A power grid digitalization project pareto optimization method and system

Abstract

This invention relates to a Pareto optimization method and system for power grid digitization projects. The method includes: establishing an evaluation index system and a shared network among projects based on quantitative evaluation information, technical characteristics of the projects, and development strategy information of the power grid company; establishing a multi-objective optimization model for the digitization project group based on the evaluation index system and the shared network, and solving for the Pareto optimal solution set; mining association rules based on the Pareto optimal solution set, and ranking the association rules using support, confidence, and lift as indicators; filtering the pre-obtained non-dominated solution set according to the ranking of the association rules, removing the worst solution in each filtering until only one optimal digitization project combination solution remains, thus completing the optimization of the digitization projects. This invention solves the project portfolio optimization problem considering project interactions.

Description

Technical Field

[0001] This invention relates to the field of power grid digitization project optimization, and in particular to a Pareto optimization method and system for power grid digitization projects. Background Technology

[0002] The realization of enterprise digitalization is a gradual process. Digitalization projects are the concrete actions taken to achieve enterprise digitalization. Selecting suitable digitalization projects from numerous projects with varying investment amounts, effectiveness, and complex relationships within a limited investment and timeframe is crucial to ensuring that enterprises can quickly and effectively improve their digitalization level and transform investment into enterprise value and benefits. Different digitalization projects have different functions, and therefore the indicators for evaluating them also differ. Almost all project benefit indicators include multiple levels and different metrics. Therefore, simply ranking projects according to a single comprehensive evaluation value is incomplete; it is necessary to optimize digitalization projects from multiple dimensions and under multiple evaluation objectives. Project portfolio selection has gained increasing interest and attention in the public administration field, including industrial companies, enterprises, and the military. The focus is on selecting project proposals with limited resources to maximize stakeholder benefits using multiple evaluation criteria. However, practitioners face two major challenges when selecting the best project portfolio. First, the extensive interactions between projects affect the actual value and risk of the project portfolio. Secondly, portfolio optimization always involves multiple objectives. Commonly used multi-objective optimization methods can effectively find non-dominant solutions, but they introduce new problems: how to further select the optimal portfolio from these solutions. Therefore, a fine-grained selection method for portfolio optimization based on project interactions is of significant research importance.

[0003] Existing literature on portfolio research and development can be summarized as follows: 1) Dividing projects into tasks, focusing on how to effectively implement projects; 2) Evaluating project value to determine funding policies aimed at maximizing overall project utility; 3) Analyzing how project synergies affect project value and expected performance. In this study, the focus of portfolio selection is project planning rather than project engineering. Literature shows that portfolio optimization attracts more attention than project interaction. This can be divided into single-objective optimization and multi-objective optimization. The latter has received more extensive research in the literature. Some studies group multi-objective optimization with multi-criteria decision-making because they directly transform multi-objectives into single-objectives through weighted operations.

[0004] Widely used algorithms for solving multi-objective problems include Non-Dominated Sorting Genetic Algorithm (NSGA), non-dominated sorting and local search, Strength Pareto Evolutionary Algorithm (SPEA), niche Pareto Genetic Algorithm, and decomposition-based multi-objective evolutionary algorithms, among others. These algorithms have been successfully applied in various fields. However, since decision-makers always expect only one solution, a common problem with these algorithms is that they are committed to finding the optimal Pareto set, rather than the optimal solution itself. Therefore, how to further refine the Pareto set to find the optimal solution is a pressing technical problem that needs to be solved. Summary of the Invention

[0005] To address the aforementioned problems, the purpose of this invention is to provide a Pareto optimization method and system for network digital projects, which overcomes the subjective experience of existing technologies, solves the problem of project portfolio optimization considering project interactions, and can be directly applied to enterprise practice.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a Pareto optimization method for power grid digitization projects, comprising: establishing an evaluation index system and a shared network among power grid digitization projects based on quantitative evaluation information, technical characteristics of power grid digitization projects, and development strategy information of power grid companies; establishing a multi-objective optimization model for a group of digitization projects based on the evaluation index system and the shared network among projects, and solving for the Pareto optimal solution set; mining association rules based on the Pareto optimal solution set, and ranking the association rules using support, confidence, and lift as indicators; filtering the pre-obtained non-dominated solution set according to the ranking of association rules, removing the worst solution in each filtering until only one optimal combination of digitization projects remains, thus completing the optimization of digitization projects.

[0007] Furthermore, establish a shared network between projects, including:

[0008] By mapping two projects that share a network, it can be determined whether the two projects have a shared use relationship.

[0009] For two projects with a shared utilization relationship, the number of technologies used simultaneously is set as the weight of the synergistic utilization advantage between the projects, thus constructing a project sharing network. ,in Indicates project The technologies required simultaneously with project y This indicates the number of technologies used in all projects.

[0010] Furthermore, the establishment of a multi-objective optimization model for a digital project group, and the solution to obtain the Pareto optimal solution set, includes:

[0011] Project portfolio solution The initial population is generated by the chromosomes of each individual in the population.

[0012] Establish and compute the fitness model and penalty function, and determine the Pareto rank in the solution set;

[0013] The initial population is sorted using a non-dominated ordering method, and all Pareto ranks are assigned.

[0014] After the initial population is sorted non-dominated, crossover, mutation, and selection operations of the genetic algorithm are performed to generate new individuals and add them to the population.

[0015] From all populations Select from individuals Individuals are grouped into groups. Based on the fitness of each individual, the individual with the best fitness value is selected to enter the offspring population. This process is repeated until the number of individuals in the offspring population reaches a certain threshold. , Population size;

[0016] Calculate the crowding degree of all individuals to make the obtained solution more uniform in the target space;

[0017] Based on the Pareto level and crowding, the parent and offspring populations are combined into a new population. A new parent population is generated from the new population according to pre-set rules, and this process is repeated until the set termination conditions are met.

[0018] Furthermore, the non-dominated sorting of the initial population, which divides all population levels, includes:

[0019] Calculate each individual Number of dominated and the set of solutions dominated by that individual. Traverse the entire population; the population size is The computational complexity of this parameter is . ;

[0020] Parameters in the population Individuals are placed into the first group set. In the middle, delete set and After identifying the individuals in the list, recalculate the number of dominated individuals for all remaining individuals. Then Individuals are placed into the second type of cluster and In this process, the process is repeated until all population ranks are defined.

[0021] Furthermore, the pre-defined rules include:

[0022] Based on the Pareto hierarchy from low to high, the entire population layer is placed into the parent population. Until a certain layer of individuals can no longer be fully incorporated into the parent population. ;

[0023] The individuals in this layer are arranged from most crowded to least crowded and then placed into the parent population in sequence. In the middle, until the parent population Fill it up.

