Multi-population co-evolution multi-target energy scheduling decision algorithm, system and application

Abstract

The invention relates to a multi-population co-evolution multi-target energy scheduling decision algorithm, and relates to the technical field of power system automation, and the multi-population co-evolution multi-target energy scheduling decision algorithm comprises the following steps: S1, dividing an energy system structure into a plurality of sub-populations; s2, obtaining chromosome codes of a plurality of sub-populations, and establishing an objective function to evaluate the fitness of the corresponding sub-populations; s3, iterating the sub-populations periodically to form a new generation of sub-populations, obtaining excellent individuals with high fitness in the new generation of sub-populations, and performing excellent individual exchange among different sub-populations; s4, a multi-target time-varying weight dynamic function is constructed, multi-target balance optimization is achieved by dynamically adjusting the weight, and the multi-target time-varying weight dynamic function is F = lambda (t) * C + mu (t) * R + gamma (t) * E; the problem that a traditional single population evolutionary algorithm is prone to falling into local optimum is solved, the probability of finding a global optimal solution is improved, and therefore comprehensive optimization of multiple energy scheduling targets is achieved.

Description

Technical Field

[0001] This application relates to the field of power system automation technology, and in particular to a multi-objective energy dispatch decision algorithm, system and application of multi-group co-evolution. Background Technology

[0002] In modern energy systems, such as power systems and integrated energy systems, energy dispatch decisions are crucial for achieving multiple objectives, including efficient energy utilization, cost reduction, and environmental pollution reduction. However, traditional energy dispatch methods often only optimize a single objective and struggle to simultaneously address multiple conflicting objectives, such as economic efficiency, reliability, and environmental friendliness.

[0003] As energy systems become increasingly complex, multi-objective energy dispatching problems become more challenging. On the one hand, there are complex coupling relationships among various energy forms in the energy system (such as electricity, heat, and gas), increasing the difficulty of dispatching decisions; on the other hand, different energy users have diverse needs and priorities, further exacerbating conflicts between objectives.

[0004] In recent years, evolutionary algorithms have demonstrated certain advantages in solving multi-objective optimization problems. However, traditional single-population evolutionary algorithms are prone to getting trapped in local optima and struggle to effectively handle complex problems in multi-objective energy scheduling. Summary of the Invention

[0005] Firstly, the multi-objective energy scheduling decision-making algorithm based on multi-population co-evolution provided in this application adopts the following technical solution:

[0006] A multi-objective energy scheduling decision-making algorithm based on multi-population co-evolution includes the following steps:

[0007] S1. Divide the energy system structure into multiple subgroups;

[0008] S2. Obtain the chromosome codes of multiple subpopulations and establish an objective function to evaluate the fitness of the corresponding subpopulations;

[0009] S3. Periodically iterate through the subpopulation to form a new generation of subpopulation, obtain high-fitness excellent individuals from the new generation of subpopulation, and exchange excellent individuals between different subpopulations.

[0010] S4. Construct a multi-objective time-varying weight dynamic function to achieve balanced optimization of multiple objectives by dynamically adjusting the weights. The multi-objective time-varying weight dynamic function is: F=λ(t)·C+μ(t)·R+γ(t)·E;

[0011] Wherein, λ(t)+μ(t)+γ(t)≡1, λ(t) is the economic weight coefficient, μ(t) is the reliability weight coefficient, γ(t) is the environmental weight coefficient, C is the economic impact index, R is the reliability impact index, and E is the environmental impact index.

[0012] Furthermore, the sub-species includes a production capacity population, a transmission capacity population, and a consumption capacity population, wherein the production capacity population includes a power generation population, a heating population, and a gas supply population;

[0013] S3 includes:

[0014] S31. With a period of 50 iterations, obtain the chromosome code of each generation of productive population within the period.

[0015] S32. Obtain the top 5% of the most fit individuals from any of the following populations: power generation, heating, or gas supply.

