Power system planning evaluation method based on random multi-criterion decision

Abstract

A power system planning evaluation method based on random multi-criterion decision comprises the following steps: 1) constructing a multi-dimensional index system, and obtaining a plurality of power system planning schemes; 2) performing initial weight assignment on three-level indexes in the multi-dimensional index system by adopting an AHP method; 3) determining positive and negative ideal solutions of each three-level index in the multi-dimensional index system based on the randomness of the power system planning scheme to the three-level index parameters; 4) constructing a nonlinear programming optimization problem; and 5) solving the nonlinear planning optimization problem to obtain the optimal weight of each three-level index, and selecting an optimal scheme from the plurality of power system planning schemes according to the optimal weight of each three-level index. According to the method, the volatility and randomness of new energy output can be fully considered, and the subjectivity of expert judgment and the objectivity of a mathematical algorithm can be combined, so that the applicable value of a planning evaluation result is improved, and more reliable decision support is provided for planning of a high-proportion renewable energy power system.

Description

Technical Field

[0001] This invention relates to the field of power systems and their automation technology, and specifically to a power system planning and evaluation method based on stochastic multi-criteria decision-making. Background Technology

[0002] With the global energy structure transformation, the integration of a high proportion of renewable energy sources, such as wind and solar power, into the power grid has become an inevitable trend. However, due to the volatility and uncertainty of renewable energy output, traditional power system planning methods face numerous challenges as the proportion of new energy gradually increases. Wind and solar power generation are significantly affected by climate, seasonality, and diurnal variations, making it difficult to accurately predict output and thus placing higher demands on the stability and security of the power grid. Existing power system operation modes are mainly designed for conventional energy generation, while the large-scale integration of new energy sources makes traditional planning and evaluation methods inadequate for adapting to this change. Therefore, in the context of high renewable energy penetration, how to scientifically and quantitatively evaluate different planning schemes and formulate reasonable power system planning schemes based on the evaluation results has become an important research direction.

[0003] Currently, evaluation methods for power system planning schemes mainly fall into two categories: traditional evaluation methods based on expert experience and quantitative evaluation methods based on mathematical modeling and optimization algorithms. However, traditional evaluation methods based on expert experience suffer from strong subjectivity and difficulty in quantitative evaluation. Furthermore, because the uncertainty of renewable energy output far exceeds that of traditional power generation methods, expert experience cannot fully consider the randomness of renewable energy, thus affecting the rationality of the planning scheme and consequently the stable operation of the power grid. Quantitative evaluation methods based on mathematical modeling and optimization algorithms significantly simplify indicators during calculation, lacking practicality and potentially leading to substantial deviations between evaluation results and actual conditions. In addition, the evaluation system is simplistic, focusing only on traditional economic and reliability indicators, failing to fully consider the randomness of renewable energy and related indicators such as output fluctuations, curtailment rates, and energy storage response capabilities. Simultaneously, the lack of assessment of system supply guarantee capabilities makes it difficult to measure the stability of the power grid under extreme weather or emergency conditions.

[0004] Existing assessment methods often fall short of quantitative support due to the lack of traditional expert evaluation, making it difficult to adapt to the trend of high renewable energy proportions. While mathematical optimization methods can provide quantitative analysis, their high degree of computational simplification makes them ill-suited to addressing multi-level decision-making needs when renewable energy output is highly uncertain. Furthermore, significant barriers exist between industry and academia in their assessment methods. Industry relies on empirical judgment and lacks quantitative evidence, while academia, though employing mathematical modeling, presents overly idealistic models that are difficult to apply directly to real-world systems. Therefore, integrating expert knowledge with data-driven analysis to construct a multi-criteria decision analysis method that is both applicable and scientific has become a crucial research direction for power system planning and assessment. Summary of the Invention

[0005] The purpose of this invention is to provide a power system planning and evaluation method based on stochastic multi-criteria decision-making, comprising the following steps:

[0006] 1) Construct a multi-dimensional indicator system and obtain multiple power system planning schemes.

[0007] The multi-dimensional indicator system includes primary indicators, secondary indicators, and tertiary indicators.

[0008] 2) Based on multiple power system planning schemes, the AHP method is used to assign initial weights to the third-level indicators in the multi-dimensional indicator system.

[0009] 3) Based on the randomness of the parameters of the third-level indicators in the power system planning scheme, determine the positive and negative ideal solutions of each third-level indicator in the multi-dimensional indicator system.

[0010] 4) Based on the initial weights and positive and negative ideal solutions of each tertiary index, construct a nonlinear programming optimization problem.

[0011] 5) Solve the nonlinear programming optimization problem to obtain the optimal weight of each tertiary index, and select the optimal scheme from multiple power system planning schemes based on the optimal weight of each tertiary index.

[0012] Furthermore, each primary indicator is further subdivided into secondary indicators, and each secondary indicator is further subdivided into tertiary indicators.

[0013] The primary indicators include economic efficiency, the penetration rate of new energy sources, the capacity for new energy absorption, the reliability of the power system, and the capacity for ensuring power supply.

[0014] The secondary indicators include, but are not limited to, electricity prices, new energy installations, system regulation capacity, load shedding probability, and demand response capacity.

[0015] The three-level indicators include, but are not limited to, electricity price, electric vehicle capacity, average regulation duration of traditional units, probability of load failure, and load shedding in extreme scenarios.

[0016] Furthermore, the power system planning scheme is a planning scheme that optimizes the power supply structure, power grid layout, and resource allocation, including a scheme that adjusts the power supply structure based on the installed capacity of new energy sources.

[0017] Furthermore, in step 2), the initial weight assignment steps for the three-level indicators are as follows:

[0018] 2.1) Divide the indicators in the multi-dimensional indicator system to obtain the evaluation indicators of power system planning schemes under different indicator levels.

[0019] The indicator hierarchy includes a primary indicator layer, a secondary indicator layer under the primary indicators, and a tertiary indicator layer under the secondary indicators.

[0020] 2.2) Compare the power system planning scheme evaluation indicators at the same indicator level pairwise, and construct judgment matrices for different indicator levels according to the importance of different indicators.

[0021] 2.3) Normalize the judgment matrix for each indicator level to obtain the approximate weight vector for each indicator level, as shown below:

[0022]

[0023] In the formula, i and j both represent indicator indices, n represents the total number of indicators at the indicator level, and a ij This indicates the degree of importance of the i-th indicator to the j-th indicator in the indicator hierarchy. This indicates the importance of the column after normalization. ω i This represents the approximate weight of the i-th indicator in the indicator hierarchy.

