CN106910141B - Complex active power distribution network decomposition scheme optimization method and device - Google Patents

Abstract

The present invention relates to a method and device for optimizing a decomposition scheme of a complex active distribution network. The method includes: obtaining a set of decomposition schemes of a complex active distribution network and evaluation indicators corresponding to each decomposition scheme in the set of decomposition schemes; according to The set of decomposition schemes and the evaluation indicators corresponding to each decomposition scheme in the set of decomposition schemes are determined by the set pair analysis method to determine the sequence of pros and cons of each decomposition scheme; the stability analysis of the sequence of pros and cons is performed and updated The superiority and inferiority sorting sequence selects the decomposition scheme corresponding to the top element in the superiority and inferiority ranking sequence as the optimal decomposition scheme; the technical scheme provided by the present invention makes the complex active distribution network decomposition scheme decision-making model The identity and oppositeness of influencing factors can be considered at the same time, which provides a new idea for dealing with the optimal decision-making problem of complex active distribution network decomposition schemes.

Description

A complex active distribution network decomposition scheme optimization method and device

Technical Field

The present invention relates to the field of distribution networks, and in particular to a method and device for optimizing a complex active distribution network decomposition scheme.

Background Art

Set Pair Analysis (SPA) is an uncertainty theory, which was proposed by Chinese scholar Zhao Keqin in 1989. It is a systematic analysis method for quantitative analysis of deterministic and uncertain systems - similarity, difference and opposite. Its core idea is to include deterministic information and uncertain information in the same system, study the certainty and uncertainty between things from the three aspects of similarity, difference and opposite, and comprehensively describe the connection and transformation between things.

The basic concept of set pair analysis is set pairs and their connection degree. A set pair is a pair of two sets with a certain connection. According to a certain characteristic of the set pair, the connection of the set pair based on this characteristic is classified and quantitatively described, and the connection degree expression of the set pair in a certain problem background is obtained as follows:

μ＝a+bi+cj

In the formula: μ is called the degree of connection. For a specific problem, the degree of connection is generally only a structural function and is only a numerical value in special cases. a represents the degree of identity between two sets, called the degree of identity; b represents the degree of uncertainty of the difference between two sets, called the degree of difference; c represents the degree of opposition between two sets, called the degree of opposition; i is the coefficient of the degree of difference, which takes values in [-1, 1]. i varies between -1 and 1, reflecting the mutual conversion between certainty and uncertainty. As i→0, the uncertainty increases significantly, and when i takes -1 and 1, the problem is deterministic; j is the symbol of the opposition mark or the corresponding coefficient, and the specified value is -1.

The degree of connection can uniformly handle the uncertainty caused by fuzziness, randomness, and incomplete information. This characterization is a quantitative description of certainty and uncertainty, where a and c are relatively certain, and b is relatively uncertain. a, b, and c meet the following normalization conditions:

a+b+c=1

This relativity is due to the complexity and variability of the objective object, as well as the uncertainty caused by the subjectivity and ambiguity of the understanding and characterization of the objective object. Therefore, in formula (1), certainty and uncertainty, identity and opposition are relative and ambiguous in cognition, and the result of the characterization is also relative and non-unique. Set pair analysis effectively describes the unity of opposites between certainty and uncertainty systems, which is in line with the dialectics of nature and human thinking and has methodological significance.

The distribution network has many devices, complex structures, large scale, and relatively complex analysis and calculation. Especially with the development of smart grids, a large number of distributed power sources are connected to the distribution network, and the traditional distribution network has gradually developed into a complex active distribution network, which further increases the computational scale of network analysis and calculation. Network decomposition can greatly reduce the computational scale of the problem and save a lot of computing time. For complex active distribution networks, there are often multiple network decomposition schemes. How to select the best solution from many feasible network decomposition schemes, in addition to considering the calculation speed, also needs to consider many factors such as calculation accuracy and resource utilization. In actual decision-making, these factors that need to be considered are often both contradictory and unified. How to make these factors well unified in a network decomposition scheme decision model is a problem worth studying.

Summary of the invention

The present invention provides a complex active distribution network decomposition scheme optimization method and device, the purpose of which is to enable the complex active distribution network decomposition scheme decision model to simultaneously consider the identity and opposition of influencing factors, and provide a new idea for dealing with the complex active distribution network decomposition scheme optimization decision problem.

The purpose of the present invention is achieved by adopting the following technical solutions:

A method for optimizing a complex active distribution network decomposition scheme, the improvement of which includes:

Obtaining a set of decomposition schemes of a complex active distribution network and an evaluation index corresponding to each decomposition scheme in the set of decomposition schemes;

According to the decomposition scheme set and the evaluation index corresponding to each decomposition scheme in the decomposition scheme set, a set pair analysis method is used to determine the superiority and inferiority ranking sequence of each decomposition scheme;

A stability analysis is performed on the merit-ranked sequence, the merit-ranked sequence is updated, and a decomposition scheme corresponding to the top-ranked element in the merit-ranked sequence is selected as an optimal decomposition scheme.

Preferably, the step of obtaining a set of decomposition schemes of a complex active distribution network and an evaluation index corresponding to each decomposition scheme in the set of decomposition schemes includes:

The resource utilization index, parallel computing complexity index, parallel computing accuracy index, parallel computing speedup ratio index, parallel computing efficiency index and parallel computing cost index corresponding to each decomposition scheme in the decomposition scheme set are determined respectively.

Furthermore, the resource utilization index y ₁ corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

In the above formula, ⁿ is the number of partitions in the network, _Pi2 is the computational scale of the ith partition in the network, i∈[1,n];

The parallel computing complexity index y ₂ corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

为网络中第i个分区的计算复杂度，l为协调级服务器数目，M为计算代价系数，i∈[1,n]，M∈[2,5]，n为网络的分区数目；In the above formula, O _min is the theoretical minimum parallel computing complexity,

is the computational complexity of the i-th partition in the network, l is the number of coordination-level servers, M is the computational cost coefficient, i∈[1,n], M∈[2,5], n is the number of partitions in the network;

The parallel computing accuracy index y ₃ corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

In the above formula, U _i ' is the voltage value of node i after network decomposition, U _i is the voltage value of node i before network decomposition, i∈[1,m], m is the total number of nodes in the network;

The parallel computing speedup ratio index S _P corresponding to each decomposition scheme in the decomposition scheme set is determined as follows:

S _P = _TS / _TP

In the above formula, _TS is the time required to solve the problem serially, and _TP is the time required to solve the problem in parallel;

The parallel computing efficiency index E corresponding to each decomposition scheme in the decomposition scheme set is determined as follows:

E＝ _SP /P

In the above formula, S _P is the parallel computing speedup ratio, and P is the number of servers required for parallel computing;

The parallel computing cost index C corresponding to each decomposition scheme in the decomposition scheme set is determined as follows:

C＝T _P *P

In the above formula, _TP is the time required to solve the problem in parallel, and P is the number of servers required for parallel computing.

