CN-122022310-A - Power operation and maintenance platform resource adaptation method based on fairness constraint game

CN122022310ACN 122022310 ACN122022310 ACN 122022310ACN-122022310-A

Abstract

The invention relates to a power operation and maintenance platform resource adaptation method based on fairness constraint game, belonging to the field of power system optimization scheduling; the method comprises the steps of firstly generating opportunity fairness rigid constraint based on historical data, limiting probability gap of a comparable group acquisition resource, establishing a fairness base line, constructing a parameterized comprehensive utility function integrating an individual cost function and a coefficient penalty term, solving by adopting a two-stage optimization process of 'first fairness screening and later efficiency preferred', screening out all feasible schemes to form a fairness feasible solution set by taking the rigidity fairness constraint as a target, and maximizing the comprehensive utility function in the solution set to obtain an optimal allocation scheme.

Inventors

WANG PING
ZHENG ZHEN
ZHANG LIN
LI JIAN
MAO YUXING
PAN JIANYU

Assignees

重庆大学
国网重庆市电力公司电力科学研究院

Dates

Publication Date: 20260512
Application Date: 20260127

Claims (8)

1. A power operation and maintenance platform resource adaptation method based on fairness constraint game is characterized by comprising the following steps: S1, acquiring available resource information of an electric power operation and maintenance platform and resource request information of multiple participants, and performing standardized processing; s2, calculating statistical probability of obtaining resources by different protected attribute subgroups in a comparable group based on historical data, and generating opportunity fairness rigid constraint; S3, constructing a logarithmic form cost function with marginal utility decreasing characteristics for each participant under the condition of guaranteeing basic requirements preferentially; s4, calculating a coefficient of a radix as an unfairness degree based on the cost function results of all the participants, and measuring fairness of the distribution result; S5, constructing a comprehensive optimization model, fusing the sum of individual values and a coefficient penalty term, building a parameterized comprehensive utility function, and adapting different operation directions through weights and sensitivity coefficients; S6, optimizing and solving the resource allocation vector which maximizes the comprehensive utility function through two stages, screening the fairness and feasibility, optimizing and selecting the efficiency, and taking the solving result as a final scheme.
2. The method for adapting resources of a power operation and maintenance platform based on fairness constraint gaming as set forth in claim 1, wherein in step S2, all participants are divided into at least one comparable group For each comparable population Determining at least one protected attribute Calculating statistical probabilities of obtaining resources for sub-populations having protected attribute values within a comparable population based on historical allocation data The absolute value of the difference between the statistical probabilities corresponding to any two protected attribute values in the internal memory is not more than a preset threshold value The method comprises the following steps: In the formula, Expressed in the group Any two participants with a protected attribute value z obtain a statistical probability of the resource, Representing any two of the protected attribute values, Representing the calculated maximum value.
3. The method for adapting resources of a power operation and maintenance platform based on fairness constraint gaming according to claim 2, wherein in step S3, each participant corresponds to a logarithmic form cost function with marginal utility decreasing characteristics Expressed as: Wherein, the Representing actual allocation to participants Is used for the amount of resources of (a), Representing participants The theoretical amount of resources for basic or normal operation is maintained, Representing priority weights, ln ()' is a logarithmic function with an exponent e; when dispensing an amount Far smaller than The function value increases faster to meet the basic requirement and generate significant value when the distribution amount is Near or exceed Then, the function value increases gradually, and the return rate of the excessive allocation of the resources is reflected to decrease.
4. The method for adapting resources of an electric power operation and maintenance platform based on fairness constraint gaming as set forth in claim 3, wherein in step S4, an unfairness is calculated based on individual basis value results of all participants using a coefficient of kunning The method comprises the following steps: In the formula, The vectors are allocated for the resources to be optimized, Representing the individual cost function, is the utility characteristic of the operation and maintenance resource to each participant, Representing actual allocation to participants Is used for the amount of resources of (a), Representing the total number of all participants, The method is characterized in that the method respectively represents different participants, the numerator part represents the absolute sum of value differences between every two participants, the discrete degree of value distribution is reflected, the denominator part is a standardized factor, the value range of G is between 0 and 1, and the smaller the G, the more fair the final result of the distribution scheme.
5. The method for adapting power operation and maintenance platform resources based on fairness constraint game according to claim 4, wherein in step S5, a comprehensive optimization model is constructed Meanwhile, the efficiency and the result fairness of resource allocation are evaluated, and the expression is as follows: In the formula, Representing an adjustable weight; for fairness sensitivity coefficients, a penalty is applied to the coefficient of kunning, Calculating an unfairness measure representing individual underlying value results of all participants by adjusting weights Coefficient of sensitivity And adapting to diversified operation scenes and policy guidance of the power operation and maintenance platform.
6. The method for adapting power operation and maintenance platform resources based on fairness constraint game according to claim 5, wherein in step S6, two-stage optimization solving processes for solving the comprehensive optimization model are respectively as follows: the first level of optimization, fair feasibility screening, searching all feasible allocation schemes under the physical limit of the total resource to form a fair feasible solution set by taking the generated rigid fair constraint as the only target ; Solving the comprehensive optimization model in a fair and feasible solution set to maximize the comprehensive utility function U and finally obtaining the optimal resource allocation vector by the output scheme ; The obtained optimal resource allocation vector Expressed as: Optimal resource allocation vector As a final solution.
7. The power operation and maintenance platform resource adaptation method based on the fairness constraint game according to claim 6, wherein the specific implementation process of the first-stage optimization is as follows: (1) Setting a fairness threshold according to policy guidance or operation requirements The allocation scheme is required to satisfy the coefficient of the foundation constraint: (2) All possible allocation schemes constitute a feasible domain under the constraint of the total amount of resources: (3) Large-scale random sampling in the feasible region A to generate Candidate allocation scheme For each candidate scheme Calculate its coefficient of background Screening out all the satisfaction Is a set of: And obtaining a fair and feasible solution set.
8. The power operation and maintenance platform resource adaptation method based on the fairness constraint game according to claim 6, wherein the specific implementation process of the second-level optimization is as follows: (1) Determining weights of areas Setting fairness sensitivity coefficient For all candidate schemes in the fair feasible solution set F, calculating the comprehensive utility value of the candidate schemes, and searching the solution with the maximum comprehensive utility in the solution set: (2) At the position of And Monte Carlo sampling is carried out within a reasonable variation range, the optimal solution distribution under different parameter combinations is calculated, and the optimal solution is evaluated Is described.