[0024] 2.4) Based on the judgment matrix and approximate weight vector of each indicator level, calculate the maximum eigenvalue of each indicator level, as shown below:

[0025]

[0026] In the formula, λ represents the largest eigenvalue of the indicator level. A represents the judgment matrix of the indicator level. ω' represents the approximate weight vector of the indicator level.

[0027] 2.5) Based on the maximum eigenvalue of each indicator level, perform a consistency check on all indicator levels. If all indicator levels pass the consistency check, obtain the approximate weight vector of all indicator levels. If any indicator level fails the consistency check, re-evaluate the importance relationship between indicator levels, adjust the judgment matrix of the indicator levels that failed the consistency check, and return to step 2.3).

[0028] 2.6) Based on the approximate weight vectors of all indicator levels, the initial weights of all third-level indicators are calculated as follows:

[0029] ω=ω c ×ω b ×ω a (4)

[0030] In the formula, ω represents the initial weight of the three-level indicators. a This indicates the weight of the indicator within its corresponding first-level indicator layer. ω b This indicates the weight of the indicator within the second-level indicator layer under its primary indicator. ωc This indicates the weight of the indicator within the tertiary indicator layer under its corresponding secondary indicator.

[0031] Furthermore, the steps for performing consistency checks on all indicator levels are as follows:

[0032] 2.5.1) Calculate the consistency index for each indicator level, as shown below:

[0033]

[0034] In the formula, n represents the total number of indicators at the indicator level, λ represents the largest eigenvalue of the indicator level, and CI represents the consistency index of the indicator level.

[0035] 2.5.2) Determine the random consistency index for each indicator level based on the total number of indicators at each indicator level, and calculate the consistency ratio for each indicator level.

[0036] The consistency ratio is as follows:

[0037]

[0038] In the formula, CR represents the consistency ratio of the indicator level. RI represents the random consistency index of the indicator level.

[0039] 2.5.3) Determine whether the consistency ratio CR < 0.1 of all indicator levels is valid. If yes, all indicator levels have passed the consistency test. If no, there are indicator levels that have not passed the consistency test.

[0040] Furthermore, in step 3), the steps for determining the positive and negative ideal solutions for each tertiary indicator in the multi-dimensional indicator system are as follows:

[0041] 3.1) Let the set of power system planning schemes be J = {J1, J2...J...} N The set of three-level indicators is K = {K1, K2, ..., K}. M}, where N represents the total number of power system planning schemes and M represents the total number of third-level indicators.

[0042] 3.2) Construct the evaluation matrix B = [b pq ] N×M Where p represents the power system planning scheme index, q represents the third-level index, and b pq J represents the p-th power system planning scheme. p For the q-th tertiary indicator K q The evaluation parameter values.

[0043] The evaluation parameter value b pq It is a normally distributed random variable, as shown below:

[0044]

[0045] In the formula, μ pq σ pq J represent the p-th power system planning scheme respectively. p For the q-th tertiary indicator K q The mean and standard deviation of the evaluation parameter values.

[0046] 3.3) For the evaluation matrix B = [b pq ] N×M After regularization, the evaluation matrix C = [c pq ] N×M As shown below:

[0047] c pq ~N(μ) pq ',σ pq ' 2 (8)

[0048]

[0049] x q =max{0,min p (μ pq -3σ pq )} (10)

[0050] y q =max p (μ pq +3σ pq (11)

[0051] In the formula, c pq J represents the p-th power system planning scheme after regularization. p For the q-th tertiary indicator K q The evaluation parameter value. μ pq '、σ pq ' represents the p-th power system planning scheme J after regularization. p For the q-th tertiary indicator K q The mean and standard deviation of the evaluation parameter values. q y q These represent the minimum and maximum values ​​of the evaluation parameter, respectively.

[0052] 3.4) Using the 3σ principle, the regularized evaluation matrix C = [c pq ] N×M Transform into an interval number matrix R = [r pq ] N×M , where r pqJ represents the p-th power system planning scheme. p For the q-th tertiary indicator K q The evaluation parameter range, and r pq =[r pq L ,r pq U ], r pq U r pq L Let r be the upper and lower limits of the interval, respectively, and r be the lower and lower limits of the interval. pq U =μ pq '+3σ pq ', r pq L =μ pq ' - 3σ pq '.

[0053] 3.5) Divide all tertiary indicators into benefit-type indicators and cost-type indicators according to their attributes, and then use the interval number matrix R = [r pq ] N×M The positive and negative ideal solutions for each tertiary indicator are determined as follows:

[0054] r q + = [r q +L ,r q +U ],r q - = [r q -L ,r q -U (12)

[0055] In the formula, r q + r q - These represent the positive and negative ideal solutions for the three-level indicators, respectively.

[0056] Among them, the upper and lower limits r of the positive ideal solution of the third-level indicator q +U r q +L The upper and lower limits r of the negative ideal solution q -U r q -L As shown below:

[0057]

[0058] Furthermore, the nonlinear programming optimization problem is as follows:

[0059]

[0060] In the formula, p represents the power system planning scheme index, q represents the tertiary indicator index, N represents the total number of power system planning schemes, and M represents the total number of tertiary indicators. ω q,opt K represents the q-th tertiary indicator. q The final weight. f(ω) q,opt ) represents a nonlinear programming optimization function. E p J represents the p-th power system planning scheme. p The degree of closeness to the ideal solution. J represent the p-th power system planning scheme respectively. p The distance to the positive ideal solution and the negative ideal solution. pq U r pq L The p-th power system planning scheme J is respectively p For the q-th tertiary indicator K q The upper and lower limits of the evaluation parameter range. q +U r q +L K represents the qth tertiary indicator. q The upper and lower limits of the positive ideal solution. q -U r q -L K represents the qth tertiary indicator. q The upper and lower limits of the negative ideal solution. p up p low These represent the upper and lower threshold parameters for the weights, respectively. ω q K represents the q-th tertiary indicator. q The initial weights.

[0061] Furthermore, in solving nonlinear programming optimization problems, the concept of multiple initial points optimization is introduced, and the simplified form is shown below:

[0062] minf(ω q,opt ),h(ω q,opt )≤0,g(ω q,opt )=0 (16)

[0063] In the formula, q represents the third-level index, and ω q,opt K represents the q-th tertiary indicator. q The weights. f(ω) q,opt ) represents the nonlinear programming optimization function. h(ω) q,opt ) represents the weight ω q,opt All inequality constraint functions. g(ω)q,opt ) represents the weight ω q,opt All equality constraint functions.

[0064] Furthermore, the algorithms for solving nonlinear programming optimization problems include gradient descent, evolutionary algorithms, and particle swarm optimization.

[0065] Furthermore, the optimal solution is the power system planning scheme that has the smallest approximation to the positive ideal solution.

