CN104484724A - Extra-high voltage drop point plan optimal selection method based on cloud model - Google Patents

Abstract

The invention belongs to the field of electric power system planning, and particularly relates to an extra-high voltage drop point plan optimal selection method based on a cloud model. The method includes the steps that (1), an extra-high voltage drop point evaluation index system about safety, regional characteristics and perspective adaptability of drop points is set up; (2), an evaluation index weighting method based on a variance maximization model is set up; (3), the extra-high voltage drop point plan optimal selection comprehensive evaluation method based on the cloud model is set up. The requirement for the complete, comprehensive and effective evaluation index system for evaluating the extra-high voltage drop points is met, an effective tool is supplied to eliminate the influence on determination of evaluation index weights by using the single weighting method, and a reference basis is provided for adding the new extra-high voltage drop point plan optimal selection method and applying the cloud model to comprehensive evaluation of extra-high voltage drop point plan optimal selection.

Description

A kind of planning of the extra-high voltage drop point based on cloud model method for optimizing

Technical field

The invention belongs to Power System Planning field, especially a kind of extra-high voltage drop point planning method for optimizing set up based on cloud model.

Background technology

The planning of extra-high voltage drop point is indispensable part in extra-high voltage grid construction, and it is intended to strengthening electric network composition, improves power supply quality and power supply reliability.The optimal programming of drop point addressing is a complicated Multiple Attribute Decision Problems, sets up System of Comprehensive Evaluation scientifically and rationally, uses enabling legislation and evaluation method to be very important.

The method that index system is set up often divides two classes, one class is expert's subjective assessment method, another kind of is compare to determine method and data statistic analysis method, the former be applicable to data limited be evaluated object, Main Basis expert rich experience knowledge carrys out agriculture products, and what the latter was applicable to quantitative assessing index is evaluated object.When setting up extra-high voltage drop point optimizing evaluation index system, the subjectivity that expert judging method is stronger may cause the index system set up not necessarily comprehensively and systematically can reflect decision problem.

After evaluation index builds, usually carry out the tax power of evaluation index.The determination of weight is mainly divided into subjective weighting method, objective weighted model and Evaluation formula.Subjective weighting method can fully demonstrate the preference of expert, makes full use of the rich experiences of expert, but subjective randomness is comparatively large, even if increase expert's number, strictly select the measures such as expert and can not improve this problem at all.Objective weighted model has tax power objectivity, is not affected by human factors, but the index weights calculated relies on sample data, changes, can not embody the importance of each index self-value with the change of sample.

After determining evaluation index and optimal weights, suitable method is selected to carry out comprehensive evaluation.Conventional integrated evaluating method has levels analytic approach, fuzzy comprehensive evaluation method and grey Relational Analysis Method and combination evaluation methods.In system evaluation application, also have some problems in these methods, such as appraisal procedure has very large randomness on selecting, and perhaps diverse ways can draw different results; Appraisal procedure self also has a lot of shortcoming, such as the uncertainty of a grey Relational Analysis Method resolution system, the ambiguity of a fuzzy comprehensive evaluation method resolution system.But in extra-high voltage drop point preferred process, often there is uncertain and ambiguity simultaneously.

Summary of the invention

The present invention mainly solves the technical matters existing for prior art; Provide one and comprehensive evaluation value sequence is carried out to scheme, intuitively can also reflect a kind of planning of the extra-high voltage drop point based on cloud model method for optimizing of good and bad grade.

Above-mentioned technical matters of the present invention is mainly solved by following technical proposals:

Based on an extra-high voltage drop point planning method for optimizing for cloud model, it is characterized in that,

Step 1, sets up the extra-high voltage drop point assessment indicator system about security, drop point region characteristic and prospect zone; Specifically: analyze extra-high voltage drop point program evaluation index and build principle and object, decision-maker tentatively chooses certain evaluation index; Adopt chromatographic assays agriculture products weight, adopt the screening of flexible strategy determining method to fall some secondary indexs; Adopt validity coefficient to portray the validity of index system, when weighing evaluate decision problem by the evaluation index set or system, produce the departure degree of understanding; The reliability and stability of index system represent the difference degree of the evaluating data that the index system of design draws and ideal data; Finally choose the index system that validity and reliability is all higher; Comprise following sub-step:

Step 1.1, the preliminary foundation of assessment indicator system:

Analyze extra-high voltage drop point program evaluation index and build principle and object, tentatively choose certain evaluation index;

Step 1.2, the screening of evaluation index:

Adopt chromatographic assays agriculture products weight, adopt the screening of flexible strategy determining method to fall some secondary indexs;

Defining the assessment indicator system tentatively determined is H={h ₁, h ₂..., h _n.

