonlinear constraints, and redundant constraints of the power dispatching model, which are considered constraint bottlenecks in the solution bottlenecks of the power dispatching model; and analyzing the power dispatching model's... Branch and bound nodes are identified as invalid and slowly converging branch and bound nodes in the power dispatch model, serving as efficiency bottlenecks in the solution process. For variable bottlenecks, high-impact integer and semi-continuous variables are relaxed to continuous variables, and high-complexity integer variables are decomposed into simpler variables. For constraint bottlenecks, highly coupled constraints are merged or dimensionality-reduced, redundant constraints are eliminated, and nonlinear constraints are transformed into linear forms. For branch and bound node bottlenecks, invalid branch and bound nodes are pruned through constraint relaxation, and heuristic searches are performed using simulated annealing, tabu search, or genetic algorithms, with the search process conducted in parallel.

[0029] Interpretive bottleneck identification aims to effectively identify the performance bottlenecks in solving power dispatch models, providing a foundation for subsequent solution optimization.

[0030] The complexity of power dispatching models increases exponentially with the number of variables and constraints. In this implementation, bottlenecks are identified using three indicators: variables, constraints, and branch and bound nodes.

[0031] Variables in power dispatch models are generally classified into three categories: continuous variables, such as power generation; discrete variables, such as unit start-up and shutdown status; and semi-continuous variables, such as special ordered set variables representing piecewise linear relationships of partial constraints. By analyzing the impact of variables on solution efficiency and their distribution and density, we can identify high-impact integer variables, high-impact semi-continuous variables, and high-complexity integer variables.

[0032] The identification methods for high-complexity integer variables include: I. Statistical analysis of integer variable distribution to clarify the quantity, distribution characteristics, and impact on model complexity of integer variables. This includes: 1) Variable quantity statistics: Calculating the proportion of integer variables in the total number of variables. If the proportion is >30%, it indicates a high proportion of integer variables, which may lead to a decrease in solution efficiency. 2) Integer variable distribution regions: Classifying integer variables by function (e.g., start / stop status and regulating equipment status), and statistically analyzing the quantity and distribution of each type of integer variable to identify key variable groups. 3) Distribution sparsity analysis: Calculating the sparsity of integer variables through the variable-constraint matrix (column sparsity). II. Variable complexity sensitivity analysis to evaluate the specific contribution of integer variables to problem complexity and solution time. This includes: 1) Variable relaxation experiments: Relaxing some integer variables into continuous variables, solving the optimization problem, and evaluating the change in solution time before and after relaxation. If the relaxation of a certain type of integer variable brings a significant reduction in time, then that type of variable is the bottleneck in the solution. 2. Integer Variable Group Sensitivity: Group integer variables by category (e.g., equipment start / stop variables, resource allocation variables, etc.), relax each group of variables, and record their contribution to the overall complexity. 3. Variable Characteristics and Performance Correlation Analysis: Use correlation analysis tools to examine the relationship between variable characteristics (e.g., sparsity) and solution time. III. Identifying High-Complexity Integer Variables: Identify the integer variables that have the greatest impact on solution complexity and clarify their role in the model. This includes: 1. Branch Contribution Analysis: During branch and bound analysis, count the number of branch expansions caused by each integer variable and calculate the branch efficiency. If the branch efficiency of a certain integer variable is significantly higher than that of other variables, then that variable is a high-complexity integer variable.

[0033] For high-impact integer variables and high-impact semi-continuous variables, the variables with the greatest impact on the convergence of the power dispatch model are identified by combining the functional characteristics of the variables (such as equipment start-up, shutdown, and charging / discharging states) and utilizing intermediate results and performance indicators from the power dispatch model solution process. For example, the convergence speed of the variables can be tested using variable relaxation methods to assess the performance improvement effect.

[0034] Constraint analysis of power dispatch models includes identifying highly coupled constraints (such as constraints spanning time, space, or equipment) and detecting nonlinear and redundant constraints. Highly coupled constraints are identified as follows: coupling is defined as the degree of correlation between variables in a constraint, obtained by calculating the column correlation of the constraint matrix. If the column correlation is greater than a threshold, it indicates that the two columns are highly coupled, corresponding to a highly coupled constraint.