[0066] The technical effectiveness of this invention is undeniable. It proposes a power system planning and evaluation method based on stochastic multi-criteria decision-making. After constructing a multi-dimensional evaluation index system adapted to the characteristics of new energy sources, it utilizes an improved Topsis-AHP method to preserve and comprehensively evaluate the randomness of the indicators. Furthermore, it combines a nonlinear optimization algorithm based on the multi-initial-point concept to objectively and dynamically adjust the subjective decision weights, ultimately providing a scientifically sound power system planning scheme. This invention not only fully considers the volatility and randomness of new energy output but also combines the subjectivity of expert judgment with the objectivity of mathematical algorithms, improving the applicability of the planning and evaluation results, thereby providing more reliable decision support for the planning of high-proportion renewable energy power systems.

[0067] Despite significantly increased uncertainties such as complex evaluation dimensions, a large number of decision variables, and random perturbations in indicators, this invention still demonstrates good robustness and stability. The scoring results remained highly consistent across multiple tests, exhibiting strong ranking stability and consistently acceptable computational efficiency, without convergence failure or abnormal fluctuations. Furthermore, in the process of integrating subjective and objective weights and optimizing multiple initial points, this invention effectively reduced the impact of expert judgment bias on the final results, ensuring that the evaluation process maintains scientific rigor while also considering practicality and flexibility.

[0068] This invention proposes a planning and evaluation method based on stochastic multi-criteria decision-making for power systems with a high proportion of renewable energy. By constructing an improved Topsis-AHP model that integrates subjective and objective weights, and combining stochastic perturbation modeling with a multi-initial-point nonlinear optimization strategy, a comprehensive quantitative evaluation of complex planning schemes is achieved. This invention can provide scientific, robust, and practically engineering-adaptable evaluation and ranking results in situations where renewable energy is highly volatile, evaluation dimensions are numerous, and decision objectives are complex, providing practical technical support for constructing safe, economical, and flexible power system planning. Attached Figure Description

[0069] Figure 1 This is a flowchart of the present invention;

[0070] Figure 2 This is a schematic diagram showing the matching of positive and negative ideal solutions for each scheme. Detailed Implementation

[0071] The present invention will be further described below with reference to embodiments, but it should not be construed that the scope of the present invention is limited to the following embodiments. Various substitutions and modifications made based on ordinary technical knowledge and common practices in the art without departing from the above-described technical concept of the present invention should be included within the scope of protection of the present invention.

[0072] Example 1:

[0073] See Figures 1 to 2 A power system planning and evaluation method based on stochastic multi-criteria decision-making includes the following steps:

[0074] 1) Construct a multi-dimensional indicator system and obtain multiple power system planning schemes. The source of the planning schemes is mainly based on the power system planning requirements, and is generated by designing and optimizing the power supply structure, power grid layout and resource allocation.

[0075] The multi-dimensional indicator system includes primary indicators, secondary indicators, and tertiary indicators.

[0076] 2) Based on multiple power system planning schemes, the AHP method is used to assign initial weights to the third-level indicators in the multi-dimensional indicator system.

[0077] 3) Based on the randomness of the parameters of the third-level indicators in the power system planning scheme, determine the positive and negative ideal solutions of each third-level indicator in the multi-dimensional indicator system.

[0078] 4) Based on the initial weights and positive and negative ideal solutions of each tertiary index, construct a nonlinear programming optimization problem.

[0079] 5) Solve the nonlinear programming optimization problem to obtain the optimal weight of each tertiary index, and select the optimal scheme from multiple power system planning schemes based on the optimal weight of each tertiary index.

[0080] Example 2:

[0081] A power system planning evaluation method based on stochastic multi-criteria decision-making, the main technical contents of which are described in Example 1, further subdivides each primary indicator into secondary indicators, and each secondary indicator into tertiary indicators.

[0082] The primary indicators include economic efficiency, the penetration rate of new energy sources, the capacity for new energy absorption, the reliability of the power system, and the capacity for ensuring power supply.

[0083] The secondary indicators include, but are not limited to, electricity prices, new energy installations, system regulation capacity, load shedding probability, and demand response capacity.

[0084] The three-level indicators include, but are not limited to, electricity price, electric vehicle capacity, average regulation duration of traditional units, probability of load failure, and load shedding in extreme scenarios.

[0085] Example 3:

[0086] A power system planning evaluation method based on stochastic multi-criteria decision-making, the main technical contents of which are described in any one of Examples 1 and 2. Furthermore, the power system planning scheme is a planning scheme that optimizes the power source structure, grid layout and resource allocation, including a scheme that adjusts the installed capacity of new energy sources as a variable.

[0087] Example 4:

[0088] A power system planning and evaluation method based on stochastic multi-criteria decision-making, the main technical contents of which are described in any one of Examples 1 to 3, further, in step 2), the initial weight assignment of the three-level indicators is as follows:

[0089] 2.1) Divide the indicators in the multi-dimensional indicator system to obtain the evaluation indicators of power system planning schemes under different indicator levels.

[0090] The indicator hierarchy includes a primary indicator layer, a secondary indicator layer under the primary indicators, and a tertiary indicator layer under the secondary indicators.

[0091] 2.2) Compare the power system planning scheme evaluation indicators at the same indicator level pairwise, and construct judgment matrices for different indicator levels according to the importance of different indicators.

[0092] 2.3) Normalize the judgment matrix for each indicator level to obtain the approximate weight vector for each indicator level, as shown below:

[0093]

[0094] In the formula, i and j both represent indicator indices, n represents the total number of indicators at the indicator level, and a ij This indicates the degree of importance of the i-th indicator to the j-th indicator in the indicator hierarchy. This indicates the importance of the column after normalization. ω i This represents the approximate weight of the i-th indicator in the indicator hierarchy.

[0095] 2.4) Based on the judgment matrix and approximate weight vector of each indicator level, calculate the maximum eigenvalue of each indicator level, as shown below:

[0096]

[0097] In the formula, λ represents the largest eigenvalue of the indicator level. A represents the judgment matrix of the indicator level. ω' represents the approximate weight vector of the indicator level.

[0098] 2.5) Based on the maximum eigenvalue of each indicator level, perform a consistency check on all indicator levels. If all indicator levels pass the consistency check, obtain the approximate weight vector for all indicator levels. If any indicator level fails the consistency check, re-evaluate the importance relationship between indicator levels, adjust the judgment matrix of the indicator levels that failed the consistency check, and return to step 2.3). An example is shown below:

[0099] For the original judgment matrix:

[0100]

[0101] Calculations showed that its CR≈0.115>0.1, and it was subsequently adjusted to:

[0102]

[0103] Calculations showed that the adjusted CR was approximately 0.04 < 0.1, indicating that the adjustment was successful.