Definition flexible strategy integrate as λ={ λ ₁, λ ₂..., λ _n, wherein λ _i∈ [0,1], (i=1,2 ..., n); If choice flexible strategy are λ _k, λ _k∈ [0,1], works as λ _i≤ λ _ktime, then index h is fallen in screening _i; Otherwise then retain this index; Usual choice flexible strategy λ _kit is 0.1 more suitable to get, and works as λ _ican think when≤0.1 that this Index Influence is less, be not enough to consider, decision-maker also can according to actual conditions, get get greatly little; After screening index, set up 2-3 initial indication system;

Step 1.3, the Effective judgement of index system:

Adopt validity coefficient to portray the validity of index system, when weighing evaluate decision problem by the evaluation index set or system, produce the departure degree of understanding; The absolute value of this value is less, and when showing to adopt this evaluation index or system appraisal decision problem, more trend is consistent for understanding, and validity is higher; Otherwise validity is lower;

Assessment indicator system after definition index screening is A={a ₁, a ₂..., a _n, participating in pricer number is m, and it is X that expert j assembles fruit to the scoring of evaluation object _j={ x _1j, x _2j..., x _nj, the validity coefficient of definition assessment indicator system A is β:

$\mathrm{\β}=\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\left(\frac{\underset{j=1}{\overset{m}{\mathrm{\Σ}}}|\stackrel{\‾}{x_i}-x_{\mathrm{ij}}|}{m*M}\right)}{n}---\left(1\right)$

Wherein, evaluation index a _ithe mean value of scoring, M is index a _icomment gathers in mark optimal value;

Step 1.4, index system reliability judges

The reliability and stability of index system represent the difference degree of the evaluating data that the index system of design draws and ideal data; According to mathematical statistics principle, adopt reliability coefficient ρ to portray this characteristic, ρ is larger, shows that difference is little, and the reliability of this index system is high; Otherwise its reliability is poor; Specifically by evaluation index a _ithe average of m evaluation result as ideal value, calculate the difference degree of m evaluating data and ideal value; The average data group of expert group's scoring is Y={y ₁, y ₂..., y _n, the reliability coefficient of assessment indicator system is ρ:

$\mathrm{\ρ}=\frac{\underset{j=1}{\overset{m}{\mathrm{\Σ}}}(\underset{i=1}{\overset{n}{\mathrm{\Σ}}}(x_{\mathrm{ij}}-\stackrel{\‾}{x_j})(y_i-\stackrel{\‾}{y})/\sqrt{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{(x_{\mathrm{ij}}-\stackrel{\‾}{x_j})}^2\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{(y_i-\stackrel{\‾}{y})}^2})}{m}---\left(2\right)$

Wherein, y _ievaluation index a _ithe mean value of scoring, the scoring average of expert j, for y _iaverage;

By above calculating, choose the index system that validity and reliability is all higher;

Step 2, sets up the evaluation index enabling legislation based on maximum variance model; Specifically: for set up extra-high voltage drop point program evaluation index system, AHP method is utilized to solve subjective weight according to an expert view, utilize entropy assessment to solve objective weight, combine with maximum variance model, try to achieve the optimal weights of extra-high voltage drop point evaluation index;

Step 2.1, the determination of subjective and objective weight:

Solve obtain subjective weight and objective weight respectively by the AHP that improves and entropy assessment, be different from traditional AHP, by analyzing each step of AHP and improving one's methods, have selected the higher scale of precision and complete weight and determine; Be different from traditional entropy assessment, fusion qualitative index and quantitative target form nondimensionalization matrix as the decision matrix solving entropy power; Finally solve and obtain subjective and objective weight;

Step 2.2, based on the Evaluation formula of maximum variance model

For a certain decision-making problem of multi-objective, decision scheme S _kat index a _ideviation that is lower and other all decision schemes may be defined as

${\mathrm{\σ}}_{\mathrm{ij}}\left(w\right)=\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{(r_{\mathrm{ij}}-r_{\mathrm{kj}})}^{2}w,i\∈n,j\∈m---\left(3\right)$

Based on this, structure departure function

$\mathrm{\σ}\left(w\right)=\underset{j=1}{\overset{m}{\mathrm{\Σ}}}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\σ}}_{\mathrm{ij}}\left(w\right)=\underset{j=1}{\overset{m}{\mathrm{\Σ}}}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{(r_{\mathrm{ij}}-r_{\mathrm{kj}})}^2{w}_j---\left(4\right)$

Thus solve weight phasor W to be equivalent to and to solve following linear programming problem:

$\left\{\begin{array}{c}\max \mathrm{\σ}\left(w\right)=\underset{j=1}{\overset{m}{\mathrm{\Σ}}}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{(r_{\mathrm{ij}}-r_{\mathrm{kj}})}^2{w}_j\\ s.t.w\∈\mathrm{\Φ}\end{array}\right.---\left(5\right)$

Wherein Weight of Expert and entropy are acted temporarily as into proportion range in this model;

Solve above-mentioned linear programming model and can obtain optimized index weight phasor W ^*=[W ^* ₁, W ^* ₂w ^* _n];

Step 3, sets up the preferred integrated evaluating method of extra-high voltage drop point based on cloud model; Specifically: after setting up assessment indicator system and enabling legislation, set up cloud model association attributes evaluation criteria level, descriptive grade cloud, utilize cloud model to carry out cloud normalization according to index attribute value and optimal weights and converge knot computing and complete the sequence of alternative drop point, then carry out drop point according to ordering scenario preferred;