Preferably, the step of determining the superiority and inferiority ranking sequence of each decomposition scheme by using a set pair analysis method according to the decomposition scheme set and the evaluation index corresponding to each decomposition scheme in the decomposition scheme set includes:

Convert the non-profit indicators in the evaluation indicators corresponding to each decomposition scheme into profit indicators;

Performing dimensionless processing on the evaluation indicators corresponding to the decomposition schemes to obtain normalized indicator values corresponding to the decomposition schemes;

Selecting an optimal normalized index value set U=(u ₁ ,u ₂ ,... _un ) and a worst normalized index value set V=(v ₁ ,v ₂ ,...v _n ) from the normalized index values corresponding to the decomposition schemes, and constructing a comparison space [V,U] of the decomposition schemes;

In the comparison space [V, U] of the decomposition schemes, determining the relative closeness of the decomposition schemes to the optimal normalized index value set U;

The decomposition schemes are sorted in descending order according to the relative closeness between the decomposition schemes and the optimal normalized index value set U, so as to obtain a ranking sequence of the advantages and disadvantages of the decomposition schemes.

Furthermore, the converting of the non-profit indicators in the evaluation indicators corresponding to each decomposition scheme into profit indicators includes:

为非收益型指标，则按下式将

转换为收益型指标：Assume that the index value of the k-th decomposition scheme with respect to the r-th index is

If it is a non-income indicator, then

Converted to profit-based indicators:

为

的收益型指标值，

为第k个分解方案关于第r个指标的最大指标值，k∈[1,m]，r∈[1,n]，m为分解方案总数，n为评价指标总数。In the above formula,

for

The income indicator value of

is the maximum index value of the k-th decomposition scheme with respect to the r-th index, k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indicators.

Furthermore, the dimensionless processing is performed on the evaluation indicators corresponding to the decomposition schemes to obtain the normalized indicator values corresponding to the decomposition schemes, including:

The evaluation indicators corresponding to the decomposition schemes are dimensionlessly processed as follows:

为第k个分解方案关于第r个指标的指标值，k∈[1,m]，r∈[1,n]，m为分解方案总数，n为评价指标总数。In the above formula, d _kr is the normalized index value of the k-th decomposition scheme with respect to the r-th index,

is the index value of the kth decomposition scheme with respect to the rth index, k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indicators.

Furthermore, determining the relative closeness of each decomposition scheme to the optimal normalized index value set U in the comparison space [V, U] of each decomposition scheme includes:

Suppose the decomposition scheme set S = {s ₁ ,s ₂ ,...,s _m }, the evaluation index set E = {e ₁ ,e ₂ ,...,e _n }, let the normalized index value of the k-th decomposition scheme with respect to the r-th index be d _kr , the optimal normalized index value among the normalized index values of each decomposition scheme with respect to the r-th index be _ur , and the worst normalized index value among the normalized index values of each decomposition scheme with respect to the r-th index be v _r , where k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indexes;

In the comparison space [V, U] of the decomposition schemes, the connection degree between the kth decomposition scheme and the optimal normalized index value set U, that is, the connection degree u{s _k ,U} of the set pair {s _k ,U} is determined as follows:

u{s _k ,U}＝ _ak + _bki + _ckj

In the above formula, s _k is the kth decomposition scheme in the decomposition scheme set, a _k is the identity between the kth decomposition scheme and the optimal normalized index value set U, b _k is the difference between the kth decomposition scheme and the optimal normalized index value set U, c _k is the opposition between the kth decomposition scheme and the optimal normalized index value set U, i is the difference coefficient, and j is the opposition mark symbol;

Among them, i∈[-1,1], j＝1,

The relative closeness γ _k between the k-th decomposition scheme and the optimal normalized index value set U is determined as follows:

Preferably, the performing stability analysis on the merit-ranking sequence, updating the merit-ranking sequence and selecting the decomposition scheme corresponding to the top-ranking element in the merit-ranking sequence as the optimal decomposition scheme comprises:

Let γ _k >γ _p , then compared with γ _p , γ _k is the element ranked higher in the superiority ranking sequence;

The difference coefficients i of the decomposition schemes corresponding to the elements in the superiority and inferiority ranking sequence are all equal. The decomposition scheme corresponding to γ _k and the decomposition scheme corresponding to γ _p in the superiority and inferiority ranking sequence are selected. When 0≤i≤1 and c _k b _p -c _p b _k ≤0, i must satisfy i∈[0,1]. If so, the ranking positions of γ _k and γ _p in the superiority and inferiority ranking sequence remain unchanged. If not, the ranking positions of γ _k and γ _p in the superiority and inferiority ranking sequence are swapped.

若满足，则所述优劣性排序序列中γ_k和γ_p的排序位置不变，若不满足，则所述优劣性排序序列中γ_k和γ_p的排序位置互换；When 0≤i≤1 and c _k b _p -c _p b _k ＞0, i must satisfy

If the condition is satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged; if the condition is not satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped;

When -1≤i＜0 and a _k b _p -a _p b _k ≥0, i must satisfy i∈[-1,0). If so, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged. If not, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped.

若满足，则所述优劣性排序序列中γ_k和γ_p的排序位置不变，若不满足，则所述优劣性排序序列中γ_k和γ_p的排序位置互换；When -1≤i＜0 and a _k b _p -a _p b _k ＜0, i must satisfy

If the condition is satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged; if the condition is not satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped;

In the above process, a _k is the identity between the kth decomposition scheme and the optimal normalized index value set U, b _k is the difference between the kth decomposition scheme and the optimal normalized index value set U, c _k is the opposition between the kth decomposition scheme and the optimal normalized index value set U, a _p is the identity between the pth decomposition scheme and the optimal normalized index value set U, b _p is the difference between the pth decomposition scheme and the optimal normalized index value set U, and c _p is the opposition between the pth decomposition scheme and the optimal normalized index value set U.

A device for optimizing a complex active power distribution network decomposition scheme, wherein the device comprises:

An acquisition module, used to acquire a set of decomposition schemes of a complex active distribution network and an evaluation index corresponding to each decomposition scheme in the set of decomposition schemes;

A determination module, used to determine the superiority and inferiority ranking sequence of each decomposition scheme by using a set pair analysis method according to the decomposition scheme set and the evaluation index corresponding to each decomposition scheme in the decomposition scheme set;

The analysis module is used to perform stability analysis on the merit-ranking sequence, update the merit-ranking sequence and select the decomposition scheme corresponding to the top-ranking element in the merit-ranking sequence as the optimal decomposition scheme.

Preferably, the acquisition module includes:

The first determination unit determines the resource utilization index, parallel computing complexity index, parallel computing accuracy index, parallel computing acceleration ratio index, parallel computing efficiency index and parallel computing cost index corresponding to each decomposition scheme in the decomposition scheme set.