[0024] Furthermore, the ranking of the association rules adopts the TOPSIS method, including:

[0025] The decision matrix is ​​obtained by normalizing the standard matrix composed of support, confidence, and lift.

[0026] Calculate the ideal solution based on the decision matrix. and negative ideal solution ;

[0027] According to the ideal solution and negative ideal solution Calculate frequent sets The distance between the standard point and the ideal point;

[0028] The relative distance from each standard point to the negative ideal solution is obtained from the distance between the standard point and the ideal point. , This describes the importance of relatively frequent items, based on relative distance. Obtain the most frequent association rules.

[0029] Furthermore, the step of filtering the pre-obtained non-dominated solution set according to the order of association rules, removing the worst solution at each stage until only one optimal digitized item combination solution remains, includes:

[0030] Analysis of the first Each association rule, in the association rule related set The system stores the project group scheme that contains this association rule;

[0031] Filtering the set of non-dominated project groups ,from Remove items not included in the set and The elements in the set are updated, and the empty set is updated. ;

[0032] Check from Filter the remaining project portfolio solutions ,if If only one project portfolio solution remains, stop and output. Otherwise, let Repeated filtering.

[0033] A Pareto optimization system for power grid digitization projects includes: a first processing module, which establishes an evaluation index system and a shared network among power grid digitization projects based on quantitative evaluation information, technical characteristics of power grid digitization projects, and development strategy information of power grid companies; a second processing module, which establishes a multi-objective optimization model for a group of digitization projects based on the evaluation index system and the shared network among projects, and solves for the Pareto optimal solution set; a third processing module, which mines association rules based on the Pareto optimal solution set, and ranks the association rules using support, confidence, and lift as indicators; and a filtering module, which filters the pre-obtained non-dominated solution set according to the ranking of association rules, removing the worst solution in each filtering until only one optimal combination of digitization projects remains, thus completing the optimization of digitization projects.

[0034] Furthermore, establish a shared network between projects, including:

[0035] By mapping two projects that share a network, it can be determined whether the two projects have a shared use relationship.

[0036] For two projects with a shared utilization relationship, the number of technologies used simultaneously is set as the weight of the synergistic utilization advantage between the projects, thus constructing a project sharing network. ,in Indicates project The technologies required simultaneously with project y This indicates the number of technologies used in all projects.

[0037] Furthermore, the establishment of a multi-objective optimization model for a digital project group, and the solution to obtain the Pareto optimal solution set, includes:

[0038] Project portfolio solution The initial population is generated by the chromosomes of each individual in the population.

[0039] Establish and compute the fitness model and penalty function, and determine the Pareto rank in the solution set;

[0040] The initial population is sorted using a non-dominated ordering method, and all Pareto ranks are assigned.

[0041] After the initial population is sorted non-dominated, crossover, mutation, and selection operations of the genetic algorithm are performed to generate new individuals and add them to the population.

[0042] From all populations Select from individuals Individuals are grouped into groups. Based on the fitness of each individual, the individual with the best fitness value is selected to enter the offspring population. This process is repeated until the number of individuals in the offspring population reaches a certain threshold. , Population size;

[0043] Calculate the crowding degree of all individuals to make the obtained solution more uniform in the target space;

[0044] Based on the Pareto level and crowding, the parent and offspring populations are combined into a new population. A new parent population is generated from the new population according to pre-set rules, and this process is repeated until the set termination conditions are met.

[0045] Furthermore, the non-dominated sorting of the initial population, which divides all population levels, includes:

[0046] Calculate each individual Number of dominated and the set of solutions dominated by that individual. Traverse the entire population; the population size is The computational complexity of this parameter is . ;

[0047] Parameters in the population Individuals are placed into the first group set. In the middle, delete set and After identifying the individuals in the list, recalculate the number of dominated individuals for all remaining individuals. Then Individuals are placed into the second type of cluster and In this process, the process is repeated until all population ranks are defined.

[0048] Furthermore, the pre-defined rules include:

[0049] Based on the Pareto hierarchy from low to high, the entire population layer is placed into the parent population. Until a certain layer of individuals can no longer be fully incorporated into the parent population. ;

[0050] The individuals in this layer are arranged from most crowded to least crowded and then placed into the parent population in sequence. In the middle, until the parent population Fill it up.

[0051] Furthermore, the ranking of the association rules adopts the TOPSIS method, including:

[0052] The decision matrix is ​​obtained by normalizing the standard matrix composed of support, confidence, and lift.

[0053] Calculate the ideal solution based on the decision matrix. and negative ideal solution ;

[0054] According to the ideal solution and negative ideal solution Calculate frequent sets The distance between the standard point and the ideal point;

[0055] The relative distance from each standard point to the negative ideal solution is obtained from the distance between the standard point and the ideal point. , This describes the importance of relatively frequent items, based on relative distance. Obtain the most frequent association rules.

[0056] Furthermore, the step of filtering the pre-obtained non-dominated solution set according to the order of association rules, removing the worst solution at each stage until only one optimal digitized item combination solution remains, includes:

[0057] Analysis of the first Each association rule, in the association rule related set The system stores the project group scheme that contains this association rule;

[0058] Filtering the set of non-dominated project groups ,from Remove items not included in the set and The elements in the set are updated, and the empty set is updated. ;

[0059] Check from Filter the remaining project portfolio solutions ,if If only one project portfolio solution remains, stop and output. Otherwise, let Repeated filtering.

[0060] A computer-readable storage medium storing one or more programs, the one or more programs including instructions that, when executed by a computing device, cause the computing device to perform any of the methods described above.

[0061] A computing device includes: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for performing any of the methods described above.

[0062] The present invention has the following advantages due to the adoption of the above technical solutions:

[0063] 1. The digital project cluster optimization method used in this invention has many advantages over traditional multi-objective decision-making methods. This invention establishes a multi-objective optimization model for digital project clusters from a shared network, rather than determining the degree of project interaction through subjective experience. The model of this invention has been proven to be effective, operable, and interpretable.

[0064] 2. This invention can directly provide decision-makers with the optimal solution without requiring them to determine the optimal solution from non-dominated sets, thereby saving decision-making costs and effort.

[0065] 3. The integrated framework adopted in this invention is clear, concise, and reasonable. It is used to solve the problem of optimizing project portfolios that consider project interactions and can be directly applied to enterprise practice. Attached Figure Description

[0066] Figure 1 This is an overall flowchart of the Pareto optimization method for power grid digitalization projects based on project relationships in one embodiment of the present invention;

[0067] Figure 2 This is a detailed flowchart of a Pareto optimization method for power grid digitization projects based on project relationships, according to one embodiment of the present invention.