Step 3.1, sets up mathematical description, comprises candidate scheme collection, index set, Zhuan Jiaji, attribute evaluation criteria collection and descriptive grade cloud;

Step 3.2, cloud normalization computing:

First, expert is formed matrix U to the result that qualitative index is marked _k', be an evaluation of estimate by multiple evaluation of estimate weighted mean normalizings of its each row, that is:

U _k′→Ex _i(6)

Merge quantitative and qualitative analysis index, and standardization obtains U _k"=(E _ki) _{y × 1}, according to U _k" and W ^*by formula (13), cloud normalization is carried out to the index of each scheme, obtain matrix

$u_{\mathrm{kd}}^{*}=\frac{\underset{i=1}{\overset{d}{\mathrm{\Σ}}}(E_{\mathrm{ki}}\×W_i)}{\underset{i=1}{\overset{d}{\mathrm{\Σ}}}{W}_i}---\left(7\right)$

According to the dimension attribute evaluating matrix U that normalization obtains _k ^*, each qualitativing concept of descriptive grade cloud obtains degree of certainty matrix u _k ^*=(u _kij') _{x × y}, 1≤i≤y, 1≤j≤x;

Step 3.3, cloud clustering algorithm:

Different desired values can be synthesized a scheme value on each qualitativing concept of descriptive grade cloud by converging knot computing, and the degree of certainty of this scheme value on this qualitativing concept can be tried to achieve, the scheme value that final maximum degree of certainty is corresponding is the evaluation of estimate of the program, and the result of therefore converging knot computing last is the evaluation of estimate of each scheme on each grade cloud and the degree of certainty of evaluation of estimate ^[20];

By the matrix U after normalization _k ^*, obtain matrix S by formula (14) _k={ s _kp, p=1,2 ... x, so each scheme will have x scheme value, each scheme value correspond to the degree of certainty of a qualitativing concept of descriptive grade cloud;

$s_{\mathrm{kj}}=\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}(u_{\mathrm{ki}}*\×{u_{\mathrm{kij}}}^{\′})}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{u_{\mathrm{kij}}}^{\′}},j\∈R---\left(8\right)$

Contrast x scheme value of t scheme and the grade degree of certainty of correspondence, higher grade, and the scheme that degree of certainty is higher is optimum.

Therefore, tool of the present invention has the following advantages: 1. complete extra-high voltage drop point by cloud model comprehensively preferred, and according to ranking results, i.e. alternative drop point WH, the sequence of HS, XN is WH, XN, HS, and WH is higher than XN and HS grade, and recommending to choose WH is extra-high voltage drop point.In addition, with Field Using Fuzzy Comprehensive Assessment and gray incidence appraisal method checking, drawn same schemes ranking, therefore demonstrated the validity of the extra-high voltage drop point method for optimizing based on membership cloud models; 2. the use of cloud model, except carrying out comprehensive evaluation value sequence to scheme, can also intuitively reflect good and bad grade.Sample calculation analysis also demonstrates the use value that this evaluation assessment preferably goes up at extra-high voltage drop point.

Accompanying drawing explanation

Fig. 1 is the schematic flow sheet of assessment indicator system construction method in the present invention.

Fig. 2 is evaluation criterion weight determination schematic flow sheet in the present invention.

Fig. 3 is the comprehensive evaluation process flow diagram schematic diagram based on cloud model involved in the present invention.

Fig. 4 is descriptive grade cloud atlas involved in the present invention.

Fig. 5 is the schematic diagram of extra-high voltage drop point program evaluation index system I in embodiment.

Fig. 6 is the schematic diagram of extra-high voltage drop point program evaluation index system II in embodiment.

Embodiment

Below by embodiment, and by reference to the accompanying drawings, technical scheme of the present invention is described in further detail.

Embodiment:

One, first, concrete steps of the present invention are introduced:

(1) assessment indicator system is set up: traditional expert's subjectivity selects evaluation index method, stronger subjectivity may cause the index system set up not necessarily comprehensively and systematically can reflect decision problem, therefore introduce statistical analysis technique to analyze the validity of index system, reliability, the final assessment indicator system building the planning of extra-high voltage drop point.

A. the preliminary foundation of assessment indicator system.

Analyze extra-high voltage drop point program evaluation index and build principle and object, decision-maker tentatively chooses certain evaluation index.

B. the screening of evaluation index.

Expert adopts chromatographic assays agriculture products weight, adopts the screening of flexible strategy determining method to fall some secondary indexs.

If the assessment indicator system tentatively determined is H={h ₁, h ₂..., h _n.

If flexible strategy integrate as λ={ λ ₁, λ ₂..., λ _n, wherein λ _i∈ [0,1], (i=1,2 ..., n).If choice flexible strategy are λ _k, λ _k∈ [0,1], works as λ _i≤ λ _ktime, then index h is fallen in screening _i; Otherwise then retain this index.Usual choice flexible strategy λ _kit is 0.1 more suitable to get, and works as λ _ican think when≤0.1 that this Index Influence is less, be not enough to consider, decision-maker also can according to actual conditions, get get greatly little.After screening index, set up 2-3 initial indication system.