[0016] S33. Obtain the average value of the top 5% of the best individuals in each generation, select the generation with the highest average value for individual exchange, and transfer the chromosome code of the corresponding best individual to the other two populations.

[0017] S34. The corresponding population updates chromosome coding using a gene fusion formula, wherein the gene fusion formula is Gene... new =αGene mig +(1-α)·Gene local ;

[0018] Among them, Gene new For the updated chromosome encoding, Gene mig Gene is the encoding of the transmitted chromosome. local The chromosome code before the update is given, and α is the dynamic fusion factor, α∈[0.3,0.7].

[0019] Furthermore, α is related to the population convergence rate ν conv Negative correlation, α = α base -β·ν conv , where α base The base value is β, which is the attenuation coefficient.

[0020] Furthermore, during the execution of S3, evolutionary state information exchange between different subgroups is achieved based on a shared database. This evolutionary state information exchange includes:

[0021] After each subpopulation undergoes iterative evolution, the population information is stored in a shared database, including the optimal solution, average fitness, and maximum fitness.

[0022] Before each subpopulation undergoes iterative evolution, population information of other populations is obtained from a shared database.

[0023] Furthermore, when periodically iterating through the subpopulation to form a new generation of subpopulations, a roulette wheel method is used to select two individuals from the original population for crossover and mutation operations.

[0024] Furthermore, in S3, the iteration terminates when any of the following conditions are met, said conditions including:

[0025] The variance of the objective function of the subpopulation is less than 0.1% for 20 consecutive generations;

[0026] Pareto front solution set coverage Cov = |A|B| / |B|≥95%, where A is the current solution set and B is the historical best solution set;

[0027] The number of iterations reached 50.

[0028] Furthermore, the chromosome coding of the power generation population includes power generation allocation information and power generation equipment start-up and shutdown status information;

[0029] The chromosome coding of the heating population includes heating power information and heating pipeline flow information;

[0030] The chromosome coding of the gas supply population includes information on natural gas supply allocation and gas pipeline pressure regulation.

[0031] Secondly, the energy dispatching system provided in this application adopts the following technical solution:

[0032] An energy dispatching system, comprising:

[0033] Data acquisition module;

[0034] The model calculation module is equipped with the aforementioned algorithms;

[0035] The visualization module is used to generate a 3D Pareto frontier decision view.

[0036] Thirdly, this application provides an application of the above-mentioned energy dispatching system in an integrated energy system.

[0037] In summary, this application includes at least one of the following beneficial technical effects:

[0038] 1. The multi-population co-evolutionary algorithm of this invention, through the collaborative cooperation of multiple subpopulations, can effectively explore the solution space of multi-objective energy scheduling, overcoming the problem of traditional single-population evolutionary algorithms easily getting trapped in local optima, increasing the probability of finding the global optimum, and thus achieving comprehensive optimization of multiple energy scheduling objectives. Population partitioning based on energy system characteristics and targeted evolutionary strategies enable the algorithm to fully consider the characteristics and needs of different links in the energy system, making it more suitable for actual energy scheduling scenarios and improving the accuracy and effectiveness of scheduling decisions.

[0039] 2. The established collaborative mechanism promotes information sharing and cooperation among populations, accelerates the convergence speed of the algorithm, and improves the efficiency of multi-objective energy scheduling. Through the transfer of superior individuals and information interaction, each subpopulation can learn from each other, evolve together, and better achieve balanced optimization of multiple objectives.

[0040] 3. The multi-objective fusion method, combined with dynamic weight adjustment, enables the algorithm to flexibly adapt to different scheduling scenarios based on the actual operation and development needs of the energy system, achieve dynamic balance among multiple objectives, and meet the optimization needs of the energy system at different stages.