Furthermore, the resource utilization index y ₁ corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

In the above formula, ⁿ is the number of partitions in the network, _Pi2 is the computational scale of the ith partition in the network, i∈[1,n];

The parallel computing complexity index y ₂ corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

为网络中第i个分区的计算复杂度，l为协调级服务器数目，M为计算代价系数，i∈[1,n]，M∈[2,5]，n为网络的分区数目；In the above formula, O _min is the theoretical minimum parallel computing complexity,

is the computational complexity of the i-th partition in the network, l is the number of coordination-level servers, M is the computational cost coefficient, i∈[1,n], M∈[2,5], n is the number of partitions in the network;

The parallel computing accuracy index y ₃ corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

In the above formula, U _i ' is the voltage value of node i after network decomposition, U _i is the voltage value of node i before network decomposition, i∈[1,m], m is the total number of nodes in the network;

The parallel computing speedup ratio index S _P corresponding to each decomposition scheme in the decomposition scheme set is determined as follows:

S _P = _TS / _TP

In the above formula, _TS is the time required to solve the problem serially, and _TP is the time required to solve the problem in parallel;

The parallel computing efficiency index E corresponding to each decomposition scheme in the decomposition scheme set is determined as follows:

E＝ _SP /P

In the above formula, S _P is the parallel computing speedup ratio, and P is the number of servers required for parallel computing;

The parallel computing cost index C corresponding to each decomposition scheme in the decomposition scheme set is determined as follows:

C＝T _P *P

In the above formula, _TP is the time required to solve the problem in parallel, and P is the number of servers required for parallel computing.

Preferably, the determining module includes:

A conversion unit, used for converting non-profit indicators in the evaluation indicators corresponding to each decomposition scheme into profit indicators;

A standardization unit, used for performing dimensionless processing on the evaluation indicators corresponding to the decomposition schemes to obtain the normalized indicator values corresponding to the decomposition schemes;

A selection unit, used for selecting an optimal normalized index value set U=(u ₁ ,u ₂ ,... _un ) and a worst normalized index value set V=(v ₁ ,v ₂ ,...v _n ) from the normalized index values corresponding to the decomposition schemes, and constructing a comparison space [V,U] of the decomposition schemes;

A second determination unit is used to determine the relative closeness of each decomposition scheme to the optimal normalized index value set U in the comparison space [V, U] of each decomposition scheme;

The sorting unit is used to sort the decomposition schemes in descending order according to the relative closeness between the decomposition schemes and the optimal normalized index value set U, so as to obtain a ranking sequence of the advantages and disadvantages of the decomposition schemes.

Furthermore, the conversion unit includes:

为非收益型指标，则按下式将

转换为收益型指标：Assume that the index value of the k-th decomposition scheme with respect to the r-th index is

If it is a non-income indicator, then

Converted to profit-based indicators:

为

的收益型指标值，

为第k个分解方案关于第r个指标的最大指标值，k∈[1,m]，r∈[1,n]，m为分解方案总数，n为评价指标总数。In the above formula,

for

The income indicator value of

is the maximum index value of the k-th decomposition scheme with respect to the r-th index, k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indicators.

Furthermore, the specification unit includes:

The evaluation indicators corresponding to the decomposition schemes are dimensionlessly processed as follows:

为第k个分解方案关于第r个指标的指标值，k∈[1,m]，r∈[1,n]，m为分解方案总数，n为评价指标总数。In the above formula, d _kr is the normalized index value of the k-th decomposition scheme with respect to the r-th index,

is the index value of the kth decomposition scheme with respect to the rth index, k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indicators.

Furthermore, the second determining unit includes:

Suppose the decomposition scheme set S = {s ₁ ,s ₂ ,...,s _m }, the evaluation index set E = {e ₁ ,e ₂ ,...,e _n }, let the normalized index value of the k-th decomposition scheme with respect to the r-th index be d _kr , the optimal normalized index value among the normalized index values of each decomposition scheme with respect to the r-th index be _ur , and the worst normalized index value among the normalized index values of each decomposition scheme with respect to the r-th index be v _r , where k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indexes;

In the comparison space [V, U] of the decomposition schemes, the connection degree between the kth decomposition scheme and the optimal normalized index value set U, that is, the connection degree u{s _k ,U} of the set pair {s _k ,U} is determined as follows:

u{s _k ,U}＝ _ak + _bki + _ckj

In the above formula, s _k is the kth decomposition scheme in the decomposition scheme set, a _k is the identity between the kth decomposition scheme and the optimal normalized index value set U, b _k is the difference between the kth decomposition scheme and the optimal normalized index value set U, c _k is the opposition between the kth decomposition scheme and the optimal normalized index value set U, i is the difference coefficient, and j is the opposition mark symbol;

Among them, i∈[-1,1], j＝1,

The relative closeness γ _k between the k-th decomposition scheme and the optimal normalized index value set U is determined as follows:

Preferably, the analysis module comprises:

Let γ _k >γ _p , then compared with γ _p , γ _k is the element ranked higher in the superiority and inferiority ranking sequence;

The difference coefficients i of the decomposition schemes corresponding to the elements in the superiority and inferiority ranking sequence are all equal, and the decomposition scheme corresponding to γ _k and the decomposition scheme corresponding to γ _p in the superiority and inferiority ranking sequence are selected;

A first judgment unit is used for, when 0≤i≤1 and c _k b _p -c _p b _k ≤0, i needs to satisfy i∈[0,1], if so, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged, if not, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped;

若满足，则所述优劣性排序序列中γ_k和γ_p的排序位置不变，若不满足，则所述优劣性排序序列中γ_k和γ_p的排序位置互换；The second judgment unit is used for, when 0≤i≤1 and c _k b _p -c _p b _k >0, i must satisfy

If the condition is satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged; if the condition is not satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped;

A third judgment unit is used for, when -1≤i＜0 and a _k b _p -a _p b _k ≥0, i needs to satisfy i∈[-1,0), if so, the ranking positions of γ _k and γ _p in the superiority and inferiority ranking sequence remain unchanged, if not, the ranking positions of γ _k and γ _p in the superiority and inferiority ranking sequence are swapped;

若满足，则所述优劣性排序序列中γ_k和γ_p的排序位置不变，若不满足，则所述优劣性排序序列中γ_k和γ_p的排序位置互换；The fourth judgment unit is used for, when -1≤i＜0 and a _k b _p -a _p b _k ＜0, i must satisfy

If the condition is satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged; if the condition is not satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped;

In the above process, a _k is the identity between the kth decomposition scheme and the optimal normalized index value set U, b _k is the difference between the kth decomposition scheme and the optimal normalized index value set U, c _k is the opposition between the kth decomposition scheme and the optimal normalized index value set U, a _p is the identity between the pth decomposition scheme and the optimal normalized index value set U, b _p is the difference between the pth decomposition scheme and the optimal normalized index value set U, and c _p is the opposition between the pth decomposition scheme and the optimal normalized index value set U.

Beneficial effects of the present invention:

The technical solution provided by the present invention adopts set pair analysis to make optimal decisions on complex active distribution network decomposition schemes, and uses a method of evaluating the pros and cons of schemes using relative closeness, which has clear concepts, simple calculations, and is easy to implement through programming; a complex active distribution network decomposition decision scheme with better credibility is obtained through set pair analysis, which meets the actual needs of achieving optimal comprehensive factors such as resource utilization, parallel computing complexity, and parallel computing accuracy; while the pros and cons of the schemes are evaluated under relatively certain conditions, the stability of the sorting results is analyzed using relative uncertainty information, i stable regions are given, and other sorting results are found, so that a relatively stable sorting can be distinguished from an unstable sorting.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a flow chart of a method for optimizing a complex active distribution network decomposition scheme according to the present invention;

FIG. 2 is a schematic diagram of the structure of a complex active distribution network decomposition scheme optimization device according to the present invention.