[0104] 2.6) Based on the approximate weight vectors of all indicator levels, the initial weights of all third-level indicators are calculated as follows:

[0105] ω=ω c ×ω b ×ω a (4)

[0106] In the formula, ω represents the initial weight of the three-level indicators. a This indicates the weight of the indicator within its corresponding first-level indicator layer. ω b This indicates the weight of the indicator within the second-level indicator layer under its primary indicator. ω c This indicates the weight of the indicator within the tertiary indicator layer under its corresponding secondary indicator.

[0107] Example 5:

[0108] A power system planning and evaluation method based on stochastic multi-criteria decision-making, the main technical contents of which are described in any one of Examples 1 to 4, further wherein the step of performing consistency verification on all index levels is as follows:

[0109] 2.5.1) Calculate the consistency index for each indicator level, as shown below:

[0110]

[0111] In the formula, n represents the total number of indicators at the indicator level, λ represents the largest eigenvalue of the indicator level, and CI represents the consistency index of the indicator level.

[0112] 2.5.2) Determine the random consistency index for each indicator level based on the total number of indicators at each indicator level, and calculate the consistency ratio for each indicator level.

[0113] The consistency ratio is as follows:

[0114]

[0115] In the formula, CR represents the consistency ratio of the indicator level. RI represents the random consistency index of the indicator level.

[0116] 2.5.3) Determine whether the consistency ratio CR < 0.1 of all indicator levels is valid. If yes, all indicator levels have passed the consistency test. If no, there are indicator levels that have not passed the consistency test.

[0117] Example 6:

[0118] A power system planning and evaluation method based on stochastic multi-criteria decision-making, the main technical contents of which are described in any one of Examples 1 to 5, further, in step 3), the steps for determining the positive and negative ideal solutions of each tertiary index in the multi-dimensional index system are as follows:

[0119] 3.1) Let the set of power system planning schemes be J = {J1, J2, ..., J...} N The set of three-level indicators is K = {K1, K2, ..., K}. M}, where N represents the total number of power system planning schemes and M represents the total number of third-level indicators.

[0120] 3.2) Construct the evaluation matrix B = [b pq ] N×M Where p represents the power system planning scheme index, q represents the third-level index, and b pq J represents the p-th power system planning scheme. p For the q-th tertiary indicator K q The evaluation parameter values.

[0121] The evaluation parameter value b pq It is a normally distributed random variable, as shown below:

[0122]

[0123] In the formula, μ pq σ pq J represent the p-th power system planning scheme respectively. p For the q-th tertiary indicator Kq The mean and standard deviation of the evaluation parameter values.

[0124] 3.3) For the evaluation matrix B = [b pq ] N×M After regularization, the evaluation matrix C = [c pq ] N×M As shown below:

[0125] c pq ~N(μ) pq ',σ pq ' 2 (8)

[0126]

[0127] x q =max{0,min p (μ pq -3σ pq )} (10)

[0128] y q =max p (μ pq +3σ pq (11)

[0129] In the formula, c pq J represents the p-th power system planning scheme after regularization. p For the q-th tertiary indicator K q The evaluation parameter value. μ pq '、σ pq ' represents the p-th power system planning scheme J after regularization. p For the q-th tertiary indicator K q The mean and standard deviation of the evaluation parameter values. q y q These represent the minimum and maximum values ​​of the evaluation parameter, respectively.

[0130] 3.4) Using the 3σ principle, the regularized evaluation matrix C = [c pq ] N×M Transform into an interval number matrix R = [r pq ] N×M , where r pq J represents the p-th power system planning scheme. p For the q-th tertiary indicator K q The evaluation parameter range, and r pq =[r pq L ,r pq U ], rpq U r pq L Let r be the upper and lower limits of the interval, respectively, and r be the lower and lower limits of the interval. pq U =μ pq '+3σ pq ', r pq L =μ pq '-3σ pq '.

[0131] 3.5) Divide all tertiary indicators into benefit-type indicators and cost-type indicators according to their attributes, and then use the interval number matrix R = [r pq ] N×M The positive and negative ideal solutions for each tertiary indicator are determined as follows:

[0132] r q + = [r q +L ,r q +U ],r q - = [r q -L ,r q -U (12)

[0133] In the formula, r q + r q - These represent the positive and negative ideal solutions for the three-level indicators, respectively.

[0134] Among them, the upper and lower limits r of the positive ideal solution of the third-level indicator q +U r q +L The upper and lower limits r of the negative ideal solution q -U r q -L As shown below:

[0135]

[0136] Example 7:

[0137] A power system planning and evaluation method based on stochastic multi-criteria decision-making, the main technical contents of which are described in any one of Examples 1 to 6, and further, the nonlinear programming optimization problem is as follows:

[0138]

[0139] In the formula, p represents the power system planning scheme index, q represents the tertiary indicator index, N represents the total number of power system planning schemes, and M represents the total number of tertiary indicators. ω q,opt K represents the q-th tertiary indicator. q The final weight. f(ω) q,opt ) represents a nonlinear programming optimization function. E p J represents the p-th power system planning scheme. p The degree of closeness to the ideal solution. J represent the p-th power system planning scheme respectively. p The distance to the positive ideal solution and the negative ideal solution. pq U r pq L The p-th power system planning scheme J is respectively p For the q-th tertiary indicator K q The upper and lower limits of the evaluation parameter range. q +U r q +L K represents the qth tertiary indicator. q The upper and lower limits of the positive ideal solution. q -U r q -L K represents the qth tertiary indicator. q The upper and lower limits of the negative ideal solution. p up p low These represent the upper and lower threshold parameters for the weights, respectively. ω q K represents the q-th tertiary indicator. q The initial weights.

[0140] Example 8:

[0141] A power system planning and evaluation method based on stochastic multi-criteria decision-making is presented. The main technical contents are described in any one of Examples 1 to 7. Furthermore, when solving the nonlinear programming optimization problem, the idea of ​​multi-initial-point optimization is introduced, and the simplified form is shown below:

[0142] minf(ω q,opt ),h(ω q,opt )≤0,g(ω q,opt )=0 (16)

[0143] In the formula, q represents the third-level index, and ω q,opt K represents the q-th tertiary indicator. q The weights. f(ω) q,opt ) represents the nonlinear programming optimization function. h(ω) q,opt ) represents the weight ω q,opt All inequality constraint functions. g(ω)q,opt ) represents the weight ω q,opt All equality constraint functions.