[0041] 4. The energy dispatching decision-making application system based on this algorithm realizes the automation, intelligence, and visualization of energy dispatching, improving the level and efficiency of energy management. Functions such as real-time monitoring, rapid decision-making, and scheme evaluation provide strong support for the stable operation and optimized dispatching of energy systems, helping to reduce energy costs, improve energy reliability, and reduce environmental pollution, thus promoting the development of energy systems towards high efficiency, reliability, and sustainability. Attached Figure Description

[0042] Figure 1 This is a flowchart illustrating a multi-objective energy dispatch decision-making algorithm;

[0043] Figure 2 This is a flowchart of step S3. Detailed Implementation

[0044] The following will be combined with the appendix Figure 1-2 The technical solution of this application is clearly and completely described. The following embodiments are exemplary and are only used to explain this application, and should not be construed as limiting this application. In the following description, the same reference numerals are used to denote the same or equivalent elements, and repeated descriptions are omitted.

[0045] In the description of this application, it should be understood that the terms "upper", "lower", "inner", "outer", "left", "right", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of this application is in use, or the orientation or positional relationship commonly understood by those skilled in the art. They are only used to facilitate the description of this application and to simplify the description, and are not intended to indicate or imply that the equipment or component referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0046] Furthermore, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0047] It should also be further understood that the term “and / or” as used in this application refers to any combination of one or more of the listed items, and all possible combinations thereof.

[0048] Example 1:

[0049] A multi-objective energy scheduling decision-making algorithm based on multi-population co-evolution, referring to Figure 1 This includes the following steps:

[0050] S1. Divide the energy system structure into multiple subgroups.

[0051] In this embodiment, energy specifically includes electricity, heat, and natural gas, and the sub-groups include energy production groups, energy transmission groups, and energy consumption groups. Depending on the energy type, the energy production groups include power generation groups, heat supply groups, and gas supply groups.

[0052] S2. Obtain the chromosome codes of multiple subpopulations and establish an objective function to evaluate the fitness of the corresponding subpopulations.

[0053] The power generation group comprises multiple power generation units, each corresponding to an individual. The chromosome code of the corresponding individual includes information such as power generation allocation and the start / stop status of power generation equipment. The heating group comprises multiple heating units, each corresponding to an individual. Each individual represents a different combination of heating equipment and operating parameters. The chromosome code of the corresponding individual includes information such as heating power setting and heating pipeline flow allocation. The gas supply group comprises multiple heating units, each representing a natural gas supply scheme. The chromosome code of the corresponding individual includes information such as natural gas supply allocation and gas pipeline pressure regulation.

[0054] S3. Periodically iterate through the subpopulation to form a new generation of subpopulations, obtain high-fitness, excellent individuals from the new generation of subpopulations, and exchange excellent individuals between different subpopulations.

[0055] In this embodiment, when periodically iterating the subpopulation to form a new generation of subpopulation, a roulette wheel is used to select two individuals from the atomic population for crossover and mutation operations.

[0056] In this embodiment, refer to Figure 2 S3 specifically includes:

[0057] S31. With a period of 50 iterations, obtain the chromosome code of the productive population for each generation within the period.

[0058] S32. Obtain the top 5% of the most fit individuals from any of the following populations: power generation, heating, or gas supply.

[0059] S33. Obtain the average value of the top 5% of the best individuals in each generation, select the generation with the highest average value for individual exchange, and transfer the chromosome code of the corresponding best individual to the other two populations.

[0060] For example, if the top 5% of the most fit individuals are selected from the power generation population, then these selected individuals will be transferred to the heating and gas supply populations.

[0061] Taking the iteration of power-generating populations as an example, it mainly includes:

[0062] A1. Initialize the population: Randomly generate a certain number of individuals, and the chromosome code of each individual represents a power generation scheduling scheme.

[0063] A2. Fitness Assessment: The fitness value of each individual is calculated based on power generation costs (including fuel costs, equipment maintenance costs, etc.) and power generation reliability indicators (such as the failure rate of power generation equipment, the probability of power outages, etc.). Power generation costs are obtained by summing the products of power generation and unit price of power generation. Power generation reliability indicators can be quantitatively assessed using methods such as the analytic hierarchy process (AHP).

