Power grid multi-objective cooperative dispatching method and device, electronic equipment and storage medium

Abstract

This invention provides a method, apparatus, electronic device, and storage medium for multi-objective coordinated scheduling of power grids. It acquires power grid operation data and establishes a multi-time-scale scheduling optimization model based on this data. The multi-time-scale scheduling optimization model includes multiple optimization objectives. The model's decision variables are mapped to the search space of an improved cyst swarm algorithm, and a candidate scheduling scheme is encoded as the position vector of a single cyst swarm individual. The fitness of each cyst swarm individual is calculated based on the multiple optimization objectives of the multi-time-scale scheduling optimization model, and iterative optimization is performed based on the fitness. When a preset iteration termination condition is met, the position of the currently globally optimal cyst swarm individual is output. The position of the currently globally optimal cyst swarm individual is decoded to output the day-ahead and real-time coordinated scheduling scheme. This invention solves the multi-objective optimization problem using the cyst swarm algorithm, improving the solution efficiency and optimization quality of multi-objective scheduling problems in power grids under high-dimensional uncertain environments.

Description

Technical Field

[0001] This invention relates to the field of power system dispatching technology, and in particular to a method, apparatus, electronic device and storage medium for multi-objective coordinated dispatching of power grids. Background Technology

[0002] With the deepening of my country's "dual carbon" goals, the penetration rate of clean energy such as wind power and photovoltaics in the power system is constantly increasing. New energy power generation exhibits significant intermittency and volatility. While its large-scale integration improves the cleanliness of the energy structure, it also brings unprecedented challenges to the safety, stability, and economic operation of the power system. Modern power grid dispatching needs to coordinate and optimize multiple objectives, including power generation cost (economic efficiency), pollutant and carbon emissions (low-carbon efficiency), and power supply reliability, under the premise of meeting complex physical and safety constraints. This forms a high-dimensional, nonlinear, strongly constrained, and uncertain multi-timescale optimization problem. Traditional mathematical programming methods, such as linear programming and mixed-integer programming, often sacrifice accuracy when dealing with such complex problems due to the need for extensive model simplification (e.g., linearizing nonlinear constraints), or face the "curse of dimensionality" due to the large problem scale, resulting in low solution efficiency or even unsolvability. In recent years, swarm intelligence optimization algorithms have been widely used in the field of power grid dispatching due to their relaxed requirements on the problem model and their parallel global search capabilities. These algorithms simulate the cooperative behavior of biological groups in nature, seeking satisfactory solutions through iterative evolution of the population in the solution space. However, when traditional swarm intelligence algorithms are directly applied to the specific complex problem of multi-objective coordinated scheduling of power grids with a high proportion of renewable energy access, the algorithm parameter adjustment strategy is simple and it is difficult to adaptively balance the global exploration and local development capabilities throughout the entire iteration process, which can easily lead to premature convergence or low search efficiency in the later stages. When dealing with conflicts between multiple objectives such as economy, low carbon, and safety, they often rely on weighted summation or inefficient Pareto maintenance mechanisms, failing to effectively construct the co-evolutionary relationship between different objectives, resulting in poor comprehensive performance of the solution set and making it difficult to improve the engineering accuracy and practicality of the final scheduling scheme. Summary of the Invention

[0003] This invention provides a method, apparatus, electronic device, and storage medium for multi-objective collaborative scheduling of power grids, which addresses the systemic defects of existing swarm intelligence optimization algorithms in handling multi-objective collaborative scheduling problems of power grids with a high proportion of new energy sources, such as an imbalance between exploration and development capabilities, a weak multi-objective collaborative mechanism, and insufficient local search accuracy. These defects make it difficult to efficiently and accurately obtain solutions that meet practical engineering requirements.

[0004] This invention provides a multi-objective cooperative scheduling method for power grids, comprising: Acquire power grid operation data, and establish a multi-time-scale scheduling optimization model based on the power grid operation data. The multi-time-scale scheduling optimization model includes multiple optimization objectives. The decision variables of the multi-timescale scheduling optimization model are mapped to the search space of the improved cysticercosis swarm algorithm, and a candidate scheduling scheme is encoded as the position vector of a cysticercosis individual in the improved cysticercosis swarm algorithm. The fitness of each individual cystidia is calculated based on multiple optimization objectives of the multi-timescale scheduling optimization model, and iterative optimization is performed based on the fitness. When the preset iteration termination condition is met, the position of the current global optimal cysticercosis individual is output, and the corresponding day-ahead and real-time coordinated scheduling scheme is output by decoding the current global optimal cysticercosis individual position.

[0005] According to the power grid multi-objective cooperative scheduling method provided by the present invention, the iterative optimization based on the fitness includes: The population in the improved cysticercosis swarm algorithm is divided into at least three subgroups, and each subgroup is independently evaluated with fitness based on the corresponding objective function. In the process of finding the best in each subgroup, the position of individual cysts is updated by integrating elite individual guidance, subgroup leader guidance, and controllable chaotic perturbation. During the iteration process, the optimal individual information is periodically exchanged between different subgroups.

[0006] The power grid multi-objective cooperative scheduling method provided by the present invention further includes: When the fitness value of the global optimal solution does not improve by more than a preset threshold over multiple consecutive generations, a local fine-tuning adjustment is triggered. The local fine-tuning includes: constructing a search neighborhood with a dynamically shrinking radius centered on the current global optimal solution, and performing a fine-tuning search within the neighborhood; For the new solutions obtained by the fine search, a repair operator based on the degree of constraint violation is used for feasibility processing; the repair operator prioritizes adjusting the decision variables that contribute the most to the constraint violation.

[0007] According to the power grid multi-objective cooperative scheduling method provided by the present invention, the fine search in the neighborhood includes: combining objective function gradient information or using a quadratic interpolation method for searching.

[0008] According to the multi-objective cooperative scheduling method for power grids provided by the present invention, the periodic exchange of optimal cysticercosis individual information among different subgroups includes: Each preset period of iteration involves one information exchange between subgroups. The information exchange includes the current optimal solution and fitness value of each subgroup. The receiving subgroup decides with probability whether to adopt partial information from the external optimal solution to update a partial dimension of its leader, wherein the probability gradually decreases as the algorithm iterates.

[0009] According to the multi-objective cooperative scheduling method for power grids provided by the present invention, the step of updating the individual position of cysticercosis by integrating elite individual guidance, subgroup leader guidance, and controllable chaotic perturbation includes: New location ; in, For the first The individual in the first The current position on the decision variable. This is the updated position; The weighted average position of the set of elite individuals with the highest fitness values ​​in the current population, based on a preset ranking. For individuals Belonging to the The leader position of a subgroup; As an elite guiding factor; As the subgroup leadership factor; It is a chaotic perturbation factor; Inertial weight; These are chaotic variables generated based on the Logistic mapping.