[0035] The branch-and-bound node analysis of the power dispatch model employs a branch-and-bound log analysis method combined with performance visualization tools to identify the size of the branch-and-bound search tree and key nodes. This method, through detailed analysis of the solver's log files and in-depth analysis of the search tree using visualization tools, can efficiently identify key nodes and performance bottlenecks. Branch-and-bound (B&B) is the core algorithm for solving mixed integer programming problems (MIP). Its performance bottlenecks typically manifest in the following aspects of the search tree: Search tree size: Too many nodes. Key nodes: Significantly long solution times for certain nodes, or a surge in expansion counts. The log analysis method obtains the following core information by parsing the branch-and-bound logs generated by the solver: Solution time of the root node: Used to evaluate the basic complexity of the model. Number of node expansions: Identifies key variables or constraints affecting the size of the search tree. Branch efficiency of each node: Analyzes the effectiveness of variable branching strategies. Subproblem solution time: Evaluates the impact of complex subproblems on the overall solution. Combined with performance visualization tools, the log analysis method can quickly identify key bottleneck nodes in the search tree.

[0036] For identified bottlenecks, select appropriate optimization methods or techniques to improve solution efficiency. For variable bottlenecks: 1. Variable relaxation: Relax integer variables into continuous variables to simplify the solution. 2. Variable decomposition: Decompose complex variables into multiple simpler variables to reduce computational complexity. For constraint bottlenecks: 1. Constraint reduction: Eliminate redundant constraints to improve the sparsity of the constraint matrix. 2. Constraint approximation: Transform complex nonlinear constraints into linear forms (e.g., DC approximation of power flow constraints). For branch and bound node bottlenecks: 1. Pruning strategies: Prune invalid nodes in advance through constraint relaxation to reduce the search space. 2. Heuristic search: Utilize intelligent methods such as simulated annealing, tabu search, or genetic algorithms for rapid convergence. 3. Parallel computing: Decompose the search process into multiple computational nodes to improve search efficiency.

[0037] In one possible implementation, the bottleneck identification module is also used to acquire power grid operation data and scheduling requirements, and to establish a power scheduling model based on the power grid operation data and scheduling requirements.

[0038] Explanatoryly, the core objective of power dispatching models is to achieve power grid supply and demand balance through optimization methods, ensuring safe and economical operation. For example, the mathematical modeling of power dispatching models can be based on the following dispatching requirements:

[0039] Minimize power generation costs:

[0040]

[0041] Among them, C i Let P be the cost function of unit i. i For the power generation of unit i, ai b i and c i These are preset coefficients.

[0042] Constraints typically include: power balance constraints: Among them, P demand For system load, P loss Network loss. Unit output constraints: in, and These represent the lower and upper limits of the generating power of unit i, respectively. Ramp-up constraint: The rate of change of unit power is limited. Network security constraints (such as power flow constraints): Based on power flow calculation equations, these constraints constrain the operational security of the power network.

[0043] Optionally, the bottleneck identification module can also be used to construct a set of typical problems, including various power dispatching scenarios, such as: large-scale generator unit combination problems, high-penetration renewable energy access problems, and multi-timescale (short-term, real-time) dispatching optimization problems. It can also be used for performance metrics and testing, including setting solution performance metrics such as computation time (total time required for solving), convergence accuracy (error between the optimal and feasible solutions), node search efficiency (number of iterations of subproblems during branch and bound processes); and building a performance testing and verification platform to evaluate the performance improvement of different solution optimization strategies.

[0044] In one possible implementation, the variable and constraint preprocessing of the mixed integer programming model in the power dispatch model includes: for unconstrained variables in the mixed integer programming model, transforming them into constrained forms through variable substitution; and for redundant constraints in the mixed integer programming model, merging constraints based on the equivalent constraint merging technique.

[0045] The mathematical form of an interpretable mixed-integer programming (MIP) model is generally expressed as:

[0046] minf(x)=c T x

[0047]

[0048] Where x is the decision variable, including integer variables x j and continuous variable x k A and b represent the constraint matrix and vector, respectively. c represents the coefficients of the objective function. X represents the feasible region of the variables.

[0049] Solving the MIP model involves optimization in a continuous space and a combined search of discrete variables, and its complexity increases exponentially with the size of the variables and constraints.

[0050] Interpretive, for the implicit free variable x (i.e., the unconstrained variable), variable substitution is performed to convert it into a constrained form:

[0051] x∈(-∞,+∞)→x′∈[0,+∞),x″=x′-x

[0052] Where x′, x″≥0≥0, the search space is reduced.

[0053] By combining equivalent constraints in a linear fashion and merging redundant constraints, the number of constraints can be reduced.