C. the Effective judgement of index system.

Adopt validity coefficient to portray the validity of index system, when measurement expert or decision-maker evaluate decision problem by a certain evaluation index or system, produce the departure degree of understanding.The absolute value of this value is less, and when showing that each expert or decision-maker adopt this evaluation index or system appraisal decision problem, more trend is consistent for understanding, and validity is higher.Otherwise validity is lower.

If the assessment indicator system after index screening is A={a ₁, a ₂..., a _n, expert's number of participating in evaluation is m, and the scoring of expert j to evaluation object integrates as X _j={ x _1j, x _2j..., x _nj, the validity coefficient of definition assessment indicator system A is β:

$\mathrm{\β}=\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\left(\frac{\underset{j=1}{\overset{m}{\mathrm{\Σ}}}|\stackrel{\‾}{x_i}-x_{\mathrm{ij}}|}{m*M}\right)}{n}---\left(1\right)$

Wherein, evaluation index a _ithe mean value of scoring, M is index a _icomment gathers in mark optimal value.

D. index system reliability judges.

The reliability and stability of index system represent the difference degree of the evaluating data that the index system of design draws and ideal data.According to mathematical statistics principle, adopt reliability coefficient ρ to portray this characteristic, ρ is larger, shows that difference is little, and the reliability of this index system is high; Otherwise its reliability is poor.

By evaluation index a _ithe average of m evaluation result as ideal value, calculate the difference degree of m evaluating data and ideal value.The average data group of expert group's scoring is Y={y ₁, y ₂..., y _n, the reliability coefficient of assessment indicator system is ρ:

$\mathrm{\ρ}=\frac{\underset{j=1}{\overset{m}{\mathrm{\Σ}}}(\underset{i=1}{\overset{n}{\mathrm{\Σ}}}(x_{\mathrm{ij}}-\stackrel{\‾}{x_j})(y_i-\stackrel{\‾}{y})/\sqrt{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{(x_{\mathrm{ij}}-\stackrel{\‾}{x_j})}^2\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{(y_i-\stackrel{\‾}{y})}^2})}{m}---\left(2\right)$

Wherein, y _ievaluation index a _ithe mean value of scoring, the scoring average of expert j, for y _iaverage.

Usually as ρ ∈ (0.90,0.95), can think that the reliability of this assessment indicator system is higher, when ρ ∈ (0.80,0.90), reliability is general, and when ρ ∈ (0,0.80), reliability is poor.

By above calculating, choose the index system that validity and reliability is all higher.

(2) the evaluation index enabling legislation based on maximum variance model is set up: combine with maximum variance model, this method not only considers the importance of index, also consider the difference effect of the same index of each scheme, regardless of the significance level of evaluation index itself, if all schemes are at index a _iunder property value difference larger, then illustrate this index to program decisions with sequence role larger, larger weight should be given.

A. the determination of subjective and objective weight.

Solve obtain subjective weight and objective weight respectively by the AHP that improves and entropy assessment, be different from traditional AHP, by analyzing each step of AHP and improving one's methods, have selected the higher scale of precision and complete weight and determine.Be different from traditional entropy assessment, fusion qualitative index and quantitative target form nondimensionalization matrix as the decision matrix solving entropy power.Finally solve and obtain subjective and objective weight.

B. based on the Evaluation formula of maximum variance model.

For a certain decision-making problem of multi-objective, decision scheme S _kat index a _ideviation that is lower and other all decision schemes may be defined as

${\mathrm{\σ}}_{\mathrm{ij}}\left(w\right)=\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{(r_{\mathrm{ij}}-r_{\mathrm{kj}})}^{2}w,i\∈n,j\∈m---\left(3\right)$

Based on this, structure departure function

$\mathrm{\σ}\left(w\right)=\underset{j=1}{\overset{m}{\mathrm{\Σ}}}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\σ}}_{\mathrm{ij}}\left(w\right)=\underset{j=1}{\overset{m}{\mathrm{\Σ}}}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{(r_{\mathrm{ij}}-r_{\mathrm{kj}})}^2{w}_j---\left(4\right)$

Thus solve weight phasor W to be equivalent to and to solve following linear programming problem:

$\left\{\begin{array}{c}\max \mathrm{\σ}\left(w\right)=\underset{j=1}{\overset{m}{\mathrm{\Σ}}}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{(r_{\mathrm{ij}}-r_{\mathrm{kj}})}^2{w}_j\\ s.t.w\∈\mathrm{\Φ}\end{array}\right.---\left(5\right)$

Wherein Weight of Expert and entropy are acted temporarily as proportion range in this model.

Solve above-mentioned linear programming model and can obtain optimized index weight phasor W ^*=[W ^* ₁, W ^* ₂w ^* _n].