[0010] According to the multi-objective cooperative scheduling method for power grids provided by the present invention, the iterative optimization based on the fitness further includes introducing a quantum rotating gate strategy to update the individual state, including: Each individual is associated with a qubit, and the state of the qubit is represented by phase. The rotation angle of the quantum rotating gate is dynamically calculated based on the relative difference between the individual's current fitness value and its historical best fitness value. The qubit state is updated based on the rotation angle to obtain a new probability amplitude, which is used to determine the bias of the individual cysticercus in searching across the decision variable dimension.

[0011] The present invention also provides a power grid multi-objective cooperative scheduling device, comprising: A module is established to acquire power grid operation data and establish a multi-time-scale scheduling optimization model based on the power grid operation data. The multi-time-scale scheduling optimization model includes multiple optimization objectives. The mapping module is used to map the decision variables of the multi-timescale scheduling optimization model to the search space of the improved cysticercosis swarm algorithm, and to encode a candidate scheduling scheme as the position vector of a cysticercosis individual in the improved cysticercosis swarm algorithm. The optimization module is used to calculate the fitness of each individual cystidia based on multiple optimization objectives of the multi-timescale scheduling optimization model, and to perform iterative optimization based on the fitness. The output module is used to output the current global optimal cysticercosis individual position when the preset iteration termination condition is met, and decode the current global optimal cysticercosis individual position to output the corresponding day-ahead and real-time coordinated scheduling scheme.

[0012] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the power grid multi-objective cooperative scheduling method as described in any of the preceding claims.

[0013] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the power grid multi-objective cooperative scheduling method described in any of the preceding claims.

[0014] The present invention provides a method, apparatus, electronic device, and storage medium for multi-objective coordinated scheduling of power grids. By acquiring power grid operation data, a multi-timescale scheduling optimization model is established based on the power grid operation data. The multi-timescale scheduling optimization model includes multiple optimization objectives. The decision variables of the multi-timescale scheduling optimization model are mapped to the search space of an improved cyst swarm algorithm. A candidate scheduling scheme is encoded as the position vector of a cyst swarm individual in the improved cyst swarm algorithm. The fitness of each cyst swarm individual is calculated based on the multiple optimization objectives of the multi-timescale scheduling optimization model, and iterative optimization is performed based on the fitness. When a preset iteration termination condition is met, the position of the currently globally optimal cyst swarm individual is output. The position of the currently globally optimal cyst swarm individual is decoded to output the corresponding day-ahead and real-time coordinated scheduling scheme. The present invention solves the multi-objective optimization problem through the cyst swarm algorithm, improving the solution efficiency and optimization quality of the multi-objective scheduling problem of power grids in high-dimensional uncertain environments. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 This is one of the flowcharts of the multi-objective cooperative scheduling method for power grids provided in the embodiments of the present invention; Figure 2 This is the second flowchart of the power grid multi-objective cooperative scheduling method provided in the embodiments of the present invention; Figure 3 This is a flowchart of the multi-subgroup cooperative optimization and quantum rotation gate strategy provided in the embodiments of the present invention; Figure 4This is a comparison curve of the convergence performance of various algorithms provided in the embodiments of the present invention; Figure 5 This is a bar chart comparing the performance of various algorithms across multiple metrics provided in this embodiment of the invention. Figure 6 This is a functional structure diagram of the power grid multi-objective cooperative scheduling device provided in an embodiment of the present invention; Figure 7 This is a functional structure diagram of the electronic device provided in the embodiments of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0018] Figure 1 The flowchart of the power grid multi-objective cooperative scheduling method provided in the embodiments of the present invention is as follows: Figure 1 As shown, the power grid multi-objective cooperative scheduling method provided in this embodiment of the invention includes: Step 101: Obtain power grid operation data, and establish a multi-time-scale scheduling optimization model based on the power grid operation data. The multi-time-scale scheduling optimization model includes multiple optimization objectives. In this embodiment of the invention, a multi-timescale scheduling optimization model is established that considers the uncertainty of new energy output, load fluctuation and power grid safety operation constraints. The model has multiple objectives, including minimizing the overall operating cost (economic efficiency), minimizing the carbon emission cost (low carbon efficiency) and maximizing the system safety margin (reliability). It also covers various constraints such as node power balance, generator output upper and lower limits, line transmission capacity, and spinning reserve, forming a high-dimensional, nonlinear mixed integer optimization problem.

[0019] Step 102: Map the decision variables of the multi-timescale scheduling optimization model to the search space of the improved cysticercosis swarm algorithm, and encode a candidate scheduling scheme as the position vector of a cysticercosis individual in the improved cysticercosis swarm algorithm. In this embodiment of the invention, the initial scale is: (For example The population of cystidia was analyzed and divided into three subgroups, each focusing on optimization objectives related to economy, low carbon footprint, and reliability. Each individual is represented by a position vector. This indicates that its dimension corresponds to the total number of decision variables in the scheduling model.

[0020] Step 103: Calculate the fitness of each individual cystidia based on the multiple optimization objectives of the multi-timescale scheduling optimization model, and perform iterative optimization based on the fitness. Step 104: When the preset iteration termination condition is met, output the current global optimal cysticercosis individual position, decode the current global optimal cysticercosis individual position and output the corresponding day-ahead and real-time coordinated scheduling scheme.

[0021] When traditional swarm intelligence algorithms are applied to the complex problem of multi-objective collaborative scheduling of power grids with a high proportion of renewable energy access, their simple parameter adjustment strategies make it difficult to adaptively balance global exploration and local development capabilities throughout the entire iteration process, which can easily lead to premature convergence or low search efficiency in the later stages. When dealing with conflicts between multiple objectives such as economy, low carbon, and safety, they often rely on weighted summation or inefficient Pareto maintenance mechanisms, failing to effectively construct the co-evolutionary relationship between different objectives, resulting in poor overall performance of the solution set. In the later stages of optimization, they lack the ability to conduct targeted and refined search and repair of the optimal solution region, making it difficult to improve the engineering accuracy and practicality of the final scheduling scheme.