[0054] a1x + a2y = b1 and

[0055] In one possible implementation, the introduction of a heuristic search strategy to determine the priority variables includes: introducing a heuristic search strategy to prioritize solving the variables that contribute significantly to the objective function in the mixed integer programming model.

[0056] Interpretively, based on the characteristics of the power dispatch model, a heuristic search rule is designed to prioritize solving variables that contribute significantly to the objective function. In the power dispatch model, variables that contribute significantly are those that have a substantial impact on the objective function (such as system cost and supply-demand balance). The methods for identifying these variables are as follows: Variable sensitivity analysis: Objective function sensitivity: Calculate the partial derivative or sensitivity of the variable with respect to the objective function. The larger the sensitivity value of a variable, the greater its contribution to the objective function. Relaxation test: Perform relaxation tests on the variables and observe the changes in the objective function after relaxation. If the change is significant, the variable is considered a major contributor. Dual critical value analysis: Calculate the improvement in the dual critical value of the variable and prioritize variables with large improvements.

[0057] In one possible implementation, in adjusting the branching order in the branch-and-bound problem-solving process based on an adaptive branch variable selection strategy and a reinforcement learning algorithm, for example, the reinforcement learning algorithm can use a pre-trained deep reinforcement learning (such as DQN) model to dynamically select branch variables to adjust the branching order. The pre-trained deep reinforcement learning model trains the agent through a state-action-reward interaction process, dynamically selecting the optimal branch variable. The specific process is as follows: State definition: The information of the current node in the search tree is used as the state, including: the variable value of the current node, the size of the remaining subproblems, and the difference between the current objective value and the dual bound. Action definition: The action is to select a variable to branch. Reward function: The reward function is defined based on the improvement of the objective function or the improvement of search efficiency after branching. Training process: Using the DQN algorithm, the following objectives are optimized: Among them, Q(S) t ,x j ) indicates that in state S t Choose variable x jThe cumulative reward value. Branch order: Predict the optimal branch variable for each state and dynamically adjust the branch order.

[0058] In one possible implementation, the parallel processing of subproblems in the branch and bound method solution process based on multithreading technology includes: assigning different subproblems to different computing threads for parallel solution; wherein, during the parallel solution process, each subproblem executes several linear programming sub-solvers in parallel, and the computing threads with large computational loads are allocated more thread resources than the computing threads with small computational loads.

[0059] Interpretive approach: Subproblems are distributed across multiple computational threads for parallel solution, accelerating search tree traversal. The linear programming sub-solver includes the primal simplex method, dual simplex method, and interior-point method. Thread resource allocation is optimized, prioritizing computationally intensive nodes and avoiding excessive node differences, thereby improving parallel solution efficiency.

[0060] In one possible implementation, introducing a cut plane in the feasible domain includes: introducing a Gomory cut plane, a cover cut plane, a power flow constraint cut plane, and a spatiotemporal characteristic cut plane in the feasible domain; dynamically checking whether the current solution violates a cut plane during the solution process, and dynamically adding cut plane constraints if it does; and generating new cut plane constraints in real time as the search tree expands.

[0061] Interpretively, the feasible region is narrowed using Gomory cut planes and covering cut planes, improving the convergence speed of the root node. Specific cut planes are dynamically generated based on power flow constraints and spatiotemporal characteristics: power flow constraint cut planes and spatiotemporal characteristic cut planes. Specifically, Gomory cut planes are extracted from the simplex tableau of the current solution. Covering cut planes are generated for sets of variables exceeding the coverage conditions in integer programming problems, based on set coverage constraints. Power flow constraint cut planes are generated based on power flow constraints, transforming power flow imbalances into cut plane constraints and reducing infeasible solutions. Spatiotemporal characteristic cut planes are dynamically generated to limit the total number of variables across time. Furthermore, cutting constraints are dynamically added during the solution process to reduce redundant iterations. This includes dynamically checking whether the current solution violates a cut plane and adding a cut plane constraint if so, as well as generating new cut plane constraints in real time as the search tree expands.

[0062] In one possible implementation, the hybrid integer programming acceleration module further includes rapidly generating initial feasible solutions using heuristic methods to shorten solution time. Examples include heuristic initial solutions (e.g., using a greedy algorithm to quickly generate initial solutions) and relaxed initial solutions (e.g., generating integer initial solutions from solutions to a continuous relaxation problem). Finally, the module uses exact algorithms to verify the results, ensuring the feasibility and accuracy of the solutions.