[0068] Figure 3 This is a diagram illustrating the conversion process between the project technology utilization network and the project mutual reference network in one embodiment of the present invention;

[0069] Figure 4 This is a flowchart of solving a multi-objective optimization model for a digital project group using the NSGA-II algorithm in one embodiment of the present invention;

[0070] Figure 5 This is a flowchart of the TOPSIS algorithm for ranking association rules in one embodiment of the present invention;

[0071] Figure 6 This is a diagram showing the usage relationships between items and technologies in one embodiment of the present invention;

[0072] Figure 7 This is a depth diagram of inter-item connections in one embodiment of the present invention;

[0073] Figure 8 This is a schematic diagram of the solution result of the target decision model in one embodiment of the present invention;

[0074] Figure 9 In one embodiment of the present invention, the Pareto optimal solution set is selected. Solution for the project portfolio at that time;

[0075] Figure 10 This refers to the relative importance of items in the selected Pareto set in one embodiment of the present invention;

[0076] Figure 11 This is a distribution diagram of the TOPSIS values ​​of all association rules in one embodiment of the present invention. Detailed Implementation

[0077] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present 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 the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention are within the scope of protection of the present invention.

[0078] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0079] The present invention provides a Pareto optimization method and system for power grid digitization projects based on project relationships, comprising: establishing an evaluation index system and a shared network among power grid digitization projects based on quantitative evaluation information, technical characteristics of power grid digitization projects, and development strategy information of power grid companies; establishing a multi-objective optimization model for a group of digitization projects based on the evaluation index system and the shared network among projects, and solving for a Pareto optimal solution set; mining association rules based on the Pareto optimal solution set, and ranking the association rules using support, confidence, and lift as indicators; filtering the pre-obtained non-dominated solution set according to the ranking of association rules, removing the worst solution in each filtering until only one optimal combination of digitization projects remains, thus completing the optimization of digitization projects.

[0080] Therefore, the objectives of this invention are: 1) to define a representative project interaction model and determine how project interactions affect the final value of the project portfolio; and 2) to construct an accurate selection method to obtain the optimal solution from the obtained non-dominated project portfolio. By applying the above methods, the final recommended project portfolio can convincingly provide benefits after implementation.

[0081] In one embodiment of the present invention, a Pareto optimization method for a power grid digitization project is provided. In this embodiment, as... Figure 1 , Figure 2 As shown, the method includes the following steps:

[0082] 1) Based on information related to the quantitative evaluation of power grid digitization projects, the technical characteristics of power grid digitization projects, and the development strategy information of power grid companies, establish an evaluation index system and a shared network among power grid digitization projects;

[0083] 2) Based on the evaluation index system and the shared network among projects, establish a multi-objective optimization model for digital project groups and solve for the Pareto optimal solution set;

[0084] 3) Based on the Pareto optimal solution set, association rules are mined, and the support, confidence, and lift are used as indicators to rank the association rules;

[0085] 4) Based on the sorting of association rules, the pre-obtained non-dominated solution set is filtered. Each filter removes the worst solution until the last optimal combination solution of digital projects remains, thus completing the selection of the optimal digital project.

[0086] In step 1) above, the basic indicators used to evaluate the power grid digitization project include two financial indicators: cost and benefits, as well as the improvement of other power grid enterprise-level indicators brought about by the implementation of the digitization project, such as the degree of compliance of the power grid company's strategic indicators, including project operation reliability, energy saving and emission reduction, project production rate, comprehensive voltage qualification rate, and power grid development resilience coefficient.

[0087] The selection and calculation of project indicators are as follows:

[0088] A1) Project Operation Reliability Rate (%): Reflects the power supply capacity of the power supply company to power users, as well as the degree to which the power industry meets the national economy's electricity demand. It is a comprehensive reflection of the quality and management level of the power supply company in terms of planning, design, infrastructure, construction, equipment manufacturing, production and operation.

[0089] (1)

[0090] In the formula, Indicates the project's operational reliability. This indicates the power outage time for the line to which the project belongs. Indicates the time period for the statistics.

[0091] A2) Energy saving and emission reduction (tons): This indicator is generally measured by the number of tons of standard coal saved to measure the effectiveness of energy saving and emission reduction.

[0092] (2)

[0093] In the formula, This indicates the saving of standard coal. This indicates that the electricity substitution project saves standard coal. This indicates that the clean energy transmission project saves standard coal equivalent.

[0094] in:

[0095] (3)

[0096] (4)

[0097] In the formula, This indicates the amount of electricity replaced by the electricity substitution project. This indicates the amount of new energy generated by the clean energy transmission project. This represents the standard coal equivalent coefficient.

[0098] A3) Project production rate (%): The percentage of projects whose actual production capacity reaches the designed production capacity, that is, the ratio of the number of projects that have reached production capacity to the total number of projects put into operation in the same year. It is used to evaluate the power grid's production completion status in digitalization projects and other operational projects.

[0099] (5)

[0100] In the formula, Indicates the project's production rate. Indicates the number of projects that have reached full production capacity. This indicates the total number of projects put into operation in the same year.

[0101] A4) Overall voltage qualification rate (%): This reflects the voltage quality of the power grid and the stability of its operation.

[0102] (6)

[0103] In the formula, Indicates the overall voltage qualification rate. , , , These represent the voltage qualification rates for voltage monitoring points of categories a, b, c, and d, respectively. Among them:

[0104] Category A refers to the 20kV and 10(6)kV bus voltages of substations and power plants supplying regional power loads.

[0105] Category B refers to dedicated power supply lines of 20kV, 35kV, and 66kV, and power supply voltages of 110kV and above.

[0106] Category C includes 20kV, 35kV, and 66kV non-dedicated power supply lines and 10(6)kV power supply voltage. At least one voltage monitoring point should be set up for every 10 MW load.

[0107] Class D is the 380 / 220 V low-voltage network supply voltage. At least two voltage monitoring points should be installed for every 100 distribution transformers. Monitoring points should be located at the beginning and end of representative low-voltage distribution networks and at some important user locations.

[0108] All monitoring points should be adjusted annually in accordance with changes in the power supply network.

[0109] A5) Power grid development elasticity coefficient: This parameter evaluates the overall relationship between the power grid and economic development from a macro perspective and can be used to measure whether power development is adapted to national economic development.

[0110] (7)

[0111] In the formula, For the power grid development flexibility coefficient, This represents the growth rate of the power grid substation capacity in the region. This represents the growth rate of the power grid line length in the region. This represents the region's GDP growth rate.

[0112] In step 1) above, the establishment of a shared network among projects is constructed using the co-reference method. Technology is interpreted as project supporter, while projects are seen as technology users, using a matrix... Indicates the technical project dependencies, where Indicates project Requires technology This dependency relationship is similar to a cross-citation network, and a shared network of projects can be constructed using co-citation methods. Specifically, it includes the following steps:

[0113] 1.1) By mapping two projects that share a network, determine whether the two projects have a shared use relationship. If they require the same technology, they have a shared use relationship.