(3) the preferred integrated evaluating method of extra-high voltage drop point based on cloud model is set up: after setting up assessment indicator system and enabling legislation, set up cloud model association attributes evaluation criteria level, descriptive grade cloud, utilize cloud model to carry out cloud normalization according to index attribute value and optimal weights and converge knot computing and complete the sequence of alternative drop point, then carry out drop point according to ordering scenario preferred.

A. set up mathematical description, comprise candidate scheme collection, index set, Zhuan Jiaji, attribute evaluation criteria collection and descriptive grade cloud;

B. cloud normalization computing

First, expert carries out scoring to qualitative index and forms matrix U _k', be an evaluation of estimate by multiple evaluation of estimate weighted mean normalizings of its each row, that is:

U _k′→Ex _i(6)

Merge quantitative and qualitative analysis index, and standardization obtains U _k"=(E _ki) _{y × 1}, according to U _k" and W ^*by formula (13), cloud normalization is carried out to the index of each scheme, obtain matrix

$u_{\mathrm{kd}}^{*}=\frac{\underset{i=1}{\overset{d}{\mathrm{\Σ}}}(E_{\mathrm{ki}}\×W_i)}{\underset{i=1}{\overset{d}{\mathrm{\Σ}}}{W}_i}---\left(7\right)$

According to the dimension attribute evaluating matrix U that normalization obtains _k ^*, each qualitativing concept of descriptive grade cloud obtains degree of certainty matrix u _k ^*=(u _kij') _{x × y}, 1≤i≤y, 1≤j≤x.

C. cloud clustering algorithm

Different desired values can be synthesized a scheme value on each qualitativing concept of descriptive grade cloud by converging knot computing, and the degree of certainty of this scheme value on this qualitativing concept can be tried to achieve, the scheme value that final maximum degree of certainty is corresponding is the evaluation of estimate of the program, and the result of therefore converging knot computing last is the evaluation of estimate of each scheme on each grade cloud and the degree of certainty of evaluation of estimate ^[20].

By the matrix U after normalization _k ^*, obtain matrix S by formula (14) _k={ s _kp, p=1,2 ... x, so each scheme will have x scheme value, each scheme value correspond to the degree of certainty of a qualitativing concept of descriptive grade cloud.

$s_{\mathrm{kj}}=\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}(u_{\mathrm{ki}}*\×{u_{\mathrm{kij}}}^{\′})}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{u_{\mathrm{kij}}}^{\′}},j\∈R---\left(8\right)$

Contrast x scheme value of t scheme and the grade degree of certainty of correspondence, higher grade, and the scheme that degree of certainty is higher is optimum.

Two, be the concrete case adopting said method below.

1. first confirm the following specifically describes:

1) candidate scheme integrates as S={s ₁, s ₂, s ₃;

2) index set A={a ₁, a ₂..., a ₆, a ₇..., a ₉, wherein a ₅, a ₆, a ₉for qualitative index, all the other are quantitative target;

3) expert collects E={e ₁, e ₂, e ₃;

4) attribute assessment collection D:D={d ₁, d ₂..., d ₅}={ VB, B, M, G, VG}

={ very poor, poor, generally, good, very good }={ 0.1,0.3,0.5,0.7,0.95}

5) according to 5 grades of attribute assessment collection, if descriptive grade cloud is expressed as follows:

$C_1=\left\{\begin{array}{cc}1& x\∈\left[\mathrm{0,0.1}\right]\\ C(\mathrm{0.1,0.2}/\mathrm{3,0.05}/9)& \mathrm{others}\end{array}\right.$

C ₂＝C(0.3，0.2/3，0.05/9)

C ₃＝C(0.5，0.2/3，0.05/9)

C ₄＝C(0.7，0.2/3，0.05/9)

$C_5=\left\{\begin{array}{cc}C(\mathrm{0.95,0.2}/\mathrm{3,0.05}/9)& \mathrm{others}\\ 1& x\∈[0,0.95]\end{array}\right.$

Descriptive grade cloud atlas is as Fig. 4:

2. set up assessment indicator system

According to and expert discussion, the basis of document [8] have selected preliminary index.Expert AHP is that each index carries out tax power, determines to accept or reject flexible strategy λ _kbe 0.1, carry out index screening, establish the extra-high voltage evaluation index system I, II that two covers are detailed, as Fig. 5, shown in 6.As space is limited, detailed calculation procedure is omitted.

3 experts give a mark to evaluation index I, II, then complete validity, the calculating of coefficient and reliability coefficient.Through calculating, draw:

β _Ⅰ＝0.165，β _Ⅱ＝0.202

The i.e. height of the effective sex ratio assessment indicator system II of assessment indicator system I.

ρ _Ⅰ＝0.808，ρ _Ⅱ＝0.447

Visible, the reliability of assessment indicator system I is than the height of assessment indicator system II.

Comprehensive above-mentioned calculating, the reliability of assessment indicator system I and validity, all higher than assessment indicator system II, therefore finally select assessment indicator system I as the assessment indicator system of this research extra-high voltage drop point planning.

3. the determination of evaluation criterion weight

1) determination of subjective weight

3 experts are also according to analytical hierarchy process when carrying out index screening, therefore can using the index system middle finger target flexible strategy of selection as subjective weight.

