Network construction type energy storage multi-objective optimization configuration method

Abstract

The invention relates to the technical field of network construction type energy storage, in particular to a network construction type energy storage multi-objective optimization configuration method. Comprising the steps of establishing a plurality of station nodes according to regional parameters; generating a plurality of initial allocation strategies according to a preset primary processing model, and generating an initial population according to all the initial allocation strategies; and generating an optimal subset according to a preset multi-objective function F and the initial population, and generating a network construction energy storage configuration strategy according to the secondary processing model and the optimal subset. The method comprises the following steps: sequencing the configuration efficiency of each station node by quantizing the voltage support effect of network construction energy storage, constructing a multi-objective function with the lowest cost, the highest new energy utilization rate and the best transient stability effect, and solving an optimal solution set through a genetic algorithm based on non-dominated sequencing and individual congestion degree calculation. And an optimal network construction energy storage configuration scheme is determined through a grey correlation projection method, and fine distribution of the network construction energy storage capacity of each station is realized.

Description

Technical Field

[0001] The present application relates to the technical field of grid-type energy storage, and in particular to a multi-objective optimization configuration method for grid-type energy storage. Background Art

[0002] In the scenario of high proportion of renewable energy, the demand for energy flexibility has increased significantly. As a high-quality flexibility resource, electrochemical energy storage can solve the randomness and volatility problems brought by the installation of new energy, ensure the stability of the power grid, reliable power supply, and safe operation, and can greatly improve the peak load, frequency regulation and voltage regulation capabilities of the power system. Adding a grid-connected energy storage system with frequency regulation and voltage control capabilities of synchronous generators or similar synchronous generators near new energy can comprehensively improve the regulation capabilities and flexibility of the power system.

[0003] In existing R&D technologies, energy storage planning generally uses network loss as the objective function and considers various characteristics to determine the location and capacity of distributed energy storage; however, the current technology does not simultaneously consider the economic and technical aspects of energy storage operation for joint optimization, ignores the various problems faced by high-proportion new energy systems, and fails to discover a suitable multi-objective method for optimizing the capacity of grid-connected energy storage power stations.

[0004] Application Contents

[0005] The purpose of this application is: to solve the above technical problems, this application provides a multi-objective optimization configuration method for grid-type energy storage, aiming to optimize the configuration scheme of grid-type energy storage and improve the operating benefits of the grid-type energy storage system.

[0006] In some embodiments of the present application, by quantifying the voltage support role of grid energy storage, the configuration efficiency of each station node is ranked, and a multi-objective function with the lowest cost, the highest new energy utilization rate, and the best temporary stabilization effect is constructed. The optimal solution set is obtained through a genetic algorithm based on non-dominated sorting and individual congestion calculation, and the optimal grid energy storage configuration scheme is determined through the grey correlation projection method, so as to achieve refined allocation of the grid energy storage capacity of each station.

[0007] In some embodiments of the present application, a multi-objective optimization configuration method for grid-type energy storage is provided, including:

[0008] Establish multiple site nodes according to regional parameters;

[0009] Generate multiple initial allocation strategies according to a preset primary processing model, and generate initial seed clusters according to all the initial allocation strategies;

[0010] An optimal subset is generated according to the preset multi-objective function F and the initial seed cluster, and a grid-building energy storage configuration strategy is generated according to the secondary processing model and the optimal subset.

[0011] In some embodiments of the present application, multiple initial allocation strategies are generated according to a preset primary processing model, including:

[0012] Get the total rated power capacity P of the grid energy storage st.N ;

[0013] Set the first total power capacity threshold and the second total power capacity threshold

[0014] Generate an initial configuration sequence {M} according to the primary processing model;

[0015]

[0016] Where N is the number of station nodes;

[0017] The first N0 subsets of the sequence {M} are intercepted to generate the primary configuration sequence {H}, {H} = {η i},i=1,2,…,N0;

[0018] Preprocessing the first-level configuration sequence;

[0019]

[0020] Set the initial P st.N P st.min ;

[0021] Generate initial allocation strategy p st.N (i) 1 ;

[0022] p st.N (i) 1 =P st.min ·η i * ,i=1,2,…,N0.