DETAILED DESCRIPTION

The specific implementation modes of the present invention are described in detail below with reference to the accompanying drawings.

In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

A method for optimizing a complex active distribution network decomposition scheme provided by the present invention, as shown in FIG1 , includes:

101. Obtain a set of decomposition schemes for a complex active distribution network and an evaluation index corresponding to each decomposition scheme in the set of decomposition schemes;

102. According to the decomposition scheme set and the evaluation index corresponding to each decomposition scheme in the decomposition scheme set, a set pair analysis method is used to determine the superiority and inferiority ranking sequence of each decomposition scheme;

103. Perform stability analysis on the merit-ranked sequence, update the merit-ranked sequence and select the decomposition scheme corresponding to the top-ranked element in the merit-ranked sequence as the optimal decomposition scheme.

Specifically, step 101 includes:

The resource utilization index, parallel computing complexity index, parallel computing accuracy index, parallel computing speedup ratio index, parallel computing efficiency index and parallel computing cost index corresponding to each decomposition scheme in the decomposition scheme set are determined respectively.

The resource utilization index y ₁ corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

为网络中第i个分区的计算规模，i∈[1,n]；In the above formula, n is the number of partitions in the network.

is the computational scale of the i-th partition in the network, i∈[1,n];

Resource utilization represents the overall utilization efficiency of the server. After each partition has completed its own calculation, the coordination server needs to use the calculation data of each sub-partition server to calculate the coordination amount. If the calculation scale of each sub-partition is too different, the partitions will wait for each other and cause a waste of resources. The closer y ₁ is to 1, the closer the calculation scale of each partition is, and the higher the resource utilization efficiency is. This indicator is a profit-oriented indicator.

The steps of parallel computing can be briefly described as three steps: first, each sub-server performs calculations in its own area and sends the data results of the coordination quantity to the coordination server, then the coordination server calculates the correction value of each coordination quantity through the data transmitted by each sub-server, and finally sends the data of the coordination quantity value to each sub-server, and each sub-server performs parallel computing on each sub-partition. The calculation in the general power system needs to be iterated multiple times, and the above steps only need to be repeated until the required accuracy is met. The parallel computing complexity index y ₂ corresponding to each decomposition scheme in the decomposition scheme set is determined as follows:

为网络中第i个分区的计算复杂度，l为协调级服务器数目，M为计算代价系数，i∈[1,n]，M∈[2,5]，n为网络的分区数目；In the above formula, O _min is the theoretical minimum parallel computing complexity,

is the computational complexity of the i-th partition in the network, l is the number of coordination-level servers, M is the computational cost coefficient, i∈[1,n], M∈[2,5], n is the number of partitions in the network;

The main resources of the coordination-level server are used to communicate with each sub-partition. The computing scale should be much smaller than that of each sub-partition computing server. Here, the second term of the denominator in formula (4) is multiplied by a coefficient M (2-5) to indicate that the computing cost of the coordination-level server is higher than that of each sub-server. The smaller the value of formula (4), the larger the parallel computing scale of the partitioning method used. When y ₂ = 1, the parallel computing complexity reaches the theoretical minimum parallel computing complexity, and the parallel computing can obtain the maximum speedup ratio. This indicator is a benefit-type indicator.

The voltage value of each node in the network is calculated by the known network, so as to obtain the voltage level of the research system. The voltage level is used to compare the voltage changes of each node before and after the network decomposition, and determine the impact of parallel calculation after partitioning on the overall calculation accuracy. The parallel calculation accuracy index y ₃ corresponding to each decomposition scheme in the decomposition scheme set is determined as follows:

In the above formula, U _i ' is the voltage value of node i after network decomposition, U _i is the voltage value of node i before network decomposition, i∈[1,m], m is the total number of nodes in the network;

The speedup ratio reflects the degree to which the parallel system's parallel capabilities are exerted when the parallel system runs a parallel program. It is related to the characteristics of the hardware, software, and application. The parallel computing speedup ratio index S _P corresponding to each decomposition scheme in the decomposition scheme set is determined as follows:

S _P = _TS / _TP

In the above formula, _TS is the time required to solve the problem serially, and _TP is the time required to solve the problem in parallel;

The parallel computing efficiency index E corresponding to each decomposition scheme in the decomposition scheme set is determined as follows:

E＝ _SP /P

In the above formula, S _P is the parallel computing speedup ratio, and P is the number of servers required for parallel computing;

The parallel computing cost index C corresponding to each decomposition scheme in the decomposition scheme set is determined as follows:

C＝T _P *P

In the above formula, _TP is the time required to solve the problem in parallel, and P is the number of servers required for parallel computing.

After obtaining a decomposition scheme set and an evaluation index corresponding to each decomposition scheme in the decomposition scheme set, step 102 includes:

Convert the non-profit indicators in the evaluation indicators corresponding to each decomposition scheme into profit indicators;

Performing dimensionless processing on the evaluation indicators corresponding to the decomposition schemes to obtain the normalized indicator values corresponding to the decomposition schemes;

Selecting an optimal normalized index value set U=(u ₁ ,u ₂ ,... _un ) and a worst normalized index value set V=(v ₁ ,v ₂ ,...v _n ) from the normalized index values corresponding to the decomposition schemes, and constructing a comparison space [V,U] of the decomposition schemes;

Among them, in order to conduct analysis within the same scope and determine the optimal and worst solutions, both can be generated from within the solution set or from outside the solution set, and can be determined based on system goals and objective conditions.

In the comparison space [V, U] of the decomposition schemes, determining the relative closeness of the decomposition schemes to the optimal normalized index value set U;

The decomposition schemes are sorted in descending order according to the relative closeness between the decomposition schemes and the optimal normalized index value set U, so as to obtain a ranking sequence of the advantages and disadvantages of the decomposition schemes.

Specifically, the non-profit indicators in the evaluation indicators corresponding to each decomposition scheme are converted into profit indicators, including:

为非收益型指标，则按下式将

转换为收益型指标：Assume that the index value of the k-th decomposition scheme with respect to the r-th index is

If it is a non-income indicator, then

Converted to profit-based indicators:

为

的收益型指标值，

为第k个分解方案关于第r个指标的最大指标值，k∈[1,m]，r∈[1,n]，m为分解方案总数，n为评价指标总数。In the above formula,

for

The income indicator value of

is the maximum index value of the k-th decomposition scheme with respect to the r-th index, k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indicators.

The step of performing dimensionless processing on the evaluation indicators corresponding to the decomposition schemes to obtain the normalized indicator values corresponding to the decomposition schemes includes:

The evaluation indicators corresponding to the decomposition schemes are dimensionlessly processed as follows:

为第k个分解方案关于第r个指标的指标值，k∈[1,m]，r∈[1,n]，m为分解方案总数，n为评价指标总数。In the above formula, d _kr is the normalized index value of the k-th decomposition scheme with respect to the r-th index,

is the index value of the kth decomposition scheme with respect to the rth index, k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indicators.