Therefore, final subjective weight is: W '=[0.3541,0.1643,0.0347,0.0736,0.0201,0.078,0.1816,0.0702,0.0234].

2) determination of objective weight

By 3 experts, 3 qualitative indexes are assessed, obtain attribute evaluating matrix, and be converted to quantitative values, specific as follows:

Average, obtain:

$U_{s_1}^{\′\′}=\begin{array}{cc}\begin{array}{c}{a}_5\\ a_6\\ a_9\end{array}& \left[\begin{array}{c}0.78\\ 0.63\\ 0.5\end{array}\right]\end{array},U_{s_2}^{\′\′}=\begin{array}{cc}\begin{array}{c}{a}_5\\ a_6\\ a_9\end{array}& \left[\begin{array}{c}0.7\\ 0.57\\ 0.43\end{array}\right]\end{array},U_{s_3}^{\′\′}=\begin{array}{cc}\begin{array}{c}{a}_5\\ a_6\\ a_9\end{array}& \left[\begin{array}{c}0.57\\ 0.37\\ 0.43\end{array}\right]\end{array}$

Then merge qualitative index and form dimensionless decision matrix:

Standardization, obtains:

Try to achieve entropy power, be final objective weight:

W″＝[0.0433,0.0460,0.0313,0.0382,0.0412,0.0336,0.2491,0.2278,0.2984]

3) combination weighting

MAT LAB program calculation obtains departure function:

σ (w)=0.0257w ₁+ 0.0559w ₂+ 0.0426w ₃+ 0.1251w ₄+ 0.3873w ₅+ 0.4467w ₆+ 0.4502w ₇+ 2.5216w ₈+ 0.0784w ₉try to achieve optimal weights according to maximum variance model, now single goal decision model is:

$\left\{\begin{array}{c}\max \mathrm{\σ}\left(w\right)=0.0257w_1+0.0559w_2+0.0426w_3+0.1251w_4\\ +0.3873w_5+0.4467w_6+0.4502w_7+2.5216w_8+0.0784w_9\\ s.t.0.0433\≤w_1\≤\mathrm{0.3541,0.046}\≤w_2\≤0.1643,\\ 0.0313\≤w_3\≤\mathrm{0.0347,0.0382}\≤w_4\≤0.0736,\\ 0.0201\≤w_5\≤\mathrm{0.0412.0.0336}\≤w_6\≤0.078,\\ 0.1816\≤w_7\≤\mathrm{0.2491,0.0702}\≤w_8\≤0.2278,\\ 0.0234\≤w_9\≤0.2984\end{array}\right.$

MAT LAB program calculation obtains optimal weights:

W ^*＝[0.0433,0.0460,0.0313,0.0736,0.0412,0.078,0.2491,0.2278,0.2097]

4. based on the comprehensive method for optimizing of cloud model

1) a dimension attribute evaluating matrix of 3 schemes is obtained after cloud normalization operation respectively and the degree of certainty matrix of element on descriptive grade cloud qualitativing concept is as follows:

$\begin{array}{ccc}{U_{s_1}}^{*}=\left[\begin{array}{c}1\\ 0.9340\\ 0.9511\\ 0.9696\\ 0.9749\\ 0.9812\\ 0.8356\\ 0.8830\\ 0.9075\end{array}\right]& {U_{s_2}}^{*}=\left[\begin{array}{c}0.9113\\ 0.9449\\ 0.9547\\ 0.9224\\ 0.9023\\ 0.8804\\ 0.7908\\ 0.6422\\ 0.6878\end{array}\right]& {U_{s_3}}^{*}=\left[\begin{array}{c}0.9325\\ 0.9673\\ 0.9470\\ 0.8908\\ 0.8471\\ 0.7891\\ 0.8825\\ 0.6718\\ 0.7112\end{array}\right]\end{array}$

2) after converging knot computing, the assembly value matrix of 3 schemes on descriptive grade cloud qualitativing concept and the degree of certainty matrix of this value as follows:

$\begin{array}{ccc}{S}_{s_1}=\begin{array}{cc}\left[\begin{array}{c}0\\ 0\\ 0\\ 0.8473\\ 0.9513\end{array}\right]& \begin{array}{c}{r}_1\\ r_2\\ r_3\\ r_4\\ r_5\end{array}\end{array}& S_{s_2}=\begin{array}{cc}\left[\begin{array}{c}0\\ 0\\ 0\\ 0.7690\\ 0.8602\end{array}\right]& \begin{array}{c}{r}_1\\ r_2\\ r_3\\ r_4\\ r_5\end{array}\end{array}& S_{s_3}=\begin{array}{cc}\left[\begin{array}{c}0\\ 0\\ 0.7112\\ 0.7433\\ 0.8815\end{array}\right]& \begin{array}{c}{r}_1\\ r_2\\ r_3\\ r_4\\ r_5\end{array}\end{array}\end{array}$

