CN105608542A - Multi-level fuzzy comprehensive evaluation method for electric power engineering project - Google Patents

Abstract

A multi-level fuzzy comprehensive evaluation method suitable for electric power engineering projects, including steps: S1, determine the evaluation object; S2, determine the evaluation factor set U; S3, determine the domain V of the comment level; S4, establish the membership function for single factor evaluation ; S5, establishing a weight distribution vector of evaluation factors; S6, obtaining a comprehensive evaluation result through compound operations; S7, analyzing the comprehensive evaluation vector. It avoids the disadvantage of assigning weights when there are too many factors, is more refined than the single-level model, and more correctly reflects the relationship between factors.

Description

A multi-level fuzzy comprehensive evaluation method suitable for electric power engineering projects

technical field

The invention relates to the field of electric power engineering evaluation, in particular to a multi-level fuzzy comprehensive evaluation method suitable for electric power engineering projects.

Background technique

In the field of power system, a considerable number of projects are evaluated every year. Compared with other comprehensive evaluation methods, fuzzy comprehensive evaluation is an absolute evaluation method, and it is also a value ranking and value classification evaluation method, so the evaluation has direct physical meaning . The evaluation result of a unit has nothing to do with the composition of participating units, which is superior to multivariate statistical evaluation. At the same time, the output results of fuzzy comprehensive evaluation of multiple comment levels usually contain a membership vector, so the content of evaluation information is relatively richer than other evaluation methods.

However, the existing fuzzy comprehensive evaluation has many disadvantages due to the uncertainty of evaluation indicators:

First, the fuzzy comprehensive evaluation is not superior to the utility function:

First, the classification of comment levels increases the risk of distorting the comprehensive evaluation conclusions. The number of ratings and the quantification method of ratings will affect the evaluation conclusions, and it is theoretically difficult to provide evaluators with basic rules that are "clear at a glance". The point value index used in the comprehensive evaluation of the quantitative calculation and ranking of comments is also inferior to the utility function method in terms of accuracy.

Second, after the ratings are divided, the membership function needs to be determined for each rating, and the determination of the membership function obviously has a great impact on the comprehensive evaluation conclusion. The utility function method only needs to determine a utility function for each indicator, while the fuzzy evaluation not only needs to determine multiple membership functions, but also needs to correctly grasp the connection between the membership functions of each comment level, and it needs to be processed in sections. It is a difficult thing for many users.

Thirdly, although fuzzy synthesis B is rich in evaluation information, it is actually not comprehensive, because B is formed by weighting the membership degree of the same comment level by each index. If there is no index in an index system to a certain If there is a subordination relationship in a comment level, the membership degree of this level must be zero after synthesis, so different comment levels cannot complement each other, so the information of B is divergent rather than comprehensive.

Secondly, in terms of classification evaluation, fuzzy comprehensive evaluation is not necessarily superior to clustering and discriminant analysis in multivariate statistics.

Contents of the invention

In order to solve the above technical problems, the present invention proposes a multi-level fuzzy comprehensive evaluation method suitable for power engineering projects, which avoids the disadvantage of assigning weights when there are too many factors, is more refined than the single-level model, and more accurately reflects the direct relationship between factors. interrelationships.

In order to achieve the above object, the scheme adopted by the present invention is:

A multi-level fuzzy comprehensive evaluation method suitable for power engineering projects, including steps:

S1, determine the evaluation object;

S2, determine the evaluation factor set U;

S3, determine the comment level discourse domain V;

S4, establishing a membership function for single-factor evaluation;

S5, establishing a weight distribution vector of evaluation factors;

S6, obtaining a comprehensive evaluation result through a compound operation;

S7, analyzing the comprehensive evaluation vector.

Described step S2 comprises the steps:

S21, list all factors of the evaluation object, the factors refer to various attributes or performances of the evaluation object;

S22. Find out factors related to the evaluation from the evaluation objects.

The hierarchical discourse domain V=(very good, good, average, bad, very poor) described in step S3.

Described step S4 comprises the steps:

S41, the evaluation factor set U is composed of several single factors u;

S42. Determine the degree of membership to the decision-making level v from the perspective of a single factor u, and the decision-making level v is a single level element in the comment universe V;

S43. Determine the evaluation matrix R, the evaluation matrix R is a fuzzy set from the evaluation factor set U to the comment domain V, and the specific elements in the evaluation matrix R are the membership degrees of the single factor u to the decision level v.

Described step S5 comprises the steps:

S51. Determine the importance degree fuzzy subset A of the evaluation factor set U, where the specific element a in A represents the measure of the degree of influence of single factor u on the overall evaluation.

The method for determining the weight in the step S5 is the method of tel illegality, factor paired comparison method, analytic hierarchy process or mathematical statistics method.

Described step S6 comprises the steps:

S61, by fuzzy transformation B=A*R, where * is an M(·,+) operator.