[0023] In some embodiments of the present application, multiple initial allocation strategies are generated according to a preset primary processing model, further comprising:

[0024] set up 0≤η i * ≤1,

[0025] P st.N Divided into K files;

[0026] from Start to search;

[0027] Randomly generate N0 random numbers between [0,1], and generate a random η based on the preprocessing results i * sequence;

[0028] P st.N Multiply by η i * Generate a single initial allocation strategy;

[0029] The loop process generates multiple initial allocation strategies.

[0030] In some embodiments of the present application, generating an initial seed cluster according to all initial allocation strategies includes:

[0031] Generate and judge each initial allocation strategy according to the preset constraint model;

[0032] Generate multiple first-level allocation strategies based on the judgment results;

[0033] Generate correction parameters for each primary allocation strategy according to a preset correction model;

[0034] Generate an initial seed cluster based on the correction results.

[0035] In some embodiments of the present application, a multi-objective function F is preset, including:

[0036] Establish the total cost function model f1;

[0037] Establish energy utilization function model f2;

[0038] Establish steady-state evaluation function model f3;

[0039] Generate a multi-objective function F according to the total cost function model f1, the energy utilization function model f2 and the steady-state evaluation function model f3;

[0040] Set F = min(f1, -f2, -f3).

[0041] In some embodiments of the present application, a total cost function model f1 is established, including:

[0042] f1=k1·E st +k2·P st ;

[0043] Among them, k1 represents the cost per unit energy storage capacity, k2 represents the cost per unit energy storage power increase, and E st is the total capacity of the grid energy storage, P st is the total power of the grid energy storage.

[0044] In some embodiments of the present application, establishing an energy utilization function model f2 includes:

[0045] f2=1-ζ dump ;

[0046]

[0047] Among them, p dump (i) is the abandoned wind and solar power in period i; Δt is the duration of a single period; T is the number of periods; Y(i) is the selection coefficient; if P dump (i)-P st (i)>0,Y(i)=1; if P dump (i)-P st (i)<0,Y(i)=0;E dump is the total amount of wind and solar power abandoned after energy storage is configured; dump is the wind and solar power abandonment rate after configuring energy storage; p ideal (i) is the total power generation in the i-th period; p st (j) Energy storage power of the i-th station node in the i-th time period.

[0048] In some embodiments of the present application, a steady-state evaluation function model f3 is established, including:

[0049]

[0050] ΔU p.u (i) represents the change of the per-unit value of the system voltage before and after the i-th configuration of grid energy storage; U p.u 1 (i) U p.u 0 (i) is the per unit value of system voltage before and after configuring grid energy storage under the same transient stability fault, p st (i) is the energy storage configuration power of the i-th station node.

[0051] In some embodiments of the present application, generating an optimal subset according to a preset multi-objective function F and an initial seed cluster includes:

[0052] Generate a set of objective function values ​​for each first-level allocation strategy;

[0053] Generate a dominance relationship mapping table between each first-level allocation strategy according to all objective function value groups;

[0054] Generate a first sub-population and a second sub-population according to the dominance relationship mapping table;

[0055] Search the first sub-population and the second sub-population according to a preset optimization model;

[0056] Generate the optimal subset based on the search results.

[0057] Compared with the prior art, the multi-objective optimization configuration method for grid-type energy storage in the embodiment of the present application has the following beneficial effects:

[0058] By quantifying the voltage support role of grid energy storage, the configuration efficiency of each station node is ranked, and a multi-objective function with the lowest cost, the highest new energy utilization rate, and the best temporary stabilization effect is constructed. The optimal solution set is obtained through a genetic algorithm based on non-dominated sorting and individual congestion calculation, and the optimal grid energy storage configuration scheme is determined through the grey correlation projection method, realizing the refined allocation of the grid energy storage capacity of each station. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It is a flow chart of a multi-objective optimization configuration method for grid-type energy storage in a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0061] The specific implementation methods of the present application are further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present application but are not intended to limit the scope of the present application.

[0062] In the description of the present application, it should be understood that the terms "center", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present application.