[0114] 1.2) For two projects with a shared utilization relationship, the number of technologies used simultaneously is set as the weight of the synergistic utilization advantage between the projects, and a project sharing network is constructed. ,in Indicates project The technologies required simultaneously with project y This indicates the number of technologies used in the projects that participated in the selection process.

[0115] The following adjacency networks, if the technology Simultaneously pointing to the project and projects ,but ;otherwise, Therefore, the project can be calculated. and The quantity of shared use. It can be proven that: And the diagonal elements conform to equation (8). equals project The required number of technologies. The process of converting project technologies using networks and project cross-referencing networks, as follows: Figure 3 As shown.

[0116] (8)

[0117] As shown in the example below, a project utilizes technology... The network, the project obtained shares a network. process.

[0118]

[0119]

[0120] In step 2) above, based on the various indicators of each project and the determination of the shared network between projects, a multi-objective optimization model for digital project groups should be established, and the Pareto optimal solution set should be obtained by using the NSGA algorithm. This includes: determining the calculation of multiple objective functions, determining the priority judgment method for different project group schemes, and screening multi-objective schemes.

[0121] Typically, the value of a digitization project comprises two main aspects: (i) the project's intrinsic benefits and (ii) its alignment with the company's strategy. The project's intrinsic benefits can be measured specifically by its net present value or revenue, while the alignment with the company's strategy can be scored using expert experience. The qualitative and quantitative indicators of this invention are used for measurement. Therefore, the multi-objective optimization model for the digitization project group is as follows:

[0122] (i) Project The benefits can be a weighted sum of multiple benefits. Such as social benefits, economic benefits, environmental benefits, etc.; project and The degree of consistency of a company's strategy is used express, Project portfolio solution The overall benefits can be expressed as:

[0123]

[0124] in:

[0125]

[0126]

[0127]

[0128]

[0129]

[0130] here Indicates project group Overall benefits; It is a project group The overall score obtained by meeting the company's various strategies; These are the weighted values ​​of various company indicators, which are scored by the company's management based on the company's future development plan; It is a coefficient used to determine the importance of a digital project in terms of its effectiveness and alignment with the company's long-term development direction; matrix It is a shared network matrix for the project. A diagonal matrix; It is the objective function The regularization term reflects the degree to which the connections between projects affect the overall value of the project. Reflects the project portfolio scheme In this context, the total number of connections between different projects; This indicates the degree of influence of the regularization term on the objective function.

[0131] (ii) Determining the risk measurement model involves two aspects: the maturity of the technologies required for the project ( This includes the degree of interconnectivity between projects. Technology maturity is widely used and is typically defined on a scale of 1 to 9. One example of how the degree of interconnectivity between projects affects portfolio risk is that if multiple projects in a portfolio require the use of an immature technology, then the risk of that project will be relatively high.

[0132] In order to establish a portfolio solution The risk measurement model needs to meet two conditions: (i) the range of risk values ​​should be within the interval. (ii) When the project portfolio scheme More low When using advanced technologies, the risks would likely be higher. Therefore, considering the connections between projects, a portfolio approach is preferable. The risk is defined as follows:

[0133]

[0134] in:

[0135]

[0136] Defined by calculating the distance between the actual technology maturity level and the ideal value of a portfolio of projects, only portfolio projects... The projects included and the technologies required for these projects are necessary to calculate the risk value of the project portfolio. It is technology Maturity; weight The weight of a technology depends on how frequently it is used; the more often a technology is needed in a project, the greater its weight. It is a vector of Norm. It is easy to prove that the above formula satisfies the two conditions of the risk measurement model.

[0137] Project portfolio solution It is by One element of all feasible portfolio options consisting of [number] projects, where the portfolio options are composed of [number] projects. The combination of values ​​determines, . Time indicates project Unselected Project Cluster Solution middle; Indicates project portfolio scheme Includes projects Only when the project portfolio solution The total cost is lower than the maximum limit At that time, the project group plan This is the feasible solution. Therefore, the set of feasible total project set solutions. It can be defined as: Considering that under multiple evaluation objectives, there is a possibility that a project group solution that satisfies all constraints and performs best may not exist, it is necessary to define the optimal solution for solving multi-objective decision problems.

[0138] The dominance relationship between project groups can be defined by the following formula. Solutions in a subordinate relationship do not need further consideration and can be directly removed from the decision space.

[0139] Definition 1: Programming Scheme quilt When exercising control, the following conditions must be met:

[0140]

[0141] Definition 2: According to Definition 1, a valid set of projects It can be defined by the following formula:

[0142]

[0143] Definition 3: Pareto Rank: In a set of solutions, the Pareto rank of a non-dominated solution is defined as 1. If the non-dominated solution is removed from the set of solutions, the Pareto rank of the remaining solution is defined as 2, and so on, so that the Pareto rank of all solutions in the set of solutions can be obtained.

[0144] The non-dominated solution space of a project portfolio solution increases significantly with the number of projects. Assuming there are m projects, the solution space will generate... There are several decision-making options. An efficient and fast method is needed to solve this multi-objective optimization problem.

[0145] NSGA is a widely used multi-objective intelligent optimization algorithm that uses operations such as crossover and mutation of parents to produce offspring and eliminate individuals that do not meet the criteria, thereby selecting a non-dominated solution. The NSGA-II algorithm, namely the fast non-dominated multi-objective optimization algorithm with an elite preservation strategy, is a multi-objective optimization algorithm based on Pareto optimality.

[0146] In this embodiment, a multi-objective optimization model for a digital project group is established, and the Pareto optimal solution set is obtained. The NSGA-II algorithm is used for solving the problem. Figure 4 As shown, it includes the following steps:

[0147] 2.1) Using a portfolio approach The initial population is generated by the chromosomes of each individual in the population.

[0148] in, , .

[0149] 2.2) Establish and calculate the fitness model and penalty function to determine the Pareto rank in the solution set;

[0150] In this embodiment, the fitness function is established based on two objective functions: maximizing the benefits of the project group and minimizing the risk. The penalty function is set by adding a penalty term to the fitness function to prevent project investment from exceeding the budget.

[0151]

[0152]

[0153]

[0154] in, and These are project portfolio solutions The algorithm aims to minimize the benefit and risk values. and The value, and Let m represent the objective functions for benefits and risks, respectively, and m represent the number of projects. Represents the investment cost vector of the project. This represents the investment budget. Furthermore, if the costs of the project portfolio exceed the budget, the excess amount will be added to both fitness functions. and It determines the extent to which the portion exceeding the budget affects the fitness function value.