[0064] A3. Genetic Operations: New populations are generated through genetic operations such as selection, crossover, and mutation. Selection uses roulette wheel selection, determining the probability of selection based on an individual's fitness value; crossover uses single-point or multi-point crossover to exchange some genes on an individual's chromosomes; mutation randomly mutates genes on an individual's chromosomes with a certain probability, introducing new genetic diversity.

[0065] Furthermore, both the heating and gas supply populations are evolved using particle swarm optimization (PSO), which mainly includes:

[0066] B1. Initialize the population: Randomly generate a certain number of particles for each population. The position and velocity of the particles represent the scheduling scheme and adjustment direction of heating or gas supply, respectively.

[0067] B2. Fitness Assessment: The fitness function of the heating population considers heating costs (such as fuel consumption costs, equipment operating costs, etc.) and heating quality indicators (such as indoor temperature compliance rate, heating uniformity, etc.); the fitness function of the gas supply population is calculated based on gas supply costs (including natural gas procurement costs, transportation costs, etc.) and gas supply stability indicators (such as gas pressure fluctuation range, etc.).

[0068] B3. Particle Update: Based on the particle's own historical best position and the population's global best position, the particle's velocity and position are adjusted using a velocity update formula. The velocity update formula typically includes the particle's own velocity, the component of the particle moving towards its own historical best position, and the component of the particle moving towards the population's global best position.

[0069] S34. The corresponding population updates chromosome coding through gene fusion formula.

[0070] Continuing with the example above, in the heating population, after receiving superior individuals from the power generation population, partial chromosome information from these superior individuals is fused with that of individuals in its own population. For instance, regarding the heating power allocation, the heating power settings of the heating population individuals are adjusted by combining the power allocation information related to the cogeneration units from the power generation population individuals, thereby achieving better coordination between electricity and heat production in the combined heat and power system.

[0071] In the gas supply population, the natural gas demand information related to gas-fired power generation equipment in the power generation population is used to optimize the allocation of natural gas supply among the gas supply population individuals, ensuring that the natural gas supply matches the power generation demand, while also taking into account the impact on other energy demands such as heating.

[0072] It should be noted that the above explanation of this application is based on the power generation population as an example only. When selecting the top 5% of the best individuals in terms of fitness from the heating population, they are transferred to the power generation population and the gas supply population; when selecting the top 5% of the best individuals in terms of fitness from the gas supply population, they are transferred to the power generation population and the heating population.

[0073] Furthermore, the gene fusion formula is Gene new =αGene mig +(1-α)·Gene local .

[0074] Among them, Gene new For the updated chromosome encoding, Gene mig Gene is the encoding of the transmitted chromosome. localThe chromosome code before the update is given, and α is the dynamic fusion factor, α∈[0.3,0.7].

[0075] α and population convergence rate ν conv Negative correlation, α = α base -β·ν conv , where α base The base value is 0.7 in this embodiment, and β is the attenuation coefficient (0.005 in this embodiment). in, Δt represents the optimal fitness value in generation t, and Δt represents the observation time window (10 generations in this embodiment).

[0076] Population convergence speed represents the efficiency of the population in approaching the optimal solution per unit time, reflecting the iterative effectiveness of the optimization algorithm. By establishing a correlation equation between the dynamic fusion factor and the population convergence speed, the optimal solution can be located quickly.

[0077] Correspondingly, in S3, the iteration terminates when any of the following conditions are met:

[0078] The variance of the objective function of the subpopulation is less than 0.1% for 20 consecutive generations.

[0079] Pareto front solution set coverage Cov = |A|B| / |B| ≥ 95%, where A is the current solution set and B is the historical best solution set.