[0063] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0064] In the description of this application, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0065] like Figure 1As shown, a multi-objective optimization configuration method for grid-type energy storage in a preferred embodiment of the present application includes:

[0066] S101: Establish multiple station nodes according to regional parameters;

[0067] S102: generating a plurality of initial allocation strategies according to a preset primary processing model, and generating an initial seed cluster according to all the initial allocation strategies;

[0068] S103: Generate an optimal subset according to the preset multi-objective function F and the initial seed cluster, and generate a grid-building energy storage configuration strategy according to the secondary processing model and the optimal subset.

[0069] Specifically, multiple station nodes are established according to the parameters of the new energy stations in the power grid area to be planned, and a single station node represents a new energy station.

[0070] Specifically, multiple initial allocation strategies are generated according to the preset primary processing model, including:

[0071] Get the total rated power capacity P of the grid energy storage st.N ;

[0072] Set the first total power capacity threshold and the second total power capacity threshold

[0073] Generate an initial configuration sequence {M} according to the primary processing model;

[0074]

[0075] Where N is the number of station nodes;

[0076] The first N0 subsets of the sequence {M} are intercepted to generate the primary configuration sequence {H}, {H} = {η i},i=1,2,…,N0;

[0077] Preprocessing the first-level configuration sequence;

[0078]

[0079] Set the initial P st.N P st.min ;

[0080] Generate initial allocation strategy p st.N (i) 1 ;

[0081] p st.N (i) 1 =P st.min ·η i *,i=1,2,…,N0.

[0082] Specifically, the voltage support role of grid energy storage is quantified according to the primary processing model, and the configuration efficiency of each station node is ranked to generate the most efficient configuration sequence {M} of grid energy storage.

[0083] Specifically, according to a single initial allocation strategy, the output power p corresponding to the grid energy storage can be further obtained: st (i) e st (i) and multi-objective function solutions.

[0084] Specifically, multiple initial allocation strategies are generated according to the preset primary processing model, including:

[0085] set up 0≤η i * ≤1,

[0086] P st.N Divided into K files;

[0087] from Start to search;

[0088] Randomly generate N0 random numbers between [0,1], and generate a random η based on the preprocessing results i * sequence;

[0089] P st.N Multiply by η i * Generate a single initial allocation strategy;

[0090] The loop process generates multiple initial allocation strategies.

[0091] Specifically, the initial sample p st.N (i) 1 It is only a potential better solution and cannot characterize the characteristics of the entire feasible solution set. Therefore, it is necessary to expand the initial calculation sample and evenly select values ​​in the feasible solution set to achieve the goal of getting as close to the optimal solution as possible.

[0092] The feasible solution set matrix is ​​a (N0)-dimensional matrix, expressed as follows:

[0093] [P st.N η1 * η2 * … η N0 * ]→[p st.N (1) p st.N (2) … p st.N(N0)]

[0094] in, 0≤η i * ≤1, The smaller the serial number, the more efficient the energy storage configuration.

[0095] P st.N Divided into K files;

[0096] from Start to Search, and randomly generate N0 random numbers between [0,1], normalize them and rearrange them in descending order to generate a random η i * sequence, P st.N Multiply by η i * A required feasible solution matrix can be generated, thereby constructing an initial allocation strategy.

[0097] In a preferred embodiment of the present application, generating an initial seed cluster according to all initial allocation strategies includes:

[0098] Generate and judge each initial allocation strategy according to the preset constraint model;

[0099] Generate multiple first-level allocation strategies based on the judgment results;

[0100] Generate correction parameters for each primary allocation strategy according to a preset correction model;

[0101] Generate an initial seed cluster based on the correction results.

[0102] Specifically, a single initial allocation strategy includes the rated power capacity of each site node.