[0155] 2.3) Perform non-dominated sorting on the initial population to divide the population into all levels;

[0156] 2.4) After the initial population is sorted non-dominated, crossover, mutation, and selection operations of the genetic algorithm are performed to generate new individuals and add them to the population;

[0157] 2.5) From the entire population Select from individuals Each individual is grouped into a set (each individual has an equal probability of being selected). Based on the fitness of each individual, the individual with the best fitness value is selected to enter the offspring population. This process is repeated until the number of individuals in the offspring population reaches a certain threshold. , Population size;

[0158] 2.6) Calculate the crowding degree of all individuals to make the obtained solution more uniform in the target space;

[0159] In this embodiment, congestion The introduction of is to make the obtained solutions more uniform in the objective space. First, the objective function value for each individual in the population is calculated. For each objective function, all individuals at each Pareto level are sorted according to the objective function, and denoted as . For individual objective function The maximum value, For individual objective function The minimum value. The crowding degree of the two boundaries after sorting. and Set as Calculate the congestion level using the following formula. :

[0160]

[0161] in, It is the objective function value of the next element after the individual is sorted. This indicates the updated congestion level.

[0162] 2.7) Based on the Pareto level and crowding, the parent and offspring populations are combined into a new population. A new parent population is generated from the new population according to the pre-set rules. This process is repeated until the set termination conditions are met.

[0163] In step 2.3) above, the initial population is subjected to non-dominated sorting, and all population ranks are divided, including:

[0164] Calculate each individual Number of dominated and the set of solutions dominated by that individual. Traverse the entire population; the population size is The computational complexity of this parameter is . ;

[0165] Parameters in the population Individuals are placed into the first group set. In the middle, delete set and After identifying the individuals in the list, recalculate the number of dominated individuals for all remaining individuals. Then in the new round Individuals are placed into the second type of cluster and In this process, the process is repeated until all population ranks are defined.

[0166] In step 2.4) above, the crossover and mutation operations are as follows:

[0167] In this embodiment, a simulated binary real number encoding is used for the crossover operation. The specific implementation formula is as follows:

[0168]

[0169]

[0170] in:

[0171]

[0172] In the formula, and For the randomly selected parent individual's first One genotype, and The genotypes of the offspring individuals generated from these two parent individuals; For the range Random numbers between; The crossover algorithm distribution index; The propagation factor is defined as the ratio of the distance between children and parents. The larger the value, the closer the offspring is to the parent.

[0173] Mutation operation: The polynomial mutation operation of the NSGA-II algorithm used in this embodiment is as follows:

[0174]

[0175] in:

[0176]

[0177] In the formula, For the first generation of the parent individual Variation operators for each genotype; For the mutated individual genotype; for A random number between [a certain number of points].

[0178] In step 2.7 above, the pre-defined rules include:

[0179] (1) According to the Pareto level from low to high, the entire population layer is placed into the parent population. Until a certain layer of individuals can no longer be fully incorporated into the parent population. ;

[0180] (2) Arrange the individuals in this layer from largest to smallest according to their crowding density, and place them into the parent population in sequence. In the middle, until the parent population Fill it up.

[0181] In step 3) above, the discovery of association rules is as follows: finding patterns in association rules can help decision-makers make decisions on numerous feasible, non-dominated options. The association rules discovered from all non-dominated project group options represent the frequency with which one or more projects appear simultaneously in the Pareto project group solution set, that is, these projects would perform better if they collaborated together.

[0182] To effectively discover association rules and build a set of association rules, it is necessary to define criteria for identifying association rules. Support, confidence, and lift are three commonly used evaluation criteria for discovering association rules. To describe these rules in detail, variables... and It was used in the calculation formula as a project example.

[0183] (1) Support

[0184] Support is the proportion of several items that appear together in a dataset out of the total dataset. This can also be interpreted as the probability of association between certain items. Support is calculated as follows:

[0185]

[0186] in, This indicates that the solutions in each project group appear simultaneously. and Project frequency, This represents the total number of items in the dataset. Generally, items with high support may not form association rules, but items with low support will definitely not form association rules. For example, if 1000 customers go to a shopping mall to buy items, and 150 of them buy both ballpoint pens and laptops, then the support of the association rule (ballpoint pen, laptop) is 150 / 1000 × 100% = 15%.

[0187] (2) Confidence level

[0188] Confidence level refers to the probability that an item will occur when another item occurs, or the conditional probability of that item. The confidence level can be expressed as:

[0189]

[0190] For example, in the example above, if 65% of customers who buy ballpoint pens also buy notebooks, then the confidence level of the association rule (ballpoint pen → notebook computer) is 65%.

[0191] (3) Lift

[0192] Lifting degree representation Probability and The ratio of probabilities. It reflects... and The degree of correlation between them. A lift greater than 1 indicates... It is a valid strong association rule. However, when When the lift is less than or equal to 1, This rule is deemed invalid due to its strong association. As a special case, when… and When independent, the lift will be equal to 1.

[0193]

[0194] For example, in the example above, the lift of the association rule (ballpoint pen → notebook) is 65% / 25%=2.6, indicating that it is a valid strong association rule.

[0195] The refinement selection strategy traverses and sorts the association rules, refining non-dominated solutions until only one solution remains. There should be a sufficient number of association rules; otherwise, these rules may have been completely traversed before the optimal solution is obtained. Therefore, when implementing the association rule mining algorithm, users should set the minimum support and confidence limits to relatively low values ​​to ensure a sufficient number of association rules are obtained.

[0196] The most frequent association rules are obtained using the ranking technique based on similarity to the ideal solution (TOPSIS). TOPSIS comprehensively considers the support, confidence, and lift indices to rank the association rules of each item. The basic idea of ​​TOPSIS is to compare the distances of all candidates to the positive and negative ideal points. It is assumed that... There are 1 frequent association rules, where the three standard vectors are represented as follows: , and ,in .

[0197] In this embodiment, the association rules are sorted using the TOPSIS method, such as... Figure 5 As shown, it includes the following steps:

[0198] 3.1) Normalize the standard matrix consisting of support, confidence, and lift to obtain the decision matrix;

[0199] Assumption Standard Decision matrix ,in It is calculated by the following formula:

[0200]

[0201] 3.2) Calculate the positive ideal solution based on the decision matrix. and negative ideal solution ;

[0202]

[0203]

[0204] 3.3) According to the positive ideal solution and negative ideal solution Calculate frequent sets The distance between the standard point and the ideal point;

[0205]

[0206]

[0207] in, This represents the distance between the standard point and the ideal point. This represents the distance between the standard point and the negative ideal point;

[0208] 3.4) The relative distance from each standard point to the negative ideal solution is obtained from the distance between the standard point and the ideal point. , This describes the importance of relatively frequent items, based on relative distance. Obtain the most frequent association rules.

[0209] in, Also called TOPSIS value.