Description

Power operation and maintenance platform resource adaptation method based on fairness constraint game Technical Field The invention belongs to the field of power system optimization scheduling, and relates to a power operation and maintenance platform resource adaptation method based on fairness constraint gaming. Background With the development of the power system to the intelligent and decentralization directions, the power operation and maintenance platform faces the core challenge of the collaborative optimization of efficiency and fairness. The platform is used as a core hub, and various resources such as calculation, communication, control and the like need to be managed, and the requirements of application scenes (such as regional micro-grids, electric automobile clusters, intelligent buildings and the like) served by the platform are complex, changeable and dynamic heterogeneous. The traditional power resource allocation method mainly adopts a centralized optimization model, such as a linear programming and dynamic pricing mechanism, and the method can ensure the system efficiency, but is difficult to adapt to strategic interaction among multiple benefit subjects. In recent years, gambling theory has been introduced into power system research for modeling competing relationships between generators, users and operators. However, the existing game model often ignores fairness constraints, so that the resource allocation result may aggravate uneven areas or user groups. In the aspect of fairness guarantee, the existing research mainly adopts two types of methods, namely firstly, rule-based post-processing adjustment, such as balancing the power utilization priority of different user groups through weight coefficients, and secondly, taking fairness indexes as one of optimization targets, such as adding a coefficient of base or variance constraint into an objective function. However, these methods have significant limitations in that the rule-based adjustment lacks dynamic adaptability, and the multi-objective optimization often faces the problem that the pareto front is difficult to solve accurately. In view of the foregoing, there is a need for a dynamic allocation method of power operation and maintenance resources that can take the fairness principle as an insurmountable constraint condition and can be efficiently solved in a complex environment, so as to promote the fair, efficient and sustainable development of intelligent services of a power system. Disclosure of Invention Therefore, the invention aims to provide the electric power operation and maintenance platform resource adaptation method based on the fairness constraint game, which fundamentally solves the problems of efficiency and fairness synergy of resource allocation under multiple subjects and multiple scenes by converting fairness ethical criteria into executable mathematical constraints and adopting an innovative optimization decision framework. In order to achieve the above purpose, the present invention provides the following technical solutions: a power operation and maintenance platform resource adaptation method based on fairness constraint game comprises the following steps: S1, acquiring available resource information of an electric power operation and maintenance platform and resource request information of multiple participants, and performing standardized processing; s2, calculating statistical probability of obtaining resources by different protected attribute subgroups in a comparable group based on historical data, and generating opportunity fairness rigid constraint; S3, constructing a logarithmic form cost function with marginal utility decreasing characteristics for each participant under the condition of guaranteeing basic requirements preferentially; s4, calculating a coefficient of a radix as an unfairness degree based on the cost function results of all the participants, and measuring fairness of the distribution result; S5, constructing a comprehensive optimization model, fusing the sum of individual values and a coefficient penalty term, building a parameterized comprehensive utility function, and adapting different operation directions through weights and sensitivity coefficients; S6, optimizing and solving the resource allocation vector which maximizes the comprehensive utility function through two stages, screening the fairness and feasibility, optimizing and selecting the efficiency, and taking the solving result as a final scheme. Further, in step S2, all participants are divided into at least one comparable populationFor each comparable populationDetermining at least one protected attributeCalculating statistical probabilities of obtaining resources for sub-populations having protected attribute values within a comparable population based on historical allocation dataThe absolute value of the difference between the statistical probabilities corresponding to any two protected attribute values in the internal