[0144] Example 9:

[0145] A power system planning and evaluation method based on stochastic multi-criteria decision-making is provided. The main technical contents are described in any one of Examples 1 to 8. Furthermore, the algorithm for solving the nonlinear programming optimization problem includes gradient descent algorithm, evolutionary algorithm, and particle swarm optimization algorithm.

[0146] Example 10:

[0147] A power system planning evaluation method based on stochastic multi-criteria decision-making, the main technical contents of which are described in any one of Examples 1 to 9, further wherein the optimal solution is the power system planning scheme with the smallest closeness to the positive ideal solution.

[0148] Example 11:

[0149] See Figures 1 to 2 A power system planning and evaluation method based on stochastic multi-criteria decision-making includes the following steps:

[0150] 1. Construct a multi-dimensional indicator system based on the characteristics of the new power system.

[0151] This embodiment first addresses the operational characteristics of the power system under the background of high proportion of renewable energy access, and constructs a five-dimensional indicator system that includes the economy of power system planning schemes, the level of new energy penetration, the capacity of new energy absorption, system reliability and supply guarantee, so as to achieve a comprehensive characterization and evaluation of planning schemes.

[0152] The economic dimension is used to assess the rationality of power system resource allocation and the cost-effectiveness of operation. This dimension covers aspects such as investment, operation and maintenance, load shedding costs, and electricity prices, and can reflect the overall cost control capability and market competitiveness of the power system. It is a key factor in measuring the economic feasibility and optimization potential of planning schemes.

[0153] The renewable energy penetration level dimension focuses on factors such as the installed capacity, capacity reliability, and power generation efficiency of renewable energy generation in the system, reflecting the proportion and operating characteristics of renewable energy in the power system. This dimension helps to measure the diversity and cleanliness of the system's energy structure.

[0154] The renewable energy absorption capacity dimension measures the system’s internal ability to absorb and regulate renewable energy output. It covers multiple aspects such as curtailment rate, system regulation performance, source-load coupling characteristics and transmission capacity, and can comprehensively reflect the operational pressure and resource coordination ability after renewable energy is connected.

[0155] The power system reliability dimension is used to characterize the degree of continuity and stability of power supply under normal or long-term operating conditions. Relevant indicators include expected power shortage, average outage time, probability of load shedding, and probability of wind and solar power curtailment.

[0156] The power system supply guarantee capacity dimension mainly assesses the system's ability to guarantee power supply under abnormal scenarios such as extreme weather or sudden disasters. It involves indicators such as maximum peak-shaving capacity, adjustable load resources, load shedding and supply guarantee rate under disaster scenarios, and is an important basis for measuring the system's resilience and emergency response level.

[0157] This embodiment further refines the indicator system under the above five dimensions into a three-tiered structure of "primary indicators → secondary indicators → tertiary indicators" to improve the resolution and adaptability of the evaluation. The complete framework composed of indicators at each level is shown in Table 1. The subsequent evaluation model will be based on this system to carry out quantitative analysis and decision optimization.

[0158] Table 1. Multi-dimensional indicator evaluation system structure used in this embodiment.

[0159]

[0160] 2. Subjective and Objective Stochastic Multi-Attribute Decision Theory Based on Improved Topsis-AHP Method

[0161] After constructing a multi-dimensional indicator system, to achieve a scientific and quantitative evaluation of various power system planning schemes, this embodiment introduces an improved Topsis-AHP multi-attribute decision-making method that integrates subjective and objective information and adapts to the uncertainties of new energy sources. This method combines the expert subjective judgment of the Analytic Hierarchy Process (AHP) with the objective distance evaluation of the Technique for Order Preference by Similarity to Ideal Solution (Topsis), and introduces stochastic perturbation modeling and a multi-initial-point optimization mechanism for the indicators, constructing a stochastic multi-criteria decision-making theoretical system that integrates subjective and objective factors and meets the requirements of actual engineering. The specific description of this theory is as follows:

[0162] First, for the established indicator system, the Analytic Hierarchy Process (AHP) is used to assign initial weights to each indicator. Then, for a given criterion, each solution is compared pairwise, and its importance is assessed to form a judgment matrix, as shown in Table 2.

[0163] Table 2. Rules for constructing the AHp method used in this embodiment.

[0164] Factor i compared to factor j Quantized value Equally important 1 Slightly important 3 Stronger and more important 5 Extremely important 7 The median value between two adjacent judgments 2，4，6，8

[0165] After construction, one judgment matrix should be established for the primary indicators, five judgment matrices for the secondary indicators, and 21 judgment matrices for the tertiary indicators. Note that each judgment matrix should be a square matrix.

[0166] For the judgment matrix A = [a ij ] n×n Normalize it:

[0167]

[0168] Then, row normalization is performed:

[0169]

[0170] The resulting column vector ω′ = (ω1, ω2, ... ω) is then obtained. n ) T This refers to the approximate weights of the indicators. Then, using the original judgment matrix A and the obtained approximate weight vector ω′, the approximate maximum eigenvalue λ is calculated:

[0171]

[0172] After calculating the weight vector and eigenvalues ​​of the judgment matrix, a consistency check must be performed on this level. Only if the check passes can the ranking weights of this level be officially confirmed; if it fails, the judgment matrix needs to be readjusted. The consistency check process is as follows:

[0173] Based on equation (3), the consistency index CI is calculated for the approximate weight vector ω':

[0174]

[0175] Then, based on Table 3, the random consistency index RI for the corresponding n is read out:

[0176] Table 3 Reference Standard Values ​​of the Random Consistency Index (CR) for Different Matrix Orders

[0177]

[0178] Then, the consistency ratio CR is calculated. If CR < 0.1, the consistency test is passed.

[0179]

[0180] After the judgment matrices of all first-, second-, and third-level indicators pass the consistency test, a cascade multiplication method is performed to obtain the weight vector of the lowest level, i.e., the third-level indicator. For a certain third-level indicator, the third-level weight under its corresponding second-level indicator is known to be ω. c The secondary weight of the secondary indicator layer is ω. bThe weight of the primary indicator to which its secondary indicator belongs is ω a Then its total ranking weight relative to the highest level, i.e., the first-level weight, is:

[0181] ω=ω c ×ω b ×ω a (6)

[0183] Let ω at this point be the initial weight of the index.