[0022] The multi-objective coordinated scheduling method for power grids provided in this invention acquires power grid operation data and establishes a multi-time-scale scheduling optimization model based on the data. This model includes multiple optimization objectives. The decision variables of the multi-time-scale scheduling optimization model are mapped to the search space of an improved cyst swarm algorithm. A candidate scheduling scheme is encoded as the position vector of a cyst swarm individual in the algorithm. The fitness of each cyst swarm individual is calculated based on the multiple optimization objectives of the multi-time-scale scheduling optimization model, and iterative optimization is performed based on the fitness. When a preset iteration termination condition is met, the position of the currently globally optimal cyst swarm individual is output. The position of the currently globally optimal cyst swarm individual is decoded to output the corresponding day-ahead and real-time coordinated scheduling scheme. This invention solves the multi-objective optimization problem using the cyst swarm algorithm, improving the solution efficiency and optimization quality of the multi-objective scheduling problem for power grids in high-dimensional uncertain environments.

[0023] Based on any of the above embodiments, the iterative optimization based on the fitness includes: Step 201: Divide the population in the improved cysticercosis swarm algorithm into at least three subgroups, and evaluate each subgroup independently using the fitness of the corresponding objective function; Step 202: During the optimization process of each subgroup, the position of individual cysts is updated by integrating elite individual guidance, subgroup leader guidance, and controllable chaotic perturbation. In this embodiment of the invention, the method of updating the individual position of a cysticercosis by fusing elite individual guidance, subgroup leader guidance, and controllable chaotic perturbation includes: New location ; in, and They represent the first The individual in the first The position of the device before and after the update; This represents the average position of the top 10% of the most fit individuals in the current population. For individuals Belonging to the The leader position of a subgroup; It is a random number in the range [-1, 1].

[0024] The weights are dynamic inertia weights, which vary with the number of iterations. From initial value =0.9 linearly decreases to the final value =0.4, the calculation formula is: This represents the maximum number of iterations. This is an elite-leading factor, with a value range of [0.25, 0.65]. The subgroup leadership factor has a value range of [0.35, 0.95]. is the chaos perturbation factor, with a value range of [0.01, 0.10].

[0025] The chaotic variable is generated based on the Logistic mapping and is used to enhance the ability to escape local optima in the later stages of algorithm iteration. Its expression is: Chaotic parameters =4.0, The value range of is [0, 1].

[0026] Step 203: During the iteration process, the optimal individual information is periodically exchanged between different subgroups.

[0027] In this embodiment of the invention, the periodic exchange of optimal cysticercosis individual information among different subgroups includes: Step 2031: Iterate at preset intervals and perform information exchange between subgroups once. The information exchange includes the current optimal solution and fitness value of each subgroup. Step 2032: The subgroup receiving the information decides with probability whether to adopt part of the information of the external optimal solution to update part of the dimension of its leader, wherein the probability gradually decreases as the algorithm iterates.

[0028] In this embodiment of the invention, a preset period is defined. The value range is 8~20, and the information exchange content is the current optimal solution and fitness value of each subgroup; the subgroup receiving the information is given a probability... The decision on whether to adopt partial information from the external optimal solution to update certain dimensions of the group leader includes... The value range is [0.255, 0.515], and the initial value is... [0.20, 0.50], and gradually decreases as the algorithm iterates; the receiving subgroup has a probability of The decision is made on whether to adopt partial information from the external optimal solution to update the group leader, thereby promoting knowledge sharing and co-evolution among different optimization objectives. Finally, based on the final global optimal solution obtained from the co-optimization process, a day-ahead and real-time coordinated scheduling scheme that satisfies multiple objectives and constraints is decoded and generated, and then executed.

[0029] Traditional algorithms lack the ability to perform targeted and refined searches and repairs of the optimal solution region in the later stages of optimization. To address this issue, the multi-objective cooperative scheduling method for power grids proposed in this invention further includes: When the fitness value of the global optimal solution does not improve by more than a preset threshold over multiple consecutive generations, a local fine-tuning adjustment is triggered. The local fine-tuning includes: constructing a search neighborhood with a dynamically shrinking radius centered on the current global optimal solution, and performing a fine-tuning search within the neighborhood; In this embodiment of the invention, the fine search within the neighborhood includes: combining the gradient information of the objective function or using a quadratic interpolation method for the search.

[0030] For the new solutions obtained by the fine search, a repair operator based on the degree of constraint violation is used for feasibility processing; the repair operator prioritizes adjusting the decision variables that contribute the most to the constraint violation.

[0031] Based on any of the above embodiments, the iterative optimization based on the fitness further includes introducing a quantum rotating gate strategy to update the individual state, including: Step 301: Associate each individual with a qubit, and the state of the qubit is represented by phase; Step 302: Dynamically calculate the rotation angle of the quantum rotating door based on the relative difference between the individual's current fitness value and its historical best fitness value; Step 303: Update the qubit state based on the rotation angle to obtain a new probability amplitude. The new probability amplitude is used to determine the bias of the individual cysticercus in searching along the decision variable dimension.

[0032] In this embodiment of the invention, a quantum rotation gate strategy is designed and applied to enhance the adaptive search capability of individuals. Each individual is associated with a qubit, whose state is determined by a probability amplitude. Description. Based on the individual's current fitness. best fitness in history The relative difference, dynamically calculating the rotation angle of the quantum rotating gate. : ; in, To prevent division by zero for extremely small constants, To prevent division by zero for extremely small constants, the range of values ​​is [1×10]. -6 1×10 -3 ]; The sign function that determines the direction of rotation. The rotation angle is limited to the range of [0.05, 0.25]. After updating the qubit state using this rotation angle, the new probability amplitude guides the individual's search bias in the corresponding decision dimension, thereby balancing exploration and development.

[0033] The optimization process is enhanced by designing a local fine-tuning method, when the global optimal solution is continuous. generation( This method is triggered when the improvement of the current global optimum does not exceed 1%. It constructs a search neighborhood with a dynamically decreasing radius, centered on the current global optimum. Number of local adjustments Decreasing: Wherein, the initial radius The attenuation coefficient is 5% to 10% of the range of values ​​for the corresponding decision variable. The value range is [0.85, 0.95]. Within this neighborhood, a fine search is performed by combining gradient information or quadratic interpolation, and the feasibility of the generated new solution is corrected using a constraint violation-based repair operator.