[0210]

[0211] Based on relative distance The formula for obtaining the most frequent association rule is:

[0212]

[0213] In step 4) above, the pre-obtained non-dominated solution set is filtered according to the order of association rules. Each filter removes the worst solution until only one optimal digital item combination solution remains. , empty set Enter all non-dominated project group schemes. This includes the following steps:

[0214] 4.1) Analysis of the first Each association rule, in the association rule related set The system stores the project group scheme that contains this association rule;

[0215] 4.2) Screening the set of non-dominated project group schemes ,from Remove items not included in the set and Elements in;

[0216] 4.3) Check from Filter the remaining project portfolio solutions ,if If only one project portfolio solution remains, stop and output. Otherwise, let Repeated filtering.

[0217] Based on the principle mentioned above, the more frequently items are selected in the non-dominant set, the more items should be retained. Therefore, this strategy follows the opposite idea: combinations of items that contain no items should be removed. Consider the following example to illustrate the process of refining the selection strategy. First, 12 association rules are abstracted from the set of case studies, as shown in the table below. Items are ranked according to the order in which they first appear in the 12 rules. That is, the earlier an item appears in the association rule, the higher its ranking. Thus, as shown in Table 1, the item rankings are [3,4], 6, 13, and 16, indicating that items 3 and 4 are the most important. Next, the non-dominant item combinations are refined. First, combinations that do not contain items 3 and 4 are removed, then combinations that do not contain item 6 are removed, and finally combinations that do not contain item 13 are removed. The removal process continues until only one combination remains.

[0218] Table 1 Association Rules and Ranking

[0219]

[0220] Implementation Example: This embodiment uses 16 specific digitalization projects during the digital transformation of a power grid company as examples. Based on the evaluation index calculation method established in step 1) and the collected information on specific digitalization projects of the power grid company, the values ​​of various indicators for each digitalization project are calculated as shown in Table 2, including project operational reliability, energy saving and emission reduction, and project production capacity. Indicators such as rate, comprehensive voltage qualification rate, and power grid development elasticity coefficient are incremental values, representing the improvement and enhancement of the company's indicators brought about by the implementation of a certain digitalization project. The unit is uniformly set as %, and the unit of cost and benefit is yuan.

[0221] Table 2. Values ​​of various indicators for each digitalization project

[0222]

[0223] Next, a technology usage table for the project was established by analyzing the specific technical information of the digital project. In this embodiment, a total of 9 key technologies were used in the 15 projects. The key technologies required by each project are different. The specific technology usage is shown in Table 3.

[0224] Table 3: Utilization of Key Technologies Required for Each Project

[0225]

[0226] In the table, 0 indicates that the digitization project did not use this technology, and 1 indicates that the project used a certain technology. If multiple projects use the same technology, it means that the technology is relatively important and the technological connections between these digitization projects are relatively close. If a project needs to use multiple technologies, it means that the project is highly dependent on technology, and the risk of the project will also increase significantly if the technology is prone to failure.

[0227] From the formula Calculate the cross-reference network matrix among 15 digital projects. The matrix is ​​shown in Table 3. The calculation results are shown in Table 4.

[0228] Table 4. Calculation results of the mutual referral network matrix among digital projects.

[0229]

[0230] like Figure 6 As shown, this illustrates the relationship between projects and technologies. The lines in the diagram represent the technologies used in a project, such as... Figure 7 As shown, this graph displays the depth of connections between projects. The numbers on the horizontal and vertical axes represent the project numbers, and the darker the color, the closer the technical connection between the projects.

[0231] Following step 2), the multi-objective decision-making model is used to solve for the Pareto optimal solution. In the NSGA-II algorithm, the population size is set to 300, the maximum number of iterations to 500, the crossover probability to 0.5, and the mutation probability to 0.1. In calculating the objective function, the weights for economic benefits and the company's strategic alignment are calculated using values ​​of 0.5 and 0.5, respectively. However, the factors that significantly influence the objective function value are... Then, different values ​​are used for calculation, and the non-dominated solution of the item group with the better result is selected. In this example... We take values ​​of 1 / 5, 1 / 4, 1 / 3, 1 / 2, 1, 2, 3, 4, and 5 respectively. The maximum investment is set to... The solution results of the multi-objective decision model are as follows: Figure 8 As shown.

[0232] Based on the obtained Pareto optimal solution set, select Taking half of the project portfolio solutions, after removing duplicate solutions, 21 digitization project portfolio solutions remain. For example... Figure 9 As shown, the vertical axis represents the project number, and the horizontal axis represents the project group scheme. A cell marked 1 indicates that the project group scheme contains the corresponding project on the horizontal axis, while a blacked-out cell indicates that the project group scheme does not contain the corresponding project. Analysis of Table 4 shows that projects 5, 14, and 15 are included in every project group scheme, while projects 6, 7, and 8 are not included in any project group scheme. Then, the projects in the selected Pareto set are calculated, and their relative importance is determined, as follows: Figure 10 As shown, items 5, 14, and 15 are the most important when viewed as individual items. However, this ranking does not consider the combination of items. Therefore, to determine the optimal combination of items, association rules should be mined from the Pareto set.

[0233] Based on the 21 item clusters remaining after the first step of screening, the Apropri algorithm was used to mine association rules. Next, using calculated support, confidence, and lift information, 1932 association rules were mined from the 21 item clusters. Then, the TOPSIS algorithm was applied to determine the rule ranking, and the ranking results are shown in Table 5. Figure 11 In the diagram, the area plot represents the distribution of TOPSIS values ​​for ranking all association rules. Most TOPSIS values ​​fall within the range of [0.3, 0.5], and the higher the TOPSIS value, the higher the ranking of the association rule.

[0234] Table 5 Ranking Results

[0235]

[0236] Finally, following the project cluster selection step in step 4), Pareto solutions that do not meet the requirements are gradually eliminated according to the ranking of association rules until only one optimal solution remains. First, project cluster solutions containing one of [14,5], 16, 12, 15, 3, 13, 2, 1, 10, 9, 11, 4 are retained, and the others are eliminated. Next, the solution cluster containing [14,5] is selected, and the rest are eliminated. This process continues until the final optimal solution is: 1, 2, 3, 5, 9, 10, 11, 12, 13, 14, 15, 16, with a benefit value of 534.4040 and a risk value of 0.2987.

[0237] In summary, the two innovative methods for project interaction and fine selection proposed in this invention are feasible and reasonable for addressing the two major challenges of project portfolio selection: (i) modeling the impact of project interaction on the final value of the project portfolio; and (ii) selecting the best solution from non-dominant project portfolios.

[0238] For the value and risk model, a project collaborative utilization network representing the interactions between projects and technologies is constructed based on the dependencies between them. Then, the project interaction degree is set as a regularization term and added to the original value function to indicate the impact of interactions on the portfolio value. Secondly, the risk model is built based on TRL (Tracking, Transactional, and Liberal Arts), where the less mature the technology upon which the portfolio depends, the higher the risk.