Described step S7 comprises the steps:

S71. Using the principle of maximum membership degree to process the comprehensive evaluation result.

The beneficial effect of the present invention is that the multi-level fuzzy comprehensive evaluation model can reflect the hierarchy of each factor of the evaluation object, and at the same time avoid the disadvantage of assigning weights when there are too many factors, and it is more refined than the single-level model, and more accurately reflects the direct relationship between the factors. interrelationships.

Description of drawings

Fig. 1 is a schematic flow chart of the present invention.

detailed description

In order to better understand the technical solution of the present invention, the present invention will be further described below in conjunction with the accompanying drawings.

As shown in Figure 1, a multi-level fuzzy comprehensive evaluation method suitable for power engineering projects, including steps:

1) Determination of the evaluation object. For the 2014 National Science and Technology Progress Award for the Power Civil Nuclear Power Group, our evaluation objects are the 14 projects shown in Table 1 in the text, ranging from "Key Technology Research and Engineering Application of Carbon Fiber Composite Core Conductors" to "±800kV UHV Innovative development and world-first application of direct current transmission technology and equipment", which will not be listed here.

2) Determine the evaluation factor set U, namely the index system. Factors refer to various attributes or performances of the evaluated object, and to find out the factor set means to propose factors related to the evaluation from the evaluated object. For the domain of factors to be evaluated, see Table 1 in the text. There are a total of 14 indicators ranging from "number of invention patents" to "degree of fit between project and planning".

3) Determining the comment level Domain V, also called the evaluation set or level judgment set, refers to the set of comments made on the evaluation object. Each level may correspond to a fuzzy subset. V=(very good, good, fair, not good, very poor).

4) Establish membership function for single factor evaluation. A single factor evaluation is performed on the single factor u in the evaluation factor set U. From the point of view of single factor u, the degree of membership to decision-making level v is determined, and the result of each single factor evaluation is obtained. Find out the evaluation matrix R. R is a fuzzy set from evaluation factor domain U to decision comment domain V. The specific elements in it represent the membership degree of factor u to decision level v.

Membership function is an index to characterize the fuzziness of factors, a mathematical index to express the uncertainty of the membership relationship of elements to fuzzy sets, and the cornerstone for fuzzy sets to be applied to practical problems. When using the fuzzy comprehensive evaluation method for actual project evaluation, the membership function must be determined first. The establishment of the membership function is objective in nature, but there is no universally applicable method in theory. Its determination is generally determined approximately by reasoning. It can be given by fuzzy statistical experiments or typical function methods summarized from actual experience. out. The common methods to determine the membership function are: fuzzy statistical method test, binary comparison sorting method and stepwise estimation method. The determination of the specific typical membership function is not introduced here, if necessary, you can Baidu yourself. Combined with the characteristics of the actual project, this paper chooses a relatively simple reduced-half trapezoidal distribution function to determine the degree of membership.

5) Establish the weight distribution vector of evaluation factors. Determine the degree of influence of single factor u in the total evaluation. The importance of each factor in the overall evaluation is different, so the focus of the evaluation should be on the fuzzy subset A on the domain U of the evaluation factor. The specific element a in A represents the measure of the influence of a single factor u on the overall evaluation. A is called is the importance degree fuzzy subset of U.

The established indicator system includes 14 indicators, not all of which are equally important to the evaluation objectives of the project. The weight of an indicator refers to the degree of importance attached to the indicator when making a comprehensive evaluation. Common methods for determining weights include: Delphi method (expert scoring method), factor pairwise comparison method, analytic hierarchy process, mathematical statistics method, etc. Due to the limited data obtained, this paper chooses the expert scoring method.

6) Perform compound operations to obtain comprehensive evaluation results. Carry out fuzzy transformation B=A*R, * represents an operator. Different models have different operators, which belong to the basic knowledge of fuzzy mathematics and will not be described in detail. The M(·, +) operator is used here.

7) Analyze the comprehensive evaluation result vector. The evaluation result obtained by using the fuzzy comprehensive evaluation method is the membership degree of each evaluated item to each grade of fuzzy subset, which constitutes a fuzzy vector. Since we want to conduct comprehensive evaluation on 14 items, we need to further process the evaluation result vector. The commonly used processing methods are: the principle of maximum membership degree, the principle of weighted average, and the single value of fuzzy vector. The above three methods can be selected according to the purpose of evaluation, and this article adopts the principle of maximum membership degree. Assuming that the fuzzy comprehensive evaluation result vector is B, and b(r)=b(max) is the largest element in B, then the evaluated items generally belong to the rth grade.