[0034] Based on any of the above embodiments, the dispatch objective of a regional power grid with a high proportion of wind and solar power generation is to collaboratively optimize the economic costs of electricity purchase and generation, the carbon emission costs from fossil fuel consumption, and the system's power supply reliability, while meeting the hard constraints of ensuring safe system operation. This problem is characterized by high dimensionality (multiple units, multiple time periods), nonlinearity (unit cost function), uncertainty (fluctuations in new energy output), and multiple conflicting objectives (economy, low carbon, reliability). Figure 2 As shown, the specific methods of this multi-objective cooperative scheduling method for the power grid include: (1) Based on the power grid topology, unit parameters, load and new energy forecast data, market information and safety standards, construct a mathematical model that includes economic, low-carbon and reliability objectives, as well as various physical and operational constraints. This step transforms the actual scheduling problem into a mathematical problem that can be solved by an optimization algorithm.

[0035] (2) Set various parameters for the improved sac swarm algorithm (such as population size, subgroup division, control factor range, etc.). Encode the scheduling scheme (i.e., the planned output of each generator unit in each future time period) into the position vector of the individual in the algorithm, and randomly generate the initial population. For example, initialize a population of size K, K=145; divide the population into 3 subgroups, and the individuals in each subgroup are independently evaluated and optimized using the reciprocal or negative value of the economic objective function value, the low carbon objective function value, and the reliability objective function value as their fitness. The position vector of each individual represents a potential scheduling scheme.

[0036] (3) The population is divided into three subgroups: economic, low-carbon, and reliable. Each subgroup is independently evaluated based on its corresponding objective function as its fitness. In each iteration, individuals update their positions according to the update formula described in claim 3 (integrating elite guidance and chaotic perturbation). Simultaneously, a quantum rotating gate strategy is applied to each individual to adaptively adjust its search direction. Information is exchanged periodically between subgroups to promote goal coordination.

[0037] (4) When it is detected that the global optimal solution has improved only slightly over multiple generations (satisfying the triggering condition of claim 5), a local fine-tuning process is initiated. A fine search is performed in a dynamically shrinking neighborhood centered on the current optimal solution, and the feasibility of the new solution is repaired to improve the accuracy of the solution.

[0038] (5) Determine whether the algorithm iteration has reached the preset maximum number of iterations or meets other convergence criteria. If not, return to step (3) to continue optimization; if it is satisfied, terminate the iteration, decode the position vector of the finally obtained global optimal individual, and restore it to the specific day-ahead scheduling scheme, including the output plan of each unit for each time period, the power flow estimation of each line, etc.

[0039] (6) Convert the optimized day-ahead scheduling scheme into a set of control instructions that can be executed by the power grid energy management system (EMS), distribute it to each power plant for execution, and prepare for the next cycle of rolling optimization or real-time adjustment.

[0040] The above process systematically demonstrates the entire process from problem modeling and intelligent optimization to solution generation, reflecting how to transform complex power grid dispatching decision-making problems into a solvable and executable standardized process.

[0041] Based on any of the above embodiments, this invention constructs a multi-time-scale collaborative scheduling optimization model for power grid systems containing a high proportion of new energy sources. This model aims to coordinate the three major objectives of economy, low carbon emissions, and reliability, and generate the optimal power generation plan under the premise of strictly meeting the physical and security constraints of the power grid. Consider the day-ahead scheduling problem of a regional power grid, with a scheduling cycle of 1 day, divided into T=24 time periods. The system includes conventional thermal power generating units, hydropower generating units, wind farms, and photovoltaic power plants. Let the total number of generating units in the system be NG, and the planned active power output of each unit in time period t be the decision variable Pi,t. Furthermore, the system load forecast, the predicted output curves of wind and photovoltaic power, the operating cost parameters of each unit, the carbon emission coefficient, and the power grid network parameters are all known input data.

[0042] The scheduling model needs to optimize the following three objectives simultaneously: Economic objective: Minimize the total system operating cost, including power generation cost and the penalty cost for wind and solar curtailment imposed to promote the consumption of renewable energy. The objective function is expressed as: ; in, The total operating cost includes the cost of generating electricity from conventional units and the cost of penalties for curtailing wind and solar power. , , Let be the power generation cost coefficient of unit i; and These are the unit penalty cost coefficients for wind curtailment and solar curtailment, respectively. and These represent the power curtailment of wind farm j and photovoltaic power station k during time period t, respectively.

[0043] Low-carbon objective: Minimize the total carbon emission cost of the system, primarily considering the carbon emissions of thermal power units. The objective function is expressed as: in, This is the sum of the carbon emission costs of thermal power units; The carbon dioxide emission intensity (tons / megawatt-hour) of unit i. Let t be the carbon trading price (yuan / ton) for period t. The summation is performed only for thermal power units.

[0044] Reliability objective: Maximize the system's positive spinning reserve margin to enhance its ability to cope with load and renewable energy fluctuations. This objective is expressed as the sum of the ratios of the system's total available reserve capacity to the total load. in, It is the negative sum of the ratio of reserve capacity to load for each time period, in order to minimize and unify it; To maximize the technical contribution of unit i Let be the load of node n during time period t. To unify the optimization framework, this maximization objective is often transformed into a minimization problem.

[0045] The three objective functions F1, F2, and F3' mentioned above together constitute a multi-objective optimization problem, which will guide the search direction of different subgroups in subsequent improved optimization algorithms.

[0046] Any feasible scheduling scheme must meet the following physical and security constraints: Power balance constraint: In each time period, the total output of all generator units must equal the sum of the total system load and network losses. Network losses can be estimated using the B-coefficient method or based on a DC power flow model. The equality constraint can be expressed as: in, and The predicted power outputs for wind power and solar power are respectively. The total system load, Let t be the network loss during time period t.

[0047] Unit operating constraints: Upper and lower limits of output: Slope rate constraint: in and These refer to the downhill and uphill climbing capabilities of unit i, respectively.

[0048] New energy operation constraints: The actual utilization power of wind farms and photovoltaic power stations must not exceed their predicted output, and the curtailed power must be non-negative. Power grid safety constraints: A DC power flow model is used to ensure that the power flow of each transmission line does not exceed the limits. For line l, the power flow constraint is: Where Fl_max is the transmission capacity limit of line l. The power transfer distribution factor is the power injected from line l to node n.

[0049] System backup constraints: The total positive spin backup capacity of the system must meet reliability requirements. in, This represents the minimum forward rotational reserve capacity required by the system during time period t.

[0050] The objective function and constraints described above together constitute a high-dimensional, nonlinear, and complex optimization model. This model fully characterizes the multi-objective conflicts and multiple constraints faced by power system dispatching under the "dual-carbon" objective, laying a clear mathematical foundation for the subsequent design of efficient intelligent solution algorithms.