[0239] For optimal item portfolio selection, association rules are mined and ranked based on the obtained non-dominated item portfolios using the TOPSIS method. An improved selection strategy is then adopted to progressively remove the worst item portfolios without high-ranking items from the non-dominated solutions until only one item portfolio remains, which will be recommended as the optimal solution.

[0240] In one embodiment of the present invention, a Pareto optimization system for a power grid digitization project is provided, comprising:

[0241] The first processing module establishes an evaluation index system and a shared network among power grid digitalization projects based on information related to the quantitative evaluation of power grid digitalization projects, the technical characteristics of power grid digitalization projects, and the development strategy information of power grid companies.

[0242] The second processing module establishes a multi-objective optimization model for digital project groups based on the evaluation index system and the shared network between projects, and solves for the Pareto optimal solution set.

[0243] The third processing module mines association rules based on the obtained Pareto optimal solution set and ranks the association rules using support, confidence, and lift as indicators.

[0244] The filtering module filters the pre-obtained non-dominated solution set according to the ranking of association rules. Each filtering removes the worst solution until the best combination of digital projects remains, thus completing the optimization of digital projects.

[0245] In the first processing module mentioned above, a shared network between projects is established, including:

[0246] By mapping two projects that share a network, it can be determined whether the two projects have a shared use relationship.

[0247] For two projects with a shared utilization relationship, the number of technologies used simultaneously is set as the weight of the synergistic utilization advantage between the projects, thus constructing a project sharing network. ,in Indicates project The technologies required simultaneously with project y This indicates the number of technologies used in all projects.

[0248] In the second processing module mentioned above, a multi-objective optimization model for the digital project group is established, and the Pareto optimal solution set is obtained, including:

[0249] Project portfolio solution The initial population is generated by the chromosomes of each individual in the population.

[0250] Establish and compute the fitness model and penalty function, and determine the Pareto rank in the solution set;

[0251] The initial population is sorted using a non-dominated ordering method, and all Pareto ranks are assigned.

[0252] After the initial population is sorted non-dominated, crossover, mutation, and selection operations of the genetic algorithm are performed to generate new individuals and add them to the population.

[0253] From all populations Select from individuals Individuals are grouped into groups. Based on the fitness of each individual, the individual with the best fitness value is selected to enter the offspring population. This process is repeated until the number of individuals in the offspring population reaches a certain threshold. , Population size;

[0254] Calculate the crowding degree of all individuals to make the obtained solution more uniform in the target space;

[0255] Based on the Pareto level and crowding, the parent and offspring populations are combined into a new population. A new parent population is generated from the new population according to pre-set rules, and this process is repeated until the set termination conditions are met.

[0256] This involves performing a non-dominated sort on the initial population, dividing the population into all levels, including:

[0257] Calculate each individual Number of dominated and the set of solutions dominated by that individual. Traverse the entire population; the population size is The computational complexity of this parameter is . ;

[0258] Parameters in the population Individuals are placed into the first group set. In the middle, delete set and After identifying the individuals in the list, recalculate the number of dominated individuals for all remaining individuals. Then Individuals are placed into the second type of cluster and In this process, the process is repeated until all population ranks are defined.

[0259] In this embodiment, the pre-defined rules include:

[0260] Based on the Pareto hierarchy from low to high, the entire population layer is placed into the parent population. Until a certain layer of individuals can no longer be fully incorporated into the parent population. ;

[0261] The individuals in this layer are arranged from most crowded to least crowded and then placed into the parent population in sequence. In the middle, until the parent population Fill it up.

[0262] In the third processing module mentioned above, the ranking of association rules adopts the TOPSIS method, including:

[0263] The decision matrix is ​​obtained by normalizing the standard matrix composed of support, confidence, and lift.

[0264] Calculate the ideal solution based on the decision matrix. and negative ideal solution ;

[0265] According to the ideal solution and negative ideal solution Calculate frequent sets The distance between the standard point and the ideal point;

[0266] The relative distance from each standard point to the negative ideal solution is obtained from the distance between the standard point and the ideal point. , This describes the importance of relatively frequent items, based on relative distance. Obtain the most frequent association rules.

[0267] In the above filtering module, the pre-obtained non-dominated solution set is filtered according to the order of association rules. Each filtering removes the worst solution until only one optimal digitized item combination solution remains, including:

[0268] Analysis of the first Each association rule, in the association rule related set The system stores the project group scheme that contains this association rule;

[0269] Filtering the set of non-dominated project groups ,from Remove items not included in the set and The elements in the set are updated, and the empty set is updated. ;

[0270] Check from Filter the remaining project portfolio solutions ,if If only one project portfolio solution remains, stop and output. Otherwise, let Repeated filtering.

[0271] The system provided in this embodiment is used to execute the above-described method embodiments. For specific processes and details, please refer to the above embodiments, which will not be repeated here.

[0272] A schematic diagram of a computing device structure is provided in one embodiment of the present invention. This computing device can be a terminal, which may include: a processor, a communication interface, memory, a display screen, and an input device. The processor, communication interface, and memory communicate with each other via a communication bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and a computer program. When executed by the processor, the computer program implements a Pareto optimization method for a power grid digitization project that considers project relationships. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The communication interface is used for wired or wireless communication with external terminals. Wireless communication can be achieved through Wi-Fi, a management network, NFC (Near Field Communication), or other technologies. The display screen can be a liquid crystal display or an e-ink display. The input device can be a touch layer covering the display screen, or buttons, a trackball, or a touchpad mounted on the casing of the computing device, or an external keyboard, touchpad, or mouse, etc. The processor can call logical instructions in memory to execute the following methods: Based on the comprehensive collection of quantitative evaluation information related to power grid digitization projects, the technical characteristics of power grid digitization projects, and the development strategy information of power grid companies, establish an evaluation index system and a shared network among projects for power grid digitization projects; based on the various indicators of each project and the determination of the shared network among projects, establish a multi-objective optimization model for the digitization project group and solve for the Pareto optimal solution set; mine association rules from the obtained Pareto optimal solution set, and rank the association rules using support, confidence, and lift as indicators; refine the pre-obtained non-dominated solution set according to the ranking of association rules, removing the worst solution each time until only one optimal combination solution of digitization projects remains, thus completing the optimization of digitization projects.

[0273] Furthermore, the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, and 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 the prior art, or a part 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.

[0274] Those skilled in the art will understand that the structure of the above-described computing device is only a part of the structure related to the solution of this application, and does not constitute a limitation on the computing device on which the solution of this application is applied. The specific computing device may include more or fewer components, or combine certain components, or have different component arrangements.