[0184] Next, the data for each scheme, which involves randomness, are processed. The uncertainty of new energy sources means that their relevant indicator parameters do not have definite values. Therefore, to enhance rigor and retain their original randomness, these variables are approximated as normally distributed. Let the set of schemes for the multi-attribute decision problem be J = {J1, J2...J...}. n The index set is K = {K1, K2, ..., K}. m}, B = [b ij ] n×m , represents the evaluation matrix, b ij It is scheme J i Corresponding to index K j The parameter value, here b ij It is a normally distributed random variable, that is:

[0185] The following is an evaluation matrix B = [b ij ] n×m To perform regularization, let:

[0186] x j =max{0,min i (μ ij -3σ ij )}(7)

[0187] y j =max i (μ ij +3σ ij )}(8)

[0188] Based on equations (7) and (8), the following operations are performed:

[0189]

[0190] c ij ~N(μ) ij ',σ ij ' 2 (10)

[0191] Using the 3σ principle, the regularized evaluation matrix C = [c ij ] n×mTransform into an interval number matrix R = [r ij ] n×m , where r ij =[r ij L ,r ij U ], r ij U =μ ij '+3σ ij ', r ij L =μ ij ' - 3σ ij '.

[0192] Each indicator attribute is categorized into benefit-type (Category A) or cost-type (Category B), and this is further analyzed using the interval number matrix R = [r...]. ij ] n×m Confirm the positive and negative ideal values ​​for each indicator. For indicator K... j Define its positive ideal index value r j + =[r j +L ,r j +U ], negative ideal index value r j - =[r j -L ,r j -U ]:

[0193] When K j For efficiency-related indicators, there are:

[0194]

[0195] When K j For cost-related indicators, we have:

[0196]

[0197] After calculating the positive and negative ideal values ​​of all indicators, we can obtain the positive and negative ideal solutions for the existing schemes and indicators:

[0198] ideal + = {r1 + r2 + r3 + ...r j +} ;ideal - = {r1 - r2 - r3 - ...rj -} (13)

[0199] Note that each r j + and r j - They are all in the form of interval numbers.

[0200] Based on the initial weights of the indicators in equation (6), initial upper and lower thresholds for disturbance are established:

[0201] p low ·w j ≤w j,opt ≤p up ·w j (14)

[0202] p low ≤1≤p up ;p low +p up =2 (15)

[0203] Using equations (11-12) and (14-15), calculate the distance from any solution to the positive and negative ideal solutions, and the distance from any solution to the positive ideal solution. for:

[0204]

[0205] Distance from any solution to the negative ideal solution for:

[0206]

[0207] The closeness of each solution to the ideal solution is as follows, with the solution having the smallest closeness being the optimal solution:

[0208]

[0209] From equations (6) and (14-18), a nonlinear programming optimization problem for solving the optimal weights is constructed:

[0210]

[0211] To ensure that the nonlinear programming model described in equation (19) does not get trapped in local optima during the solution process, the idea of ​​multi-initial-point optimization is introduced. For the optimization problem:

[0212] min f(ω),h(ω)≤0,g(ω)=0(20)

[0213] Given k initial points ω randomly (1) ,ω (2) ,ω(3) ......ω (k) For each point ω (k) Execute optimization algorithms (such as gradient descent, evolutionary algorithms, particle swarm optimization, etc.) to obtain a local optimum or approximate solution. Select the best solution as the final global optimal solution:

[0214]

[0215] Based on equations (19-20), this nonlinear programming optimization problem is solved, and the weights that minimize the sum of the closeness of all solutions to the positive ideal solution are obtained, denoted as the optimal weights ω. opt =(ω 1,opt ,ω 2,opt ...ω n,opt ) T The solution with the smallest sum of proximity is denoted as the optimal solution J. opt .

[0216] Example 12:

[0217] A power system planning evaluation method based on stochastic multi-criteria decision-making is presented. The main technical content is described in Example 11. Furthermore, to verify the applicability and effectiveness of the proposed power system planning evaluation method based on stochastic multi-criteria decision-making, this example selects the HRP-38 power system provided by the Tsinghua University team as a test platform to simulate and evaluate different planning schemes. The HRP-38 system is a simplified version of the 2030 planning scheme for a 750kV power grid in a certain region of my country, exhibiting typical characteristics of high-proportion renewable energy integration. The system contains 38 nodes and designs multiple candidate transmission lines, covering both AC and DC channels. The installed capacity of non-hydro renewable energy accounts for over 47%, and the electricity penetration rate is approximately 30%, making it highly representative and suitable as a public test platform for research on complex power system planning and operation.

[0218] While keeping the network structure and load demand unchanged, this embodiment designs four planning schemes, adjusting the installed capacity of new energy sources as variables to reflect the impact of different proportions of new energy sources on the evaluation indicators:

[0219] Case 1 (M1): Baseline solution, with installed capacity set according to existing plans;

[0220] Case 2 (M2): Based on the baseline plan, the renewable energy capacity is reduced to 80% of the original plan;

[0221] Case 3 (M3): The renewable energy capacity was increased to 110% of the original plan;

[0222] Case 4 (M4): The capacity of new energy sources was increased to 150% of the original plan.

[0223] Using the Matlab simulation platform, 36 tertiary indicator data under each scheme were obtained. Combining the normalization and stochastic modeling mechanism proposed in this embodiment, some new energy-related indicator parameters were modeled as random variables that follow a normal distribution, and an interval number matrix was constructed accordingly for subsequent Topsis calculation.

[0224] Based on the Analytic Hierarchy Process (AHP), judgment matrices for first-level, second-level, and third-level indicators were constructed, and the initial weights of each indicator were obtained after consistency testing. Subsequently, a nonlinear optimization model was constructed, and the initial weight parameter p was set. low =0.9, p up =1.1, and a multi-initial-point optimization strategy was introduced to finally obtain the optimal index weight, as shown in Table 4.

[0225] Substitute the optimal weights into the improved Topsis model to calculate the similarity E of each scheme. i The results are shown in Table 5.

[0226] Based on the proximity calculation results in Table 5, and according to the ranking principle that "the smaller the proximity, the better the solution," solution 3 can be determined as the optimal solution, and the order of superiority is as follows:

[0227] Option 3 > Option 1 > Option 4 > Option 2

[0228] The results demonstrate that, under the indicator system and evaluation model constructed in this embodiment, appropriately increasing the installed capacity of new energy sources can effectively balance the power system's ability to absorb the volatility of new energy sources while improving the system's economy and reliability. This conclusion is highly consistent with the current qualitative understanding of new energy planning in the power industry, further validating the rationality and engineering applicability of the method in this embodiment.

[0229] In addition, a positive and negative ideal solution matching analysis is performed on the final optimization results, such as... Figure 1 As shown, Scheme 3 is closer to the positive ideal solution in multiple indicator dimensions and has the fewest negative ideal solution matches, further verifying the practicality and rationality of this method in dealing with multi-objective decision-making and random fluctuation scenarios.