[0103] Specifically, the constraint model includes:

[0104] 1) Economic constraints

[0105] The economic constraint on the total cost of grid energy storage can be expressed as st.max The constraints are expressed as follows:

[0106] f1=k1·E st +k2·P st ≤Ψ

[0107] 2) Actual energy storage power and capacity constraints

[0108] The actual power of the grid-building energy storage of each station does not exceed the rated power. If the abandoned power is less than the rated power of the energy storage, the new energy power will be stored according to the abandoned power. i is the station number:

[0109]

[0110] The actual storage capacity of the grid-connected energy storage at each station shall not exceed the rated capacity. st (i) Charging until the rated capacity of the energy storage is reached or the power abandonment is cut off:

[0111]

[0112] 3) Constraints on the rate of abandonment of new energy wind and solar power

[0113] The power grid has certain constraints on the wind and solar power abandonment rate of renewable energy, which is generally required not to exceed 5% or 10%. Considering that the wind and solar power abandonment rate itself is included in the multi-objective optimization function, it can be optimized to 10% to prevent the loss of potential excellent solutions during the optimization process:

[0114] ζ dump ≤10%

[0115] At the same time, considering reasonable economic efficiency, it is not necessary to completely store the limited new energy through energy storage, so the wind and solar power abandonment rate needs to be greater than 0, with a lower limit of 2%:

[0116] ζ dump ≥ 2%

[0117] According to dump The upper and lower limits of the grid energy storage power can be obtained by stmin and P stmax .

[0118] 4) Rated energy storage power and capacity constraints

[0119] The sum of the rated power of the grid-connected energy storage of each station must meet the power capacity constraint:

[0120]

[0121] At the same time, the sum of the rated capacity of the grid-connected energy storage of each station must meet the energy capacity constraint:

[0122]

[0123] Specifically, the rated power sequence corresponding to each initial allocation strategy must satisfy all the above constraints. If not, the corresponding initial allocation strategy must be eliminated.

[0124] Specifically, since the initial population is randomly generated, some solutions may be too dense or too sparse in the solution space. Only uniformly distributed solutions can more effectively represent the characteristics of the overall solution space. Therefore, it is necessary to define the distance between each point:

[0125]

[0126] Search for the nearest point of each point and calculate the distance between two points as a representation of the density of the point.

[0127] Screen the points with unreasonable density in the initial group, delete some points that are too dense, and take the median to supplement the points that are too far away:

[0128] p st.N (i) xy =(p st.N (i) x +p st.N (i) y ) / 2;

[0129] The initial population is reorganized according to this principle to ensure that it is distributed as evenly as possible in the solution space.

[0130] In a preferred embodiment of the present application, a multi-objective function F is preset, including:

[0131] Establish the total cost function model f1;

[0132] Establish energy utilization function model f2;

[0133] Establish steady-state evaluation function model f3;

[0134] Generate a multi-objective function F according to the total cost function model f1, the energy utilization function model f2 and the steady-state evaluation function model f3;

[0135] Set F = min(f1, -f2, -f3).

[0136] Specifically, the total cost function model f1 is established, including:

[0137] f1=k1·E st +k2·P st ;

[0138] Among them, k1 represents the cost per unit energy storage capacity, k2 represents the cost per unit energy storage power increase, and E st is the total capacity of the grid energy storage, P st is the total power of the grid energy storage.

[0139] Specifically, the energy utilization function model f2 is established, including:

[0140] f2=1-ζ dump ;

[0141]

[0142] Among them, p dump(i) is the abandoned wind and solar power in period i; Δt is the duration of a single period; T is the number of periods; Y(i) is the selection coefficient; if P dump (i)-P st (i)>0,Y(i)=1; if P dump (i)-P st (i)<0,Y(i)=0;E dump is the total amount of wind and solar power abandoned after energy storage is configured; dump is the wind and solar power abandonment rate after configuring energy storage; p ideal (i) is the total power generation in the i-th period; p st (j) Energy storage power of the i-th station node in the i-th time period.

[0143] Specifically, a steady-state evaluation function model f3 is established, including:

[0144]

[0145] ΔU p.u (i) represents the change of the per-unit value of the system voltage before and after the i-th configuration of grid energy storage; U p.u 1 (i) U p.u 0 (i) is the per unit value of system voltage before and after configuring grid energy storage under the same transient stability fault, p st (i) is the energy storage configuration power of the i-th station node.