[0063] In one possible implementation, the multi-random variable modeling of the stochastic mixed integer programming model in the power dispatch model includes: adopting a two-stage stochastic programming approach, wherein the first stage solves for deterministic variables, the second stage sets the probability distribution of uncertain variables, and solves for uncertain variables through stochastic scenario compensation.

[0064] Interpretive methods define the probability distribution of uncertain variables, such as defining random variables for wind power, photovoltaic output, and load fluctuations, assuming they satisfy a certain probability distribution, such as a normal distribution or a Beta distribution.

[0065] In one possible implementation, the uncertainty constraint transformation includes: transforming the uncertainty constraint into an equivalent deterministic constraint by introducing a confidence level to the uncertainty constraint.

[0066] Interpretively, a confidence level of 1-α is introduced into the uncertainty constraint: P(Ax≤b(ξ))≥1-α, transforming the uncertainty constraint into an equivalent deterministic constraint. in, Let z be the expected value, σ be the standard deviation of the random variable, and z be the expected value. α The quantiles are at the confidence level.

[0067] In one possible implementation, the random scenario simulation and parallel solution of random scenarios include: generating several initial random scenarios using Monte Carlo simulation, and clustering the several initial random scenarios using K-means clustering to obtain several typical random scenarios; solving the several typical random scenarios in parallel to obtain the optimal solution of each typical random scenario and performing scenario weighted merging to obtain the final global weighted optimal solution.

[0068] Interpretive, a large number of initial random scenarios are generated through Monte Carlo simulation, each scenario representing the specific values ​​of uncertain variables, such as the output of new energy sources and load demand.

[0069] This explanation describes how to cluster several initial random scenes using K-means clustering. The process involves using the K-means clustering algorithm to cluster a large number of initial random scenes, grouping similar scenes together. Input: A scene data matrix, where each row represents a feature vector for a scene. Output: K cluster centers, i.e., typical random scenes, and the classification label for each typical random scene.

[0070] Interpretively, several typical random scenarios are solved in parallel to obtain the optimal solutions for each typical random scenario. These scenarios are then weighted and merged to obtain the final globally weighted optimal solution, which includes:

[0071] Each typical random scene is assigned a weight, the weight value of which is determined by the frequency of occurrence of that type of typical random scene in the initial set of random scenes: The optimal solution for each typical random scenario is weighted according to its weight to obtain the globally weighted optimal solution, taking into account all scenarios.

[0072] Optionally, it can also be verified whether the globally weighted optimal solution satisfies global constraints. If not, the result is adjusted through post-processing. Post-processing adjustment aims to improve the feasibility and robustness of the scheduling scheme. The specific process is as follows: Feasibility verification: Check whether the globally weighted optimal solution satisfies all constraints. If constraints are violated (such as supply and demand balance or equipment capacity limitations), the adjustment process begins. Adjustment methods: 1. Heuristic correction: For variables that violate constraints, use a greedy algorithm or local search to adjust the solution. 2. Lagrange relaxation method: Introduce relaxation variables to relax and optimize the constraints. 3. Robustness enhancement method: Increase redundant adjustment capacity to address new energy fluctuations and load uncertainties. 4. Re-optimization method: Perform local optimization on the adjusted solution to ensure that the adjusted scheme satisfies all constraints and is close to the optimal solution.

[0073] The interpretable optimization engine integration module integrates linear programming, mixed-integer programming, and stochastic mixed-integer programming algorithms to form a unified optimization engine. This provides service-oriented encapsulation of the algorithms, offering flexible modular combination and invocation. A multi-threaded parallel solution framework is designed, supporting node parallelism and algorithm parallelism to improve computational resource utilization. Algorithm encapsulation interfaces are provided to enable data interaction and invocation with the power grid dispatching system. Furthermore, sparse storage structures and memory optimization techniques are employed to enhance the algorithm's solution efficiency in large-scale models.

[0074] The interpretable computational resources used in the dynamic invocation of the power dispatch model solution process include: 1. Optimizing thread allocation through a task scheduler to ensure efficient utilization of computational resources and avoid idle resources. 2. Reusing already computed solutions in branch and bound operations through a warm-start strategy to accelerate the solution of subproblems and avoid redundant computation.

[0075] See Figure 2 In another embodiment of the present invention, a power dispatching model solving optimization method is provided, which is implemented based on the above-mentioned power dispatching model solving optimization system.