[0051] This invention employs a deeply improved sac-like swarm optimization algorithm to efficiently solve the aforementioned complex scheduling model. This algorithm significantly improves global optimization capability and convergence accuracy by constructing a multi-objective cooperative search framework, designing an adaptive update mechanism, and introducing a quantum optimization strategy. The detailed execution flow of the algorithm is as follows: Figure 3 As shown, it specifically includes: The scheduling scheme is encoded as the position vectors of individual algorithmic entities (i.e., "cysticercosis"). For a given set of... For a system with 10 generator units and a scheduling period of T, a complete scheduling scheme consists of the planned output values ​​of each unit in each time period. Define the position vector of the k-th individual. Let D be a real vector, where D = ×T, each dimension in the vector Corresponding to a specific decision variable .

[0052] During the initialization phase, an initial population of size N is randomly generated. The initial value of each dimension is within its corresponding physical output range. Generated by random uniform sampling within the internal region, with the following encoding method: ; To ensure the feasibility of the initial solution, a fast repair operator based on constraint violation degree is used to fine-tune each initial individual to satisfy the unit output constraint and power balance constraint, laying the foundation for subsequent efficient optimization.

[0053] To coordinate the independent optimization and co-evolution of the three objectives of economy, low carbon, and reliability, the entire population is divided into three subgroups: the economic subgroup. Low-carbon subgroup and reliability subgroup The size of the subgroup can be set as needed, for example, by dividing it equally.

[0054] Each subgroup uses the corresponding objective function defined in the second part as its independent fitness evaluation criterion: Fitness of economically-oriented subgroups: ; Low-carbon subgroup fitness: ; Reliability subgroup fitness: ; Within each subgroup, individuals are ranked according to their fitness, and the best individual at that time is selected as the subgroup leader. Simultaneously, the top 10% of individuals with the best overall performance across all targets are selected from the entire population to form an elite set, and their average position is denoted as... It is used to guide the overall evolutionary direction of a population.

[0055] The position update strategy is the core of the algorithm's balance between global exploration and local exploitation. For individuals in a subgroup S... Its update formula at the t-th iteration is as follows: ; The functions and settings of each component are as follows: Inertial term : Simulates an individual's ability to maintain its original motion tendency. Inertia weight. From initial value linearly decreasing to This allows for a search approach that emphasizes exploration in the early stages and development in the later stages.

[0056] Elite Guide This guides individuals to learn from the elite average level of the entire population, leveraging global high-quality information to accelerate convergence. (Elite guiding factor) The value is [0.25, 0.65].

[0057] Subgroup Leader Item Subgroup leadership factor: Drives individuals to learn towards the optimal solution of their subgroup, deepening their understanding of specific goals. The value is [0.35, 0.95].

[0058] Chaotic perturbation term Introducing controllable chaotic perturbations enhances the ability to escape local optima. Chaotic variables. Logistic mapping Generation, perturbation factor The value is [0.01, 0.10].

[0059] After the update, for the new location Perform constraint verification and feasibility repair to ensure that it meets all operational constraints.

[0060] To further improve the adaptability and precision of the search, each individual is associated with a qubit, and its state is represented as follows: Phase angle This implies information about the individual's search tendencies in the solution space.

[0061] In each iteration, based on the individual's current fitness Best fit with historical global performance The gap, dynamically calculating the adjustment angle Δ of the quantum rotating gate. : in, It is a very small constant. The sign function determines the direction of the search, whether it tends towards or away from the target. Rotation angle. The phase is limited to the range [0.05π, 0.25π]. The phase of the qubit is then updated. New probability amplitude and This will be used to modulate the search step size or random behavior of individuals in the next iteration, achieving adaptive adjustment: individuals far from the optimal solution will conduct large-scale exploration, while individuals close to the optimal solution will undergo fine-tuning.

[0062] The embodiments of the present invention also include a local fine-tuning strategy for improving the accuracy of the final solution, and an enhanced subgroup cooperation mechanism for achieving multi-objective knowledge sharing. The two work together to ensure that the algorithm can stably converge to a high-quality, high-precision global compromise solution.

[0063] As the algorithm's main search progresses into its later stages, population diversity declines, making it difficult for traditional update strategies to effectively "fine-tune" the current optimal solution. To address this issue, embodiments of the present invention also include a dynamically triggered local fine-tuning method, specifically comprising: Continuously monitor the global historical best solution The evolutionary state. If its fitness value is continuously... If the improvement in each iteration does not exceed η%, the algorithm is considered to have entered a "plateau," and the local fine-tuning module is activated. Recommended parameter settings are... ∈[15,30], η=1.0.

[0064] The adjustment process uses the current optimal solution. Centered on a given variable, a dynamically shrinking search neighborhood is constructed for each decision variable. The neighborhood radius is... It decays exponentially with increasing local adjustment number m: ; in, The initial neighborhood radius of the d-th dimension variable is usually set to the range of values ​​allowed for that variable. The attenuation factor is 5% to 10%. γ is the attenuation coefficient, ranging from [0.85, 0.95], for example, 0.9. This design allows the search range to shrink rapidly as the adjustment deepens, focusing on a fine region near the optimal solution.

[0065] Fine-grained search and repair within the neighborhood: in the constructed compact neighborhood [ , Within this area, a more deterministic strategy is employed for the search: If the objective function is differentiable, its first-order gradient information is used to perform steepest descent methods to quickly locate better points in the neighborhood. A quadratic interpolation model is constructed between the current optimal solution and its neighborhood sampling points, and new candidate solutions are obtained by finding its extreme points.

[0066] Since neighborhood search may produce infeasible solutions, a feasibility repair operator is implemented. This operator prioritizes adjusting the decision variables that contribute most to constraint violations (such as power imbalance and power flow violation), ensuring a new solution through small, directional adjustments. All operational constraints must be strictly satisfied. If the repaired solution is better than the current one... Then the replacement is complete.

[0067] The adjustment section is embedded as a separate subroutine within the main algorithm loop. It starts and runs once the trigger condition is met. Wheel (such as) =5) The "construction-search-repair" process. After completion, the updated optimal solution is reinjected into the main population, allowing the main algorithm to continue iterating or terminate. This mechanism effectively breaks the search stagnation and significantly improves the accuracy of the final scheduling scheme on complex constraint boundaries.

[0068] To promote deep synergy among the three objectives of economy, low carbon emissions, and reliability, this invention, in its embodiments, designs a more refined and adaptive synergy mechanism based on fundamental periodic information exchange. Three subgroups , , While evolving independently, it is also governed by a central coordinating controller. Every interval... (Recommended value) (∈[8,20]), the controller initiates a round of information exchange. The exchange content is the current leader solution of each subgroup. And its corresponding total objective function values ​​{F1, F2, F3}.