[0230] In contrast, even under conditions of significantly increased uncertainty, such as complex evaluation dimensions, a large number of decision variables, and random perturbations in the indicators, the method in this embodiment still demonstrates good robustness and stability. The scoring results remained highly consistent across multiple test experiments, exhibiting strong ranking stability, and the computational efficiency remained within an acceptable range, without any convergence failures or abnormal fluctuations. Furthermore, in the process of integrating subjective and objective weights and optimizing multiple initial points, this method effectively reduced the impact of expert judgment bias on the final results, ensuring that the evaluation process maintains scientific rigor while also considering practicality and flexibility.

[0231] Table 4 Comparison of Indicator Attributes and Weights in This Embodiment

[0232] Third-level indicator name Indicator Attributes Initial weights Optimal weight Annual investment cost B 0.066924 0.0736 Operation and maintenance costs B 0.066924 0.0602 Shedding Costs B 0.066924 0.0736 Electricity price A 0.066924 0.068 Total installed capacity of wind power and photovoltaic power A 0.020897 0.023 Electric vehicle capacity A 0.003017 0.0027 Total installed capacity of centralized renewable energy A 0.011201 0.0123 Total installed capacity of distributed renewable energy A 0.0056 0.0062 Credibility of wind power generation A 0.010179 0.0092 Reliability of photovoltaic power generation A 0.010179 0.0112 Annual power generation A 0.027143 0.0244 Annual utilization hours A 0.054287 0.0597 New energy curtailment rate B 0.020671 0.0186 The average annual increase in power generation for traditional generating units A 0.012855 0.0116 Traditional generating units reduce average annual power generation A 0.012855 0.0141 Average regulation duration of traditional units B 0.004285 0.0047 Average regulation duration of energy storage within the system to be evaluated B 0.004285 0.0047 Average load regulation duration within the system to be evaluated B 0.004285 0.0039 Maximum peak-shaving capacity of wind power A 0.001856 0.002 Minimum peak-shaving capacity of wind power A 0.000928 0.001 Average photovoltaic output during peak load period A 0.001856 0.002 Average photovoltaic output during the mid-load period A 0.000928 0.001 Intra-provincial transmission capacity A 0.003445 0.0038 Inter-provincial transmission capacity A 0.00689 0.0062 Power shortage expectation B 0.008734 0.0079 Average power outage time B 0.008734 0.0079 Probability of load loss B 0.008734 0.0096 wind curtailment probability B 0.008734 0.0096 Probability of discarded light B 0.008734 0.0096 Peak energy storage capacity in extreme scenarios A 0.023594 0.026 Duration of energy storage supply in extreme scenarios B 0.011797 0.013 Interruptible load power A 0.013712 0.0123 Transferable load power A 0.006856 0.0075 Load shedding in extreme scenarios B 0.126082 0.1387 Extreme scenario load shedding loss B 0.221833 0.1996 Disaster load supply guarantee rate A 0.067117 0.0604

[0233] Table 5 Comparison of the similarity of each scheme in this embodiment.

[0234] Solution Name M1 M2 M3 M4 <![CDATA[E i ]]> 0.3182 0.6184 0.2670 0.3817

[0235] In summary, this embodiment proposes a planning and evaluation method based on stochastic multi-criteria decision-making for power systems with a high proportion of renewable energy. By constructing an improved Topsis-AHP model that integrates subjective and objective weights, and combining stochastic perturbation modeling with a multi-initial-point nonlinear optimization strategy, a comprehensive quantitative evaluation of complex planning schemes is achieved. This method can provide scientific, robust, and practically engineering-adaptable evaluation and ranking results in situations where renewable energy is highly volatile, evaluation dimensions are numerous, and decision objectives are complex. It provides practical technical support for constructing safe, economical, and flexible power system planning.

Claims

1. A power system planning and evaluation method based on stochastic multi-criteria decision-making, characterized in that, Includes the following steps: 1) Construct a multi-dimensional indicator system and obtain multiple power system planning schemes; The multi-dimensional indicator system includes primary indicators, secondary indicators, and tertiary indicators. 2) Based on multiple power system planning schemes, the AHP method is used to assign initial weights to the third-level indicators in the multi-dimensional indicator system; 3) Based on the randomness of the parameters of the third-level indicators in the power system planning scheme, determine the positive and negative ideal solutions for each third-level indicator in the multi-dimensional indicator system; 4) Based on the initial weights and positive and negative ideal solutions of each tertiary index, construct a nonlinear programming optimization problem; 5) Solve the nonlinear programming optimization problem to obtain the optimal weight of each tertiary index, and select the optimal scheme from multiple power system planning schemes based on the optimal weight of each tertiary index.

2. The power system planning and evaluation method based on stochastic multi-criteria decision-making according to claim 1, characterized in that, Each primary indicator is further subdivided into secondary indicators, and each secondary indicator is further subdivided into tertiary indicators. The primary indicators include economic efficiency, new energy penetration level, new energy absorption capacity, power system reliability, and supply guarantee capacity. The secondary indicators include, but are not limited to, electricity price, new energy installed capacity, system regulation capacity, probability of load shedding, and demand response capacity; The three-level indicators include, but are not limited to, electricity price, electric vehicle capacity, average regulation duration of traditional units, probability of load failure, and load shedding in extreme scenarios.

3. The power system planning and evaluation method based on stochastic multi-criteria decision-making according to claim 1, characterized in that, The power system planning scheme is a planning scheme that optimizes the power supply structure, power grid layout and resource allocation, including schemes that adjust based on the installed capacity of new energy sources.

4. The power system planning and evaluation method based on stochastic multi-criteria decision-making according to claim 1, characterized in that, In step 2), the initial weight assignment steps for the three-level indicators are as follows: 2.1) Divide the indicators in the multi-dimensional indicator system to obtain the evaluation indicators of power system planning schemes under different indicator levels; The indicator hierarchy includes a primary indicator layer, a secondary indicator layer under the primary indicators, and a tertiary indicator layer under the secondary indicators. 2.2) Compare the evaluation indicators of power system planning schemes under the same indicator level pairwise, and construct judgment matrices for different indicator levels according to the importance of different indicators. 2.3) Normalize the judgment matrix for each indicator level to obtain the approximate weight vector for each indicator level, as shown below: In the formula, i and j both represent indicator indices, n represents the total number of indicators at the indicator level, and a ij This indicates the degree of importance of the i-th indicator to the j-th indicator in the indicator hierarchy; Indicates the importance of the column after normalization; ω i This represents the approximate weight of the i-th indicator in the indicator hierarchy; 2.4) Based on the judgment matrix and approximate weight vector of each indicator level, calculate the maximum eigenvalue of each indicator level, as shown below: In the formula, λ represents the largest eigenvalue of the indicator level; A represents the judgment matrix of the indicator level; ω' represents the approximate weight vector of the indicator level. 2.5) Based on the maximum eigenvalue of each indicator level, perform a consistency test on all indicator levels. If all indicator levels pass the consistency test, obtain the approximate weight vector of all indicator levels. If any indicator level fails the consistency test, re-evaluate the importance relationship between indicator levels, adjust the judgment matrix of the indicator levels that failed the consistency test, and return to step 2.3). 2.6) Based on the approximate weight vectors of all indicator levels, the initial weights of all third-level indicators are calculated as follows: oh = oh c ×ω b ×ω a (4) In the formula, ω represents the initial weight of the three-level indicators; ω a This indicates the weight of the indicator within its corresponding first-level indicator layer; ω b This indicates the weight of the indicator within the secondary indicator layer under its primary indicator; ω c This indicates the weight of the indicator within the tertiary indicator layer under its corresponding secondary indicator.