[0076] Specifically, the power dispatching model solution optimization method includes the following steps: During the power dispatching model solution process, a bottleneck identification module analyzes the variables, constraints, and branch-and-bound nodes of the power dispatching model to identify solution bottlenecks. Based on these bottlenecks, a solution optimization strategy is generated, and the solution process is optimized according to the optimization strategy. During the power dispatching model solution process, a mixed-integer programming acceleration module preprocesses the mixed-integer programming model in the power dispatching model for variables and constraints; a heuristic search strategy is introduced to determine priority solution variables; and an adaptive branch variable selection strategy and reinforcement learning algorithm are used to adjust the branch-and-bound process. The algorithm implements several key features: it optimizes the branch order, employs multi-threading technology for parallel processing of subproblems in the branch-and-bound method, and introduces cutting planes within the feasible region. In the power dispatch model solution process, it utilizes a stochastic mixed-integer programming acceleration module to perform multi-random variable modeling, uncertainty constraint transformation, stochastic scenario simulation, parallel solution of stochastic scenarios, and adjust the branch order in the branch-and-bound method based on reinforcement learning algorithms. Furthermore, it uses a unified interface to call various algorithm modules within the optimization engine integration module and dynamically accesses computational resources for the power dispatch model solution process.

[0077] In another embodiment of the present invention, a power dispatching system is provided, comprising: a dispatching data input terminal for receiving demand data for power grid dispatching, including network topology, load forecasting, renewable energy output, etc.; a power dispatching model solving and optimization system for modeling a power dispatching model and executing the aforementioned power dispatching model solving and optimization method to efficiently solve the power dispatching model; and a result output terminal for outputting the optimal power dispatching scheme, including unit start-up and shutdown status, power allocation, power flow scheme, etc.

[0078] The present invention provides an optimization system and method for solving power dispatching models, achieving high performance and high reliability in solving power dispatching models. Specific beneficial effects are as follows:

[0079] 1. Improve solution efficiency and overcome computational bottlenecks. Through modeling-algorithm collaboration, variable preprocessing, and constraint aggregation, large-scale power dispatch optimization problems are preprocessed, reducing the size of variables and the number of constraints. Adaptive branch variable selection and dynamic cutting plane generation techniques, combined with deep reinforcement learning to optimize variable branching order, effectively reduce the size of the search tree and improve the search efficiency of the branch and bound method. Parallel branch and bound algorithms and multi-threaded resource scheduling mechanisms enable parallel task solving, accelerating the overall convergence speed. Parallel execution of multilinear programming solution algorithms (primitive simplex method, dual simplex method, interior point method) and dynamic selection of the optimal method further improve computational efficiency. Under the same problem scale, the solution speed of this invention is significantly improved compared to traditional commercial solvers (such as CPLEX and Gurobi), enabling the solution of large-scale mixed integer programming problems within a specified time, meeting the real-time requirements of power grid dispatch.

[0080] 2. Effectively handles uncertainty and enhances the robustness of the scheme. Through chance-constrained programming, confidence level transformation, and stochastic scenario simulation, it effectively handles stochastic factors such as renewable energy output and load fluctuations. Scenario clustering and screening techniques simplify the computational complexity introduced by random variables and retain representative stochastic scenarios. Deep reinforcement learning is used to select auxiliary variables, improving the efficiency of solving key variables under stochastic scenarios and optimizing the robustness of the scheduling scheme. Under conditions of renewable energy output fluctuations and random load changes, the power dispatching scheme generated by this invention satisfies the constraints at a given confidence level, exhibiting stronger robustness and reliability.

[0081] 3. Ensuring System Security. This invention eliminates reliance on foreign commercial solvers (such as CPLEX and Gurobi) and proposes an independently controllable optimization engine. Through modular encapsulation and service-oriented interfaces, it supports efficient solutions for linear programming, mixed-integer programming, and stochastic programming. It achieves algorithm control and multi-scenario adaptation, flexibly combining algorithm modules according to the actual needs of the power system to adapt to various power grid dispatch optimization scenarios. This invention provides autonomous optimization capabilities, solving the "black box" problem and potential information security risks of traditional commercial solvers, ensuring the security and controllability of power dispatch optimization.

[0082] 4. Efficient resource utilization and improved system adaptability. Through multi-threaded parallel computing and task scheduling mechanisms, dynamic allocation and efficient utilization of computing resources are achieved, avoiding resource waste. A warm-start mechanism is designed to reduce redundant calculations and improve the solution speed of subproblems. The algorithm module is encapsulated as a service, providing standard interfaces to support data interaction and invocation with external scheduling systems, exhibiting good scalability and compatibility. This invention maximizes the utilization of computing resources, flexibly adapting to the power dispatch optimization needs of different scales and scenarios, and possesses strong practicality and scalability.