$\begin{array}{ccc}{u}_{s_1}=\begin{array}{cc}\left[\begin{array}{c}0\\ 0\\ 0\\ 0.0871\\ 0.9998\end{array}\right]& \begin{array}{c}{r}_1\\ r_2\\ r_3\\ r_4\\ r_5\end{array}\end{array}& u_{s_2}=\begin{array}{cc}\left[\begin{array}{c}0\\ 0\\ 0\\ 0.5853\\ 0.4037\end{array}\right]& \begin{array}{c}{r}_1\\ r_2\\ r_3\\ r_4\\ r_5\end{array}\end{array}& u_{s_3}=\begin{array}{cc}\left[\begin{array}{c}0\\ 0\\ 0.0066\\ 0.8098\\ 0.5899\end{array}\right]& \begin{array}{c}{r}_1\\ r_2\\ r_3\\ r_4\\ r_5\end{array}\end{array}\end{array}$

As can be seen here,

Table 1 extra-high voltage drop point optimization selection evaluation result

Specific embodiment described herein is only to the explanation for example of the present invention's spirit.Those skilled in the art can make various amendment or supplement or adopt similar mode to substitute to described specific embodiment, but can't depart from spirit of the present invention or surmount the scope that appended claims defines.

Claims (1)

1., based on an extra-high voltage drop point planning method for optimizing for cloud model, it is characterized in that,

Step 1, sets up the extra-high voltage drop point assessment indicator system about security, drop point region characteristic and prospect zone; Specifically: analyze extra-high voltage drop point program evaluation index and build principle and object, decision-maker tentatively chooses certain evaluation index; Adopt chromatographic assays agriculture products weight, adopt the screening of flexible strategy determining method to fall some secondary indexs; Adopt validity coefficient to portray the validity of index system, when weighing evaluate decision problem by the evaluation index set or system, produce the departure degree of understanding; The reliability and stability of index system represent the difference degree of the evaluating data that the index system of design draws and ideal data; Finally choose the index system that validity and reliability is all higher; Comprise following sub-step:

Step 1.1, the preliminary foundation of assessment indicator system:

Analyze extra-high voltage drop point program evaluation index and build principle and object, tentatively choose certain evaluation index;

Step 1.2, the screening of evaluation index:

Adopt chromatographic assays agriculture products weight, adopt the screening of flexible strategy determining method to fall some secondary indexs;

Defining the assessment indicator system tentatively determined is H={h ₁, h ₂..., h _n.

Definition flexible strategy integrate as λ={ λ ₁, λ ₂..., λ _n, wherein λ _i∈ [0,1], (i=1,2 ..., n); If choice flexible strategy are λ _k, λ _k∈ [0,1], works as λ _i≤ λ _ktime, then index h is fallen in screening _i; Otherwise then retain this index; Usual choice flexible strategy λ _kit is 0.1 more suitable to get, and works as λ _ican think when≤0.1 that this Index Influence is less, be not enough to consider, decision-maker also can according to actual conditions, get get greatly little; After screening index, set up 2-3 initial indication system;

Step 1.3, the Effective judgement of index system:

Adopt validity coefficient to portray the validity of index system, when weighing evaluate decision problem by the evaluation index set or system, produce the departure degree of understanding; The absolute value of this value is less, and when showing to adopt this evaluation index or system appraisal decision problem, more trend is consistent for understanding, and validity is higher; Otherwise validity is lower;

Assessment indicator system after definition index screening is A={a ₁, a ₂..., a _n, participating in pricer number is m, and it is X that expert j assembles fruit to the scoring of evaluation object _j={ x _1j, x _2j..., x _nj, the validity coefficient of definition assessment indicator system A is β:

$\mathrm{\β}=\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\left(\frac{\underset{j=1}{\overset{m}{\mathrm{\Σ}}}|\stackrel{\‾}{x_i}-x_{\mathrm{ij}}|}{m*M}\right)}{n}---\left(1\right)$

Wherein, evaluation index a _ithe mean value of scoring, M is index a _icomment gathers in mark optimal value;

Step 1.4, index system reliability judges

The reliability and stability of index system represent the difference degree of the evaluating data that the index system of design draws and ideal data; According to mathematical statistics principle, adopt reliability coefficient ρ to portray this characteristic, ρ is larger, shows that difference is little, and the reliability of this index system is high; Otherwise its reliability is poor; Specifically by evaluation index a _ithe average of m evaluation result as ideal value, calculate the difference degree of m evaluating data and ideal value; The average data group of expert group's scoring is Y={y ₁, y ₂..., y _n, the reliability coefficient of assessment indicator system is ρ:

$\mathrm{\ρ}=\frac{\underset{j=1}{\overset{m}{\mathrm{\Σ}}}(\underset{i=1}{\overset{n}{\mathrm{\Σ}}}(x_{\mathrm{ij}}-\stackrel{\‾}{x_j})(y_i-\stackrel{\‾}{y})/\sqrt{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{(x_{\mathrm{ij}}-\stackrel{\‾}{x_j})}^2\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{(y_i-\stackrel{\‾}{y})}^2}}{m}---\left(2\right)$

Wherein, y _ievaluation index a _ithe mean value of scoring, the scoring average of expert j, for y _iaverage;