[0146] It can be understood that in the above embodiment, the optimization goal is to maximize the overall new energy absorption capacity by building network energy storage, and a multi-objective function with the lowest cost, the highest new energy utilization rate, and the best temporary stabilization effect is constructed. All initial allocation strategies are judged to achieve refined allocation of the network energy storage capacity of each station.

[0147] In a preferred embodiment of the present application, the optimal subset is generated according to the preset multi-objective function F and the initial seed cluster, including:

[0148] Generate a set of objective function values ​​for each first-level allocation strategy;

[0149] Generate a dominance relationship mapping table between each first-level allocation strategy according to all objective function value groups;

[0150] Generate a first sub-population and a second sub-population according to the dominance relationship mapping table;

[0151] Search the first sub-population and the second sub-population according to a preset optimization model;

[0152] Generate the optimal subset based on the search results.

[0153] The initial population pst.N (i) x Substitute the multi-objective function F into the calculation, and we can obtain a set of objective function values ​​(f1, -f2, -f3) of the same population size. There is a mutual dominance relationship between the function values. Definition: 1) f1 x <f1 y , -f2 x <-f2 y and -f3 x <-f3 y In the case of x Completely superior to F y ; 2)f1 x >f1 y , -f2 x >-f2 y and -f3 x >-f3 y In the case of x Completely inferior to F y ; 3) If the above two relationships are not satisfied, then F x With F y There is no dominance relationship.

[0154] According to the above definition, the objective function value F is stratified and carried out in a cyclic manner. In the first round, all function values ​​are screened and the sub-population with the highest dominance order is selected. Any individual in this sub-population cannot be dominated by any other individual, and there is no mutual dominance relationship within the sub-population; then the sub-population is deleted from the population, and the sub-population that cannot be dominated by other individuals is selected again as the sub-population with the second dominance order; the steps are subsequently repeated until all individuals in the population are traversed.

[0155] According to the above non-dominated sorting method, the sub-populations with the first and second dominant orders in the objective function value F are selected as the optimal objective function value set corresponding to the initial population. At the same time, the initial population subset corresponding to the sub-population is used as the elite subset to enter the initial population for the next round of fine search.

[0156] Specifically, the subpopulation with the first non-dominated order is set as the first subpopulation, the subpopulation with the second non-dominated order is set as the second subpopulation, and the first subpopulation and the second subpopulation are set as elite subsets. It is believed that they are relatively close to the optimal network energy storage configuration in the solution space, so a second round of fine search is carried out near this subset.

[0157] Crossover operator: select an individual from the elite subset, average it with the individual closest to it, and take the average of the two as the new feasible solution;

[0158] Mutation operator: select an individual from the elite subset, mutate any of its values, and use the mutated value as a new feasible solution;

[0159] Perform crossover mutation operation on the elite subset, incorporate the newly generated subset into the elite subset, repeat the above steps until the first subset in non-dominated order does not change or the cycle reaches a limited number of times, and output the first subset in non-dominated order as the final optimal subset.

[0160] Specifically, in the optimal subset, any two individuals cannot dominate each other, which means that it is impossible to simply identify a certain individual as optimal. The grey correlation projection method is used to establish a secondary processing model to determine the optimal individual. The network energy storage configuration strategy is generated based on the allocation strategy corresponding to the optimal individual.

[0161] According to the first concept of the present application, by quantifying the voltage support role of grid energy storage, the configuration efficiency of each station node is ranked, and a multi-objective function with the lowest cost, the highest new energy utilization rate, and the best temporary stabilization effect is constructed. The optimal solution set is obtained through a genetic algorithm based on non-dominated sorting and individual congestion calculation, and the optimal grid energy storage configuration scheme is determined through the grey correlation projection method.

[0162] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and substitutions can be made without departing from the technical principles of the present application. These improvements and substitutions should also be regarded as the scope of protection of the present application.

Claims

1. A multi-objective optimization configuration method for grid-type energy storage, characterized in that: include: Establish multiple site nodes according to regional parameters; Generate multiple initial allocation strategies according to a preset primary processing model, and generate initial seed clusters according to all the initial allocation strategies; An optimal subset is generated according to the preset multi-objective function F and the initial seed cluster, and a grid-building energy storage configuration strategy is generated according to the secondary processing model and the optimal subset.