[0080] The number of iterations reached 50.

[0081] Furthermore, during the execution of S3, evolutionary state information exchange between different subgroups is achieved based on a shared database. This evolutionary state information exchange includes:

[0082] After each subpopulation undergoes iterative evolution, the population information is stored in a shared database, including the optimal solution, average fitness, and maximum fitness.

[0083] Before each subpopulation undergoes iterative evolution, population information of other populations is obtained from a shared database.

[0084] S4. Construct a multi-objective time-varying weight dynamic function to achieve multi-objective balance optimization by dynamically adjusting the weights. The multi-objective time-varying weight dynamic function is: F=λ(t)·C+μ(t)·R+γ(t)·E.

[0085] Wherein, λ(t)+μ(t)+γ(t)≡1, λ(t) is the economic weight coefficient, μ(t) is the reliability weight coefficient, γ(t) is the environmental weight coefficient, C is the economic impact index, R is the reliability impact index, and E is the environmental impact index.

[0086] Continuing with the example of power-generating populations, the logic for dynamic adjustment is as follows:

[0087] During the off-season for energy supply, assuming an economic weighting coefficient of 0.6, a reliability weighting coefficient of 0.3, and an environmental weighting coefficient of 0.1, the algorithm focuses more on reducing power generation costs. By optimizing the scheduling of power generation equipment, such as adjusting power allocation and start-up / shutdown strategies, the algorithm aims to minimize power generation costs while ensuring a certain level of reliability and environmental impact.

[0088] During peak energy supply periods, to ensure a stable and reliable energy supply, the reliability weighting coefficient is increased to 0.5, the economic weighting coefficient is adjusted to 0.4, and the environmental weighting coefficient remains at 0.1. The algorithm will prioritize the reliable operation of power generation equipment and may appropriately increase some reserve power generation capacity or adopt a more conservative dispatching strategy to reduce the risk of power outages, while taking into account power generation costs and environmental impact.

[0089] Specifically, The values ​​of k1 and k2 should be set according to the actual situation.

[0090] In this embodiment, when the real-time electricity price P > the benchmark electricity price P, k1 is set to 0.6, reflecting the strong dominance of the economic objective. When the real-time electricity price ≤ the benchmark electricity price, k1 is set to 0.3 by default, at which point the weight of economic objectives decreases, while the weight of other objectives (such as environmental protection) increases.

[0091] Load factor <70%: k2 is forced to zero, and the impact of load factor on economic weight is ignored.

[0092] Load factor ≥ 70%; k2 value is positively correlated with load factor, and the specific rules are as follows:

[0093] Base value: k2 = 0.4 (default value when load factor ≥ 70%).

[0094] Segmented adjustments:

[0095] When 70% ≤ load factor < 85%: k2 is calculated using linear interpolation, and the formula is: For example, when the load factor is 80%, k2 = 0.27.

[0096] When the load factor is ≥85%, k2 takes the upper limit of 0.4, at which point the load factor's contribution to the economic weight is maximized.

[0097] By setting an economic weighting coefficient, the impact of load factors on economic objectives (such as optimizing unit output allocation) is strengthened when the load rate is high, while the load rate weight is reduced when the load rate is low, thus avoiding excessive constraints on scheduling flexibility.

[0098] Furthermore, when the load factor exceeds 85% of the peak value, μ(t) is fixed at 0.7, and the reliability of the population is improved by activating backup energy sources; under other conditions, Where, N fault N represents the number of faulty devices (e.g., transformers, transmission lines, etc.). total This represents the total number of critical system devices. μ(t) increases as the number of faulty devices increases. This is to ensure energy safety under high load or high failure rate conditions.

[0099] Furthermore, the environmental weighting coefficient of γ(t) is constructed as a function based on the actual situation, for example, reducing thermal power generation and increasing photovoltaic power generation to improve the coefficient, increasing thermal power generation and reducing photovoltaic power generation to reduce the coefficient.