By above calculating, choose the index system that validity and reliability is all higher;

Step 2, sets up the evaluation index enabling legislation based on maximum variance model; Specifically: for set up extra-high voltage drop point program evaluation index system, AHP method is utilized to solve subjective weight according to an expert view, utilize entropy assessment to solve objective weight, combine with maximum variance model, try to achieve the optimal weights of extra-high voltage drop point evaluation index;

Step 2.1, the determination of subjective and objective weight:

Solve obtain subjective weight and objective weight respectively by the AHP that improves and entropy assessment, be different from traditional AHP, by analyzing each step of AHP and improving one's methods, have selected the higher scale of precision and complete weight and determine; Be different from traditional entropy assessment, fusion qualitative index and quantitative target form nondimensionalization matrix as the decision matrix solving entropy power; Finally solve and obtain subjective and objective weight;

Step 2.2, based on the Evaluation formula of maximum variance model

For a certain decision-making problem of multi-objective, decision scheme S _kat index a _ideviation that is lower and other all decision schemes may be defined as

${\mathrm{\σ}}_{\mathrm{ij}}\left(w\right)=\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{(r_{\mathrm{ij}}-r_{\mathrm{kj}})}^2{w}_j,i\∈n,j\∈m---\left(3\right)$

Based on this, structure departure function

$\mathrm{\σ}\left(w\right)=\underset{j=1}{\overset{m}{\mathrm{\Σ}}}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\σ}}_{\mathrm{ij}}\left(w\right)=\underset{j=1}{\overset{m}{\mathrm{\Σ}}}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{(r_{\mathrm{ij}}-r_{\mathrm{kj}})}^2{w}_j---\left(4\right)$

Thus solve weight phasor W to be equivalent to and to solve following linear programming problem:

$\left\{\begin{array}{c}\max \mathrm{\σ}\left(w\right)=\underset{j=1}{\overset{m}{\mathrm{\Σ}}}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{(r_{\mathrm{ij}}-r_{\mathrm{kj}})}^2{w}_j\\ s.t.w\∈\mathrm{\Φ}\end{array}\right.---\left(5\right)$

Wherein Weight of Expert and entropy are acted temporarily as into proportion range in this model;

Solve above-mentioned linear programming model and can obtain optimized index weight phasor W ^*=[W ^* ₁, W ^* ₂w ^* _n];

Step 3, sets up the preferred integrated evaluating method of extra-high voltage drop point based on cloud model; Specifically: after setting up assessment indicator system and enabling legislation, set up cloud model association attributes evaluation criteria level, descriptive grade cloud, utilize cloud model to carry out cloud normalization according to index attribute value and optimal weights and converge knot computing and complete the sequence of alternative drop point, then carry out drop point according to ordering scenario preferred;

Step 3.1, sets up mathematical description, comprises candidate scheme collection, index set, Zhuan Jiaji, attribute evaluation criteria collection and descriptive grade cloud;

Step 3.2, cloud normalization computing:

First, expert is formed matrix U to the result that qualitative index is marked _k', be an evaluation of estimate by multiple evaluation of estimate weighted mean normalizings of its each row, that is:

U _k′→Ex _i(6)

Merge quantitative and qualitative analysis index, and standardization obtains U _k"=(E _ki) _{y × 1}, according to U _k" and W ^*by formula (13), cloud normalization is carried out to the index of each scheme, obtain matrix

$u_{\mathrm{kd}}^{*}=\frac{\underset{i=1}{\overset{d}{\mathrm{\Σ}}}(E_{\mathrm{ki}}\×W_i)}{\underset{i=1}{\overset{d}{\mathrm{\Σ}}}{W}_i}---\left(7\right)$

According to the dimension attribute evaluating matrix U that normalization obtains _k ^*, each qualitativing concept of descriptive grade cloud obtains degree of certainty matrix u _k ^*=(u _kij') _{x × y}, 1≤i≤y, 1≤j≤x;

Step 3.3, cloud clustering algorithm:

Different desired values can be synthesized a scheme value on each qualitativing concept of descriptive grade cloud by converging knot computing, and the degree of certainty of this scheme value on this qualitativing concept can be tried to achieve, the scheme value that final maximum degree of certainty is corresponding is the evaluation of estimate of the program, and the result of therefore converging knot computing last is the evaluation of estimate of each scheme on each grade cloud and the degree of certainty of evaluation of estimate ^[20];

By the matrix U after normalization _k ^*, obtain matrix S by formula (14) _k={ s _kp, p=1,2 ... x, so each scheme will have x scheme value, each scheme value correspond to the degree of certainty of a qualitativing concept of descriptive grade cloud;

$s_{\mathrm{kj}}=\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}(u_{\mathrm{ki}}*\×{u_{\mathrm{kij}}}^{\′})}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{u_{\mathrm{kij}}}^{\′}},j\∈R---\left(8\right)$

Contrast x scheme value of t scheme and the grade degree of certainty of correspondence, higher grade, and the scheme that degree of certainty is higher is optimum.