[0069] A subgroup such as After receiving leader information from other subgroups, a probabilistic acceptance strategy is used to decide whether to perform knowledge fusion. Acceptance probability. It is not fixed, but rather decreases adaptively as the iteration progresses: Wherein, the initial probability =0.5, =0.2. Initially, there's a high probability of encouraging cross-target exploration; later, there's a low probability of ensuring in-depth exploration of each target. If the random number is less than the current value... If so, then fusion is accepted. The fusion operation is not a complete replacement, but rather a random selection of superior external solutions, such as those from economical subgroups. The decision-making dimensions corresponding to certain groups or time periods are used to replace their own leaders. The corresponding dimension. This process is equivalent to injecting a fragment of the "low-cost" scheduling mode into the "low-carbon" search, directly guiding the algorithm to explore the Pareto equilibrium region between cost and emissions.

[0070] Through the aforementioned mechanism, the three subgroups are no longer isolated optimizers. The "cost-efficiency" information of the economic subgroup, the "emission control" information of the low-carbon subgroup, and the "safety margin" information of the reliability subgroup flow and merge in a directional and controlled manner under algorithmic control. This continuous co-evolution can more efficiently drive the species to explore complex target spaces, leading to the final globally optimal solution. It aims to become a high-quality compromise solution that performs well across multiple key metrics, rather than an extreme solution with a single objective.

[0071] Local fine-tuning and enhanced subgroup cooperation mechanisms are seamlessly integrated into the main workflow of the improved sac swarm algorithm. In each iteration, the algorithm sequentially executes: independent fitness evaluation and leader update for multiple subgroups → position update incorporating elites and chaotic perturbations → adaptive search driven by quantum rotation gates → periodic exchange of cooperative information between subgroups. When the main iteration loop ends (reaching the maximum number of iterations), the algorithm automatically checks whether the triggering conditions for local fine-tuning are met. If so, this module is activated for final optimization and "polishing." The final output... The corresponding complete scheduling scheme, which is the day-ahead optimal power generation plan obtained by the method of this invention, can be directly sent to the power grid control system for execution.

[0072] To verify the effectiveness and superiority of the method proposed in this invention, a simulation example was constructed based on a standard test system. The method of this invention (denoted as I-SBO) was compared and analyzed with several mainstream intelligent optimization algorithms. All simulations were performed in the same hardware and software environment to ensure the fairness of the comparison. The simulation results are shown in Table 1.

[0073] Table 1: Comparison of performance metrics for each algorithm (average of 30 runs) I-SBO (This invention) 128.45 2456.3 115.7 PSO 135.82 2632.1 121.5 NSGA-II 132.15 2518.7 188.3 MOGWO 134.91 2594.5 165.4 The simulation uses a modified IEEE 30-bus system as the test platform. This system comprises 41 transmission lines and, in addition to the original six traditional thermal power units, connects a wind farm at nodes 5, 11, and 13, and a photovoltaic power station at nodes 8 and 24, respectively, to simulate a high-proportion renewable energy integration scenario. The system's base power is 100 MVA, and the scheduling cycle is 24 hours, divided into 1-hour intervals. The predicted output curves for wind and photovoltaic power, as well as the total system load curve, are based on actual operating data from a specific region.

[0074] The key parameters of the algorithm (I-SBO) of this invention are set as follows: the total population size is 120, which is evenly divided into three subgroups: economic, low-carbon, and reliable, with each subgroup having a size of 40; the maximum number of iterations is set to 300; the elite guidance factor α is set to 0.45, the subgroup leadership factor β is set to 0.75, and the chaos perturbation factor δ is set to 0.05; the inertia weight is linearly decreased from 0.9 to 0.4; the subgroup information exchange cycle is 12 generations; the trigger condition for local fine-tuning adjustment is that the global optimal solution has improved by less than 1% for 20 consecutive generations; the initial radius of the neighborhood search is 8% of the variable range, and the decay coefficient is 0.9.

[0075] Four representative intelligent optimization algorithms were selected as benchmarks for comparison: Standard Particle Swarm Optimization (PSO), Non-Dominated Sorting Genetic Algorithm II (NSGA-II), Multi-Objective Gray Wolf Optimization Algorithm (MOGWO), and Basic Cystic Swarm Optimization (BSO). The population size for all comparison algorithms was set to 120, and the maximum number of iterations was 300. Their key parameters were referenced from classic literature or preliminarily optimized to ensure their performance was fully utilized.

[0076] To eliminate the influence of randomness, each algorithm was run independently 30 times, and the average value of its key performance indicators was compared. The experimental results are presented in Figure 4 as a comparison curve of the convergence performance of each algorithm and in Figure 5 as a bar chart comparing the performance of each algorithm across multiple indicators.

[0077] Multi-indicator performance analysis: (1) Overall operating cost: The overall operating cost of the I-SBO algorithm of this invention is RMB 1,284,500, which is 2.8% lower than NSGA-II (RMB 1,321,500), 4.8% lower than MOGWO (RMB 1,349,100), and 5.4% lower than PSO (RMB 1,358,200), fully demonstrating the optimization accuracy advantage of the algorithm of this invention.

[0078] (2) Carbon emissions: The carbon emissions of the I-SBO algorithm are 2456.3 tons, which is 2.5% lower than NSGA-II and 6.7% lower than PSO, verifying the effectiveness of the algorithm of this invention in achieving the goal of low-carbon scheduling.

[0079] (3) Computational efficiency: The computation time of the I-SBO algorithm is 115.7 seconds, which is significantly faster than NSGA-II (188.3 seconds) and MOGWO (165.4 seconds), demonstrating higher solution efficiency.

[0080] Comparison of convergence performance of various algorithms, for example Figure 4 As shown, the convergence performance analysis is as follows: (1) Convergence speed: The I-SBO algorithm of this invention shows a rapid decline trend in the first 20 iterations, enters a stable optimization stage around the 80th iteration, and fully converges and stabilizes at the optimal cost at the 150th iteration; while the NSGA-II, MOGWO, PSO and BSO algorithms require 180, 220, 200 and 250 iterations respectively to achieve stable convergence. The convergence speed of I-SBO is 16.7% higher than that of the second-best NSGA-II and 40% higher than that of the basic BSO algorithm, which fully demonstrates the role of elite guidance and quantum rotating door strategy in accelerating optimization.