Determining the relative closeness between each decomposition scheme and the optimal normalized index value set U in the comparison space [V, U] of each decomposition scheme includes:

Suppose the decomposition scheme set S = {s ₁ ,s ₂ ,...,s _m }, the evaluation index set E = {e ₁ ,e ₂ ,...,e _n }, let the normalized index value of the k-th decomposition scheme with respect to the r-th index be d _kr , the optimal normalized index value among the normalized index values of each decomposition scheme with respect to the r-th index be _ur , and the worst normalized index value among the normalized index values of each decomposition scheme with respect to the r-th index be v _r , where k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indexes;

In the comparison space [V, U] of the decomposition schemes, the connection degree between the kth decomposition scheme and the optimal normalized index value set U, that is, the connection degree u{s _k ,U} of the set pair {s _k ,U} is determined as follows:

u{s _k ,U}＝ _ak + _bki + _ckj

In the above formula, s _k is the kth decomposition scheme in the decomposition scheme set, a _k is the identity between the kth decomposition scheme and the optimal normalized index value set U, b _k is the difference between the kth decomposition scheme and the optimal normalized index value set U, c _k is the opposition between the kth decomposition scheme and the optimal normalized index value set U, i is the difference coefficient, and j is the opposition mark symbol;

Among them, i∈[-1,1], j＝1,

The relative closeness γ _k between the k-th decomposition scheme and the optimal normalized index value set U is determined as follows:

Before determining the relative closeness between each decomposition scheme and the optimal normalized index value set U, the present invention needs to first determine the relative closeness between each decomposition scheme and a single optimal index in the optimal normalized index value set, including:

Let the evaluated scheme be s _k =(d _k1 ,d _k2 ,...,d _kn )(k=1,2,...,m), and determine the connection degree of the set pair {d _kr , _ur } in the comparison interval [v _r , _ur ] of _er .

可表示d_kr和u_r的接近程度；

可表示d_kr和v_r的接近程度。

It can indicate the closeness between d _kr and u _r ;

It can indicate how close d _kr and v _r are.

的数值：当

时取最小值

当d_kr＝v_r或u_r时取最大值

When d _kr ∈ [v _r , u _r ]

The value of:

Take the minimum value when

When d _kr = v _r or u _r, it takes the maximum value

进行归一化，即用

除

和

分别得到

二者可视为对d_kr与u_r接近程度的肯定和否定，可将它们分别定义为集对{d_kr,u_r}的同一度和对立度。To make

Normalize, that is,

remove

and

Get

The two can be regarded as the affirmation and negation of the closeness between d _kr and _ur , and they can be defined as the identity and opposition of the set pair {d _kr , _ur } respectively.

According to a+b+c=1, the difference between the set pair {d _kr , _ur } is calculated as:

Therefore, the connection degree of {d _kr , _ur } is:

时，差异度最大为

From the above formula, we can see that when d _kr = _ur or v _r , the difference is minimum and equals to zero; when

When , the maximum difference is

In the decision-making method of complex active distribution network decomposition scheme in this paper, the relative closeness γ _k of relatively stable a _k and c _k in μ _k = a _k + b _k i + c _k j is used to evaluate the pros and cons of the scheme. However, b _k is relatively uncertain, and its value indicates the size of uncertainty, and the sign and value of i can be regarded as the correction direction and degree of b _k to a _k or c _k , which will affect the evaluation results of the scheme. Therefore, it is necessary to analyze the ranking stability of the evaluation results of the complex active distribution network decomposition scheme, and try to find other ranking results except the basic ranking, that is, to obtain the extended order.

The analytical formula μ _k = a _k + b _k i + c _k j shows that when i>0, it is a positive correction to a _k , indicating a positive attitude towards _Sk approaching the ideal optimal solution U, and the closer i is to 1, the stronger this correction effect is. On the contrary, when i<0, it is a positive correction to c _k , indicating a negative attitude towards _Sk approaching U, and the closer i is to -1, the stronger this correction effect is. Therefore, regarding the change of i in [-1, 1], the stability analysis of the solution ranking can be performed. Therefore, the step 103 includes:

Let γ _k >γ _p , then compared with γ _p , γ _k is the element ranked higher in the superiority and inferiority ranking sequence;

The difference coefficients i of the decomposition schemes corresponding to the elements in the superiority and inferiority ranking sequence are all equal. The decomposition scheme corresponding to γ _k and the decomposition scheme corresponding to γ _p in the superiority and inferiority ranking sequence are selected. When 0≤i≤1 and c _k b _p -c _p b _k ≤0, i must satisfy i∈[0,1]. If so, the ranking positions of γ _k and γ _p in the superiority and inferiority ranking sequence remain unchanged. If not, the ranking positions of γ _k and γ _p in the superiority and inferiority ranking sequence are swapped.

若满足，则所述优劣性排序序列中γ_k和γ_p的排序位置不变，若不满足，则所述优劣性排序序列中γ_k和γ_p的排序位置互换；When 0≤i≤1 and c _k b _p -c _p b _k ＞0, i must satisfy

If the condition is satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged; if the condition is not satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped;

When -1≤i＜0 and a _k b _p -a _p b _k ≥0, i must satisfy i∈[-1,0). If so, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged. If not, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped.

若满足，则所述优劣性排序序列中γ_k和γ_p的排序位置不变，若不满足，则所述优劣性排序序列中γ_k和γ_p的排序位置互换；When -1≤i＜0 and a _k b _p -a _p b _k ＜0, i must satisfy

If the condition is satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged; if the condition is not satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped;

In the above process, a _k is the identity between the kth decomposition scheme and the optimal normalized index value set U, b _k is the difference between the kth decomposition scheme and the optimal normalized index value set U, c _k is the opposition between the kth decomposition scheme and the optimal normalized index value set U, a _p is the identity between the pth decomposition scheme and the optimal normalized index value set U, b _p is the difference between the pth decomposition scheme and the optimal normalized index value set U, and c _p is the opposition between the pth decomposition scheme and the optimal normalized index value set U.

The present invention also provides a complex active distribution network decomposition scheme optimization device, as shown in FIG2, the device comprises:

An acquisition module, used to acquire a set of decomposition schemes of a complex active distribution network and an evaluation index corresponding to each decomposition scheme in the set of decomposition schemes;

A determination module, used to determine the superiority and inferiority ranking sequence of each decomposition scheme by using a set pair analysis method according to the decomposition scheme set and the evaluation index corresponding to each decomposition scheme in the decomposition scheme set;

The analysis module is used to perform stability analysis on the merit-ranking sequence, update the merit-ranking sequence and select the decomposition scheme corresponding to the top-ranking element in the merit-ranking sequence as the optimal decomposition scheme.

The acquisition module comprises:

The first determination unit determines the resource utilization index, parallel computing complexity index, parallel computing accuracy index, parallel computing acceleration ratio index, parallel computing efficiency index and parallel computing cost index corresponding to each decomposition scheme in the decomposition scheme set.