[0083] 5. Comprehensive Dispatch Effect, Enhancing the Economy and Stability of Power Grid Operation. This invention generates optimized power dispatch schemes through efficient solution algorithms, achieving the minimization of generation costs and a balance between power supply and demand. Under complex and uncertain conditions, stochastic optimization improves the robustness of power grid dispatch, ensuring the safe operation of the power grid. By modeling the uncertainties of renewable energy output and load forecasting, it effectively addresses stochastic disturbances and improves the power grid's adaptability to large-scale renewable energy integration. The optimized dispatch scheme reduces system operating costs and enhances the economy, stability, and reliability of power grid dispatch, aligning with the goals of efficient and low-carbon operation of the new power system.

[0084] 6. Adaptable to the diverse needs of new power systems. This invention is applicable to various power grid dispatching scenarios, such as large-scale unit combination problems: solved through an efficient mixed-integer programming approach; high-penetration renewable energy scenarios: addressing the uncertainty of wind and solar power output to achieve dynamic supply-demand balance; and multi-timescale dispatching: supporting multi-scale optimization needs for long-term, day-ahead, and real-time dispatching. It possesses high adaptability and flexibility, capable of meeting the dispatching and operation requirements of different scenarios in new power systems.

[0085] The power dispatching model optimization system and method of this invention can be applied to the following scenarios:

[0086] 1. Large-scale power system dispatching and operation scenarios. In large-scale power systems, the dispatch center needs to coordinate the power flow between hundreds of generating units and multiple grid areas in real time, while meeting supply and demand balance, security constraints, and economic objectives. However, large-scale power dispatching involves a large number of mixed integer programming problems, including unit start-up and shutdown, power allocation, and power flow optimization. Traditional solution methods have bottlenecks in terms of computational speed and accuracy. This paper proposes a power dispatching model-based optimization system and method to quickly solve large-scale unit combination and economic dispatching problems, ensuring that the dispatch center can generate optimal solutions in real time. Through variable preprocessing and adaptive branch and bound techniques, the solution time is significantly reduced, meeting real-time operation requirements.

[0087] 2. High-Penetration Renewable Energy Grid Dispatch Scenarios. With the large-scale integration of new energy sources such as wind and solar power, the uncertainties faced by the power system have increased significantly, including fluctuations in renewable energy output and load forecasting errors. In this situation, traditional dispatching methods are difficult to effectively address, and dispatching schemes are prone to deviating from actual needs, affecting the safe and stable operation of the power grid. Based on the power dispatching model to solve the optimization system and methods, a stochastic mixed-integer programming acceleration module, through chance-constrained programming and stochastic scenario simulation, fully considers the volatility of renewable energy output and the uncertainty of load demand, generating highly robust dispatching schemes. Through confidence level transformation, it ensures that the grid dispatching scheme meets supply-demand balance and security constraints under a given default probability. This improves the grid's adaptability to fluctuations in renewable energy output and reduces wind and solar curtailment rates.

[0088] 3. Multi-timescale dispatching scenarios in regional power grids. Dispatching requirements in regional power grids are typically categorized into multiple timescales, including long-term dispatching, day-ahead dispatching, and real-time dispatching. Different timescales present varying optimization complexities and accuracy requirements. Current dispatching systems lack a unified optimization engine, making it difficult to adapt to the need for efficient solutions across multiple timescales. Based on power dispatching model-based optimization systems and methods, different algorithm modules can be flexibly invoked across multiple timescales. For example, for long-term dispatching: a mixed-integer programming method is used to solve unit combination and long-term supply-demand balance problems; for day-ahead dispatching: generation plans and network power flow schemes are quickly generated; for real-time dispatching: rapid solutions are achieved through parallel computing, ensuring the real-time safe operation of the power grid. This improves dispatching efficiency and reliability, meeting the dispatching accuracy requirements at different timescales.