5. The power system planning and evaluation method based on stochastic multi-criteria decision-making according to claim 4, characterized in that, The steps for performing consistency checks on all indicator levels are as follows: 2.5.1) Calculate the consistency index for each indicator level, as shown below: In the formula, n represents the total number of indicators at the indicator level, λ represents the largest eigenvalue of the indicator level, and CI represents the consistency index of the indicator level. 2.5.2) Determine the random consistency index for each indicator level based on the total number of indicators at each indicator level, and calculate the consistency ratio for each indicator level. The consistency ratio is as follows: In the formula, CR represents the consistency ratio of the indicator level; RI represents the random consistency index of the indicator level. 2.5.3) Determine whether the consistency ratio CR < 0.1 of all indicator levels is valid. If yes, all indicator levels have passed the consistency test. If no, there are indicator levels that have not passed the consistency test.

6. The power system planning and evaluation method based on stochastic multi-criteria decision-making according to claim 1, characterized in that, In step 3), the steps for determining the positive and negative ideal solutions for each tertiary indicator in the multi-dimensional indicator system are as follows: 3.1) Let the set of power system planning schemes be J = {J1, J2...J...} N The set of three-level indicators is K = {K1, K2, ..., K}. M }, where N represents the total number of power system planning schemes and M represents the total number of third-level indicators; 3.2) Construct the evaluation matrix B = [b pq ] N×M Where p represents the power system planning scheme index, q represents the third-level index, and b pq J represents the p-th power system planning scheme. p For the q-th tertiary indicator K q Evaluation parameter values; The evaluation parameter value b pq It is a normally distributed random variable, as shown below: In the formula, μ pq σ pq J represent the p-th power system planning scheme respectively. p For the q-th tertiary indicator K q The mean and standard deviation of the evaluation parameter values; 3.3) For the evaluation matrix B = [b pq ] N×M After regularization, the evaluation matrix C = [c pq ] N×M As shown below: c pq ~N(μ pq ',s pq ' 2 ) (8) x q =max{0,min p (m pq -3s pq )} (10) y q =max p (m pq +3s pq )} (11) In the formula, c pq J represents the p-th power system planning scheme after regularization. p For the q-th tertiary indicator K q Evaluation parameter value; μ pq '、σ pq ' represents the p-th power system planning scheme J after regularization. p For the q-th tertiary indicator K q The mean and standard deviation of the evaluation parameter values; x q y q These represent the minimum and maximum values ​​of the evaluation parameter, respectively. 3.4) Using the 3σ principle, the regularized evaluation matrix C = [c pq ] N×M Transform into an interval number matrix R = [r pq ] N×M , where r pq J represents the p-th power system planning scheme. p For the q-th tertiary indicator K q The evaluation parameter range, and r pq =[r pq L ,r pq U ], r pq U r pq L Let r be the upper and lower limits of the interval, respectively, and r be the lower and lower limits of the interval. pq U =μ pq '+3σ pq ', r pq L =μ pq ' - 3σ pq '; 3.5) Divide all tertiary indicators into benefit-type indicators and cost-type indicators according to their attributes, and then use the interval number matrix R = [r pq ] N×M The positive and negative ideal solutions for each tertiary indicator are determined as follows: r q + ＝[r q +L ,r q +U ],r q - ＝[r q -L ,r q -U ] (12) In the formula, r q + r q - These represent the positive and negative ideal solutions for the three-level indicators, respectively. Among them, the upper and lower limits r of the positive ideal solution of the third-level indicator q +U r q +L The upper and lower limits r of the negative ideal solution q -U r q -L As shown below:

7. The power system planning and evaluation method based on stochastic multi-criteria decision-making according to claim 1, characterized in that, The nonlinear programming optimization problem is as follows: In the formula, p represents the power system planning scheme index, q represents the tertiary indicator index, N represents the total number of power system planning schemes, and M represents the total number of tertiary indicators; ω q,opt K represents the q-th tertiary indicator. q The final weight; f(ω) q,opt ) represents a nonlinear programming optimization function; E p J represents the p-th power system planning scheme. p The degree of closeness to the ideal solution; J represent the p-th power system planning scheme respectively. p Distance to the positive ideal solution and the negative ideal solution; r pq U r pq L The p-th power system planning scheme J is respectively p For the q-th tertiary indicator K q The upper and lower limits of the evaluation parameter range; r q +U r q +L K represents the qth tertiary indicator. q The upper and lower limits of the positive ideal solution; r q -U r q -L K represents the qth tertiary indicator. q The upper and lower limits of the negative ideal solution; p up p low These represent the upper and lower threshold parameters of the weight, respectively; ω q K represents the q-th tertiary indicator. q The initial weights.

8. The power system planning and evaluation method based on stochastic multi-criteria decision-making according to claim 1, characterized in that, In solving nonlinear programming optimization problems, the concept of multiple initial points optimization is also introduced, and the simplified form is shown below: minf(ω q,opt ),h(ω q,opt )≤0,g(ω q,opt )=0 (16) In the formula, q represents the three-level index, ω q,opt K represents the q-th tertiary indicator. q The weights; f(ω) q,opt ) represents the nonlinear programming optimization function; h(ω) q,opt ) represents the weight ω q,opt All inequality constraint functions; g(ω) q,opt ) represents the weight ω q,opt All equality constraint functions.

9. A power system planning and evaluation method based on stochastic multi-criteria decision-making according to claim 1, characterized in that, The algorithms for solving nonlinear programming optimization problems include gradient descent, evolutionary algorithms, and particle swarm optimization.

10. A power system planning and evaluation method based on stochastic multi-criteria decision-making according to claim 1, characterized in that, The optimal solution is the power system planning scheme that has the smallest approximation to the positive ideal solution.

Images