[0100] Example 2:

[0101] An energy dispatching system includes a data acquisition module, a model computation module with the aforementioned algorithm deployed, and a visualization module for generating a three-dimensional Pareto front decision view.

[0102] The data acquisition module collects real-time operating parameters of the power generation equipment, such as power, temperature, and pressure, through a sensor network, as well as power flow data of the transmission network (such as voltage, current, and power factor) and real-time electricity demand data of electricity users (including electricity consumption and usage time of different types of users).

[0103] For heating and gas supply, parameters such as temperature, flow rate, and pressure of heating equipment, as well as data on pressure, flow rate, and composition of natural gas supply pipelines, are collected. This data is acquired in real time using data acquisition devices and transmitted to the system's database for storage.

[0104] Furthermore, the model computation module performs real-time calculations on the server side. Based on the collected data, the server initializes each subpopulation and performs iterative calculations according to the set evolutionary strategy and cooperative mechanism. During the computation process, the individual information of the population is continuously updated, the fitness value of each individual is evaluated, and new populations are generated through genetic operations or particle updates, until the convergence condition is met or the predetermined number of generations is reached.

[0105] In addition, the visualization module displays energy dispatching schemes in the form of charts and reports. For example, bar charts show the power distribution of different power generation units, line charts show the power flow trends of the transmission network, and tables show the distribution of users' electricity load and energy consumption.

[0106] For heating and gas supply, the interface also displays the operating status of heating equipment, heating power allocation, and information such as natural gas flow and pressure in a clear and intuitive way. Users can interact with the interface to view scheduling plans for different time periods and compare the performance indicators of different plans.

[0107] With the visualization module, users can input different scheduling scenarios and target requirements through the system interface. For example, they can set the peak electricity demand for a certain period of tomorrow, while also considering the goals of reducing power generation costs and reducing pollutant emissions.

[0108] Based on the user's input, the system quickly activates the model calculation module, runs a multi-population co-evolutionary algorithm, and generates an energy dispatch scheme that meets the requirements. The scheme includes detailed information such as the power generation plan of the power generation equipment, the power allocation of the transmission network, the adjustment strategy of the electricity load, and the supply plan of heating and gas, which is displayed to the user on the interface.

[0109] Simultaneously, the system performs multi-objective performance evaluations on the generated energy dispatch schemes. It calculates the power generation cost, reliability indicators, and environmental impact indicators for each scheme and compares them with other available schemes.

[0110] By using charts and reports to illustrate the performance differences of various solutions across different objectives, users can intuitively understand the advantages and disadvantages of each solution. For example, radar charts can be used to display the comprehensive performance of different solutions across multiple objectives such as power generation cost, reliability, and environmental impact. Users can quickly determine which solution best suits their needs based on the shape and size of the radar chart.

[0111] Building upon this, the energy dispatching system also includes a monitoring module, which monitors the operational status of the energy system in real time and promptly detects anomalies by setting thresholds and alarm rules. For example, when the power flow on a transmission line exceeds 80% of its rated capacity, the system automatically issues an overload alarm and highlights the relevant data for that line on the interface.

[0112] For power generation equipment, monitor changes in key operating parameters, such as excessively high temperatures or abnormal power fluctuations, and promptly notify maintenance personnel for inspection and handling. In heating and gas supply systems, monitor parameters such as heating temperature and gas pressure in real time to ensure the stability of energy supply.

[0113] The embodiments described in this specific implementation are preferred embodiments of this application and are not intended to limit the scope of protection of this application. Identical components are represented by the same reference numerals. Therefore, all equivalent changes made to the structure, shape, and principle of this application should be covered within the scope of protection of this application.