2. The multi-objective optimization configuration method for grid-type energy storage according to claim 1, characterized in that: Generate multiple initial allocation strategies based on the preset primary processing model, including: Get the total rated power capacity P of the grid energy storage st.N ; Set the first total power capacity threshold P stmin and the second total power capacity threshold P stmax ; Generate an initial configuration sequence {M} according to the primary processing model; Where N is the number of station nodes; The first N0 subsets of the sequence {M} are intercepted to generate the primary configuration sequence {H}, {H} = {η i },i=1,2,…,N0; Preprocessing the first-level configuration sequence; Set the initial P st.N P st.min ; Generate initial allocation strategy p st.N (i) 1 ; p st.N (i) 1 =P st.min ·η i * ,i=1,2,…,N0。 3. The multi-objective optimization configuration method for grid-type energy storage according to claim 2, characterized in that: Generate multiple initial allocation strategies based on the preset primary processing model, including: Set P stmin ≤P st.N ≤P stmax , P st.N Divided into K files; from Start to search; Randomly generate N0 random numbers between [0,1], and generate a random η based on the preprocessing results i * sequence; P st.N Multiply by η i * Generate a single initial allocation strategy; The loop process generates multiple initial allocation strategies.

4. The multi-objective optimization configuration method for grid-type energy storage according to claim 3, characterized in that: Generate an initial seed population according to all initial allocation strategies, including: Generate and judge each initial allocation strategy according to the preset constraint model; Generate multiple first-level allocation strategies based on the judgment results; Generate correction parameters for each primary allocation strategy according to a preset correction model; Generate an initial seed cluster based on the correction results.

5. The multi-objective optimization configuration method for grid-type energy storage according to claim 3, characterized in that: The preset multi-objective function F includes: Establish the total cost function model f1; Establish energy utilization function model f2; Establish steady-state evaluation function model f3; Generate a multi-objective function F according to the total cost function model f1, the energy utilization function model f2 and the steady-state evaluation function model f3; Set F = min(f1, -f2, -f3).

6. The multi-objective optimization configuration method for grid-type energy storage according to claim 5, characterized in that: Establish the total cost function model f1, including: f1=k1·E st +k2·P st ; Among them, k1 represents the cost per unit energy storage capacity, k2 represents the cost per unit energy storage power increase, and E st is the total capacity of the grid energy storage, P st is the total power of the grid energy storage.

7. The multi-objective optimization configuration method for grid-type energy storage according to claim 6, characterized in that: Establish energy utilization function model f2, including: f2=1-ζ dump ; Among them, p dump (i) is the abandoned wind and solar power in period i; Δt is the duration of a single period; T is the number of periods; Y(i) is the selection coefficient; if P dump (i)-P st (i)>0,Y(i)=1; if P dump (i)-P st (i)<0,Y(i)=0;E dump is the total amount of wind and solar power abandoned after energy storage is configured; dump is the wind and solar power abandonment rate after configuring energy storage; p ideal (i) is the total power generation in the i-th period; p st (j) Energy storage power of the i-th station node in the i-th period.

8. The multi-objective optimization configuration method for grid-type energy storage according to claim 7, characterized in that: Establish a steady-state evaluation function model f3, including: ΔU p.u (i) represents the change of the per-unit value of the system voltage before and after the i-th configuration of grid energy storage; U p.u 1 (i) U p.u 0 (i) is the per unit value of system voltage before and after configuring grid energy storage under the same transient stability fault, p st (i) is the energy storage configuration power of the i-th station node.

9. The multi-objective optimization configuration method for grid-type energy storage according to claim 8, characterized in that: Generate the optimal subset according to the preset multi-objective function F and the initial seed cluster, including: Generate a set of objective function values ​​for each first-level allocation strategy; Generate a dominance relationship mapping table between each first-level allocation strategy according to all objective function value groups; Generate a first sub-population and a second sub-population according to the dominance relationship mapping table; Search the first sub-population and the second sub-population according to a preset optimization model; Generate the optimal subset based on the search results.

Images