[0089] 4. Economic Dispatch Scenario in Electricity Market Environment. In an electricity market environment, grid dispatch needs to simultaneously satisfy supply and demand balance and the optimal goal of market economy, involving multi-stakeholder game theory and cost minimization. Furthermore, dispatch optimization needs to be completed rapidly within the market clearing time. Based on a power dispatch model-based optimization system and method, through a parallel computing framework and multi-algorithm integration, the power market dispatch optimization process is significantly accelerated. By introducing deep reinforcement learning to assist decision-making, the branch-and-bound search path is optimized, and the optimal market clearing scheme is quickly generated. This improves market clearing efficiency, ensures fair competition and optimal resource allocation, and meets the requirements of real-time market settlement.

[0090] 5. Grid Emergency Dispatch Scenarios under Extreme Weather. Under extreme weather conditions such as typhoons and blizzards, the grid's operation is easily disturbed, further increasing the uncertainty of renewable energy output and load demand. Traditional methods are insufficient for rapid response to emergencies, impacting grid stability. This paper proposes an optimization system and method based on a power dispatch model to simulate various disturbance scenarios under extreme weather conditions and formulate emergency dispatch plans. Confidence level transformation and reinforcement learning are used to assist optimization, quickly identifying key variables and improving the robustness and response speed of the dispatch plan. This ensures the safe and stable operation of the grid under extreme conditions and reduces the risk of power supply interruptions.

[0091] 6. Distributed Energy and Microgrid Dispatch Scenarios. With the widespread application of distributed energy, regional power grids or microgrids need to achieve coordinated and optimized dispatch of distributed energy sources (such as photovoltaics, wind power, and energy storage) to address the random fluctuations of distributed resources and grid connection issues. Based on power dispatch models, optimization systems and methods are developed. Through chance-constrained programming and scenario simulation, resources such as photovoltaics, energy storage, and loads are effectively coordinated to achieve source-grid-load-storage collaborative optimization. In microgrid operation, energy storage charging and discharging strategies and distributed resource dispatch are optimized to improve microgrid adaptability and power supply reliability. This reduces regional power grid energy consumption costs and increases the utilization rate of new energy sources.

[0092] The power dispatch model solving optimization system and method of the present invention can be widely applied to new power systems. In particular, it has significant economic efficiency, real-time performance and stability when dealing with uncertainties and complex constraints, and provides technical support for building intelligent, low-carbon and reliable power systems.

[0093] The module division in this embodiment of the invention is illustrative and represents only one logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional modules in the various embodiments of the invention can be integrated into a single processor, exist as separate physical entities, or be integrated into a single module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0094] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0095] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0096] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0097] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0098] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A power dispatching model solving optimization system, characterized in that, include: The bottleneck identification module is used to analyze the variables, constraints, and branch and bound nodes of the power dispatch model, identify the solution bottlenecks of the power dispatch model, and generate solution optimization strategies based on the solution bottlenecks. The mixed-integer programming acceleration module is used to perform variable and constraint preprocessing in the mixed-integer programming model of power dispatch model, introduce heuristic search strategy to determine priority solution variables, adjust the branch order in the branch and bound method solution process based on adaptive branch variable selection strategy and reinforcement learning algorithm, process sub-problems in the branch and bound method solution process in parallel based on multi-threading technology, and introduce cutting plane in feasible region. The Stochastic Mixed Integer Programming Acceleration Module is used to perform multiple random variable modeling, uncertainty constraint transformation, random scenario simulation, parallel solution of random scenarios, and adjustment of the branch order in the branch and bound method solution process based on reinforcement learning algorithm for the stochastic mixed integer programming model in power dispatching model. The optimization engine integration module is used to modularly encapsulate the solution algorithms of various power dispatching models, obtain each algorithm module, and establish a unified calling interface for calling, as well as dynamically calling the computing resources of the power dispatching model solution process.

2. The power dispatching model solving optimization system according to claim 1, characterized in that, The analysis of variables, constraints, and branch and bound nodes in the power dispatching model, identification of solution bottlenecks in the power dispatching model, and generation of solution optimization strategies based on the solution bottlenecks include: By analyzing the impact of variables on solution efficiency and the distribution and density of variables, high-impact integer variables, high-impact semi-continuous variables, and high-complexity integer variables in the power dispatch model are identified as variable bottlenecks in the solution bottleneck of the power dispatch model. By analyzing the constraints of the power dispatching model, we can identify the highly coupled constraints, nonlinear constraints, and redundant constraints of the power dispatching model, which serve as the constraint bottlenecks in solving the power dispatching model. By analyzing the branch and bound nodes of the power dispatching model, we can identify the invalid branch and bound nodes and the slowly converging branch and bound nodes, which serve as the efficiency bottlenecks in solving the power dispatching model. To address the variable bottleneck, high-impact integer variables and high-impact semi-continuous variables are relaxed into continuous variables, and high-complexity integer variables are decomposed into several simple variables. To address constraint bottlenecks, highly coupled constraints are merged or dimensionality reduced, redundant constraints are eliminated, and nonlinear constraints are transformed into linearized forms. To address the bottleneck of branch and bound nodes, we employ methods such as constraint relaxation to prune invalid branch and bound nodes, heuristic search using simulated annealing, tabu search, or genetic algorithms, and parallelizing the search process.