[0081] (2) Optimization stability: The I-SBO algorithm has no obvious fluctuations after convergence and remains stable at the optimal cost of RMB 1.2845 million. In contrast, the MOGWO, PSO and other algorithms experienced a brief cost rebound during the convergence process, and the BSO algorithm still had a slight fluctuation even after convergence. This shows that the algorithm of this invention effectively avoids getting trapped in local optima through elite guidance, quantum rotating gate strategy and local fine adjustment, and has stronger optimization stability.

[0082] (3) Robustness: The convergence curve of the I-SBO algorithm is smooth overall and without drastic fluctuations. Even in the middle of the iteration (50-100 generations), it maintains a stable downward trend. This is due to the dynamic adjustment of the chaotic perturbation factor on the search process, which makes it more resistant to interference in the complex solution space.

[0083] A detailed analysis of the optimal scheduling scheme obtained by the method of this invention reveals that it makes fuller use of hydropower and new energy output during peak load periods, while rationally arranging the start-up and shutdown and output curves of thermal power units, achieving zero wind and solar curtailment, and fully meeting all network security constraints, demonstrating higher energy utilization efficiency and system operation safety.

[0084] To verify the robustness of the algorithm in the face of uncertainties in renewable energy output, additional scenarios with different prediction error levels were tested. The results show that, even with ±10% fluctuations in predicted wind and solar power output, the scheduling scheme generated by the I-SBO method exhibits a significantly smaller increase in overall operating cost compared to other comparative algorithms. This demonstrates that by introducing chaotic perturbations and adaptive mechanisms, the scheme obtained by the method of this invention possesses stronger anti-interference capabilities and adaptability, making it more suitable for uncertain operating environments in practical engineering.

[0085] The multi-objective cooperative scheduling method for power grids provided in this invention divides a sac swarm into subgroups oriented towards different objectives and designs a periodic information interaction and probability acceptance mechanism. This effectively simulates and promotes the cooperative optimization process among multiple objectives such as economy, low carbon emissions, and reliability, overcoming the shortcomings of traditional weighted summation methods where objective weights are difficult to determine and Pareto search methods are inefficient. The method innovatively integrates dynamic inertia weights, elite individual guidance, subgroup leader following, and controllable chaotic perturbations into the sac swarm position update formula. This achieves an adaptive balance between global exploration and local development capabilities throughout the iteration process, significantly improving convergence speed and reducing the risk of getting trapped in local optima. The introduced quantum rotating gate strategy dynamically maps the difference between individual fitness and the global optimum to a rotation angle, enabling intelligent adaptive adjustment of the search step size and direction, further enhancing the algorithm's optimization accuracy and robustness in complex solution spaces. The designed local fine-tuning method performs targeted fine-tuning searches of the optimal solution region in the later stages of algorithm convergence, combined with feasibility repair, effectively improving the accuracy and engineering practicality of the final scheduling scheme and ensuring its safety and reliability. An improved sac swarm algorithm that deeply integrates elite chaos mechanism, quantum intelligence strategy and co-evolution idea provides an efficient, accurate and practical new method for solving the high-dimensional, uncertain and strongly constrained multi-objective cooperative scheduling problem of power grid.

[0086] The following describes the power grid multi-objective cooperative scheduling device provided by the present invention. The power grid multi-objective cooperative scheduling device described below and the power grid multi-objective cooperative scheduling method described above can be referred to in correspondence.

[0087] Figure 6 The functional structure diagram of the power grid multi-objective cooperative scheduling device provided in the embodiments of the present invention is as follows: Figure 6 As shown, the power grid multi-objective cooperative scheduling device provided in this embodiment of the invention includes: A module 601 is established to acquire power grid operation data and establish a multi-time-scale scheduling optimization model based on the power grid operation data. The multi-time-scale scheduling optimization model includes multiple optimization objectives. The mapping module 602 is used to map the decision variables of the multi-timescale scheduling optimization model to the search space of the improved cysticercosis swarm algorithm, and to encode a candidate scheduling scheme as the position vector of a cysticercosis individual in the improved cysticercosis swarm algorithm. The optimization module 603 is used to calculate the fitness of each individual cystidia based on multiple optimization objectives of the multi-timescale scheduling optimization model, and to perform iterative optimization based on the fitness. The output module 604 is used to output the current global optimal cysticercosis individual position when the preset iteration termination condition is met, and decode the current global optimal cysticercosis individual position to output the corresponding day-ahead and real-time coordinated scheduling scheme.

[0088] This invention simulates independent search and knowledge exchange among different optimization objectives by constructing a multi-subgroup co-evolutionary architecture; effectively balances global exploration and local development by designing a hybrid update strategy that integrates elite guidance and controllable chaotic perturbation; achieves adaptive adjustment of individual search step size and direction by introducing a quantum rotation gate strategy; and significantly improves the engineering accuracy of the final solution by adding a local fine-tuning mechanism. The organic combination of these mechanisms enables the algorithm to efficiently and accurately approximate the Pareto optimal front for multi-objective systems under complex constraints.

[0089] Simulation results show that, compared with existing mainstream intelligent optimization algorithms, the method of this invention can achieve a scheduling scheme with lower overall operating cost, less carbon emissions, and higher security on a standard test system. It also shows significant advantages in terms of convergence speed, solution set quality, and robustness to uncertainty.

[0090] The multi-objective coordinated scheduling device for power grids provided in this invention acquires power grid operation data and establishes a multi-time-scale scheduling optimization model based on the data. This model includes multiple optimization objectives. The decision variables of the multi-time-scale scheduling optimization model are mapped to the search space of an improved cyst swarm algorithm. A candidate scheduling scheme is encoded as the position vector of a cyst swarm individual in the algorithm. The fitness of each cyst swarm individual is calculated based on the multiple optimization objectives, and iterative optimization is performed according to the fitness. When a preset iteration termination condition is met, the position of the currently globally optimal cyst swarm individual is output. The position of the currently globally optimal cyst swarm individual is decoded, and the corresponding day-ahead and real-time coordinated scheduling scheme is output. This invention solves multi-objective optimization problems using the cyst swarm algorithm, improving the solution efficiency and optimization quality of multi-objective scheduling problems in power grids under high-dimensional uncertain environments.