The resource utilization index y ₁ corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

为网络中第i个分区的计算规模，i∈[1,n]；In the above formula, n is the number of partitions in the network.

is the computational scale of the i-th partition in the network, i∈[1,n];

The parallel computing complexity index y ₂ corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

为网络中第i个分区的计算复杂度，l为协调级服务器数目，M为计算代价系数，i∈[1,n]，M∈[2,5]，n为网络的分区数目；In the above formula, O _min is the theoretical minimum parallel computing complexity,

is the computational complexity of the i-th partition in the network, l is the number of coordination-level servers, M is the computational cost coefficient, i∈[1,n], M∈[2,5], n is the number of partitions in the network;

The parallel computing accuracy index y ₃ corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

In the above formula, U _i ' is the voltage value of node i after network decomposition, U _i is the voltage value of node i before network decomposition, i∈[1,m], m is the total number of nodes in the network;

The parallel computing speedup ratio index S _P corresponding to each decomposition scheme in the decomposition scheme set is determined as follows:

S _P = _TS / _TP

In the above formula, _TS is the time required to solve the problem serially, and _TP is the time required to solve the problem in parallel;

The parallel computing efficiency index E corresponding to each decomposition scheme in the decomposition scheme set is determined as follows:

E＝ _SP /P

In the above formula, S _P is the parallel computing speedup ratio, and P is the number of servers required for parallel computing;

The parallel computing cost index C corresponding to each decomposition scheme in the decomposition scheme set is determined as follows:

C＝T _P *P

In the above formula, _TP is the time required to solve the problem in parallel, and P is the number of servers required for parallel computing.

The determining module comprises:

A conversion unit, used for converting non-profit indicators in the evaluation indicators corresponding to each decomposition scheme into profit indicators;

A standardization unit, used for performing dimensionless processing on the evaluation indicators corresponding to the decomposition schemes to obtain the normalized indicator values corresponding to the decomposition schemes;

A selection unit, used for selecting an optimal normalized index value set U=(u ₁ ,u ₂ ,... _un ) and a worst normalized index value set V=(v ₁ ,v ₂ ,...v _n ) from the normalized index values corresponding to the decomposition schemes, and constructing a comparison space [V,U] of the decomposition schemes;

A second determination unit is used to determine the relative closeness of each decomposition scheme to the optimal normalized index value set U in the comparison space [V, U] of each decomposition scheme;

The sorting unit is used to sort the decomposition schemes in descending order according to the relative closeness between the decomposition schemes and the optimal normalized index value set U, so as to obtain a ranking sequence of the advantages and disadvantages of the decomposition schemes.

The conversion unit comprises:

为非收益型指标，则按下式将

转换为收益型指标：Assume that the index value of the k-th decomposition scheme with respect to the r-th index is

If it is a non-income indicator, then

Converted to profit-based indicators:

为

的收益型指标值，

为第k个分解方案关于第r个指标的最大指标值，k∈[1,m]，r∈[1,n]，m为分解方案总数，n为评价指标总数。In the above formula,

for

The income indicator value of

is the maximum index value of the k-th decomposition scheme with respect to the r-th index, k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indicators.

The specification unit includes:

The evaluation indicators corresponding to the decomposition schemes are dimensionlessly processed as follows:

为第k个分解方案关于第r个指标的指标值，k∈[1,m]，r∈[1,n]，m为分解方案总数，n为评价指标总数。In the above formula, d _kr is the normalized index value of the k-th decomposition scheme with respect to the r-th index,

is the index value of the kth decomposition scheme with respect to the rth index, k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indicators.

The second determining unit includes:

Suppose the decomposition scheme set S = {s ₁ ,s ₂ ,...,s _m }, the evaluation index set E = {e ₁ ,e ₂ ,...,e _n }, let the normalized index value of the k-th decomposition scheme with respect to the r-th index be d _kr , the optimal normalized index value among the normalized index values of each decomposition scheme with respect to the r-th index be _ur , and the worst normalized index value among the normalized index values of each decomposition scheme with respect to the r-th index be v _r , where k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indexes;

In the comparison space [V, U] of the decomposition schemes, the connection degree between the kth decomposition scheme and the optimal normalized index value set U, that is, the connection degree u{s _k ,U} of the set pair {s _k ,U} is determined as follows:

u{s _k ,U}＝ _ak + _bki + _ckj

In the above formula, s _k is the kth decomposition scheme in the decomposition scheme set, a _k is the identity between the kth decomposition scheme and the optimal normalized index value set U, b _k is the difference between the kth decomposition scheme and the optimal normalized index value set U, c _k is the opposition between the kth decomposition scheme and the optimal normalized index value set U, i is the difference coefficient, and j is the opposition mark symbol;

Among them, i∈[-1,1], j＝1,

The relative closeness γ _k between the k-th decomposition scheme and the optimal normalized index value set U is determined as follows:

The analysis module comprises:

Let γ _k >γ _p , then compared with γ _p , γ _k is the element ranked higher in the superiority and inferiority ranking sequence;

The difference coefficients i of the decomposition schemes corresponding to the elements in the superiority and inferiority ranking sequence are all equal, and the decomposition scheme corresponding to γ _k and the decomposition scheme corresponding to γ _p in the superiority and inferiority ranking sequence are selected;

A first judgment unit is used for, when 0≤i≤1 and c _k b _p -c _p b _k ≤0, i needs to satisfy i∈[0,1], if so, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged, if not, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped;

若满足，则所述优劣性排序序列中γ_k和γ_p的排序位置不变，若不满足，则所述优劣性排序序列中γ_k和γ_p的排序位置互换；The second judgment unit is used for when 0≤i≤1 and c _k b _p -c _p b _k >0, i must satisfy

If the condition is satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged; if the condition is not satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped;

A third judgment unit is used for, when -1≤i＜0 and a _k b _p -a _p b _k ≥0, i needs to satisfy i∈[-1,0), if so, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged, if not, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped;

若满足，则所述优劣性排序序列中γ_k和γ_p的排序位置不变，若不满足，则所述优劣性排序序列中γ_k和γ_p的排序位置互换；The fourth judgment unit is used for, when -1≤i＜0 and a _k b _p -a _p b _k ＜0, i must satisfy

If the condition is satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged; if the condition is not satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped;

In the above process, a _k is the identity between the kth decomposition scheme and the optimal normalized index value set U, b _k is the difference between the kth decomposition scheme and the optimal normalized index value set U, c _k is the opposition between the kth decomposition scheme and the optimal normalized index value set U, a _p is the identity between the pth decomposition scheme and the optimal normalized index value set U, b _p is the difference between the pth decomposition scheme and the optimal normalized index value set U, and c _p is the opposition between the pth decomposition scheme and the optimal normalized index value set U.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims (14)