Claims

1. A multi-objective energy scheduling decision-making algorithm based on multi-population co-evolution, characterized in that, Includes the following steps: S1. Divide the energy system structure into multiple subgroups; S2. Obtain the chromosome codes of multiple subpopulations and establish an objective function to evaluate the fitness of the corresponding subpopulations; S3. Periodically iterate through the subpopulation to form a new generation of subpopulation, obtain high-fitness excellent individuals from the new generation of subpopulation, and exchange excellent individuals between different subpopulations. S4. Construct a multi-objective time-varying weight dynamic function to achieve balanced optimization of multiple objectives by dynamically adjusting the weights. The multi-objective time-varying weight dynamic function is: F=λ(t)·C+μ(t)·R+γ(t)·E; Wherein, λ(t)+μ(t)+γ(t)≡1, λ(t) is the economic weight coefficient, μ(t) is the reliability weight coefficient, γ(t) is the environmental weight coefficient, C is the economic impact index, R is the reliability impact index, and E is the environmental impact index.

2. The multi-objective energy scheduling decision-making algorithm based on multi-population co-evolution as described in claim 1, characterized in that, The sub-groups include a production capacity group, a transmission capacity group, and a consumption capacity group. The production capacity group includes a power generation group, a heating group, and a gas supply group. S3 includes: S31. With a period of 50 iterations, obtain the chromosome code of each generation of productive population within the period. S32. Obtain the top 5% of the most fit individuals from any of the following populations: power generation, heating, or gas supply. S33. Obtain the average value of the top 5% of the best individuals in each generation, select the generation with the highest average value for individual exchange, and transfer the chromosome code of the corresponding best individual to the other two populations. S34. The corresponding population updates chromosome coding through gene fusion formula; The gene fusion formula is Gene new =αGene mig +(1-α)·Gene local Among them, Gene new For the updated chromosome encoding, Gene mig Gene is the encoding of the transmitted chromosome. local The chromosome code before the update is given, and α is the dynamic fusion factor, α∈[0.3,0.7].

3. The multi-objective energy scheduling decision-making algorithm based on multi-population co-evolution as described in claim 2, characterized in that, The relationship between α and the population convergence rate ν conv Negative correlation, α = α base -β·ν conv , where α base The base value is β, which is the attenuation coefficient.

4. The multi-objective energy scheduling decision-making algorithm based on multi-population co-evolution as described in claim 2, characterized in that, When executing S3, evolutionary state information exchange between different subgroups is realized based on a shared database. This evolutionary state information exchange includes: After each subpopulation undergoes iterative evolution, the population information is stored in a shared database, including the optimal solution, average fitness, and maximum fitness. Before each subpopulation undergoes iterative evolution, population information of other populations is obtained from a shared database.

5. The multi-objective energy scheduling decision-making algorithm based on multi-population co-evolution as described in claim 4, characterized in that, When periodically iterating through the subpopulation to form a new generation of subpopulations, a roulette wheel method is used to select two individuals from the original population for crossover and mutation operations.

6. The multi-objective energy scheduling decision-making algorithm based on multi-population co-evolution as described in claim 5, characterized in that, In S3, the iteration terminates when any of the following conditions are met: The variance of the objective function of the subpopulation is less than 0.1% for 20 consecutive generations; Pareto front solution set coverage Cov = |A|B| / |B|≥95%, where A is the current solution set and B is the historical best solution set; The number of iterations reached 50.

7. The multi-objective energy scheduling decision algorithm based on multi-population co-evolution as described in claim 5, characterized in that, The chromosome coding of the power generation population includes power generation allocation information and power generation equipment start-up and shutdown status information; The chromosome coding of the heating population includes heating power information and heating pipeline flow information; The chromosome coding of the gas supply population includes information on natural gas supply allocation and gas pipeline pressure regulation.

8. An energy dispatching system, characterized in that, include: Data acquisition module; The model computation module is equipped with the algorithm described in any one of claims 1-7; The visualization module is used to generate a 3D Pareto frontier decision view.

9. The application of the energy dispatching system according to claim 8 in an integrated energy system.

Images