In the specific implementation process, the steps are:

1) According to the specific power engineering review items, give the evaluated object set X={x ₁ ,x ₂ ,…,x _k }

2) Determine the domain of discourse U of the factors to be judged, U={u ₁ ,u ₂ ,…,u _n }

If there are many factors, U={u ₁ ,u ₂ ,…,u _n } is often divided into s subsets according to certain attributes, And meet the conditions:

U _i ∩ U _j = φ,i≠j

3), determine the comment set V={v ₁ ,v ₂ ,…,v _m }

4), from the factor set U _i and the comment set V, an evaluation matrix can be obtained

${\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\left[\begin{array}{cccc}{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 11</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}& {\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 12</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}& \mathrm{<span\; class="notranslate">\; ...</span>}& {\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; m</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}\\ {\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; one</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}& {\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; two</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}& \mathrm{<span\; class="notranslate">\; ...</span>}& {\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; m</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}\\ <span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span>\\ <span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span>\\ <span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span>& <span\; class="notranslate">\; .</span>\\ {\mathrm{<span\; class="notranslate">\; r</span>}}_{{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}& {\mathrm{<span\; class="notranslate">\; r</span>}}_{{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}& \mathrm{<span\; class="notranslate">\; ...</span>}& {\mathrm{<span\; class="notranslate">\; r</span>}}_{{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}\mathrm{<span\; class="notranslate">\; m</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span>}\end{array}\right]$

5) For each U _i , make a comprehensive decision respectively. Let the distribution of the weight of each factor in U _i (called fuzzy weight vector) be in

If R _i is a single factor matrix, the first-level evaluation vector is obtained as

6), regard each U _i as a factor, record U={U ₁ ,U ₂ ,…,U _s }, then U is a single factor set, and the single factor judgment matrix of U is As a part of U, each U _i reflects a certain attribute of U, and the weight distribution A={a ₁ ,a ₂ , _… ,as } can be given according to their importance, so a two-level fuzzy comprehensive evaluation model can be obtained for

7) If each sub-factor U _i , i=1, 2,..., s still has more factors, then U _i can be subdivided, so there is a three-level or higher-level model.

The multi-level fuzzy comprehensive evaluation model can reflect the hierarchy of each factor of the evaluation object, and at the same time avoid the disadvantage of assigning weights when there are too many factors. It is more precise than the single-level model and more correctly reflects the relationship between factors.

Although the specific implementation of the present invention has been described above in conjunction with the accompanying drawings, it does not limit the protection scope of the present invention. Those skilled in the art should understand that on the basis of the technical solution of the present invention, those skilled in the art do not need to pay creative work Various modifications or variations that can be made are still within the protection scope of the present invention.

Claims (8)

1. A multi-level fuzzy comprehensive evaluation method applicable to electric power engineering projects is characterized in that it comprises steps: S1, determine the evaluation object; S2, determine the evaluation factor set U; S3, determine the comment level discourse domain V; S4, establishing a membership function for single-factor evaluation; S5, establishing a weight distribution vector of evaluation factors; S6, obtaining a comprehensive evaluation result through a compound operation; S7, analyzing the comprehensive evaluation vector. 2. A kind of multi-level fuzzy comprehensive evaluation method applicable to power engineering projects as claimed in claim 1, wherein said step S2 comprises the steps of: S21, list all factors of the evaluation object, the factors refer to various attributes or performances of the evaluation object; S22. Find out factors related to the evaluation from the evaluation objects. 3. a kind of multi-level fuzzy comprehensive evaluation method applicable to electric power engineering projects as claimed in claim 1 or 2, is characterized in that, the domain of discourse V=(very good, good, general, no good, bad). 4. A kind of multi-level fuzzy comprehensive evaluation method applicable to power engineering projects as claimed in claim 3, characterized in that, said step S4 comprises the steps of: S41, the evaluation factor set U is composed of several single factors u; S42. Determine the degree of membership to the decision-making level v from the perspective of a single factor u, and the decision-making level v is a single level element in the comment domain V; S43. Determine the evaluation matrix R, the evaluation matrix R is a fuzzy set from the evaluation factor set U to the comment domain V, and the specific elements in the evaluation matrix R are the membership degrees of the single factor u to the decision level v. 5. A kind of multi-level fuzzy comprehensive evaluation method applicable to power engineering projects as claimed in claim 3, characterized in that, said step S5 comprises the steps of: S51. Determine the importance degree fuzzy subset A of the evaluation factor set U, where the specific element a in A represents the measure of the degree of influence of single factor u on the overall evaluation. 6. A kind of multi-level fuzzy comprehensive evaluation method applicable to electric power engineering projects as claimed in claim 3, is characterized in that, the method for determining weight in the described step S5 is special law, factor paired comparison method, hierarchical analysis or mathematical statistics. 7. A kind of multi-level fuzzy comprehensive evaluation method applicable to power engineering projects as claimed in claim 5, characterized in that, said step S6 comprises the steps of: S61, by fuzzy transformation B=A*R, where * is an M(·,+) operator. 8. A kind of multi-level fuzzy comprehensive evaluation method applicable to power engineering projects as claimed in claim 5, characterized in that, said step S7 comprises the steps of: S71. Using the principle of maximum membership degree to process the comprehensive evaluation result.