3. The power dispatching model solving optimization system according to claim 1, characterized in that, The bottleneck identification module is also used to acquire power grid operation data and scheduling requirements, and to establish a power scheduling model based on the power grid operation data and scheduling requirements.

4. The power dispatching model solving optimization system according to claim 1, characterized in that, The variable and constraint preprocessing of the mixed integer programming model in the power dispatch model includes: for unconstrained variables in the mixed integer programming model, transforming them into constrained forms through variable substitution; for redundant constraints in the mixed integer programming model, merging constraints based on the equivalent constraint merging technique. The introduction of heuristic search strategies to determine priority variables includes: prioritizing the solutions of variables that contribute significantly to the objective function in the mixed integer programming model.

5. The power dispatching model solving optimization system according to claim 1, characterized in that, The sub-problems in the branch-and-bound method solution process, which is based on multi-threading technology for parallel processing, include: Different subproblems are assigned to different computational threads for parallel solution; during the parallel solution process, several linear programming sub-solvers are executed in parallel for each subproblem, and the computational threads with large computational loads are allocated more thread resources than the computational threads with small computational loads.

6. The power dispatching model solving optimization system according to claim 1, characterized in that, The introduction of a cutting plane in the feasible region includes: Introducing Gomory cut plane, covering cut plane, power flow constraint cut plane, and spatiotemporal characteristic cut plane into the feasible region; dynamically checking whether the current solution violates the cut plane during the solution process, and dynamically adding cut plane constraints if it does; and generating new cut plane constraints in real time as the search tree expands.

7. The power dispatching model solving optimization system according to claim 1, characterized in that, The multi-random variable modeling of the stochastic mixed-integer programming model in the power dispatching model includes: A two-stage stochastic programming approach is adopted, in which the first stage solves for deterministic variables, and the second stage sets the probability distribution of uncertain variables and solves for uncertain variables through stochastic scenario compensation.

8. The power dispatching model solving optimization system according to claim 1, characterized in that, The transformation of the uncertainty constraint includes: By introducing confidence levels into the uncertainty constraints, the uncertainty constraints are transformed into equivalent deterministic constraints.

9. The power dispatching model solving optimization system according to claim 1, characterized in that, The random scenario simulation and parallel solution of random scenarios include: Several initial random scenarios were generated using Monte Carlo simulation, and K-means clustering was used to cluster these initial random scenarios to obtain several typical random scenarios. Several typical random scenarios are solved in parallel to obtain the optimal solution for each typical random scenario. The scenarios are then weighted and merged to obtain the final global weighted optimal solution.

10. A method for optimizing a power dispatch model based on the power dispatch model optimization system according to claim 1, characterized in that, include: In the process of solving the power dispatch model, the bottleneck identification module analyzes the variables, constraints and branch and bound nodes of the power dispatch model, identifies the bottlenecks in solving the power dispatch model, generates a solution optimization strategy based on the bottlenecks, and optimizes the solution process based on the solution optimization strategy. In the process of solving the power dispatch model, the mixed integer programming acceleration module is used to preprocess the variables and constraints of the mixed integer programming model in the power dispatch model, introduce a heuristic search strategy to determine the priority solution variables, adjust the branch order in the branch and bound method solution process based on the adaptive branch variable selection strategy and reinforcement learning algorithm, process the sub-problems in the branch and bound method solution process in parallel based on multi-threading technology, and introduce a cutting plane in the feasible region. In the process of solving the power dispatch model, the stochastic mixed integer programming acceleration module is used to perform multiple random variable modeling, uncertainty constraint transformation, stochastic scenario simulation, parallel solution of stochastic scenarios, and adjustment of the branch order in the branch and bound method solution process based on reinforcement learning algorithm. During the solution process of the power dispatch model, the algorithm modules in the optimization engine integration module are called through a unified calling interface, and the computational resources for the power dispatch model solution process are dynamically called through the optimization engine integration module.

Images