1. A method for optimizing a complex active distribution network decomposition scheme, characterized in that the method comprises: Obtaining a set of decomposition schemes of a complex active distribution network and an evaluation index corresponding to each decomposition scheme in the set of decomposition schemes; According to the decomposition scheme set and the evaluation index corresponding to each decomposition scheme in the decomposition scheme set, a set pair analysis method is used to determine the superiority and inferiority ranking sequence of each decomposition scheme; Performing stability analysis on the merit-ranking sequence, updating the merit-ranking sequence and selecting the decomposition scheme corresponding to the first-ranked element in the merit-ranking sequence as the optimal decomposition scheme; The step of performing stability analysis on the sequence of merit ranking, updating the sequence of merit ranking, and selecting the decomposition scheme corresponding to the first-ranked element in the sequence of merit ranking as the optimal decomposition scheme includes: Let γ _k >γ _p , then compared with γ _p , γ _k is the element ranked higher in the superiority and inferiority ranking sequence; The difference coefficients i of the decomposition schemes corresponding to the elements in the superiority and inferiority ranking sequence are all equal. The decomposition scheme corresponding to γ _k and the decomposition scheme corresponding to γ _p in the superiority and inferiority ranking sequence are selected. When 0≤i≤1 and c _k b _p -c _p b _k ≤0, i must satisfy i∈[0,1]. If so, the ranking positions of γ _k and γ _p in the superiority and inferiority ranking sequence remain unchanged. If not, the ranking positions of γ _k and γ _p in the superiority and inferiority ranking sequence are swapped. When 0≤i≤1 and c _k b _p -c _p b _k >0, i must satisfy

If the condition is satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged; if the condition is not satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped; When -1≤i<0 and a _k b _p -a _p b _k ≥0, i must satisfy i∈[-1,0). If so, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged. If not, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped. When -1≤i<0 and a _k b _p -a _p b _k <0, i must satisfy

If the condition is satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged; if the condition is not satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped; In the above process, a _k is the identity between the kth decomposition scheme and the optimal normalized index value set U, b _k is the difference between the kth decomposition scheme and the optimal normalized index value set U, c _k is the opposition between the kth decomposition scheme and the optimal normalized index value set U, a _p is the identity between the pth decomposition scheme and the optimal normalized index value set U, b _p is the difference between the pth decomposition scheme and the optimal normalized index value set U, and c _p is the opposition between the pth decomposition scheme and the optimal normalized index value set U. 2. The method according to claim 1, wherein the step of obtaining a set of decomposition schemes of a complex active distribution network and an evaluation index corresponding to each decomposition scheme in the set of decomposition schemes comprises: The resource utilization index, parallel computing complexity index, parallel computing accuracy index, parallel computing speedup ratio index, parallel computing efficiency index and parallel computing cost index corresponding to each decomposition scheme in the decomposition scheme set are determined respectively. 3. The method according to claim 2, characterized in that the resource utilization index y ₁ corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

In the above formula, ⁿ is the number of partitions in the network, _Pi2 is the computational scale of the ith partition in the network, i∈[1,n]; The parallel computing complexity index y ₂ corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

In the above formula, O _min is the theoretical minimum parallel computing complexity, _Pi ² is the computing complexity of the i-th partition in the network, l is the number of coordination-level servers, M is the computing cost coefficient, i∈[1,n], M∈[2,5], and n is the number of partitions in the network; The parallel computing accuracy index y ₃ corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

In the above formula, _U'i is the voltage value of node i after network decomposition, _Ui is the voltage value of node i before network decomposition, i∈[1,m], m is the total number of nodes in the network; The parallel computing speedup ratio index S _P corresponding to each decomposition scheme in the decomposition scheme set is determined as follows: S _P = _TS / _TP In the above formula, _TS is the time required to solve the problem serially, and _TP is the time required to solve the problem in parallel; The parallel computing efficiency index E corresponding to each decomposition scheme in the decomposition scheme set is determined as follows: E＝ _SP /P In the above formula, S _P is the parallel computing speedup ratio, and P is the number of servers required for parallel computing; The parallel computing cost index C corresponding to each decomposition scheme in the decomposition scheme set is determined as follows: C＝T _P *P In the above formula, _TP is the time required to solve the problem in parallel, and P is the number of servers required for parallel computing. 4. The method according to claim 1, characterized in that the step of determining the superiority and inferiority ranking sequence of each decomposition scheme by using a set pair analysis method according to the decomposition scheme set and the evaluation index corresponding to each decomposition scheme in the decomposition scheme set comprises: Convert the non-profit indicators in the evaluation indicators corresponding to each decomposition scheme into profit indicators; Performing dimensionless processing on the evaluation indicators corresponding to the decomposition schemes to obtain the normalized indicator values corresponding to the decomposition schemes; Selecting an optimal normalized index value set U=(u ₁ ,u ₂ ,... _un ) and a worst normalized index value set V=(v ₁ ,v ₂ ,...v _n ) from the normalized index values corresponding to the decomposition schemes, and constructing a comparison space [V,U] of the decomposition schemes; In the comparison space [V, U] of the decomposition schemes, determining the relative closeness of the decomposition schemes to the optimal normalized index value set U; The decomposition schemes are sorted in descending order according to the relative closeness between the decomposition schemes and the optimal normalized index value set U, so as to obtain a ranking sequence of the advantages and disadvantages of the decomposition schemes. 5. The method according to claim 4, characterized in that the step of converting the non-profit indicators in the evaluation indicators corresponding to each decomposition scheme into profit indicators comprises: Assume that the index value of the k-th decomposition scheme with respect to the r-th index is

If it is a non-income indicator, then

Converted to profit-based indicators:

In the above formula,

for

The income indicator value of

is the maximum index value of the k-th decomposition scheme with respect to the r-th index, k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indicators. 6. The method according to claim 4, characterized in that the step of performing dimensionless processing on the evaluation indicators corresponding to the decomposition schemes to obtain the normalized indicator values corresponding to the decomposition schemes comprises: The evaluation indicators corresponding to the decomposition schemes are dimensionlessly processed as follows:

In the above formula, d _kr is the normalized index value of the k-th decomposition scheme with respect to the r-th index,

is the index value of the kth decomposition scheme with respect to the rth index, k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indicators. 7. The method according to claim 4, characterized in that the step of determining the relative closeness of each decomposition scheme to the optimal normalized index value set U in the comparison space [V, U] of each decomposition scheme comprises: Suppose the decomposition scheme set S = {s ₁ ,s ₂ ,...,s _m }, the evaluation index set E = {e ₁ ,e ₂ ,...,e _n }, let the normalized index value of the k-th decomposition scheme with respect to the r-th index be d _kr , the optimal normalized index value among the normalized index values of each decomposition scheme with respect to the r-th index be _ur , and the worst normalized index value among the normalized index values of each decomposition scheme with respect to the r-th index be v _r , where k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indexes; In the comparison space [V, U] of the decomposition schemes, the connection degree between the kth decomposition scheme and the optimal normalized index value set U, that is, the connection degree u{s _k ,U} of the set pair {s _k ,U} is determined according to the following formula: u{s _k ,U}＝ _ak + _bki + _ckj In the above formula, s _k is the kth decomposition scheme in the decomposition scheme set, a _k is the identity between the kth decomposition scheme and the optimal normalized index value set U, b _k is the difference between the kth decomposition scheme and the optimal normalized index value set U, c _k is the opposition between the kth decomposition scheme and the optimal normalized index value set U, i is the difference coefficient, and j is the opposition mark symbol; Among them, i∈[-1,1], j＝1,

The relative closeness γ _k between the k-th decomposition scheme and the optimal normalized index value set U is determined as follows:

8. A device for optimizing a complex active distribution network decomposition scheme, characterized in that the device comprises: An acquisition module, used to acquire a set of decomposition schemes of a complex active distribution network and an evaluation index corresponding to each decomposition scheme in the set of decomposition schemes; A determination module, used to determine the superiority and inferiority ranking sequence of each decomposition scheme by using a set pair analysis method according to the decomposition scheme set and the evaluation index corresponding to each decomposition scheme in the decomposition scheme set; An analysis module, used for performing stability analysis on the merit-ranking sequence, updating the merit-ranking sequence and selecting the decomposition scheme corresponding to the top-ranking element in the merit-ranking sequence as the optimal decomposition scheme; The analysis module comprises: Let γ _k >γ _p , then compared with γ _p , γ _k is the element ranked higher in the superiority ranking sequence; The difference coefficients i of the decomposition schemes corresponding to the elements in the superiority and inferiority ranking sequence are all equal, and the decomposition scheme corresponding to γ _k and the decomposition scheme corresponding to γ _p in the superiority and inferiority ranking sequence are selected; A first judgment unit is used for, when 0≤i≤1 and c _k b _p -c _p b _k ≤0, i needs to satisfy i∈[0,1], if so, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged, if not, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped; The second judgment unit is used for when 0≤i≤1 and c _k b _p -c _p b _k >0, i must satisfy

If the condition is satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged; if the condition is not satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped; A third judgment unit is used for, when -1≤i<0 and a _k b _p -a _p b _k ≥0, i needs to satisfy i∈[-1,0), if so, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged, if not, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped; The fourth judgment unit is used for, when -1≤i<0 and a _k b _p -a _p b _k <0, i must satisfy

If the condition is satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence remain unchanged; if the condition is not satisfied, the ranking positions of γ _k and γ _p in the superiority ranking sequence are swapped; In the above process, a _k is the identity between the kth decomposition scheme and the optimal normalized index value set U, b _k is the difference between the kth decomposition scheme and the optimal normalized index value set U, c _k is the opposition between the kth decomposition scheme and the optimal normalized index value set U, a _p is the identity between the pth decomposition scheme and the optimal normalized index value set U, b _p is the difference between the pth decomposition scheme and the optimal normalized index value set U, and c _p is the opposition between the pth decomposition scheme and the optimal normalized index value set U. 9. The device according to claim 8, wherein the acquisition module comprises: The first determination unit determines the resource utilization index, parallel computing complexity index, parallel computing accuracy index, parallel computing acceleration ratio index, parallel computing efficiency index and parallel computing cost index corresponding to each decomposition scheme in the decomposition scheme set. 10. The device according to claim 9, characterized in that the resource utilization index _y1 corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

In the above formula, ⁿ is the number of partitions in the network, _Pi2 is the computational scale of the ith partition in the network, i∈[1,n]; The parallel computing complexity index y ₂ corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

In the above formula, O _min is the theoretical minimum parallel computing complexity, _Pi ² is the computing complexity of the i-th partition in the network, l is the number of coordination-level servers, M is the computing cost coefficient, i∈[1,n], M∈[2,5], and n is the number of partitions in the network; The parallel computing accuracy index y ₃ corresponding to each decomposition scheme in the decomposition scheme set is determined according to the following formula:

In the above formula, U _i ' is the voltage value of node i after network decomposition, U _i is the voltage value of node i before network decomposition, i∈[1,m], m is the total number of nodes in the network; The parallel computing speedup ratio index S _P corresponding to each decomposition scheme in the decomposition scheme set is determined as follows: S _P = _TS / _TP In the above formula, _TS is the time required to solve the problem serially, and _TP is the time required to solve the problem in parallel; The parallel computing efficiency index E corresponding to each decomposition scheme in the decomposition scheme set is determined as follows: E＝ _SP /P In the above formula, S _P is the parallel computing speedup ratio, and P is the number of servers required for parallel computing; The parallel computing cost index C corresponding to each decomposition scheme in the decomposition scheme set is determined as follows: C＝T _P *P In the above formula, _TP is the time required to solve the problem in parallel, and P is the number of servers required for parallel computing. 11. The device according to claim 8, wherein the determining module comprises: A conversion unit, used for converting non-profit indicators in the evaluation indicators corresponding to each decomposition scheme into profit indicators; A standardization unit, used for performing dimensionless processing on the evaluation indicators corresponding to the decomposition schemes to obtain the normalized indicator values corresponding to the decomposition schemes; A selection unit, used for selecting an optimal normalized index value set U=(u ₁ ,u ₂ ,... _un ) and a worst normalized index value set V=(v ₁ ,v ₂ ,...v _n ) from the normalized index values corresponding to the decomposition schemes, and constructing a comparison space [V,U] of the decomposition schemes; A second determination unit is used to determine the relative closeness of each decomposition scheme to the optimal normalized index value set U in the comparison space [V, U] of each decomposition scheme; The sorting unit is used to sort the decomposition schemes in descending order according to the relative closeness between the decomposition schemes and the optimal normalized index value set U, so as to obtain a ranking sequence of the advantages and disadvantages of the decomposition schemes. 12. The device according to claim 11, wherein the conversion unit comprises: Assume that the index value of the k-th decomposition scheme with respect to the r-th index is

If it is a non-income indicator, then

Converted to profit-based indicators:

In the above formula,

for

The income indicator value of

is the maximum index value of the k-th decomposition scheme with respect to the r-th index, k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indicators. 13. The device according to claim 11, wherein the standardization unit comprises: The evaluation indicators corresponding to the decomposition schemes are dimensionlessly processed as follows:

In the above formula, d _kr is the normalized index value of the k-th decomposition scheme with respect to the r-th index,

is the index value of the kth decomposition scheme with respect to the rth index, k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indicators. 14. The device according to claim 11, wherein the second determining unit comprises: Suppose the decomposition scheme set S = {s ₁ ,s ₂ ,...,s _m }, the evaluation index set E = {e ₁ ,e ₂ ,...,e _n }, let the normalized index value of the k-th decomposition scheme with respect to the r-th index be d _kr , the optimal normalized index value among the normalized index values of each decomposition scheme with respect to the r-th index be _ur , and the worst normalized index value among the normalized index values of each decomposition scheme with respect to the r-th index be v _r , where k∈[1,m], r∈[1,n], m is the total number of decomposition schemes, and n is the total number of evaluation indexes; In the comparison space [V, U] of the decomposition schemes, the connection degree between the kth decomposition scheme and the optimal normalized index value set U, that is, the connection degree u{s _k ,U} of the set pair {s _k ,U} is determined as follows: u{s _k ,U}＝ _ak + _bki + _ckj In the above formula, s _k is the kth decomposition scheme in the decomposition scheme set, a _k is the identity between the kth decomposition scheme and the optimal normalized index value set U, b _k is the difference between the kth decomposition scheme and the optimal normalized index value set U, c _k is the opposition between the kth decomposition scheme and the optimal normalized index value set U, i is the difference coefficient, and j is the opposition mark symbol; Among them, i∈[-1,1], j＝1,

The relative closeness γ _k between the k-th decomposition scheme and the optimal normalized index value set U is determined as follows:

Images