[0275] In one embodiment of the present invention, a computer program product is provided, the computer program product including a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, when the program instructions are executed by a computer, the computer is able to execute the methods provided in the above-described method embodiments, such as including: establishing an evaluation index system and a shared network among projects for power grid digitization projects based on comprehensively collected quantitative evaluation information related to power grid digitization projects, technical characteristics of power grid digitization projects, and development strategy information of power grid companies; establishing a multi-objective optimization model for digitization project groups based on various indicators of each project and on the basis of determining the shared network among projects, and solving for Pareto optimal solution sets; mining association rules based on the obtained Pareto optimal solution sets, and ranking the association rules using support, confidence, and lift as indicators; selecting the pre-obtained non-dominated solution sets according to the order of the association rules, removing the worst solution each time until only one optimal combination solution of digitization projects remains, thus completing the selection of digitization projects.

[0276] In one embodiment of the present invention, a non-transitory computer-readable storage medium is provided, which stores server instructions that cause a computer to execute the methods provided in the above embodiments. These instructions include, for example,: establishing an evaluation index system and a shared network among projects for the power grid digitization project based on comprehensively collected quantitative evaluation information related to the power grid digitization project, the technical characteristics of the project, and the development strategy information of the power grid company; establishing a multi-objective optimization model for the digitization project group based on the indicators of each project and the determined shared network among projects, and solving for a Pareto optimal solution set; mining association rules from the obtained Pareto optimal solution set, and ranking the association rules using support, confidence, and lift as indicators; and selecting the pre-obtained non-dominated solution set according to the order of the association rules, removing the worst solution each time until only one optimal combination of digitization projects remains, thus completing the optimal selection of digitization projects.

[0277] The computer-readable storage medium provided in the above embodiments has a similar implementation principle and technical effect to the above method embodiments, and will not be described again here.

[0278] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0279] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0280] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0281] 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 Pareto optimization method for a power grid digitization project, characterized in that, include: Based on information related to the quantitative evaluation of power grid digitization projects, the technical characteristics of power grid digitization projects, and the development strategy information of power grid companies, establish an evaluation index system and a shared network among power grid digitization projects. Establishing a shared network among projects includes: processing two projects sharing a network through mapping to determine if they have a shared utilization relationship; for two projects with a shared utilization relationship, the number of technologies used simultaneously is set as a weight for the synergistic utilization advantage between the projects, thus constructing the shared network. ,in Indicates project The technologies required simultaneously with project y This indicates the number of technologies used in all projects; Based on the evaluation index system and the shared network among projects, a multi-objective optimization model for digital project groups is established, and the Pareto optimal solution set is obtained. Association rules are mined based on the Pareto optimal solution set, and ranked using support, confidence, and lift as indicators. The ranking of association rules employs the TOPSIS method, which includes: normalizing the standard matrix composed of support, confidence, and lift to obtain a decision matrix; and calculating the positive ideal solution based on the decision matrix. and negative ideal solution According to the ideal solution and negative ideal solution Calculate frequent sets The distance between the standard point and the ideal point; the relative distance from each standard point to the negative ideal solution is obtained from the distance between the standard point and the ideal point. , This describes the importance of relatively frequent items, based on relative distance. Obtain the most frequent association rules; The pre-obtained non-dominated solution set is filtered according to the association rule ranking. Each filter removes the worst solution until only one optimal combination of digitized items remains, thus completing the optimal selection of digitized items, including: Analysis of the first Each association rule, in the association rule related set The system stores the project group scheme that contains this association rule; Filtering the set of non-dominated project groups ,from Remove items not included in the set and The elements in the set are updated, and the empty set is updated. ; Check from Filter the remaining project portfolio solutions ,if If only one project portfolio solution remains, stop and output. Otherwise, let Repeated filtering.

2. The Pareto optimization method for power grid digitization projects as described in claim 1, characterized in that, The establishment of a multi-objective optimization model for a group of digital projects, and the solution to obtain the Pareto optimal solution set, includes: Project portfolio solution The initial population is generated by the chromosomes of each individual in the population. Establish and compute the fitness model and penalty function, and determine the Pareto rank in the solution set; The initial population is sorted using a non-dominated ordering method, and all Pareto ranks are assigned. After the initial population is sorted non-dominated, crossover, mutation, and selection operations of the genetic algorithm are performed to generate new individuals and add them to the population. From all populations Select from individuals Individuals are grouped into groups. Based on the fitness of each individual, the individual with the best fitness value is selected to enter the offspring population. This process is repeated until the number of individuals in the offspring population reaches a certain threshold. , Population size; Calculate the crowding degree of all individuals to make the obtained solution more uniform in the target space; Based on the Pareto level and crowding, the parent and offspring populations are combined into a new population. A new parent population is generated from the new population according to pre-set rules, and this process is repeated until the set termination conditions are met.

3. The Pareto optimization method for power grid digitization projects as described in claim 2, characterized in that, The process of performing non-dominated sorting on the initial population, which divides the population into all levels, includes: Calculate each individual Number of dominated and the set of solutions dominated by that individual. Traverse the entire population; the population size is The computational complexity of this parameter is . ; Parameters in the population Individuals are placed into the first group set. In the middle, delete set and After identifying the individuals in the list, recalculate the number of dominated individuals for all remaining individuals. Then Individuals are placed into the second type of cluster and This process is repeated until all population ranks are defined.

4. The Pareto optimization method for power grid digitization projects as described in claim 2, characterized in that, The pre-defined rules include: Based on the Pareto hierarchy from low to high, the entire population layer is placed into the parent population. Until a certain layer of individuals can no longer be fully incorporated into the parent population. ; The individuals in this layer are arranged from most crowded to least crowded and then placed into the parent population in sequence. In the middle, until the parent population Fill it up.

5. A Pareto optimization system for a power grid digitization project, used to implement the Pareto optimization method for a power grid digitization project as described in any one of claims 1 to 4, characterized in that, include: The first processing module establishes an evaluation index system and a shared network among power grid digitalization projects based on information related to the quantitative evaluation of power grid digitalization projects, the technical characteristics of power grid digitalization projects, and the development strategy information of power grid companies. The second processing module establishes a multi-objective optimization model for digital project groups based on the evaluation index system and the shared network between projects, and solves for the Pareto optimal solution set. The third processing module mines association rules based on the Pareto optimal solution set and ranks the association rules using support, confidence, and lift as indicators. The filtering module filters the pre-obtained non-dominated solution set according to the ranking of association rules. Each filtering removes the worst solution until the best combination of digital projects remains, thus completing the optimization of digital projects.

6. A computer-readable storage medium for storing one or more programs, characterized in that, The one or more programs include instructions that, when executed by a computing device, cause the computing device to perform any of the methods described in claims 1 to 4.

7. A computing device, characterized in that, include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for performing any of the methods described in claims 1 to 4.

Images