[0091] Figure 7 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 7As shown, the electronic device may include: a processor 710, a communication interface 720, a memory 730, and a communication bus 740, wherein the processor 710, the communication interface 720, and the memory 730 communicate with each other through the communication bus 740. The memory 730 includes a computer program, an operating system, and acquired data. The processor 710 can call the logical instructions in the memory 730 to execute a power grid multi-objective coordinated scheduling method, device, electronic device, and storage medium. The method includes: acquiring power grid operation data; establishing a multi-time-scale scheduling optimization model based on the power grid operation data, wherein the multi-time-scale scheduling optimization model includes multiple optimization objectives; mapping the decision variables of the multi-time-scale scheduling optimization model to the search space of an improved cysticercosis swarm algorithm; encoding a candidate scheduling scheme as the position vector of a cysticercosis individual in the improved cysticercosis swarm algorithm; calculating the fitness of each cysticercosis individual based on the multiple optimization objectives of the multi-time-scale scheduling optimization model; and iteratively optimizing based on the fitness; when a preset iteration termination condition is met, outputting the current globally optimal cysticercosis individual position; and decoding the current globally optimal cysticercosis individual position to output the corresponding day-ahead and real-time coordinated scheduling scheme.

[0092] Furthermore, the logical instructions in the aforementioned memory 730 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to related technologies, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0093] On the other hand, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the power grid multi-objective coordinated scheduling method provided by the above methods. The method includes: acquiring power grid operation data; establishing a multi-timescale scheduling optimization model based on the power grid operation data, the multi-timescale scheduling optimization model including multiple optimization objectives; mapping the decision variables of the multi-timescale scheduling optimization model to the search space of an improved cysticercosis swarm algorithm; encoding a candidate scheduling scheme as the position vector of a cysticercosis individual in the improved cysticercosis swarm algorithm; calculating the fitness of each cysticercosis individual based on the multiple optimization objectives of the multi-timescale scheduling optimization model, and iteratively optimizing based on the fitness; when a preset iteration termination condition is met, outputting the position of the current globally optimal cysticercosis individual; decoding the position of the current globally optimal cysticercosis individual and outputting the corresponding day-ahead and real-time coordinated scheduling scheme.

[0094] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0095] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of software products. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0096] 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 them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-objective cooperative scheduling method for power grids, characterized in that, include: Acquire power grid operation data, and establish a multi-time-scale scheduling optimization model based on the power grid operation data. The multi-time-scale scheduling optimization model includes multiple optimization objectives. The decision variables of the multi-timescale scheduling optimization model are mapped to the search space of the improved cysticercosis swarm algorithm, and a candidate scheduling scheme is encoded as the position vector of a cysticercosis individual in the improved cysticercosis swarm algorithm. The fitness of each individual cystidia is calculated based on multiple optimization objectives of the multi-timescale scheduling optimization model, and iterative optimization is performed based on the fitness. When the preset iteration termination condition is met, the position of the current global optimal cysticercosis individual is output, and the corresponding day-ahead and real-time coordinated scheduling scheme is output by decoding the current global optimal cysticercosis individual position.

2. The power grid multi-objective cooperative scheduling method according to claim 1, characterized in that, The iterative optimization based on the fitness includes: The population in the improved cysticercosis swarm algorithm is divided into at least three subgroups, and each subgroup is independently evaluated with fitness based on the corresponding objective function. In the process of finding the best in each subgroup, the position of individual cysts is updated by integrating elite individual guidance, subgroup leader guidance, and controllable chaotic perturbation. During the iteration process, the optimal individual information is periodically exchanged between different subgroups.

3. The power grid multi-objective cooperative scheduling method according to claim 2, characterized in that, Also includes: When the fitness value of the global optimal solution does not improve by more than a preset threshold over multiple consecutive generations, a local fine-tuning adjustment is triggered. The local fine-tuning includes: constructing a search neighborhood with a dynamically shrinking radius centered on the current global optimal solution, and performing a fine-tuning search within the neighborhood; For the new solutions obtained by the fine search, a repair operator based on the degree of constraint violation is used for feasibility processing; the repair operator prioritizes adjusting the decision variables that contribute the most to the constraint violation.

4. The power grid multi-objective cooperative scheduling method according to claim 3, characterized in that, The fine search within the neighborhood includes: combining the gradient information of the objective function or using a quadratic interpolation method for the search.

5. The power grid multi-objective cooperative scheduling method according to claim 2, characterized in that, The periodic exchange of optimal cysticercosis individual information among different subgroups includes: Each preset period of iteration involves one information exchange between subgroups. The information exchange includes the current optimal solution and fitness value of each subgroup. The receiving subgroup decides with probability whether to adopt partial information from the external optimal solution to update a partial dimension of its leader, wherein the probability gradually decreases as the algorithm iterates.

6. The power grid multi-objective cooperative scheduling method according to claim 2, characterized in that, The method of updating the individual positions of cystidia by integrating elite individual guidance, subgroup leader guidance, and controllable chaotic perturbation includes: New location ; in, For the first The individual in the first The current position on the decision variable. This is the updated position; The weighted average position of the set of elite individuals with the highest fitness values ​​in the current population, based on a preset ranking. For individuals Belonging to the The leader position of a subgroup; As an elite guiding factor; As the subgroup leadership factor; It is a chaotic perturbation factor; Inertial weights; These are chaotic variables generated based on the Logistic mapping.

7. The power grid multi-objective cooperative scheduling method according to claim 1 or 2, characterized in that, The iterative optimization based on the fitness also includes introducing a quantum rotating gate strategy to update the individual state, including: Each individual is associated with a qubit, and the state of the qubit is represented by phase. The rotation angle of the quantum rotating gate is dynamically calculated based on the relative difference between the individual's current fitness value and its historical best fitness value. The qubit state is updated based on the rotation angle to obtain a new probability amplitude, which is used to determine the bias of the individual cysticercus in searching across the decision variable dimension.

8. A power grid multi-objective cooperative dispatching device, characterized in that, include: A module is established to acquire power grid operation data and establish a multi-time-scale scheduling optimization model based on the power grid operation data. The multi-time-scale scheduling optimization model includes multiple optimization objectives. The mapping module is used to map the decision variables of the multi-timescale scheduling optimization model to the search space of the improved cysticercosis swarm algorithm, and to encode a candidate scheduling scheme as the position vector of a cysticercosis individual in the improved cysticercosis swarm algorithm. The optimization module is used to calculate the fitness of each individual cystidia based on multiple optimization objectives of the multi-timescale scheduling optimization model, and to perform iterative optimization based on the fitness. The output module is used to output the current global optimal cysticercosis individual position when the preset iteration termination condition is met, and decode the current global optimal cysticercosis individual position to output the corresponding day-ahead and real-time coordinated scheduling scheme.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the power grid multi-objective cooperative scheduling method as described in any one of claims 1 to 7.

10. A non-transitory readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the power grid multi-objective cooperative scheduling method as described in any one of claims 1 to 7.

Images